From f6e9456c9272bb570df6e217cdbe007e270b1c4e Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 21 Jul 2010 19:03:58 +0200 Subject: x86, cleanup: Remove obsolete boot_cpu_id variable boot_cpu_id is there for historical reasons and was renamed to boot_cpu_physical_apicid in patch: c70dcb7 x86: change boot_cpu_id to boot_cpu_physical_apicid However, there are some remaining occurrences of boot_cpu_id that are never touched in the kernel and thus its value is always 0. This patch removes boot_cpu_id completely. Signed-off-by: Robert Richter LKML-Reference: <1279731838-1522-8-git-send-email-robert.richter@amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/apb_timer.h b/arch/x86/include/asm/apb_timer.h index a69b1ac..2fefa50 100644 --- a/arch/x86/include/asm/apb_timer.h +++ b/arch/x86/include/asm/apb_timer.h @@ -54,7 +54,6 @@ extern struct clock_event_device *global_clock_event; extern unsigned long apbt_quick_calibrate(void); extern int arch_setup_apbt_irqs(int irq, int trigger, int mask, int cpu); extern void apbt_setup_secondary_clock(void); -extern unsigned int boot_cpu_id; extern struct sfi_timer_table_entry *sfi_get_mtmr(int hint); extern void sfi_free_mtmr(struct sfi_timer_table_entry *mtmr); diff --git a/arch/x86/include/asm/cpu.h b/arch/x86/include/asm/cpu.h index b185091..4fab24d 100644 --- a/arch/x86/include/asm/cpu.h +++ b/arch/x86/include/asm/cpu.h @@ -32,6 +32,5 @@ extern void arch_unregister_cpu(int); DECLARE_PER_CPU(int, cpu_state); -extern unsigned int boot_cpu_id; #endif /* _ASM_X86_CPU_H */ diff --git a/arch/x86/kernel/apb_timer.c b/arch/x86/kernel/apb_timer.c index 8dd7780..08f75fb 100644 --- a/arch/x86/kernel/apb_timer.c +++ b/arch/x86/kernel/apb_timer.c @@ -343,7 +343,7 @@ void apbt_setup_secondary_clock(void) /* Don't register boot CPU clockevent */ cpu = smp_processor_id(); - if (cpu == boot_cpu_id) + if (!cpu) return; /* * We need to calculate the scaled math multiplication factor for diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 4dc0084..8884928 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -162,7 +162,7 @@ int __init arch_early_irq_init(void) cfg = irq_cfgx; count = ARRAY_SIZE(irq_cfgx); - node= cpu_to_node(boot_cpu_id); + node = cpu_to_node(0); for (i = 0; i < count; i++) { desc = irq_to_desc(i); @@ -1483,7 +1483,7 @@ static void __init setup_IO_APIC_irqs(void) int notcon = 0; struct irq_desc *desc; struct irq_cfg *cfg; - int node = cpu_to_node(boot_cpu_id); + int node = cpu_to_node(0); apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); @@ -1548,7 +1548,7 @@ static void __init setup_IO_APIC_irqs(void) void setup_IO_APIC_irq_extra(u32 gsi) { int apic_id = 0, pin, idx, irq; - int node = cpu_to_node(boot_cpu_id); + int node = cpu_to_node(0); struct irq_desc *desc; struct irq_cfg *cfg; @@ -2925,7 +2925,7 @@ static inline void __init check_timer(void) { struct irq_desc *desc = irq_to_desc(0); struct irq_cfg *cfg = desc->chip_data; - int node = cpu_to_node(boot_cpu_id); + int node = cpu_to_node(0); int apic1, pin1, apic2, pin2; unsigned long flags; int no_pin1 = 0; @@ -3279,7 +3279,7 @@ unsigned int create_irq_nr(unsigned int irq_want, int node) int create_irq(void) { - int node = cpu_to_node(boot_cpu_id); + int node = cpu_to_node(0); unsigned int irq_want; int irq; @@ -3901,7 +3901,7 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq, if (dev) node = dev_to_node(dev); else - node = cpu_to_node(boot_cpu_id); + node = cpu_to_node(0); desc = irq_to_desc_alloc_node(irq, node); if (!desc) { diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 60a57b1..3a7c852f 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -148,7 +148,7 @@ static void __cpuinit amd_k7_smp_check(struct cpuinfo_x86 *c) { #ifdef CONFIG_SMP /* calling is from identify_secondary_cpu() ? */ - if (c->cpu_index == boot_cpu_id) + if (!c->cpu_index) return; /* diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 490dac6..787b3c7 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -665,7 +665,7 @@ static void __init early_identify_cpu(struct cpuinfo_x86 *c) this_cpu->c_early_init(c); #ifdef CONFIG_SMP - c->cpu_index = boot_cpu_id; + c->cpu_index = 0; #endif filter_cpuid_features(c, false); } diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index 85f69cd..3a683ea 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -169,7 +169,7 @@ static void __cpuinit intel_smp_check(struct cpuinfo_x86 *c) { #ifdef CONFIG_SMP /* calling is from identify_secondary_cpu() ? */ - if (c->cpu_index == boot_cpu_id) + if (!c->cpu_index) return; /* diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index e3af342..7a4cf14 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -84,7 +84,7 @@ static int __init reboot_setup(char *str) } /* we will leave sorting out the final value when we are ready to reboot, since we might not - have set up boot_cpu_id or smp_num_cpu */ + have detected BSP APIC ID or smp_num_cpu */ break; #endif /* CONFIG_SMP */ diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index b008e78..dede5c4 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -125,7 +125,6 @@ unsigned long max_pfn_mapped; RESERVE_BRK(dmi_alloc, 65536); #endif -unsigned int boot_cpu_id __read_mostly; static __initdata unsigned long _brk_start = (unsigned long)__brk_base; unsigned long _brk_end = (unsigned long)__brk_base; diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index a60df9a..2335c15 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -253,7 +253,7 @@ void __init setup_per_cpu_areas(void) * Up to this point, the boot CPU has been using .init.data * area. Reload any changed state for the boot CPU. */ - if (cpu == boot_cpu_id) + if (!cpu) switch_to_new_gdt(cpu); } diff --git a/arch/x86/mm/k8topology_64.c b/arch/x86/mm/k8topology_64.c index 970ed57..240f864 100644 --- a/arch/x86/mm/k8topology_64.c +++ b/arch/x86/mm/k8topology_64.c @@ -54,8 +54,8 @@ static __init int find_northbridge(void) static __init void early_get_boot_cpu_id(void) { /* - * need to get boot_cpu_id so can use that to create apicid_to_node - * in k8_scan_nodes() + * need to get the APIC ID of the BSP so can use that to + * create apicid_to_node in k8_scan_nodes() */ #ifdef CONFIG_X86_MPPARSE /* @@ -212,7 +212,7 @@ int __init k8_scan_nodes(void) bits = boot_cpu_data.x86_coreid_bits; cores = (1< 0) { pr_info("BSP APIC ID: %02x\n", boot_cpu_physical_apicid); -- cgit v0.10.2 From 1f999ab5a1360afc388868cc0ef9afe8edeef3be Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 21 Jul 2010 19:03:57 +0200 Subject: x86, xsave: Disable xsave in i387 emulation mode xsave is broken for (!HAVE_HWFP). This is the case if config MATH_EMULATION is enabled, 'no387' kernel parameter is set and xsave exists. xsave will not work because x86/math-emu and xsave share the same memory. As this case can be treated as corner case we simply disable xsave then. Signed-off-by: Robert Richter LKML-Reference: <1279731838-1522-7-git-send-email-robert.richter@amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index 1f11f5c..2605c50 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -67,6 +67,12 @@ static void __cpuinit init_thread_xstate(void) */ if (!HAVE_HWFP) { + /* + * Disable xsave as we do not support it if i387 + * emulation is enabled. + */ + setup_clear_cpu_cap(X86_FEATURE_XSAVE); + setup_clear_cpu_cap(X86_FEATURE_XSAVEOPT); xstate_size = sizeof(struct i387_soft_struct); return; } -- cgit v0.10.2 From f1b499f029c5dde85d46a8811353c62f29157541 Mon Sep 17 00:00:00 2001 From: John Kacur Date: Thu, 5 Aug 2010 17:10:53 +0200 Subject: lockdep: Remove __debug_show_held_locks There is no longer any functional difference between __debug_show_held_locks() and debug_show_held_locks(), so remove the former. Signed-off-by: John Kacur Cc: Peter Zijlstra LKML-Reference: <1281021054-4228-1-git-send-email-jkacur@redhat.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/debug_locks.h b/include/linux/debug_locks.h index 29b3ce3..2833452 100644 --- a/include/linux/debug_locks.h +++ b/include/linux/debug_locks.h @@ -49,7 +49,6 @@ struct task_struct; #ifdef CONFIG_LOCKDEP extern void debug_show_all_locks(void); -extern void __debug_show_held_locks(struct task_struct *task); extern void debug_show_held_locks(struct task_struct *task); extern void debug_check_no_locks_freed(const void *from, unsigned long len); extern void debug_check_no_locks_held(struct task_struct *task); @@ -58,10 +57,6 @@ static inline void debug_show_all_locks(void) { } -static inline void __debug_show_held_locks(struct task_struct *task) -{ -} - static inline void debug_show_held_locks(struct task_struct *task) { } diff --git a/kernel/hung_task.c b/kernel/hung_task.c index 0c642d5..bca9423 100644 --- a/kernel/hung_task.c +++ b/kernel/hung_task.c @@ -98,7 +98,7 @@ static void check_hung_task(struct task_struct *t, unsigned long timeout) printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\"" " disables this message.\n"); sched_show_task(t); - __debug_show_held_locks(t); + debug_show_held_locks(t); touch_nmi_watchdog(); diff --git a/kernel/lockdep.c b/kernel/lockdep.c index f2852a5..84baa71 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3775,7 +3775,7 @@ EXPORT_SYMBOL_GPL(debug_show_all_locks); * Careful: only use this function if you are sure that * the task cannot run in parallel! */ -void __debug_show_held_locks(struct task_struct *task) +void debug_show_held_locks(struct task_struct *task) { if (unlikely(!debug_locks)) { printk("INFO: lockdep is turned off.\n"); @@ -3783,12 +3783,6 @@ void __debug_show_held_locks(struct task_struct *task) } lockdep_print_held_locks(task); } -EXPORT_SYMBOL_GPL(__debug_show_held_locks); - -void debug_show_held_locks(struct task_struct *task) -{ - __debug_show_held_locks(task); -} EXPORT_SYMBOL_GPL(debug_show_held_locks); void lockdep_sys_exit(void) -- cgit v0.10.2 From 6a103b0d44e9f97dc430002cf3ac7a7defa3819f Mon Sep 17 00:00:00 2001 From: John Kacur Date: Thu, 5 Aug 2010 17:10:54 +0200 Subject: lockup detector: Fix grammar by adding a missing "to" in the comments This fixes a minor grammar problem in the comments in hung_task.c Signed-off-by: John Kacur Cc: Peter Zijlstra LKML-Reference: <1281021054-4228-2-git-send-email-jkacur@redhat.com> Signed-off-by: Ingo Molnar diff --git a/kernel/hung_task.c b/kernel/hung_task.c index bca9423..53ead17 100644 --- a/kernel/hung_task.c +++ b/kernel/hung_task.c @@ -111,7 +111,7 @@ static void check_hung_task(struct task_struct *t, unsigned long timeout) * periodically exit the critical section and enter a new one. * * For preemptible RCU it is sufficient to call rcu_read_unlock in order - * exit the grace period. For classic RCU, a reschedule is required. + * to exit the grace period. For classic RCU, a reschedule is required. */ static void rcu_lock_break(struct task_struct *g, struct task_struct *t) { -- cgit v0.10.2 From d34a16661ed0fed433c9469d7cfa3ca4d30ca42e Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 14 Jun 2010 17:06:21 -0700 Subject: net: convert to rcu_dereference_index_check() The task_cls_classid() function applies rcu_dereference() to integers, which does not work with the shiny new sparse-based checking in rcu_dereference(). This commit therefore moves to the new RCU API rcu_dereference_index_check(). Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett Acked-by: David S. Miller Acked-by: Herbert Xu diff --git a/include/net/cls_cgroup.h b/include/net/cls_cgroup.h index 726cc35..dd1fdb8 100644 --- a/include/net/cls_cgroup.h +++ b/include/net/cls_cgroup.h @@ -45,7 +45,8 @@ static inline u32 task_cls_classid(struct task_struct *p) return 0; rcu_read_lock(); - id = rcu_dereference(net_cls_subsys_id); + id = rcu_dereference_index_check(net_cls_subsys_id, + rcu_read_lock_held()); if (id >= 0) classid = container_of(task_subsys_state(p, id), struct cgroup_cls_state, css)->classid; -- cgit v0.10.2 From ca5ecddfa8fcbd948c95530e7e817cee9fb43a3d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 28 Apr 2010 14:39:09 -0700 Subject: rcu: define __rcu address space modifier for sparse This commit provides definitions for the __rcu annotation defined earlier. This annotation permits sparse to check for correct use of RCU-protected pointers. If a pointer that is annotated with __rcu is accessed directly (as opposed to via rcu_dereference(), rcu_assign_pointer(), or one of their variants), sparse can be made to complain. To enable such complaints, use the new default-disabled CONFIG_SPARSE_RCU_POINTER kernel configuration option. Please note that these sparse complaints are intended to be a debugging aid, -not- a code-style-enforcement mechanism. There are special rcu_dereference_protected() and rcu_access_pointer() accessors for use when RCU read-side protection is not required, for example, when no other CPU has access to the data structure in question or while the current CPU hold the update-side lock. This patch also updates a number of docbook comments that were showing their age. Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Christopher Li Reviewed-by: Josh Triplett diff --git a/include/linux/compiler.h b/include/linux/compiler.h index c1a62c5..320d6c9 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -16,7 +16,11 @@ # define __release(x) __context__(x,-1) # define __cond_lock(x,c) ((c) ? ({ __acquire(x); 1; }) : 0) # define __percpu __attribute__((noderef, address_space(3))) +#ifdef CONFIG_SPARSE_RCU_POINTER +# define __rcu __attribute__((noderef, address_space(4))) +#else # define __rcu +#endif extern void __chk_user_ptr(const volatile void __user *); extern void __chk_io_ptr(const volatile void __iomem *); #else diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index 9fbc54a..b973dea 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -41,6 +41,7 @@ #include #include #include +#include #ifdef CONFIG_RCU_TORTURE_TEST extern int rcutorture_runnable; /* for sysctl */ @@ -120,14 +121,15 @@ extern struct lockdep_map rcu_sched_lock_map; extern int debug_lockdep_rcu_enabled(void); /** - * rcu_read_lock_held - might we be in RCU read-side critical section? + * rcu_read_lock_held() - might we be in RCU read-side critical section? * * If CONFIG_DEBUG_LOCK_ALLOC is selected, returns nonzero iff in an RCU * read-side critical section. In absence of CONFIG_DEBUG_LOCK_ALLOC, * this assumes we are in an RCU read-side critical section unless it can - * prove otherwise. + * prove otherwise. This is useful for debug checks in functions that + * require that they be called within an RCU read-side critical section. * - * Check debug_lockdep_rcu_enabled() to prevent false positives during boot + * Checks debug_lockdep_rcu_enabled() to prevent false positives during boot * and while lockdep is disabled. */ static inline int rcu_read_lock_held(void) @@ -144,14 +146,16 @@ static inline int rcu_read_lock_held(void) extern int rcu_read_lock_bh_held(void); /** - * rcu_read_lock_sched_held - might we be in RCU-sched read-side critical section? + * rcu_read_lock_sched_held() - might we be in RCU-sched read-side critical section? * * If CONFIG_DEBUG_LOCK_ALLOC is selected, returns nonzero iff in an * RCU-sched read-side critical section. In absence of * CONFIG_DEBUG_LOCK_ALLOC, this assumes we are in an RCU-sched read-side * critical section unless it can prove otherwise. Note that disabling * of preemption (including disabling irqs) counts as an RCU-sched - * read-side critical section. + * read-side critical section. This is useful for debug checks in functions + * that required that they be called within an RCU-sched read-side + * critical section. * * Check debug_lockdep_rcu_enabled() to prevent false positives during boot * and while lockdep is disabled. @@ -220,41 +224,155 @@ extern int rcu_my_thread_group_empty(void); } \ } while (0) +#else /* #ifdef CONFIG_PROVE_RCU */ + +#define __do_rcu_dereference_check(c) do { } while (0) + +#endif /* #else #ifdef CONFIG_PROVE_RCU */ + +/* + * Helper functions for rcu_dereference_check(), rcu_dereference_protected() + * and rcu_assign_pointer(). Some of these could be folded into their + * callers, but they are left separate in order to ease introduction of + * multiple flavors of pointers to match the multiple flavors of RCU + * (e.g., __rcu_bh, * __rcu_sched, and __srcu), should this make sense in + * the future. + */ +#define __rcu_access_pointer(p, space) \ + ({ \ + typeof(*p) *_________p1 = (typeof(*p)*__force )ACCESS_ONCE(p); \ + (void) (((typeof (*p) space *)p) == p); \ + ((typeof(*p) __force __kernel *)(_________p1)); \ + }) +#define __rcu_dereference_check(p, c, space) \ + ({ \ + typeof(*p) *_________p1 = (typeof(*p)*__force )ACCESS_ONCE(p); \ + __do_rcu_dereference_check(c); \ + (void) (((typeof (*p) space *)p) == p); \ + smp_read_barrier_depends(); \ + ((typeof(*p) __force __kernel *)(_________p1)); \ + }) +#define __rcu_dereference_protected(p, c, space) \ + ({ \ + __do_rcu_dereference_check(c); \ + (void) (((typeof (*p) space *)p) == p); \ + ((typeof(*p) __force __kernel *)(p)); \ + }) + +#define __rcu_dereference_index_check(p, c) \ + ({ \ + typeof(p) _________p1 = ACCESS_ONCE(p); \ + __do_rcu_dereference_check(c); \ + smp_read_barrier_depends(); \ + (_________p1); \ + }) +#define __rcu_assign_pointer(p, v, space) \ + ({ \ + if (!__builtin_constant_p(v) || \ + ((v) != NULL)) \ + smp_wmb(); \ + (p) = (typeof(*v) __force space *)(v); \ + }) + + +/** + * rcu_access_pointer() - fetch RCU pointer with no dereferencing + * @p: The pointer to read + * + * Return the value of the specified RCU-protected pointer, but omit the + * smp_read_barrier_depends() and keep the ACCESS_ONCE(). This is useful + * when the value of this pointer is accessed, but the pointer is not + * dereferenced, for example, when testing an RCU-protected pointer against + * NULL. Although rcu_access_pointer() may also be used in cases where + * update-side locks prevent the value of the pointer from changing, you + * should instead use rcu_dereference_protected() for this use case. + */ +#define rcu_access_pointer(p) __rcu_access_pointer((p), __rcu) + /** - * rcu_dereference_check - rcu_dereference with debug checking + * rcu_dereference_check() - rcu_dereference with debug checking * @p: The pointer to read, prior to dereferencing * @c: The conditions under which the dereference will take place * * Do an rcu_dereference(), but check that the conditions under which the - * dereference will take place are correct. Typically the conditions indicate - * the various locking conditions that should be held at that point. The check - * should return true if the conditions are satisfied. + * dereference will take place are correct. Typically the conditions + * indicate the various locking conditions that should be held at that + * point. The check should return true if the conditions are satisfied. + * An implicit check for being in an RCU read-side critical section + * (rcu_read_lock()) is included. * * For example: * - * bar = rcu_dereference_check(foo->bar, rcu_read_lock_held() || - * lockdep_is_held(&foo->lock)); + * bar = rcu_dereference_check(foo->bar, lockdep_is_held(&foo->lock)); * * could be used to indicate to lockdep that foo->bar may only be dereferenced - * if either the RCU read lock is held, or that the lock required to replace + * if either rcu_read_lock() is held, or that the lock required to replace * the bar struct at foo->bar is held. * * Note that the list of conditions may also include indications of when a lock * need not be held, for example during initialisation or destruction of the * target struct: * - * bar = rcu_dereference_check(foo->bar, rcu_read_lock_held() || - * lockdep_is_held(&foo->lock) || + * bar = rcu_dereference_check(foo->bar, lockdep_is_held(&foo->lock) || * atomic_read(&foo->usage) == 0); + * + * Inserts memory barriers on architectures that require them + * (currently only the Alpha), prevents the compiler from refetching + * (and from merging fetches), and, more importantly, documents exactly + * which pointers are protected by RCU and checks that the pointer is + * annotated as __rcu. */ #define rcu_dereference_check(p, c) \ - ({ \ - __do_rcu_dereference_check(c); \ - rcu_dereference_raw(p); \ - }) + __rcu_dereference_check((p), rcu_read_lock_held() || (c), __rcu) + +/** + * rcu_dereference_bh_check() - rcu_dereference_bh with debug checking + * @p: The pointer to read, prior to dereferencing + * @c: The conditions under which the dereference will take place + * + * This is the RCU-bh counterpart to rcu_dereference_check(). + */ +#define rcu_dereference_bh_check(p, c) \ + __rcu_dereference_check((p), rcu_read_lock_bh_held() || (c), __rcu) /** - * rcu_dereference_protected - fetch RCU pointer when updates prevented + * rcu_dereference_sched_check() - rcu_dereference_sched with debug checking + * @p: The pointer to read, prior to dereferencing + * @c: The conditions under which the dereference will take place + * + * This is the RCU-sched counterpart to rcu_dereference_check(). + */ +#define rcu_dereference_sched_check(p, c) \ + __rcu_dereference_check((p), rcu_read_lock_sched_held() || (c), \ + __rcu) + +#define rcu_dereference_raw(p) rcu_dereference_check(p, 1) /*@@@ needed? @@@*/ + +/** + * rcu_dereference_index_check() - rcu_dereference for indices with debug checking + * @p: The pointer to read, prior to dereferencing + * @c: The conditions under which the dereference will take place + * + * Similar to rcu_dereference_check(), but omits the sparse checking. + * This allows rcu_dereference_index_check() to be used on integers, + * which can then be used as array indices. Attempting to use + * rcu_dereference_check() on an integer will give compiler warnings + * because the sparse address-space mechanism relies on dereferencing + * the RCU-protected pointer. Dereferencing integers is not something + * that even gcc will put up with. + * + * Note that this function does not implicitly check for RCU read-side + * critical sections. If this function gains lots of uses, it might + * make sense to provide versions for each flavor of RCU, but it does + * not make sense as of early 2010. + */ +#define rcu_dereference_index_check(p, c) \ + __rcu_dereference_index_check((p), (c)) + +/** + * rcu_dereference_protected() - fetch RCU pointer when updates prevented + * @p: The pointer to read, prior to dereferencing + * @c: The conditions under which the dereference will take place * * Return the value of the specified RCU-protected pointer, but omit * both the smp_read_barrier_depends() and the ACCESS_ONCE(). This @@ -263,35 +381,61 @@ extern int rcu_my_thread_group_empty(void); * prevent the compiler from repeating this reference or combining it * with other references, so it should not be used without protection * of appropriate locks. + * + * This function is only for update-side use. Using this function + * when protected only by rcu_read_lock() will result in infrequent + * but very ugly failures. */ #define rcu_dereference_protected(p, c) \ - ({ \ - __do_rcu_dereference_check(c); \ - (p); \ - }) + __rcu_dereference_protected((p), (c), __rcu) -#else /* #ifdef CONFIG_PROVE_RCU */ +/** + * rcu_dereference_bh_protected() - fetch RCU-bh pointer when updates prevented + * @p: The pointer to read, prior to dereferencing + * @c: The conditions under which the dereference will take place + * + * This is the RCU-bh counterpart to rcu_dereference_protected(). + */ +#define rcu_dereference_bh_protected(p, c) \ + __rcu_dereference_protected((p), (c), __rcu) -#define rcu_dereference_check(p, c) rcu_dereference_raw(p) -#define rcu_dereference_protected(p, c) (p) +/** + * rcu_dereference_sched_protected() - fetch RCU-sched pointer when updates prevented + * @p: The pointer to read, prior to dereferencing + * @c: The conditions under which the dereference will take place + * + * This is the RCU-sched counterpart to rcu_dereference_protected(). + */ +#define rcu_dereference_sched_protected(p, c) \ + __rcu_dereference_protected((p), (c), __rcu) -#endif /* #else #ifdef CONFIG_PROVE_RCU */ /** - * rcu_access_pointer - fetch RCU pointer with no dereferencing + * rcu_dereference() - fetch RCU-protected pointer for dereferencing + * @p: The pointer to read, prior to dereferencing * - * Return the value of the specified RCU-protected pointer, but omit the - * smp_read_barrier_depends() and keep the ACCESS_ONCE(). This is useful - * when the value of this pointer is accessed, but the pointer is not - * dereferenced, for example, when testing an RCU-protected pointer against - * NULL. This may also be used in cases where update-side locks prevent - * the value of the pointer from changing, but rcu_dereference_protected() - * is a lighter-weight primitive for this use case. + * This is a simple wrapper around rcu_dereference_check(). */ -#define rcu_access_pointer(p) ACCESS_ONCE(p) +#define rcu_dereference(p) rcu_dereference_check(p, 0) /** - * rcu_read_lock - mark the beginning of an RCU read-side critical section. + * rcu_dereference_bh() - fetch an RCU-bh-protected pointer for dereferencing + * @p: The pointer to read, prior to dereferencing + * + * Makes rcu_dereference_check() do the dirty work. + */ +#define rcu_dereference_bh(p) rcu_dereference_bh_check(p, 0) + +/** + * rcu_dereference_sched() - fetch RCU-sched-protected pointer for dereferencing + * @p: The pointer to read, prior to dereferencing + * + * Makes rcu_dereference_check() do the dirty work. + */ +#define rcu_dereference_sched(p) rcu_dereference_sched_check(p, 0) + +/** + * rcu_read_lock() - mark the beginning of an RCU read-side critical section * * When synchronize_rcu() is invoked on one CPU while other CPUs * are within RCU read-side critical sections, then the @@ -337,7 +481,7 @@ static inline void rcu_read_lock(void) */ /** - * rcu_read_unlock - marks the end of an RCU read-side critical section. + * rcu_read_unlock() - marks the end of an RCU read-side critical section. * * See rcu_read_lock() for more information. */ @@ -349,15 +493,16 @@ static inline void rcu_read_unlock(void) } /** - * rcu_read_lock_bh - mark the beginning of a softirq-only RCU critical section + * rcu_read_lock_bh() - mark the beginning of an RCU-bh critical section * * This is equivalent of rcu_read_lock(), but to be used when updates - * are being done using call_rcu_bh(). Since call_rcu_bh() callbacks - * consider completion of a softirq handler to be a quiescent state, - * a process in RCU read-side critical section must be protected by - * disabling softirqs. Read-side critical sections in interrupt context - * can use just rcu_read_lock(). - * + * are being done using call_rcu_bh() or synchronize_rcu_bh(). Since + * both call_rcu_bh() and synchronize_rcu_bh() consider completion of a + * softirq handler to be a quiescent state, a process in RCU read-side + * critical section must be protected by disabling softirqs. Read-side + * critical sections in interrupt context can use just rcu_read_lock(), + * though this should at least be commented to avoid confusing people + * reading the code. */ static inline void rcu_read_lock_bh(void) { @@ -379,13 +524,12 @@ static inline void rcu_read_unlock_bh(void) } /** - * rcu_read_lock_sched - mark the beginning of a RCU-classic critical section + * rcu_read_lock_sched() - mark the beginning of a RCU-sched critical section * - * Should be used with either - * - synchronize_sched() - * or - * - call_rcu_sched() and rcu_barrier_sched() - * on the write-side to insure proper synchronization. + * This is equivalent of rcu_read_lock(), but to be used when updates + * are being done using call_rcu_sched() or synchronize_rcu_sched(). + * Read-side critical sections can also be introduced by anything that + * disables preemption, including local_irq_disable() and friends. */ static inline void rcu_read_lock_sched(void) { @@ -420,54 +564,14 @@ static inline notrace void rcu_read_unlock_sched_notrace(void) preempt_enable_notrace(); } - /** - * rcu_dereference_raw - fetch an RCU-protected pointer + * rcu_assign_pointer() - assign to RCU-protected pointer + * @p: pointer to assign to + * @v: value to assign (publish) * - * The caller must be within some flavor of RCU read-side critical - * section, or must be otherwise preventing the pointer from changing, - * for example, by holding an appropriate lock. This pointer may later - * be safely dereferenced. It is the caller's responsibility to have - * done the right thing, as this primitive does no checking of any kind. - * - * Inserts memory barriers on architectures that require them - * (currently only the Alpha), and, more importantly, documents - * exactly which pointers are protected by RCU. - */ -#define rcu_dereference_raw(p) ({ \ - typeof(p) _________p1 = ACCESS_ONCE(p); \ - smp_read_barrier_depends(); \ - (_________p1); \ - }) - -/** - * rcu_dereference - fetch an RCU-protected pointer, checking for RCU - * - * Makes rcu_dereference_check() do the dirty work. - */ -#define rcu_dereference(p) \ - rcu_dereference_check(p, rcu_read_lock_held()) - -/** - * rcu_dereference_bh - fetch an RCU-protected pointer, checking for RCU-bh - * - * Makes rcu_dereference_check() do the dirty work. - */ -#define rcu_dereference_bh(p) \ - rcu_dereference_check(p, rcu_read_lock_bh_held()) - -/** - * rcu_dereference_sched - fetch RCU-protected pointer, checking for RCU-sched - * - * Makes rcu_dereference_check() do the dirty work. - */ -#define rcu_dereference_sched(p) \ - rcu_dereference_check(p, rcu_read_lock_sched_held()) - -/** - * rcu_assign_pointer - assign (publicize) a pointer to a newly - * initialized structure that will be dereferenced by RCU read-side - * critical sections. Returns the value assigned. + * Assigns the specified value to the specified RCU-protected + * pointer, ensuring that any concurrent RCU readers will see + * any prior initialization. Returns the value assigned. * * Inserts memory barriers on architectures that require them * (pretty much all of them other than x86), and also prevents @@ -476,14 +580,17 @@ static inline notrace void rcu_read_unlock_sched_notrace(void) * call documents which pointers will be dereferenced by RCU read-side * code. */ - #define rcu_assign_pointer(p, v) \ - ({ \ - if (!__builtin_constant_p(v) || \ - ((v) != NULL)) \ - smp_wmb(); \ - (p) = (v); \ - }) + __rcu_assign_pointer((p), (v), __rcu) + +/** + * RCU_INIT_POINTER() - initialize an RCU protected pointer + * + * Initialize an RCU-protected pointer in such a way to avoid RCU-lockdep + * splats. + */ +#define RCU_INIT_POINTER(p, v) \ + p = (typeof(*v) __force __rcu *)(v) /* Infrastructure to implement the synchronize_() primitives. */ @@ -495,7 +602,7 @@ struct rcu_synchronize { extern void wakeme_after_rcu(struct rcu_head *head); /** - * call_rcu - Queue an RCU callback for invocation after a grace period. + * call_rcu() - Queue an RCU callback for invocation after a grace period. * @head: structure to be used for queueing the RCU updates. * @func: actual update function to be invoked after the grace period * @@ -509,7 +616,7 @@ extern void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *head)); /** - * call_rcu_bh - Queue an RCU for invocation after a quicker grace period. + * call_rcu_bh() - Queue an RCU for invocation after a quicker grace period. * @head: structure to be used for queueing the RCU updates. * @func: actual update function to be invoked after the grace period * @@ -566,37 +673,4 @@ static inline void debug_rcu_head_unqueue(struct rcu_head *head) } #endif /* #else !CONFIG_DEBUG_OBJECTS_RCU_HEAD */ -#ifndef CONFIG_PROVE_RCU -#define __do_rcu_dereference_check(c) do { } while (0) -#endif /* #ifdef CONFIG_PROVE_RCU */ - -#define __rcu_dereference_index_check(p, c) \ - ({ \ - typeof(p) _________p1 = ACCESS_ONCE(p); \ - __do_rcu_dereference_check(c); \ - smp_read_barrier_depends(); \ - (_________p1); \ - }) - -/** - * rcu_dereference_index_check() - rcu_dereference for indices with debug checking - * @p: The pointer to read, prior to dereferencing - * @c: The conditions under which the dereference will take place - * - * Similar to rcu_dereference_check(), but omits the sparse checking. - * This allows rcu_dereference_index_check() to be used on integers, - * which can then be used as array indices. Attempting to use - * rcu_dereference_check() on an integer will give compiler warnings - * because the sparse address-space mechanism relies on dereferencing - * the RCU-protected pointer. Dereferencing integers is not something - * that even gcc will put up with. - * - * Note that this function does not implicitly check for RCU read-side - * critical sections. If this function gains lots of uses, it might - * make sense to provide versions for each flavor of RCU, but it does - * not make sense as of early 2010. - */ -#define rcu_dereference_index_check(p, c) \ - __rcu_dereference_index_check((p), (c)) - #endif /* __LINUX_RCUPDATE_H */ diff --git a/include/linux/srcu.h b/include/linux/srcu.h index 4d5d2f5..6f456a7 100644 --- a/include/linux/srcu.h +++ b/include/linux/srcu.h @@ -108,12 +108,31 @@ static inline int srcu_read_lock_held(struct srcu_struct *sp) #endif /* #else #ifdef CONFIG_DEBUG_LOCK_ALLOC */ /** - * srcu_dereference - fetch SRCU-protected pointer with checking + * srcu_dereference_check - fetch SRCU-protected pointer for later dereferencing + * @p: the pointer to fetch and protect for later dereferencing + * @sp: pointer to the srcu_struct, which is used to check that we + * really are in an SRCU read-side critical section. + * @c: condition to check for update-side use * - * Makes rcu_dereference_check() do the dirty work. + * If PROVE_RCU is enabled, invoking this outside of an RCU read-side + * critical section will result in an RCU-lockdep splat, unless @c evaluates + * to 1. The @c argument will normally be a logical expression containing + * lockdep_is_held() calls. */ -#define srcu_dereference(p, sp) \ - rcu_dereference_check(p, srcu_read_lock_held(sp)) +#define srcu_dereference_check(p, sp, c) \ + __rcu_dereference_check((p), srcu_read_lock_held(sp) || (c), __rcu) + +/** + * srcu_dereference - fetch SRCU-protected pointer for later dereferencing + * @p: the pointer to fetch and protect for later dereferencing + * @sp: pointer to the srcu_struct, which is used to check that we + * really are in an SRCU read-side critical section. + * + * Makes rcu_dereference_check() do the dirty work. If PROVE_RCU + * is enabled, invoking this outside of an RCU read-side critical + * section will result in an RCU-lockdep splat. + */ +#define srcu_dereference(p, sp) srcu_dereference_check((p), (sp), 0) /** * srcu_read_lock - register a new reader for an SRCU-protected structure. diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 4d16983..6c79e85 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -73,12 +73,14 @@ int debug_lockdep_rcu_enabled(void) EXPORT_SYMBOL_GPL(debug_lockdep_rcu_enabled); /** - * rcu_read_lock_bh_held - might we be in RCU-bh read-side critical section? + * rcu_read_lock_bh_held() - might we be in RCU-bh read-side critical section? * * Check for bottom half being disabled, which covers both the * CONFIG_PROVE_RCU and not cases. Note that if someone uses * rcu_read_lock_bh(), but then later enables BH, lockdep (if enabled) - * will show the situation. + * will show the situation. This is useful for debug checks in functions + * that require that they be called within an RCU read-side critical + * section. * * Check debug_lockdep_rcu_enabled() to prevent false positives during boot. */ diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 1b4afd2..12465f2 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -539,6 +539,19 @@ config PROVE_RCU_REPEATEDLY disabling, allowing multiple RCU-lockdep warnings to be printed on a single reboot. +config SPARSE_RCU_POINTER + bool "RCU debugging: sparse-based checks for pointer usage" + default n + help + This feature enables the __rcu sparse annotation for + RCU-protected pointers. This annotation will cause sparse + to flag any non-RCU used of annotated pointers. This can be + helpful when debugging RCU usage. Please note that this feature + is not intended to enforce code cleanliness; it is instead merely + a debugging aid. + + Say Y to make sparse flag questionable use of RCU-protected pointers + Say N if you are unsure. config LOCKDEP -- cgit v0.10.2 From 67bdbffd696f29a0b68aa8daa285783a06651583 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Thu, 25 Feb 2010 16:55:13 +0100 Subject: rculist: avoid __rcu annotations This avoids warnings from missing __rcu annotations in the rculist implementation, making it possible to use the same lists in both RCU and non-RCU cases. We can add rculist annotations later, together with lockdep support for rculist, which is missing as well, but that may involve changing all the users. Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Pavel Emelyanov Cc: Sukadev Bhattiprolu Reviewed-by: Josh Triplett diff --git a/include/linux/rculist.h b/include/linux/rculist.h index 4ec3b38..c10b105 100644 --- a/include/linux/rculist.h +++ b/include/linux/rculist.h @@ -10,6 +10,12 @@ #include /* + * return the ->next pointer of a list_head in an rcu safe + * way, we must not access it directly + */ +#define list_next_rcu(list) (*((struct list_head __rcu **)(&(list)->next))) + +/* * Insert a new entry between two known consecutive entries. * * This is only for internal list manipulation where we know @@ -20,7 +26,7 @@ static inline void __list_add_rcu(struct list_head *new, { new->next = next; new->prev = prev; - rcu_assign_pointer(prev->next, new); + rcu_assign_pointer(list_next_rcu(prev), new); next->prev = new; } @@ -138,7 +144,7 @@ static inline void list_replace_rcu(struct list_head *old, { new->next = old->next; new->prev = old->prev; - rcu_assign_pointer(new->prev->next, new); + rcu_assign_pointer(list_next_rcu(new->prev), new); new->next->prev = new; old->prev = LIST_POISON2; } @@ -193,7 +199,7 @@ static inline void list_splice_init_rcu(struct list_head *list, */ last->next = at; - rcu_assign_pointer(head->next, first); + rcu_assign_pointer(list_next_rcu(head), first); first->prev = head; at->prev = last; } @@ -208,7 +214,9 @@ static inline void list_splice_init_rcu(struct list_head *list, * primitives such as list_add_rcu() as long as it's guarded by rcu_read_lock(). */ #define list_entry_rcu(ptr, type, member) \ - container_of(rcu_dereference_raw(ptr), type, member) + ({typeof (*ptr) __rcu *__ptr = (typeof (*ptr) __rcu __force *)ptr; \ + container_of((typeof(ptr))rcu_dereference_raw(__ptr), type, member); \ + }) /** * list_first_entry_rcu - get the first element from a list @@ -225,9 +233,9 @@ static inline void list_splice_init_rcu(struct list_head *list, list_entry_rcu((ptr)->next, type, member) #define __list_for_each_rcu(pos, head) \ - for (pos = rcu_dereference_raw((head)->next); \ + for (pos = rcu_dereference_raw(list_next_rcu(head)); \ pos != (head); \ - pos = rcu_dereference_raw(pos->next)) + pos = rcu_dereference_raw(list_next_rcu((pos))) /** * list_for_each_entry_rcu - iterate over rcu list of given type @@ -257,9 +265,9 @@ static inline void list_splice_init_rcu(struct list_head *list, * as long as the traversal is guarded by rcu_read_lock(). */ #define list_for_each_continue_rcu(pos, head) \ - for ((pos) = rcu_dereference_raw((pos)->next); \ + for ((pos) = rcu_dereference_raw(list_next_rcu(pos)); \ prefetch((pos)->next), (pos) != (head); \ - (pos) = rcu_dereference_raw((pos)->next)) + (pos) = rcu_dereference_raw(list_next_rcu(pos))) /** * list_for_each_entry_continue_rcu - continue iteration over list of given type @@ -314,12 +322,19 @@ static inline void hlist_replace_rcu(struct hlist_node *old, new->next = next; new->pprev = old->pprev; - rcu_assign_pointer(*new->pprev, new); + rcu_assign_pointer(*(struct hlist_node __rcu **)new->pprev, new); if (next) new->next->pprev = &new->next; old->pprev = LIST_POISON2; } +/* + * return the first or the next element in an RCU protected hlist + */ +#define hlist_first_rcu(head) (*((struct hlist_node __rcu **)(&(head)->first))) +#define hlist_next_rcu(node) (*((struct hlist_node __rcu **)(&(node)->next))) +#define hlist_pprev_rcu(node) (*((struct hlist_node __rcu **)((node)->pprev))) + /** * hlist_add_head_rcu * @n: the element to add to the hash list. @@ -346,7 +361,7 @@ static inline void hlist_add_head_rcu(struct hlist_node *n, n->next = first; n->pprev = &h->first; - rcu_assign_pointer(h->first, n); + rcu_assign_pointer(hlist_first_rcu(h), n); if (first) first->pprev = &n->next; } @@ -374,7 +389,7 @@ static inline void hlist_add_before_rcu(struct hlist_node *n, { n->pprev = next->pprev; n->next = next; - rcu_assign_pointer(*(n->pprev), n); + rcu_assign_pointer(hlist_pprev_rcu(n), n); next->pprev = &n->next; } @@ -401,15 +416,15 @@ static inline void hlist_add_after_rcu(struct hlist_node *prev, { n->next = prev->next; n->pprev = &prev->next; - rcu_assign_pointer(prev->next, n); + rcu_assign_pointer(hlist_next_rcu(prev), n); if (n->next) n->next->pprev = &n->next; } -#define __hlist_for_each_rcu(pos, head) \ - for (pos = rcu_dereference((head)->first); \ - pos && ({ prefetch(pos->next); 1; }); \ - pos = rcu_dereference(pos->next)) +#define __hlist_for_each_rcu(pos, head) \ + for (pos = rcu_dereference(hlist_first_rcu(head)); \ + pos && ({ prefetch(pos->next); 1; }); \ + pos = rcu_dereference(hlist_next_rcu(pos))) /** * hlist_for_each_entry_rcu - iterate over rcu list of given type @@ -422,11 +437,11 @@ static inline void hlist_add_after_rcu(struct hlist_node *prev, * the _rcu list-mutation primitives such as hlist_add_head_rcu() * as long as the traversal is guarded by rcu_read_lock(). */ -#define hlist_for_each_entry_rcu(tpos, pos, head, member) \ - for (pos = rcu_dereference_raw((head)->first); \ +#define hlist_for_each_entry_rcu(tpos, pos, head, member) \ + for (pos = rcu_dereference_raw(hlist_first_rcu(head)); \ pos && ({ prefetch(pos->next); 1; }) && \ ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1; }); \ - pos = rcu_dereference_raw(pos->next)) + pos = rcu_dereference_raw(hlist_next_rcu(pos))) /** * hlist_for_each_entry_rcu_bh - iterate over rcu list of given type diff --git a/include/linux/rculist_nulls.h b/include/linux/rculist_nulls.h index b70ffe5..2ae1371 100644 --- a/include/linux/rculist_nulls.h +++ b/include/linux/rculist_nulls.h @@ -37,6 +37,12 @@ static inline void hlist_nulls_del_init_rcu(struct hlist_nulls_node *n) } } +#define hlist_nulls_first_rcu(head) \ + (*((struct hlist_nulls_node __rcu __force **)&(head)->first)) + +#define hlist_nulls_next_rcu(node) \ + (*((struct hlist_nulls_node __rcu __force **)&(node)->next)) + /** * hlist_nulls_del_rcu - deletes entry from hash list without re-initialization * @n: the element to delete from the hash list. @@ -88,7 +94,7 @@ static inline void hlist_nulls_add_head_rcu(struct hlist_nulls_node *n, n->next = first; n->pprev = &h->first; - rcu_assign_pointer(h->first, n); + rcu_assign_pointer(hlist_nulls_first_rcu(h), n); if (!is_a_nulls(first)) first->pprev = &n->next; } @@ -100,11 +106,11 @@ static inline void hlist_nulls_add_head_rcu(struct hlist_nulls_node *n, * @member: the name of the hlist_nulls_node within the struct. * */ -#define hlist_nulls_for_each_entry_rcu(tpos, pos, head, member) \ - for (pos = rcu_dereference_raw((head)->first); \ - (!is_a_nulls(pos)) && \ +#define hlist_nulls_for_each_entry_rcu(tpos, pos, head, member) \ + for (pos = rcu_dereference_raw(hlist_nulls_first_rcu(head)); \ + (!is_a_nulls(pos)) && \ ({ tpos = hlist_nulls_entry(pos, typeof(*tpos), member); 1; }); \ - pos = rcu_dereference_raw(pos->next)) + pos = rcu_dereference_raw(hlist_nulls_next_rcu(pos))) #endif #endif diff --git a/kernel/pid.c b/kernel/pid.c index d55c6fb..0f90c2f 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -401,7 +401,7 @@ struct task_struct *pid_task(struct pid *pid, enum pid_type type) struct task_struct *result = NULL; if (pid) { struct hlist_node *first; - first = rcu_dereference_check(pid->tasks[type].first, + first = rcu_dereference_check(hlist_first_rcu(&pid->tasks[type]), rcu_read_lock_held() || lockdep_tasklist_lock_is_held()); if (first) -- cgit v0.10.2 From 2c392b8c3450ceb69ba1b93cb0cddb3998fb8cdc Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 24 Feb 2010 19:41:39 +0100 Subject: cgroups: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Acked-by: Paul Menage Cc: Li Zefan Reviewed-by: Josh Triplett diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h index ed3e92e..3cb7d04 100644 --- a/include/linux/cgroup.h +++ b/include/linux/cgroup.h @@ -75,7 +75,7 @@ struct cgroup_subsys_state { unsigned long flags; /* ID for this css, if possible */ - struct css_id *id; + struct css_id __rcu *id; }; /* bits in struct cgroup_subsys_state flags field */ @@ -205,7 +205,7 @@ struct cgroup { struct list_head children; /* my children */ struct cgroup *parent; /* my parent */ - struct dentry *dentry; /* cgroup fs entry, RCU protected */ + struct dentry __rcu *dentry; /* cgroup fs entry, RCU protected */ /* Private pointers for each registered subsystem */ struct cgroup_subsys_state *subsys[CGROUP_SUBSYS_COUNT]; diff --git a/include/linux/sched.h b/include/linux/sched.h index 1e2a6db..bbffd08 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1418,7 +1418,7 @@ struct task_struct { #endif #ifdef CONFIG_CGROUPS /* Control Group info protected by css_set_lock */ - struct css_set *cgroups; + struct css_set __rcu *cgroups; /* cg_list protected by css_set_lock and tsk->alloc_lock */ struct list_head cg_list; #endif diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 192f88c..e5c5497 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -138,7 +138,7 @@ struct css_id { * is called after synchronize_rcu(). But for safe use, css_is_removed() * css_tryget() should be used for avoiding race. */ - struct cgroup_subsys_state *css; + struct cgroup_subsys_state __rcu *css; /* * ID of this css. */ -- cgit v0.10.2 From 1b0ba1c9037b2265d6e5d0165d31e4c0269b603b Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 24 Feb 2010 19:45:09 +0100 Subject: credentials: rcu annotation Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Peter Zijlstra Cc: Ingo Molnar Acked-by: David Howells Reviewed-by: Josh Triplett diff --git a/include/linux/sched.h b/include/linux/sched.h index bbffd08..2c75666 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1288,9 +1288,9 @@ struct task_struct { struct list_head cpu_timers[3]; /* process credentials */ - const struct cred *real_cred; /* objective and real subjective task + const struct cred __rcu *real_cred; /* objective and real subjective task * credentials (COW) */ - const struct cred *cred; /* effective (overridable) subjective task + const struct cred __rcu *cred; /* effective (overridable) subjective task * credentials (COW) */ struct mutex cred_guard_mutex; /* guard against foreign influences on * credential calculations -- cgit v0.10.2 From e63ba744a64d234c8a07c469ab1806443cb0a6ff Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 26 Feb 2010 18:01:20 +0100 Subject: keys: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Acked-by: David Howells Reviewed-by: Josh Triplett diff --git a/include/linux/cred.h b/include/linux/cred.h index 4d2c395..4aaeab3 100644 --- a/include/linux/cred.h +++ b/include/linux/cred.h @@ -84,7 +84,7 @@ struct thread_group_cred { atomic_t usage; pid_t tgid; /* thread group process ID */ spinlock_t lock; - struct key *session_keyring; /* keyring inherited over fork */ + struct key __rcu *session_keyring; /* keyring inherited over fork */ struct key *process_keyring; /* keyring private to this process */ struct rcu_head rcu; /* RCU deletion hook */ }; diff --git a/include/linux/key.h b/include/linux/key.h index cd50dfa..3db0adc 100644 --- a/include/linux/key.h +++ b/include/linux/key.h @@ -178,8 +178,9 @@ struct key { */ union { unsigned long value; + void __rcu *rcudata; void *data; - struct keyring_list *subscriptions; + struct keyring_list __rcu *subscriptions; } payload; }; -- cgit v0.10.2 From 5b22216e11f717adc344abc7f97b42e03127c6c0 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 3 Mar 2010 10:20:10 +0100 Subject: nfs: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Acked-by: Trond Myklebust diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 508f8cf..d0edf7d 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -185,7 +185,7 @@ struct nfs_inode { struct nfs4_cached_acl *nfs4_acl; /* NFSv4 state */ struct list_head open_states; - struct nfs_delegation *delegation; + struct nfs_delegation __rcu *delegation; fmode_t delegation_state; struct rw_semaphore rwsem; #endif /* CONFIG_NFS_V4*/ diff --git a/include/linux/sunrpc/auth_gss.h b/include/linux/sunrpc/auth_gss.h index 671538d..8eee9db 100644 --- a/include/linux/sunrpc/auth_gss.h +++ b/include/linux/sunrpc/auth_gss.h @@ -69,7 +69,7 @@ struct gss_cl_ctx { enum rpc_gss_proc gc_proc; u32 gc_seq; spinlock_t gc_seq_lock; - struct gss_ctx *gc_gss_ctx; + struct gss_ctx __rcu *gc_gss_ctx; struct xdr_netobj gc_wire_ctx; u32 gc_win; unsigned long gc_expiry; @@ -80,7 +80,7 @@ struct gss_upcall_msg; struct gss_cred { struct rpc_cred gc_base; enum rpc_gss_svc gc_service; - struct gss_cl_ctx *gc_ctx; + struct gss_cl_ctx __rcu *gc_ctx; struct gss_upcall_msg *gc_upcall; unsigned long gc_upcall_timestamp; unsigned char gc_machine_cred : 1; -- cgit v0.10.2 From 2be85279281bafe7de808ca99de59af4fd474c49 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Thu, 4 Mar 2010 15:50:28 +0100 Subject: input: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Dmitry Torokhov Acked-by: Dmitry Torokhov Reviewed-by: Josh Triplett diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c index c908c5f..5808731 100644 --- a/drivers/input/evdev.c +++ b/drivers/input/evdev.c @@ -28,7 +28,7 @@ struct evdev { int minor; struct input_handle handle; wait_queue_head_t wait; - struct evdev_client *grab; + struct evdev_client __rcu *grab; struct list_head client_list; spinlock_t client_lock; /* protects client_list */ struct mutex mutex; diff --git a/include/linux/input.h b/include/linux/input.h index 896a922..d6ae176 100644 --- a/include/linux/input.h +++ b/include/linux/input.h @@ -1196,7 +1196,7 @@ struct input_dev { int (*flush)(struct input_dev *dev, struct file *file); int (*event)(struct input_dev *dev, unsigned int type, unsigned int code, int value); - struct input_handle *grab; + struct input_handle __rcu *grab; spinlock_t event_lock; struct mutex mutex; -- cgit v0.10.2 From 0906a372f2aa0fec1e59bd12b896883b6e41307a Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 9 Mar 2010 20:59:15 +0100 Subject: net/netfilter: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Acked-by: Patrick McHardy Cc: "David S. Miller" Cc: Eric Dumazet Reviewed-by: Josh Triplett diff --git a/include/net/netfilter/nf_conntrack.h b/include/net/netfilter/nf_conntrack.h index e624dae..caf17db 100644 --- a/include/net/netfilter/nf_conntrack.h +++ b/include/net/netfilter/nf_conntrack.h @@ -75,7 +75,7 @@ struct nf_conntrack_helper; /* nf_conn feature for connections that have a helper */ struct nf_conn_help { /* Helper. if any */ - struct nf_conntrack_helper *helper; + struct nf_conntrack_helper __rcu *helper; union nf_conntrack_help help; diff --git a/net/ipv4/netfilter/nf_nat_core.c b/net/ipv4/netfilter/nf_nat_core.c index 8c8632d..957c924 100644 --- a/net/ipv4/netfilter/nf_nat_core.c +++ b/net/ipv4/netfilter/nf_nat_core.c @@ -38,7 +38,7 @@ static DEFINE_SPINLOCK(nf_nat_lock); static struct nf_conntrack_l3proto *l3proto __read_mostly; #define MAX_IP_NAT_PROTO 256 -static const struct nf_nat_protocol *nf_nat_protos[MAX_IP_NAT_PROTO] +static const struct nf_nat_protocol __rcu *nf_nat_protos[MAX_IP_NAT_PROTO] __read_mostly; static inline const struct nf_nat_protocol * diff --git a/net/netfilter/core.c b/net/netfilter/core.c index 78b505d..fdaec7d 100644 --- a/net/netfilter/core.c +++ b/net/netfilter/core.c @@ -27,7 +27,7 @@ static DEFINE_MUTEX(afinfo_mutex); -const struct nf_afinfo *nf_afinfo[NFPROTO_NUMPROTO] __read_mostly; +const struct nf_afinfo __rcu *nf_afinfo[NFPROTO_NUMPROTO] __read_mostly; EXPORT_SYMBOL(nf_afinfo); int nf_register_afinfo(const struct nf_afinfo *afinfo) diff --git a/net/netfilter/nf_conntrack_ecache.c b/net/netfilter/nf_conntrack_ecache.c index cdcc764..5702de3 100644 --- a/net/netfilter/nf_conntrack_ecache.c +++ b/net/netfilter/nf_conntrack_ecache.c @@ -26,10 +26,10 @@ static DEFINE_MUTEX(nf_ct_ecache_mutex); -struct nf_ct_event_notifier *nf_conntrack_event_cb __read_mostly; +struct nf_ct_event_notifier __rcu *nf_conntrack_event_cb __read_mostly; EXPORT_SYMBOL_GPL(nf_conntrack_event_cb); -struct nf_exp_event_notifier *nf_expect_event_cb __read_mostly; +struct nf_exp_event_notifier __rcu *nf_expect_event_cb __read_mostly; EXPORT_SYMBOL_GPL(nf_expect_event_cb); /* deliver cached events and clear cache entry - must be called with locally diff --git a/net/netfilter/nf_conntrack_extend.c b/net/netfilter/nf_conntrack_extend.c index 7dcf7a4..1d9bdae 100644 --- a/net/netfilter/nf_conntrack_extend.c +++ b/net/netfilter/nf_conntrack_extend.c @@ -16,7 +16,7 @@ #include #include -static struct nf_ct_ext_type *nf_ct_ext_types[NF_CT_EXT_NUM]; +static struct nf_ct_ext_type __rcu *nf_ct_ext_types[NF_CT_EXT_NUM]; static DEFINE_MUTEX(nf_ct_ext_type_mutex); void __nf_ct_ext_destroy(struct nf_conn *ct) diff --git a/net/netfilter/nf_conntrack_proto.c b/net/netfilter/nf_conntrack_proto.c index 5886ba1..ed6d929 100644 --- a/net/netfilter/nf_conntrack_proto.c +++ b/net/netfilter/nf_conntrack_proto.c @@ -28,8 +28,8 @@ #include #include -static struct nf_conntrack_l4proto **nf_ct_protos[PF_MAX] __read_mostly; -struct nf_conntrack_l3proto *nf_ct_l3protos[AF_MAX] __read_mostly; +static struct nf_conntrack_l4proto __rcu **nf_ct_protos[PF_MAX] __read_mostly; +struct nf_conntrack_l3proto __rcu *nf_ct_l3protos[AF_MAX] __read_mostly; EXPORT_SYMBOL_GPL(nf_ct_l3protos); static DEFINE_MUTEX(nf_ct_proto_mutex); diff --git a/net/netfilter/nf_log.c b/net/netfilter/nf_log.c index 7df37fd..b07393e 100644 --- a/net/netfilter/nf_log.c +++ b/net/netfilter/nf_log.c @@ -16,7 +16,7 @@ #define NF_LOG_PREFIXLEN 128 #define NFLOGGER_NAME_LEN 64 -static const struct nf_logger *nf_loggers[NFPROTO_NUMPROTO] __read_mostly; +static const struct nf_logger __rcu *nf_loggers[NFPROTO_NUMPROTO] __read_mostly; static struct list_head nf_loggers_l[NFPROTO_NUMPROTO] __read_mostly; static DEFINE_MUTEX(nf_log_mutex); diff --git a/net/netfilter/nf_queue.c b/net/netfilter/nf_queue.c index 78b3cf9c..74aebed 100644 --- a/net/netfilter/nf_queue.c +++ b/net/netfilter/nf_queue.c @@ -18,7 +18,7 @@ * long term mutex. The handler must provide an an outfn() to accept packets * for queueing and must reinject all packets it receives, no matter what. */ -static const struct nf_queue_handler *queue_handler[NFPROTO_NUMPROTO] __read_mostly; +static const struct nf_queue_handler __rcu *queue_handler[NFPROTO_NUMPROTO] __read_mostly; static DEFINE_MUTEX(queue_handler_mutex); -- cgit v0.10.2 From 4b6a2872a2a00042ee50024822ab706e5456aad8 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Thu, 4 Mar 2010 15:59:23 +0100 Subject: kvm: add __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Avi Kivity Cc: Marcelo Tosatti Reviewed-by: Josh Triplett diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index c13cc48..ac740b2 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -205,7 +205,7 @@ struct kvm { struct mutex irq_lock; #ifdef CONFIG_HAVE_KVM_IRQCHIP - struct kvm_irq_routing_table *irq_routing; + struct kvm_irq_routing_table __rcu *irq_routing; struct hlist_head mask_notifier_list; struct hlist_head irq_ack_notifier_list; #endif -- cgit v0.10.2 From 84483ea42cd4f2781d6e97a83ab3ebd0ff19fb10 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 16 Jun 2010 16:48:13 -0700 Subject: rcu: add shiny new debug assists to Documentation/RCU/checklist.txt Add a section describing PROVE_RCU, DEBUG_OBJECTS_RCU_HEAD, and the __rcu sparse checking to the RCU checklist. Suggested-by: David Miller Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/Documentation/RCU/checklist.txt b/Documentation/RCU/checklist.txt index 790d1a8..c7c6788 100644 --- a/Documentation/RCU/checklist.txt +++ b/Documentation/RCU/checklist.txt @@ -365,3 +365,26 @@ over a rather long period of time, but improvements are always welcome! and the compiler to freely reorder code into and out of RCU read-side critical sections. It is the responsibility of the RCU update-side primitives to deal with this. + +17. Use CONFIG_PROVE_RCU, CONFIG_DEBUG_OBJECTS_RCU_HEAD, and + the __rcu sparse checks to validate your RCU code. These + can help find problems as follows: + + CONFIG_PROVE_RCU: check that accesses to RCU-protected data + structures are carried out under the proper RCU + read-side critical section, while holding the right + combination of locks, or whatever other conditions + are appropriate. + + CONFIG_DEBUG_OBJECTS_RCU_HEAD: check that you don't pass the + same object to call_rcu() (or friends) before an RCU + grace period has elapsed since the last time that you + passed that same object to call_rcu() (or friends). + + __rcu sparse checks: tag the pointer to the RCU-protected data + structure with __rcu, and sparse will warn you if you + access that pointer without the services of one of the + variants of rcu_dereference(). + + These debugging aids can help you find problems that are + otherwise extremely difficult to spot. -- cgit v0.10.2 From e546f485e1d7520ca0200502cdcc11b503f4805c Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 21 Jun 2010 16:57:42 +0800 Subject: rcutorture: add random preemption Add random preemption to help we to torture the preemptable rcu. srcu_read_delay() also calls rcu_read_delay() for shorter delays. Added comment to preempt_schedule() call indicating that no quiescent states happen if preemption is disabled. Signed-off-by: Lai Jiangshan Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 2e2726d..7297102 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -303,6 +303,10 @@ static void rcu_read_delay(struct rcu_random_state *rrsp) mdelay(longdelay_ms); if (!(rcu_random(rrsp) % (nrealreaders * 2 * shortdelay_us))) udelay(shortdelay_us); +#ifdef CONFIG_PREEMPT + if (!preempt_count() && !(rcu_random(rrsp) % (nrealreaders * 20000))) + preempt_schedule(); /* No QS if preempt_disable() in effect */ +#endif } static void rcu_torture_read_unlock(int idx) __releases(RCU) @@ -536,6 +540,8 @@ static void srcu_read_delay(struct rcu_random_state *rrsp) delay = rcu_random(rrsp) % (nrealreaders * 2 * longdelay * uspertick); if (!delay) schedule_timeout_interruptible(longdelay); + else + rcu_read_delay(rrsp); } static void srcu_torture_read_unlock(int idx) __releases(&srcu_ctl) -- cgit v0.10.2 From 394f99a9007d4274f7076bb8553ab0ff9707688b Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 28 Jun 2010 16:25:04 +0800 Subject: rcu: simplify the usage of percpu data &percpu_data is compatible with allocated percpu data. And we use it and remove the "->rda[NR_CPUS]" array, saving significant storage on systems with large numbers of CPUs. This does add an additional level of indirection and thus an additional cache line referenced, but because ->rda is not used on the read side, this is OK. Signed-off-by: Lai Jiangshan Reviewed-by: Tejun Heo Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/kernel/rcutree.c b/kernel/rcutree.c index d5bc439..5b1c3c2 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -712,7 +712,7 @@ static void rcu_start_gp(struct rcu_state *rsp, unsigned long flags) __releases(rcu_get_root(rsp)->lock) { - struct rcu_data *rdp = rsp->rda[smp_processor_id()]; + struct rcu_data *rdp = this_cpu_ptr(rsp->rda); struct rcu_node *rnp = rcu_get_root(rsp); if (!cpu_needs_another_gp(rsp, rdp) || rsp->fqs_active) { @@ -960,7 +960,7 @@ rcu_check_quiescent_state(struct rcu_state *rsp, struct rcu_data *rdp) static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp) { int i; - struct rcu_data *rdp = rsp->rda[smp_processor_id()]; + struct rcu_data *rdp = this_cpu_ptr(rsp->rda); if (rdp->nxtlist == NULL) return; /* irqs disabled, so comparison is stable. */ @@ -984,7 +984,7 @@ static void rcu_adopt_orphan_cbs(struct rcu_state *rsp) struct rcu_data *rdp; raw_spin_lock_irqsave(&rsp->onofflock, flags); - rdp = rsp->rda[smp_processor_id()]; + rdp = this_cpu_ptr(rsp->rda); if (rsp->orphan_cbs_list == NULL) { raw_spin_unlock_irqrestore(&rsp->onofflock, flags); return; @@ -1007,7 +1007,7 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) unsigned long flags; unsigned long mask; int need_report = 0; - struct rcu_data *rdp = rsp->rda[cpu]; + struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); struct rcu_node *rnp; /* Exclude any attempts to start a new grace period. */ @@ -1226,7 +1226,8 @@ static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) cpu = rnp->grplo; bit = 1; for (; cpu <= rnp->grphi; cpu++, bit <<= 1) { - if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) + if ((rnp->qsmask & bit) != 0 && + f(per_cpu_ptr(rsp->rda, cpu))) mask |= bit; } if (mask != 0) { @@ -1402,7 +1403,7 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), * a quiescent state betweentimes. */ local_irq_save(flags); - rdp = rsp->rda[smp_processor_id()]; + rdp = this_cpu_ptr(rsp->rda); rcu_process_gp_end(rsp, rdp); check_for_new_grace_period(rsp, rdp); @@ -1701,7 +1702,7 @@ rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) { unsigned long flags; int i; - struct rcu_data *rdp = rsp->rda[cpu]; + struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ @@ -1729,7 +1730,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) { unsigned long flags; unsigned long mask; - struct rcu_data *rdp = rsp->rda[cpu]; + struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ @@ -1865,7 +1866,8 @@ static void __init rcu_init_levelspread(struct rcu_state *rsp) /* * Helper function for rcu_init() that initializes one rcu_state structure. */ -static void __init rcu_init_one(struct rcu_state *rsp) +static void __init rcu_init_one(struct rcu_state *rsp, + struct rcu_data __percpu *rda) { static char *buf[] = { "rcu_node_level_0", "rcu_node_level_1", @@ -1918,37 +1920,23 @@ static void __init rcu_init_one(struct rcu_state *rsp) } } + rsp->rda = rda; rnp = rsp->level[NUM_RCU_LVLS - 1]; for_each_possible_cpu(i) { while (i > rnp->grphi) rnp++; - rsp->rda[i]->mynode = rnp; + per_cpu_ptr(rsp->rda, i)->mynode = rnp; rcu_boot_init_percpu_data(i, rsp); } } -/* - * Helper macro for __rcu_init() and __rcu_init_preempt(). To be used - * nowhere else! Assigns leaf node pointers into each CPU's rcu_data - * structure. - */ -#define RCU_INIT_FLAVOR(rsp, rcu_data) \ -do { \ - int i; \ - \ - for_each_possible_cpu(i) { \ - (rsp)->rda[i] = &per_cpu(rcu_data, i); \ - } \ - rcu_init_one(rsp); \ -} while (0) - void __init rcu_init(void) { int cpu; rcu_bootup_announce(); - RCU_INIT_FLAVOR(&rcu_sched_state, rcu_sched_data); - RCU_INIT_FLAVOR(&rcu_bh_state, rcu_bh_data); + rcu_init_one(&rcu_sched_state, &rcu_sched_data); + rcu_init_one(&rcu_bh_state, &rcu_bh_data); __rcu_init_preempt(); open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 14c040b..5ce197e 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -283,7 +283,7 @@ struct rcu_state { struct rcu_node *level[NUM_RCU_LVLS]; /* Hierarchy levels. */ u32 levelcnt[MAX_RCU_LVLS + 1]; /* # nodes in each level. */ u8 levelspread[NUM_RCU_LVLS]; /* kids/node in each level. */ - struct rcu_data *rda[NR_CPUS]; /* array of rdp pointers. */ + struct rcu_data __percpu *rda; /* pointer of percu rcu_data. */ /* The following fields are guarded by the root rcu_node's lock. */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 0e4f420..9906f85 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -154,7 +154,7 @@ static void rcu_preempt_note_context_switch(int cpu) (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) { /* Possibly blocking in an RCU read-side critical section. */ - rdp = rcu_preempt_state.rda[cpu]; + rdp = per_cpu_ptr(rcu_preempt_state.rda, cpu); rnp = rdp->mynode; raw_spin_lock_irqsave(&rnp->lock, flags); t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED; @@ -771,7 +771,7 @@ static void rcu_preempt_send_cbs_to_orphanage(void) */ static void __init __rcu_init_preempt(void) { - RCU_INIT_FLAVOR(&rcu_preempt_state, rcu_preempt_data); + rcu_init_one(&rcu_preempt_state, &rcu_preempt_data); } /* diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c index 36c95b4..458e032 100644 --- a/kernel/rcutree_trace.c +++ b/kernel/rcutree_trace.c @@ -262,7 +262,7 @@ static void print_rcu_pendings(struct seq_file *m, struct rcu_state *rsp) struct rcu_data *rdp; for_each_possible_cpu(cpu) { - rdp = rsp->rda[cpu]; + rdp = per_cpu_ptr(rsp->rda, cpu); if (rdp->beenonline) print_one_rcu_pending(m, rdp); } -- cgit v0.10.2 From 4221a9918e38b7494cee341dda7b7b4bb8c04bde Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Sat, 26 Jun 2010 01:08:19 +0900 Subject: Add RCU check for find_task_by_vpid(). find_task_by_vpid() says "Must be called under rcu_read_lock().". But due to commit 3120438 "rcu: Disable lockdep checking in RCU list-traversal primitives", we are currently unable to catch "find_task_by_vpid() with tasklist_lock held but RCU lock not held" errors due to the RCU-lockdep checks being suppressed in the RCU variants of the struct list_head traversals. This commit therefore places an explicit check for being in an RCU read-side critical section in find_task_by_pid_ns(). =================================================== [ INFO: suspicious rcu_dereference_check() usage. ] --------------------------------------------------- kernel/pid.c:386 invoked rcu_dereference_check() without protection! other info that might help us debug this: rcu_scheduler_active = 1, debug_locks = 1 1 lock held by rc.sysinit/1102: #0: (tasklist_lock){.+.+..}, at: [] sys_setpgid+0x40/0x160 stack backtrace: Pid: 1102, comm: rc.sysinit Not tainted 2.6.35-rc3-dirty #1 Call Trace: [] lockdep_rcu_dereference+0x94/0xb0 [] find_task_by_pid_ns+0x6d/0x70 [] find_task_by_vpid+0x18/0x20 [] sys_setpgid+0x47/0x160 [] sysenter_do_call+0x12/0x36 Commit updated to use a new rcu_lockdep_assert() exported API rather than the old internal __do_rcu_dereference(). Signed-off-by: Tetsuo Handa Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index b973dea..b124bc6 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -215,7 +215,11 @@ static inline int rcu_read_lock_sched_held(void) extern int rcu_my_thread_group_empty(void); -#define __do_rcu_dereference_check(c) \ +/** + * rcu_lockdep_assert - emit lockdep splat if specified condition not met + * @c: condition to check + */ +#define rcu_lockdep_assert(c) \ do { \ static bool __warned; \ if (debug_lockdep_rcu_enabled() && !__warned && !(c)) { \ @@ -226,7 +230,7 @@ extern int rcu_my_thread_group_empty(void); #else /* #ifdef CONFIG_PROVE_RCU */ -#define __do_rcu_dereference_check(c) do { } while (0) +#define rcu_lockdep_assert(c) do { } while (0) #endif /* #else #ifdef CONFIG_PROVE_RCU */ @@ -247,14 +251,14 @@ extern int rcu_my_thread_group_empty(void); #define __rcu_dereference_check(p, c, space) \ ({ \ typeof(*p) *_________p1 = (typeof(*p)*__force )ACCESS_ONCE(p); \ - __do_rcu_dereference_check(c); \ + rcu_lockdep_assert(c); \ (void) (((typeof (*p) space *)p) == p); \ smp_read_barrier_depends(); \ ((typeof(*p) __force __kernel *)(_________p1)); \ }) #define __rcu_dereference_protected(p, c, space) \ ({ \ - __do_rcu_dereference_check(c); \ + rcu_lockdep_assert(c); \ (void) (((typeof (*p) space *)p) == p); \ ((typeof(*p) __force __kernel *)(p)); \ }) @@ -262,7 +266,7 @@ extern int rcu_my_thread_group_empty(void); #define __rcu_dereference_index_check(p, c) \ ({ \ typeof(p) _________p1 = ACCESS_ONCE(p); \ - __do_rcu_dereference_check(c); \ + rcu_lockdep_assert(c); \ smp_read_barrier_depends(); \ (_________p1); \ }) diff --git a/kernel/pid.c b/kernel/pid.c index 0f90c2f..39b65b6 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -416,6 +416,7 @@ EXPORT_SYMBOL(pid_task); */ struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns) { + rcu_lockdep_assert(rcu_read_lock_held()); return pid_task(find_pid_ns(nr, ns), PIDTYPE_PID); } -- cgit v0.10.2 From b163760e37047781b37c412cde54d146ac4b651f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 2 Jun 2010 16:21:38 -0700 Subject: rcu: make CPU stall warning timeout configurable Also set the default to 60 seconds, up from the previous hard-coded timeout of 10 seconds. This allows people who care to set short timeouts, while avoiding people with unusual configurations (make randconfig!!!) from being bothered with spurious CPU stall warnings. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 5ce197e..183ebf4 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -254,9 +254,10 @@ struct rcu_data { #define RCU_STALL_DELAY_DELTA 0 #endif -#define RCU_SECONDS_TILL_STALL_CHECK (10 * HZ + RCU_STALL_DELAY_DELTA) +#define RCU_SECONDS_TILL_STALL_CHECK (CONFIG_RCU_CPU_STALL_TIMEOUT * HZ + \ + RCU_STALL_DELAY_DELTA) /* for rsp->jiffies_stall */ -#define RCU_SECONDS_TILL_STALL_RECHECK (30 * HZ + RCU_STALL_DELAY_DELTA) +#define RCU_SECONDS_TILL_STALL_RECHECK (3 * RCU_SECONDS_TILL_STALL_CHECK + 30) /* for rsp->jiffies_stall */ #define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ /* to take at least one */ diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 12465f2..b5005b7 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -845,6 +845,17 @@ config RCU_CPU_STALL_DETECTOR Say Y if you are unsure. +config RCU_CPU_STALL_TIMEOUT + int "RCU CPU stall timeout in seconds" + depends on RCU_CPU_STALL_DETECTOR + range 3 300 + default 60 + help + If a given RCU grace period extends more than the specified + number of seconds, a CPU stall warning is printed. If the + RCU grace period persists, additional CPU stall warnings are + printed at more widely spaced intervals. + config RCU_CPU_STALL_VERBOSE bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR" depends on RCU_CPU_STALL_DETECTOR && TREE_PREEMPT_RCU -- cgit v0.10.2 From 742734eea0cf5314cde5945963ed964be167bd84 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 30 Jun 2010 11:43:52 -0700 Subject: rcu: add boot parameter to suppress RCU CPU stall warning messages Although the RCU CPU stall warning messages are a very good way to alert people to a problem, once alerted, it is sometimes helpful to shut them off in order to avoid obscuring other messages that might be being used to track down the problem. Although you can rebuild the kernel with CONFIG_RCU_CPU_STALL_DETECTOR=n, this is sometimes inconvenient. This commit therefore adds a boot parameter named "rcu_cpu_stall_suppress" that shuts these messages off without requiring a rebuild (though a reboot might be needed for those not brave enough to patch their kernel while it is running). This message-suppression was already in place for the panic case, so this commit need only rename the variable and export it via module_param(). Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 5b1c3c2..f3d5906 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -143,6 +143,11 @@ module_param(blimit, int, 0); module_param(qhimark, int, 0); module_param(qlowmark, int, 0); +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR +int rcu_cpu_stall_suppress __read_mostly; +module_param(rcu_cpu_stall_suppress, int, 0); +#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ + static void force_quiescent_state(struct rcu_state *rsp, int relaxed); static int rcu_pending(int cpu); @@ -450,7 +455,7 @@ static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) #ifdef CONFIG_RCU_CPU_STALL_DETECTOR -int rcu_cpu_stall_panicking __read_mostly; +int rcu_cpu_stall_suppress __read_mostly; static void record_gp_stall_check_time(struct rcu_state *rsp) { @@ -530,7 +535,7 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp) long delta; struct rcu_node *rnp; - if (rcu_cpu_stall_panicking) + if (rcu_cpu_stall_suppress) return; delta = jiffies - rsp->jiffies_stall; rnp = rdp->mynode; @@ -548,7 +553,7 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp) static int rcu_panic(struct notifier_block *this, unsigned long ev, void *ptr) { - rcu_cpu_stall_panicking = 1; + rcu_cpu_stall_suppress = 1; return NOTIFY_DONE; } -- cgit v0.10.2 From 77d8485a8b5416c615b6acd95f01bfcacd7d81ff Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 8 Jul 2010 17:38:59 -0700 Subject: rcu: improve kerneldoc for rcu_read_lock(), call_rcu(), and synchronize_rcu() Make it explicit that new RCU read-side critical sections that start after call_rcu() and synchronize_rcu() start might still be running after the end of the relevant grace period. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index b124bc6..3e1b662 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -450,7 +450,7 @@ extern int rcu_my_thread_group_empty(void); * until after the all the other CPUs exit their critical sections. * * Note, however, that RCU callbacks are permitted to run concurrently - * with RCU read-side critical sections. One way that this can happen + * with new RCU read-side critical sections. One way that this can happen * is via the following sequence of events: (1) CPU 0 enters an RCU * read-side critical section, (2) CPU 1 invokes call_rcu() to register * an RCU callback, (3) CPU 0 exits the RCU read-side critical section, @@ -608,11 +608,13 @@ extern void wakeme_after_rcu(struct rcu_head *head); /** * call_rcu() - Queue an RCU callback for invocation after a grace period. * @head: structure to be used for queueing the RCU updates. - * @func: actual update function to be invoked after the grace period + * @func: actual callback function to be invoked after the grace period * - * The update function will be invoked some time after a full grace - * period elapses, in other words after all currently executing RCU - * read-side critical sections have completed. RCU read-side critical + * The callback function will be invoked some time after a full grace + * period elapses, in other words after all pre-existing RCU read-side + * critical sections have completed. However, the callback function + * might well execute concurrently with RCU read-side critical sections + * that started after call_rcu() was invoked. RCU read-side critical * sections are delimited by rcu_read_lock() and rcu_read_unlock(), * and may be nested. */ @@ -622,9 +624,9 @@ extern void call_rcu(struct rcu_head *head, /** * call_rcu_bh() - Queue an RCU for invocation after a quicker grace period. * @head: structure to be used for queueing the RCU updates. - * @func: actual update function to be invoked after the grace period + * @func: actual callback function to be invoked after the grace period * - * The update function will be invoked some time after a full grace + * The callback function will be invoked some time after a full grace * period elapses, in other words after all currently executing RCU * read-side critical sections have completed. call_rcu_bh() assumes * that the read-side critical sections end on completion of a softirq diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 9906f85..63bb771 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -546,9 +546,11 @@ EXPORT_SYMBOL_GPL(call_rcu); * * Control will return to the caller some time after a full grace * period has elapsed, in other words after all currently executing RCU - * read-side critical sections have completed. RCU read-side critical - * sections are delimited by rcu_read_lock() and rcu_read_unlock(), - * and may be nested. + * read-side critical sections have completed. Note, however, that + * upon return from synchronize_rcu(), the caller might well be executing + * concurrently with new RCU read-side critical sections that began while + * synchronize_rcu() was waiting. RCU read-side critical sections are + * delimited by rcu_read_lock() and rcu_read_unlock(), and may be nested. */ void synchronize_rcu(void) { -- cgit v0.10.2 From 374a8e0dc33c984fac284de7d57d77af3cfdbfb7 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 24 Feb 2010 20:00:13 +0100 Subject: notifiers: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Alan Cox Reviewed-by: Josh Triplett diff --git a/include/linux/notifier.h b/include/linux/notifier.h index b2f1a4d..2026f9e 100644 --- a/include/linux/notifier.h +++ b/include/linux/notifier.h @@ -49,28 +49,28 @@ struct notifier_block { int (*notifier_call)(struct notifier_block *, unsigned long, void *); - struct notifier_block *next; + struct notifier_block __rcu *next; int priority; }; struct atomic_notifier_head { spinlock_t lock; - struct notifier_block *head; + struct notifier_block __rcu *head; }; struct blocking_notifier_head { struct rw_semaphore rwsem; - struct notifier_block *head; + struct notifier_block __rcu *head; }; struct raw_notifier_head { - struct notifier_block *head; + struct notifier_block __rcu *head; }; struct srcu_notifier_head { struct mutex mutex; struct srcu_struct srcu; - struct notifier_block *head; + struct notifier_block __rcu *head; }; #define ATOMIC_INIT_NOTIFIER_HEAD(name) do { \ -- cgit v0.10.2 From a1115570b31091f3e3ab9e6cf7ee8d320a42be84 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Thu, 25 Feb 2010 23:43:52 +0100 Subject: radix-tree: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Nick Piggin Reviewed-by: Josh Triplett diff --git a/include/linux/radix-tree.h b/include/linux/radix-tree.h index 634b8e6..a39cbed 100644 --- a/include/linux/radix-tree.h +++ b/include/linux/radix-tree.h @@ -47,6 +47,8 @@ static inline void *radix_tree_indirect_to_ptr(void *ptr) { return (void *)((unsigned long)ptr & ~RADIX_TREE_INDIRECT_PTR); } +#define radix_tree_indirect_to_ptr(ptr) \ + radix_tree_indirect_to_ptr((void __force *)(ptr)) static inline int radix_tree_is_indirect_ptr(void *ptr) { @@ -61,7 +63,7 @@ static inline int radix_tree_is_indirect_ptr(void *ptr) struct radix_tree_root { unsigned int height; gfp_t gfp_mask; - struct radix_tree_node *rnode; + struct radix_tree_node __rcu *rnode; }; #define RADIX_TREE_INIT(mask) { \ diff --git a/lib/radix-tree.c b/lib/radix-tree.c index e907858..899fb75 100644 --- a/lib/radix-tree.c +++ b/lib/radix-tree.c @@ -49,7 +49,7 @@ struct radix_tree_node { unsigned int height; /* Height from the bottom */ unsigned int count; struct rcu_head rcu_head; - void *slots[RADIX_TREE_MAP_SIZE]; + void __rcu *slots[RADIX_TREE_MAP_SIZE]; unsigned long tags[RADIX_TREE_MAX_TAGS][RADIX_TREE_TAG_LONGS]; }; -- cgit v0.10.2 From d2c2486bc8e185548490e8edbc84d185de9eaff1 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 26 Feb 2010 14:53:26 +0100 Subject: idr: __rcu annotations Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Manfred Spraul Reviewed-by: Josh Triplett diff --git a/include/linux/idr.h b/include/linux/idr.h index e968db7..cdb715e 100644 --- a/include/linux/idr.h +++ b/include/linux/idr.h @@ -50,14 +50,14 @@ struct idr_layer { unsigned long bitmap; /* A zero bit means "space here" */ - struct idr_layer *ary[1< Date: Wed, 24 Feb 2010 20:01:56 +0100 Subject: kernel: __rcu annotations This adds annotations for RCU operations in core kernel components Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: Al Viro Cc: Jens Axboe Cc: Andrew Morton Reviewed-by: Josh Triplett diff --git a/include/linux/fdtable.h b/include/linux/fdtable.h index f59ed29..133c0ba 100644 --- a/include/linux/fdtable.h +++ b/include/linux/fdtable.h @@ -31,7 +31,7 @@ struct embedded_fd_set { struct fdtable { unsigned int max_fds; - struct file ** fd; /* current fd array */ + struct file __rcu **fd; /* current fd array */ fd_set *close_on_exec; fd_set *open_fds; struct rcu_head rcu; @@ -46,7 +46,7 @@ struct files_struct { * read mostly part */ atomic_t count; - struct fdtable *fdt; + struct fdtable __rcu *fdt; struct fdtable fdtab; /* * written part on a separate cache line in SMP @@ -55,7 +55,7 @@ struct files_struct { int next_fd; struct embedded_fd_set close_on_exec_init; struct embedded_fd_set open_fds_init; - struct file * fd_array[NR_OPEN_DEFAULT]; + struct file __rcu * fd_array[NR_OPEN_DEFAULT]; }; #define rcu_dereference_check_fdtable(files, fdtfd) \ diff --git a/include/linux/fs.h b/include/linux/fs.h index 76041b6..aa3dc8d 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1380,7 +1380,7 @@ struct super_block { * Saved mount options for lazy filesystems using * generic_show_options() */ - char *s_options; + char __rcu *s_options; }; extern struct timespec current_fs_time(struct super_block *sb); diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 5f2f4c4..af3f06b 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -129,8 +129,8 @@ struct blk_scsi_cmd_filter { struct disk_part_tbl { struct rcu_head rcu_head; int len; - struct hd_struct *last_lookup; - struct hd_struct *part[]; + struct hd_struct __rcu *last_lookup; + struct hd_struct __rcu *part[]; }; struct gendisk { @@ -149,7 +149,7 @@ struct gendisk { * non-critical accesses use RCU. Always access through * helpers. */ - struct disk_part_tbl *part_tbl; + struct disk_part_tbl __rcu *part_tbl; struct hd_struct part0; const struct block_device_operations *fops; diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 1f43fa5..6460fc6 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -137,8 +137,8 @@ extern struct cred init_cred; .children = LIST_HEAD_INIT(tsk.children), \ .sibling = LIST_HEAD_INIT(tsk.sibling), \ .group_leader = &tsk, \ - .real_cred = &init_cred, \ - .cred = &init_cred, \ + RCU_INIT_POINTER(.real_cred, &init_cred), \ + RCU_INIT_POINTER(.cred, &init_cred), \ .cred_guard_mutex = \ __MUTEX_INITIALIZER(tsk.cred_guard_mutex), \ .comm = "swapper", \ diff --git a/include/linux/iocontext.h b/include/linux/iocontext.h index 64d5291..3e70b21 100644 --- a/include/linux/iocontext.h +++ b/include/linux/iocontext.h @@ -53,7 +53,7 @@ struct io_context { struct radix_tree_root radix_root; struct hlist_head cic_list; - void *ioc_data; + void __rcu *ioc_data; }; static inline struct io_context *ioc_task_link(struct io_context *ioc) diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index b8bb9a6..05537a5 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -299,7 +299,7 @@ struct mm_struct { * new_owner->mm == mm * new_owner->alloc_lock is held */ - struct task_struct *owner; + struct task_struct __rcu *owner; #endif #ifdef CONFIG_PROC_FS -- cgit v0.10.2 From 5e8067adfdbaf97039a97540765b1e16eb8d61cc Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Sat, 17 Apr 2010 08:48:41 -0400 Subject: rcu head remove init RCU heads really don't need to be initialized. Their state before call_rcu() really does not matter. We need to keep init/destroy_rcu_head_on_stack() though, since we want debugobjects to be able to keep track of these objects. Signed-off-by: Alexey Dobriyan Signed-off-by: Mathieu Desnoyers CC: David S. Miller CC: "Paul E. McKenney" CC: akpm@linux-foundation.org CC: mingo@elte.hu CC: laijs@cn.fujitsu.com CC: dipankar@in.ibm.com CC: josh@joshtriplett.org CC: dvhltc@us.ibm.com CC: niv@us.ibm.com CC: tglx@linutronix.de CC: peterz@infradead.org CC: rostedt@goodmis.org CC: Valdis.Kletnieks@vt.edu CC: dhowells@redhat.com CC: eric.dumazet@gmail.com CC: Alexey Dobriyan Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index 3e1b662..27b44b3 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -75,12 +75,6 @@ extern void rcu_init(void); #error "Unknown RCU implementation specified to kernel configuration" #endif -#define RCU_HEAD_INIT { .next = NULL, .func = NULL } -#define RCU_HEAD(head) struct rcu_head head = RCU_HEAD_INIT -#define INIT_RCU_HEAD(ptr) do { \ - (ptr)->next = NULL; (ptr)->func = NULL; \ -} while (0) - /* * init_rcu_head_on_stack()/destroy_rcu_head_on_stack() are needed for dynamic * initialization and destruction of rcu_head on the stack. rcu_head structures -- cgit v0.10.2 From 65e423f8ee5843e1ea3f2d94adf4ba3560a17f7b Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 19 May 2010 10:42:16 -0700 Subject: Update documentation to note the passage of INIT_RCU_HEAD() Signed-off-by: Alexey Dobriyan Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/Documentation/DocBook/kernel-locking.tmpl b/Documentation/DocBook/kernel-locking.tmpl index 084f6ad..e6cc574 100644 --- a/Documentation/DocBook/kernel-locking.tmpl +++ b/Documentation/DocBook/kernel-locking.tmpl @@ -1725,14 +1725,6 @@ the amount of locking which needs to be done. if (++cache_num > MAX_CACHE_SIZE) { struct object *i, *outcast = NULL; list_for_each_entry(i, &cache, list) { -@@ -85,6 +94,7 @@ - obj->popularity = 0; - atomic_set(&obj->refcnt, 1); /* The cache holds a reference */ - spin_lock_init(&obj->lock); -+ INIT_RCU_HEAD(&obj->rcu); - - spin_lock_irqsave(&cache_lock, flags); - __cache_add(obj); @@ -104,12 +114,11 @@ struct object *cache_find(int id) { -- cgit v0.10.2 From ded5e5ed2f3348ba2f9a319c6497e46c22850e97 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 19 May 2010 10:46:55 -0700 Subject: Update call_rcu() usage, add synchronize_rcu() Reported-by: Kyle Hubert Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett diff --git a/Documentation/DocBook/kernel-locking.tmpl b/Documentation/DocBook/kernel-locking.tmpl index e6cc574..ed64d22 100644 --- a/Documentation/DocBook/kernel-locking.tmpl +++ b/Documentation/DocBook/kernel-locking.tmpl @@ -1645,7 +1645,9 @@ the amount of locking which needs to be done. all the readers who were traversing the list when we deleted the element are finished. We use call_rcu() to register a callback which will actually destroy the object once - the readers are finished. + all pre-existing readers are finished. Alternatively, + synchronize_rcu() may be used to block until + all pre-existing are finished. But how does Read Copy Update know when the readers are @@ -1714,7 +1716,7 @@ the amount of locking which needs to be done. - object_put(obj); + list_del_rcu(&obj->list); cache_num--; -+ call_rcu(&obj->rcu, cache_delete_rcu, obj); ++ call_rcu(&obj->rcu, cache_delete_rcu); } /* Must be holding cache_lock */ -- cgit v0.10.2 From f2e0dd7090eddef427ab9d9f81de122244cded51 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 14 Jul 2010 14:38:30 -0700 Subject: rcu: allow RCU CPU stall warning messages to be controlled in /sys Set the permissions of the rcu_cpu_stall_suppress to 644 to enable RCU CPU stall warnings to be enabled and disabled at runtime via sysfs. Suggested-by: Josh Triplett Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutree.c b/kernel/rcutree.c index f3d5906..5d910be 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -145,7 +145,7 @@ module_param(qlowmark, int, 0); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR int rcu_cpu_stall_suppress __read_mostly; -module_param(rcu_cpu_stall_suppress, int, 0); +module_param(rcu_cpu_stall_suppress, int, 0644); #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ static void force_quiescent_state(struct rcu_state *rsp, int relaxed); -- cgit v0.10.2 From 687d7a960aea46e016182c7ce346d62c4dbd0366 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 21 Jul 2010 06:52:40 -0700 Subject: rcu: restrict TREE_RCU to SMP builds with !PREEMPT Because both TINY_RCU and TREE_PREEMPT_RCU have been in mainline for several releases, it is time to restrict the use of TREE_RCU to SMP non-preemptible systems. This reduces testing/validation effort. This commit is a first step towards driving the selection of RCU implementation directly off of the SMP and PREEMPT configuration parameters. Signed-off-by: Paul E. McKenney diff --git a/init/Kconfig b/init/Kconfig index 2de5b1c..79ac52b 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -340,6 +340,7 @@ choice config TREE_RCU bool "Tree-based hierarchical RCU" + depends on !PREEMPT && SMP help This option selects the RCU implementation that is designed for very large SMP system with hundreds or -- cgit v0.10.2 From 910b1b7e19a292ff685001caf1bf1a9775b771a1 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 21 Jul 2010 08:05:56 -0700 Subject: rcu: Allow RCU CPU stall warnings to be off at boot, but manually enablable Currently, if RCU CPU stall warnings are enabled, they are enabled immediately upon boot. They can be manually disabled via /sys (and also re-enabled via /sys), and are automatically disabled upon panic. However, some users need RCU CPU stalls to be disabled at boot time, but to be enabled without rebuilding/rebooting. For example, someone running a real-time application in production might not want the additional latency of RCU CPU stall detection in normal operation, but might need to enable it at any point for fault isolation purposes. This commit therefore provides a new CONFIG_RCU_CPU_STALL_DETECTOR_RUNNABLE kernel configuration parameter that maintains the current behavior (enable at boot) by default, but allows a kernel to be configured with RCU CPU stall detection built into the kernel, but disabled at boot time. Requested-by: Clark Williams Requested-by: John Kacur Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 5d910be..5aab7da 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -144,7 +144,7 @@ module_param(qhimark, int, 0); module_param(qlowmark, int, 0); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR -int rcu_cpu_stall_suppress __read_mostly; +int rcu_cpu_stall_suppress __read_mostly = RCU_CPU_STALL_SUPPRESS_INIT; module_param(rcu_cpu_stall_suppress, int, 0644); #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 183ebf4..bb4d086 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -264,6 +264,12 @@ struct rcu_data { /* scheduling clock irq */ /* before ratting on them. */ +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR_RUNNABLE +#define RCU_CPU_STALL_SUPPRESS_INIT 0 +#else +#define RCU_CPU_STALL_SUPPRESS_INIT 1 +#endif + #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ #define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index b5005b7..52c2172 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -856,6 +856,19 @@ config RCU_CPU_STALL_TIMEOUT RCU grace period persists, additional CPU stall warnings are printed at more widely spaced intervals. +config RCU_CPU_STALL_DETECTOR_RUNNABLE + bool "RCU CPU stall checking starts automatically at boot" + depends on RCU_CPU_STALL_DETECTOR + default y + help + If set, start checking for RCU CPU stalls immediately on + boot. Otherwise, RCU CPU stall checking must be manually + enabled. + + Say Y if you are unsure. + + Say N if you wish to suppress RCU CPU stall checking during boot. + config RCU_CPU_STALL_VERBOSE bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR" depends on RCU_CPU_STALL_DETECTOR && TREE_PREEMPT_RCU -- cgit v0.10.2 From 4d87ffadbba88105f33271bef5f2c79366c6a4e1 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 4 Aug 2010 17:31:12 -0700 Subject: rcu: Fix RCU_FANOUT help message Commit cf244dc01bf68 added a fourth level to the TREE_RCU hierarchy, but the RCU_FANOUT help message still said "cube root". This commit fixes this to "fourth root" and also emphasizes that production systems are well-served by the default. (Stress-testing RCU itself uses small RCU_FANOUT values in order to test large-system code paths on small(er) systems.) Located-by: John Kacur Signed-off-by: Paul E. McKenney diff --git a/init/Kconfig b/init/Kconfig index 79ac52b..dbc08ba 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -388,9 +388,12 @@ config RCU_FANOUT help This option controls the fanout of hierarchical implementations of RCU, allowing RCU to work efficiently on machines with - large numbers of CPUs. This value must be at least the cube - root of NR_CPUS, which allows NR_CPUS up to 32,768 for 32-bit - systems and up to 262,144 for 64-bit systems. + large numbers of CPUs. This value must be at least the fourth + root of NR_CPUS, which allows NR_CPUS to be insanely large. + The default value of RCU_FANOUT should be used for production + systems, but if you are stress-testing the RCU implementation + itself, small RCU_FANOUT values allow you to test large-system + code paths on small(er) systems. Select a specific number if testing RCU itself. Take the default if unsure. -- cgit v0.10.2 From 277b199800ac90811ac86d215063df1984f51619 Mon Sep 17 00:00:00 2001 From: Lin Ming Date: Fri, 20 Aug 2010 11:03:51 +0800 Subject: lockup_detector: Make callback function static watchdog_overflow_callback() is only used in kernel/watchdog.c. Signed-off-by: Lin Ming Cc: Peter Zijlstra Cc: Don Zickus LKML-Reference: <1282273431.16443.32.camel@minggr.sh.intel.com> Signed-off-by: Ingo Molnar diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 613bc1f..b60e2a8 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -202,7 +202,7 @@ static struct perf_event_attr wd_hw_attr = { }; /* Callback function for perf event subsystem */ -void watchdog_overflow_callback(struct perf_event *event, int nmi, +static void watchdog_overflow_callback(struct perf_event *event, int nmi, struct perf_sample_data *data, struct pt_regs *regs) { -- cgit v0.10.2 From a57eb940d130477a799dfb24a570ee04979c0f7f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 29 Jun 2010 16:49:16 -0700 Subject: rcu: Add a TINY_PREEMPT_RCU MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Implement a small-memory-footprint uniprocessor-only implementation of preemptible RCU. This implementation uses but a single blocked-tasks list rather than the combinatorial number used per leaf rcu_node by TREE_PREEMPT_RCU, which reduces memory consumption and greatly simplifies processing. This version also takes advantage of uniprocessor execution to accelerate grace periods in the case where there are no readers. The general design is otherwise broadly similar to that of TREE_PREEMPT_RCU. This implementation is a step towards having RCU implementation driven off of the SMP and PREEMPT kernel configuration variables, which can happen once this implementation has accumulated sufficient experience. Removed ACCESS_ONCE() from __rcu_read_unlock() and added barrier() as suggested by Steve Rostedt in order to avoid the compiler-reordering issue noted by Mathieu Desnoyers (http://lkml.org/lkml/2010/8/16/183). As can be seen below, CONFIG_TINY_PREEMPT_RCU represents almost 5Kbyte savings compared to CONFIG_TREE_PREEMPT_RCU. Of course, for non-real-time workloads, CONFIG_TINY_RCU is even better. CONFIG_TREE_PREEMPT_RCU text data bss dec filename 13 0 0 13 kernel/rcupdate.o 6170 825 28 7023 kernel/rcutree.o ---- 7026 Total CONFIG_TINY_PREEMPT_RCU text data bss dec filename 13 0 0 13 kernel/rcupdate.o 2081 81 8 2170 kernel/rcutiny.o ---- 2183 Total CONFIG_TINY_RCU (non-preemptible) text data bss dec filename 13 0 0 13 kernel/rcupdate.o 719 25 0 744 kernel/rcutiny.o --- 757 Total Requested-by: Loïc Minier Signed-off-by: Paul E. McKenney diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index d5b3876..1f4517d 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -139,7 +139,7 @@ static inline void account_system_vtime(struct task_struct *tsk) #endif #if defined(CONFIG_NO_HZ) -#if defined(CONFIG_TINY_RCU) +#if defined(CONFIG_TINY_RCU) || defined(CONFIG_TINY_PREEMPT_RCU) extern void rcu_enter_nohz(void); extern void rcu_exit_nohz(void); diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 6460fc6..2fea6c8 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -82,11 +82,17 @@ extern struct group_info init_groups; # define CAP_INIT_BSET CAP_FULL_SET #ifdef CONFIG_TREE_PREEMPT_RCU +#define INIT_TASK_RCU_TREE_PREEMPT() \ + .rcu_blocked_node = NULL, +#else +#define INIT_TASK_RCU_TREE_PREEMPT(tsk) +#endif +#ifdef CONFIG_PREEMPT_RCU #define INIT_TASK_RCU_PREEMPT(tsk) \ .rcu_read_lock_nesting = 0, \ .rcu_read_unlock_special = 0, \ - .rcu_blocked_node = NULL, \ - .rcu_node_entry = LIST_HEAD_INIT(tsk.rcu_node_entry), + .rcu_node_entry = LIST_HEAD_INIT(tsk.rcu_node_entry), \ + INIT_TASK_RCU_TREE_PREEMPT() #else #define INIT_TASK_RCU_PREEMPT(tsk) #endif diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index 27b44b3..24b8966 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -58,7 +58,6 @@ struct rcu_head { }; /* Exported common interfaces */ -extern void rcu_barrier(void); extern void rcu_barrier_bh(void); extern void rcu_barrier_sched(void); extern void synchronize_sched_expedited(void); @@ -69,7 +68,7 @@ extern void rcu_init(void); #if defined(CONFIG_TREE_RCU) || defined(CONFIG_TREE_PREEMPT_RCU) #include -#elif defined(CONFIG_TINY_RCU) +#elif defined(CONFIG_TINY_RCU) || defined(CONFIG_TINY_PREEMPT_RCU) #include #else #error "Unknown RCU implementation specified to kernel configuration" diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index e2e8931..4cc5eba 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -29,66 +29,51 @@ void rcu_sched_qs(int cpu); void rcu_bh_qs(int cpu); -static inline void rcu_note_context_switch(int cpu) -{ - rcu_sched_qs(cpu); -} +#ifdef CONFIG_TINY_RCU #define __rcu_read_lock() preempt_disable() #define __rcu_read_unlock() preempt_enable() +#else /* #ifdef CONFIG_TINY_RCU */ +void __rcu_read_lock(void); +void __rcu_read_unlock(void); +#endif /* #else #ifdef CONFIG_TINY_RCU */ #define __rcu_read_lock_bh() local_bh_disable() #define __rcu_read_unlock_bh() local_bh_enable() -#define call_rcu_sched call_rcu +extern void call_rcu_sched(struct rcu_head *head, + void (*func)(struct rcu_head *rcu)); #define rcu_init_sched() do { } while (0) -extern void rcu_check_callbacks(int cpu, int user); -static inline int rcu_needs_cpu(int cpu) -{ - return 0; -} +extern void synchronize_sched(void); -/* - * Return the number of grace periods. - */ -static inline long rcu_batches_completed(void) -{ - return 0; -} +#ifdef CONFIG_TINY_RCU -/* - * Return the number of bottom-half grace periods. - */ -static inline long rcu_batches_completed_bh(void) -{ - return 0; -} +#define call_rcu call_rcu_sched -static inline void rcu_force_quiescent_state(void) +static inline void synchronize_rcu(void) { + synchronize_sched(); } -static inline void rcu_bh_force_quiescent_state(void) +static inline void synchronize_rcu_expedited(void) { + synchronize_sched(); /* Only one CPU, so pretty fast anyway!!! */ } -static inline void rcu_sched_force_quiescent_state(void) +static inline void rcu_barrier(void) { + rcu_barrier_sched(); /* Only one CPU, so only one list of callbacks! */ } -extern void synchronize_sched(void); +#else /* #ifdef CONFIG_TINY_RCU */ -static inline void synchronize_rcu(void) -{ - synchronize_sched(); -} +void synchronize_rcu(void); +void rcu_barrier(void); +void synchronize_rcu_expedited(void); -static inline void synchronize_rcu_bh(void) -{ - synchronize_sched(); -} +#endif /* #else #ifdef CONFIG_TINY_RCU */ -static inline void synchronize_rcu_expedited(void) +static inline void synchronize_rcu_bh(void) { synchronize_sched(); } @@ -117,15 +102,82 @@ static inline void rcu_exit_nohz(void) #endif /* #else #ifdef CONFIG_NO_HZ */ +#ifdef CONFIG_TINY_RCU + +static inline void rcu_preempt_note_context_switch(void) +{ +} + static inline void exit_rcu(void) { } +static inline int rcu_needs_cpu(int cpu) +{ + return 0; +} + static inline int rcu_preempt_depth(void) { return 0; } +#else /* #ifdef CONFIG_TINY_RCU */ + +void rcu_preempt_note_context_switch(void); +extern void exit_rcu(void); +int rcu_preempt_needs_cpu(void); + +static inline int rcu_needs_cpu(int cpu) +{ + return rcu_preempt_needs_cpu(); +} + +/* + * Defined as macro as it is a very low level header + * included from areas that don't even know about current + * FIXME: combine with include/linux/rcutree.h into rcupdate.h. + */ +#define rcu_preempt_depth() (current->rcu_read_lock_nesting) + +#endif /* #else #ifdef CONFIG_TINY_RCU */ + +static inline void rcu_note_context_switch(int cpu) +{ + rcu_sched_qs(cpu); + rcu_preempt_note_context_switch(); +} + +extern void rcu_check_callbacks(int cpu, int user); + +/* + * Return the number of grace periods. + */ +static inline long rcu_batches_completed(void) +{ + return 0; +} + +/* + * Return the number of bottom-half grace periods. + */ +static inline long rcu_batches_completed_bh(void) +{ + return 0; +} + +static inline void rcu_force_quiescent_state(void) +{ +} + +static inline void rcu_bh_force_quiescent_state(void) +{ +} + +static inline void rcu_sched_force_quiescent_state(void) +{ +} + #ifdef CONFIG_DEBUG_LOCK_ALLOC extern int rcu_scheduler_active __read_mostly; diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h index c0ed1c0..c13b85d 100644 --- a/include/linux/rcutree.h +++ b/include/linux/rcutree.h @@ -95,6 +95,8 @@ static inline void synchronize_rcu_bh_expedited(void) synchronize_sched_expedited(); } +extern void rcu_barrier(void); + extern void rcu_check_callbacks(int cpu, int user); extern long rcu_batches_completed(void); diff --git a/include/linux/sched.h b/include/linux/sched.h index 2c75666..e18473f 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1202,11 +1202,13 @@ struct task_struct { unsigned int policy; cpumask_t cpus_allowed; -#ifdef CONFIG_TREE_PREEMPT_RCU +#ifdef CONFIG_PREEMPT_RCU int rcu_read_lock_nesting; char rcu_read_unlock_special; - struct rcu_node *rcu_blocked_node; struct list_head rcu_node_entry; +#endif /* #ifdef CONFIG_PREEMPT_RCU */ +#ifdef CONFIG_TREE_PREEMPT_RCU + struct rcu_node *rcu_blocked_node; #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) @@ -1740,7 +1742,7 @@ extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t * #define tsk_used_math(p) ((p)->flags & PF_USED_MATH) #define used_math() tsk_used_math(current) -#ifdef CONFIG_TREE_PREEMPT_RCU +#ifdef CONFIG_PREEMPT_RCU #define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */ #define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */ @@ -1749,7 +1751,9 @@ static inline void rcu_copy_process(struct task_struct *p) { p->rcu_read_lock_nesting = 0; p->rcu_read_unlock_special = 0; +#ifdef CONFIG_TREE_PREEMPT_RCU p->rcu_blocked_node = NULL; +#endif INIT_LIST_HEAD(&p->rcu_node_entry); } diff --git a/init/Kconfig b/init/Kconfig index dbc08ba..a619a1a 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -348,7 +348,7 @@ config TREE_RCU smaller systems. config TREE_PREEMPT_RCU - bool "Preemptable tree-based hierarchical RCU" + bool "Preemptible tree-based hierarchical RCU" depends on PREEMPT help This option selects the RCU implementation that is @@ -366,8 +366,22 @@ config TINY_RCU is not required. This option greatly reduces the memory footprint of RCU. +config TINY_PREEMPT_RCU + bool "Preemptible UP-only small-memory-footprint RCU" + depends on !SMP && PREEMPT + help + This option selects the RCU implementation that is designed + for real-time UP systems. This option greatly reduces the + memory footprint of RCU. + endchoice +config PREEMPT_RCU + def_bool ( TREE_PREEMPT_RCU || TINY_PREEMPT_RCU ) + help + This option enables preemptible-RCU code that is common between + the TREE_PREEMPT_RCU and TINY_PREEMPT_RCU implementations. + config RCU_TRACE bool "Enable tracing for RCU" depends on TREE_RCU || TREE_PREEMPT_RCU diff --git a/kernel/Makefile b/kernel/Makefile index 0b72d1a..17046b6 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -86,6 +86,7 @@ obj-$(CONFIG_TREE_RCU) += rcutree.o obj-$(CONFIG_TREE_PREEMPT_RCU) += rcutree.o obj-$(CONFIG_TREE_RCU_TRACE) += rcutree_trace.o obj-$(CONFIG_TINY_RCU) += rcutiny.o +obj-$(CONFIG_TINY_PREEMPT_RCU) += rcutiny.o obj-$(CONFIG_RELAY) += relay.o obj-$(CONFIG_SYSCTL) += utsname_sysctl.o obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index 196ec02..d806735 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -59,6 +59,14 @@ int rcu_scheduler_active __read_mostly; EXPORT_SYMBOL_GPL(rcu_scheduler_active); #endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ +/* Forward declarations for rcutiny_plugin.h. */ +static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp); +static void __call_rcu(struct rcu_head *head, + void (*func)(struct rcu_head *rcu), + struct rcu_ctrlblk *rcp); + +#include "rcutiny_plugin.h" + #ifdef CONFIG_NO_HZ static long rcu_dynticks_nesting = 1; @@ -140,6 +148,7 @@ void rcu_check_callbacks(int cpu, int user) rcu_sched_qs(cpu); else if (!in_softirq()) rcu_bh_qs(cpu); + rcu_preempt_check_callbacks(); } /* @@ -162,6 +171,7 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) *rcp->donetail = NULL; if (rcp->curtail == rcp->donetail) rcp->curtail = &rcp->rcucblist; + rcu_preempt_remove_callbacks(rcp); rcp->donetail = &rcp->rcucblist; local_irq_restore(flags); @@ -182,6 +192,7 @@ static void rcu_process_callbacks(struct softirq_action *unused) { __rcu_process_callbacks(&rcu_sched_ctrlblk); __rcu_process_callbacks(&rcu_bh_ctrlblk); + rcu_preempt_process_callbacks(); } /* @@ -223,15 +234,15 @@ static void __call_rcu(struct rcu_head *head, } /* - * Post an RCU callback to be invoked after the end of an RCU grace + * Post an RCU callback to be invoked after the end of an RCU-sched grace * period. But since we have but one CPU, that would be after any * quiescent state. */ -void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) +void call_rcu_sched(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) { __call_rcu(head, func, &rcu_sched_ctrlblk); } -EXPORT_SYMBOL_GPL(call_rcu); +EXPORT_SYMBOL_GPL(call_rcu_sched); /* * Post an RCU bottom-half callback to be invoked after any subsequent @@ -243,20 +254,6 @@ void call_rcu_bh(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) } EXPORT_SYMBOL_GPL(call_rcu_bh); -void rcu_barrier(void) -{ - struct rcu_synchronize rcu; - - init_rcu_head_on_stack(&rcu.head); - init_completion(&rcu.completion); - /* Will wake me after RCU finished. */ - call_rcu(&rcu.head, wakeme_after_rcu); - /* Wait for it. */ - wait_for_completion(&rcu.completion); - destroy_rcu_head_on_stack(&rcu.head); -} -EXPORT_SYMBOL_GPL(rcu_barrier); - void rcu_barrier_bh(void) { struct rcu_synchronize rcu; @@ -289,5 +286,3 @@ void __init rcu_init(void) { open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); } - -#include "rcutiny_plugin.h" diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index d223a92bc..e6bc1b4 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -1,7 +1,7 @@ /* - * Read-Copy Update mechanism for mutual exclusion (tree-based version) + * Read-Copy Update mechanism for mutual exclusion, the Bloatwatch edition * Internal non-public definitions that provide either classic - * or preemptable semantics. + * or preemptible semantics. * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by @@ -17,11 +17,587 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. * - * Copyright IBM Corporation, 2009 + * Copyright (c) 2010 Linaro * * Author: Paul E. McKenney */ +#ifdef CONFIG_TINY_PREEMPT_RCU + +#include + +/* FIXME: merge with definitions in kernel/rcutree.h. */ +#define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) +#define ULONG_CMP_LT(a, b) (ULONG_MAX / 2 < (a) - (b)) + +/* Global control variables for preemptible RCU. */ +struct rcu_preempt_ctrlblk { + struct rcu_ctrlblk rcb; /* curtail: ->next ptr of last CB for GP. */ + struct rcu_head **nexttail; + /* Tasks blocked in a preemptible RCU */ + /* read-side critical section while an */ + /* preemptible-RCU grace period is in */ + /* progress must wait for a later grace */ + /* period. This pointer points to the */ + /* ->next pointer of the last task that */ + /* must wait for a later grace period, or */ + /* to &->rcb.rcucblist if there is no */ + /* such task. */ + struct list_head blkd_tasks; + /* Tasks blocked in RCU read-side critical */ + /* section. Tasks are placed at the head */ + /* of this list and age towards the tail. */ + struct list_head *gp_tasks; + /* Pointer to the first task blocking the */ + /* current grace period, or NULL if there */ + /* is not such task. */ + struct list_head *exp_tasks; + /* Pointer to first task blocking the */ + /* current expedited grace period, or NULL */ + /* if there is no such task. If there */ + /* is no current expedited grace period, */ + /* then there cannot be any such task. */ + u8 gpnum; /* Current grace period. */ + u8 gpcpu; /* Last grace period blocked by the CPU. */ + u8 completed; /* Last grace period completed. */ + /* If all three are equal, RCU is idle. */ +}; + +static struct rcu_preempt_ctrlblk rcu_preempt_ctrlblk = { + .rcb.donetail = &rcu_preempt_ctrlblk.rcb.rcucblist, + .rcb.curtail = &rcu_preempt_ctrlblk.rcb.rcucblist, + .nexttail = &rcu_preempt_ctrlblk.rcb.rcucblist, + .blkd_tasks = LIST_HEAD_INIT(rcu_preempt_ctrlblk.blkd_tasks), +}; + +static int rcu_preempted_readers_exp(void); +static void rcu_report_exp_done(void); + +/* + * Return true if the CPU has not yet responded to the current grace period. + */ +static int rcu_cpu_cur_gp(void) +{ + return rcu_preempt_ctrlblk.gpcpu != rcu_preempt_ctrlblk.gpnum; +} + +/* + * Check for a running RCU reader. Because there is only one CPU, + * there can be but one running RCU reader at a time. ;-) + */ +static int rcu_preempt_running_reader(void) +{ + return current->rcu_read_lock_nesting; +} + +/* + * Check for preempted RCU readers blocking any grace period. + * If the caller needs a reliable answer, it must disable hard irqs. + */ +static int rcu_preempt_blocked_readers_any(void) +{ + return !list_empty(&rcu_preempt_ctrlblk.blkd_tasks); +} + +/* + * Check for preempted RCU readers blocking the current grace period. + * If the caller needs a reliable answer, it must disable hard irqs. + */ +static int rcu_preempt_blocked_readers_cgp(void) +{ + return rcu_preempt_ctrlblk.gp_tasks != NULL; +} + +/* + * Return true if another preemptible-RCU grace period is needed. + */ +static int rcu_preempt_needs_another_gp(void) +{ + return *rcu_preempt_ctrlblk.rcb.curtail != NULL; +} + +/* + * Return true if a preemptible-RCU grace period is in progress. + * The caller must disable hardirqs. + */ +static int rcu_preempt_gp_in_progress(void) +{ + return rcu_preempt_ctrlblk.completed != rcu_preempt_ctrlblk.gpnum; +} + +/* + * Record a preemptible-RCU quiescent state for the specified CPU. Note + * that this just means that the task currently running on the CPU is + * in a quiescent state. There might be any number of tasks blocked + * while in an RCU read-side critical section. + * + * Unlike the other rcu_*_qs() functions, callers to this function + * must disable irqs in order to protect the assignment to + * ->rcu_read_unlock_special. + * + * Because this is a single-CPU implementation, the only way a grace + * period can end is if the CPU is in a quiescent state. The reason is + * that a blocked preemptible-RCU reader can exit its critical section + * only if the CPU is running it at the time. Therefore, when the + * last task blocking the current grace period exits its RCU read-side + * critical section, neither the CPU nor blocked tasks will be stopping + * the current grace period. (In contrast, SMP implementations + * might have CPUs running in RCU read-side critical sections that + * block later grace periods -- but this is not possible given only + * one CPU.) + */ +static void rcu_preempt_cpu_qs(void) +{ + /* Record both CPU and task as having responded to current GP. */ + rcu_preempt_ctrlblk.gpcpu = rcu_preempt_ctrlblk.gpnum; + current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS; + + /* + * If there is no GP, or if blocked readers are still blocking GP, + * then there is nothing more to do. + */ + if (!rcu_preempt_gp_in_progress() || rcu_preempt_blocked_readers_cgp()) + return; + + /* Advance callbacks. */ + rcu_preempt_ctrlblk.completed = rcu_preempt_ctrlblk.gpnum; + rcu_preempt_ctrlblk.rcb.donetail = rcu_preempt_ctrlblk.rcb.curtail; + rcu_preempt_ctrlblk.rcb.curtail = rcu_preempt_ctrlblk.nexttail; + + /* If there are no blocked readers, next GP is done instantly. */ + if (!rcu_preempt_blocked_readers_any()) + rcu_preempt_ctrlblk.rcb.donetail = rcu_preempt_ctrlblk.nexttail; + + /* If there are done callbacks, make RCU_SOFTIRQ process them. */ + if (*rcu_preempt_ctrlblk.rcb.donetail != NULL) + raise_softirq(RCU_SOFTIRQ); +} + +/* + * Start a new RCU grace period if warranted. Hard irqs must be disabled. + */ +static void rcu_preempt_start_gp(void) +{ + if (!rcu_preempt_gp_in_progress() && rcu_preempt_needs_another_gp()) { + + /* Official start of GP. */ + rcu_preempt_ctrlblk.gpnum++; + + /* Any blocked RCU readers block new GP. */ + if (rcu_preempt_blocked_readers_any()) + rcu_preempt_ctrlblk.gp_tasks = + rcu_preempt_ctrlblk.blkd_tasks.next; + + /* If there is no running reader, CPU is done with GP. */ + if (!rcu_preempt_running_reader()) + rcu_preempt_cpu_qs(); + } +} + +/* + * We have entered the scheduler, and the current task might soon be + * context-switched away from. If this task is in an RCU read-side + * critical section, we will no longer be able to rely on the CPU to + * record that fact, so we enqueue the task on the blkd_tasks list. + * If the task started after the current grace period began, as recorded + * by ->gpcpu, we enqueue at the beginning of the list. Otherwise + * before the element referenced by ->gp_tasks (or at the tail if + * ->gp_tasks is NULL) and point ->gp_tasks at the newly added element. + * The task will dequeue itself when it exits the outermost enclosing + * RCU read-side critical section. Therefore, the current grace period + * cannot be permitted to complete until the ->gp_tasks pointer becomes + * NULL. + * + * Caller must disable preemption. + */ +void rcu_preempt_note_context_switch(void) +{ + struct task_struct *t = current; + unsigned long flags; + + local_irq_save(flags); /* must exclude scheduler_tick(). */ + if (rcu_preempt_running_reader() && + (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) { + + /* Possibly blocking in an RCU read-side critical section. */ + t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED; + + /* + * If this CPU has already checked in, then this task + * will hold up the next grace period rather than the + * current grace period. Queue the task accordingly. + * If the task is queued for the current grace period + * (i.e., this CPU has not yet passed through a quiescent + * state for the current grace period), then as long + * as that task remains queued, the current grace period + * cannot end. + */ + list_add(&t->rcu_node_entry, &rcu_preempt_ctrlblk.blkd_tasks); + if (rcu_cpu_cur_gp()) + rcu_preempt_ctrlblk.gp_tasks = &t->rcu_node_entry; + } + + /* + * Either we were not in an RCU read-side critical section to + * begin with, or we have now recorded that critical section + * globally. Either way, we can now note a quiescent state + * for this CPU. Again, if we were in an RCU read-side critical + * section, and if that critical section was blocking the current + * grace period, then the fact that the task has been enqueued + * means that current grace period continues to be blocked. + */ + rcu_preempt_cpu_qs(); + local_irq_restore(flags); +} + +/* + * Tiny-preemptible RCU implementation for rcu_read_lock(). + * Just increment ->rcu_read_lock_nesting, shared state will be updated + * if we block. + */ +void __rcu_read_lock(void) +{ + current->rcu_read_lock_nesting++; + barrier(); /* needed if we ever invoke rcu_read_lock in rcutiny.c */ +} +EXPORT_SYMBOL_GPL(__rcu_read_lock); + +/* + * Handle special cases during rcu_read_unlock(), such as needing to + * notify RCU core processing or task having blocked during the RCU + * read-side critical section. + */ +static void rcu_read_unlock_special(struct task_struct *t) +{ + int empty; + int empty_exp; + unsigned long flags; + struct list_head *np; + int special; + + /* + * NMI handlers cannot block and cannot safely manipulate state. + * They therefore cannot possibly be special, so just leave. + */ + if (in_nmi()) + return; + + local_irq_save(flags); + + /* + * If RCU core is waiting for this CPU to exit critical section, + * let it know that we have done so. + */ + special = t->rcu_read_unlock_special; + if (special & RCU_READ_UNLOCK_NEED_QS) + rcu_preempt_cpu_qs(); + + /* Hardware IRQ handlers cannot block. */ + if (in_irq()) { + local_irq_restore(flags); + return; + } + + /* Clean up if blocked during RCU read-side critical section. */ + if (special & RCU_READ_UNLOCK_BLOCKED) { + t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BLOCKED; + + /* + * Remove this task from the ->blkd_tasks list and adjust + * any pointers that might have been referencing it. + */ + empty = !rcu_preempt_blocked_readers_cgp(); + empty_exp = rcu_preempt_ctrlblk.exp_tasks == NULL; + np = t->rcu_node_entry.next; + if (np == &rcu_preempt_ctrlblk.blkd_tasks) + np = NULL; + list_del(&t->rcu_node_entry); + if (&t->rcu_node_entry == rcu_preempt_ctrlblk.gp_tasks) + rcu_preempt_ctrlblk.gp_tasks = np; + if (&t->rcu_node_entry == rcu_preempt_ctrlblk.exp_tasks) + rcu_preempt_ctrlblk.exp_tasks = np; + INIT_LIST_HEAD(&t->rcu_node_entry); + + /* + * If this was the last task on the current list, and if + * we aren't waiting on the CPU, report the quiescent state + * and start a new grace period if needed. + */ + if (!empty && !rcu_preempt_blocked_readers_cgp()) { + rcu_preempt_cpu_qs(); + rcu_preempt_start_gp(); + } + + /* + * If this was the last task on the expedited lists, + * then we need wake up the waiting task. + */ + if (!empty_exp && rcu_preempt_ctrlblk.exp_tasks == NULL) + rcu_report_exp_done(); + } + local_irq_restore(flags); +} + +/* + * Tiny-preemptible RCU implementation for rcu_read_unlock(). + * Decrement ->rcu_read_lock_nesting. If the result is zero (outermost + * rcu_read_unlock()) and ->rcu_read_unlock_special is non-zero, then + * invoke rcu_read_unlock_special() to clean up after a context switch + * in an RCU read-side critical section and other special cases. + */ +void __rcu_read_unlock(void) +{ + struct task_struct *t = current; + + barrier(); /* needed if we ever invoke rcu_read_unlock in rcutiny.c */ + --t->rcu_read_lock_nesting; + barrier(); /* decrement before load of ->rcu_read_unlock_special */ + if (t->rcu_read_lock_nesting == 0 && + unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) + rcu_read_unlock_special(t); +#ifdef CONFIG_PROVE_LOCKING + WARN_ON_ONCE(t->rcu_read_lock_nesting < 0); +#endif /* #ifdef CONFIG_PROVE_LOCKING */ +} +EXPORT_SYMBOL_GPL(__rcu_read_unlock); + +/* + * Check for a quiescent state from the current CPU. When a task blocks, + * the task is recorded in the rcu_preempt_ctrlblk structure, which is + * checked elsewhere. This is called from the scheduling-clock interrupt. + * + * Caller must disable hard irqs. + */ +static void rcu_preempt_check_callbacks(void) +{ + struct task_struct *t = current; + + if (!rcu_preempt_running_reader() && rcu_preempt_gp_in_progress()) + rcu_preempt_cpu_qs(); + if (&rcu_preempt_ctrlblk.rcb.rcucblist != + rcu_preempt_ctrlblk.rcb.donetail) + raise_softirq(RCU_SOFTIRQ); + if (rcu_preempt_gp_in_progress() && rcu_preempt_running_reader()) + t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS; +} + +/* + * TINY_PREEMPT_RCU has an extra callback-list tail pointer to + * update, so this is invoked from __rcu_process_callbacks() to + * handle that case. Of course, it is invoked for all flavors of + * RCU, but RCU callbacks can appear only on one of the lists, and + * neither ->nexttail nor ->donetail can possibly be NULL, so there + * is no need for an explicit check. + */ +static void rcu_preempt_remove_callbacks(struct rcu_ctrlblk *rcp) +{ + if (rcu_preempt_ctrlblk.nexttail == rcp->donetail) + rcu_preempt_ctrlblk.nexttail = &rcp->rcucblist; +} + +/* + * Process callbacks for preemptible RCU. + */ +static void rcu_preempt_process_callbacks(void) +{ + __rcu_process_callbacks(&rcu_preempt_ctrlblk.rcb); +} + +/* + * Queue a preemptible -RCU callback for invocation after a grace period. + */ +void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) +{ + unsigned long flags; + + debug_rcu_head_queue(head); + head->func = func; + head->next = NULL; + + local_irq_save(flags); + *rcu_preempt_ctrlblk.nexttail = head; + rcu_preempt_ctrlblk.nexttail = &head->next; + rcu_preempt_start_gp(); /* checks to see if GP needed. */ + local_irq_restore(flags); +} +EXPORT_SYMBOL_GPL(call_rcu); + +void rcu_barrier(void) +{ + struct rcu_synchronize rcu; + + init_rcu_head_on_stack(&rcu.head); + init_completion(&rcu.completion); + /* Will wake me after RCU finished. */ + call_rcu(&rcu.head, wakeme_after_rcu); + /* Wait for it. */ + wait_for_completion(&rcu.completion); + destroy_rcu_head_on_stack(&rcu.head); +} +EXPORT_SYMBOL_GPL(rcu_barrier); + +/* + * synchronize_rcu - wait until a grace period has elapsed. + * + * Control will return to the caller some time after a full grace + * period has elapsed, in other words after all currently executing RCU + * read-side critical sections have completed. RCU read-side critical + * sections are delimited by rcu_read_lock() and rcu_read_unlock(), + * and may be nested. + */ +void synchronize_rcu(void) +{ +#ifdef CONFIG_DEBUG_LOCK_ALLOC + if (!rcu_scheduler_active) + return; +#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + + WARN_ON_ONCE(rcu_preempt_running_reader()); + if (!rcu_preempt_blocked_readers_any()) + return; + + /* Once we get past the fastpath checks, same code as rcu_barrier(). */ + rcu_barrier(); +} +EXPORT_SYMBOL_GPL(synchronize_rcu); + +static DECLARE_WAIT_QUEUE_HEAD(sync_rcu_preempt_exp_wq); +static unsigned long sync_rcu_preempt_exp_count; +static DEFINE_MUTEX(sync_rcu_preempt_exp_mutex); + +/* + * Return non-zero if there are any tasks in RCU read-side critical + * sections blocking the current preemptible-RCU expedited grace period. + * If there is no preemptible-RCU expedited grace period currently in + * progress, returns zero unconditionally. + */ +static int rcu_preempted_readers_exp(void) +{ + return rcu_preempt_ctrlblk.exp_tasks != NULL; +} + +/* + * Report the exit from RCU read-side critical section for the last task + * that queued itself during or before the current expedited preemptible-RCU + * grace period. + */ +static void rcu_report_exp_done(void) +{ + wake_up(&sync_rcu_preempt_exp_wq); +} + +/* + * Wait for an rcu-preempt grace period, but expedite it. The basic idea + * is to rely in the fact that there is but one CPU, and that it is + * illegal for a task to invoke synchronize_rcu_expedited() while in a + * preemptible-RCU read-side critical section. Therefore, any such + * critical sections must correspond to blocked tasks, which must therefore + * be on the ->blkd_tasks list. So just record the current head of the + * list in the ->exp_tasks pointer, and wait for all tasks including and + * after the task pointed to by ->exp_tasks to drain. + */ +void synchronize_rcu_expedited(void) +{ + unsigned long flags; + struct rcu_preempt_ctrlblk *rpcp = &rcu_preempt_ctrlblk; + unsigned long snap; + + barrier(); /* ensure prior action seen before grace period. */ + + WARN_ON_ONCE(rcu_preempt_running_reader()); + + /* + * Acquire lock so that there is only one preemptible RCU grace + * period in flight. Of course, if someone does the expedited + * grace period for us while we are acquiring the lock, just leave. + */ + snap = sync_rcu_preempt_exp_count + 1; + mutex_lock(&sync_rcu_preempt_exp_mutex); + if (ULONG_CMP_LT(snap, sync_rcu_preempt_exp_count)) + goto unlock_mb_ret; /* Others did our work for us. */ + + local_irq_save(flags); + + /* + * All RCU readers have to already be on blkd_tasks because + * we cannot legally be executing in an RCU read-side critical + * section. + */ + + /* Snapshot current head of ->blkd_tasks list. */ + rpcp->exp_tasks = rpcp->blkd_tasks.next; + if (rpcp->exp_tasks == &rpcp->blkd_tasks) + rpcp->exp_tasks = NULL; + local_irq_restore(flags); + + /* Wait for tail of ->blkd_tasks list to drain. */ + if (rcu_preempted_readers_exp()) + wait_event(sync_rcu_preempt_exp_wq, + !rcu_preempted_readers_exp()); + + /* Clean up and exit. */ + barrier(); /* ensure expedited GP seen before counter increment. */ + sync_rcu_preempt_exp_count++; +unlock_mb_ret: + mutex_unlock(&sync_rcu_preempt_exp_mutex); + barrier(); /* ensure subsequent action seen after grace period. */ +} +EXPORT_SYMBOL_GPL(synchronize_rcu_expedited); + +/* + * Does preemptible RCU need the CPU to stay out of dynticks mode? + */ +int rcu_preempt_needs_cpu(void) +{ + if (!rcu_preempt_running_reader()) + rcu_preempt_cpu_qs(); + return rcu_preempt_ctrlblk.rcb.rcucblist != NULL; +} + +/* + * Check for a task exiting while in a preemptible -RCU read-side + * critical section, clean up if so. No need to issue warnings, + * as debug_check_no_locks_held() already does this if lockdep + * is enabled. + */ +void exit_rcu(void) +{ + struct task_struct *t = current; + + if (t->rcu_read_lock_nesting == 0) + return; + t->rcu_read_lock_nesting = 1; + rcu_read_unlock(); +} + +#else /* #ifdef CONFIG_TINY_PREEMPT_RCU */ + +/* + * Because preemptible RCU does not exist, it never has any callbacks + * to check. + */ +static void rcu_preempt_check_callbacks(void) +{ +} + +/* + * Because preemptible RCU does not exist, it never has any callbacks + * to remove. + */ +static void rcu_preempt_remove_callbacks(struct rcu_ctrlblk *rcp) +{ +} + +/* + * Because preemptible RCU does not exist, it never has any callbacks + * to process. + */ +static void rcu_preempt_process_callbacks(void) +{ +} + +#endif /* #else #ifdef CONFIG_TINY_PREEMPT_RCU */ + #ifdef CONFIG_DEBUG_LOCK_ALLOC #include -- cgit v0.10.2 From 9079fd7c2e06a92cf27d05224a1f478581916c5b Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Sat, 7 Aug 2010 21:59:54 -0700 Subject: rcu: update obsolete rcu_read_lock() comment. The comment says that blocking is illegal in rcu_read_lock()-style RCU read-side critical sections, which is no longer entirely true given preemptible RCU. This commit provides a fix. Suggested-by: David Miller Signed-off-by: Paul E. McKenney diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index 24b8966..d7af96e 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -458,7 +458,20 @@ extern int rcu_my_thread_group_empty(void); * will be deferred until the outermost RCU read-side critical section * completes. * - * It is illegal to block while in an RCU read-side critical section. + * You can avoid reading and understanding the next paragraph by + * following this rule: don't put anything in an rcu_read_lock() RCU + * read-side critical section that would block in a !PREEMPT kernel. + * But if you want the full story, read on! + * + * In non-preemptible RCU implementations (TREE_RCU and TINY_RCU), it + * is illegal to block while in an RCU read-side critical section. In + * preemptible RCU implementations (TREE_PREEMPT_RCU and TINY_PREEMPT_RCU) + * in CONFIG_PREEMPT kernel builds, RCU read-side critical sections may + * be preempted, but explicit blocking is illegal. Finally, in preemptible + * RCU implementations in real-time (CONFIG_PREEMPT_RT) kernel builds, + * RCU read-side critical sections may be preempted and they may also + * block, but only when acquiring spinlocks that are subject to priority + * inheritance. */ static inline void rcu_read_lock(void) { -- cgit v0.10.2 From 8cdd32a918350430483751feaae1c19cef816f69 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 9 Aug 2010 14:23:03 -0700 Subject: rcu: refer RCU CPU stall-warning victims to stallwarn.txt There is some documentation on RCU CPU stall warnings contained in Documentation/RCU/stallwarn.txt, but it will not be apparent to someone who runs into such a warning while under time pressure. This commit therefore adds comments preceding the printk()s pointing out the location of this documentation. Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 5aab7da..ff21411 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -487,8 +487,11 @@ static void print_other_cpu_stall(struct rcu_state *rsp) rcu_print_task_stall(rnp); raw_spin_unlock_irqrestore(&rnp->lock, flags); - /* OK, time to rat on our buddy... */ - + /* + * OK, time to rat on our buddy... + * See Documentation/RCU/stallwarn.txt for info on how to debug + * RCU CPU stall warnings. + */ printk(KERN_ERR "INFO: %s detected stalls on CPUs/tasks: {", rsp->name); rcu_for_each_leaf_node(rsp, rnp) { @@ -517,6 +520,11 @@ static void print_cpu_stall(struct rcu_state *rsp) unsigned long flags; struct rcu_node *rnp = rcu_get_root(rsp); + /* + * OK, time to rat on ourselves... + * See Documentation/RCU/stallwarn.txt for info on how to debug + * RCU CPU stall warnings. + */ printk(KERN_ERR "INFO: %s detected stall on CPU %d (t=%lu jiffies)\n", rsp->name, smp_processor_id(), jiffies - rsp->gp_start); trigger_all_cpu_backtrace(); -- cgit v0.10.2 From 53d84e004d5e8c018be395c4330dc72fd60bd13e Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 10 Aug 2010 14:28:53 -0700 Subject: rcu: permit suppressing current grace period's CPU stall warnings When using a kernel debugger, a long sojourn in the debugger can get you lots of RCU CPU stall warnings once you resume. This might not be helpful, especially if you are using the system console. This patch therefore allows RCU CPU stall warnings to be suppressed, but only for the duration of the current set of grace periods. This differs from Jason's original patch in that it adds support for tiny RCU and preemptible RCU, and uses a slightly different method for suppressing the RCU CPU stall warning messages. Signed-off-by: Jason Wessel Signed-off-by: Paul E. McKenney Tested-by: Jason Wessel diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index 4cc5eba..3fa1797 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -178,6 +178,10 @@ static inline void rcu_sched_force_quiescent_state(void) { } +static inline void rcu_cpu_stall_reset(void) +{ +} + #ifdef CONFIG_DEBUG_LOCK_ALLOC extern int rcu_scheduler_active __read_mostly; diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h index c13b85d..0726809 100644 --- a/include/linux/rcutree.h +++ b/include/linux/rcutree.h @@ -36,6 +36,7 @@ extern void rcu_sched_qs(int cpu); extern void rcu_bh_qs(int cpu); extern void rcu_note_context_switch(int cpu); extern int rcu_needs_cpu(int cpu); +extern void rcu_cpu_stall_reset(void); #ifdef CONFIG_TREE_PREEMPT_RCU diff --git a/kernel/rcutree.c b/kernel/rcutree.c index ff21411..42140a8 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -565,6 +565,22 @@ static int rcu_panic(struct notifier_block *this, unsigned long ev, void *ptr) return NOTIFY_DONE; } +/** + * rcu_cpu_stall_reset - prevent further stall warnings in current grace period + * + * Set the stall-warning timeout way off into the future, thus preventing + * any RCU CPU stall-warning messages from appearing in the current set of + * RCU grace periods. + * + * The caller must disable hard irqs. + */ +void rcu_cpu_stall_reset(void) +{ + rcu_sched_state.jiffies_stall = jiffies + ULONG_MAX / 2; + rcu_bh_state.jiffies_stall = jiffies + ULONG_MAX / 2; + rcu_preempt_stall_reset(); +} + static struct notifier_block rcu_panic_block = { .notifier_call = rcu_panic, }; @@ -584,6 +600,10 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp) { } +void rcu_cpu_stall_reset(void) +{ +} + static void __init check_cpu_stall_init(void) { } diff --git a/kernel/rcutree.h b/kernel/rcutree.h index bb4d086..7abd439 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -372,6 +372,7 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, #ifdef CONFIG_RCU_CPU_STALL_DETECTOR static void rcu_print_detail_task_stall(struct rcu_state *rsp); static void rcu_print_task_stall(struct rcu_node *rnp); +static void rcu_preempt_stall_reset(void); #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp); #ifdef CONFIG_HOTPLUG_CPU diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 63bb771..561410f 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -417,6 +417,16 @@ static void rcu_print_task_stall(struct rcu_node *rnp) } } +/* + * Suppress preemptible RCU's CPU stall warnings by pushing the + * time of the next stall-warning message comfortably far into the + * future. + */ +static void rcu_preempt_stall_reset(void) +{ + rcu_preempt_state.jiffies_stall = jiffies + ULONG_MAX / 2; +} + #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ /* @@ -867,6 +877,14 @@ static void rcu_print_task_stall(struct rcu_node *rnp) { } +/* + * Because preemptible RCU does not exist, there is no need to suppress + * its CPU stall warnings. + */ +static void rcu_preempt_stall_reset(void) +{ +} + #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ /* -- cgit v0.10.2 From a3dc3fb161f9b4066c0fce22db72638af8baf83b Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 13 Aug 2010 16:16:25 -0700 Subject: rcu: repair code-duplication FIXMEs Combine the duplicate definitions of ULONG_CMP_GE(), ULONG_CMP_LT(), and rcu_preempt_depth() into include/linux/rcupdate.h. Signed-off-by: Paul E. McKenney diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index d7af96e..325bad7 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -47,6 +47,9 @@ extern int rcutorture_runnable; /* for sysctl */ #endif /* #ifdef CONFIG_RCU_TORTURE_TEST */ +#define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) +#define ULONG_CMP_LT(a, b) (ULONG_MAX / 2 < (a) - (b)) + /** * struct rcu_head - callback structure for use with RCU * @next: next update requests in a list @@ -66,6 +69,18 @@ extern int sched_expedited_torture_stats(char *page); /* Internal to kernel */ extern void rcu_init(void); +#ifdef CONFIG_PREEMPT_RCU + +/* + * Defined as a macro as it is a very low level header included from + * areas that don't even know about current. This gives the rcu_read_lock() + * nesting depth, but makes sense only if CONFIG_PREEMPT_RCU -- in other + * types of kernel builds, the rcu_read_lock() nesting depth is unknowable. + */ +#define rcu_preempt_depth() (current->rcu_read_lock_nesting) + +#endif /* #ifdef CONFIG_PREEMPT_RCU */ + #if defined(CONFIG_TREE_RCU) || defined(CONFIG_TREE_PREEMPT_RCU) #include #elif defined(CONFIG_TINY_RCU) || defined(CONFIG_TINY_PREEMPT_RCU) diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index 3fa1797..c6b11dc 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -133,13 +133,6 @@ static inline int rcu_needs_cpu(int cpu) return rcu_preempt_needs_cpu(); } -/* - * Defined as macro as it is a very low level header - * included from areas that don't even know about current - * FIXME: combine with include/linux/rcutree.h into rcupdate.h. - */ -#define rcu_preempt_depth() (current->rcu_read_lock_nesting) - #endif /* #else #ifdef CONFIG_TINY_RCU */ static inline void rcu_note_context_switch(int cpu) diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h index 0726809..54a20c1 100644 --- a/include/linux/rcutree.h +++ b/include/linux/rcutree.h @@ -45,12 +45,6 @@ extern void __rcu_read_unlock(void); extern void synchronize_rcu(void); extern void exit_rcu(void); -/* - * Defined as macro as it is a very low level header - * included from areas that don't even know about current - */ -#define rcu_preempt_depth() (current->rcu_read_lock_nesting) - #else /* #ifdef CONFIG_TREE_PREEMPT_RCU */ static inline void __rcu_read_lock(void) diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index e6bc1b4..c5bea11 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -26,10 +26,6 @@ #include -/* FIXME: merge with definitions in kernel/rcutree.h. */ -#define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) -#define ULONG_CMP_LT(a, b) (ULONG_MAX / 2 < (a) - (b)) - /* Global control variables for preemptible RCU. */ struct rcu_preempt_ctrlblk { struct rcu_ctrlblk rcb; /* curtail: ->next ptr of last CB for GP. */ diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 7abd439..7918ba6 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -272,9 +272,6 @@ struct rcu_data { #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ -#define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) -#define ULONG_CMP_LT(a, b) (ULONG_MAX / 2 < (a) - (b)) - /* * RCU global state, including node hierarchy. This hierarchy is * represented in "heap" form in a dense array. The root (first level) -- cgit v0.10.2 From 5cc6517abdeccb6690b344a43b5ce8eaee82da3c Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 13 Aug 2010 16:34:22 -0700 Subject: rcu: document ways of stalling updates in low-memory situations Signed-off-by: Paul E. McKenney diff --git a/Documentation/RCU/checklist.txt b/Documentation/RCU/checklist.txt index c7c6788..0c134f8 100644 --- a/Documentation/RCU/checklist.txt +++ b/Documentation/RCU/checklist.txt @@ -218,13 +218,22 @@ over a rather long period of time, but improvements are always welcome! include: a. Keeping a count of the number of data-structure elements - used by the RCU-protected data structure, including those - waiting for a grace period to elapse. Enforce a limit - on this number, stalling updates as needed to allow - previously deferred frees to complete. - - Alternatively, limit only the number awaiting deferred - free rather than the total number of elements. + used by the RCU-protected data structure, including + those waiting for a grace period to elapse. Enforce a + limit on this number, stalling updates as needed to allow + previously deferred frees to complete. Alternatively, + limit only the number awaiting deferred free rather than + the total number of elements. + + One way to stall the updates is to acquire the update-side + mutex. (Don't try this with a spinlock -- other CPUs + spinning on the lock could prevent the grace period + from ever ending.) Another way to stall the updates + is for the updates to use a wrapper function around + the memory allocator, so that this wrapper function + simulates OOM when there is too much memory awaiting an + RCU grace period. There are of course many other + variations on this theme. b. Limiting update rate. For example, if updates occur only once per hour, then no explicit rate limiting is required, -- cgit v0.10.2 From 73d4da4d360136826b36f78f5cf72b29da82c8a6 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 16 Aug 2010 10:50:54 -0700 Subject: rcu: Upgrade srcu_read_lock() docbook about SRCU grace periods It is illegal to wait for an SRCU grace period while within the corresponding flavor of SRCU read-side critical section. Therefore, this commit updates the srcu_read_lock() docbook accordingly. Signed-off-by: Paul E. McKenney diff --git a/include/linux/srcu.h b/include/linux/srcu.h index 6f456a7..58971e8 100644 --- a/include/linux/srcu.h +++ b/include/linux/srcu.h @@ -139,7 +139,12 @@ static inline int srcu_read_lock_held(struct srcu_struct *sp) * @sp: srcu_struct in which to register the new reader. * * Enter an SRCU read-side critical section. Note that SRCU read-side - * critical sections may be nested. + * critical sections may be nested. However, it is illegal to + * call anything that waits on an SRCU grace period for the same + * srcu_struct, whether directly or indirectly. Please note that + * one way to indirectly wait on an SRCU grace period is to acquire + * a mutex that is held elsewhere while calling synchronize_srcu() or + * synchronize_srcu_expedited(). */ static inline int srcu_read_lock(struct srcu_struct *sp) __acquires(sp) { -- cgit v0.10.2 From 7b0b759b65247cbc66384a912be9acf8d4800636 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 17 Aug 2010 14:18:46 -0700 Subject: rcu: combine duplicate code, courtesy of CONFIG_PREEMPT_RCU The CONFIG_PREEMPT_RCU kernel configuration parameter was recently re-introduced, but as an indication of the type of RCU (preemptible vs. non-preemptible) instead of as selecting a given implementation. This commit uses CONFIG_PREEMPT_RCU to combine duplicate code from include/linux/rcutiny.h and include/linux/rcutree.h into include/linux/rcupdate.h. This commit also combines a few other pieces of duplicate code that have accumulated. Signed-off-by: Paul E. McKenney diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index 325bad7..89414d6 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -61,16 +61,30 @@ struct rcu_head { }; /* Exported common interfaces */ +extern void call_rcu_sched(struct rcu_head *head, + void (*func)(struct rcu_head *rcu)); +extern void synchronize_sched(void); extern void rcu_barrier_bh(void); extern void rcu_barrier_sched(void); extern void synchronize_sched_expedited(void); extern int sched_expedited_torture_stats(char *page); -/* Internal to kernel */ -extern void rcu_init(void); +static inline void __rcu_read_lock_bh(void) +{ + local_bh_disable(); +} + +static inline void __rcu_read_unlock_bh(void) +{ + local_bh_enable(); +} #ifdef CONFIG_PREEMPT_RCU +extern void __rcu_read_lock(void); +extern void __rcu_read_unlock(void); +void synchronize_rcu(void); + /* * Defined as a macro as it is a very low level header included from * areas that don't even know about current. This gives the rcu_read_lock() @@ -79,7 +93,53 @@ extern void rcu_init(void); */ #define rcu_preempt_depth() (current->rcu_read_lock_nesting) -#endif /* #ifdef CONFIG_PREEMPT_RCU */ +#else /* #ifdef CONFIG_PREEMPT_RCU */ + +static inline void __rcu_read_lock(void) +{ + preempt_disable(); +} + +static inline void __rcu_read_unlock(void) +{ + preempt_enable(); +} + +static inline void synchronize_rcu(void) +{ + synchronize_sched(); +} + +static inline int rcu_preempt_depth(void) +{ + return 0; +} + +#endif /* #else #ifdef CONFIG_PREEMPT_RCU */ + +/* Internal to kernel */ +extern void rcu_init(void); +extern void rcu_sched_qs(int cpu); +extern void rcu_bh_qs(int cpu); +extern void rcu_check_callbacks(int cpu, int user); +struct notifier_block; + +#ifdef CONFIG_NO_HZ + +extern void rcu_enter_nohz(void); +extern void rcu_exit_nohz(void); + +#else /* #ifdef CONFIG_NO_HZ */ + +static inline void rcu_enter_nohz(void) +{ +} + +static inline void rcu_exit_nohz(void) +{ +} + +#endif /* #else #ifdef CONFIG_NO_HZ */ #if defined(CONFIG_TREE_RCU) || defined(CONFIG_TREE_PREEMPT_RCU) #include @@ -626,6 +686,8 @@ struct rcu_synchronize { extern void wakeme_after_rcu(struct rcu_head *head); +#ifdef CONFIG_PREEMPT_RCU + /** * call_rcu() - Queue an RCU callback for invocation after a grace period. * @head: structure to be used for queueing the RCU updates. @@ -642,6 +704,13 @@ extern void wakeme_after_rcu(struct rcu_head *head); extern void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *head)); +#else /* #ifdef CONFIG_PREEMPT_RCU */ + +/* In classic RCU, call_rcu() is just call_rcu_sched(). */ +#define call_rcu call_rcu_sched + +#endif /* #else #ifdef CONFIG_PREEMPT_RCU */ + /** * call_rcu_bh() - Queue an RCU for invocation after a quicker grace period. * @head: structure to be used for queueing the RCU updates. diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index c6b11dc..13877cb 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -27,34 +27,10 @@ #include -void rcu_sched_qs(int cpu); -void rcu_bh_qs(int cpu); - -#ifdef CONFIG_TINY_RCU -#define __rcu_read_lock() preempt_disable() -#define __rcu_read_unlock() preempt_enable() -#else /* #ifdef CONFIG_TINY_RCU */ -void __rcu_read_lock(void); -void __rcu_read_unlock(void); -#endif /* #else #ifdef CONFIG_TINY_RCU */ -#define __rcu_read_lock_bh() local_bh_disable() -#define __rcu_read_unlock_bh() local_bh_enable() -extern void call_rcu_sched(struct rcu_head *head, - void (*func)(struct rcu_head *rcu)); - #define rcu_init_sched() do { } while (0) -extern void synchronize_sched(void); - #ifdef CONFIG_TINY_RCU -#define call_rcu call_rcu_sched - -static inline void synchronize_rcu(void) -{ - synchronize_sched(); -} - static inline void synchronize_rcu_expedited(void) { synchronize_sched(); /* Only one CPU, so pretty fast anyway!!! */ @@ -67,7 +43,6 @@ static inline void rcu_barrier(void) #else /* #ifdef CONFIG_TINY_RCU */ -void synchronize_rcu(void); void rcu_barrier(void); void synchronize_rcu_expedited(void); @@ -83,25 +58,6 @@ static inline void synchronize_rcu_bh_expedited(void) synchronize_sched(); } -struct notifier_block; - -#ifdef CONFIG_NO_HZ - -extern void rcu_enter_nohz(void); -extern void rcu_exit_nohz(void); - -#else /* #ifdef CONFIG_NO_HZ */ - -static inline void rcu_enter_nohz(void) -{ -} - -static inline void rcu_exit_nohz(void) -{ -} - -#endif /* #else #ifdef CONFIG_NO_HZ */ - #ifdef CONFIG_TINY_RCU static inline void rcu_preempt_note_context_switch(void) @@ -117,11 +73,6 @@ static inline int rcu_needs_cpu(int cpu) return 0; } -static inline int rcu_preempt_depth(void) -{ - return 0; -} - #else /* #ifdef CONFIG_TINY_RCU */ void rcu_preempt_note_context_switch(void); @@ -141,8 +92,6 @@ static inline void rcu_note_context_switch(int cpu) rcu_preempt_note_context_switch(); } -extern void rcu_check_callbacks(int cpu, int user); - /* * Return the number of grace periods. */ diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h index 54a20c1..95518e6 100644 --- a/include/linux/rcutree.h +++ b/include/linux/rcutree.h @@ -30,59 +30,23 @@ #ifndef __LINUX_RCUTREE_H #define __LINUX_RCUTREE_H -struct notifier_block; - -extern void rcu_sched_qs(int cpu); -extern void rcu_bh_qs(int cpu); extern void rcu_note_context_switch(int cpu); extern int rcu_needs_cpu(int cpu); extern void rcu_cpu_stall_reset(void); #ifdef CONFIG_TREE_PREEMPT_RCU -extern void __rcu_read_lock(void); -extern void __rcu_read_unlock(void); -extern void synchronize_rcu(void); extern void exit_rcu(void); #else /* #ifdef CONFIG_TREE_PREEMPT_RCU */ -static inline void __rcu_read_lock(void) -{ - preempt_disable(); -} - -static inline void __rcu_read_unlock(void) -{ - preempt_enable(); -} - -#define synchronize_rcu synchronize_sched - static inline void exit_rcu(void) { } -static inline int rcu_preempt_depth(void) -{ - return 0; -} - #endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */ -static inline void __rcu_read_lock_bh(void) -{ - local_bh_disable(); -} -static inline void __rcu_read_unlock_bh(void) -{ - local_bh_enable(); -} - -extern void call_rcu_sched(struct rcu_head *head, - void (*func)(struct rcu_head *rcu)); extern void synchronize_rcu_bh(void); -extern void synchronize_sched(void); extern void synchronize_rcu_expedited(void); static inline void synchronize_rcu_bh_expedited(void) @@ -92,8 +56,6 @@ static inline void synchronize_rcu_bh_expedited(void) extern void rcu_barrier(void); -extern void rcu_check_callbacks(int cpu, int user); - extern long rcu_batches_completed(void); extern long rcu_batches_completed_bh(void); extern long rcu_batches_completed_sched(void); @@ -101,18 +63,6 @@ extern void rcu_force_quiescent_state(void); extern void rcu_bh_force_quiescent_state(void); extern void rcu_sched_force_quiescent_state(void); -#ifdef CONFIG_NO_HZ -void rcu_enter_nohz(void); -void rcu_exit_nohz(void); -#else /* CONFIG_NO_HZ */ -static inline void rcu_enter_nohz(void) -{ -} -static inline void rcu_exit_nohz(void) -{ -} -#endif /* CONFIG_NO_HZ */ - /* A context switch is a grace period for RCU-sched and RCU-bh. */ static inline int rcu_blocking_is_gp(void) { diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 561410f..87f60f0 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -939,15 +939,6 @@ static void rcu_preempt_process_callbacks(void) } /* - * In classic RCU, call_rcu() is just call_rcu_sched(). - */ -void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) -{ - call_rcu_sched(head, func); -} -EXPORT_SYMBOL_GPL(call_rcu); - -/* * Wait for an rcu-preempt grace period, but make it happen quickly. * But because preemptable RCU does not exist, map to rcu-sched. */ -- cgit v0.10.2 From 80dcf60e6b97c7363971e7a0a788d8484d35f8a6 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 19 Aug 2010 16:57:45 -0700 Subject: rcu: apply TINY_PREEMPT_RCU read-side speedup to TREE_PREEMPT_RCU Replace one of the ACCESS_ONCE() calls in each of __rcu_read_lock() and __rcu_read_unlock() with barrier() as suggested by Steve Rostedt in order to avoid the potential compiler-optimization-induced bug noted by Mathieu Desnoyers. Located-by: Mathieu Desnoyers Suggested-by: Steven Rostedt Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 87f60f0..e9e0bc74 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -201,7 +201,7 @@ static void rcu_preempt_note_context_switch(int cpu) */ void __rcu_read_lock(void) { - ACCESS_ONCE(current->rcu_read_lock_nesting)++; + current->rcu_read_lock_nesting++; barrier(); /* needed if we ever invoke rcu_read_lock in rcutree.c */ } EXPORT_SYMBOL_GPL(__rcu_read_lock); @@ -344,7 +344,9 @@ void __rcu_read_unlock(void) struct task_struct *t = current; barrier(); /* needed if we ever invoke rcu_read_unlock in rcutree.c */ - if (--ACCESS_ONCE(t->rcu_read_lock_nesting) == 0 && + --t->rcu_read_lock_nesting; + barrier(); /* decrement before load of ->rcu_read_unlock_special */ + if (t->rcu_read_lock_nesting == 0 && unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) rcu_read_unlock_special(t); #ifdef CONFIG_PROVE_LOCKING -- cgit v0.10.2 From 65e6bf484c497f02d47a0faae69ee398cd59cfda Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 19 Aug 2010 21:43:09 -0700 Subject: rcu: add comment stating that list_empty() applies to RCU-protected lists Because list_empty() does not dereference any RCU-protected pointers, and further does not pass such pointers to the caller (so that the caller does not dereference them either), it is safe to use list_empty() on RCU-protected lists. There is no need for a list_empty_rcu(). This commit adds a comment stating this explicitly. Requested-by: Andrew Morton Signed-off-by: Paul E. McKenney diff --git a/include/linux/rculist.h b/include/linux/rculist.h index c10b105..f31ef61 100644 --- a/include/linux/rculist.h +++ b/include/linux/rculist.h @@ -10,6 +10,15 @@ #include /* + * Why is there no list_empty_rcu()? Because list_empty() serves this + * purpose. The list_empty() function fetches the RCU-protected pointer + * and compares it to the address of the list head, but neither dereferences + * this pointer itself nor provides this pointer to the caller. Therefore, + * it is not necessary to use rcu_dereference(), so that list_empty() can + * be used anywhere you would want to use a list_empty_rcu(). + */ + +/* * return the ->next pointer of a list_head in an rcu safe * way, we must not access it directly */ -- cgit v0.10.2 From 28457ee69c473a903e51e26c7bcd6f1e9eceb93e Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 9 Mar 2010 19:24:45 +0100 Subject: vhost: add __rcu annotations Also add rcu_dereference_protected() for code paths where locks are held. Signed-off-by: Arnd Bergmann Signed-off-by: Paul E. McKenney Cc: "Michael S. Tsirkin" diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index 29e850a..1318ee0 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -127,7 +127,10 @@ static void handle_tx(struct vhost_net *net) size_t len, total_len = 0; int err, wmem; size_t hdr_size; - struct socket *sock = rcu_dereference(vq->private_data); + struct socket *sock; + + sock = rcu_dereference_check(vq->private_data, + lockdep_is_held(&vq->mutex)); if (!sock) return; @@ -582,7 +585,10 @@ static void vhost_net_disable_vq(struct vhost_net *n, static void vhost_net_enable_vq(struct vhost_net *n, struct vhost_virtqueue *vq) { - struct socket *sock = vq->private_data; + struct socket *sock; + + sock = rcu_dereference_protected(vq->private_data, + lockdep_is_held(&vq->mutex)); if (!sock) return; if (vq == n->vqs + VHOST_NET_VQ_TX) { @@ -598,7 +604,8 @@ static struct socket *vhost_net_stop_vq(struct vhost_net *n, struct socket *sock; mutex_lock(&vq->mutex); - sock = vq->private_data; + sock = rcu_dereference_protected(vq->private_data, + lockdep_is_held(&vq->mutex)); vhost_net_disable_vq(n, vq); rcu_assign_pointer(vq->private_data, NULL); mutex_unlock(&vq->mutex); @@ -736,7 +743,8 @@ static long vhost_net_set_backend(struct vhost_net *n, unsigned index, int fd) } /* start polling new socket */ - oldsock = vq->private_data; + oldsock = rcu_dereference_protected(vq->private_data, + lockdep_is_held(&vq->mutex)); if (sock != oldsock) { vhost_net_disable_vq(n, vq); rcu_assign_pointer(vq->private_data, sock); diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index e05557d..b5c4947 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -284,7 +284,7 @@ long vhost_dev_reset_owner(struct vhost_dev *dev) vhost_dev_cleanup(dev); memory->nregions = 0; - dev->memory = memory; + RCU_INIT_POINTER(dev->memory, memory); return 0; } @@ -316,8 +316,9 @@ void vhost_dev_cleanup(struct vhost_dev *dev) fput(dev->log_file); dev->log_file = NULL; /* No one will access memory at this point */ - kfree(dev->memory); - dev->memory = NULL; + kfree(rcu_dereference_protected(dev->memory, + lockdep_is_held(&dev->mutex))); + RCU_INIT_POINTER(dev->memory, NULL); if (dev->mm) mmput(dev->mm); dev->mm = NULL; @@ -401,14 +402,22 @@ static int vq_access_ok(unsigned int num, /* Caller should have device mutex but not vq mutex */ int vhost_log_access_ok(struct vhost_dev *dev) { - return memory_access_ok(dev, dev->memory, 1); + struct vhost_memory *mp; + + mp = rcu_dereference_protected(dev->memory, + lockdep_is_held(&dev->mutex)); + return memory_access_ok(dev, mp, 1); } /* Verify access for write logging. */ /* Caller should have vq mutex and device mutex */ static int vq_log_access_ok(struct vhost_virtqueue *vq, void __user *log_base) { - return vq_memory_access_ok(log_base, vq->dev->memory, + struct vhost_memory *mp; + + mp = rcu_dereference_protected(vq->dev->memory, + lockdep_is_held(&vq->mutex)); + return vq_memory_access_ok(log_base, mp, vhost_has_feature(vq->dev, VHOST_F_LOG_ALL)) && (!vq->log_used || log_access_ok(log_base, vq->log_addr, sizeof *vq->used + @@ -448,7 +457,8 @@ static long vhost_set_memory(struct vhost_dev *d, struct vhost_memory __user *m) kfree(newmem); return -EFAULT; } - oldmem = d->memory; + oldmem = rcu_dereference_protected(d->memory, + lockdep_is_held(&d->mutex)); rcu_assign_pointer(d->memory, newmem); synchronize_rcu(); kfree(oldmem); diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h index afd7729..af3c11d 100644 --- a/drivers/vhost/vhost.h +++ b/drivers/vhost/vhost.h @@ -106,7 +106,7 @@ struct vhost_virtqueue { * vhost_work execution acts instead of rcu_read_lock() and the end of * vhost_work execution acts instead of rcu_read_lock(). * Writers use virtqueue mutex. */ - void *private_data; + void __rcu *private_data; /* Log write descriptors */ void __user *log_base; struct vhost_log log[VHOST_NET_MAX_SG]; @@ -116,7 +116,7 @@ struct vhost_dev { /* Readers use RCU to access memory table pointer * log base pointer and features. * Writers use mutex below.*/ - struct vhost_memory *memory; + struct vhost_memory __rcu *memory; struct mm_struct *mm; struct mutex mutex; unsigned acked_features; @@ -173,7 +173,11 @@ enum { static inline int vhost_has_feature(struct vhost_dev *dev, int bit) { - unsigned acked_features = rcu_dereference(dev->acked_features); + unsigned acked_features; + + acked_features = + rcu_dereference_index_check(dev->acked_features, + lockdep_is_held(&dev->mutex)); return acked_features & (1 << bit); } -- cgit v0.10.2 From 61c77326d1df079f202fa79403c3ccd8c5966a81 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Mon, 16 Aug 2010 09:16:55 +0800 Subject: x86, mm: Avoid unnecessary TLB flush In x86, access and dirty bits are set automatically by CPU when CPU accesses memory. When we go into the code path of below flush_tlb_fix_spurious_fault(), we already set dirty bit for pte and don't need flush tlb. This might mean tlb entry in some CPUs hasn't dirty bit set, but this doesn't matter. When the CPUs do page write, they will automatically check the bit and no software involved. On the other hand, flush tlb in below position is harmful. Test creates CPU number of threads, each thread writes to a same but random address in same vma range and we measure the total time. Under a 4 socket system, original time is 1.96s, while with the patch, the time is 0.8s. Under a 2 socket system, there is 20% time cut too. perf shows a lot of time are taking to send ipi/handle ipi for tlb flush. Signed-off-by: Shaohua Li LKML-Reference: <20100816011655.GA362@sli10-desk.sh.intel.com> Acked-by: Suresh Siddha Cc: Andrea Archangeli Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index a34c785..2d0a33b 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -603,6 +603,8 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, pte_update(mm, addr, ptep); } +#define flush_tlb_fix_spurious_fault(vma, address) + /* * clone_pgd_range(pgd_t *dst, pgd_t *src, int count); * diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h index e2bd73e..f4d4120 100644 --- a/include/asm-generic/pgtable.h +++ b/include/asm-generic/pgtable.h @@ -129,6 +129,10 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addres #define move_pte(pte, prot, old_addr, new_addr) (pte) #endif +#ifndef flush_tlb_fix_spurious_fault +#define flush_tlb_fix_spurious_fault(vma, address) flush_tlb_page(vma, address) +#endif + #ifndef pgprot_noncached #define pgprot_noncached(prot) (prot) #endif diff --git a/mm/memory.c b/mm/memory.c index 2ed2267..a40da69 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -3147,7 +3147,7 @@ static inline int handle_pte_fault(struct mm_struct *mm, * with threads. */ if (flags & FAULT_FLAG_WRITE) - flush_tlb_page(vma, address); + flush_tlb_fix_spurious_fault(vma, address); } unlock: pte_unmap_unlock(pte, ptl); -- cgit v0.10.2 From fdf4289679fd41d76553ce224750e9737cd80eea Mon Sep 17 00:00:00 2001 From: "Ma, Ling" Date: Mon, 23 Aug 2010 14:11:12 -0700 Subject: x86, mem: Don't implement forward memmove() as memcpy() memmove() allow source and destination address to be overlap, but there is no such limitation for memcpy(). Therefore, explicitly implement memmove() in both the forwards and backward directions, to give us the ability to optimize memcpy(). Signed-off-by: Ma Ling LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/arch/x86/lib/memcpy_32.c b/arch/x86/lib/memcpy_32.c index 5415a9d..be424df 100644 --- a/arch/x86/lib/memcpy_32.c +++ b/arch/x86/lib/memcpy_32.c @@ -25,19 +25,35 @@ void *memmove(void *dest, const void *src, size_t n) int d0, d1, d2; if (dest < src) { - memcpy(dest, src, n); + if ((dest + n) < src) + return memcpy(dest, src, n); + else + __asm__ __volatile__( + "rep\n\t" + "movsb\n\t" + : "=&c" (d0), "=&S" (d1), "=&D" (d2) + :"0" (n), + "1" (src), + "2" (dest) + :"memory"); + } else { - __asm__ __volatile__( - "std\n\t" - "rep\n\t" - "movsb\n\t" - "cld" - : "=&c" (d0), "=&S" (d1), "=&D" (d2) - :"0" (n), - "1" (n-1+src), - "2" (n-1+dest) - :"memory"); + + if((src + count) < dest) + return memcpy(dest, src, count); + else + __asm__ __volatile__( + "std\n\t" + "rep\n\t" + "movsb\n\t" + "cld" + : "=&c" (d0), "=&S" (d1), "=&D" (d2) + :"0" (n), + "1" (n-1+src), + "2" (n-1+dest) + :"memory"); } + return dest; } EXPORT_SYMBOL(memmove); diff --git a/arch/x86/lib/memmove_64.c b/arch/x86/lib/memmove_64.c index 0a33909..ecacc4b 100644 --- a/arch/x86/lib/memmove_64.c +++ b/arch/x86/lib/memmove_64.c @@ -8,13 +8,49 @@ #undef memmove void *memmove(void *dest, const void *src, size_t count) { + unsigned long d0, d1, d2, d3; if (dest < src) { - return memcpy(dest, src, count); + if ((dest + count) < src) + return memcpy(dest, src, count); + else + __asm__ __volatile__( + "movq %0, %3\n\t" + "shr $3, %0\n\t" + "andq $7, %3\n\t" + "rep\n\t" + "movsq\n\t" + "movq %3, %0\n\t" + "rep\n\t" + "movsb" + : "=&c" (d0), "=&S" (d1), "=&D" (d2), "=r" (d3) + :"0" (count), + "1" (src), + "2" (dest) + :"memory"); } else { - char *p = dest + count; - const char *s = src + count; - while (count--) - *--p = *--s; + if((src + count) < dest) + return memcpy(dest, src, count); + else + __asm__ __volatile__( + "movq %0, %3\n\t" + "lea -8(%1, %0), %1\n\t" + "lea -8(%2, %0), %2\n\t" + "shr $3, %0\n\t" + "andq $7, %3\n\t" + "std\n\t" + "rep\n\t" + "movsq\n\t" + "lea 7(%1), %1\n\t" + "lea 7(%2), %2\n\t" + "movq %3, %0\n\t" + "rep\n\t" + "movsb\n\t" + "cld" + : "=&c" (d0), "=&S" (d1), "=&D" (d2), "=r" (d3) + :"0" (count), + "1" (src), + "2" (dest) + :"memory"); } return dest; } -- cgit v0.10.2 From 59daa706fbec745684702741b9f5373142dd9fdc Mon Sep 17 00:00:00 2001 From: Ma Ling Date: Tue, 29 Jun 2010 03:24:25 +0800 Subject: x86, mem: Optimize memcpy by avoiding memory false dependece All read operations after allocation stage can run speculatively, all write operation will run in program order, and if addresses are different read may run before older write operation, otherwise wait until write commit. However CPU don't check each address bit, so read could fail to recognize different address even they are in different page.For example if rsi is 0xf004, rdi is 0xe008, in following operation there will generate big performance latency. 1. movq (%rsi), %rax 2. movq %rax, (%rdi) 3. movq 8(%rsi), %rax 4. movq %rax, 8(%rdi) If %rsi and rdi were in really the same meory page, there are TRUE read-after-write dependence because instruction 2 write 0x008 and instruction 3 read 0x00c, the two address are overlap partially. Actually there are in different page and no any issues, but without checking each address bit CPU could think they are in the same page, and instruction 3 have to wait for instruction 2 to write data into cache from write buffer, then load data from cache, the cost time read spent is equal to mfence instruction. We may avoid it by tuning operation sequence as follow. 1. movq 8(%rsi), %rax 2. movq %rax, 8(%rdi) 3. movq (%rsi), %rax 4. movq %rax, (%rdi) Instruction 3 read 0x004, instruction 2 write address 0x010, no any dependence. At last on Core2 we gain 1.83x speedup compared with original instruction sequence. In this patch we first handle small size(less 20bytes), then jump to different copy mode. Based on our micro-benchmark small bytes from 1 to 127 bytes, we got up to 2X improvement, and up to 1.5X improvement for 1024 bytes on Corei7. (We use our micro-benchmark, and will do further test according to your requirment) Signed-off-by: Ma Ling LKML-Reference: <1277753065-18610-1-git-send-email-ling.ma@intel.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/lib/memcpy_32.c b/arch/x86/lib/memcpy_32.c index be424df..81130d4 100644 --- a/arch/x86/lib/memcpy_32.c +++ b/arch/x86/lib/memcpy_32.c @@ -36,11 +36,9 @@ void *memmove(void *dest, const void *src, size_t n) "1" (src), "2" (dest) :"memory"); - } else { - - if((src + count) < dest) - return memcpy(dest, src, count); + if((src + n) < dest) + return memcpy(dest, src, n); else __asm__ __volatile__( "std\n\t" diff --git a/arch/x86/lib/memcpy_64.S b/arch/x86/lib/memcpy_64.S index bcbcd1e..75ef61e 100644 --- a/arch/x86/lib/memcpy_64.S +++ b/arch/x86/lib/memcpy_64.S @@ -40,84 +40,132 @@ ENTRY(__memcpy) ENTRY(memcpy) CFI_STARTPROC + movq %rdi, %rax /* - * Put the number of full 64-byte blocks into %ecx. - * Tail portion is handled at the end: + * Use 32bit CMP here to avoid long NOP padding. */ - movq %rdi, %rax - movl %edx, %ecx - shrl $6, %ecx - jz .Lhandle_tail + cmp $0x20, %edx + jb .Lhandle_tail - .p2align 4 -.Lloop_64: /* - * We decrement the loop index here - and the zero-flag is - * checked at the end of the loop (instructions inbetween do - * not change the zero flag): + * We check whether memory false dependece could occur, + * then jump to corresponding copy mode. */ - decl %ecx + cmp %dil, %sil + jl .Lcopy_backward + subl $0x20, %edx +.Lcopy_forward_loop: + subq $0x20, %rdx /* - * Move in blocks of 4x16 bytes: + * Move in blocks of 4x8 bytes: */ - movq 0*8(%rsi), %r11 - movq 1*8(%rsi), %r8 - movq %r11, 0*8(%rdi) - movq %r8, 1*8(%rdi) - - movq 2*8(%rsi), %r9 - movq 3*8(%rsi), %r10 - movq %r9, 2*8(%rdi) - movq %r10, 3*8(%rdi) - - movq 4*8(%rsi), %r11 - movq 5*8(%rsi), %r8 - movq %r11, 4*8(%rdi) - movq %r8, 5*8(%rdi) - - movq 6*8(%rsi), %r9 - movq 7*8(%rsi), %r10 - movq %r9, 6*8(%rdi) - movq %r10, 7*8(%rdi) - - leaq 64(%rsi), %rsi - leaq 64(%rdi), %rdi - - jnz .Lloop_64 + movq 0*8(%rsi), %r8 + movq 1*8(%rsi), %r9 + movq 2*8(%rsi), %r10 + movq 3*8(%rsi), %r11 + leaq 4*8(%rsi), %rsi + + movq %r8, 0*8(%rdi) + movq %r9, 1*8(%rdi) + movq %r10, 2*8(%rdi) + movq %r11, 3*8(%rdi) + leaq 4*8(%rdi), %rdi + jae .Lcopy_forward_loop + addq $0x20, %rdx + jmp .Lhandle_tail + +.Lcopy_backward: + /* + * Calculate copy position to tail. + */ + addq %rdx, %rsi + addq %rdx, %rdi + subq $0x20, %rdx + /* + * At most 3 ALU operations in one cycle, + * so append NOPS in the same 16bytes trunk. + */ + .p2align 4 +.Lcopy_backward_loop: + subq $0x20, %rdx + movq -1*8(%rsi), %r8 + movq -2*8(%rsi), %r9 + movq -3*8(%rsi), %r10 + movq -4*8(%rsi), %r11 + leaq -4*8(%rsi), %rsi + movq %r8, -1*8(%rdi) + movq %r9, -2*8(%rdi) + movq %r10, -3*8(%rdi) + movq %r11, -4*8(%rdi) + leaq -4*8(%rdi), %rdi + jae .Lcopy_backward_loop + /* + * Calculate copy position to head. + */ + addq $0x20, %rdx + subq %rdx, %rsi + subq %rdx, %rdi .Lhandle_tail: - movl %edx, %ecx - andl $63, %ecx - shrl $3, %ecx - jz .Lhandle_7 + cmpq $16, %rdx + jb .Lless_16bytes + /* + * Move data from 16 bytes to 31 bytes. + */ + movq 0*8(%rsi), %r8 + movq 1*8(%rsi), %r9 + movq -2*8(%rsi, %rdx), %r10 + movq -1*8(%rsi, %rdx), %r11 + movq %r8, 0*8(%rdi) + movq %r9, 1*8(%rdi) + movq %r10, -2*8(%rdi, %rdx) + movq %r11, -1*8(%rdi, %rdx) + retq .p2align 4 -.Lloop_8: - decl %ecx - movq (%rsi), %r8 - movq %r8, (%rdi) - leaq 8(%rdi), %rdi - leaq 8(%rsi), %rsi - jnz .Lloop_8 - -.Lhandle_7: - movl %edx, %ecx - andl $7, %ecx - jz .Lend +.Lless_16bytes: + cmpq $8, %rdx + jb .Lless_8bytes + /* + * Move data from 8 bytes to 15 bytes. + */ + movq 0*8(%rsi), %r8 + movq -1*8(%rsi, %rdx), %r9 + movq %r8, 0*8(%rdi) + movq %r9, -1*8(%rdi, %rdx) + retq + .p2align 4 +.Lless_8bytes: + cmpq $4, %rdx + jb .Lless_3bytes + /* + * Move data from 4 bytes to 7 bytes. + */ + movl (%rsi), %ecx + movl -4(%rsi, %rdx), %r8d + movl %ecx, (%rdi) + movl %r8d, -4(%rdi, %rdx) + retq .p2align 4 +.Lless_3bytes: + cmpl $0, %edx + je .Lend + /* + * Move data from 1 bytes to 3 bytes. + */ .Lloop_1: movb (%rsi), %r8b movb %r8b, (%rdi) incq %rdi incq %rsi - decl %ecx + decl %edx jnz .Lloop_1 .Lend: - ret + retq CFI_ENDPROC ENDPROC(memcpy) ENDPROC(__memcpy) -- cgit v0.10.2 From 9863c90f682fba34cdc26c3437e8c00da6c83fa4 Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Mon, 23 Aug 2010 14:49:11 -0700 Subject: x86, vmware: Remove deprecated VMI kernel support With the recent innovations in CPU hardware acceleration technologies from Intel and AMD, VMware ran a few experiments to compare these techniques to guest paravirtualization technique on VMware's platform. These hardware assisted virtualization techniques have outperformed the performance benefits provided by VMI in most of the workloads. VMware expects that these hardware features will be ubiquitous in a couple of years, as a result, VMware has started a phased retirement of this feature from the hypervisor. Please note that VMI has always been an optimization and non-VMI kernels still work fine on VMware's platform. Latest versions of VMware's product which support VMI are, Workstation 7.0 and VSphere 4.0 on ESX side, future maintainence releases for these products will continue supporting VMI. For more details about VMI retirement take a look at this, http://blogs.vmware.com/guestosguide/2009/09/vmi-retirement.html This feature removal was scheduled for 2.6.37 back in September 2009. Signed-off-by: Alok N Kataria LKML-Reference: <1282600151.19396.22.camel@ank32.eng.vmware.com> Signed-off-by: H. Peter Anvin diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt index 842aa9d..5e2bc4a 100644 --- a/Documentation/feature-removal-schedule.txt +++ b/Documentation/feature-removal-schedule.txt @@ -386,34 +386,6 @@ Who: Tejun Heo ---------------------------- -What: Support for VMware's guest paravirtuliazation technique [VMI] will be - dropped. -When: 2.6.37 or earlier. -Why: With the recent innovations in CPU hardware acceleration technologies - from Intel and AMD, VMware ran a few experiments to compare these - techniques to guest paravirtualization technique on VMware's platform. - These hardware assisted virtualization techniques have outperformed the - performance benefits provided by VMI in most of the workloads. VMware - expects that these hardware features will be ubiquitous in a couple of - years, as a result, VMware has started a phased retirement of this - feature from the hypervisor. We will be removing this feature from the - Kernel too. Right now we are targeting 2.6.37 but can retire earlier if - technical reasons (read opportunity to remove major chunk of pvops) - arise. - - Please note that VMI has always been an optimization and non-VMI kernels - still work fine on VMware's platform. - Latest versions of VMware's product which support VMI are, - Workstation 7.0 and VSphere 4.0 on ESX side, future maintainence - releases for these products will continue supporting VMI. - - For more details about VMI retirement take a look at this, - http://blogs.vmware.com/guestosguide/2009/09/vmi-retirement.html - -Who: Alok N Kataria - ----------------------------- - What: Support for lcd_switch and display_get in asus-laptop driver When: March 2010 Why: These two features use non-standard interfaces. There are the diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 2c85c06..5824098 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -455,7 +455,7 @@ and is between 256 and 4096 characters. It is defined in the file [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, pxa_timer,timer3,32k_counter,timer0_1 [AVR32] avr32 - [X86-32] pit,hpet,tsc,vmi-timer; + [X86-32] pit,hpet,tsc; scx200_hrt on Geode; cyclone on IBM x440 [MIPS] MIPS [PARISC] cr16 diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..f0ee331 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -517,25 +517,6 @@ if PARAVIRT_GUEST source "arch/x86/xen/Kconfig" -config VMI - bool "VMI Guest support (DEPRECATED)" - select PARAVIRT - depends on X86_32 - ---help--- - VMI provides a paravirtualized interface to the VMware ESX server - (it could be used by other hypervisors in theory too, but is not - at the moment), by linking the kernel to a GPL-ed ROM module - provided by the hypervisor. - - As of September 2009, VMware has started a phased retirement - of this feature from VMware's products. Please see - feature-removal-schedule.txt for details. If you are - planning to enable this option, please note that you cannot - live migrate a VMI enabled VM to a future VMware product, - which doesn't support VMI. So if you expect your kernel to - seamlessly migrate to newer VMware products, keep this - disabled. - config KVM_CLOCK bool "KVM paravirtualized clock" select PARAVIRT diff --git a/arch/x86/include/asm/vmi.h b/arch/x86/include/asm/vmi.h deleted file mode 100644 index 61e08c0..0000000 --- a/arch/x86/include/asm/vmi.h +++ /dev/null @@ -1,269 +0,0 @@ -/* - * VMI interface definition - * - * Copyright (C) 2005, VMware, Inc. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, but - * WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or - * NON INFRINGEMENT. See the GNU General Public License for more - * details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - * - * Maintained by: Zachary Amsden zach@vmware.com - * - */ -#include - -/* - *--------------------------------------------------------------------- - * - * VMI Option ROM API - * - *--------------------------------------------------------------------- - */ -#define VMI_SIGNATURE 0x696d5663 /* "cVmi" */ - -#define PCI_VENDOR_ID_VMWARE 0x15AD -#define PCI_DEVICE_ID_VMWARE_VMI 0x0801 - -/* - * We use two version numbers for compatibility, with the major - * number signifying interface breakages, and the minor number - * interface extensions. - */ -#define VMI_API_REV_MAJOR 3 -#define VMI_API_REV_MINOR 0 - -#define VMI_CALL_CPUID 0 -#define VMI_CALL_WRMSR 1 -#define VMI_CALL_RDMSR 2 -#define VMI_CALL_SetGDT 3 -#define VMI_CALL_SetLDT 4 -#define VMI_CALL_SetIDT 5 -#define VMI_CALL_SetTR 6 -#define VMI_CALL_GetGDT 7 -#define VMI_CALL_GetLDT 8 -#define VMI_CALL_GetIDT 9 -#define VMI_CALL_GetTR 10 -#define VMI_CALL_WriteGDTEntry 11 -#define VMI_CALL_WriteLDTEntry 12 -#define VMI_CALL_WriteIDTEntry 13 -#define VMI_CALL_UpdateKernelStack 14 -#define VMI_CALL_SetCR0 15 -#define VMI_CALL_SetCR2 16 -#define VMI_CALL_SetCR3 17 -#define VMI_CALL_SetCR4 18 -#define VMI_CALL_GetCR0 19 -#define VMI_CALL_GetCR2 20 -#define VMI_CALL_GetCR3 21 -#define VMI_CALL_GetCR4 22 -#define VMI_CALL_WBINVD 23 -#define VMI_CALL_SetDR 24 -#define VMI_CALL_GetDR 25 -#define VMI_CALL_RDPMC 26 -#define VMI_CALL_RDTSC 27 -#define VMI_CALL_CLTS 28 -#define VMI_CALL_EnableInterrupts 29 -#define VMI_CALL_DisableInterrupts 30 -#define VMI_CALL_GetInterruptMask 31 -#define VMI_CALL_SetInterruptMask 32 -#define VMI_CALL_IRET 33 -#define VMI_CALL_SYSEXIT 34 -#define VMI_CALL_Halt 35 -#define VMI_CALL_Reboot 36 -#define VMI_CALL_Shutdown 37 -#define VMI_CALL_SetPxE 38 -#define VMI_CALL_SetPxELong 39 -#define VMI_CALL_UpdatePxE 40 -#define VMI_CALL_UpdatePxELong 41 -#define VMI_CALL_MachineToPhysical 42 -#define VMI_CALL_PhysicalToMachine 43 -#define VMI_CALL_AllocatePage 44 -#define VMI_CALL_ReleasePage 45 -#define VMI_CALL_InvalPage 46 -#define VMI_CALL_FlushTLB 47 -#define VMI_CALL_SetLinearMapping 48 - -#define VMI_CALL_SetIOPLMask 61 -#define VMI_CALL_SetInitialAPState 62 -#define VMI_CALL_APICWrite 63 -#define VMI_CALL_APICRead 64 -#define VMI_CALL_IODelay 65 -#define VMI_CALL_SetLazyMode 73 - -/* - *--------------------------------------------------------------------- - * - * MMU operation flags - * - *--------------------------------------------------------------------- - */ - -/* Flags used by VMI_{Allocate|Release}Page call */ -#define VMI_PAGE_PAE 0x10 /* Allocate PAE shadow */ -#define VMI_PAGE_CLONE 0x20 /* Clone from another shadow */ -#define VMI_PAGE_ZEROED 0x40 /* Page is pre-zeroed */ - - -/* Flags shared by Allocate|Release Page and PTE updates */ -#define VMI_PAGE_PT 0x01 -#define VMI_PAGE_PD 0x02 -#define VMI_PAGE_PDP 0x04 -#define VMI_PAGE_PML4 0x08 - -#define VMI_PAGE_NORMAL 0x00 /* for debugging */ - -/* Flags used by PTE updates */ -#define VMI_PAGE_CURRENT_AS 0x10 /* implies VMI_PAGE_VA_MASK is valid */ -#define VMI_PAGE_DEFER 0x20 /* may queue update until TLB inval */ -#define VMI_PAGE_VA_MASK 0xfffff000 - -#ifdef CONFIG_X86_PAE -#define VMI_PAGE_L1 (VMI_PAGE_PT | VMI_PAGE_PAE | VMI_PAGE_ZEROED) -#define VMI_PAGE_L2 (VMI_PAGE_PD | VMI_PAGE_PAE | VMI_PAGE_ZEROED) -#else -#define VMI_PAGE_L1 (VMI_PAGE_PT | VMI_PAGE_ZEROED) -#define VMI_PAGE_L2 (VMI_PAGE_PD | VMI_PAGE_ZEROED) -#endif - -/* Flags used by VMI_FlushTLB call */ -#define VMI_FLUSH_TLB 0x01 -#define VMI_FLUSH_GLOBAL 0x02 - -/* - *--------------------------------------------------------------------- - * - * VMI relocation definitions for ROM call get_reloc - * - *--------------------------------------------------------------------- - */ - -/* VMI Relocation types */ -#define VMI_RELOCATION_NONE 0 -#define VMI_RELOCATION_CALL_REL 1 -#define VMI_RELOCATION_JUMP_REL 2 -#define VMI_RELOCATION_NOP 3 - -#ifndef __ASSEMBLY__ -struct vmi_relocation_info { - unsigned char *eip; - unsigned char type; - unsigned char reserved[3]; -}; -#endif - - -/* - *--------------------------------------------------------------------- - * - * Generic ROM structures and definitions - * - *--------------------------------------------------------------------- - */ - -#ifndef __ASSEMBLY__ - -struct vrom_header { - u16 rom_signature; /* option ROM signature */ - u8 rom_length; /* ROM length in 512 byte chunks */ - u8 rom_entry[4]; /* 16-bit code entry point */ - u8 rom_pad0; /* 4-byte align pad */ - u32 vrom_signature; /* VROM identification signature */ - u8 api_version_min;/* Minor version of API */ - u8 api_version_maj;/* Major version of API */ - u8 jump_slots; /* Number of jump slots */ - u8 reserved1; /* Reserved for expansion */ - u32 virtual_top; /* Hypervisor virtual address start */ - u16 reserved2; /* Reserved for expansion */ - u16 license_offs; /* Offset to License string */ - u16 pci_header_offs;/* Offset to PCI OPROM header */ - u16 pnp_header_offs;/* Offset to PnP OPROM header */ - u32 rom_pad3; /* PnP reserverd / VMI reserved */ - u8 reserved[96]; /* Reserved for headers */ - char vmi_init[8]; /* VMI_Init jump point */ - char get_reloc[8]; /* VMI_GetRelocationInfo jump point */ -} __attribute__((packed)); - -struct pnp_header { - char sig[4]; - char rev; - char size; - short next; - short res; - long devID; - unsigned short manufacturer_offset; - unsigned short product_offset; -} __attribute__((packed)); - -struct pci_header { - char sig[4]; - short vendorID; - short deviceID; - short vpdData; - short size; - char rev; - char class; - char subclass; - char interface; - short chunks; - char rom_version_min; - char rom_version_maj; - char codetype; - char lastRom; - short reserved; -} __attribute__((packed)); - -/* Function prototypes for bootstrapping */ -#ifdef CONFIG_VMI -extern void vmi_init(void); -extern void vmi_activate(void); -extern void vmi_bringup(void); -#else -static inline void vmi_init(void) {} -static inline void vmi_activate(void) {} -static inline void vmi_bringup(void) {} -#endif - -/* State needed to start an application processor in an SMP system. */ -struct vmi_ap_state { - u32 cr0; - u32 cr2; - u32 cr3; - u32 cr4; - - u64 efer; - - u32 eip; - u32 eflags; - u32 eax; - u32 ebx; - u32 ecx; - u32 edx; - u32 esp; - u32 ebp; - u32 esi; - u32 edi; - u16 cs; - u16 ss; - u16 ds; - u16 es; - u16 fs; - u16 gs; - u16 ldtr; - - u16 gdtr_limit; - u32 gdtr_base; - u32 idtr_base; - u16 idtr_limit; -}; - -#endif diff --git a/arch/x86/include/asm/vmi_time.h b/arch/x86/include/asm/vmi_time.h deleted file mode 100644 index c6e0bee..0000000 --- a/arch/x86/include/asm/vmi_time.h +++ /dev/null @@ -1,98 +0,0 @@ -/* - * VMI Time wrappers - * - * Copyright (C) 2006, VMware, Inc. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, but - * WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or - * NON INFRINGEMENT. See the GNU General Public License for more - * details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - * - * Send feedback to dhecht@vmware.com - * - */ - -#ifndef _ASM_X86_VMI_TIME_H -#define _ASM_X86_VMI_TIME_H - -/* - * Raw VMI call indices for timer functions - */ -#define VMI_CALL_GetCycleFrequency 66 -#define VMI_CALL_GetCycleCounter 67 -#define VMI_CALL_SetAlarm 68 -#define VMI_CALL_CancelAlarm 69 -#define VMI_CALL_GetWallclockTime 70 -#define VMI_CALL_WallclockUpdated 71 - -/* Cached VMI timer operations */ -extern struct vmi_timer_ops { - u64 (*get_cycle_frequency)(void); - u64 (*get_cycle_counter)(int); - u64 (*get_wallclock)(void); - int (*wallclock_updated)(void); - void (*set_alarm)(u32 flags, u64 expiry, u64 period); - void (*cancel_alarm)(u32 flags); -} vmi_timer_ops; - -/* Prototypes */ -extern void __init vmi_time_init(void); -extern unsigned long vmi_get_wallclock(void); -extern int vmi_set_wallclock(unsigned long now); -extern unsigned long long vmi_sched_clock(void); -extern unsigned long vmi_tsc_khz(void); - -#ifdef CONFIG_X86_LOCAL_APIC -extern void __devinit vmi_time_bsp_init(void); -extern void __devinit vmi_time_ap_init(void); -#endif - -/* - * When run under a hypervisor, a vcpu is always in one of three states: - * running, halted, or ready. The vcpu is in the 'running' state if it - * is executing. When the vcpu executes the halt interface, the vcpu - * enters the 'halted' state and remains halted until there is some work - * pending for the vcpu (e.g. an alarm expires, host I/O completes on - * behalf of virtual I/O). At this point, the vcpu enters the 'ready' - * state (waiting for the hypervisor to reschedule it). Finally, at any - * time when the vcpu is not in the 'running' state nor the 'halted' - * state, it is in the 'ready' state. - * - * Real time is advances while the vcpu is 'running', 'ready', or - * 'halted'. Stolen time is the time in which the vcpu is in the - * 'ready' state. Available time is the remaining time -- the vcpu is - * either 'running' or 'halted'. - * - * All three views of time are accessible through the VMI cycle - * counters. - */ - -/* The cycle counters. */ -#define VMI_CYCLES_REAL 0 -#define VMI_CYCLES_AVAILABLE 1 -#define VMI_CYCLES_STOLEN 2 - -/* The alarm interface 'flags' bits */ -#define VMI_ALARM_COUNTERS 2 - -#define VMI_ALARM_COUNTER_MASK 0x000000ff - -#define VMI_ALARM_WIRED_IRQ0 0x00000000 -#define VMI_ALARM_WIRED_LVTT 0x00010000 - -#define VMI_ALARM_IS_ONESHOT 0x00000000 -#define VMI_ALARM_IS_PERIODIC 0x00000100 - -#define CONFIG_VMI_ALARM_HZ 100 - -#endif /* _ASM_X86_VMI_TIME_H */ diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 0925676..801127c 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -91,7 +91,6 @@ obj-$(CONFIG_K8_NB) += k8.o obj-$(CONFIG_DEBUG_RODATA_TEST) += test_rodata.o obj-$(CONFIG_DEBUG_NX_TEST) += test_nx.o -obj-$(CONFIG_VMI) += vmi_32.o vmiclock_32.o obj-$(CONFIG_KVM_GUEST) += kvm.o obj-$(CONFIG_KVM_CLOCK) += kvmclock.o obj-$(CONFIG_PARAVIRT) += paravirt.o paravirt_patch_$(BITS).o diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index c3a4fbb..feb4c21 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -83,7 +83,6 @@ #include #include #include -#include #include #include #include @@ -734,10 +733,10 @@ void __init setup_arch(char **cmdline_p) printk(KERN_INFO "Command line: %s\n", boot_command_line); #endif - /* VMI may relocate the fixmap; do this before touching ioremap area */ - vmi_init(); - - /* OFW also may relocate the fixmap */ + /* + * If we have OLPC OFW, we might end up relocating the fixmap due to + * reserve_top(), so do this before touching the ioremap area. + */ olpc_ofw_detect(); early_trap_init(); @@ -838,9 +837,6 @@ void __init setup_arch(char **cmdline_p) x86_report_nx(); - /* Must be before kernel pagetables are setup */ - vmi_activate(); - /* after early param, so could get panic from serial */ reserve_early_setup_data(); diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 8b3bfc4..63a1a55 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -62,7 +62,6 @@ #include #include #include -#include #include #include #include @@ -311,7 +310,6 @@ notrace static void __cpuinit start_secondary(void *unused) __flush_tlb_all(); #endif - vmi_bringup(); cpu_init(); preempt_disable(); smp_callin(); diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c deleted file mode 100644 index ce9fbac..0000000 --- a/arch/x86/kernel/vmi_32.c +++ /dev/null @@ -1,893 +0,0 @@ -/* - * VMI specific paravirt-ops implementation - * - * Copyright (C) 2005, VMware, Inc. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, but - * WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or - * NON INFRINGEMENT. See the GNU General Public License for more - * details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - * - * Send feedback to zach@vmware.com - * - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* Convenient for calling VMI functions indirectly in the ROM */ -typedef u32 __attribute__((regparm(1))) (VROMFUNC)(void); -typedef u64 __attribute__((regparm(2))) (VROMLONGFUNC)(int); - -#define call_vrom_func(rom,func) \ - (((VROMFUNC *)(rom->func))()) - -#define call_vrom_long_func(rom,func,arg) \ - (((VROMLONGFUNC *)(rom->func)) (arg)) - -static struct vrom_header *vmi_rom; -static int disable_pge; -static int disable_pse; -static int disable_sep; -static int disable_tsc; -static int disable_mtrr; -static int disable_noidle; -static int disable_vmi_timer; - -/* Cached VMI operations */ -static struct { - void (*cpuid)(void /* non-c */); - void (*_set_ldt)(u32 selector); - void (*set_tr)(u32 selector); - void (*write_idt_entry)(struct desc_struct *, int, u32, u32); - void (*write_gdt_entry)(struct desc_struct *, int, u32, u32); - void (*write_ldt_entry)(struct desc_struct *, int, u32, u32); - void (*set_kernel_stack)(u32 selector, u32 sp0); - void (*allocate_page)(u32, u32, u32, u32, u32); - void (*release_page)(u32, u32); - void (*set_pte)(pte_t, pte_t *, unsigned); - void (*update_pte)(pte_t *, unsigned); - void (*set_linear_mapping)(int, void *, u32, u32); - void (*_flush_tlb)(int); - void (*set_initial_ap_state)(int, int); - void (*halt)(void); - void (*set_lazy_mode)(int mode); -} vmi_ops; - -/* Cached VMI operations */ -struct vmi_timer_ops vmi_timer_ops; - -/* - * VMI patching routines. - */ -#define MNEM_CALL 0xe8 -#define MNEM_JMP 0xe9 -#define MNEM_RET 0xc3 - -#define IRQ_PATCH_INT_MASK 0 -#define IRQ_PATCH_DISABLE 5 - -static inline void patch_offset(void *insnbuf, - unsigned long ip, unsigned long dest) -{ - *(unsigned long *)(insnbuf+1) = dest-ip-5; -} - -static unsigned patch_internal(int call, unsigned len, void *insnbuf, - unsigned long ip) -{ - u64 reloc; - struct vmi_relocation_info *const rel = (struct vmi_relocation_info *)&reloc; - reloc = call_vrom_long_func(vmi_rom, get_reloc, call); - switch(rel->type) { - case VMI_RELOCATION_CALL_REL: - BUG_ON(len < 5); - *(char *)insnbuf = MNEM_CALL; - patch_offset(insnbuf, ip, (unsigned long)rel->eip); - return 5; - - case VMI_RELOCATION_JUMP_REL: - BUG_ON(len < 5); - *(char *)insnbuf = MNEM_JMP; - patch_offset(insnbuf, ip, (unsigned long)rel->eip); - return 5; - - case VMI_RELOCATION_NOP: - /* obliterate the whole thing */ - return 0; - - case VMI_RELOCATION_NONE: - /* leave native code in place */ - break; - - default: - BUG(); - } - return len; -} - -/* - * Apply patch if appropriate, return length of new instruction - * sequence. The callee does nop padding for us. - */ -static unsigned vmi_patch(u8 type, u16 clobbers, void *insns, - unsigned long ip, unsigned len) -{ - switch (type) { - case PARAVIRT_PATCH(pv_irq_ops.irq_disable): - return patch_internal(VMI_CALL_DisableInterrupts, len, - insns, ip); - case PARAVIRT_PATCH(pv_irq_ops.irq_enable): - return patch_internal(VMI_CALL_EnableInterrupts, len, - insns, ip); - case PARAVIRT_PATCH(pv_irq_ops.restore_fl): - return patch_internal(VMI_CALL_SetInterruptMask, len, - insns, ip); - case PARAVIRT_PATCH(pv_irq_ops.save_fl): - return patch_internal(VMI_CALL_GetInterruptMask, len, - insns, ip); - case PARAVIRT_PATCH(pv_cpu_ops.iret): - return patch_internal(VMI_CALL_IRET, len, insns, ip); - case PARAVIRT_PATCH(pv_cpu_ops.irq_enable_sysexit): - return patch_internal(VMI_CALL_SYSEXIT, len, insns, ip); - default: - break; - } - return len; -} - -/* CPUID has non-C semantics, and paravirt-ops API doesn't match hardware ISA */ -static void vmi_cpuid(unsigned int *ax, unsigned int *bx, - unsigned int *cx, unsigned int *dx) -{ - int override = 0; - if (*ax == 1) - override = 1; - asm volatile ("call *%6" - : "=a" (*ax), - "=b" (*bx), - "=c" (*cx), - "=d" (*dx) - : "0" (*ax), "2" (*cx), "r" (vmi_ops.cpuid)); - if (override) { - if (disable_pse) - *dx &= ~X86_FEATURE_PSE; - if (disable_pge) - *dx &= ~X86_FEATURE_PGE; - if (disable_sep) - *dx &= ~X86_FEATURE_SEP; - if (disable_tsc) - *dx &= ~X86_FEATURE_TSC; - if (disable_mtrr) - *dx &= ~X86_FEATURE_MTRR; - } -} - -static inline void vmi_maybe_load_tls(struct desc_struct *gdt, int nr, struct desc_struct *new) -{ - if (gdt[nr].a != new->a || gdt[nr].b != new->b) - write_gdt_entry(gdt, nr, new, 0); -} - -static void vmi_load_tls(struct thread_struct *t, unsigned int cpu) -{ - struct desc_struct *gdt = get_cpu_gdt_table(cpu); - vmi_maybe_load_tls(gdt, GDT_ENTRY_TLS_MIN + 0, &t->tls_array[0]); - vmi_maybe_load_tls(gdt, GDT_ENTRY_TLS_MIN + 1, &t->tls_array[1]); - vmi_maybe_load_tls(gdt, GDT_ENTRY_TLS_MIN + 2, &t->tls_array[2]); -} - -static void vmi_set_ldt(const void *addr, unsigned entries) -{ - unsigned cpu = smp_processor_id(); - struct desc_struct desc; - - pack_descriptor(&desc, (unsigned long)addr, - entries * sizeof(struct desc_struct) - 1, - DESC_LDT, 0); - write_gdt_entry(get_cpu_gdt_table(cpu), GDT_ENTRY_LDT, &desc, DESC_LDT); - vmi_ops._set_ldt(entries ? GDT_ENTRY_LDT*sizeof(struct desc_struct) : 0); -} - -static void vmi_set_tr(void) -{ - vmi_ops.set_tr(GDT_ENTRY_TSS*sizeof(struct desc_struct)); -} - -static void vmi_write_idt_entry(gate_desc *dt, int entry, const gate_desc *g) -{ - u32 *idt_entry = (u32 *)g; - vmi_ops.write_idt_entry(dt, entry, idt_entry[0], idt_entry[1]); -} - -static void vmi_write_gdt_entry(struct desc_struct *dt, int entry, - const void *desc, int type) -{ - u32 *gdt_entry = (u32 *)desc; - vmi_ops.write_gdt_entry(dt, entry, gdt_entry[0], gdt_entry[1]); -} - -static void vmi_write_ldt_entry(struct desc_struct *dt, int entry, - const void *desc) -{ - u32 *ldt_entry = (u32 *)desc; - vmi_ops.write_ldt_entry(dt, entry, ldt_entry[0], ldt_entry[1]); -} - -static void vmi_load_sp0(struct tss_struct *tss, - struct thread_struct *thread) -{ - tss->x86_tss.sp0 = thread->sp0; - - /* This can only happen when SEP is enabled, no need to test "SEP"arately */ - if (unlikely(tss->x86_tss.ss1 != thread->sysenter_cs)) { - tss->x86_tss.ss1 = thread->sysenter_cs; - wrmsr(MSR_IA32_SYSENTER_CS, thread->sysenter_cs, 0); - } - vmi_ops.set_kernel_stack(__KERNEL_DS, tss->x86_tss.sp0); -} - -static void vmi_flush_tlb_user(void) -{ - vmi_ops._flush_tlb(VMI_FLUSH_TLB); -} - -static void vmi_flush_tlb_kernel(void) -{ - vmi_ops._flush_tlb(VMI_FLUSH_TLB | VMI_FLUSH_GLOBAL); -} - -/* Stub to do nothing at all; used for delays and unimplemented calls */ -static void vmi_nop(void) -{ -} - -static void vmi_allocate_pte(struct mm_struct *mm, unsigned long pfn) -{ - vmi_ops.allocate_page(pfn, VMI_PAGE_L1, 0, 0, 0); -} - -static void vmi_allocate_pmd(struct mm_struct *mm, unsigned long pfn) -{ - /* - * This call comes in very early, before mem_map is setup. - * It is called only for swapper_pg_dir, which already has - * data on it. - */ - vmi_ops.allocate_page(pfn, VMI_PAGE_L2, 0, 0, 0); -} - -static void vmi_allocate_pmd_clone(unsigned long pfn, unsigned long clonepfn, unsigned long start, unsigned long count) -{ - vmi_ops.allocate_page(pfn, VMI_PAGE_L2 | VMI_PAGE_CLONE, clonepfn, start, count); -} - -static void vmi_release_pte(unsigned long pfn) -{ - vmi_ops.release_page(pfn, VMI_PAGE_L1); -} - -static void vmi_release_pmd(unsigned long pfn) -{ - vmi_ops.release_page(pfn, VMI_PAGE_L2); -} - -/* - * We use the pgd_free hook for releasing the pgd page: - */ -static void vmi_pgd_free(struct mm_struct *mm, pgd_t *pgd) -{ - unsigned long pfn = __pa(pgd) >> PAGE_SHIFT; - - vmi_ops.release_page(pfn, VMI_PAGE_L2); -} - -/* - * Helper macros for MMU update flags. We can defer updates until a flush - * or page invalidation only if the update is to the current address space - * (otherwise, there is no flush). We must check against init_mm, since - * this could be a kernel update, which usually passes init_mm, although - * sometimes this check can be skipped if we know the particular function - * is only called on user mode PTEs. We could change the kernel to pass - * current->active_mm here, but in particular, I was unsure if changing - * mm/highmem.c to do this would still be correct on other architectures. - */ -#define is_current_as(mm, mustbeuser) ((mm) == current->active_mm || \ - (!mustbeuser && (mm) == &init_mm)) -#define vmi_flags_addr(mm, addr, level, user) \ - ((level) | (is_current_as(mm, user) ? \ - (VMI_PAGE_CURRENT_AS | ((addr) & VMI_PAGE_VA_MASK)) : 0)) -#define vmi_flags_addr_defer(mm, addr, level, user) \ - ((level) | (is_current_as(mm, user) ? \ - (VMI_PAGE_DEFER | VMI_PAGE_CURRENT_AS | ((addr) & VMI_PAGE_VA_MASK)) : 0)) - -static void vmi_update_pte(struct mm_struct *mm, unsigned long addr, pte_t *ptep) -{ - vmi_ops.update_pte(ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); -} - -static void vmi_update_pte_defer(struct mm_struct *mm, unsigned long addr, pte_t *ptep) -{ - vmi_ops.update_pte(ptep, vmi_flags_addr_defer(mm, addr, VMI_PAGE_PT, 0)); -} - -static void vmi_set_pte(pte_t *ptep, pte_t pte) -{ - /* XXX because of set_pmd_pte, this can be called on PT or PD layers */ - vmi_ops.set_pte(pte, ptep, VMI_PAGE_PT); -} - -static void vmi_set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) -{ - vmi_ops.set_pte(pte, ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); -} - -static void vmi_set_pmd(pmd_t *pmdp, pmd_t pmdval) -{ -#ifdef CONFIG_X86_PAE - const pte_t pte = { .pte = pmdval.pmd }; -#else - const pte_t pte = { pmdval.pud.pgd.pgd }; -#endif - vmi_ops.set_pte(pte, (pte_t *)pmdp, VMI_PAGE_PD); -} - -#ifdef CONFIG_X86_PAE - -static void vmi_set_pte_atomic(pte_t *ptep, pte_t pteval) -{ - /* - * XXX This is called from set_pmd_pte, but at both PT - * and PD layers so the VMI_PAGE_PT flag is wrong. But - * it is only called for large page mapping changes, - * the Xen backend, doesn't support large pages, and the - * ESX backend doesn't depend on the flag. - */ - set_64bit((unsigned long long *)ptep,pte_val(pteval)); - vmi_ops.update_pte(ptep, VMI_PAGE_PT); -} - -static void vmi_set_pud(pud_t *pudp, pud_t pudval) -{ - /* Um, eww */ - const pte_t pte = { .pte = pudval.pgd.pgd }; - vmi_ops.set_pte(pte, (pte_t *)pudp, VMI_PAGE_PDP); -} - -static void vmi_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) -{ - const pte_t pte = { .pte = 0 }; - vmi_ops.set_pte(pte, ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); -} - -static void vmi_pmd_clear(pmd_t *pmd) -{ - const pte_t pte = { .pte = 0 }; - vmi_ops.set_pte(pte, (pte_t *)pmd, VMI_PAGE_PD); -} -#endif - -#ifdef CONFIG_SMP -static void __devinit -vmi_startup_ipi_hook(int phys_apicid, unsigned long start_eip, - unsigned long start_esp) -{ - struct vmi_ap_state ap; - - /* Default everything to zero. This is fine for most GPRs. */ - memset(&ap, 0, sizeof(struct vmi_ap_state)); - - ap.gdtr_limit = GDT_SIZE - 1; - ap.gdtr_base = (unsigned long) get_cpu_gdt_table(phys_apicid); - - ap.idtr_limit = IDT_ENTRIES * 8 - 1; - ap.idtr_base = (unsigned long) idt_table; - - ap.ldtr = 0; - - ap.cs = __KERNEL_CS; - ap.eip = (unsigned long) start_eip; - ap.ss = __KERNEL_DS; - ap.esp = (unsigned long) start_esp; - - ap.ds = __USER_DS; - ap.es = __USER_DS; - ap.fs = __KERNEL_PERCPU; - ap.gs = __KERNEL_STACK_CANARY; - - ap.eflags = 0; - -#ifdef CONFIG_X86_PAE - /* efer should match BSP efer. */ - if (cpu_has_nx) { - unsigned l, h; - rdmsr(MSR_EFER, l, h); - ap.efer = (unsigned long long) h << 32 | l; - } -#endif - - ap.cr3 = __pa(swapper_pg_dir); - /* Protected mode, paging, AM, WP, NE, MP. */ - ap.cr0 = 0x80050023; - ap.cr4 = mmu_cr4_features; - vmi_ops.set_initial_ap_state((u32)&ap, phys_apicid); -} -#endif - -static void vmi_start_context_switch(struct task_struct *prev) -{ - paravirt_start_context_switch(prev); - vmi_ops.set_lazy_mode(2); -} - -static void vmi_end_context_switch(struct task_struct *next) -{ - vmi_ops.set_lazy_mode(0); - paravirt_end_context_switch(next); -} - -static void vmi_enter_lazy_mmu(void) -{ - paravirt_enter_lazy_mmu(); - vmi_ops.set_lazy_mode(1); -} - -static void vmi_leave_lazy_mmu(void) -{ - vmi_ops.set_lazy_mode(0); - paravirt_leave_lazy_mmu(); -} - -static inline int __init check_vmi_rom(struct vrom_header *rom) -{ - struct pci_header *pci; - struct pnp_header *pnp; - const char *manufacturer = "UNKNOWN"; - const char *product = "UNKNOWN"; - const char *license = "unspecified"; - - if (rom->rom_signature != 0xaa55) - return 0; - if (rom->vrom_signature != VMI_SIGNATURE) - return 0; - if (rom->api_version_maj != VMI_API_REV_MAJOR || - rom->api_version_min+1 < VMI_API_REV_MINOR+1) { - printk(KERN_WARNING "VMI: Found mismatched rom version %d.%d\n", - rom->api_version_maj, - rom->api_version_min); - return 0; - } - - /* - * Relying on the VMI_SIGNATURE field is not 100% safe, so check - * the PCI header and device type to make sure this is really a - * VMI device. - */ - if (!rom->pci_header_offs) { - printk(KERN_WARNING "VMI: ROM does not contain PCI header.\n"); - return 0; - } - - pci = (struct pci_header *)((char *)rom+rom->pci_header_offs); - if (pci->vendorID != PCI_VENDOR_ID_VMWARE || - pci->deviceID != PCI_DEVICE_ID_VMWARE_VMI) { - /* Allow it to run... anyways, but warn */ - printk(KERN_WARNING "VMI: ROM from unknown manufacturer\n"); - } - - if (rom->pnp_header_offs) { - pnp = (struct pnp_header *)((char *)rom+rom->pnp_header_offs); - if (pnp->manufacturer_offset) - manufacturer = (const char *)rom+pnp->manufacturer_offset; - if (pnp->product_offset) - product = (const char *)rom+pnp->product_offset; - } - - if (rom->license_offs) - license = (char *)rom+rom->license_offs; - - printk(KERN_INFO "VMI: Found %s %s, API version %d.%d, ROM version %d.%d\n", - manufacturer, product, - rom->api_version_maj, rom->api_version_min, - pci->rom_version_maj, pci->rom_version_min); - - /* Don't allow BSD/MIT here for now because we don't want to end up - with any binary only shim layers */ - if (strcmp(license, "GPL") && strcmp(license, "GPL v2")) { - printk(KERN_WARNING "VMI: Non GPL license `%s' found for ROM. Not used.\n", - license); - return 0; - } - - return 1; -} - -/* - * Probe for the VMI option ROM - */ -static inline int __init probe_vmi_rom(void) -{ - unsigned long base; - - /* VMI ROM is in option ROM area, check signature */ - for (base = 0xC0000; base < 0xE0000; base += 2048) { - struct vrom_header *romstart; - romstart = (struct vrom_header *)isa_bus_to_virt(base); - if (check_vmi_rom(romstart)) { - vmi_rom = romstart; - return 1; - } - } - return 0; -} - -/* - * VMI setup common to all processors - */ -void vmi_bringup(void) -{ - /* We must establish the lowmem mapping for MMU ops to work */ - if (vmi_ops.set_linear_mapping) - vmi_ops.set_linear_mapping(0, (void *)__PAGE_OFFSET, MAXMEM_PFN, 0); -} - -/* - * Return a pointer to a VMI function or NULL if unimplemented - */ -static void *vmi_get_function(int vmicall) -{ - u64 reloc; - const struct vmi_relocation_info *rel = (struct vmi_relocation_info *)&reloc; - reloc = call_vrom_long_func(vmi_rom, get_reloc, vmicall); - BUG_ON(rel->type == VMI_RELOCATION_JUMP_REL); - if (rel->type == VMI_RELOCATION_CALL_REL) - return (void *)rel->eip; - else - return NULL; -} - -/* - * Helper macro for making the VMI paravirt-ops fill code readable. - * For unimplemented operations, fall back to default, unless nop - * is returned by the ROM. - */ -#define para_fill(opname, vmicall) \ -do { \ - reloc = call_vrom_long_func(vmi_rom, get_reloc, \ - VMI_CALL_##vmicall); \ - if (rel->type == VMI_RELOCATION_CALL_REL) \ - opname = (void *)rel->eip; \ - else if (rel->type == VMI_RELOCATION_NOP) \ - opname = (void *)vmi_nop; \ - else if (rel->type != VMI_RELOCATION_NONE) \ - printk(KERN_WARNING "VMI: Unknown relocation " \ - "type %d for " #vmicall"\n",\ - rel->type); \ -} while (0) - -/* - * Helper macro for making the VMI paravirt-ops fill code readable. - * For cached operations which do not match the VMI ROM ABI and must - * go through a tranlation stub. Ignore NOPs, since it is not clear - * a NOP * VMI function corresponds to a NOP paravirt-op when the - * functions are not in 1-1 correspondence. - */ -#define para_wrap(opname, wrapper, cache, vmicall) \ -do { \ - reloc = call_vrom_long_func(vmi_rom, get_reloc, \ - VMI_CALL_##vmicall); \ - BUG_ON(rel->type == VMI_RELOCATION_JUMP_REL); \ - if (rel->type == VMI_RELOCATION_CALL_REL) { \ - opname = wrapper; \ - vmi_ops.cache = (void *)rel->eip; \ - } \ -} while (0) - -/* - * Activate the VMI interface and switch into paravirtualized mode - */ -static inline int __init activate_vmi(void) -{ - short kernel_cs; - u64 reloc; - const struct vmi_relocation_info *rel = (struct vmi_relocation_info *)&reloc; - - /* - * Prevent page tables from being allocated in highmem, even if - * CONFIG_HIGHPTE is enabled. - */ - __userpte_alloc_gfp &= ~__GFP_HIGHMEM; - - if (call_vrom_func(vmi_rom, vmi_init) != 0) { - printk(KERN_ERR "VMI ROM failed to initialize!"); - return 0; - } - savesegment(cs, kernel_cs); - - pv_info.paravirt_enabled = 1; - pv_info.kernel_rpl = kernel_cs & SEGMENT_RPL_MASK; - pv_info.name = "vmi [deprecated]"; - - pv_init_ops.patch = vmi_patch; - - /* - * Many of these operations are ABI compatible with VMI. - * This means we can fill in the paravirt-ops with direct - * pointers into the VMI ROM. If the calling convention for - * these operations changes, this code needs to be updated. - * - * Exceptions - * CPUID paravirt-op uses pointers, not the native ISA - * halt has no VMI equivalent; all VMI halts are "safe" - * no MSR support yet - just trap and emulate. VMI uses the - * same ABI as the native ISA, but Linux wants exceptions - * from bogus MSR read / write handled - * rdpmc is not yet used in Linux - */ - - /* CPUID is special, so very special it gets wrapped like a present */ - para_wrap(pv_cpu_ops.cpuid, vmi_cpuid, cpuid, CPUID); - - para_fill(pv_cpu_ops.clts, CLTS); - para_fill(pv_cpu_ops.get_debugreg, GetDR); - para_fill(pv_cpu_ops.set_debugreg, SetDR); - para_fill(pv_cpu_ops.read_cr0, GetCR0); - para_fill(pv_mmu_ops.read_cr2, GetCR2); - para_fill(pv_mmu_ops.read_cr3, GetCR3); - para_fill(pv_cpu_ops.read_cr4, GetCR4); - para_fill(pv_cpu_ops.write_cr0, SetCR0); - para_fill(pv_mmu_ops.write_cr2, SetCR2); - para_fill(pv_mmu_ops.write_cr3, SetCR3); - para_fill(pv_cpu_ops.write_cr4, SetCR4); - - para_fill(pv_irq_ops.save_fl.func, GetInterruptMask); - para_fill(pv_irq_ops.restore_fl.func, SetInterruptMask); - para_fill(pv_irq_ops.irq_disable.func, DisableInterrupts); - para_fill(pv_irq_ops.irq_enable.func, EnableInterrupts); - - para_fill(pv_cpu_ops.wbinvd, WBINVD); - para_fill(pv_cpu_ops.read_tsc, RDTSC); - - /* The following we emulate with trap and emulate for now */ - /* paravirt_ops.read_msr = vmi_rdmsr */ - /* paravirt_ops.write_msr = vmi_wrmsr */ - /* paravirt_ops.rdpmc = vmi_rdpmc */ - - /* TR interface doesn't pass TR value, wrap */ - para_wrap(pv_cpu_ops.load_tr_desc, vmi_set_tr, set_tr, SetTR); - - /* LDT is special, too */ - para_wrap(pv_cpu_ops.set_ldt, vmi_set_ldt, _set_ldt, SetLDT); - - para_fill(pv_cpu_ops.load_gdt, SetGDT); - para_fill(pv_cpu_ops.load_idt, SetIDT); - para_fill(pv_cpu_ops.store_gdt, GetGDT); - para_fill(pv_cpu_ops.store_idt, GetIDT); - para_fill(pv_cpu_ops.store_tr, GetTR); - pv_cpu_ops.load_tls = vmi_load_tls; - para_wrap(pv_cpu_ops.write_ldt_entry, vmi_write_ldt_entry, - write_ldt_entry, WriteLDTEntry); - para_wrap(pv_cpu_ops.write_gdt_entry, vmi_write_gdt_entry, - write_gdt_entry, WriteGDTEntry); - para_wrap(pv_cpu_ops.write_idt_entry, vmi_write_idt_entry, - write_idt_entry, WriteIDTEntry); - para_wrap(pv_cpu_ops.load_sp0, vmi_load_sp0, set_kernel_stack, UpdateKernelStack); - para_fill(pv_cpu_ops.set_iopl_mask, SetIOPLMask); - para_fill(pv_cpu_ops.io_delay, IODelay); - - para_wrap(pv_cpu_ops.start_context_switch, vmi_start_context_switch, - set_lazy_mode, SetLazyMode); - para_wrap(pv_cpu_ops.end_context_switch, vmi_end_context_switch, - set_lazy_mode, SetLazyMode); - - para_wrap(pv_mmu_ops.lazy_mode.enter, vmi_enter_lazy_mmu, - set_lazy_mode, SetLazyMode); - para_wrap(pv_mmu_ops.lazy_mode.leave, vmi_leave_lazy_mmu, - set_lazy_mode, SetLazyMode); - - /* user and kernel flush are just handled with different flags to FlushTLB */ - para_wrap(pv_mmu_ops.flush_tlb_user, vmi_flush_tlb_user, _flush_tlb, FlushTLB); - para_wrap(pv_mmu_ops.flush_tlb_kernel, vmi_flush_tlb_kernel, _flush_tlb, FlushTLB); - para_fill(pv_mmu_ops.flush_tlb_single, InvalPage); - - /* - * Until a standard flag format can be agreed on, we need to - * implement these as wrappers in Linux. Get the VMI ROM - * function pointers for the two backend calls. - */ -#ifdef CONFIG_X86_PAE - vmi_ops.set_pte = vmi_get_function(VMI_CALL_SetPxELong); - vmi_ops.update_pte = vmi_get_function(VMI_CALL_UpdatePxELong); -#else - vmi_ops.set_pte = vmi_get_function(VMI_CALL_SetPxE); - vmi_ops.update_pte = vmi_get_function(VMI_CALL_UpdatePxE); -#endif - - if (vmi_ops.set_pte) { - pv_mmu_ops.set_pte = vmi_set_pte; - pv_mmu_ops.set_pte_at = vmi_set_pte_at; - pv_mmu_ops.set_pmd = vmi_set_pmd; -#ifdef CONFIG_X86_PAE - pv_mmu_ops.set_pte_atomic = vmi_set_pte_atomic; - pv_mmu_ops.set_pud = vmi_set_pud; - pv_mmu_ops.pte_clear = vmi_pte_clear; - pv_mmu_ops.pmd_clear = vmi_pmd_clear; -#endif - } - - if (vmi_ops.update_pte) { - pv_mmu_ops.pte_update = vmi_update_pte; - pv_mmu_ops.pte_update_defer = vmi_update_pte_defer; - } - - vmi_ops.allocate_page = vmi_get_function(VMI_CALL_AllocatePage); - if (vmi_ops.allocate_page) { - pv_mmu_ops.alloc_pte = vmi_allocate_pte; - pv_mmu_ops.alloc_pmd = vmi_allocate_pmd; - pv_mmu_ops.alloc_pmd_clone = vmi_allocate_pmd_clone; - } - - vmi_ops.release_page = vmi_get_function(VMI_CALL_ReleasePage); - if (vmi_ops.release_page) { - pv_mmu_ops.release_pte = vmi_release_pte; - pv_mmu_ops.release_pmd = vmi_release_pmd; - pv_mmu_ops.pgd_free = vmi_pgd_free; - } - - /* Set linear is needed in all cases */ - vmi_ops.set_linear_mapping = vmi_get_function(VMI_CALL_SetLinearMapping); - - /* - * These MUST always be patched. Don't support indirect jumps - * through these operations, as the VMI interface may use either - * a jump or a call to get to these operations, depending on - * the backend. They are performance critical anyway, so requiring - * a patch is not a big problem. - */ - pv_cpu_ops.irq_enable_sysexit = (void *)0xfeedbab0; - pv_cpu_ops.iret = (void *)0xbadbab0; - -#ifdef CONFIG_SMP - para_wrap(pv_apic_ops.startup_ipi_hook, vmi_startup_ipi_hook, set_initial_ap_state, SetInitialAPState); -#endif - -#ifdef CONFIG_X86_LOCAL_APIC - para_fill(apic->read, APICRead); - para_fill(apic->write, APICWrite); -#endif - - /* - * Check for VMI timer functionality by probing for a cycle frequency method - */ - reloc = call_vrom_long_func(vmi_rom, get_reloc, VMI_CALL_GetCycleFrequency); - if (!disable_vmi_timer && rel->type != VMI_RELOCATION_NONE) { - vmi_timer_ops.get_cycle_frequency = (void *)rel->eip; - vmi_timer_ops.get_cycle_counter = - vmi_get_function(VMI_CALL_GetCycleCounter); - vmi_timer_ops.get_wallclock = - vmi_get_function(VMI_CALL_GetWallclockTime); - vmi_timer_ops.wallclock_updated = - vmi_get_function(VMI_CALL_WallclockUpdated); - vmi_timer_ops.set_alarm = vmi_get_function(VMI_CALL_SetAlarm); - vmi_timer_ops.cancel_alarm = - vmi_get_function(VMI_CALL_CancelAlarm); - x86_init.timers.timer_init = vmi_time_init; -#ifdef CONFIG_X86_LOCAL_APIC - x86_init.timers.setup_percpu_clockev = vmi_time_bsp_init; - x86_cpuinit.setup_percpu_clockev = vmi_time_ap_init; -#endif - pv_time_ops.sched_clock = vmi_sched_clock; - x86_platform.calibrate_tsc = vmi_tsc_khz; - x86_platform.get_wallclock = vmi_get_wallclock; - x86_platform.set_wallclock = vmi_set_wallclock; - - /* We have true wallclock functions; disable CMOS clock sync */ - no_sync_cmos_clock = 1; - } else { - disable_noidle = 1; - disable_vmi_timer = 1; - } - - para_fill(pv_irq_ops.safe_halt, Halt); - - /* - * Alternative instruction rewriting doesn't happen soon enough - * to convert VMI_IRET to a call instead of a jump; so we have - * to do this before IRQs get reenabled. Fortunately, it is - * idempotent. - */ - apply_paravirt(__parainstructions, __parainstructions_end); - - vmi_bringup(); - - return 1; -} - -#undef para_fill - -void __init vmi_init(void) -{ - if (!vmi_rom) - probe_vmi_rom(); - else - check_vmi_rom(vmi_rom); - - /* In case probing for or validating the ROM failed, basil */ - if (!vmi_rom) - return; - - reserve_top_address(-vmi_rom->virtual_top); - -#ifdef CONFIG_X86_IO_APIC - /* This is virtual hardware; timer routing is wired correctly */ - no_timer_check = 1; -#endif -} - -void __init vmi_activate(void) -{ - unsigned long flags; - - if (!vmi_rom) - return; - - local_irq_save(flags); - activate_vmi(); - local_irq_restore(flags & X86_EFLAGS_IF); -} - -static int __init parse_vmi(char *arg) -{ - if (!arg) - return -EINVAL; - - if (!strcmp(arg, "disable_pge")) { - clear_cpu_cap(&boot_cpu_data, X86_FEATURE_PGE); - disable_pge = 1; - } else if (!strcmp(arg, "disable_pse")) { - clear_cpu_cap(&boot_cpu_data, X86_FEATURE_PSE); - disable_pse = 1; - } else if (!strcmp(arg, "disable_sep")) { - clear_cpu_cap(&boot_cpu_data, X86_FEATURE_SEP); - disable_sep = 1; - } else if (!strcmp(arg, "disable_tsc")) { - clear_cpu_cap(&boot_cpu_data, X86_FEATURE_TSC); - disable_tsc = 1; - } else if (!strcmp(arg, "disable_mtrr")) { - clear_cpu_cap(&boot_cpu_data, X86_FEATURE_MTRR); - disable_mtrr = 1; - } else if (!strcmp(arg, "disable_timer")) { - disable_vmi_timer = 1; - disable_noidle = 1; - } else if (!strcmp(arg, "disable_noidle")) - disable_noidle = 1; - return 0; -} - -early_param("vmi", parse_vmi); diff --git a/arch/x86/kernel/vmiclock_32.c b/arch/x86/kernel/vmiclock_32.c deleted file mode 100644 index 5e1ff66..0000000 --- a/arch/x86/kernel/vmiclock_32.c +++ /dev/null @@ -1,317 +0,0 @@ -/* - * VMI paravirtual timer support routines. - * - * Copyright (C) 2007, VMware, Inc. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, but - * WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or - * NON INFRINGEMENT. See the GNU General Public License for more - * details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - * - */ - -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include - -#define VMI_ONESHOT (VMI_ALARM_IS_ONESHOT | VMI_CYCLES_REAL | vmi_get_alarm_wiring()) -#define VMI_PERIODIC (VMI_ALARM_IS_PERIODIC | VMI_CYCLES_REAL | vmi_get_alarm_wiring()) - -static DEFINE_PER_CPU(struct clock_event_device, local_events); - -static inline u32 vmi_counter(u32 flags) -{ - /* Given VMI_ONESHOT or VMI_PERIODIC, return the corresponding - * cycle counter. */ - return flags & VMI_ALARM_COUNTER_MASK; -} - -/* paravirt_ops.get_wallclock = vmi_get_wallclock */ -unsigned long vmi_get_wallclock(void) -{ - unsigned long long wallclock; - wallclock = vmi_timer_ops.get_wallclock(); // nsec - (void)do_div(wallclock, 1000000000); // sec - - return wallclock; -} - -/* paravirt_ops.set_wallclock = vmi_set_wallclock */ -int vmi_set_wallclock(unsigned long now) -{ - return 0; -} - -/* paravirt_ops.sched_clock = vmi_sched_clock */ -unsigned long long vmi_sched_clock(void) -{ - return cycles_2_ns(vmi_timer_ops.get_cycle_counter(VMI_CYCLES_AVAILABLE)); -} - -/* x86_platform.calibrate_tsc = vmi_tsc_khz */ -unsigned long vmi_tsc_khz(void) -{ - unsigned long long khz; - khz = vmi_timer_ops.get_cycle_frequency(); - (void)do_div(khz, 1000); - return khz; -} - -static inline unsigned int vmi_get_timer_vector(void) -{ - return IRQ0_VECTOR; -} - -/** vmi clockchip */ -#ifdef CONFIG_X86_LOCAL_APIC -static unsigned int startup_timer_irq(unsigned int irq) -{ - unsigned long val = apic_read(APIC_LVTT); - apic_write(APIC_LVTT, vmi_get_timer_vector()); - - return (val & APIC_SEND_PENDING); -} - -static void mask_timer_irq(unsigned int irq) -{ - unsigned long val = apic_read(APIC_LVTT); - apic_write(APIC_LVTT, val | APIC_LVT_MASKED); -} - -static void unmask_timer_irq(unsigned int irq) -{ - unsigned long val = apic_read(APIC_LVTT); - apic_write(APIC_LVTT, val & ~APIC_LVT_MASKED); -} - -static void ack_timer_irq(unsigned int irq) -{ - ack_APIC_irq(); -} - -static struct irq_chip vmi_chip __read_mostly = { - .name = "VMI-LOCAL", - .startup = startup_timer_irq, - .mask = mask_timer_irq, - .unmask = unmask_timer_irq, - .ack = ack_timer_irq -}; -#endif - -/** vmi clockevent */ -#define VMI_ALARM_WIRED_IRQ0 0x00000000 -#define VMI_ALARM_WIRED_LVTT 0x00010000 -static int vmi_wiring = VMI_ALARM_WIRED_IRQ0; - -static inline int vmi_get_alarm_wiring(void) -{ - return vmi_wiring; -} - -static void vmi_timer_set_mode(enum clock_event_mode mode, - struct clock_event_device *evt) -{ - cycle_t now, cycles_per_hz; - BUG_ON(!irqs_disabled()); - - switch (mode) { - case CLOCK_EVT_MODE_ONESHOT: - case CLOCK_EVT_MODE_RESUME: - break; - case CLOCK_EVT_MODE_PERIODIC: - cycles_per_hz = vmi_timer_ops.get_cycle_frequency(); - (void)do_div(cycles_per_hz, HZ); - now = vmi_timer_ops.get_cycle_counter(vmi_counter(VMI_PERIODIC)); - vmi_timer_ops.set_alarm(VMI_PERIODIC, now, cycles_per_hz); - break; - case CLOCK_EVT_MODE_UNUSED: - case CLOCK_EVT_MODE_SHUTDOWN: - switch (evt->mode) { - case CLOCK_EVT_MODE_ONESHOT: - vmi_timer_ops.cancel_alarm(VMI_ONESHOT); - break; - case CLOCK_EVT_MODE_PERIODIC: - vmi_timer_ops.cancel_alarm(VMI_PERIODIC); - break; - default: - break; - } - break; - default: - break; - } -} - -static int vmi_timer_next_event(unsigned long delta, - struct clock_event_device *evt) -{ - /* Unfortunately, set_next_event interface only passes relative - * expiry, but we want absolute expiry. It'd be better if were - * were passed an absolute expiry, since a bunch of time may - * have been stolen between the time the delta is computed and - * when we set the alarm below. */ - cycle_t now = vmi_timer_ops.get_cycle_counter(vmi_counter(VMI_ONESHOT)); - - BUG_ON(evt->mode != CLOCK_EVT_MODE_ONESHOT); - vmi_timer_ops.set_alarm(VMI_ONESHOT, now + delta, 0); - return 0; -} - -static struct clock_event_device vmi_clockevent = { - .name = "vmi-timer", - .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, - .shift = 22, - .set_mode = vmi_timer_set_mode, - .set_next_event = vmi_timer_next_event, - .rating = 1000, - .irq = 0, -}; - -static irqreturn_t vmi_timer_interrupt(int irq, void *dev_id) -{ - struct clock_event_device *evt = &__get_cpu_var(local_events); - evt->event_handler(evt); - return IRQ_HANDLED; -} - -static struct irqaction vmi_clock_action = { - .name = "vmi-timer", - .handler = vmi_timer_interrupt, - .flags = IRQF_DISABLED | IRQF_NOBALANCING | IRQF_TIMER, -}; - -static void __devinit vmi_time_init_clockevent(void) -{ - cycle_t cycles_per_msec; - struct clock_event_device *evt; - - int cpu = smp_processor_id(); - evt = &__get_cpu_var(local_events); - - /* Use cycles_per_msec since div_sc params are 32-bits. */ - cycles_per_msec = vmi_timer_ops.get_cycle_frequency(); - (void)do_div(cycles_per_msec, 1000); - - memcpy(evt, &vmi_clockevent, sizeof(*evt)); - /* Must pick .shift such that .mult fits in 32-bits. Choosing - * .shift to be 22 allows 2^(32-22) cycles per nano-seconds - * before overflow. */ - evt->mult = div_sc(cycles_per_msec, NSEC_PER_MSEC, evt->shift); - /* Upper bound is clockevent's use of ulong for cycle deltas. */ - evt->max_delta_ns = clockevent_delta2ns(ULONG_MAX, evt); - evt->min_delta_ns = clockevent_delta2ns(1, evt); - evt->cpumask = cpumask_of(cpu); - - printk(KERN_WARNING "vmi: registering clock event %s. mult=%u shift=%u\n", - evt->name, evt->mult, evt->shift); - clockevents_register_device(evt); -} - -void __init vmi_time_init(void) -{ - unsigned int cpu; - /* Disable PIT: BIOSes start PIT CH0 with 18.2hz peridic. */ - outb_pit(0x3a, PIT_MODE); /* binary, mode 5, LSB/MSB, ch 0 */ - - vmi_time_init_clockevent(); - setup_irq(0, &vmi_clock_action); - for_each_possible_cpu(cpu) - per_cpu(vector_irq, cpu)[vmi_get_timer_vector()] = 0; -} - -#ifdef CONFIG_X86_LOCAL_APIC -void __devinit vmi_time_bsp_init(void) -{ - /* - * On APIC systems, we want local timers to fire on each cpu. We do - * this by programming LVTT to deliver timer events to the IRQ handler - * for IRQ-0, since we can't re-use the APIC local timer handler - * without interfering with that code. - */ - clockevents_notify(CLOCK_EVT_NOTIFY_SUSPEND, NULL); - local_irq_disable(); -#ifdef CONFIG_SMP - /* - * XXX handle_percpu_irq only defined for SMP; we need to switch over - * to using it, since this is a local interrupt, which each CPU must - * handle individually without locking out or dropping simultaneous - * local timers on other CPUs. We also don't want to trigger the - * quirk workaround code for interrupts which gets invoked from - * handle_percpu_irq via eoi, so we use our own IRQ chip. - */ - set_irq_chip_and_handler_name(0, &vmi_chip, handle_percpu_irq, "lvtt"); -#else - set_irq_chip_and_handler_name(0, &vmi_chip, handle_edge_irq, "lvtt"); -#endif - vmi_wiring = VMI_ALARM_WIRED_LVTT; - apic_write(APIC_LVTT, vmi_get_timer_vector()); - local_irq_enable(); - clockevents_notify(CLOCK_EVT_NOTIFY_RESUME, NULL); -} - -void __devinit vmi_time_ap_init(void) -{ - vmi_time_init_clockevent(); - apic_write(APIC_LVTT, vmi_get_timer_vector()); -} -#endif - -/** vmi clocksource */ -static struct clocksource clocksource_vmi; - -static cycle_t read_real_cycles(struct clocksource *cs) -{ - cycle_t ret = (cycle_t)vmi_timer_ops.get_cycle_counter(VMI_CYCLES_REAL); - return max(ret, clocksource_vmi.cycle_last); -} - -static struct clocksource clocksource_vmi = { - .name = "vmi-timer", - .rating = 450, - .read = read_real_cycles, - .mask = CLOCKSOURCE_MASK(64), - .mult = 0, /* to be set */ - .shift = 22, - .flags = CLOCK_SOURCE_IS_CONTINUOUS, -}; - -static int __init init_vmi_clocksource(void) -{ - cycle_t cycles_per_msec; - - if (!vmi_timer_ops.get_cycle_frequency) - return 0; - /* Use khz2mult rather than hz2mult since hz arg is only 32-bits. */ - cycles_per_msec = vmi_timer_ops.get_cycle_frequency(); - (void)do_div(cycles_per_msec, 1000); - - /* Note that clocksource.{mult, shift} converts in the opposite direction - * as clockevents. */ - clocksource_vmi.mult = clocksource_khz2mult(cycles_per_msec, - clocksource_vmi.shift); - - printk(KERN_WARNING "vmi: registering clock source khz=%lld\n", cycles_per_msec); - return clocksource_register(&clocksource_vmi); - -} -module_init(init_vmi_clocksource); -- cgit v0.10.2 From 2c96c7751d2bb822542b03ddfaca70933f5aaf02 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 23 Aug 2010 16:34:02 -0700 Subject: rcu: upgrade stallwarn.txt documentation for CPU-bound RT processes CPU-bound real-time processes can cause RCU CPU stall warnings, and much other trouble as well. Document the fact that they can cause RCU CPU stall warnings. Suggested-by: Darren Hart Signed-off-by: Paul E. McKenney diff --git a/Documentation/RCU/stallwarn.txt b/Documentation/RCU/stallwarn.txt index 44c6dcc9..862c08e 100644 --- a/Documentation/RCU/stallwarn.txt +++ b/Documentation/RCU/stallwarn.txt @@ -80,6 +80,24 @@ o A CPU looping with bottom halves disabled. This condition can o For !CONFIG_PREEMPT kernels, a CPU looping anywhere in the kernel without invoking schedule(). +o A CPU-bound real-time task in a CONFIG_PREEMPT kernel, which might + happen to preempt a low-priority task in the middle of an RCU + read-side critical section. This is especially damaging if + that low-priority task is not permitted to run on any other CPU, + in which case the next RCU grace period can never complete, which + will eventually cause the system to run out of memory and hang. + While the system is in the process of running itself out of + memory, you might see stall-warning messages. + +o A CPU-bound real-time task in a CONFIG_PREEMPT_RT kernel that + is running at a higher priority than the RCU softirq threads. + This will prevent RCU callbacks from ever being invoked, + and in a CONFIG_TREE_PREEMPT_RCU kernel will further prevent + RCU grace periods from ever completing. Either way, the + system will eventually run out of memory and hang. In the + CONFIG_TREE_PREEMPT_RCU case, you might see stall-warning + messages. + o A bug in the RCU implementation. o A hardware failure. This is quite unlikely, but has occurred -- cgit v0.10.2 From b0f4c062fb6dd4c02b1fe6de73319ed50a09b27d Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Mon, 23 Aug 2010 17:05:57 -0700 Subject: x86, paravirt: Remove alloc_pmd_clone hook, only used by VMI VMI was the only user of the alloc_pmd_clone hook, given that VMI is now removed we can also remove this hook. Signed-off-by: Alok N Kataria LKML-Reference: <1282608357.19396.36.camel@ank32.eng.vmware.com> Cc: Jeremy Fitzhardinge Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h index 5653f43..edecb4e 100644 --- a/arch/x86/include/asm/paravirt.h +++ b/arch/x86/include/asm/paravirt.h @@ -416,11 +416,6 @@ static inline void paravirt_alloc_pmd(struct mm_struct *mm, unsigned long pfn) PVOP_VCALL2(pv_mmu_ops.alloc_pmd, mm, pfn); } -static inline void paravirt_alloc_pmd_clone(unsigned long pfn, unsigned long clonepfn, - unsigned long start, unsigned long count) -{ - PVOP_VCALL4(pv_mmu_ops.alloc_pmd_clone, pfn, clonepfn, start, count); -} static inline void paravirt_release_pmd(unsigned long pfn) { PVOP_VCALL1(pv_mmu_ops.release_pmd, pfn); diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h index db9ef55..b82bac9 100644 --- a/arch/x86/include/asm/paravirt_types.h +++ b/arch/x86/include/asm/paravirt_types.h @@ -255,7 +255,6 @@ struct pv_mmu_ops { */ void (*alloc_pte)(struct mm_struct *mm, unsigned long pfn); void (*alloc_pmd)(struct mm_struct *mm, unsigned long pfn); - void (*alloc_pmd_clone)(unsigned long pfn, unsigned long clonepfn, unsigned long start, unsigned long count); void (*alloc_pud)(struct mm_struct *mm, unsigned long pfn); void (*release_pte)(unsigned long pfn); void (*release_pmd)(unsigned long pfn); diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c index 1db183e..c5b2500 100644 --- a/arch/x86/kernel/paravirt.c +++ b/arch/x86/kernel/paravirt.c @@ -413,7 +413,6 @@ struct pv_mmu_ops pv_mmu_ops = { .alloc_pte = paravirt_nop, .alloc_pmd = paravirt_nop, - .alloc_pmd_clone = paravirt_nop, .alloc_pud = paravirt_nop, .release_pte = paravirt_nop, .release_pmd = paravirt_nop, diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c index 5c4ee42..a96023e 100644 --- a/arch/x86/mm/pgtable.c +++ b/arch/x86/mm/pgtable.c @@ -98,10 +98,6 @@ static void pgd_ctor(pgd_t *pgd) clone_pgd_range(pgd + KERNEL_PGD_BOUNDARY, swapper_pg_dir + KERNEL_PGD_BOUNDARY, KERNEL_PGD_PTRS); - paravirt_alloc_pmd_clone(__pa(pgd) >> PAGE_SHIFT, - __pa(swapper_pg_dir) >> PAGE_SHIFT, - KERNEL_PGD_BOUNDARY, - KERNEL_PGD_PTRS); } /* list required to sync kernel mapping updates */ diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 42086ac..b2363fc 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -1969,7 +1969,6 @@ static const struct pv_mmu_ops xen_mmu_ops __initdata = { .alloc_pte = xen_alloc_pte_init, .release_pte = xen_release_pte_init, .alloc_pmd = xen_alloc_pmd_init, - .alloc_pmd_clone = paravirt_nop, .release_pmd = xen_release_pmd_init, #ifdef CONFIG_X86_64 -- cgit v0.10.2 From d0cd7425fab774a480cce17c2f649984312d0b55 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Tue, 24 Aug 2010 17:32:04 -0700 Subject: x86, bios: By default, reserve the low 64K for all BIOSes The laundry list of BIOSes that need the low 64K reserved is getting very long, so make it the default across all BIOSes. This also allows the code to be simplified and unified with the reservation code for the first 4K. This resolves kernel bugzilla 16661 and who knows what else... Signed-off-by: H. Peter Anvin LKML-Reference: diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..683ae8f 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1326,25 +1326,34 @@ config X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK Set whether the default state of memory_corruption_check is on or off. -config X86_RESERVE_LOW_64K - bool "Reserve low 64K of RAM on AMI/Phoenix BIOSen" - default y - ---help--- - Reserve the first 64K of physical RAM on BIOSes that are known - to potentially corrupt that memory range. A numbers of BIOSes are - known to utilize this area during suspend/resume, so it must not - be used by the kernel. - - Set this to N if you are absolutely sure that you trust the BIOS - to get all its memory reservations and usages right. - - If you have doubts about the BIOS (e.g. suspend/resume does not - work or there's kernel crashes after certain hardware hotplug - events) and it's not AMI or Phoenix, then you might want to enable - X86_CHECK_BIOS_CORRUPTION=y to allow the kernel to check typical - corruption patterns. - - Say Y if unsure. +config X86_LOW_RESERVE + int "Amount of low memory, in kilobytes, to reserve for the BIOS" + default 64 + range 4 640 + ---help--- + Specify the amount of low memory to reserve for the BIOS. + + The first page contains BIOS data structures that the kernel + must not use, so that page must always be reserved. + + By default we reserve the first 64K of physical RAM, as a + number of BIOSes are known to corrupt that memory range + during events such as suspend/resume or monitor cable + insertion, so it must not be used by the kernel. + + You can set this to 4 if you are absolutely sure that you + trust the BIOS to get all its memory reservations and usages + right. If you know your BIOS have problems beyond the + default 64K area, you can set this to 640 to avoid using the + entire low memory range. + + If you have doubts about the BIOS (e.g. suspend/resume does + not work or there's kernel crashes after certain hardware + hotplug events) then you might want to enable + X86_CHECK_BIOS_CORRUPTION=y to allow the kernel to check + typical corruption patterns. + + Leave this to the default value of 64 if you are unsure. config MATH_EMULATION bool diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index c3a4fbb..eb87f1c 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -618,88 +618,20 @@ static __init void reserve_ibft_region(void) reserve_early_overlap_ok(addr, addr + size, "ibft"); } -#ifdef CONFIG_X86_RESERVE_LOW_64K -static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) -{ - printk(KERN_NOTICE - "%s detected: BIOS may corrupt low RAM, working around it.\n", - d->ident); - - e820_update_range(0, 0x10000, E820_RAM, E820_RESERVED); - sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map); - - return 0; -} -#endif - -/* List of systems that have known low memory corruption BIOS problems */ -static struct dmi_system_id __initdata bad_bios_dmi_table[] = { -#ifdef CONFIG_X86_RESERVE_LOW_64K - { - .callback = dmi_low_memory_corruption, - .ident = "AMI BIOS", - .matches = { - DMI_MATCH(DMI_BIOS_VENDOR, "American Megatrends Inc."), - }, - }, - { - .callback = dmi_low_memory_corruption, - .ident = "Phoenix BIOS", - .matches = { - DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies"), - }, - }, - { - .callback = dmi_low_memory_corruption, - .ident = "Phoenix/MSC BIOS", - .matches = { - DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix/MSC"), - }, - }, - /* - * AMI BIOS with low memory corruption was found on Intel DG45ID and - * DG45FC boards. - * It has a different DMI_BIOS_VENDOR = "Intel Corp.", for now we will - * match only DMI_BOARD_NAME and see if there is more bad products - * with this vendor. - */ - { - .callback = dmi_low_memory_corruption, - .ident = "AMI BIOS", - .matches = { - DMI_MATCH(DMI_BOARD_NAME, "DG45ID"), - }, - }, - { - .callback = dmi_low_memory_corruption, - .ident = "AMI BIOS", - .matches = { - DMI_MATCH(DMI_BOARD_NAME, "DG45FC"), - }, - }, - /* - * The Dell Inspiron Mini 1012 has DMI_BIOS_VENDOR = "Dell Inc.", so - * match on the product name. - */ - { - .callback = dmi_low_memory_corruption, - .ident = "Phoenix BIOS", - .matches = { - DMI_MATCH(DMI_PRODUCT_NAME, "Inspiron 1012"), - }, - }, -#endif - {} -}; - static void __init trim_bios_range(void) { /* * A special case is the first 4Kb of memory; * This is a BIOS owned area, not kernel ram, but generally * not listed as such in the E820 table. + * + * This typically reserves additional memory (64KiB by default) + * since some BIOSes are known to corrupt low memory. See the + * Kconfig help text for X86_LOW_RESERVE. */ - e820_update_range(0, PAGE_SIZE, E820_RAM, E820_RESERVED); + e820_update_range(0, ALIGN(CONFIG_X86_LOW_RESERVE << 10, PAGE_SIZE), + E820_RAM, E820_RESERVED); + /* * special case: Some BIOSen report the PC BIOS * area (640->1Mb) as ram even though it is not. @@ -863,8 +795,6 @@ void __init setup_arch(char **cmdline_p) dmi_scan_machine(); - dmi_check_system(bad_bios_dmi_table); - /* * VMware detection requires dmi to be available, so this * needs to be done after dmi_scan_machine, for the BP. -- cgit v0.10.2 From acf01734b1747b1ec4be6f159aff579ea5f7f8e2 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 25 Aug 2010 18:28:23 +0200 Subject: x86, tsc: Remove CPU frequency calibration on AMD 6b37f5a20c0e5c334c010a587058354215433e92 introduced the CPU frequency calibration code for AMD CPUs whose TSCs didn't increment with the core's P0 frequency. From F10h, revB onward, however, the TSC increment rate is denoted by MSRC001_0015[24] and when this bit is set (which should be done by the BIOS) the TSC increments with the P0 frequency so the calibration is not needed and booting can be a couple of mcecs faster on those machines. Besides, there should be virtually no machines out there which don't have this bit set, therefore this calibration can be safely removed. It is a shaky hack anyway since it assumes implicitly that the core is in P0 when BIOS hands off to the OS, which might not always be the case. Signed-off-by: Borislav Petkov LKML-Reference: <20100825162823.GE26438@aftab> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index ba5f62f..fc563fa 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -412,6 +412,23 @@ static void __cpuinit early_init_amd(struct cpuinfo_x86 *c) set_cpu_cap(c, X86_FEATURE_EXTD_APICID); } #endif + + /* We need to do the following only once */ + if (c != &boot_cpu_data) + return; + + if (cpu_has(c, X86_FEATURE_CONSTANT_TSC)) { + + if (c->x86 > 0x10 || + (c->x86 == 0x10 && c->x86_model >= 0x2)) { + u64 val; + + rdmsrl(MSR_K7_HWCR, val); + if (!(val & BIT(24))) + printk(KERN_WARNING FW_BUG "TSC doesn't count " + "with P0 frequency!\n"); + } + } } static void __cpuinit init_amd(struct cpuinfo_x86 *c) diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index ce8e502..13b6a6c 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -854,60 +854,6 @@ static void __init init_tsc_clocksource(void) clocksource_register_khz(&clocksource_tsc, tsc_khz); } -#ifdef CONFIG_X86_64 -/* - * calibrate_cpu is used on systems with fixed rate TSCs to determine - * processor frequency - */ -#define TICK_COUNT 100000000 -static unsigned long __init calibrate_cpu(void) -{ - int tsc_start, tsc_now; - int i, no_ctr_free; - unsigned long evntsel3 = 0, pmc3 = 0, pmc_now = 0; - unsigned long flags; - - for (i = 0; i < 4; i++) - if (avail_to_resrv_perfctr_nmi_bit(i)) - break; - no_ctr_free = (i == 4); - if (no_ctr_free) { - WARN(1, KERN_WARNING "Warning: AMD perfctrs busy ... " - "cpu_khz value may be incorrect.\n"); - i = 3; - rdmsrl(MSR_K7_EVNTSEL3, evntsel3); - wrmsrl(MSR_K7_EVNTSEL3, 0); - rdmsrl(MSR_K7_PERFCTR3, pmc3); - } else { - reserve_perfctr_nmi(MSR_K7_PERFCTR0 + i); - reserve_evntsel_nmi(MSR_K7_EVNTSEL0 + i); - } - local_irq_save(flags); - /* start measuring cycles, incrementing from 0 */ - wrmsrl(MSR_K7_PERFCTR0 + i, 0); - wrmsrl(MSR_K7_EVNTSEL0 + i, 1 << 22 | 3 << 16 | 0x76); - rdtscl(tsc_start); - do { - rdmsrl(MSR_K7_PERFCTR0 + i, pmc_now); - tsc_now = get_cycles(); - } while ((tsc_now - tsc_start) < TICK_COUNT); - - local_irq_restore(flags); - if (no_ctr_free) { - wrmsrl(MSR_K7_EVNTSEL3, 0); - wrmsrl(MSR_K7_PERFCTR3, pmc3); - wrmsrl(MSR_K7_EVNTSEL3, evntsel3); - } else { - release_perfctr_nmi(MSR_K7_PERFCTR0 + i); - release_evntsel_nmi(MSR_K7_EVNTSEL0 + i); - } - - return pmc_now * tsc_khz / (tsc_now - tsc_start); -} -#else -static inline unsigned long calibrate_cpu(void) { return cpu_khz; } -#endif - void __init tsc_init(void) { u64 lpj; @@ -926,10 +872,6 @@ void __init tsc_init(void) return; } - if (cpu_has(&boot_cpu_data, X86_FEATURE_CONSTANT_TSC) && - (boot_cpu_data.x86_vendor == X86_VENDOR_AMD)) - cpu_khz = calibrate_cpu(); - printk("Detected %lu.%03lu MHz processor.\n", (unsigned long)cpu_khz / 1000, (unsigned long)cpu_khz % 1000); -- cgit v0.10.2 From 9ea77bdb39b62c9bf9fd3cdd1c25a9420bccd380 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Wed, 25 Aug 2010 16:38:20 -0700 Subject: x86, bios: Make the x86 early memory reservation a kernel option Add a kernel command-line option so the x86 early memory reservation size can be adjusted at runtime instead of only at compile time. Suggested-by: Andrew Morton LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 2c85c06..41ce93e 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -2150,6 +2150,11 @@ and is between 256 and 4096 characters. It is defined in the file Reserves a hole at the top of the kernel virtual address space. + reservelow= [X86] + Format: nn[K] + Set the amount of memory to reserve for BIOS at + the bottom of the address space. + reset_devices [KNL] Force drivers to reset the underlying device during initialization. diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 683ae8f..d359000 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1326,7 +1326,7 @@ config X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK Set whether the default state of memory_corruption_check is on or off. -config X86_LOW_RESERVE +config X86_RESERVE_LOW int "Amount of low memory, in kilobytes, to reserve for the BIOS" default 64 range 4 640 diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index eb87f1c..af277e3 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -618,6 +618,8 @@ static __init void reserve_ibft_region(void) reserve_early_overlap_ok(addr, addr + size, "ibft"); } +static unsigned reserve_low = CONFIG_X86_RESERVE_LOW << 10; + static void __init trim_bios_range(void) { /* @@ -627,9 +629,9 @@ static void __init trim_bios_range(void) * * This typically reserves additional memory (64KiB by default) * since some BIOSes are known to corrupt low memory. See the - * Kconfig help text for X86_LOW_RESERVE. + * Kconfig help text for X86_RESERVE_LOW. */ - e820_update_range(0, ALIGN(CONFIG_X86_LOW_RESERVE << 10, PAGE_SIZE), + e820_update_range(0, ALIGN(reserve_low, PAGE_SIZE), E820_RAM, E820_RESERVED); /* @@ -641,6 +643,28 @@ static void __init trim_bios_range(void) sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map); } +static int __init parse_reservelow(char *p) +{ + unsigned long long size; + + if (!p) + return -EINVAL; + + size = memparse(p, &p); + + if (size < 4096) + size = 4096; + + if (size > 640*1024) + size = 640*1024; + + reserve_low = size; + + return 0; +} + +early_param("reservelow", parse_reservelow); + /* * Determine if we were loaded by an EFI loader. If so, then we have also been * passed the efi memmap, systab, etc., so we should use these data structures -- cgit v0.10.2 From 6afb5157b9eba4092e2f0f54d24a3806409bdde5 Mon Sep 17 00:00:00 2001 From: Haicheng Li Date: Wed, 19 May 2010 17:42:14 +0800 Subject: x86, mm: Separate x86_64 vmalloc_sync_all() into separate functions No behavior change. Move some of vmalloc_sync_all() code into a new function sync_global_pgds() that will be useful for memory hotplug. Signed-off-by: Haicheng Li LKML-Reference: <4C6E4ECD.1090607@linux.intel.com> Reviewed-by: Wu Fengguang Reviewed-by: Andi Kleen Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h index 076052c..f96ac9b 100644 --- a/arch/x86/include/asm/pgtable_64.h +++ b/arch/x86/include/asm/pgtable_64.h @@ -102,6 +102,8 @@ static inline void native_pgd_clear(pgd_t *pgd) native_set_pgd(pgd, native_make_pgd(0)); } +extern void sync_global_pgds(unsigned long start, unsigned long end); + /* * Conversion functions: convert a page and protection to a page entry, * and a page entry and page directory to the page they refer to. diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 4c4508e..51f7ee7 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -326,29 +326,7 @@ out: void vmalloc_sync_all(void) { - unsigned long address; - - for (address = VMALLOC_START & PGDIR_MASK; address <= VMALLOC_END; - address += PGDIR_SIZE) { - - const pgd_t *pgd_ref = pgd_offset_k(address); - unsigned long flags; - struct page *page; - - if (pgd_none(*pgd_ref)) - continue; - - spin_lock_irqsave(&pgd_lock, flags); - list_for_each_entry(page, &pgd_list, lru) { - pgd_t *pgd; - pgd = (pgd_t *)page_address(page) + pgd_index(address); - if (pgd_none(*pgd)) - set_pgd(pgd, *pgd_ref); - else - BUG_ON(pgd_page_vaddr(*pgd) != pgd_page_vaddr(*pgd_ref)); - } - spin_unlock_irqrestore(&pgd_lock, flags); - } + sync_global_pgds(VMALLOC_START & PGDIR_MASK, VMALLOC_END); } /* diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 9a66746..61a1b4f 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -98,6 +98,36 @@ static int __init nonx32_setup(char *str) __setup("noexec32=", nonx32_setup); /* + * When memory was added/removed make sure all the processes MM have + * suitable PGD entries in the local PGD level page. + */ +void sync_global_pgds(unsigned long start, unsigned long end) +{ + unsigned long address; + + for (address = start; address <= end; address += PGDIR_SIZE) { + const pgd_t *pgd_ref = pgd_offset_k(address); + unsigned long flags; + struct page *page; + + if (pgd_none(*pgd_ref)) + continue; + + spin_lock_irqsave(&pgd_lock, flags); + list_for_each_entry(page, &pgd_list, lru) { + pgd_t *pgd; + pgd = (pgd_t *)page_address(page) + pgd_index(address); + if (pgd_none(*pgd)) + set_pgd(pgd, *pgd_ref); + else + BUG_ON(pgd_page_vaddr(*pgd) + != pgd_page_vaddr(*pgd_ref)); + } + spin_unlock_irqrestore(&pgd_lock, flags); + } +} + +/* * NOTE: This function is marked __ref because it calls __init function * (alloc_bootmem_pages). It's safe to do it ONLY when after_bootmem == 0. */ -- cgit v0.10.2 From 9b861528a8012e7bc4d1f7bae07395b225331477 Mon Sep 17 00:00:00 2001 From: Haicheng Li Date: Fri, 20 Aug 2010 17:50:16 +0800 Subject: x86-64, mem: Update all PGDs for direct mapping and vmemmap mapping changes When memory hotplug-adding happens for a large enough area that a new PGD entry is needed for the direct mapping, the PGDs of other processes would not get updated. This leads to some CPUs oopsing like below when they have to access the unmapped areas. [ 1139.243192] BUG: soft lockup - CPU#0 stuck for 61s! [bash:6534] [ 1139.243195] Modules linked in: ipv6 autofs4 rfcomm l2cap crc16 bluetooth rfkill binfmt_misc dm_mirror dm_region_hash dm_log dm_multipath dm_mod video output sbs sbshc fan battery ac parport_pc lp parport joydev usbhid processor thermal thermal_sys container button rtc_cmos rtc_core rtc_lib i2c_i801 i2c_core pcspkr uhci_hcd ohci_hcd ehci_hcd usbcore [ 1139.243229] irq event stamp: 8538759 [ 1139.243230] hardirqs last enabled at (8538759): [] restore_args+0x0/0x30 [ 1139.243236] hardirqs last disabled at (8538757): [] __do_softirq+0x106/0x146 [ 1139.243240] softirqs last enabled at (8538758): [] __do_softirq+0x137/0x146 [ 1139.243245] softirqs last disabled at (8538743): [] call_softirq+0x1c/0x34 [ 1139.243249] CPU 0: [ 1139.243250] Modules linked in: ipv6 autofs4 rfcomm l2cap crc16 bluetooth rfkill binfmt_misc dm_mirror dm_region_hash dm_log dm_multipath dm_mod video output sbs sbshc fan battery ac parport_pc lp parport joydev usbhid processor thermal thermal_sys container button rtc_cmos rtc_core rtc_lib i2c_i801 i2c_core pcspkr uhci_hcd ohci_hcd ehci_hcd usbcore [ 1139.243284] Pid: 6534, comm: bash Tainted: G M 2.6.32-haicheng-cpuhp #7 QSSC-S4R [ 1139.243287] RIP: 0010:[] [] alloc_arraycache+0x35/0x69 [ 1139.243292] RSP: 0018:ffff8802799f9d78 EFLAGS: 00010286 [ 1139.243295] RAX: ffff8884ffc00000 RBX: ffff8802799f9d98 RCX: 0000000000000000 [ 1139.243297] RDX: 0000000000190018 RSI: 0000000000000001 RDI: ffff8884ffc00010 [ 1139.243300] RBP: ffffffff8100c34e R08: 0000000000000002 R09: 0000000000000000 [ 1139.243303] R10: ffffffff8246dda0 R11: 000000d08246dda0 R12: ffff8802599bfff0 [ 1139.243305] R13: ffff88027904c040 R14: ffff8802799f8000 R15: 0000000000000001 [ 1139.243308] FS: 00007fe81bfe86e0(0000) GS:ffff88000d800000(0000) knlGS:0000000000000000 [ 1139.243311] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 1139.243313] CR2: ffff8884ffc00000 CR3: 000000026cf2d000 CR4: 00000000000006f0 [ 1139.243316] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 1139.243318] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 1139.243321] Call Trace: [ 1139.243324] [] ? alloc_arraycache+0x29/0x69 [ 1139.243328] [] ? cpuup_callback+0x1b0/0x32a [ 1139.243333] [] ? notifier_call_chain+0x33/0x5b [ 1139.243337] [] ? __raw_notifier_call_chain+0x9/0xb [ 1139.243340] [] ? cpu_up+0xb3/0x152 [ 1139.243344] [] ? store_online+0x4d/0x75 [ 1139.243348] [] ? sysdev_store+0x1b/0x1d [ 1139.243351] [] ? sysfs_write_file+0xe5/0x121 [ 1139.243355] [] ? vfs_write+0xae/0x14a [ 1139.243358] [] ? sys_write+0x47/0x6f [ 1139.243362] [] ? system_call_fastpath+0x16/0x1b This patch makes sure to always replicate new direct mapping PGD entries to the PGDs of all processes, as well as ensures corresponding vmemmap mapping gets synced. V1: initial code by Andi Kleen. V2: fix several issues found in testing. V3: as suggested by Wu Fengguang, reuse common code of vmalloc_sync_all(). [ hpa: changed pgd_change from int to bool ] Originally-by: Andi Kleen Signed-off-by: Haicheng Li LKML-Reference: <4C6E4FD8.6080100@linux.intel.com> Reviewed-by: Wu Fengguang Reviewed-by: Andi Kleen Signed-off-by: H. Peter Anvin diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 61a1b4f..64e7bc2 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -564,8 +564,9 @@ kernel_physical_mapping_init(unsigned long start, unsigned long end, unsigned long page_size_mask) { - + bool pgd_changed = false; unsigned long next, last_map_addr = end; + unsigned long addr; start = (unsigned long)__va(start); end = (unsigned long)__va(end); @@ -593,7 +594,12 @@ kernel_physical_mapping_init(unsigned long start, spin_lock(&init_mm.page_table_lock); pgd_populate(&init_mm, pgd, __va(pud_phys)); spin_unlock(&init_mm.page_table_lock); + pgd_changed = true; } + + if (pgd_changed) + sync_global_pgds(addr, end); + __flush_tlb_all(); return last_map_addr; @@ -1033,6 +1039,7 @@ vmemmap_populate(struct page *start_page, unsigned long size, int node) } } + sync_global_pgds((unsigned long)start_page, end); return 0; } -- cgit v0.10.2 From 660a293ea9be709b893d371fbc0328fcca33c33a Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Tue, 27 Jul 2010 16:06:28 +0800 Subject: x86, mm: Make spurious_fault check explicitly check the PRESENT bit pte_present() returns true even present bit isn't set but _PAGE_PROTNONE (global bit) bit is set. While with CONFIG_DEBUG_PAGEALLOC, free pages have global bit set but present bit clear. This patch makes we could catch free pages access with CONFIG_DEBUG_PAGEALLOC enabled. [ hpa: added a comment in the code as a warning to janitors ] Signed-off-by: Shaohua Li LKML-Reference: <1280217988.32400.75.camel@sli10-desk.sh.intel.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 51f7ee7..caec229 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -872,8 +872,14 @@ spurious_fault(unsigned long error_code, unsigned long address) if (pmd_large(*pmd)) return spurious_fault_check(error_code, (pte_t *) pmd); + /* + * Note: don't use pte_present() here, since it returns true + * if the _PAGE_PROTNONE bit is set. However, this aliases the + * _PAGE_GLOBAL bit, which for kernel pages give false positives + * when CONFIG_DEBUG_PAGEALLOC is used. + */ pte = pte_offset_kernel(pmd, address); - if (!pte_present(*pte)) + if (!(pte_flags(*pte) & _PAGE_PRESENT)) return 0; ret = spurious_fault_check(error_code, pte); -- cgit v0.10.2 From dd7c4d89730a1be2c1d361a8ae1f0fe9465ccf9c Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 27 Aug 2010 10:51:17 -0700 Subject: rcu: performance fixes to TINY_PREEMPT_RCU callback checking This commit tightens up checks in rcu_preempt_check_callbacks() to avoid unnecessary special handling at rcu_read_unlock() time. Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index c5bea11..6ceca4f 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -72,7 +72,7 @@ static void rcu_report_exp_done(void); /* * Return true if the CPU has not yet responded to the current grace period. */ -static int rcu_cpu_cur_gp(void) +static int rcu_cpu_blocking_cur_gp(void) { return rcu_preempt_ctrlblk.gpcpu != rcu_preempt_ctrlblk.gpnum; } @@ -229,7 +229,7 @@ void rcu_preempt_note_context_switch(void) * cannot end. */ list_add(&t->rcu_node_entry, &rcu_preempt_ctrlblk.blkd_tasks); - if (rcu_cpu_cur_gp()) + if (rcu_cpu_blocking_cur_gp()) rcu_preempt_ctrlblk.gp_tasks = &t->rcu_node_entry; } @@ -368,12 +368,16 @@ static void rcu_preempt_check_callbacks(void) { struct task_struct *t = current; - if (!rcu_preempt_running_reader() && rcu_preempt_gp_in_progress()) + if (rcu_preempt_gp_in_progress() && + (!rcu_preempt_running_reader() || + !rcu_cpu_blocking_cur_gp())) rcu_preempt_cpu_qs(); if (&rcu_preempt_ctrlblk.rcb.rcucblist != rcu_preempt_ctrlblk.rcb.donetail) raise_softirq(RCU_SOFTIRQ); - if (rcu_preempt_gp_in_progress() && rcu_preempt_running_reader()) + if (rcu_preempt_gp_in_progress() && + rcu_cpu_blocking_cur_gp() && + rcu_preempt_running_reader()) t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS; } -- cgit v0.10.2 From 9fbaf49c7f717740002d49eee1bbd03d89d8766a Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Sat, 28 Aug 2010 17:41:03 +0200 Subject: x86, kmemcheck: Remove double test The opcodes 0x2e and 0x3e are tested for in the first Group 2 line as well. The sematic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // @expression@ expression E; @@ ( * E || ... || E | * E && ... && E ) // Signed-off-by: Julia Lawall Reviewed-by: Pekka Enberg Cc: Vegard Nossum LKML-Reference: <1283010066-20935-5-git-send-email-julia@diku.dk> Signed-off-by: Ingo Molnar diff --git a/arch/x86/mm/kmemcheck/opcode.c b/arch/x86/mm/kmemcheck/opcode.c index 63c19e2..324aa3f 100644 --- a/arch/x86/mm/kmemcheck/opcode.c +++ b/arch/x86/mm/kmemcheck/opcode.c @@ -9,7 +9,7 @@ static bool opcode_is_prefix(uint8_t b) b == 0xf0 || b == 0xf2 || b == 0xf3 /* Group 2 */ || b == 0x2e || b == 0x36 || b == 0x3e || b == 0x26 - || b == 0x64 || b == 0x65 || b == 0x2e || b == 0x3e + || b == 0x64 || b == 0x65 /* Group 3 */ || b == 0x66 /* Group 4 */ -- cgit v0.10.2 From 81a294c44e973dc7182e4733421b7cb2daba3c29 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 30 Aug 2010 09:52:50 -0700 Subject: rcu: fix _oddness handling of verbose stall warnings CONFIG_RCU_CPU_STALL_VERBOSE depends on CONFIG_TREE_PREEMPT_RCU, but rcu_bootup_announce_oddness() complains if CONFIG_RCU_CPU_STALL_VERBOSE is not set even in the case of CONFIG_TREE_RCU. This commit therefore fixes rcu_bootup_announce_oddness() to avoid insisting on impossibilities. Reported-by: Guy Martin Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index e9e0bc74..71a4147 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -57,7 +57,7 @@ static void __init rcu_bootup_announce_oddness(void) printk(KERN_INFO "\tRCU-based detection of stalled CPUs is disabled.\n"); #endif -#ifndef CONFIG_RCU_CPU_STALL_VERBOSE +#if defined(CONFIG_TREE_PREEMPT_RCU) && !defined(CONFIG_RCU_CPU_STALL_VERBOSE) printk(KERN_INFO "\tVerbose stalled-CPUs detection is disabled.\n"); #endif #if NUM_RCU_LVL_4 != 0 -- cgit v0.10.2 From e6b04b6b5a3182ae36cf9a69f1aaaee432edc8ad Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 2 Sep 2010 13:52:45 +0100 Subject: x86-64: Fix unwind annotations in syscall stubs With the return address removed from the stack, these should really refer to their caller's register state. Signed-off-by: Jan Beulich Acked-by: Alexander van Heukelum LKML-Reference: <4C7FBA3D0200007800013F61@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 17be5ec..16aeff0 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -714,9 +714,8 @@ END(ptregscall_common) ENTRY(stub_execve) CFI_STARTPROC - popq %r11 - CFI_ADJUST_CFA_OFFSET -8 - CFI_REGISTER rip, r11 + addq $8, %rsp + PARTIAL_FRAME 0 SAVE_REST FIXUP_TOP_OF_STACK %r11 movq %rsp, %rcx @@ -735,7 +734,7 @@ END(stub_execve) ENTRY(stub_rt_sigreturn) CFI_STARTPROC addq $8, %rsp - CFI_ADJUST_CFA_OFFSET -8 + PARTIAL_FRAME 0 SAVE_REST movq %rsp,%rdi FIXUP_TOP_OF_STACK %r11 @@ -1445,7 +1444,6 @@ error_swapgs: error_sti: TRACE_IRQS_OFF ret - CFI_ENDPROC /* * There are two places in the kernel that can potentially fault with @@ -1470,6 +1468,7 @@ bstep_iret: /* Fix truncated RIP */ movq %rcx,RIP+8(%rsp) jmp error_swapgs + CFI_ENDPROC END(error_entry) -- cgit v0.10.2 From 1f130a783a796f147b080c594488b566c86007d0 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 2 Sep 2010 13:54:32 +0100 Subject: x86-64: Adjust frame type at paranoid_exit: As this isn't an exception or interrupt entry point, it doesn't have any of the hardware provide frame layouts active. Signed-off-by: Jan Beulich Acked-by: Alexander van Heukelum LKML-Reference: <4C7FBAA80200007800013F67@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 16aeff0..64dfe30 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1367,7 +1367,7 @@ paranoidzeroentry machine_check *machine_check_vector(%rip) /* ebx: no swapgs flag */ ENTRY(paranoid_exit) - INTR_FRAME + DEFAULT_FRAME DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF testl %ebx,%ebx /* swapgs needed? */ -- cgit v0.10.2 From b1cccb1bb01dc1cb89f58723a58c3d4988d44d94 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 2 Sep 2010 13:55:11 +0100 Subject: x86-64: Use symbolics instead of raw numbers in entry_64.S ... making the code a little less fragile. Also use pushq_cfi instead of raw CFI annotations in two more places, and add two missing annotations after stack pointer adjustments which got modified here anyway. Signed-off-by: Jan Beulich Acked-by: Alexander van Heukelum LKML-Reference: <4C7FBACF0200007800013F6A@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 64dfe30..6f30583 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -795,8 +795,8 @@ END(interrupt) /* 0(%rsp): ~(interrupt number) */ .macro interrupt func - subq $10*8, %rsp - CFI_ADJUST_CFA_OFFSET 10*8 + subq $ORIG_RAX-ARGOFFSET+8, %rsp + CFI_ADJUST_CFA_OFFSET ORIG_RAX-ARGOFFSET+8 call save_args PARTIAL_FRAME 0 call \func @@ -1035,8 +1035,8 @@ ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq_cfi $-1 /* ORIG_RAX: no syscall to restart */ - subq $15*8,%rsp - CFI_ADJUST_CFA_OFFSET 15*8 + subq $ORIG_RAX-R15, %rsp + CFI_ADJUST_CFA_OFFSET ORIG_RAX-R15 call error_entry DEFAULT_FRAME 0 movq %rsp,%rdi /* pt_regs pointer */ @@ -1051,9 +1051,9 @@ END(\sym) ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $-1 /* ORIG_RAX: no syscall to restart */ - CFI_ADJUST_CFA_OFFSET 8 - subq $15*8, %rsp + pushq_cfi $-1 /* ORIG_RAX: no syscall to restart */ + subq $ORIG_RAX-R15, %rsp + CFI_ADJUST_CFA_OFFSET ORIG_RAX-R15 call save_paranoid TRACE_IRQS_OFF movq %rsp,%rdi /* pt_regs pointer */ @@ -1069,9 +1069,9 @@ END(\sym) ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $-1 /* ORIG_RAX: no syscall to restart */ - CFI_ADJUST_CFA_OFFSET 8 - subq $15*8, %rsp + pushq_cfi $-1 /* ORIG_RAX: no syscall to restart */ + subq $ORIG_RAX-R15, %rsp + CFI_ADJUST_CFA_OFFSET ORIG_RAX-R15 call save_paranoid TRACE_IRQS_OFF movq %rsp,%rdi /* pt_regs pointer */ @@ -1088,8 +1088,8 @@ END(\sym) ENTRY(\sym) XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - subq $15*8,%rsp - CFI_ADJUST_CFA_OFFSET 15*8 + subq $ORIG_RAX-R15, %rsp + CFI_ADJUST_CFA_OFFSET ORIG_RAX-R15 call error_entry DEFAULT_FRAME 0 movq %rsp,%rdi /* pt_regs pointer */ @@ -1106,8 +1106,8 @@ END(\sym) ENTRY(\sym) XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - subq $15*8,%rsp - CFI_ADJUST_CFA_OFFSET 15*8 + subq $ORIG_RAX-R15, %rsp + CFI_ADJUST_CFA_OFFSET ORIG_RAX-R15 call save_paranoid DEFAULT_FRAME 0 TRACE_IRQS_OFF @@ -1497,8 +1497,8 @@ ENTRY(nmi) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq_cfi $-1 - subq $15*8, %rsp - CFI_ADJUST_CFA_OFFSET 15*8 + subq $ORIG_RAX-R15, %rsp + CFI_ADJUST_CFA_OFFSET ORIG_RAX-R15 call save_paranoid DEFAULT_FRAME 0 /* paranoidentry do_nmi, 0; without TRACE_IRQS_OFF */ -- cgit v0.10.2 From a34107b5577968dc53cf9c2195c7c2d4a2caf9ce Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 2 Sep 2010 14:04:16 +0100 Subject: i386: Add unwind directives to syscall ptregs stubs When these stubs are actual functions (i.e. having a return instruction) and have stack manipulation instructions in them, they should also be annotated to allow unwinding through them. Signed-off-by: Jan Beulich Acked-by: Alexander van Heukelum LKML-Reference: <4C7FBCF00200007800013F99@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 227d009..d9b950e 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -750,14 +750,18 @@ ptregs_##name: \ #define PTREGSCALL3(name) \ ALIGN; \ ptregs_##name: \ + CFI_STARTPROC; \ leal 4(%esp),%eax; \ - pushl %eax; \ + pushl_cfi %eax; \ movl PT_EDX(%eax),%ecx; \ movl PT_ECX(%eax),%edx; \ movl PT_EBX(%eax),%eax; \ call sys_##name; \ addl $4,%esp; \ - ret + CFI_ADJUST_CFA_OFFSET -4; \ + ret; \ + CFI_ENDPROC; \ +ENDPROC(ptregs_##name) PTREGSCALL1(iopl) PTREGSCALL0(fork) @@ -772,15 +776,19 @@ PTREGSCALL1(vm86old) /* Clone is an oddball. The 4th arg is in %edi */ ALIGN; ptregs_clone: + CFI_STARTPROC leal 4(%esp),%eax - pushl %eax - pushl PT_EDI(%eax) + pushl_cfi %eax + pushl_cfi PT_EDI(%eax) movl PT_EDX(%eax),%ecx movl PT_ECX(%eax),%edx movl PT_EBX(%eax),%eax call sys_clone addl $8,%esp + CFI_ADJUST_CFA_OFFSET -8 ret + CFI_ENDPROC +ENDPROC(ptregs_clone) .macro FIXUP_ESPFIX_STACK /* -- cgit v0.10.2 From df5d1874ce1a1f0e0eceff4fa3a9d45620243a68 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 2 Sep 2010 14:07:16 +0100 Subject: x86: Use {push,pop}{l,q}_cfi in more places ... plus additionally introduce {push,pop}f{l,q}_cfi. All in the hope that the code becomes better readable this way (it gets quite a bit smaller in any case). Signed-off-by: Jan Beulich Acked-by: Alexander van Heukelum LKML-Reference: <4C7FBDA40200007800013FAF@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/dwarf2.h b/arch/x86/include/asm/dwarf2.h index 733f7e9..3260991 100644 --- a/arch/x86/include/asm/dwarf2.h +++ b/arch/x86/include/asm/dwarf2.h @@ -89,6 +89,16 @@ CFI_ADJUST_CFA_OFFSET -8 .endm + .macro pushfq_cfi + pushfq + CFI_ADJUST_CFA_OFFSET 8 + .endm + + .macro popfq_cfi + popfq + CFI_ADJUST_CFA_OFFSET -8 + .endm + .macro movq_cfi reg offset=0 movq %\reg, \offset(%rsp) CFI_REL_OFFSET \reg, \offset @@ -109,6 +119,16 @@ CFI_ADJUST_CFA_OFFSET -4 .endm + .macro pushfl_cfi + pushfl + CFI_ADJUST_CFA_OFFSET 4 + .endm + + .macro popfl_cfi + popfl + CFI_ADJUST_CFA_OFFSET -4 + .endm + .macro movl_cfi reg offset=0 movl %\reg, \offset(%esp) CFI_REL_OFFSET \reg, \offset diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index d9b950e..9fb188d 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -115,8 +115,7 @@ /* unfortunately push/pop can't be no-op */ .macro PUSH_GS - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 .endm .macro POP_GS pop=0 addl $(4 + \pop), %esp @@ -140,14 +139,12 @@ #else /* CONFIG_X86_32_LAZY_GS */ .macro PUSH_GS - pushl %gs - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %gs /*CFI_REL_OFFSET gs, 0*/ .endm .macro POP_GS pop=0 -98: popl %gs - CFI_ADJUST_CFA_OFFSET -4 +98: popl_cfi %gs /*CFI_RESTORE gs*/ .if \pop <> 0 add $\pop, %esp @@ -195,35 +192,25 @@ .macro SAVE_ALL cld PUSH_GS - pushl %fs - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %fs /*CFI_REL_OFFSET fs, 0;*/ - pushl %es - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %es /*CFI_REL_OFFSET es, 0;*/ - pushl %ds - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ds /*CFI_REL_OFFSET ds, 0;*/ - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax CFI_REL_OFFSET eax, 0 - pushl %ebp - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ebp CFI_REL_OFFSET ebp, 0 - pushl %edi - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %edi CFI_REL_OFFSET edi, 0 - pushl %esi - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %esi CFI_REL_OFFSET esi, 0 - pushl %edx - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %edx CFI_REL_OFFSET edx, 0 - pushl %ecx - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ecx CFI_REL_OFFSET ecx, 0 - pushl %ebx - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ebx CFI_REL_OFFSET ebx, 0 movl $(__USER_DS), %edx movl %edx, %ds @@ -234,39 +221,29 @@ .endm .macro RESTORE_INT_REGS - popl %ebx - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %ebx CFI_RESTORE ebx - popl %ecx - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %ecx CFI_RESTORE ecx - popl %edx - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %edx CFI_RESTORE edx - popl %esi - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %esi CFI_RESTORE esi - popl %edi - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %edi CFI_RESTORE edi - popl %ebp - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %ebp CFI_RESTORE ebp - popl %eax - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %eax CFI_RESTORE eax .endm .macro RESTORE_REGS pop=0 RESTORE_INT_REGS -1: popl %ds - CFI_ADJUST_CFA_OFFSET -4 +1: popl_cfi %ds /*CFI_RESTORE ds;*/ -2: popl %es - CFI_ADJUST_CFA_OFFSET -4 +2: popl_cfi %es /*CFI_RESTORE es;*/ -3: popl %fs - CFI_ADJUST_CFA_OFFSET -4 +3: popl_cfi %fs /*CFI_RESTORE fs;*/ POP_GS \pop .pushsection .fixup, "ax" @@ -320,16 +297,12 @@ ENTRY(ret_from_fork) CFI_STARTPROC - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax call schedule_tail GET_THREAD_INFO(%ebp) - popl %eax - CFI_ADJUST_CFA_OFFSET -4 - pushl $0x0202 # Reset kernel eflags - CFI_ADJUST_CFA_OFFSET 4 - popfl - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %eax + pushl_cfi $0x0202 # Reset kernel eflags + popfl_cfi jmp syscall_exit CFI_ENDPROC END(ret_from_fork) @@ -409,29 +382,23 @@ sysenter_past_esp: * enough kernel state to call TRACE_IRQS_OFF can be called - but * we immediately enable interrupts at that point anyway. */ - pushl $(__USER_DS) - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $(__USER_DS) /*CFI_REL_OFFSET ss, 0*/ - pushl %ebp - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ebp CFI_REL_OFFSET esp, 0 - pushfl + pushfl_cfi orl $X86_EFLAGS_IF, (%esp) - CFI_ADJUST_CFA_OFFSET 4 - pushl $(__USER_CS) - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $(__USER_CS) /*CFI_REL_OFFSET cs, 0*/ /* * Push current_thread_info()->sysenter_return to the stack. * A tiny bit of offset fixup is necessary - 4*4 means the 4 words * pushed above; +8 corresponds to copy_thread's esp0 setting. */ - pushl (TI_sysenter_return-THREAD_SIZE+8+4*4)(%esp) - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi (TI_sysenter_return-THREAD_SIZE+8+4*4)(%esp) CFI_REL_OFFSET eip, 0 - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax SAVE_ALL ENABLE_INTERRUPTS(CLBR_NONE) @@ -486,8 +453,7 @@ sysenter_audit: movl %eax,%edx /* 2nd arg: syscall number */ movl $AUDIT_ARCH_I386,%eax /* 1st arg: audit arch */ call audit_syscall_entry - pushl %ebx - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ebx movl PT_EAX(%esp),%eax /* reload syscall number */ jmp sysenter_do_call @@ -529,8 +495,7 @@ ENDPROC(ia32_sysenter_target) # system call handler stub ENTRY(system_call) RING0_INT_FRAME # can't unwind into user space anyway - pushl %eax # save orig_eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax # save orig_eax SAVE_ALL GET_THREAD_INFO(%ebp) # system call tracing in operation / emulation @@ -566,7 +531,6 @@ restore_all_notrace: je ldt_ss # returning to user-space with LDT SS restore_nocheck: RESTORE_REGS 4 # skip orig_eax/error_code - CFI_ADJUST_CFA_OFFSET -4 irq_return: INTERRUPT_RETURN .section .fixup,"ax" @@ -619,10 +583,8 @@ ldt_ss: shr $16, %edx mov %dl, GDT_ESPFIX_SS + 4 /* bits 16..23 */ mov %dh, GDT_ESPFIX_SS + 7 /* bits 24..31 */ - pushl $__ESPFIX_SS - CFI_ADJUST_CFA_OFFSET 4 - push %eax /* new kernel esp */ - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $__ESPFIX_SS + pushl_cfi %eax /* new kernel esp */ /* Disable interrupts, but do not irqtrace this section: we * will soon execute iret and the tracer was already set to * the irqstate after the iret */ @@ -666,11 +628,9 @@ work_notifysig: # deal with pending signals and ALIGN work_notifysig_v86: - pushl %ecx # save ti_flags for do_notify_resume - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ecx # save ti_flags for do_notify_resume call save_v86_state # %eax contains pt_regs pointer - popl %ecx - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %ecx movl %eax, %esp #else movl %esp, %eax @@ -803,10 +763,8 @@ ENDPROC(ptregs_clone) mov GDT_ESPFIX_SS + 7, %ah /* bits 24..31 */ shl $16, %eax addl %esp, %eax /* the adjusted stack pointer */ - pushl $__KERNEL_DS - CFI_ADJUST_CFA_OFFSET 4 - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $__KERNEL_DS + pushl_cfi %eax lss (%esp), %esp /* switch to the normal stack segment */ CFI_ADJUST_CFA_OFFSET -8 .endm @@ -843,8 +801,7 @@ vector=FIRST_EXTERNAL_VECTOR .if vector <> FIRST_EXTERNAL_VECTOR CFI_ADJUST_CFA_OFFSET -4 .endif -1: pushl $(~vector+0x80) /* Note: always in signed byte range */ - CFI_ADJUST_CFA_OFFSET 4 +1: pushl_cfi $(~vector+0x80) /* Note: always in signed byte range */ .if ((vector-FIRST_EXTERNAL_VECTOR)%7) <> 6 jmp 2f .endif @@ -884,8 +841,7 @@ ENDPROC(common_interrupt) #define BUILD_INTERRUPT3(name, nr, fn) \ ENTRY(name) \ RING0_INT_FRAME; \ - pushl $~(nr); \ - CFI_ADJUST_CFA_OFFSET 4; \ + pushl_cfi $~(nr); \ SAVE_ALL; \ TRACE_IRQS_OFF \ movl %esp,%eax; \ @@ -901,21 +857,18 @@ ENDPROC(name) ENTRY(coprocessor_error) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_coprocessor_error - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 + pushl_cfi $do_coprocessor_error jmp error_code CFI_ENDPROC END(coprocessor_error) ENTRY(simd_coprocessor_error) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 #ifdef CONFIG_X86_INVD_BUG /* AMD 486 bug: invd from userspace calls exception 19 instead of #GP */ -661: pushl $do_general_protection +661: pushl_cfi $do_general_protection 662: .section .altinstructions,"a" .balign 4 @@ -930,19 +883,16 @@ ENTRY(simd_coprocessor_error) 664: .previous #else - pushl $do_simd_coprocessor_error + pushl_cfi $do_simd_coprocessor_error #endif - CFI_ADJUST_CFA_OFFSET 4 jmp error_code CFI_ENDPROC END(simd_coprocessor_error) ENTRY(device_not_available) RING0_INT_FRAME - pushl $-1 # mark this as an int - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_device_not_available - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $-1 # mark this as an int + pushl_cfi $do_device_not_available jmp error_code CFI_ENDPROC END(device_not_available) @@ -964,82 +914,68 @@ END(native_irq_enable_sysexit) ENTRY(overflow) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_overflow - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 + pushl_cfi $do_overflow jmp error_code CFI_ENDPROC END(overflow) ENTRY(bounds) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_bounds - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 + pushl_cfi $do_bounds jmp error_code CFI_ENDPROC END(bounds) ENTRY(invalid_op) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_invalid_op - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 + pushl_cfi $do_invalid_op jmp error_code CFI_ENDPROC END(invalid_op) ENTRY(coprocessor_segment_overrun) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_coprocessor_segment_overrun - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 + pushl_cfi $do_coprocessor_segment_overrun jmp error_code CFI_ENDPROC END(coprocessor_segment_overrun) ENTRY(invalid_TSS) RING0_EC_FRAME - pushl $do_invalid_TSS - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $do_invalid_TSS jmp error_code CFI_ENDPROC END(invalid_TSS) ENTRY(segment_not_present) RING0_EC_FRAME - pushl $do_segment_not_present - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $do_segment_not_present jmp error_code CFI_ENDPROC END(segment_not_present) ENTRY(stack_segment) RING0_EC_FRAME - pushl $do_stack_segment - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $do_stack_segment jmp error_code CFI_ENDPROC END(stack_segment) ENTRY(alignment_check) RING0_EC_FRAME - pushl $do_alignment_check - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $do_alignment_check jmp error_code CFI_ENDPROC END(alignment_check) ENTRY(divide_error) RING0_INT_FRAME - pushl $0 # no error code - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_divide_error - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 # no error code + pushl_cfi $do_divide_error jmp error_code CFI_ENDPROC END(divide_error) @@ -1047,10 +983,8 @@ END(divide_error) #ifdef CONFIG_X86_MCE ENTRY(machine_check) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 - pushl machine_check_vector - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 + pushl_cfi machine_check_vector jmp error_code CFI_ENDPROC END(machine_check) @@ -1058,10 +992,8 @@ END(machine_check) ENTRY(spurious_interrupt_bug) RING0_INT_FRAME - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 - pushl $do_spurious_interrupt_bug - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 + pushl_cfi $do_spurious_interrupt_bug jmp error_code CFI_ENDPROC END(spurious_interrupt_bug) @@ -1092,8 +1024,7 @@ ENTRY(xen_sysenter_target) ENTRY(xen_hypervisor_callback) CFI_STARTPROC - pushl $0 - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $0 SAVE_ALL TRACE_IRQS_OFF @@ -1129,23 +1060,20 @@ ENDPROC(xen_hypervisor_callback) # We distinguish between categories by maintaining a status value in EAX. ENTRY(xen_failsafe_callback) CFI_STARTPROC - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax movl $1,%eax 1: mov 4(%esp),%ds 2: mov 8(%esp),%es 3: mov 12(%esp),%fs 4: mov 16(%esp),%gs testl %eax,%eax - popl %eax - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %eax lea 16(%esp),%esp CFI_ADJUST_CFA_OFFSET -16 jz 5f addl $16,%esp jmp iret_exc # EAX != 0 => Category 2 (Bad IRET) -5: pushl $0 # EAX == 0 => Category 1 (Bad segment) - CFI_ADJUST_CFA_OFFSET 4 +5: pushl_cfi $0 # EAX == 0 => Category 1 (Bad segment) SAVE_ALL jmp ret_from_exception CFI_ENDPROC @@ -1295,40 +1223,29 @@ syscall_table_size=(.-sys_call_table) ENTRY(page_fault) RING0_EC_FRAME - pushl $do_page_fault - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $do_page_fault ALIGN error_code: /* the function address is in %gs's slot on the stack */ - pushl %fs - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %fs /*CFI_REL_OFFSET fs, 0*/ - pushl %es - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %es /*CFI_REL_OFFSET es, 0*/ - pushl %ds - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ds /*CFI_REL_OFFSET ds, 0*/ - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax CFI_REL_OFFSET eax, 0 - pushl %ebp - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ebp CFI_REL_OFFSET ebp, 0 - pushl %edi - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %edi CFI_REL_OFFSET edi, 0 - pushl %esi - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %esi CFI_REL_OFFSET esi, 0 - pushl %edx - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %edx CFI_REL_OFFSET edx, 0 - pushl %ecx - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ecx CFI_REL_OFFSET ecx, 0 - pushl %ebx - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ebx CFI_REL_OFFSET ebx, 0 cld movl $(__KERNEL_PERCPU), %ecx @@ -1370,12 +1287,9 @@ END(page_fault) movl TSS_sysenter_sp0 + \offset(%esp), %esp CFI_DEF_CFA esp, 0 CFI_UNDEFINED eip - pushfl - CFI_ADJUST_CFA_OFFSET 4 - pushl $__KERNEL_CS - CFI_ADJUST_CFA_OFFSET 4 - pushl $sysenter_past_esp - CFI_ADJUST_CFA_OFFSET 4 + pushfl_cfi + pushl_cfi $__KERNEL_CS + pushl_cfi $sysenter_past_esp CFI_REL_OFFSET eip, 0 .endm @@ -1385,8 +1299,7 @@ ENTRY(debug) jne debug_stack_correct FIX_STACK 12, debug_stack_correct, debug_esp_fix_insn debug_stack_correct: - pushl $-1 # mark this as an int - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $-1 # mark this as an int SAVE_ALL TRACE_IRQS_OFF xorl %edx,%edx # error code 0 @@ -1406,32 +1319,27 @@ END(debug) */ ENTRY(nmi) RING0_INT_FRAME - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax movl %ss, %eax cmpw $__ESPFIX_SS, %ax - popl %eax - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %eax je nmi_espfix_stack cmpl $ia32_sysenter_target,(%esp) je nmi_stack_fixup - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax movl %esp,%eax /* Do not access memory above the end of our stack page, * it might not exist. */ andl $(THREAD_SIZE-1),%eax cmpl $(THREAD_SIZE-20),%eax - popl %eax - CFI_ADJUST_CFA_OFFSET -4 + popl_cfi %eax jae nmi_stack_correct cmpl $ia32_sysenter_target,12(%esp) je nmi_debug_stack_check nmi_stack_correct: /* We have a RING0_INT_FRAME here */ - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax SAVE_ALL xorl %edx,%edx # zero error code movl %esp,%eax # pt_regs pointer @@ -1460,18 +1368,14 @@ nmi_espfix_stack: * * create the pointer to lss back */ - pushl %ss - CFI_ADJUST_CFA_OFFSET 4 - pushl %esp - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %ss + pushl_cfi %esp addl $4, (%esp) /* copy the iret frame of 12 bytes */ .rept 3 - pushl 16(%esp) - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi 16(%esp) .endr - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi %eax SAVE_ALL FIXUP_ESPFIX_STACK # %eax == %esp xorl %edx,%edx # zero error code @@ -1485,8 +1389,7 @@ END(nmi) ENTRY(int3) RING0_INT_FRAME - pushl $-1 # mark this as an int - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $-1 # mark this as an int SAVE_ALL TRACE_IRQS_OFF xorl %edx,%edx # zero error code @@ -1498,8 +1401,7 @@ END(int3) ENTRY(general_protection) RING0_EC_FRAME - pushl $do_general_protection - CFI_ADJUST_CFA_OFFSET 4 + pushl_cfi $do_general_protection jmp error_code CFI_ENDPROC END(general_protection) diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 6f30583..8851a2b 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -213,23 +213,17 @@ ENDPROC(native_usergs_sysret64) .macro FAKE_STACK_FRAME child_rip /* push in order ss, rsp, eflags, cs, rip */ xorl %eax, %eax - pushq $__KERNEL_DS /* ss */ - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi $__KERNEL_DS /* ss */ /*CFI_REL_OFFSET ss,0*/ - pushq %rax /* rsp */ - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi %rax /* rsp */ CFI_REL_OFFSET rsp,0 - pushq $X86_EFLAGS_IF /* eflags - interrupts on */ - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi $X86_EFLAGS_IF /* eflags - interrupts on */ /*CFI_REL_OFFSET rflags,0*/ - pushq $__KERNEL_CS /* cs */ - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi $__KERNEL_CS /* cs */ /*CFI_REL_OFFSET cs,0*/ - pushq \child_rip /* rip */ - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi \child_rip /* rip */ CFI_REL_OFFSET rip,0 - pushq %rax /* orig rax */ - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi %rax /* orig rax */ .endm .macro UNFAKE_STACK_FRAME @@ -398,10 +392,8 @@ ENTRY(ret_from_fork) LOCK ; btr $TIF_FORK,TI_flags(%r8) - push kernel_eflags(%rip) - CFI_ADJUST_CFA_OFFSET 8 - popf # reset kernel eflags - CFI_ADJUST_CFA_OFFSET -8 + pushq_cfi kernel_eflags(%rip) + popfq_cfi # reset kernel eflags call schedule_tail # rdi: 'prev' task parameter @@ -521,11 +513,9 @@ sysret_careful: jnc sysret_signal TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) - pushq %rdi - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi %rdi call schedule - popq %rdi - CFI_ADJUST_CFA_OFFSET -8 + popq_cfi %rdi jmp sysret_check /* Handle a signal */ @@ -634,11 +624,9 @@ int_careful: jnc int_very_careful TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) - pushq %rdi - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi %rdi call schedule - popq %rdi - CFI_ADJUST_CFA_OFFSET -8 + popq_cfi %rdi DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF jmp int_with_check @@ -652,12 +640,10 @@ int_check_syscall_exit_work: /* Check for syscall exit trace */ testl $_TIF_WORK_SYSCALL_EXIT,%edx jz int_signal - pushq %rdi - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi %rdi leaq 8(%rsp),%rdi # &ptregs -> arg1 call syscall_trace_leave - popq %rdi - CFI_ADJUST_CFA_OFFSET -8 + popq_cfi %rdi andl $~(_TIF_WORK_SYSCALL_EXIT|_TIF_SYSCALL_EMU),%edi jmp int_restore_rest @@ -765,8 +751,7 @@ vector=FIRST_EXTERNAL_VECTOR .if vector <> FIRST_EXTERNAL_VECTOR CFI_ADJUST_CFA_OFFSET -8 .endif -1: pushq $(~vector+0x80) /* Note: always in signed byte range */ - CFI_ADJUST_CFA_OFFSET 8 +1: pushq_cfi $(~vector+0x80) /* Note: always in signed byte range */ .if ((vector-FIRST_EXTERNAL_VECTOR)%7) <> 6 jmp 2f .endif @@ -821,6 +806,7 @@ ret_from_intr: TRACE_IRQS_OFF decl PER_CPU_VAR(irq_count) leaveq + CFI_RESTORE rbp CFI_DEF_CFA_REGISTER rsp CFI_ADJUST_CFA_OFFSET -8 exit_intr: @@ -902,11 +888,9 @@ retint_careful: jnc retint_signal TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) - pushq %rdi - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi %rdi call schedule - popq %rdi - CFI_ADJUST_CFA_OFFSET -8 + popq_cfi %rdi GET_THREAD_INFO(%rcx) DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF @@ -955,8 +939,7 @@ END(common_interrupt) .macro apicinterrupt num sym do_sym ENTRY(\sym) INTR_FRAME - pushq $~(\num) - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi $~(\num) interrupt \do_sym jmp ret_from_intr CFI_ENDPROC @@ -1138,16 +1121,14 @@ zeroentry simd_coprocessor_error do_simd_coprocessor_error /* edi: new selector */ ENTRY(native_load_gs_index) CFI_STARTPROC - pushf - CFI_ADJUST_CFA_OFFSET 8 + pushfq_cfi DISABLE_INTERRUPTS(CLBR_ANY & ~CLBR_RDI) SWAPGS gs_change: movl %edi,%gs 2: mfence /* workaround */ SWAPGS - popf - CFI_ADJUST_CFA_OFFSET -8 + popfq_cfi ret CFI_ENDPROC END(native_load_gs_index) @@ -1214,8 +1195,7 @@ END(kernel_execve) /* Call softirq on interrupt stack. Interrupts are off. */ ENTRY(call_softirq) CFI_STARTPROC - push %rbp - CFI_ADJUST_CFA_OFFSET 8 + pushq_cfi %rbp CFI_REL_OFFSET rbp,0 mov %rsp,%rbp CFI_DEF_CFA_REGISTER rbp @@ -1224,6 +1204,7 @@ ENTRY(call_softirq) push %rbp # backlink for old unwinder call __do_softirq leaveq + CFI_RESTORE rbp CFI_DEF_CFA_REGISTER rsp CFI_ADJUST_CFA_OFFSET -8 decl PER_CPU_VAR(irq_count) -- cgit v0.10.2 From 7fe977dab356fbd7e86aa10bf83891761107c57c Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 2 Sep 2010 14:01:58 +0100 Subject: i386: Make kernel_execve() suitable for stack unwinding The explicit saving and restoring of %ebx was confusing stack unwind data consumers, and it is plain unnecessary to do this within the asm(), since that was only introduced for PIC user mode consumers of the original _syscall3() macro this was derived from. Signed-off-by: Jan Beulich Cc: Arnd Bergmann LKML-Reference: <4C7FBC660200007800013F95@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/sys_i386_32.c b/arch/x86/kernel/sys_i386_32.c index d5e0662..0b0cb5f 100644 --- a/arch/x86/kernel/sys_i386_32.c +++ b/arch/x86/kernel/sys_i386_32.c @@ -33,8 +33,8 @@ int kernel_execve(const char *filename, const char *const envp[]) { long __res; - asm volatile ("push %%ebx ; movl %2,%%ebx ; int $0x80 ; pop %%ebx" + asm volatile ("int $0x80" : "=a" (__res) - : "0" (__NR_execve), "ri" (filename), "c" (argv), "d" (envp) : "memory"); + : "0" (__NR_execve), "b" (filename), "c" (argv), "d" (envp) : "memory"); return __res; } -- cgit v0.10.2 From 1c5f50ee347daea013671f718b70cd6bf497bef9 Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Fri, 3 Sep 2010 17:04:07 +0800 Subject: x86, mm: fix uninitialized addr in kernel_physical_mapping_init() This re-adds the lost chunk in commit 9b861528a80. Reported-by: Stephen Rothwell Signed-off-by: Wu Fengguang Cc: Peter Zijlstra Cc: Haicheng Li Cc: Andi Kleen LKML-Reference: <20100903090407.GA19771@localhost> Signed-off-by: Ingo Molnar diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 64e7bc2..74f0f35 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -570,6 +570,7 @@ kernel_physical_mapping_init(unsigned long start, start = (unsigned long)__va(start); end = (unsigned long)__va(end); + addr = start; for (; start < end; start = next) { pgd_t *pgd = pgd_offset_k(start); -- cgit v0.10.2 From 57ab43e33122ffdc2eebca5d6de035699f0a8c06 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 3 Sep 2010 18:39:39 +0200 Subject: x86, GART: Remove superfluous AMD64_GARTEN There is a GARTEN so use that and drop the duplicate. Signed-off-by: Borislav Petkov Cc: Dave Airlie Cc: FUJITA Tomonori LKML-Reference: <1283531981-7495-2-git-send-email-bp@amd64.org> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/gart.h b/arch/x86/include/asm/gart.h index 4ac5b0f..fba0a72 100644 --- a/arch/x86/include/asm/gart.h +++ b/arch/x86/include/asm/gart.h @@ -27,7 +27,6 @@ extern int fix_aperture; #define AMD64_GARTAPERTUREBASE 0x94 #define AMD64_GARTTABLEBASE 0x98 #define AMD64_GARTCACHECTL 0x9c -#define AMD64_GARTEN (1<<0) #ifdef CONFIG_GART_IOMMU extern int gart_iommu_aperture; diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c index a2e0caf..6fabd40 100644 --- a/arch/x86/kernel/aperture_64.c +++ b/arch/x86/kernel/aperture_64.c @@ -307,7 +307,7 @@ void __init early_gart_iommu_check(void) continue; ctl = read_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL); - aper_enabled = ctl & AMD64_GARTEN; + aper_enabled = ctl & GARTEN; aper_order = (ctl >> 1) & 7; aper_size = (32 * 1024 * 1024) << aper_order; aper_base = read_pci_config(bus, slot, 3, AMD64_GARTAPERTUREBASE) & 0x7fff; @@ -362,7 +362,7 @@ void __init early_gart_iommu_check(void) continue; ctl = read_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL); - ctl &= ~AMD64_GARTEN; + ctl &= ~GARTEN; write_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL, ctl); } } diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c index 70312da..bfe372b 100644 --- a/drivers/char/agp/amd64-agp.c +++ b/drivers/char/agp/amd64-agp.c @@ -199,7 +199,7 @@ static void amd64_cleanup(void) struct pci_dev *dev = k8_northbridges[i]; /* disable gart translation */ pci_read_config_dword(dev, AMD64_GARTAPERTURECTL, &tmp); - tmp &= ~AMD64_GARTEN; + tmp &= ~GARTEN; pci_write_config_dword(dev, AMD64_GARTAPERTURECTL, tmp); } } -- cgit v0.10.2 From 260133ab658bd2b80e07832a878e00405e19ff43 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 3 Sep 2010 18:39:40 +0200 Subject: x86, GART: Disable GART table walk probes Current code tramples over bit F3x90[6] which can be used to disable GART table walk probes. However, this bit should be set for performance reasons (speed up GART table walks). We are allowed to do that since we put GART tables in UC memory later anyway. Make it so. Signed-off-by: Borislav Petkov Cc: Dave Airlie Cc: FUJITA Tomonori LKML-Reference: <1283531981-7495-3-git-send-email-bp@amd64.org> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/gart.h b/arch/x86/include/asm/gart.h index fba0a72..bf357f9 100644 --- a/arch/x86/include/asm/gart.h +++ b/arch/x86/include/asm/gart.h @@ -17,6 +17,7 @@ extern int fix_aperture; #define GARTEN (1<<0) #define DISGARTCPU (1<<4) #define DISGARTIO (1<<5) +#define DISTLBWALKPRB (1<<6) /* GART cache control register bits. */ #define INVGART (1<<0) @@ -56,6 +57,19 @@ static inline void gart_iommu_hole_init(void) extern int agp_amd64_init(void); +static inline void gart_set_size_and_enable(struct pci_dev *dev, u32 order) +{ + u32 ctl; + + /* + * Don't enable translation but enable GART IO and CPU accesses. + * Also, set DISTLBWALKPRB since GART tables memory is UC. + */ + ctl = DISTLBWALKPRB | order << 1; + + pci_write_config_dword(dev, AMD64_GARTAPERTURECTL, ctl); +} + static inline void enable_gart_translation(struct pci_dev *dev, u64 addr) { u32 tmp, ctl; diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c index 6fabd40..c9cb173 100644 --- a/arch/x86/kernel/aperture_64.c +++ b/arch/x86/kernel/aperture_64.c @@ -505,8 +505,13 @@ out: /* Fix up the north bridges */ for (i = 0; i < ARRAY_SIZE(bus_dev_ranges); i++) { - int bus; - int dev_base, dev_limit; + int bus, dev_base, dev_limit; + + /* + * Don't enable translation yet but enable GART IO and CPU + * accesses and set DISTLBWALKPRB since GART table memory is UC. + */ + u32 ctl = DISTLBWALKPRB | aper_order << 1; bus = bus_dev_ranges[i].bus; dev_base = bus_dev_ranges[i].dev_base; @@ -515,10 +520,7 @@ out: if (!early_is_k8_nb(read_pci_config(bus, slot, 3, 0x00))) continue; - /* Don't enable translation yet. That is done later. - Assume this BIOS didn't initialise the GART so - just overwrite all previous bits */ - write_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL, aper_order << 1); + write_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL, ctl); write_pci_config(bus, slot, 3, AMD64_GARTAPERTUREBASE, aper_alloc >> 25); } } diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index 0f7f130..6015ee1 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -601,7 +601,7 @@ static void gart_fixup_northbridges(struct sys_device *dev) * Don't enable translations just yet. That is the next * step. Restore the pre-suspend aperture settings. */ - pci_write_config_dword(dev, AMD64_GARTAPERTURECTL, aperture_order << 1); + gart_set_size_and_enable(dev, aperture_order); pci_write_config_dword(dev, AMD64_GARTAPERTUREBASE, aperture_alloc >> 25); } } diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c index bfe372b..564808a 100644 --- a/drivers/char/agp/amd64-agp.c +++ b/drivers/char/agp/amd64-agp.c @@ -313,7 +313,7 @@ static __devinit int fix_northbridge(struct pci_dev *nb, struct pci_dev *agp, if (order < 0 || !agp_aperture_valid(aper, (32*1024*1024)<> 25); return 0; -- cgit v0.10.2 From 1b13fe6a6e9986dbc079cbb05090be75edbffa5d Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 3 Sep 2010 18:39:41 +0200 Subject: AGP: Warn when GATT memory cannot be set to UC This is one of those paranoid checks which should at least tell us that something is about to go haywire after we've disabled GART table walk probes which is done by default now on AMD. Signed-off-by: Borislav Petkov Cc: Dave Airlie Cc: FUJITA Tomonori LKML-Reference: <1283531981-7495-4-git-send-email-bp@amd64.org> Signed-off-by: Ingo Molnar diff --git a/drivers/char/agp/generic.c b/drivers/char/agp/generic.c index d2abf51..64255ce 100644 --- a/drivers/char/agp/generic.c +++ b/drivers/char/agp/generic.c @@ -984,7 +984,9 @@ int agp_generic_create_gatt_table(struct agp_bridge_data *bridge) bridge->driver->cache_flush(); #ifdef CONFIG_X86 - set_memory_uc((unsigned long)table, 1 << page_order); + if (set_memory_uc((unsigned long)table, 1 << page_order)) + printk(KERN_WARNING "Could not set GATT table memory to UC!"); + bridge->gatt_table = (void *)table; #else bridge->gatt_table = ioremap_nocache(virt_to_phys(table), -- cgit v0.10.2 From d9fadd7ba99a67030783a212bcb17d11f0678433 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Thu, 2 Sep 2010 15:37:10 +0200 Subject: x86, AMD: Remove needless CPU family check (for L3 cache info) Old 32-bit AMD CPUs (all w/o L3 cache) should always return 0 for cpuid_edx(0x80000006). For unknown reason the 32-bit implementation differed from the 64-bit implementation. See commit 67cddd94799 ("i386: Add L3 cache support to AMD CPUID4 emulation"). The current check is the result of the x86 merge. Signed-off-by: Andreas Herrmann Cc: Andi Kleen LKML-Reference: <20100902133710.GA5449@loge.amd.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index fc563fa..0f0ace5d 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -540,7 +540,7 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c) #endif if (c->extended_cpuid_level >= 0x80000006) { - if ((c->x86 >= 0x0f) && (cpuid_edx(0x80000006) & 0xf000)) + if (cpuid_edx(0x80000006) & 0xf000) num_cache_leaves = 4; else num_cache_leaves = 3; -- cgit v0.10.2 From 592091c0e21655bfbdf68741dd5a920c2ac2bbe6 Mon Sep 17 00:00:00 2001 From: Jin Dongming Date: Tue, 31 Aug 2010 09:13:33 +0900 Subject: therm_throt.c: Trivial printk message fix for a unsuitable abbreviation of 'thermal' In unexpected_thermal_interrupt(), "LVT TMR interrupt" is used in error message. I don't think TMR is a suitable abbreviation for thermal. 1.TMR has been used in IA32 Architectures Software Developer's Manual, and is the abbreviation for Trigger Mode Register. 2.There is not an standard abbreviation "TMR" defined for thermal in IA32 Architectures Software Developer's Manual. 3.Though we could understand it as Thermal Monitor Register, it is easy to be misunderstood as a *TIMER* interrupt also. I think this patch will fix it. Signed-off-by: Jin Dongming Reviewed-by: Jean Delvare Cc: Brown Len Cc: Hidetoshi Seto Cc: Fenghua Yu LKML-Reference: <4C7C492D.5020704@np.css.fujitsu.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/mcheck/therm_throt.c b/arch/x86/kernel/cpu/mcheck/therm_throt.c index c2a8b26..1d0f743 100644 --- a/arch/x86/kernel/cpu/mcheck/therm_throt.c +++ b/arch/x86/kernel/cpu/mcheck/therm_throt.c @@ -348,7 +348,7 @@ static void intel_thermal_interrupt(void) static void unexpected_thermal_interrupt(void) { - printk(KERN_ERR "CPU%d: Unexpected LVT TMR interrupt!\n", + printk(KERN_ERR "CPU%d: Unexpected LVT thermal interrupt!\n", smp_processor_id()); add_taint(TAINT_MACHINE_CHECK); } -- cgit v0.10.2 From 0f1cf415f00286a38f5ce35b459342dbfc895b50 Mon Sep 17 00:00:00 2001 From: Christian Dietrich Date: Mon, 6 Sep 2010 16:36:18 +0200 Subject: x86: Remove unnecessary #ifdef ACPI/X86_IO_ACPI The ACPI/X86_IO_ACPI ifdef isn't necessary at this point, because it is checked in an outer ifdef level already and has no effect here. Cleanup only, no functional effect. Signed-off-by: Christian Dietrich Acked-by: Borislav Petkov Cc: vamos-dev@i4.informatik.uni-erlangen.de LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index e5cc7e8..e4bd78c 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -98,7 +98,6 @@ static void __init nvidia_bugs(int num, int slot, int func) } #if defined(CONFIG_ACPI) && defined(CONFIG_X86_IO_APIC) -#if defined(CONFIG_ACPI) && defined(CONFIG_X86_IO_APIC) static u32 __init ati_ixp4x0_rev(int num, int slot, int func) { u32 d; @@ -116,7 +115,6 @@ static u32 __init ati_ixp4x0_rev(int num, int slot, int func) d &= 0xff; return d; } -#endif static void __init ati_bugs(int num, int slot, int func) { -- cgit v0.10.2 From ed2d372c0738386b8a184a6a6bea9c16df6ffb68 Mon Sep 17 00:00:00 2001 From: Christian Dietrich Date: Mon, 6 Sep 2010 16:37:05 +0200 Subject: sched: Remove unnecessary #ifdef CONFIG_SMP The CONFIG_SMP ifdef isn't necessary at this point, because it is checked in an outer ifdef level already and has no effect here. Cleanup only, no functional effect. Signed-off-by: Christian Dietrich Cc: vamos-dev@i4.informatik.uni-erlangen.de Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Tejun Heo LKML-Reference: <7a3a39ef3f765a4473cb026b1f204059568a7098.1283782701.git.qy03fugy@stud.informatik.uni-erlangen.de> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 09b574e..8eef8e5 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -426,9 +426,7 @@ struct root_domain { */ cpumask_var_t rto_mask; atomic_t rto_count; -#ifdef CONFIG_SMP struct cpupri cpupri; -#endif }; /* @@ -437,7 +435,7 @@ struct root_domain { */ static struct root_domain def_root_domain; -#endif +#endif /* CONFIG_SMP */ /* * This is the main, per-CPU runqueue data structure. -- cgit v0.10.2 From 7ef8aa72ab176e0288f363d1247079732c5d5792 Mon Sep 17 00:00:00 2001 From: Andre Przywara Date: Mon, 6 Sep 2010 15:14:17 +0200 Subject: x86, cpu: Fix renamed, not-yet-shipping AMD CPUID feature bit The AMD SSE5 feature set as-it has been replaced by some extensions to the AVX instruction set. Thus the bit formerly advertised as SSE5 is re-used for one of these extensions (XOP). Although this changes the /proc/cpuinfo output, it is not user visible, as there are no CPUs (yet) having this feature. To avoid confusion this should be added to the stable series, too. Cc: stable@kernel.org [.32.x .34.x, .35.x] Signed-off-by: Andre Przywara LKML-Reference: <1283778860-26843-2-git-send-email-andre.przywara@amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index 781a50b..c9c73d8 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -152,7 +152,7 @@ #define X86_FEATURE_3DNOWPREFETCH (6*32+ 8) /* 3DNow prefetch instructions */ #define X86_FEATURE_OSVW (6*32+ 9) /* OS Visible Workaround */ #define X86_FEATURE_IBS (6*32+10) /* Instruction Based Sampling */ -#define X86_FEATURE_SSE5 (6*32+11) /* SSE-5 */ +#define X86_FEATURE_XOP (6*32+11) /* extended AVX instructions */ #define X86_FEATURE_SKINIT (6*32+12) /* SKINIT/STGI instructions */ #define X86_FEATURE_WDT (6*32+13) /* Watchdog timer */ #define X86_FEATURE_NODEID_MSR (6*32+19) /* NodeId MSR */ diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 3a09c62..dd54779 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -1996,7 +1996,7 @@ static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, const u32 kvm_supported_word6_x86_features = F(LAHF_LM) | F(CMP_LEGACY) | F(SVM) | 0 /* ExtApicSpace */ | F(CR8_LEGACY) | F(ABM) | F(SSE4A) | F(MISALIGNSSE) | - F(3DNOWPREFETCH) | 0 /* OSVW */ | 0 /* IBS */ | F(SSE5) | + F(3DNOWPREFETCH) | 0 /* OSVW */ | 0 /* IBS */ | F(XOP) | 0 /* SKINIT */ | 0 /* WDT */; /* all calls to cpuid_count() should be made on the same cpu */ -- cgit v0.10.2 From 33ed82fb6c5f032151f7e9f1ac7b667f78f426b8 Mon Sep 17 00:00:00 2001 From: Andre Przywara Date: Mon, 6 Sep 2010 15:14:18 +0200 Subject: x86, cpu: Update AMD CPUID feature bits AMD's public CPUID specification has been updated and some bits have got names. Add them to properly describe new CPU features. Signed-off-by: Andre Przywara LKML-Reference: <1283778860-26843-3-git-send-email-andre.przywara@amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index c9c73d8..341835df 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -155,7 +155,11 @@ #define X86_FEATURE_XOP (6*32+11) /* extended AVX instructions */ #define X86_FEATURE_SKINIT (6*32+12) /* SKINIT/STGI instructions */ #define X86_FEATURE_WDT (6*32+13) /* Watchdog timer */ +#define X86_FEATURE_LWP (6*32+15) /* Light Weight Profiling */ +#define X86_FEATURE_FMA4 (6*32+16) /* 4 operands MAC instructions */ #define X86_FEATURE_NODEID_MSR (6*32+19) /* NodeId MSR */ +#define X86_FEATURE_TBM (6*32+21) /* trailing bit manipulations */ +#define X86_FEATURE_TOPOEXT (6*32+22) /* topology extensions CPUID leafs */ /* * Auxiliary flags: Linux defined - For features scattered in various -- cgit v0.10.2 From 6d886fd042634c0d3312bace63a5d0c541b721dc Mon Sep 17 00:00:00 2001 From: Andre Przywara Date: Mon, 6 Sep 2010 15:14:19 +0200 Subject: x86, cpu: Fix allowed CPUID bits for KVM guests The AMD extensions to AVX (FMA4, XOP) work on the same YMM register set as AVX, so they are safe for guests to use, as long as AVX itself is allowed. Add F16C and AES on the way for the same reasons. Signed-off-by: Andre Przywara LKML-Reference: <1283778860-26843-4-git-send-email-andre.przywara@amd.com> Acked-by: Avi Kivity Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index dd54779..6c2ecf0 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -1991,13 +1991,14 @@ static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, 0 /* Reserved */ | F(CX16) | 0 /* xTPR Update, PDCM */ | 0 /* Reserved, DCA */ | F(XMM4_1) | F(XMM4_2) | F(X2APIC) | F(MOVBE) | F(POPCNT) | - 0 /* Reserved, AES */ | F(XSAVE) | 0 /* OSXSAVE */ | F(AVX); + 0 /* Reserved*/ | F(AES) | F(XSAVE) | 0 /* OSXSAVE */ | F(AVX) | + F(F16C); /* cpuid 0x80000001.ecx */ const u32 kvm_supported_word6_x86_features = F(LAHF_LM) | F(CMP_LEGACY) | F(SVM) | 0 /* ExtApicSpace */ | F(CR8_LEGACY) | F(ABM) | F(SSE4A) | F(MISALIGNSSE) | F(3DNOWPREFETCH) | 0 /* OSVW */ | 0 /* IBS */ | F(XOP) | - 0 /* SKINIT */ | 0 /* WDT */; + 0 /* SKINIT, WDT, LWP */ | F(FMA4) | F(TBM); /* all calls to cpuid_count() should be made on the same cpu */ get_cpu(); -- cgit v0.10.2 From aeb9c7d618264dcf6eea39142fefee096c3b09e2 Mon Sep 17 00:00:00 2001 From: Andre Przywara Date: Mon, 6 Sep 2010 15:14:20 +0200 Subject: x86, kvm: add new AMD SVM feature bits The recently updated CPUID specification names new SVM feature bits. Add them to the list of reported features. Signed-off-by: Andre Przywara LKML-Reference: <1283778860-26843-5-git-send-email-andre.przywara@amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index 341835df..bffeab7 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -183,6 +183,13 @@ #define X86_FEATURE_LBRV (8*32+ 6) /* AMD LBR Virtualization support */ #define X86_FEATURE_SVML (8*32+ 7) /* "svm_lock" AMD SVM locking MSR */ #define X86_FEATURE_NRIPS (8*32+ 8) /* "nrip_save" AMD SVM next_rip save */ +#define X86_FEATURE_TSCRATEMSR (8*32+ 9) /* "tsc_scale" AMD TSC scaling support */ +#define X86_FEATURE_VMCBCLEAN (8*32+10) /* "vmcb_clean" AMD VMCB clean bits support */ +#define X86_FEATURE_FLUSHBYASID (8*32+11) /* AMD flush-by-ASID support */ +#define X86_FEATURE_DECODEASSISTS (8*32+12) /* AMD Decode Assists support */ +#define X86_FEATURE_PAUSEFILTER (8*32+13) /* AMD filtered pause intercept */ +#define X86_FEATURE_PFTHRESHOLD (8*32+14) /* AMD pause filter threshold */ + /* Intel-defined CPU features, CPUID level 0x00000007:0 (ebx), word 9 */ #define X86_FEATURE_FSGSBASE (9*32+ 0) /* {RD/WR}{FS/GS}BASE instructions*/ diff --git a/arch/x86/kernel/cpu/scattered.c b/arch/x86/kernel/cpu/scattered.c index 34b4dad..2c77931 100644 --- a/arch/x86/kernel/cpu/scattered.c +++ b/arch/x86/kernel/cpu/scattered.c @@ -43,6 +43,12 @@ void __cpuinit init_scattered_cpuid_features(struct cpuinfo_x86 *c) { X86_FEATURE_LBRV, CR_EDX, 1, 0x8000000a, 0 }, { X86_FEATURE_SVML, CR_EDX, 2, 0x8000000a, 0 }, { X86_FEATURE_NRIPS, CR_EDX, 3, 0x8000000a, 0 }, + { X86_FEATURE_TSCRATEMSR, CR_EDX, 4, 0x8000000a, 0 }, + { X86_FEATURE_VMCBCLEAN, CR_EDX, 5, 0x8000000a, 0 }, + { X86_FEATURE_FLUSHBYASID, CR_EDX, 6, 0x8000000a, 0 }, + { X86_FEATURE_DECODEASSISTS, CR_EDX, 7, 0x8000000a, 0 }, + { X86_FEATURE_PAUSEFILTER, CR_EDX,10, 0x8000000a, 0 }, + { X86_FEATURE_PFTHRESHOLD, CR_EDX,12, 0x8000000a, 0 }, { 0, 0, 0, 0, 0 } }; -- cgit v0.10.2 From 37a2f9f30a360fb03522d15c85c78265ccd80287 Mon Sep 17 00:00:00 2001 From: Cliff Wickman Date: Wed, 8 Sep 2010 10:14:27 -0500 Subject: x86, kdump: Change copy_oldmem_page() to use cached addressing The copy of /proc/vmcore to a user buffer proceeds much faster if the kernel addresses memory as cached. With this patch we have seen an increase in transfer rate from less than 15MB/s to 80-460MB/s, depending on size of the transfer. This makes a big difference in time needed to save a system dump. Signed-off-by: Cliff Wickman Acked-by: "Eric W. Biederman" Cc: kexec@lists.infradead.org Cc: # as far back as it would apply LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c index 045b36c..bf43188 100644 --- a/arch/x86/kernel/crash_dump_64.c +++ b/arch/x86/kernel/crash_dump_64.c @@ -34,7 +34,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, if (!csize) return 0; - vaddr = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE); + vaddr = ioremap_cache(pfn << PAGE_SHIFT, PAGE_SIZE); if (!vaddr) return -ENOMEM; -- cgit v0.10.2 From f269893c575167447cc9f6d1867e639fb5b6f0c5 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 31 Aug 2010 10:28:15 +0200 Subject: sched: Merge cpu_to_core_group functions Merge and simplify the two cpu_to_core_group variants so that the resulting function follows the same pattern like cpu_to_phys_group. Signed-off-by: Heiko Carstens Signed-off-by: Peter Zijlstra LKML-Reference: <20100831082843.953617555@de.ibm.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 8eef8e5..1a0c084 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -6552,31 +6552,23 @@ cpu_to_cpu_group(int cpu, const struct cpumask *cpu_map, #ifdef CONFIG_SCHED_MC static DEFINE_PER_CPU(struct static_sched_domain, core_domains); static DEFINE_PER_CPU(struct static_sched_group, sched_group_core); -#endif /* CONFIG_SCHED_MC */ -#if defined(CONFIG_SCHED_MC) && defined(CONFIG_SCHED_SMT) static int cpu_to_core_group(int cpu, const struct cpumask *cpu_map, struct sched_group **sg, struct cpumask *mask) { int group; - +#ifdef CONFIG_SCHED_SMT cpumask_and(mask, topology_thread_cpumask(cpu), cpu_map); group = cpumask_first(mask); +#else + group = cpu; +#endif if (sg) *sg = &per_cpu(sched_group_core, group).sg; return group; } -#elif defined(CONFIG_SCHED_MC) -static int -cpu_to_core_group(int cpu, const struct cpumask *cpu_map, - struct sched_group **sg, struct cpumask *unused) -{ - if (sg) - *sg = &per_cpu(sched_group_core, cpu).sg; - return cpu; -} -#endif +#endif /* CONFIG_SCHED_MC */ static DEFINE_PER_CPU(struct static_sched_domain, phys_domains); static DEFINE_PER_CPU(struct static_sched_group, sched_group_phys); -- cgit v0.10.2 From 01a08546af311c065f34727787dd0cc8dc0c216f Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 31 Aug 2010 10:28:16 +0200 Subject: sched: Add book scheduling domain On top of the SMT and MC scheduling domains this adds the BOOK scheduling domain. This is useful for NUMA like machines which do not have an interface which tells which piece of memory is attached to which node or where the hardware performs striping. Signed-off-by: Heiko Carstens Signed-off-by: Peter Zijlstra LKML-Reference: <20100831082844.253053798@de.ibm.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/sched.h b/include/linux/sched.h index 1e2a6db..b51c53c 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -875,6 +875,7 @@ enum sched_domain_level { SD_LV_NONE = 0, SD_LV_SIBLING, SD_LV_MC, + SD_LV_BOOK, SD_LV_CPU, SD_LV_NODE, SD_LV_ALLNODES, diff --git a/include/linux/topology.h b/include/linux/topology.h index 64e084f..b91a40e 100644 --- a/include/linux/topology.h +++ b/include/linux/topology.h @@ -201,6 +201,12 @@ int arch_update_cpu_topology(void); .balance_interval = 64, \ } +#ifdef CONFIG_SCHED_BOOK +#ifndef SD_BOOK_INIT +#error Please define an appropriate SD_BOOK_INIT in include/asm/topology.h!!! +#endif +#endif /* CONFIG_SCHED_BOOK */ + #ifdef CONFIG_NUMA #ifndef SD_NODE_INIT #error Please define an appropriate SD_NODE_INIT in include/asm/topology.h!!! diff --git a/kernel/sched.c b/kernel/sched.c index 1a0c084..26f83e2 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -6506,6 +6506,7 @@ struct s_data { cpumask_var_t nodemask; cpumask_var_t this_sibling_map; cpumask_var_t this_core_map; + cpumask_var_t this_book_map; cpumask_var_t send_covered; cpumask_var_t tmpmask; struct sched_group **sched_group_nodes; @@ -6517,6 +6518,7 @@ enum s_alloc { sa_rootdomain, sa_tmpmask, sa_send_covered, + sa_this_book_map, sa_this_core_map, sa_this_sibling_map, sa_nodemask, @@ -6570,6 +6572,31 @@ cpu_to_core_group(int cpu, const struct cpumask *cpu_map, } #endif /* CONFIG_SCHED_MC */ +/* + * book sched-domains: + */ +#ifdef CONFIG_SCHED_BOOK +static DEFINE_PER_CPU(struct static_sched_domain, book_domains); +static DEFINE_PER_CPU(struct static_sched_group, sched_group_book); + +static int +cpu_to_book_group(int cpu, const struct cpumask *cpu_map, + struct sched_group **sg, struct cpumask *mask) +{ + int group = cpu; +#ifdef CONFIG_SCHED_MC + cpumask_and(mask, cpu_coregroup_mask(cpu), cpu_map); + group = cpumask_first(mask); +#elif defined(CONFIG_SCHED_SMT) + cpumask_and(mask, topology_thread_cpumask(cpu), cpu_map); + group = cpumask_first(mask); +#endif + if (sg) + *sg = &per_cpu(sched_group_book, group).sg; + return group; +} +#endif /* CONFIG_SCHED_BOOK */ + static DEFINE_PER_CPU(struct static_sched_domain, phys_domains); static DEFINE_PER_CPU(struct static_sched_group, sched_group_phys); @@ -6578,7 +6605,10 @@ cpu_to_phys_group(int cpu, const struct cpumask *cpu_map, struct sched_group **sg, struct cpumask *mask) { int group; -#ifdef CONFIG_SCHED_MC +#ifdef CONFIG_SCHED_BOOK + cpumask_and(mask, cpu_book_mask(cpu), cpu_map); + group = cpumask_first(mask); +#elif defined(CONFIG_SCHED_MC) cpumask_and(mask, cpu_coregroup_mask(cpu), cpu_map); group = cpumask_first(mask); #elif defined(CONFIG_SCHED_SMT) @@ -6839,6 +6869,9 @@ SD_INIT_FUNC(CPU) #ifdef CONFIG_SCHED_MC SD_INIT_FUNC(MC) #endif +#ifdef CONFIG_SCHED_BOOK + SD_INIT_FUNC(BOOK) +#endif static int default_relax_domain_level = -1; @@ -6888,6 +6921,8 @@ static void __free_domain_allocs(struct s_data *d, enum s_alloc what, free_cpumask_var(d->tmpmask); /* fall through */ case sa_send_covered: free_cpumask_var(d->send_covered); /* fall through */ + case sa_this_book_map: + free_cpumask_var(d->this_book_map); /* fall through */ case sa_this_core_map: free_cpumask_var(d->this_core_map); /* fall through */ case sa_this_sibling_map: @@ -6934,8 +6969,10 @@ static enum s_alloc __visit_domain_allocation_hell(struct s_data *d, return sa_nodemask; if (!alloc_cpumask_var(&d->this_core_map, GFP_KERNEL)) return sa_this_sibling_map; - if (!alloc_cpumask_var(&d->send_covered, GFP_KERNEL)) + if (!alloc_cpumask_var(&d->this_book_map, GFP_KERNEL)) return sa_this_core_map; + if (!alloc_cpumask_var(&d->send_covered, GFP_KERNEL)) + return sa_this_book_map; if (!alloc_cpumask_var(&d->tmpmask, GFP_KERNEL)) return sa_send_covered; d->rd = alloc_rootdomain(); @@ -6993,6 +7030,23 @@ static struct sched_domain *__build_cpu_sched_domain(struct s_data *d, return sd; } +static struct sched_domain *__build_book_sched_domain(struct s_data *d, + const struct cpumask *cpu_map, struct sched_domain_attr *attr, + struct sched_domain *parent, int i) +{ + struct sched_domain *sd = parent; +#ifdef CONFIG_SCHED_BOOK + sd = &per_cpu(book_domains, i).sd; + SD_INIT(sd, BOOK); + set_domain_attribute(sd, attr); + cpumask_and(sched_domain_span(sd), cpu_map, cpu_book_mask(i)); + sd->parent = parent; + parent->child = sd; + cpu_to_book_group(i, cpu_map, &sd->groups, d->tmpmask); +#endif + return sd; +} + static struct sched_domain *__build_mc_sched_domain(struct s_data *d, const struct cpumask *cpu_map, struct sched_domain_attr *attr, struct sched_domain *parent, int i) @@ -7050,6 +7104,15 @@ static void build_sched_groups(struct s_data *d, enum sched_domain_level l, d->send_covered, d->tmpmask); break; #endif +#ifdef CONFIG_SCHED_BOOK + case SD_LV_BOOK: /* set up book groups */ + cpumask_and(d->this_book_map, cpu_map, cpu_book_mask(cpu)); + if (cpu == cpumask_first(d->this_book_map)) + init_sched_build_groups(d->this_book_map, cpu_map, + &cpu_to_book_group, + d->send_covered, d->tmpmask); + break; +#endif case SD_LV_CPU: /* set up physical groups */ cpumask_and(d->nodemask, cpumask_of_node(cpu), cpu_map); if (!cpumask_empty(d->nodemask)) @@ -7097,12 +7160,14 @@ static int __build_sched_domains(const struct cpumask *cpu_map, sd = __build_numa_sched_domains(&d, cpu_map, attr, i); sd = __build_cpu_sched_domain(&d, cpu_map, attr, sd, i); + sd = __build_book_sched_domain(&d, cpu_map, attr, sd, i); sd = __build_mc_sched_domain(&d, cpu_map, attr, sd, i); sd = __build_smt_sched_domain(&d, cpu_map, attr, sd, i); } for_each_cpu(i, cpu_map) { build_sched_groups(&d, SD_LV_SIBLING, cpu_map, i); + build_sched_groups(&d, SD_LV_BOOK, cpu_map, i); build_sched_groups(&d, SD_LV_MC, cpu_map, i); } @@ -7133,6 +7198,12 @@ static int __build_sched_domains(const struct cpumask *cpu_map, init_sched_groups_power(i, sd); } #endif +#ifdef CONFIG_SCHED_BOOK + for_each_cpu(i, cpu_map) { + sd = &per_cpu(book_domains, i).sd; + init_sched_groups_power(i, sd); + } +#endif for_each_cpu(i, cpu_map) { sd = &per_cpu(phys_domains, i).sd; @@ -7158,6 +7229,8 @@ static int __build_sched_domains(const struct cpumask *cpu_map, sd = &per_cpu(cpu_domains, i).sd; #elif defined(CONFIG_SCHED_MC) sd = &per_cpu(core_domains, i).sd; +#elif defined(CONFIG_SCHED_BOOK) + sd = &per_cpu(book_domains, i).sd; #else sd = &per_cpu(phys_domains, i).sd; #endif -- cgit v0.10.2 From b40d8ed4e42c79f8ed1cf345eed0888f4a2f0678 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 31 Aug 2010 10:28:17 +0200 Subject: topology/sysfs: Provide book id and siblings attributes Create attributes: /sys/devices/system/cpu/cpuX/topology/book_id /sys/devices/system/cpu/cpuX/topology/book_siblings which show the book id and the book siblings of a cpu. Unlike the attributes for SMT and MC these attributes are only present if CONFIG_SCHED_BOOK is set. There is no reason to pollute sysfs for every architecture with unused attributes. Signed-off-by: Heiko Carstens Signed-off-by: Peter Zijlstra LKML-Reference: <20100831082844.435648457@de.ibm.com> Signed-off-by: Ingo Molnar diff --git a/Documentation/cputopology.txt b/Documentation/cputopology.txt index f1c5c4b..902d315 100644 --- a/Documentation/cputopology.txt +++ b/Documentation/cputopology.txt @@ -14,25 +14,39 @@ to /proc/cpuinfo. identifier (rather than the kernel's). The actual value is architecture and platform dependent. -3) /sys/devices/system/cpu/cpuX/topology/thread_siblings: +3) /sys/devices/system/cpu/cpuX/topology/book_id: + + the book ID of cpuX. Typically it is the hardware platform's + identifier (rather than the kernel's). The actual value is + architecture and platform dependent. + +4) /sys/devices/system/cpu/cpuX/topology/thread_siblings: internel kernel map of cpuX's hardware threads within the same core as cpuX -4) /sys/devices/system/cpu/cpuX/topology/core_siblings: +5) /sys/devices/system/cpu/cpuX/topology/core_siblings: internal kernel map of cpuX's hardware threads within the same physical_package_id. +6) /sys/devices/system/cpu/cpuX/topology/book_siblings: + + internal kernel map of cpuX's hardware threads within the same + book_id. + To implement it in an architecture-neutral way, a new source file, -drivers/base/topology.c, is to export the 4 attributes. +drivers/base/topology.c, is to export the 4 or 6 attributes. The two book +related sysfs files will only be created if CONFIG_SCHED_BOOK is selected. For an architecture to support this feature, it must define some of these macros in include/asm-XXX/topology.h: #define topology_physical_package_id(cpu) #define topology_core_id(cpu) +#define topology_book_id(cpu) #define topology_thread_cpumask(cpu) #define topology_core_cpumask(cpu) +#define topology_book_cpumask(cpu) The type of **_id is int. The type of siblings is (const) struct cpumask *. @@ -45,6 +59,9 @@ not defined by include/asm-XXX/topology.h: 3) thread_siblings: just the given CPU 4) core_siblings: just the given CPU +For architectures that don't support books (CONFIG_SCHED_BOOK) there are no +default definitions for topology_book_id() and topology_book_cpumask(). + Additionally, CPU topology information is provided under /sys/devices/system/cpu and includes these files. The internal source for the output is in brackets ("[]"). diff --git a/drivers/base/topology.c b/drivers/base/topology.c index 9fc630c..f6f37a0 100644 --- a/drivers/base/topology.c +++ b/drivers/base/topology.c @@ -45,7 +45,8 @@ static ssize_t show_##name(struct sys_device *dev, \ return sprintf(buf, "%d\n", topology_##name(cpu)); \ } -#if defined(topology_thread_cpumask) || defined(topology_core_cpumask) +#if defined(topology_thread_cpumask) || defined(topology_core_cpumask) || \ + defined(topology_book_cpumask) static ssize_t show_cpumap(int type, const struct cpumask *mask, char *buf) { ptrdiff_t len = PTR_ALIGN(buf + PAGE_SIZE - 1, PAGE_SIZE) - buf; @@ -114,6 +115,14 @@ define_siblings_show_func(core_cpumask); define_one_ro_named(core_siblings, show_core_cpumask); define_one_ro_named(core_siblings_list, show_core_cpumask_list); +#ifdef CONFIG_SCHED_BOOK +define_id_show_func(book_id); +define_one_ro(book_id); +define_siblings_show_func(book_cpumask); +define_one_ro_named(book_siblings, show_book_cpumask); +define_one_ro_named(book_siblings_list, show_book_cpumask_list); +#endif + static struct attribute *default_attrs[] = { &attr_physical_package_id.attr, &attr_core_id.attr, @@ -121,6 +130,11 @@ static struct attribute *default_attrs[] = { &attr_thread_siblings_list.attr, &attr_core_siblings.attr, &attr_core_siblings_list.attr, +#ifdef CONFIG_SCHED_BOOK + &attr_book_id.attr, + &attr_book_siblings.attr, + &attr_book_siblings_list.attr, +#endif NULL }; -- cgit v0.10.2 From 4cb14bc8c54f8489251d5cba8328cb410ead0cea Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 31 Aug 2010 10:28:18 +0200 Subject: topology, s390: Add z11 cpu topology support Use the extended cpu topology information that z11 machines provide to improve the scheduler's decision making. Signed-off-by: Heiko Carstens Signed-off-by: Peter Zijlstra LKML-Reference: <20100831082844.604956770@de.ibm.com> Signed-off-by: Ingo Molnar diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index f0777a4..74a2f1b 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -198,6 +198,13 @@ config HOTPLUG_CPU can be controlled through /sys/devices/system/cpu/cpu#. Say N if you want to disable CPU hotplug. +config SCHED_BOOK + bool "Book scheduler support" + depends on SMP + help + Book scheduler support improves the CPU scheduler's decision making + when dealing with machines that have several books. + config MATHEMU bool "IEEE FPU emulation" depends on MARCH_G5 diff --git a/arch/s390/include/asm/topology.h b/arch/s390/include/asm/topology.h index 831bd03..051107a 100644 --- a/arch/s390/include/asm/topology.h +++ b/arch/s390/include/asm/topology.h @@ -3,15 +3,32 @@ #include -#define mc_capable() (1) - -const struct cpumask *cpu_coregroup_mask(unsigned int cpu); - extern unsigned char cpu_core_id[NR_CPUS]; extern cpumask_t cpu_core_map[NR_CPUS]; +static inline const struct cpumask *cpu_coregroup_mask(unsigned int cpu) +{ + return &cpu_core_map[cpu]; +} + #define topology_core_id(cpu) (cpu_core_id[cpu]) #define topology_core_cpumask(cpu) (&cpu_core_map[cpu]) +#define mc_capable() (1) + +#ifdef CONFIG_SCHED_BOOK + +extern unsigned char cpu_book_id[NR_CPUS]; +extern cpumask_t cpu_book_map[NR_CPUS]; + +static inline const struct cpumask *cpu_book_mask(unsigned int cpu) +{ + return &cpu_book_map[cpu]; +} + +#define topology_book_id(cpu) (cpu_book_id[cpu]) +#define topology_book_cpumask(cpu) (&cpu_book_map[cpu]) + +#endif /* CONFIG_SCHED_BOOK */ int topology_set_cpu_management(int fc); void topology_schedule_update(void); @@ -30,6 +47,8 @@ static inline void s390_init_cpu_topology(void) }; #endif +#define SD_BOOK_INIT SD_CPU_INIT + #include #endif /* _ASM_S390_TOPOLOGY_H */ diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c index bcef007..13559c9 100644 --- a/arch/s390/kernel/topology.c +++ b/arch/s390/kernel/topology.c @@ -57,8 +57,8 @@ struct tl_info { union tl_entry tle[0]; }; -struct core_info { - struct core_info *next; +struct mask_info { + struct mask_info *next; unsigned char id; cpumask_t mask; }; @@ -66,7 +66,6 @@ struct core_info { static int topology_enabled; static void topology_work_fn(struct work_struct *work); static struct tl_info *tl_info; -static struct core_info core_info; static int machine_has_topology; static struct timer_list topology_timer; static void set_topology_timer(void); @@ -74,38 +73,37 @@ static DECLARE_WORK(topology_work, topology_work_fn); /* topology_lock protects the core linked list */ static DEFINE_SPINLOCK(topology_lock); +static struct mask_info core_info; cpumask_t cpu_core_map[NR_CPUS]; unsigned char cpu_core_id[NR_CPUS]; -static cpumask_t cpu_coregroup_map(unsigned int cpu) +#ifdef CONFIG_SCHED_BOOK +static struct mask_info book_info; +cpumask_t cpu_book_map[NR_CPUS]; +unsigned char cpu_book_id[NR_CPUS]; +#endif + +static cpumask_t cpu_group_map(struct mask_info *info, unsigned int cpu) { - struct core_info *core = &core_info; - unsigned long flags; cpumask_t mask; cpus_clear(mask); if (!topology_enabled || !machine_has_topology) return cpu_possible_map; - spin_lock_irqsave(&topology_lock, flags); - while (core) { - if (cpu_isset(cpu, core->mask)) { - mask = core->mask; + while (info) { + if (cpu_isset(cpu, info->mask)) { + mask = info->mask; break; } - core = core->next; + info = info->next; } - spin_unlock_irqrestore(&topology_lock, flags); if (cpus_empty(mask)) mask = cpumask_of_cpu(cpu); return mask; } -const struct cpumask *cpu_coregroup_mask(unsigned int cpu) -{ - return &cpu_core_map[cpu]; -} - -static void add_cpus_to_core(struct tl_cpu *tl_cpu, struct core_info *core) +static void add_cpus_to_mask(struct tl_cpu *tl_cpu, struct mask_info *book, + struct mask_info *core) { unsigned int cpu; @@ -117,23 +115,35 @@ static void add_cpus_to_core(struct tl_cpu *tl_cpu, struct core_info *core) rcpu = CPU_BITS - 1 - cpu + tl_cpu->origin; for_each_present_cpu(lcpu) { - if (cpu_logical_map(lcpu) == rcpu) { - cpu_set(lcpu, core->mask); - cpu_core_id[lcpu] = core->id; - smp_cpu_polarization[lcpu] = tl_cpu->pp; - } + if (cpu_logical_map(lcpu) != rcpu) + continue; +#ifdef CONFIG_SCHED_BOOK + cpu_set(lcpu, book->mask); + cpu_book_id[lcpu] = book->id; +#endif + cpu_set(lcpu, core->mask); + cpu_core_id[lcpu] = core->id; + smp_cpu_polarization[lcpu] = tl_cpu->pp; } } } -static void clear_cores(void) +static void clear_masks(void) { - struct core_info *core = &core_info; + struct mask_info *info; - while (core) { - cpus_clear(core->mask); - core = core->next; + info = &core_info; + while (info) { + cpus_clear(info->mask); + info = info->next; + } +#ifdef CONFIG_SCHED_BOOK + info = &book_info; + while (info) { + cpus_clear(info->mask); + info = info->next; } +#endif } static union tl_entry *next_tle(union tl_entry *tle) @@ -146,29 +156,36 @@ static union tl_entry *next_tle(union tl_entry *tle) static void tl_to_cores(struct tl_info *info) { +#ifdef CONFIG_SCHED_BOOK + struct mask_info *book = &book_info; +#else + struct mask_info *book = NULL; +#endif + struct mask_info *core = &core_info; union tl_entry *tle, *end; - struct core_info *core = &core_info; + spin_lock_irq(&topology_lock); - clear_cores(); + clear_masks(); tle = info->tle; end = (union tl_entry *)((unsigned long)info + info->length); while (tle < end) { switch (tle->nl) { - case 5: - case 4: - case 3: +#ifdef CONFIG_SCHED_BOOK case 2: + book = book->next; + book->id = tle->container.id; break; +#endif case 1: core = core->next; core->id = tle->container.id; break; case 0: - add_cpus_to_core(&tle->cpu, core); + add_cpus_to_mask(&tle->cpu, book, core); break; default: - clear_cores(); + clear_masks(); machine_has_topology = 0; goto out; } @@ -221,10 +238,29 @@ int topology_set_cpu_management(int fc) static void update_cpu_core_map(void) { + unsigned long flags; int cpu; - for_each_possible_cpu(cpu) - cpu_core_map[cpu] = cpu_coregroup_map(cpu); + spin_lock_irqsave(&topology_lock, flags); + for_each_possible_cpu(cpu) { + cpu_core_map[cpu] = cpu_group_map(&core_info, cpu); +#ifdef CONFIG_SCHED_BOOK + cpu_book_map[cpu] = cpu_group_map(&book_info, cpu); +#endif + } + spin_unlock_irqrestore(&topology_lock, flags); +} + +static void store_topology(struct tl_info *info) +{ +#ifdef CONFIG_SCHED_BOOK + int rc; + + rc = stsi(info, 15, 1, 3); + if (rc != -ENOSYS) + return; +#endif + stsi(info, 15, 1, 2); } int arch_update_cpu_topology(void) @@ -238,7 +274,7 @@ int arch_update_cpu_topology(void) topology_update_polarization_simple(); return 0; } - stsi(info, 15, 1, 2); + store_topology(info); tl_to_cores(info); update_cpu_core_map(); for_each_online_cpu(cpu) { @@ -299,12 +335,24 @@ out: } __initcall(init_topology_update); +static void alloc_masks(struct tl_info *info, struct mask_info *mask, int offset) +{ + int i, nr_masks; + + nr_masks = info->mag[NR_MAG - offset]; + for (i = 0; i < info->mnest - offset; i++) + nr_masks *= info->mag[NR_MAG - offset - 1 - i]; + nr_masks = max(nr_masks, 1); + for (i = 0; i < nr_masks; i++) { + mask->next = alloc_bootmem(sizeof(struct mask_info)); + mask = mask->next; + } +} + void __init s390_init_cpu_topology(void) { unsigned long long facility_bits; struct tl_info *info; - struct core_info *core; - int nr_cores; int i; if (stfle(&facility_bits, 1) <= 0) @@ -315,25 +363,13 @@ void __init s390_init_cpu_topology(void) tl_info = alloc_bootmem_pages(PAGE_SIZE); info = tl_info; - stsi(info, 15, 1, 2); - - nr_cores = info->mag[NR_MAG - 2]; - for (i = 0; i < info->mnest - 2; i++) - nr_cores *= info->mag[NR_MAG - 3 - i]; - + store_topology(info); pr_info("The CPU configuration topology of the machine is:"); for (i = 0; i < NR_MAG; i++) printk(" %d", info->mag[i]); printk(" / %d\n", info->mnest); - - core = &core_info; - for (i = 0; i < nr_cores; i++) { - core->next = alloc_bootmem(sizeof(struct core_info)); - core = core->next; - if (!core) - goto error; - } - return; -error: - machine_has_topology = 0; + alloc_masks(info, &core_info, 2); +#ifdef CONFIG_SCHED_BOOK + alloc_masks(info, &book_info, 3); +#endif } -- cgit v0.10.2 From 8af3c153baf95374eff20a37f00c59a295b52756 Mon Sep 17 00:00:00 2001 From: Miroslav Lichvar Date: Tue, 7 Sep 2010 16:43:46 +0200 Subject: ntp: Clamp PLL update interval Clamp update interval to reduce PLL gain with low sampling rate (e.g. intermittent network connection) to avoid instability. The clamp roughly corresponds to the loop time constant, it's 8 * poll interval for SHIFT_PLL 2 and 32 * poll interval for SHIFT_PLL 4. This gives good results without affecting the gain in normal conditions where ntpd skips only up to seven consecutive samples. Signed-off-by: Miroslav Lichvar Acked-by: john stultz LKML-Reference: <1283870626-9472-1-git-send-email-mlichvar@redhat.com> Signed-off-by: Thomas Gleixner diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index c631168..d232189 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -149,10 +149,18 @@ static void ntp_update_offset(long offset) time_reftime = get_seconds(); offset64 = offset; - freq_adj = (offset64 * secs) << - (NTP_SCALE_SHIFT - 2 * (SHIFT_PLL + 2 + time_constant)); + freq_adj = ntp_update_offset_fll(offset64, secs); - freq_adj += ntp_update_offset_fll(offset64, secs); + /* + * Clamp update interval to reduce PLL gain with low + * sampling rate (e.g. intermittent network connection) + * to avoid instability. + */ + if (unlikely(secs > 1 << (SHIFT_PLL + 1 + time_constant))) + secs = 1 << (SHIFT_PLL + 1 + time_constant); + + freq_adj += (offset64 * secs) << + (NTP_SCALE_SHIFT - 2 * (SHIFT_PLL + 2 + time_constant)); freq_adj = min(freq_adj + time_freq, MAXFREQ_SCALED); -- cgit v0.10.2 From 2df7a6e9e8e67c19e5fe2eac3f2d2223b7bb4a7b Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:08 -0400 Subject: x86: Use correct type for %cr4 %cr4 is 64-bit in 64-bit mode (although the upper 32-bits are currently reserved). Use unsigned long for the temporary variable to get the right size. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-2-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index 325b7bd..396b80f 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -602,7 +602,7 @@ extern unsigned long mmu_cr4_features; static inline void set_in_cr4(unsigned long mask) { - unsigned cr4; + unsigned long cr4; mmu_cr4_features |= mask; cr4 = read_cr4(); @@ -612,7 +612,7 @@ static inline void set_in_cr4(unsigned long mask) static inline void clear_in_cr4(unsigned long mask) { - unsigned cr4; + unsigned long cr4; mmu_cr4_features &= ~mask; cr4 = read_cr4(); -- cgit v0.10.2 From 6ac8bac2684235f4caf22a410549c582aa7327d6 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:09 -0400 Subject: x86, fpu: Merge fpu_init() Make fpu_init() handle 32-bit setup. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-3-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 490dac6..f9e23e8 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -1264,13 +1264,6 @@ void __cpuinit cpu_init(void) clear_all_debug_regs(); dbg_restore_debug_regs(); - /* - * Force FPU initialization: - */ - current_thread_info()->status = 0; - clear_used_math(); - mxcsr_feature_mask_init(); - fpu_init(); xsave_init(); } diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index 2605c50..8216651 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -85,7 +85,6 @@ static void __cpuinit init_thread_xstate(void) #endif } -#ifdef CONFIG_X86_64 /* * Called at bootup to set up the initial FPU state that is later cloned * into all processes. @@ -93,12 +92,21 @@ static void __cpuinit init_thread_xstate(void) void __cpuinit fpu_init(void) { - unsigned long oldcr0 = read_cr0(); - - set_in_cr4(X86_CR4_OSFXSR); - set_in_cr4(X86_CR4_OSXMMEXCPT); + unsigned long cr0; + unsigned long cr4_mask = 0; - write_cr0(oldcr0 & ~(X86_CR0_TS|X86_CR0_EM)); /* clear TS and EM */ + if (cpu_has_fxsr) + cr4_mask |= X86_CR4_OSFXSR; + if (cpu_has_xmm) + cr4_mask |= X86_CR4_OSXMMEXCPT; + if (cr4_mask) + set_in_cr4(cr4_mask); + + cr0 = read_cr0(); + cr0 &= ~(X86_CR0_TS|X86_CR0_EM); /* clear TS and EM */ + if (!HAVE_HWFP) + cr0 |= X86_CR0_EM; + write_cr0(cr0); if (!smp_processor_id()) init_thread_xstate(); @@ -109,16 +117,6 @@ void __cpuinit fpu_init(void) clear_used_math(); } -#else /* CONFIG_X86_64 */ - -void __cpuinit fpu_init(void) -{ - if (!smp_processor_id()) - init_thread_xstate(); -} - -#endif /* CONFIG_X86_32 */ - void fpu_finit(struct fpu *fpu) { #ifdef CONFIG_X86_32 diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 60788de..d0029eb 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -881,18 +881,6 @@ void __init trap_init(void) #endif #ifdef CONFIG_X86_32 - if (cpu_has_fxsr) { - printk(KERN_INFO "Enabling fast FPU save and restore... "); - set_in_cr4(X86_CR4_OSFXSR); - printk("done.\n"); - } - if (cpu_has_xmm) { - printk(KERN_INFO - "Enabling unmasked SIMD FPU exception support... "); - set_in_cr4(X86_CR4_OSXMMEXCPT); - printk("done.\n"); - } - set_system_trap_gate(SYSCALL_VECTOR, &system_call); set_bit(SYSCALL_VECTOR, used_vectors); #endif -- cgit v0.10.2 From 51115d4d45700fc7c08306f7ba6e68551f526ae5 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:10 -0400 Subject: x86, fpu: Merge tolerant_fwait() Commit e2e75c91 merged the math exception handler, allowing both 32-bit and 64-bit to handle math exceptions from kernel mode. Switch to using the 64-bit version of tolerant_fwait() without fnclex, which simply ignores the exception if one is still pending from userspace. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-4-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index a73a8d5..5d8f9a7 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -77,15 +77,6 @@ static inline void sanitize_i387_state(struct task_struct *tsk) } #ifdef CONFIG_X86_64 - -/* Ignore delayed exceptions from user space */ -static inline void tolerant_fwait(void) -{ - asm volatile("1: fwait\n" - "2:\n" - _ASM_EXTABLE(1b, 2b)); -} - static inline int fxrstor_checking(struct i387_fxsave_struct *fx) { int err; @@ -220,11 +211,6 @@ extern void finit_soft_fpu(struct i387_soft_struct *soft); static inline void finit_soft_fpu(struct i387_soft_struct *soft) {} #endif -static inline void tolerant_fwait(void) -{ - asm volatile("fnclex ; fwait"); -} - /* perform fxrstor iff the processor has extended states, otherwise frstor */ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) { @@ -344,7 +330,10 @@ static inline void __unlazy_fpu(struct task_struct *tsk) static inline void __clear_fpu(struct task_struct *tsk) { if (task_thread_info(tsk)->status & TS_USEDFPU) { - tolerant_fwait(); + /* Ignore delayed exceptions from user space */ + asm volatile("1: fwait\n" + "2:\n" + _ASM_EXTABLE(1b, 2b)); task_thread_info(tsk)->status &= ~TS_USEDFPU; stts(); } -- cgit v0.10.2 From bfd946cb891800d408decaae268a3480775178a3 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:11 -0400 Subject: x86, fpu: Merge __save_init_fpu() __save_init_fpu() is identical for 32-bit and 64-bit. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-5-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index 5d8f9a7..88065e3 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -197,12 +197,6 @@ static inline void fpu_save_init(struct fpu *fpu) fpu_clear(fpu); } -static inline void __save_init_fpu(struct task_struct *tsk) -{ - fpu_save_init(&tsk->thread.fpu); - task_thread_info(tsk)->status &= ~TS_USEDFPU; -} - #else /* CONFIG_X86_32 */ #ifdef CONFIG_MATH_EMULATION @@ -285,15 +279,14 @@ end: ; } +#endif /* CONFIG_X86_64 */ + static inline void __save_init_fpu(struct task_struct *tsk) { fpu_save_init(&tsk->thread.fpu); task_thread_info(tsk)->status &= ~TS_USEDFPU; } - -#endif /* CONFIG_X86_64 */ - static inline int fpu_fxrstor_checking(struct fpu *fpu) { return fxrstor_checking(&fpu->state->fxsave); -- cgit v0.10.2 From a4d4fbc7735bba6654b20f859135f9d3f8fe7f76 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:12 -0400 Subject: x86-64, fpu: Disable preemption when using TS_USEDFPU Consolidates code and fixes the below race for 64-bit. commit 9fa2f37bfeb798728241cc4a19578ce6e4258f25 Author: torvalds Date: Tue Sep 2 07:37:25 2003 +0000 Be a lot more careful about TS_USEDFPU and preemption We had some races where we testecd (or set) TS_USEDFPU together with sequences that depended on the setting (like clearing or setting the TS flag in %cr0) and we could be preempted in between, which screws up the FPU state, since preemption will itself change USEDFPU and the TS flag. This makes it a lot more explicit: the "internal" low-level FPU functions ("__xxxx_fpu()") all require preemption to be disabled, and the exported "real" functions will make sure that is the case. One case - in __switch_to() - was switched to the non-preempt-safe internal version, since the scheduler itself has already disabled preemption. BKrev: 3f5448b5WRiQuyzAlbajs3qoQjSobw Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-6-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index 88065e3..8b40a83 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -387,19 +387,6 @@ static inline void irq_ts_restore(int TS_state) stts(); } -#ifdef CONFIG_X86_64 - -static inline void save_init_fpu(struct task_struct *tsk) -{ - __save_init_fpu(tsk); - stts(); -} - -#define unlazy_fpu __unlazy_fpu -#define clear_fpu __clear_fpu - -#else /* CONFIG_X86_32 */ - /* * These disable preemption on their own and are safe */ @@ -425,8 +412,6 @@ static inline void clear_fpu(struct task_struct *tsk) preempt_enable(); } -#endif /* CONFIG_X86_64 */ - /* * i387 state interaction */ diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 3d9ea53..b3d7a3a 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -424,7 +424,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) load_TLS(next, cpu); /* Must be after DS reload */ - unlazy_fpu(prev_p); + __unlazy_fpu(prev_p); /* Make sure cpu is ready for new context */ if (preload_fpu) -- cgit v0.10.2 From 10c11f304986a1f84201c2261a428701f9d2dffc Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:13 -0400 Subject: x86-64, fpu: Fix %cs value in convert_from_fxsr() While %ds still contains the userspace selector, %cs is KERNEL_CS at this point. Always get %cs from pt_regs even for the current task. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-7-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index 8216651..f3775f5 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -389,19 +389,17 @@ convert_from_fxsr(struct user_i387_ia32_struct *env, struct task_struct *tsk) #ifdef CONFIG_X86_64 env->fip = fxsave->rip; env->foo = fxsave->rdp; + /* + * should be actually ds/cs at fpu exception time, but + * that information is not available in 64bit mode. + */ + env->fcs = task_pt_regs(tsk)->cs; if (tsk == current) { - /* - * should be actually ds/cs at fpu exception time, but - * that information is not available in 64bit mode. - */ - asm("mov %%ds, %[fos]" : [fos] "=r" (env->fos)); - asm("mov %%cs, %[fcs]" : [fcs] "=r" (env->fcs)); + savesegment(ds, env->fos); } else { - struct pt_regs *regs = task_pt_regs(tsk); - - env->fos = 0xffff0000 | tsk->thread.ds; - env->fcs = regs->cs; + env->fos = tsk->thread.ds; } + env->fos |= 0xffff0000; #else env->fip = fxsave->fip; env->fcs = (u16) fxsave->fcs | ((u32) fxsave->fop << 16); -- cgit v0.10.2 From 820241356d6aa9a895fc10def15794a5a5bfcd98 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:14 -0400 Subject: x86-64, fpu: Simplify constraints for fxsave/fxtstor Use the "R" constraint (legacy register) instead of listing all the possible registers. Clean up the comments as well. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-8-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index 8b40a83..768fcb2 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -81,6 +81,7 @@ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) { int err; + /* See comment in fxsave() below. */ asm volatile("1: rex64/fxrstor (%[fx])\n\t" "2:\n" ".section .fixup,\"ax\"\n" @@ -89,11 +90,7 @@ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) ".previous\n" _ASM_EXTABLE(1b, 3b) : [err] "=r" (err) -#if 0 /* See comment in fxsave() below. */ - : [fx] "r" (fx), "m" (*fx), "0" (0)); -#else - : [fx] "cdaSDb" (fx), "m" (*fx), "0" (0)); -#endif + : [fx] "R" (fx), "m" (*fx), "0" (0)); return err; } @@ -140,6 +137,7 @@ static inline int fxsave_user(struct i387_fxsave_struct __user *fx) if (unlikely(err)) return -EFAULT; + /* See comment in fxsave() below. */ asm volatile("1: rex64/fxsave (%[fx])\n\t" "2:\n" ".section .fixup,\"ax\"\n" @@ -148,11 +146,7 @@ static inline int fxsave_user(struct i387_fxsave_struct __user *fx) ".previous\n" _ASM_EXTABLE(1b, 3b) : [err] "=r" (err), "=m" (*fx) -#if 0 /* See comment in fxsave() below. */ - : [fx] "r" (fx), "0" (0)); -#else - : [fx] "cdaSDb" (fx), "0" (0)); -#endif + : [fx] "R" (fx), "0" (0)); if (unlikely(err) && __clear_user(fx, sizeof(struct i387_fxsave_struct))) err = -EFAULT; @@ -165,26 +159,22 @@ static inline void fpu_fxsave(struct fpu *fpu) /* Using "rex64; fxsave %0" is broken because, if the memory operand uses any extended registers for addressing, a second REX prefix will be generated (to the assembler, rex64 followed by semicolon - is a separate instruction), and hence the 64-bitness is lost. */ -#if 0 - /* Using "fxsaveq %0" would be the ideal choice, but is only supported - starting with gas 2.16. */ - __asm__ __volatile__("fxsaveq %0" - : "=m" (fpu->state->fxsave)); -#elif 0 - /* Using, as a workaround, the properly prefixed form below isn't + is a separate instruction), and hence the 64-bitness is lost. + Using "fxsaveq %0" would be the ideal choice, but is only supported + starting with gas 2.16. + asm volatile("fxsaveq %0" + : "=m" (fpu->state->fxsave)); + Using, as a workaround, the properly prefixed form below isn't accepted by any binutils version so far released, complaining that the same type of prefix is used twice if an extended register is - needed for addressing (fix submitted to mainline 2005-11-21). */ - __asm__ __volatile__("rex64/fxsave %0" - : "=m" (fpu->state->fxsave)); -#else - /* This, however, we can work around by forcing the compiler to select + needed for addressing (fix submitted to mainline 2005-11-21). + asm volatile("rex64/fxsave %0" + : "=m" (fpu->state->fxsave)); + This, however, we can work around by forcing the compiler to select an addressing mode that doesn't require extended registers. */ - __asm__ __volatile__("rex64/fxsave (%1)" - : "=m" (fpu->state->fxsave) - : "cdaSDb" (&fpu->state->fxsave)); -#endif + asm volatile("rex64/fxsave (%[fx])" + : "=m" (fpu->state->fxsave) + : [fx] "R" (&fpu->state->fxsave)); } static inline void fpu_save_init(struct fpu *fpu) -- cgit v0.10.2 From a334fe43d85f570ae907acf988a053c5eff78d6e Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:15 -0400 Subject: x86-32, fpu: Remove math_emulate stub check_fpu() in bugs.c halts boot if no FPU is found and math emulation isn't enabled. Therefore this stub will never be used. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-9-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index d0029eb..d439685 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -776,21 +776,10 @@ asmlinkage void math_state_restore(void) } EXPORT_SYMBOL_GPL(math_state_restore); -#ifndef CONFIG_MATH_EMULATION -void math_emulate(struct math_emu_info *info) -{ - printk(KERN_EMERG - "math-emulation not enabled and no coprocessor found.\n"); - printk(KERN_EMERG "killing %s.\n", current->comm); - force_sig(SIGFPE, current); - schedule(); -} -#endif /* CONFIG_MATH_EMULATION */ - dotraplinkage void __kprobes do_device_not_available(struct pt_regs *regs, long error_code) { -#ifdef CONFIG_X86_32 +#ifdef CONFIG_MATH_EMULATION if (read_cr0() & X86_CR0_EM) { struct math_emu_info info = { }; @@ -798,12 +787,12 @@ do_device_not_available(struct pt_regs *regs, long error_code) info.regs = regs; math_emulate(&info); - } else { - math_state_restore(); /* interrupts still off */ - conditional_sti(regs); + return; } -#else - math_state_restore(); +#endif + math_state_restore(); /* interrupts still off */ +#ifdef CONFIG_X86_32 + conditional_sti(regs); #endif } -- cgit v0.10.2 From 8eb91a577d7763d21628f6761045328784b1911c Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:16 -0400 Subject: x86, fpu: Remove unnecessary ifdefs from i387 code. Remove ifdefs for code that the compiler can optimize away on 64-bit. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-10-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index 768fcb2..42b507e 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -55,6 +55,12 @@ extern int save_i387_xstate_ia32(void __user *buf); extern int restore_i387_xstate_ia32(void __user *buf); #endif +#ifdef CONFIG_MATH_EMULATION +extern void finit_soft_fpu(struct i387_soft_struct *soft); +#else +static inline void finit_soft_fpu(struct i387_soft_struct *soft) {} +#endif + #define X87_FSW_ES (1 << 7) /* Exception Summary */ static __always_inline __pure bool use_xsaveopt(void) @@ -189,12 +195,6 @@ static inline void fpu_save_init(struct fpu *fpu) #else /* CONFIG_X86_32 */ -#ifdef CONFIG_MATH_EMULATION -extern void finit_soft_fpu(struct i387_soft_struct *soft); -#else -static inline void finit_soft_fpu(struct i387_soft_struct *soft) {} -#endif - /* perform fxrstor iff the processor has extended states, otherwise frstor */ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) { diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index f3775f5..e795e36 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -79,10 +79,8 @@ static void __cpuinit init_thread_xstate(void) if (cpu_has_fxsr) xstate_size = sizeof(struct i387_fxsave_struct); -#ifdef CONFIG_X86_32 else xstate_size = sizeof(struct i387_fsave_struct); -#endif } /* @@ -119,12 +117,10 @@ void __cpuinit fpu_init(void) void fpu_finit(struct fpu *fpu) { -#ifdef CONFIG_X86_32 if (!HAVE_HWFP) { finit_soft_fpu(&fpu->state->soft); return; } -#endif if (cpu_has_fxsr) { struct i387_fxsave_struct *fx = &fpu->state->fxsave; -- cgit v0.10.2 From eec73f813ab0954253e5e2168119c4555f83f07d Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:17 -0400 Subject: x86, fpu: Remove PSHUFB_XMM5_* macros The PSHUFB_XMM5_* macros are no longer used. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-11-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index 42b507e..907967e 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -465,7 +465,4 @@ extern void fpu_finit(struct fpu *fpu); #endif /* __ASSEMBLY__ */ -#define PSHUFB_XMM5_XMM0 .byte 0x66, 0x0f, 0x38, 0x00, 0xc5 -#define PSHUFB_XMM5_XMM6 .byte 0x66, 0x0f, 0x38, 0x00, 0xf5 - #endif /* _ASM_X86_I387_H */ -- cgit v0.10.2 From 58a992b9cbaf449aeebd3575c3695a9eb5d95b5e Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:18 -0400 Subject: x86-32, fpu: Rewrite fpu_save_init() Rewrite fpu_save_init() to prepare for merging with 64-bit. Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-12-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index 907967e..b45abef 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -73,6 +73,11 @@ static __always_inline __pure bool use_xsave(void) return static_cpu_has(X86_FEATURE_XSAVE); } +static __always_inline __pure bool use_fxsr(void) +{ + return static_cpu_has(X86_FEATURE_FXSR); +} + extern void __sanitize_i387_state(struct task_struct *); static inline void sanitize_i387_state(struct task_struct *tsk) @@ -211,6 +216,12 @@ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) return 0; } +static inline void fpu_fxsave(struct fpu *fpu) +{ + asm volatile("fxsave %[fx]" + : [fx] "=m" (fpu->state->fxsave)); +} + /* We need a safe address that is cheap to find and that is already in L1 during context switch. The best choices are unfortunately different for UP and SMP */ @@ -226,36 +237,24 @@ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) static inline void fpu_save_init(struct fpu *fpu) { if (use_xsave()) { - struct xsave_struct *xstate = &fpu->state->xsave; - struct i387_fxsave_struct *fx = &fpu->state->fxsave; - fpu_xsave(fpu); /* * xsave header may indicate the init state of the FP. */ - if (!(xstate->xsave_hdr.xstate_bv & XSTATE_FP)) - goto end; - - if (unlikely(fx->swd & X87_FSW_ES)) - asm volatile("fnclex"); - - /* - * we can do a simple return here or be paranoid :) - */ - goto clear_state; + if (!(fpu->state->xsave.xsave_hdr.xstate_bv & XSTATE_FP)) + return; + } else if (use_fxsr()) { + fpu_fxsave(fpu); + } else { + asm volatile("fsave %[fx]; fwait" + : [fx] "=m" (fpu->state->fsave)); + return; } - /* Use more nops than strictly needed in case the compiler - varies code */ - alternative_input( - "fnsave %[fx] ;fwait;" GENERIC_NOP8 GENERIC_NOP4, - "fxsave %[fx]\n" - "bt $7,%[fsw] ; jnc 1f ; fnclex\n1:", - X86_FEATURE_FXSR, - [fx] "m" (fpu->state->fxsave), - [fsw] "m" (fpu->state->fxsave.swd) : "memory"); -clear_state: + if (unlikely(fpu->state->fxsave.swd & X87_FSW_ES)) + asm volatile("fnclex"); + /* AMD K7/K8 CPUs don't save/restore FDP/FIP/FOP unless an exception is pending. Clear the x87 state here by setting it to fixed values. safe_address is a random variable that should be in L1 */ @@ -265,8 +264,6 @@ clear_state: "fildl %[addr]", /* set F?P to defined value */ X86_FEATURE_FXSAVE_LEAK, [addr] "m" (safe_address)); -end: - ; } #endif /* CONFIG_X86_64 */ -- cgit v0.10.2 From b2b57fe053c9cf8b8af5a0e826a465996afed0ff Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 3 Sep 2010 21:17:19 -0400 Subject: x86, fpu: Merge fpu_save_init() Make 64-bit use the 32-bit version of fpu_save_init(). Remove unused clear_fpu_state(). Signed-off-by: Brian Gerst Acked-by: Pekka Enberg Cc: Suresh Siddha LKML-Reference: <1283563039-3466-13-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index b45abef..70626ed 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -105,36 +105,6 @@ static inline int fxrstor_checking(struct i387_fxsave_struct *fx) return err; } -/* AMD CPUs don't save/restore FDP/FIP/FOP unless an exception - is pending. Clear the x87 state here by setting it to fixed - values. The kernel data segment can be sometimes 0 and sometimes - new user value. Both should be ok. - Use the PDA as safe address because it should be already in L1. */ -static inline void fpu_clear(struct fpu *fpu) -{ - struct xsave_struct *xstate = &fpu->state->xsave; - struct i387_fxsave_struct *fx = &fpu->state->fxsave; - - /* - * xsave header may indicate the init state of the FP. - */ - if (use_xsave() && - !(xstate->xsave_hdr.xstate_bv & XSTATE_FP)) - return; - - if (unlikely(fx->swd & X87_FSW_ES)) - asm volatile("fnclex"); - alternative_input(ASM_NOP8 ASM_NOP2, - " emms\n" /* clear stack tags */ - " fildl %%gs:0", /* load to clear state */ - X86_FEATURE_FXSAVE_LEAK); -} - -static inline void clear_fpu_state(struct task_struct *tsk) -{ - fpu_clear(&tsk->thread.fpu); -} - static inline int fxsave_user(struct i387_fxsave_struct __user *fx) { int err; @@ -188,16 +158,6 @@ static inline void fpu_fxsave(struct fpu *fpu) : [fx] "R" (&fpu->state->fxsave)); } -static inline void fpu_save_init(struct fpu *fpu) -{ - if (use_xsave()) - fpu_xsave(fpu); - else - fpu_fxsave(fpu); - - fpu_clear(fpu); -} - #else /* CONFIG_X86_32 */ /* perform fxrstor iff the processor has extended states, otherwise frstor */ @@ -222,6 +182,8 @@ static inline void fpu_fxsave(struct fpu *fpu) : [fx] "=m" (fpu->state->fxsave)); } +#endif /* CONFIG_X86_64 */ + /* We need a safe address that is cheap to find and that is already in L1 during context switch. The best choices are unfortunately different for UP and SMP */ @@ -259,15 +221,13 @@ static inline void fpu_save_init(struct fpu *fpu) is pending. Clear the x87 state here by setting it to fixed values. safe_address is a random variable that should be in L1 */ alternative_input( - GENERIC_NOP8 GENERIC_NOP2, + ASM_NOP8 ASM_NOP2, "emms\n\t" /* clear stack tags */ - "fildl %[addr]", /* set F?P to defined value */ + "fildl %P[addr]", /* set F?P to defined value */ X86_FEATURE_FXSAVE_LEAK, [addr] "m" (safe_address)); } -#endif /* CONFIG_X86_64 */ - static inline void __save_init_fpu(struct task_struct *tsk) { fpu_save_init(&tsk->thread.fpu); -- cgit v0.10.2 From a7f07cfbaa1dd5bf9e615948f280c92e7928e6f7 Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Fri, 10 Sep 2010 15:55:49 -0700 Subject: x86, mtrr: Refactor MTRR type overlap check code Move the MTRR type overlap check into a new function. No functional change in this patch. Just making it easier to add multiple region overlap check in the following patch. Signed-off-by: Venkatesh Pallipadi LKML-Reference: <1284159350-19841-2-git-send-email-venki@google.com> Reviewed-by: Suresh Siddha Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 7d28d7d..14f4f0c 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -65,6 +65,33 @@ static inline void k8_check_syscfg_dram_mod_en(void) } /* + * Check and return the effective type for MTRR-MTRR type overlap. + * Returns 1 if the effective type is UNCACHEABLE, else returns 0 + */ +static int check_type_overlap(u8 *prev, u8 *curr) +{ + if (*prev == MTRR_TYPE_UNCACHABLE || *curr == MTRR_TYPE_UNCACHABLE) { + *prev = MTRR_TYPE_UNCACHABLE; + *curr = MTRR_TYPE_UNCACHABLE; + return 1; + } + + if ((*prev == MTRR_TYPE_WRBACK && *curr == MTRR_TYPE_WRTHROUGH) || + (*prev == MTRR_TYPE_WRTHROUGH && *curr == MTRR_TYPE_WRBACK)) { + *prev = MTRR_TYPE_WRTHROUGH; + *curr = MTRR_TYPE_WRTHROUGH; + } + + if (*prev != *curr) { + *prev = MTRR_TYPE_UNCACHABLE; + *curr = MTRR_TYPE_UNCACHABLE; + return 1; + } + + return 0; +} + +/* * Returns the effective MTRR type for the region * Error returns: * - 0xFE - when the range is "not entirely covered" by _any_ var range MTRR @@ -138,21 +165,8 @@ u8 mtrr_type_lookup(u64 start, u64 end) continue; } - if (prev_match == MTRR_TYPE_UNCACHABLE || - curr_match == MTRR_TYPE_UNCACHABLE) { - return MTRR_TYPE_UNCACHABLE; - } - - if ((prev_match == MTRR_TYPE_WRBACK && - curr_match == MTRR_TYPE_WRTHROUGH) || - (prev_match == MTRR_TYPE_WRTHROUGH && - curr_match == MTRR_TYPE_WRBACK)) { - prev_match = MTRR_TYPE_WRTHROUGH; - curr_match = MTRR_TYPE_WRTHROUGH; - } - - if (prev_match != curr_match) - return MTRR_TYPE_UNCACHABLE; + if (check_type_overlap(&prev_match, &curr_match)) + return curr_match; } if (mtrr_tom2) { -- cgit v0.10.2 From 351e5a703ad994405bd900da330823d3b4a372e0 Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Fri, 10 Sep 2010 15:55:50 -0700 Subject: x86, mtrr: Support mtrr lookup for range spanning across MTRR range mtrr_type_lookup [start:end] looked up the resultant MTRR type for that range, based on fixed and all variable MTRR ranges. It did check for multiple MTRR var ranges overlapping [start:end] and returned the net type. However, if the [start:end] range spanned across any var MTRR range, mtrr_type_lookup would return an error return of 0xFE. This was based on typical usage of mtrr_type_lookup in PAT mapping, where region being mapped would not normally span across MTRR ranges and also trying to keep the code simple. Mark recently reported the problem with this limitation. When there are two continguous MTRR's of type "writeback" and if there is a memory mapping over a region starting in one MTRR range and ending in another MTRR range, such mapping will fallback to "uncached" due to the above limitation. Change below adds support for such lookups spanning multiple MTRR ranges. We now have a wrapper mtrr_type_lookup that dynamically splits such a region into smaller chunks that fit within one MTRR range and does a __mtrr_type_lookup on it and combine the results later. Reported-by: Mark Langsdorf Signed-off-by: Venkatesh Pallipadi LKML-Reference: <1284159350-19841-3-git-send-email-venki@google.com> Reviewed-by: Suresh Siddha Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 14f4f0c..9f27228 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -64,6 +64,18 @@ static inline void k8_check_syscfg_dram_mod_en(void) } } +/* Get the size of contiguous MTRR range */ +static u64 get_mtrr_size(u64 mask) +{ + u64 size; + + mask >>= PAGE_SHIFT; + mask |= size_or_mask; + size = -mask; + size <<= PAGE_SHIFT; + return size; +} + /* * Check and return the effective type for MTRR-MTRR type overlap. * Returns 1 if the effective type is UNCACHEABLE, else returns 0 @@ -92,17 +104,19 @@ static int check_type_overlap(u8 *prev, u8 *curr) } /* - * Returns the effective MTRR type for the region - * Error returns: - * - 0xFE - when the range is "not entirely covered" by _any_ var range MTRR - * - 0xFF - when MTRR is not enabled + * Error/Semi-error returns: + * 0xFF - when MTRR is not enabled + * *repeat == 1 implies [start:end] spanned across MTRR range and type returned + * corresponds only to [start:*partial_end]. + * Caller has to lookup again for [*partial_end:end]. */ -u8 mtrr_type_lookup(u64 start, u64 end) +static u8 __mtrr_type_lookup(u64 start, u64 end, u64 *partial_end, int *repeat) { int i; u64 base, mask; u8 prev_match, curr_match; + *repeat = 0; if (!mtrr_state_set) return 0xFF; @@ -153,8 +167,34 @@ u8 mtrr_type_lookup(u64 start, u64 end) start_state = ((start & mask) == (base & mask)); end_state = ((end & mask) == (base & mask)); - if (start_state != end_state) - return 0xFE; + + if (start_state != end_state) { + /* + * We have start:end spanning across an MTRR. + * We split the region into + * either + * (start:mtrr_end) (mtrr_end:end) + * or + * (start:mtrr_start) (mtrr_start:end) + * depending on kind of overlap. + * Return the type for first region and a pointer to + * the start of second region so that caller will + * lookup again on the second region. + * Note: This way we handle multiple overlaps as well. + */ + if (start_state) + *partial_end = base + get_mtrr_size(mask); + else + *partial_end = base; + + if (unlikely(*partial_end <= start)) { + WARN_ON(1); + *partial_end = start + PAGE_SIZE; + } + + end = *partial_end - 1; /* end is inclusive */ + *repeat = 1; + } if ((start & mask) != (base & mask)) continue; @@ -180,6 +220,36 @@ u8 mtrr_type_lookup(u64 start, u64 end) return mtrr_state.def_type; } +/* + * Returns the effective MTRR type for the region + * Error return: + * 0xFF - when MTRR is not enabled + */ +u8 mtrr_type_lookup(u64 start, u64 end) +{ + u8 type, prev_type; + int repeat; + u64 partial_end; + + type = __mtrr_type_lookup(start, end, &partial_end, &repeat); + + /* + * Common path is with repeat = 0. + * However, we can have cases where [start:end] spans across some + * MTRR range. Do repeated lookups for that case here. + */ + while (repeat) { + prev_type = type; + start = partial_end; + type = __mtrr_type_lookup(start, end, &partial_end, &repeat); + + if (check_type_overlap(&prev_type, &type)) + return type; + } + + return type; +} + /* Get the MSR pair relating to a var range */ static void get_mtrr_var_range(unsigned int index, struct mtrr_var_range *vr) -- cgit v0.10.2 From d0ed0c32662e756e7daf85e70a5a27a9c1111331 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Sat, 11 Sep 2010 22:10:54 -0700 Subject: x86: Remove pr_ uses of KERN_ Signed-off-by: Joe Perches Cc: Jiri Kosina LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/apb_timer.c b/arch/x86/kernel/apb_timer.c index 8dd7780..4004417 100644 --- a/arch/x86/kernel/apb_timer.c +++ b/arch/x86/kernel/apb_timer.c @@ -398,7 +398,7 @@ static int apbt_cpuhp_notify(struct notifier_block *n, } break; default: - pr_debug(KERN_INFO "APBT notified %lu, no action\n", action); + pr_debug("APBT notified %lu, no action\n", action); } return NOTIFY_OK; } @@ -552,7 +552,7 @@ bad_count: pr_debug("APB CS going back %lx:%lx:%lx ", t2, last_read, t2 - last_read); bad_count_x3: - pr_debug(KERN_INFO "tripple check enforced\n"); + pr_debug("triple check enforced\n"); t0 = apbt_readl(phy_cs_timer_id, APBTMR_N_CURRENT_VALUE); udelay(1); -- cgit v0.10.2 From 637bbdc5b83615ef9f45f50399d1c7f27473c713 Mon Sep 17 00:00:00 2001 From: Dave Young Date: Mon, 13 Sep 2010 20:19:03 +0800 Subject: sched: Remove unused PF_ALIGNWARN flag PF_ALIGNWARN is not implemented and it is for 486 as the comment. It is not likely someone will implement this flag feature. So here remove this flag and leave the valuable 0x00000001 for future use. Signed-off-by: Dave Young Cc: Peter Zijlstra Cc: Linus Torvalds LKML-Reference: <20100913121903.GB22238@darkstar> Signed-off-by: Ingo Molnar diff --git a/include/linux/sched.h b/include/linux/sched.h index b51c53c..cdf5669 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1682,8 +1682,6 @@ extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t * /* * Per process flags */ -#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */ - /* Not implemented yet, only for 486*/ #define PF_STARTING 0x00000002 /* being created */ #define PF_EXITING 0x00000004 /* getting shut down */ #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */ -- cgit v0.10.2 From 7740191cd909b75d75685fb08a5d1f54b8a9d28b Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 13 Sep 2010 17:47:00 -0400 Subject: sched: Fix string comparison in /proc/sched_features Fix incorrect handling of the following case: INTERACTIVE INTERACTIVE_SOMETHING_ELSE The comparison only checks up to each element's length. Changelog since v1: - Embellish using some Rostedtisms. [ mingo: ^^ == smaller and cleaner ] Signed-off-by: Mathieu Desnoyers Reviewed-by: Steven Rostedt Cc: Cc: Peter Zijlstra Cc: Tony Lindgren LKML-Reference: <20100913214700.GB16118@Krystal> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 26f83e2..b40b82e 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -721,7 +721,7 @@ sched_feat_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { char buf[64]; - char *cmp = buf; + char *cmp; int neg = 0; int i; @@ -732,6 +732,7 @@ sched_feat_write(struct file *filp, const char __user *ubuf, return -EFAULT; buf[cnt] = 0; + cmp = strstrip(buf); if (strncmp(buf, "NO_", 3) == 0) { neg = 1; @@ -739,9 +740,7 @@ sched_feat_write(struct file *filp, const char __user *ubuf, } for (i = 0; sched_feat_names[i]; i++) { - int len = strlen(sched_feat_names[i]); - - if (strncmp(cmp, sched_feat_names[i], len) == 0) { + if (strcmp(cmp, sched_feat_names[i]) == 0) { if (neg) sysctl_sched_features &= ~(1UL << i); else -- cgit v0.10.2 From 75e3cfbed6f71a8f151dc6e413b6ce3c390030cb Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Fri, 27 Aug 2010 11:09:48 -0700 Subject: x86, intr-remap: Set redirection hint in the IRTE Currently the redirection hint in the interrupt-remapping table entry is set to 0, which means the remapped interrupt is directed to the processors listed in the destination. So in logical flat mode in the presence of intr-remapping, this results in a single interrupt multi-casted to multiple cpu's as specified by the destination bit mask. But what we really want is to send that interrupt to one of the cpus based on the lowest priority delivery mode. Set the redirection hint in the IRTE to '1' to indicate that we want the remapped interrupt to be directed to only one of the processors listed in the destination. This fixes the issue of same interrupt getting delivered to multiple cpu's in the logical flat mode in the presence of interrupt-remapping. While there is no functional issue observed with this behavior, this will impact performance of such configurations (<=8 cpu's using logical flat mode in the presence of interrupt-remapping) Signed-off-by: Suresh Siddha LKML-Reference: <20100827181049.013051492@sbsiddha-MOBL3.sc.intel.com> Cc: Weidong Han Cc: # [v2.6.32+] Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index f1efeba..90f8a75 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1392,6 +1392,7 @@ int setup_ioapic_entry(int apic_id, int irq, irte.dlvry_mode = apic->irq_delivery_mode; irte.vector = vector; irte.dest_id = IRTE_DEST(destination); + irte.redir_hint = 1; /* Set source-id of interrupt request */ set_ioapic_sid(&irte, apic_id); @@ -3343,6 +3344,7 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, irte.dlvry_mode = apic->irq_delivery_mode; irte.vector = cfg->vector; irte.dest_id = IRTE_DEST(dest); + irte.redir_hint = 1; /* Set source-id of interrupt request */ if (pdev) -- cgit v0.10.2 From 62a92f4c69cd1d9361ad8c16be1dd16e6821bc15 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Fri, 27 Aug 2010 11:09:49 -0700 Subject: x86, intr-remap: Remove IRTE setup duplicate code Remove IRTE setup duplicate code with prepare_irte(). Signed-off-by: Suresh Siddha LKML-Reference: <20100827181049.095067319@sbsiddha-MOBL3.sc.intel.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/irq_remapping.h b/arch/x86/include/asm/irq_remapping.h index f275e22..8d84150 100644 --- a/arch/x86/include/asm/irq_remapping.h +++ b/arch/x86/include/asm/irq_remapping.h @@ -3,4 +3,31 @@ #define IRTE_DEST(dest) ((x2apic_mode) ? dest : dest << 8) +#ifdef CONFIG_INTR_REMAP +static inline void prepare_irte(struct irte *irte, int vector, + unsigned int dest) +{ + memset(irte, 0, sizeof(*irte)); + + irte->present = 1; + irte->dst_mode = apic->irq_dest_mode; + /* + * Trigger mode in the IRTE will always be edge, and for IO-APIC, the + * actual level or edge trigger will be setup in the IO-APIC + * RTE. This will help simplify level triggered irq migration. + * For more details, see the comments (in io_apic.c) explainig IO-APIC + * irq migration in the presence of interrupt-remapping. + */ + irte->trigger_mode = 0; + irte->dlvry_mode = apic->irq_delivery_mode; + irte->vector = vector; + irte->dest_id = IRTE_DEST(dest); + irte->redir_hint = 1; +} +#else +static void prepare_irte(struct irte *irte, int vector, unsigned int dest) +{ +} +#endif + #endif /* _ASM_X86_IRQ_REMAPPING_H */ diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 90f8a75..e8c95a2 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1377,22 +1377,7 @@ int setup_ioapic_entry(int apic_id, int irq, if (index < 0) panic("Failed to allocate IRTE for ioapic %d\n", apic_id); - memset(&irte, 0, sizeof(irte)); - - irte.present = 1; - irte.dst_mode = apic->irq_dest_mode; - /* - * Trigger mode in the IRTE will always be edge, and the - * actual level or edge trigger will be setup in the IO-APIC - * RTE. This will help simplify level triggered irq migration. - * For more details, see the comments above explainig IO-APIC - * irq migration in the presence of interrupt-remapping. - */ - irte.trigger_mode = 0; - irte.dlvry_mode = apic->irq_delivery_mode; - irte.vector = vector; - irte.dest_id = IRTE_DEST(destination); - irte.redir_hint = 1; + prepare_irte(&irte, vector, destination); /* Set source-id of interrupt request */ set_ioapic_sid(&irte, apic_id); @@ -3336,15 +3321,7 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, ir_index = map_irq_to_irte_handle(irq, &sub_handle); BUG_ON(ir_index == -1); - memset (&irte, 0, sizeof(irte)); - - irte.present = 1; - irte.dst_mode = apic->irq_dest_mode; - irte.trigger_mode = 0; /* edge */ - irte.dlvry_mode = apic->irq_delivery_mode; - irte.vector = cfg->vector; - irte.dest_id = IRTE_DEST(dest); - irte.redir_hint = 1; + prepare_irte(&irte, cfg->vector, dest); /* Set source-id of interrupt request */ if (pdev) -- cgit v0.10.2 From fa47f7e52874683a9659df2f1f143105f676dc0f Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Fri, 27 Aug 2010 11:09:50 -0700 Subject: x86, x2apic: Simplify apic init in SMP and UP builds Move enable_IR_x2apic() inside the default_setup_apic_routing(), and for SMP platforms, move the default_setup_apic_routing() after smp_sanity_check(). This cleans up the code that tries to avoid multiple calls to default_setup_apic_routing() when smp_sanity_check() fails (which goes through the APIC_init_uniprocessor() path). Signed-off-by: Suresh Siddha LKML-Reference: <20100827181049.173087246@sbsiddha-MOBL3.sc.intel.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index e3b534c..8cf86fb 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -1665,10 +1665,7 @@ int __init APIC_init_uniprocessor(void) } #endif -#ifndef CONFIG_SMP - enable_IR_x2apic(); default_setup_apic_routing(); -#endif verify_local_APIC(); connect_bsp_APIC(); diff --git a/arch/x86/kernel/apic/probe_64.c b/arch/x86/kernel/apic/probe_64.c index 83e9be4..f9e4e6a 100644 --- a/arch/x86/kernel/apic/probe_64.c +++ b/arch/x86/kernel/apic/probe_64.c @@ -54,6 +54,9 @@ static int apicid_phys_pkg_id(int initial_apic_id, int index_msb) */ void __init default_setup_apic_routing(void) { + + enable_IR_x2apic(); + #ifdef CONFIG_X86_X2APIC if (x2apic_mode #ifdef CONFIG_X86_UV diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 8b3bfc4..87a8c6b 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1109,8 +1109,6 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus) } set_cpu_sibling_map(0); - enable_IR_x2apic(); - default_setup_apic_routing(); if (smp_sanity_check(max_cpus) < 0) { printk(KERN_INFO "SMP disabled\n"); @@ -1118,6 +1116,8 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus) goto out; } + default_setup_apic_routing(); + preempt_disable(); if (read_apic_id() != boot_cpu_physical_apicid) { panic("Boot APIC ID in local APIC unexpected (%d vs %d)", -- cgit v0.10.2 From 31915ab4cbf507aadab40847cf9989da5e88b090 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Thu, 16 Sep 2010 14:42:25 +0200 Subject: sched: Remove branch hints within context_switch() With 710390d9 "sched: Optimize branch hint in context_switch()" the branch hint logic within context_switch() got inversed. In fact the hints "if (likely(!mm))" and "if (likely(!prev->mm))" mean that it is likely that the previous and next task are kernel threads. That assumption is certainly counter intuitive, but Tim has shown that at least with his workload this is true. Nevertheless the truth is: it depends on the current workload. So just remove the annotations which also improves readability. Reported-by: Tim Blechmann Signed-off-by: Heiko Carstens Cc: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20100916124225.GA2209@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index b40b82e..16a1129 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2845,14 +2845,14 @@ context_switch(struct rq *rq, struct task_struct *prev, */ arch_start_context_switch(prev); - if (likely(!mm)) { + if (!mm) { next->active_mm = oldmm; atomic_inc(&oldmm->mm_count); enter_lazy_tlb(oldmm, next); } else switch_mm(oldmm, mm, next); - if (likely(!prev->mm)) { + if (!prev->mm) { prev->active_mm = NULL; rq->prev_mm = oldmm; } -- cgit v0.10.2 From 16d3ea26f82271fef9b1c4523b5e1ea31fa39eec Mon Sep 17 00:00:00 2001 From: "Martin K. Petersen" Date: Fri, 10 Sep 2010 01:20:27 -0400 Subject: [SCSI] Fix VPD inquiry page wrapper Fix two bugs in the VPD page wrapper: - Don't return failure if the user asked for page 0 - The end of buffer check failed to account for the page header size and consequently didn't work Signed-off-by: Martin K. Petersen Cc: Stable Tree Signed-off-by: James Bottomley diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c index ad0ed21..348fba0 100644 --- a/drivers/scsi/scsi.c +++ b/drivers/scsi/scsi.c @@ -1046,13 +1046,13 @@ int scsi_get_vpd_page(struct scsi_device *sdev, u8 page, unsigned char *buf, /* If the user actually wanted this page, we can skip the rest */ if (page == 0) - return -EINVAL; + return 0; for (i = 0; i < min((int)buf[3], buf_len - 4); i++) if (buf[i + 4] == page) goto found; - if (i < buf[3] && i > buf_len) + if (i < buf[3] && i >= buf_len - 4) /* ran off the end of the buffer, give us benefit of doubt */ goto found; /* The device claims it doesn't support the requested page */ -- cgit v0.10.2 From 3ee48b6af49cf534ca2f481ecc484b156a41451d Mon Sep 17 00:00:00 2001 From: Cliff Wickman Date: Thu, 16 Sep 2010 11:44:02 -0500 Subject: mm, x86: Saving vmcore with non-lazy freeing of vmas During the reading of /proc/vmcore the kernel is doing ioremap()/iounmap() repeatedly. And the buildup of un-flushed vm_area_struct's is causing a great deal of overhead. (rb_next() is chewing up most of that time). This solution is to provide function set_iounmap_nonlazy(). It causes a subsequent call to iounmap() to immediately purge the vma area (with try_purge_vmap_area_lazy()). With this patch we have seen the time for writing a 250MB compressed dump drop from 71 seconds to 44 seconds. Signed-off-by: Cliff Wickman Cc: Andrew Morton Cc: kexec@lists.infradead.org Cc: LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h index 30a3e97..6a45ec4 100644 --- a/arch/x86/include/asm/io.h +++ b/arch/x86/include/asm/io.h @@ -206,6 +206,7 @@ static inline void __iomem *ioremap(resource_size_t offset, unsigned long size) extern void iounmap(volatile void __iomem *addr); +extern void set_iounmap_nonlazy(void); #ifdef __KERNEL__ diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c index bf43188..9948288 100644 --- a/arch/x86/kernel/crash_dump_64.c +++ b/arch/x86/kernel/crash_dump_64.c @@ -46,6 +46,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, } else memcpy(buf, vaddr + offset, csize); + set_iounmap_nonlazy(); iounmap(vaddr); return csize; } diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 6b8889d..d8087f0 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -517,6 +517,15 @@ static atomic_t vmap_lazy_nr = ATOMIC_INIT(0); static void purge_fragmented_blocks_allcpus(void); /* + * called before a call to iounmap() if the caller wants vm_area_struct's + * immediately freed. + */ +void set_iounmap_nonlazy(void) +{ + atomic_set(&vmap_lazy_nr, lazy_max_pages()+1); +} + +/* * Purges all lazily-freed vmap areas. * * If sync is 0 then don't purge if there is already a purge in progress. -- cgit v0.10.2 From 3518dd14ca888085797ca8d3a9e11c8ef9e7ae68 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Fri, 17 Sep 2010 18:07:45 +0200 Subject: x86, cacheinfo: Fix dependency of AMD L3 CID L3 cache index disable code uses PCI accesses to AMD northbridge functions. Currently the code is #ifdef CONFIG_CPU_SUP_AMD. But it should be #if (defined(CONFIG_CPU_SUP_AMD) && defined(CONFIG_PCI)) which in the end is a dependency to K8_NB. Signed-off-by: Andreas Herrmann LKML-Reference: <20100917160744.GF4958@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 898c2f4..2521cdc 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -306,7 +306,7 @@ struct _cache_attr { ssize_t (*store)(struct _cpuid4_info *, const char *, size_t count); }; -#ifdef CONFIG_CPU_SUP_AMD +#ifdef CONFIG_K8_NB /* * L3 cache descriptors @@ -556,12 +556,12 @@ static struct _cache_attr cache_disable_0 = __ATTR(cache_disable_0, 0644, static struct _cache_attr cache_disable_1 = __ATTR(cache_disable_1, 0644, show_cache_disable_1, store_cache_disable_1); -#else /* CONFIG_CPU_SUP_AMD */ +#else /* CONFIG_K8_NB */ static void __cpuinit amd_check_l3_disable(struct _cpuid4_info_regs *this_leaf, int index) { }; -#endif /* CONFIG_CPU_SUP_AMD */ +#endif /* CONFIG_K8_NB */ static int __cpuinit cpuid4_cache_lookup_regs(int index, @@ -1000,7 +1000,7 @@ static struct attribute *default_attrs[] = { static struct attribute *default_l3_attrs[] = { DEFAULT_SYSFS_CACHE_ATTRS, -#ifdef CONFIG_CPU_SUP_AMD +#ifdef CONFIG_K8_NB &cache_disable_0.attr, &cache_disable_1.attr, #endif -- cgit v0.10.2 From 900f9ac9f12dc3dd6fc8e33e16df172eafcaead6 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Fri, 17 Sep 2010 18:02:54 +0200 Subject: x86, k8-gart: Decouple handling of garts and northbridges So far we only provide num_k8_northbridges. This is required in different areas (e.g. L3 cache index disable, GART). But not all AMD CPUs provide a GART. Thus it is useful to split off the GART handling from the generic caching of AMD northbridge misc devices. Signed-off-by: Andreas Herrmann LKML-Reference: <20100917160254.GC4958@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/k8.h b/arch/x86/include/asm/k8.h index af00bd1..9cee145 100644 --- a/arch/x86/include/asm/k8.h +++ b/arch/x86/include/asm/k8.h @@ -7,24 +7,27 @@ extern struct pci_device_id k8_nb_ids[]; struct bootnode; extern int early_is_k8_nb(u32 value); -extern struct pci_dev **k8_northbridges; -extern int num_k8_northbridges; extern int cache_k8_northbridges(void); extern void k8_flush_garts(void); extern int k8_get_nodes(struct bootnode *nodes); extern int k8_numa_init(unsigned long start_pfn, unsigned long end_pfn); extern int k8_scan_nodes(void); +struct k8_northbridge_info { + u16 num; + u8 gart_supported; + struct pci_dev **nb_misc; +}; +extern struct k8_northbridge_info k8_northbridges; + #ifdef CONFIG_K8_NB -extern int num_k8_northbridges; static inline struct pci_dev *node_to_k8_nb_misc(int node) { - return (node < num_k8_northbridges) ? k8_northbridges[node] : NULL; + return (node < k8_northbridges.num) ? k8_northbridges.nb_misc[node] : NULL; } #else -#define num_k8_northbridges 0 static inline struct pci_dev *node_to_k8_nb_misc(int node) { diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 2521cdc..6fdfb0b2 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -369,7 +369,7 @@ static void __cpuinit amd_check_l3_disable(struct _cpuid4_info_regs *this_leaf, return; /* not in virtualized environments */ - if (num_k8_northbridges == 0) + if (k8_northbridges.num == 0) return; /* @@ -377,7 +377,7 @@ static void __cpuinit amd_check_l3_disable(struct _cpuid4_info_regs *this_leaf, * never freed but this is done only on shutdown so it doesn't matter. */ if (!l3_caches) { - int size = num_k8_northbridges * sizeof(struct amd_l3_cache *); + int size = k8_northbridges.num * sizeof(struct amd_l3_cache *); l3_caches = kzalloc(size, GFP_ATOMIC); if (!l3_caches) diff --git a/arch/x86/kernel/k8.c b/arch/x86/kernel/k8.c index 0f7bc20..5de1b6b 100644 --- a/arch/x86/kernel/k8.c +++ b/arch/x86/kernel/k8.c @@ -10,9 +10,6 @@ #include #include -int num_k8_northbridges; -EXPORT_SYMBOL(num_k8_northbridges); - static u32 *flush_words; struct pci_device_id k8_nb_ids[] = { @@ -22,7 +19,7 @@ struct pci_device_id k8_nb_ids[] = { }; EXPORT_SYMBOL(k8_nb_ids); -struct pci_dev **k8_northbridges; +struct k8_northbridge_info k8_northbridges; EXPORT_SYMBOL(k8_northbridges); static struct pci_dev *next_k8_northbridge(struct pci_dev *dev) @@ -40,36 +37,44 @@ int cache_k8_northbridges(void) int i; struct pci_dev *dev; - if (num_k8_northbridges) + if (k8_northbridges.num) return 0; dev = NULL; while ((dev = next_k8_northbridge(dev)) != NULL) - num_k8_northbridges++; + k8_northbridges.num++; + + /* some CPU families (e.g. family 0x11) do not support GART */ + if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10) + k8_northbridges.gart_supported = 1; - k8_northbridges = kmalloc((num_k8_northbridges + 1) * sizeof(void *), - GFP_KERNEL); - if (!k8_northbridges) + k8_northbridges.nb_misc = kmalloc((k8_northbridges.num + 1) * + sizeof(void *), GFP_KERNEL); + if (!k8_northbridges.nb_misc) return -ENOMEM; - if (!num_k8_northbridges) { - k8_northbridges[0] = NULL; + if (!k8_northbridges.num) { + k8_northbridges.nb_misc[0] = NULL; return 0; } - flush_words = kmalloc(num_k8_northbridges * sizeof(u32), GFP_KERNEL); - if (!flush_words) { - kfree(k8_northbridges); - return -ENOMEM; + if (k8_northbridges.gart_supported) { + flush_words = kmalloc(k8_northbridges.num * sizeof(u32), + GFP_KERNEL); + if (!flush_words) { + kfree(k8_northbridges.nb_misc); + return -ENOMEM; + } } dev = NULL; i = 0; while ((dev = next_k8_northbridge(dev)) != NULL) { - k8_northbridges[i] = dev; - pci_read_config_dword(dev, 0x9c, &flush_words[i++]); + k8_northbridges.nb_misc[i] = dev; + if (k8_northbridges.gart_supported) + pci_read_config_dword(dev, 0x9c, &flush_words[i++]); } - k8_northbridges[i] = NULL; + k8_northbridges.nb_misc[i] = NULL; return 0; } EXPORT_SYMBOL_GPL(cache_k8_northbridges); @@ -93,22 +98,25 @@ void k8_flush_garts(void) unsigned long flags; static DEFINE_SPINLOCK(gart_lock); + if (!k8_northbridges.gart_supported) + return; + /* Avoid races between AGP and IOMMU. In theory it's not needed but I'm not sure if the hardware won't lose flush requests when another is pending. This whole thing is so expensive anyways that it doesn't matter to serialize more. -AK */ spin_lock_irqsave(&gart_lock, flags); flushed = 0; - for (i = 0; i < num_k8_northbridges; i++) { - pci_write_config_dword(k8_northbridges[i], 0x9c, + for (i = 0; i < k8_northbridges.num; i++) { + pci_write_config_dword(k8_northbridges.nb_misc[i], 0x9c, flush_words[i]|1); flushed++; } - for (i = 0; i < num_k8_northbridges; i++) { + for (i = 0; i < k8_northbridges.num; i++) { u32 w; /* Make sure the hardware actually executed the flush*/ for (;;) { - pci_read_config_dword(k8_northbridges[i], + pci_read_config_dword(k8_northbridges.nb_misc[i], 0x9c, &w); if (!(w & 1)) break; diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index 0f7f130..8f214a2 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -560,8 +560,11 @@ static void enable_gart_translations(void) { int i; - for (i = 0; i < num_k8_northbridges; i++) { - struct pci_dev *dev = k8_northbridges[i]; + if (!k8_northbridges.gart_supported) + return; + + for (i = 0; i < k8_northbridges.num; i++) { + struct pci_dev *dev = k8_northbridges.nb_misc[i]; enable_gart_translation(dev, __pa(agp_gatt_table)); } @@ -592,10 +595,13 @@ static void gart_fixup_northbridges(struct sys_device *dev) if (!fix_up_north_bridges) return; + if (!k8_northbridges.gart_supported) + return; + pr_info("PCI-DMA: Restoring GART aperture settings\n"); - for (i = 0; i < num_k8_northbridges; i++) { - struct pci_dev *dev = k8_northbridges[i]; + for (i = 0; i < k8_northbridges.num; i++) { + struct pci_dev *dev = k8_northbridges.nb_misc[i]; /* * Don't enable translations just yet. That is the next @@ -649,8 +655,8 @@ static __init int init_k8_gatt(struct agp_kern_info *info) aper_size = aper_base = info->aper_size = 0; dev = NULL; - for (i = 0; i < num_k8_northbridges; i++) { - dev = k8_northbridges[i]; + for (i = 0; i < k8_northbridges.num; i++) { + dev = k8_northbridges.nb_misc[i]; new_aper_base = read_aperture(dev, &new_aper_size); if (!new_aper_base) goto nommu; @@ -718,10 +724,13 @@ static void gart_iommu_shutdown(void) if (!no_agp) return; - for (i = 0; i < num_k8_northbridges; i++) { + if (!k8_northbridges.gart_supported) + return; + + for (i = 0; i < k8_northbridges.num; i++) { u32 ctl; - dev = k8_northbridges[i]; + dev = k8_northbridges.nb_misc[i]; pci_read_config_dword(dev, AMD64_GARTAPERTURECTL, &ctl); ctl &= ~GARTEN; @@ -739,7 +748,7 @@ int __init gart_iommu_init(void) unsigned long scratch; long i; - if (num_k8_northbridges == 0) + if (!k8_northbridges.gart_supported) return 0; #ifndef CONFIG_AGP_AMD64 diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c index 70312da..bdf00d58 100644 --- a/drivers/char/agp/amd64-agp.c +++ b/drivers/char/agp/amd64-agp.c @@ -124,7 +124,7 @@ static int amd64_fetch_size(void) u32 temp; struct aper_size_info_32 *values; - dev = k8_northbridges[0]; + dev = k8_northbridges.nb_misc[0]; if (dev==NULL) return 0; @@ -181,10 +181,14 @@ static int amd_8151_configure(void) unsigned long gatt_bus = virt_to_phys(agp_bridge->gatt_table_real); int i; + if (!k8_northbridges.gart_supported) + return 0; + /* Configure AGP regs in each x86-64 host bridge. */ - for (i = 0; i < num_k8_northbridges; i++) { + for (i = 0; i < k8_northbridges.num; i++) { agp_bridge->gart_bus_addr = - amd64_configure(k8_northbridges[i], gatt_bus); + amd64_configure(k8_northbridges.nb_misc[i], + gatt_bus); } k8_flush_garts(); return 0; @@ -195,8 +199,12 @@ static void amd64_cleanup(void) { u32 tmp; int i; - for (i = 0; i < num_k8_northbridges; i++) { - struct pci_dev *dev = k8_northbridges[i]; + + if (!k8_northbridges.gart_supported) + return; + + for (i = 0; i < k8_northbridges.num; i++) { + struct pci_dev *dev = k8_northbridges.nb_misc[i]; /* disable gart translation */ pci_read_config_dword(dev, AMD64_GARTAPERTURECTL, &tmp); tmp &= ~AMD64_GARTEN; @@ -319,16 +327,19 @@ static __devinit int fix_northbridge(struct pci_dev *nb, struct pci_dev *agp, return 0; } -static __devinit int cache_nbs (struct pci_dev *pdev, u32 cap_ptr) +static __devinit int cache_nbs(struct pci_dev *pdev, u32 cap_ptr) { int i; if (cache_k8_northbridges() < 0) return -ENODEV; + if (!k8_northbridges.gart_supported) + return -ENODEV; + i = 0; - for (i = 0; i < num_k8_northbridges; i++) { - struct pci_dev *dev = k8_northbridges[i]; + for (i = 0; i < k8_northbridges.num; i++) { + struct pci_dev *dev = k8_northbridges.nb_misc[i]; if (fix_northbridge(dev, pdev, cap_ptr) < 0) { dev_err(&dev->dev, "no usable aperture found\n"); #ifdef __x86_64__ @@ -405,7 +416,8 @@ static int __devinit uli_agp_init(struct pci_dev *pdev) } /* shadow x86-64 registers into ULi registers */ - pci_read_config_dword (k8_northbridges[0], AMD64_GARTAPERTUREBASE, &httfea); + pci_read_config_dword (k8_northbridges.nb_misc[0], AMD64_GARTAPERTUREBASE, + &httfea); /* if x86-64 aperture base is beyond 4G, exit here */ if ((httfea & 0x7fff) >> (32 - 25)) { @@ -472,7 +484,8 @@ static int nforce3_agp_init(struct pci_dev *pdev) pci_write_config_dword(dev1, NVIDIA_X86_64_1_APSIZE, tmp); /* shadow x86-64 registers into NVIDIA registers */ - pci_read_config_dword (k8_northbridges[0], AMD64_GARTAPERTUREBASE, &apbase); + pci_read_config_dword (k8_northbridges.nb_misc[0], AMD64_GARTAPERTUREBASE, + &apbase); /* if x86-64 aperture base is beyond 4G, exit here */ if ( (apbase & 0x7fff) >> (32 - 25) ) { diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c index e7d5d6b..5babf6f 100644 --- a/drivers/edac/amd64_edac.c +++ b/drivers/edac/amd64_edac.c @@ -2927,7 +2927,7 @@ static int __init amd64_edac_init(void) * to finish initialization of the MC instances. */ err = -ENODEV; - for (nb = 0; nb < num_k8_northbridges; nb++) { + for (nb = 0; nb < k8_northbridges.num; nb++) { if (!pvt_lookup[nb]) continue; -- cgit v0.10.2 From bc83cccc761953f878088cdfa682de0970b5561f Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Fri, 17 Sep 2010 15:36:40 -0700 Subject: x86, mwait: Move mwait constants to a common header file We have MWAIT constants spread across three different .c files, for no good reason. Move them all into a common header file. Signed-off-by: H. Peter Anvin Reviewed-by: Arjan van de Ven Cc: Len Brown LKML-Reference: diff --git a/arch/x86/include/asm/mwait.h b/arch/x86/include/asm/mwait.h new file mode 100644 index 0000000..bcdff99 --- /dev/null +++ b/arch/x86/include/asm/mwait.h @@ -0,0 +1,15 @@ +#ifndef _ASM_X86_MWAIT_H +#define _ASM_X86_MWAIT_H + +#define MWAIT_SUBSTATE_MASK 0xf +#define MWAIT_CSTATE_MASK 0xf +#define MWAIT_SUBSTATE_SIZE 4 +#define MWAIT_MAX_NUM_CSTATES 8 + +#define CPUID_MWAIT_LEAF 5 +#define CPUID5_ECX_EXTENSIONS_SUPPORTED 0x1 +#define CPUID5_ECX_INTERRUPT_BREAK 0x2 + +#define MWAIT_ECX_INTERRUPT_BREAK 0x1 + +#endif /* _ASM_X86_MWAIT_H */ diff --git a/arch/x86/kernel/acpi/cstate.c b/arch/x86/kernel/acpi/cstate.c index fb7a5f0..bcc4add 100644 --- a/arch/x86/kernel/acpi/cstate.c +++ b/arch/x86/kernel/acpi/cstate.c @@ -13,6 +13,7 @@ #include #include +#include /* * Initialize bm_flags based on the CPU cache properties @@ -65,16 +66,6 @@ static struct cstate_entry *cpu_cstate_entry; /* per CPU ptr */ static short mwait_supported[ACPI_PROCESSOR_MAX_POWER]; -#define MWAIT_SUBSTATE_MASK (0xf) -#define MWAIT_CSTATE_MASK (0xf) -#define MWAIT_SUBSTATE_SIZE (4) - -#define CPUID_MWAIT_LEAF (5) -#define CPUID5_ECX_EXTENSIONS_SUPPORTED (0x1) -#define CPUID5_ECX_INTERRUPT_BREAK (0x2) - -#define MWAIT_ECX_INTERRUPT_BREAK (0x1) - #define NATIVE_CSTATE_BEYOND_HALT (2) static long acpi_processor_ffh_cstate_probe_cpu(void *_cx) diff --git a/drivers/acpi/acpi_pad.c b/drivers/acpi/acpi_pad.c index b76848c..ca2694d 100644 --- a/drivers/acpi/acpi_pad.c +++ b/drivers/acpi/acpi_pad.c @@ -30,18 +30,13 @@ #include #include #include +#include #define ACPI_PROCESSOR_AGGREGATOR_CLASS "acpi_pad" #define ACPI_PROCESSOR_AGGREGATOR_DEVICE_NAME "Processor Aggregator" #define ACPI_PROCESSOR_AGGREGATOR_NOTIFY 0x80 static DEFINE_MUTEX(isolated_cpus_lock); -#define MWAIT_SUBSTATE_MASK (0xf) -#define MWAIT_CSTATE_MASK (0xf) -#define MWAIT_SUBSTATE_SIZE (4) -#define CPUID_MWAIT_LEAF (5) -#define CPUID5_ECX_EXTENSIONS_SUPPORTED (0x1) -#define CPUID5_ECX_INTERRUPT_BREAK (0x2) static unsigned long power_saving_mwait_eax; static unsigned char tsc_detected_unstable; diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c index a10152b..065c804 100755 --- a/drivers/idle/intel_idle.c +++ b/drivers/idle/intel_idle.c @@ -59,18 +59,11 @@ #include /* ktime_get_real() */ #include #include +#include #define INTEL_IDLE_VERSION "0.4" #define PREFIX "intel_idle: " -#define MWAIT_SUBSTATE_MASK (0xf) -#define MWAIT_CSTATE_MASK (0xf) -#define MWAIT_SUBSTATE_SIZE (4) -#define MWAIT_MAX_NUM_CSTATES 8 -#define CPUID_MWAIT_LEAF (5) -#define CPUID5_ECX_EXTENSIONS_SUPPORTED (0x1) -#define CPUID5_ECX_INTERRUPT_BREAK (0x2) - static struct cpuidle_driver intel_idle_driver = { .name = "intel_idle", .owner = THIS_MODULE, -- cgit v0.10.2 From ea53069231f9317062910d6e772cca4ce93de8c8 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Fri, 17 Sep 2010 15:39:11 -0700 Subject: x86, hotplug: Use mwait to offline a processor, fix the legacy case The code in native_play_dead() has a number of problems: 1. We should use MWAIT when available, to put ourselves into a deeper sleep state. 2. We use the existence of CLFLUSH to determine if WBINVD is safe, but that is totally bogus -- WBINVD is 486+, whereas CLFLUSH is a much later addition. 3. We should do WBINVD inside the loop, just in case of something like setting an A bit on page tables. Pointed out by Arjan van de Ven. This code is based in part of a previous patch by Venki Pallipadi, but unlike that patch this one keeps all the detection code local instead of pre-caching a bunch of information. We're shutting down the CPU; there is absolutely no hurry. This patch moves all the code to C and deletes the global wbinvd_halt() which is broken anyway. Originally-by: Venkatesh Pallipadi Signed-off-by: H. Peter Anvin Reviewed-by: Arjan van de Ven Cc: Len Brown Cc: Venkatesh Pallipadi Cc: Peter Zijlstra LKML-Reference: <20090522232230.162239000@intel.com> diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index 325b7bd..f358241 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -764,29 +764,6 @@ extern unsigned long idle_halt; extern unsigned long idle_nomwait; extern bool c1e_detected; -/* - * on systems with caches, caches must be flashed as the absolute - * last instruction before going into a suspended halt. Otherwise, - * dirty data can linger in the cache and become stale on resume, - * leading to strange errors. - * - * perform a variety of operations to guarantee that the compiler - * will not reorder instructions. wbinvd itself is serializing - * so the processor will not reorder. - * - * Systems without cache can just go into halt. - */ -static inline void wbinvd_halt(void) -{ - mb(); - /* check for clflush to determine if wbinvd is legal */ - if (cpu_has_clflush) - asm volatile("cli; wbinvd; 1: hlt; jmp 1b" : : : "memory"); - else - while (1) - halt(); -} - extern void enable_sep_cpu(void); extern int sysenter_setup(void); diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 8b3bfc4..07bf423 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -62,6 +62,7 @@ #include #include #include +#include #include #include #include @@ -1383,11 +1384,71 @@ void play_dead_common(void) local_irq_disable(); } +/* + * We need to flush the caches before going to sleep, lest we have + * dirty data in our caches when we come back up. + */ +static inline void mwait_play_dead(void) +{ + unsigned int eax, ebx, ecx, edx; + unsigned int highest_cstate = 0; + unsigned int highest_subcstate = 0; + int i; + + if (!cpu_has(¤t_cpu_data, X86_FEATURE_MWAIT)) + return; + if (current_cpu_data.cpuid_level < CPUID_MWAIT_LEAF) + return; + + eax = CPUID_MWAIT_LEAF; + ecx = 0; + native_cpuid(&eax, &ebx, &ecx, &edx); + + /* + * eax will be 0 if EDX enumeration is not valid. + * Initialized below to cstate, sub_cstate value when EDX is valid. + */ + if (!(ecx & CPUID5_ECX_EXTENSIONS_SUPPORTED)) { + eax = 0; + } else { + edx >>= MWAIT_SUBSTATE_SIZE; + for (i = 0; i < 7 && edx; i++, edx >>= MWAIT_SUBSTATE_SIZE) { + if (edx & MWAIT_SUBSTATE_MASK) { + highest_cstate = i; + highest_subcstate = edx & MWAIT_SUBSTATE_MASK; + } + } + eax = (highest_cstate << MWAIT_SUBSTATE_SIZE) | + (highest_subcstate - 1); + } + + while (1) { + mb(); + wbinvd(); + __monitor(¤t_thread_info()->flags, 0, 0); + mb(); + __mwait(eax, 0); + } +} + +static inline void hlt_play_dead(void) +{ + while (1) { + mb(); + if (current_cpu_data.x86 >= 4) + wbinvd(); + mb(); + native_halt(); + } +} + void native_play_dead(void) { play_dead_common(); tboot_shutdown(TB_SHUTDOWN_WFS); - wbinvd_halt(); + + mwait_play_dead(); /* Only returns on failure */ + hlt_play_dead(); } #else /* ... !CONFIG_HOTPLUG_CPU */ -- cgit v0.10.2 From a68e5c94f7d3dd64fef34dd5d97e365cae4bb42a Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Fri, 17 Sep 2010 17:06:46 -0700 Subject: x86, hotplug: Move WBINVD back outside the play_dead loop On processors with hyperthreading, when only one thread is offlined the other thread can cause a spurious wakeup on the idled thread. We do not want to re-WBINVD when that happens. Ideally, we should simply skip WBINVD unless we're the last thread on a particular core to shut down, but there might be similar issues elsewhere in the system. Thus, revert to previous behavior of only WBINVD outside the loop. Partly as a result, remove the mb()'s around it: they are not necessary since wbinvd() is a serializing instruction, but they were intended to make sure the compiler didn't do any funny loop optimizations. Reported-by: Asit Mallick Signed-off-by: H. Peter Anvin Cc: Arjan van de Ven Cc: Len Brown Cc: Venkatesh Pallipadi Cc: Peter Zijlstra LKML-Reference: diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 07bf423..55c80ffb 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1422,9 +1422,9 @@ static inline void mwait_play_dead(void) (highest_subcstate - 1); } + wbinvd(); + while (1) { - mb(); - wbinvd(); __monitor(¤t_thread_info()->flags, 0, 0); mb(); __mwait(eax, 0); @@ -1433,11 +1433,10 @@ static inline void mwait_play_dead(void) static inline void hlt_play_dead(void) { + if (current_cpu_data.x86 >= 4) + wbinvd(); + while (1) { - mb(); - if (current_cpu_data.x86 >= 4) - wbinvd(); - mb(); native_halt(); } } -- cgit v0.10.2 From 1dcc41bb037533839753df983d31778b30b67d93 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Tue, 14 Sep 2010 21:43:46 +0900 Subject: futex: Change 3rd arg of fetch_robust_entry() to unsigned int* Sparse complains: kernel/futex.c:2495:59: warning: incorrect type in argument 3 (different signedness) Make 3rd argument of fetch_robust_entry() 'unsigned int'. Signed-off-by: Namhyung Kim Cc: Peter Zijlstra Cc: Darren Hart LKML-Reference: <1284468228-8723-1-git-send-email-namhyung@gmail.com> Signed-off-by: Thomas Gleixner diff --git a/kernel/futex.c b/kernel/futex.c index 6a3a5fa..464de27 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -2458,7 +2458,7 @@ retry: */ static inline int fetch_robust_entry(struct robust_list __user **entry, struct robust_list __user * __user *head, - int *pi) + unsigned int *pi) { unsigned long uentry; diff --git a/kernel/futex_compat.c b/kernel/futex_compat.c index d49afb2..06da4df 100644 --- a/kernel/futex_compat.c +++ b/kernel/futex_compat.c @@ -19,7 +19,7 @@ */ static inline int fetch_robust_entry(compat_uptr_t *uentry, struct robust_list __user **entry, - compat_uptr_t __user *head, int *pi) + compat_uptr_t __user *head, unsigned int *pi) { if (get_user(*uentry, head)) return -EFAULT; -- cgit v0.10.2 From a3c74c52570c0c4ac90c9a0216de800c39089ba7 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Tue, 14 Sep 2010 21:43:47 +0900 Subject: futex: Mark restart_block.futex.uaddr[2] __user @uaddr and @uaddr2 fields in restart_block.futex are user pointers. Add __user and remove unnecessary casts. Signed-off-by: Namhyung Kim Cc: Peter Zijlstra Cc: Darren Hart LKML-Reference: <1284468228-8723-2-git-send-email-namhyung@gmail.com> Signed-off-by: Thomas Gleixner diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h index a8cc4e1..c906965 100644 --- a/include/linux/thread_info.h +++ b/include/linux/thread_info.h @@ -23,12 +23,12 @@ struct restart_block { }; /* For futex_wait and futex_wait_requeue_pi */ struct { - u32 *uaddr; + u32 __user *uaddr; u32 val; u32 flags; u32 bitset; u64 time; - u32 *uaddr2; + u32 __user *uaddr2; } futex; /* For nanosleep */ struct { diff --git a/kernel/futex.c b/kernel/futex.c index 464de27..45e448a 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -1843,7 +1843,7 @@ retry: restart = ¤t_thread_info()->restart_block; restart->fn = futex_wait_restart; - restart->futex.uaddr = (u32 *)uaddr; + restart->futex.uaddr = uaddr; restart->futex.val = val; restart->futex.time = abs_time->tv64; restart->futex.bitset = bitset; @@ -1869,7 +1869,7 @@ out: static long futex_wait_restart(struct restart_block *restart) { - u32 __user *uaddr = (u32 __user *)restart->futex.uaddr; + u32 __user *uaddr = restart->futex.uaddr; int fshared = 0; ktime_t t, *tp = NULL; -- cgit v0.10.2 From 15e408cd6ccc3f4f453d87ccd5bc7a84d59feb96 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Tue, 14 Sep 2010 21:43:48 +0900 Subject: futex: Add lock context annotations queue_lock/unlock/me() and unqueue_me_pi() grab/release spinlocks but are missing proper annotations. Add them. Signed-off-by: Namhyung Kim Cc: Peter Zijlstra Cc: Darren Hart LKML-Reference: <1284468228-8723-3-git-send-email-namhyung@gmail.com> Signed-off-by: Thomas Gleixner diff --git a/kernel/futex.c b/kernel/futex.c index 45e448a..92a31d4 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -1360,6 +1360,7 @@ out: /* The key must be already stored in q->key. */ static inline struct futex_hash_bucket *queue_lock(struct futex_q *q) + __acquires(&hb->lock) { struct futex_hash_bucket *hb; @@ -1373,6 +1374,7 @@ static inline struct futex_hash_bucket *queue_lock(struct futex_q *q) static inline void queue_unlock(struct futex_q *q, struct futex_hash_bucket *hb) + __releases(&hb->lock) { spin_unlock(&hb->lock); drop_futex_key_refs(&q->key); @@ -1391,6 +1393,7 @@ queue_unlock(struct futex_q *q, struct futex_hash_bucket *hb) * an example). */ static inline void queue_me(struct futex_q *q, struct futex_hash_bucket *hb) + __releases(&hb->lock) { int prio; @@ -1471,6 +1474,7 @@ retry: * and dropped here. */ static void unqueue_me_pi(struct futex_q *q) + __releases(q->lock_ptr) { WARN_ON(plist_node_empty(&q->list)); plist_del(&q->list, &q->list.plist); -- cgit v0.10.2 From 38e7afe96c7c0ad900824911c61fdb04078033dc Mon Sep 17 00:00:00 2001 From: Henrik Rydberg Date: Sun, 19 Sep 2010 16:25:36 -0700 Subject: Input: uinput - setup MT usage during device creation The input devices created by uinput do not currently handle multitouch properly. All events will appear as if they came from slot zero, and the input event buffers are not adjusted. This patch creates the MT slots during setup, and sets the number of events per packet based on the MT usage. Signed-off-by: Henrik Rydberg Signed-off-by: Dmitry Torokhov diff --git a/drivers/input/misc/uinput.c b/drivers/input/misc/uinput.c index 0d4266a..3606985 100644 --- a/drivers/input/misc/uinput.c +++ b/drivers/input/misc/uinput.c @@ -404,6 +404,13 @@ static int uinput_setup_device(struct uinput_device *udev, const char __user *bu retval = uinput_validate_absbits(dev); if (retval < 0) goto exit; + if (test_bit(ABS_MT_SLOT, dev->absbit)) { + int nslot = input_abs_get_max(dev, ABS_MT_SLOT) + 1; + input_mt_create_slots(dev, nslot); + input_set_events_per_packet(dev, 6 * nslot); + } else if (test_bit(ABS_MT_POSITION_X, dev->absbit)) { + input_set_events_per_packet(dev, 60); + } } udev->state = UIST_SETUP_COMPLETE; -- cgit v0.10.2 From ff8f33c8b30d7b7efdcf2548c7f6e64db6a89b29 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 11 Aug 2010 09:37:53 +0100 Subject: GFS2: New truncate sequence This updates GFS2's truncate code to use the new truncate sequence correctly. This is a stepping stone to being able to remove ip->i_disksize in favour of using i_size everywhere now that the two sizes are always identical. Signed-off-by: Steven Whitehouse Cc: Nick Piggin Cc: Christoph Hellwig diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 194fe16..f687f25 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -696,13 +696,11 @@ out: page_cache_release(page); - /* - * XXX(truncate): the call below should probably be replaced with - * a call to the gfs2-specific truncate blocks helper to actually - * release disk blocks.. - */ + gfs2_trans_end(sdp); if (pos + len > ip->i_inode.i_size) - truncate_setsize(&ip->i_inode, ip->i_inode.i_size); + gfs2_trim_blocks(&ip->i_inode); + goto out_trans_fail; + out_endtrans: gfs2_trans_end(sdp); out_trans_fail: diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 6f48280..20b971a 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -50,7 +50,7 @@ struct strip_mine { * @ip: the inode * @dibh: the dinode buffer * @block: the block number that was allocated - * @private: any locked page held by the caller process + * @page: The (optional) page. This is looked up if @page is NULL * * Returns: errno */ @@ -109,8 +109,7 @@ static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh, /** * gfs2_unstuff_dinode - Unstuff a dinode when the data has grown too big * @ip: The GFS2 inode to unstuff - * @unstuffer: the routine that handles unstuffing a non-zero length file - * @private: private data for the unstuffer + * @page: The (optional) page. This is looked up if the @page is NULL * * This routine unstuffs a dinode and returns it to a "normal" state such * that the height can be grown in the traditional way. @@ -885,83 +884,14 @@ out: } /** - * do_grow - Make a file look bigger than it is - * @ip: the inode - * @size: the size to set the file to - * - * Called with an exclusive lock on @ip. - * - * Returns: errno - */ - -static int do_grow(struct gfs2_inode *ip, u64 size) -{ - struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); - struct gfs2_alloc *al; - struct buffer_head *dibh; - int error; - - al = gfs2_alloc_get(ip); - if (!al) - return -ENOMEM; - - error = gfs2_quota_lock_check(ip); - if (error) - goto out; - - al->al_requested = sdp->sd_max_height + RES_DATA; - - error = gfs2_inplace_reserve(ip); - if (error) - goto out_gunlock_q; - - error = gfs2_trans_begin(sdp, - sdp->sd_max_height + al->al_rgd->rd_length + - RES_JDATA + RES_DINODE + RES_STATFS + RES_QUOTA, 0); - if (error) - goto out_ipres; - - error = gfs2_meta_inode_buffer(ip, &dibh); - if (error) - goto out_end_trans; - - if (size > sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)) { - if (gfs2_is_stuffed(ip)) { - error = gfs2_unstuff_dinode(ip, NULL); - if (error) - goto out_brelse; - } - } - - ip->i_disksize = size; - ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; - gfs2_trans_add_bh(ip->i_gl, dibh, 1); - gfs2_dinode_out(ip, dibh->b_data); - -out_brelse: - brelse(dibh); -out_end_trans: - gfs2_trans_end(sdp); -out_ipres: - gfs2_inplace_release(ip); -out_gunlock_q: - gfs2_quota_unlock(ip); -out: - gfs2_alloc_put(ip); - return error; -} - - -/** * gfs2_block_truncate_page - Deal with zeroing out data for truncate * * This is partly borrowed from ext3. */ -static int gfs2_block_truncate_page(struct address_space *mapping) +static int gfs2_block_truncate_page(struct address_space *mapping, loff_t from) { struct inode *inode = mapping->host; struct gfs2_inode *ip = GFS2_I(inode); - loff_t from = inode->i_size; unsigned long index = from >> PAGE_CACHE_SHIFT; unsigned offset = from & (PAGE_CACHE_SIZE-1); unsigned blocksize, iblock, length, pos; @@ -1023,9 +953,11 @@ unlock: return err; } -static int trunc_start(struct gfs2_inode *ip, u64 size) +static int trunc_start(struct inode *inode, u64 oldsize, u64 newsize) { - struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_sbd *sdp = GFS2_SB(inode); + struct address_space *mapping = inode->i_mapping; struct buffer_head *dibh; int journaled = gfs2_is_jdata(ip); int error; @@ -1039,31 +971,27 @@ static int trunc_start(struct gfs2_inode *ip, u64 size) if (error) goto out; + gfs2_trans_add_bh(ip->i_gl, dibh, 1); + if (gfs2_is_stuffed(ip)) { - u64 dsize = size + sizeof(struct gfs2_dinode); - ip->i_disksize = size; - ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; - gfs2_trans_add_bh(ip->i_gl, dibh, 1); - gfs2_dinode_out(ip, dibh->b_data); - if (dsize > dibh->b_size) - dsize = dibh->b_size; - gfs2_buffer_clear_tail(dibh, dsize); - error = 1; + gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode) + newsize); } else { - if (size & (u64)(sdp->sd_sb.sb_bsize - 1)) - error = gfs2_block_truncate_page(ip->i_inode.i_mapping); - - if (!error) { - ip->i_disksize = size; - ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; - ip->i_diskflags |= GFS2_DIF_TRUNC_IN_PROG; - gfs2_trans_add_bh(ip->i_gl, dibh, 1); - gfs2_dinode_out(ip, dibh->b_data); + if (newsize & (u64)(sdp->sd_sb.sb_bsize - 1)) { + error = gfs2_block_truncate_page(mapping, newsize); + if (error) + goto out_brelse; } + ip->i_diskflags |= GFS2_DIF_TRUNC_IN_PROG; } - brelse(dibh); + i_size_write(inode, newsize); + ip->i_disksize = newsize; + ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; + gfs2_dinode_out(ip, dibh->b_data); + truncate_pagecache(inode, oldsize, newsize); +out_brelse: + brelse(dibh); out: gfs2_trans_end(sdp); return error; @@ -1143,86 +1071,149 @@ out: /** * do_shrink - make a file smaller - * @ip: the inode - * @size: the size to make the file - * @truncator: function to truncate the last partial block + * @inode: the inode + * @oldsize: the current inode size + * @newsize: the size to make the file * - * Called with an exclusive lock on @ip. + * Called with an exclusive lock on @inode. The @size must + * be equal to or smaller than the current inode size. * * Returns: errno */ -static int do_shrink(struct gfs2_inode *ip, u64 size) +static int do_shrink(struct inode *inode, u64 oldsize, u64 newsize) { + struct gfs2_inode *ip = GFS2_I(inode); int error; - error = trunc_start(ip, size); + error = trunc_start(inode, oldsize, newsize); if (error < 0) return error; - if (error > 0) + if (gfs2_is_stuffed(ip)) return 0; - error = trunc_dealloc(ip, size); - if (!error) + error = trunc_dealloc(ip, newsize); + if (error == 0) error = trunc_end(ip); return error; } -static int do_touch(struct gfs2_inode *ip, u64 size) +void gfs2_trim_blocks(struct inode *inode) { - struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); + u64 size = inode->i_size; + int ret; + + ret = do_shrink(inode, size, size); + WARN_ON(ret != 0); +} + +/** + * do_grow - Touch and update inode size + * @inode: The inode + * @size: The new size + * + * This function updates the timestamps on the inode and + * may also increase the size of the inode. This function + * must not be called with @size any smaller than the current + * inode size. + * + * Although it is not strictly required to unstuff files here, + * earlier versions of GFS2 have a bug in the stuffed file reading + * code which will result in a buffer overrun if the size is larger + * than the max stuffed file size. In order to prevent this from + * occuring, such files are unstuffed, but in other cases we can + * just update the inode size directly. + * + * Returns: 0 on success, or -ve on error + */ + +static int do_grow(struct inode *inode, u64 size) +{ + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_sbd *sdp = GFS2_SB(inode); struct buffer_head *dibh; + struct gfs2_alloc *al = NULL; int error; - error = gfs2_trans_begin(sdp, RES_DINODE, 0); + if (gfs2_is_stuffed(ip) && + (size > (sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)))) { + al = gfs2_alloc_get(ip); + if (al == NULL) + return -ENOMEM; + + error = gfs2_quota_lock_check(ip); + if (error) + goto do_grow_alloc_put; + + al->al_requested = 1; + error = gfs2_inplace_reserve(ip); + if (error) + goto do_grow_qunlock; + } + + error = gfs2_trans_begin(sdp, RES_DINODE + 1, 0); if (error) - return error; + goto do_grow_release; - down_write(&ip->i_rw_mutex); + if (al) { + error = gfs2_unstuff_dinode(ip, NULL); + if (error) + goto do_end_trans; + } error = gfs2_meta_inode_buffer(ip, &dibh); if (error) - goto do_touch_out; + goto do_end_trans; + i_size_write(inode, size); + ip->i_disksize = size; ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; gfs2_trans_add_bh(ip->i_gl, dibh, 1); gfs2_dinode_out(ip, dibh->b_data); brelse(dibh); -do_touch_out: - up_write(&ip->i_rw_mutex); +do_end_trans: gfs2_trans_end(sdp); +do_grow_release: + if (al) { + gfs2_inplace_release(ip); +do_grow_qunlock: + gfs2_quota_unlock(ip); +do_grow_alloc_put: + gfs2_alloc_put(ip); + } return error; } /** - * gfs2_truncatei - make a file a given size - * @ip: the inode - * @size: the size to make the file - * @truncator: function to truncate the last partial block + * gfs2_setattr_size - make a file a given size + * @inode: the inode + * @newsize: the size to make the file * - * The file size can grow, shrink, or stay the same size. + * The file size can grow, shrink, or stay the same size. This + * is called holding i_mutex and an exclusive glock on the inode + * in question. * * Returns: errno */ -int gfs2_truncatei(struct gfs2_inode *ip, u64 size) +int gfs2_setattr_size(struct inode *inode, u64 newsize) { - int error; + int ret; + u64 oldsize; - if (gfs2_assert_warn(GFS2_SB(&ip->i_inode), S_ISREG(ip->i_inode.i_mode))) - return -EINVAL; + BUG_ON(!S_ISREG(inode->i_mode)); - if (size > ip->i_disksize) - error = do_grow(ip, size); - else if (size < ip->i_disksize) - error = do_shrink(ip, size); - else - /* update time stamps */ - error = do_touch(ip, size); + ret = inode_newsize_ok(inode, newsize); + if (ret) + return ret; - return error; + oldsize = inode->i_size; + if (newsize >= oldsize) + return do_grow(inode, newsize); + + return do_shrink(inode, oldsize, newsize); } int gfs2_truncatei_resume(struct gfs2_inode *ip) diff --git a/fs/gfs2/bmap.h b/fs/gfs2/bmap.h index a20a521..42fea03 100644 --- a/fs/gfs2/bmap.h +++ b/fs/gfs2/bmap.h @@ -44,14 +44,16 @@ static inline void gfs2_write_calc_reserv(const struct gfs2_inode *ip, } } -int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page); -int gfs2_block_map(struct inode *inode, sector_t lblock, struct buffer_head *bh, int create); -int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, u64 *dblock, unsigned *extlen); - -int gfs2_truncatei(struct gfs2_inode *ip, u64 size); -int gfs2_truncatei_resume(struct gfs2_inode *ip); -int gfs2_file_dealloc(struct gfs2_inode *ip); -int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset, - unsigned int len); +extern int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page); +extern int gfs2_block_map(struct inode *inode, sector_t lblock, + struct buffer_head *bh, int create); +extern int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, + u64 *dblock, unsigned *extlen); +extern int gfs2_setattr_size(struct inode *inode, u64 size); +extern void gfs2_trim_blocks(struct inode *inode); +extern int gfs2_truncatei_resume(struct gfs2_inode *ip); +extern int gfs2_file_dealloc(struct gfs2_inode *ip); +extern int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset, + unsigned int len); #endif /* __BMAP_DOT_H__ */ diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 1009be2..1d3f2fb 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -1071,30 +1071,6 @@ int gfs2_permission(struct inode *inode, int mask) return error; } -/* - * XXX(truncate): the truncate_setsize calls should be moved to the end. - */ -static int setattr_size(struct inode *inode, struct iattr *attr) -{ - struct gfs2_inode *ip = GFS2_I(inode); - struct gfs2_sbd *sdp = GFS2_SB(inode); - int error; - - if (attr->ia_size != ip->i_disksize) { - error = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks); - if (error) - return error; - truncate_setsize(inode, attr->ia_size); - gfs2_trans_end(sdp); - } - - error = gfs2_truncatei(ip, attr->ia_size); - if (error && (inode->i_size != ip->i_disksize)) - i_size_write(inode, ip->i_disksize); - - return error; -} - static int setattr_chown(struct inode *inode, struct iattr *attr) { struct gfs2_inode *ip = GFS2_I(inode); @@ -1195,7 +1171,7 @@ static int gfs2_setattr(struct dentry *dentry, struct iattr *attr) goto out; if (attr->ia_valid & ATTR_SIZE) - error = setattr_size(inode, attr); + error = gfs2_setattr_size(inode, attr->ia_size); else if (attr->ia_valid & (ATTR_UID | ATTR_GID)) error = setattr_chown(inode, attr); else if ((attr->ia_valid & ATTR_MODE) && IS_POSIXACL(inode)) -- cgit v0.10.2 From a2e0f79939e09e74698564b88dee709db208e1e2 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 11 Aug 2010 09:53:11 +0100 Subject: GFS2: Remove i_disksize With the update of the truncate code, ip->i_disksize and inode->i_size are merely copies of each other. This means we can remove ip->i_disksize and use inode->i_size exclusively reducing the size of a GFS2 inode by 8 bytes. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index f687f25..c92f36b 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -800,10 +800,8 @@ static int gfs2_stuffed_write_end(struct inode *inode, struct buffer_head *dibh, page_cache_release(page); if (copied) { - if (inode->i_size < to) { + if (inode->i_size < to) i_size_write(inode, to); - ip->i_disksize = inode->i_size; - } gfs2_dinode_out(ip, di); mark_inode_dirty(inode); } @@ -874,8 +872,6 @@ static int gfs2_write_end(struct file *file, struct address_space *mapping, ret = generic_write_end(file, mapping, pos, len, copied, page, fsdata); if (ret > 0) { - if (inode->i_size > ip->i_disksize) - ip->i_disksize = inode->i_size; gfs2_dinode_out(ip, dibh->b_data); mark_inode_dirty(inode); } diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 20b971a..04513e9 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -131,7 +131,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) if (error) goto out; - if (ip->i_disksize) { + if (i_size_read(&ip->i_inode)) { /* Get a free block, fill it with the stuffed data, and write it out to disk */ @@ -160,7 +160,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) di = (struct gfs2_dinode *)dibh->b_data; gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); - if (ip->i_disksize) { + if (i_size_read(&ip->i_inode)) { *(__be64 *)(di + 1) = cpu_to_be64(block); gfs2_add_inode_blocks(&ip->i_inode, 1); di->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode)); @@ -985,7 +985,6 @@ static int trunc_start(struct inode *inode, u64 oldsize, u64 newsize) } i_size_write(inode, newsize); - ip->i_disksize = newsize; ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; gfs2_dinode_out(ip, dibh->b_data); @@ -1051,7 +1050,7 @@ static int trunc_end(struct gfs2_inode *ip) if (error) goto out; - if (!ip->i_disksize) { + if (!i_size_read(&ip->i_inode)) { ip->i_height = 0; ip->i_goal = ip->i_no_addr; gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); @@ -1167,7 +1166,6 @@ static int do_grow(struct inode *inode, u64 size) goto do_end_trans; i_size_write(inode, size); - ip->i_disksize = size; ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; gfs2_trans_add_bh(ip->i_gl, dibh, 1); gfs2_dinode_out(ip, dibh->b_data); @@ -1219,7 +1217,7 @@ int gfs2_setattr_size(struct inode *inode, u64 newsize) int gfs2_truncatei_resume(struct gfs2_inode *ip) { int error; - error = trunc_dealloc(ip, ip->i_disksize); + error = trunc_dealloc(ip, i_size_read(&ip->i_inode)); if (!error) error = trunc_end(ip); return error; @@ -1260,7 +1258,7 @@ int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset, shift = sdp->sd_sb.sb_bsize_shift; BUG_ON(gfs2_is_dir(ip)); - end_of_file = (ip->i_disksize + sdp->sd_sb.sb_bsize - 1) >> shift; + end_of_file = (i_size_read(&ip->i_inode) + sdp->sd_sb.sb_bsize - 1) >> shift; lblock = offset >> shift; lblock_stop = (offset + len + sdp->sd_sb.sb_bsize - 1) >> shift; if (lblock_stop > end_of_file) diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index b9dd88a..c1042ae 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -127,8 +127,8 @@ static int gfs2_dir_write_stuffed(struct gfs2_inode *ip, const char *buf, gfs2_trans_add_bh(ip->i_gl, dibh, 1); memcpy(dibh->b_data + offset + sizeof(struct gfs2_dinode), buf, size); - if (ip->i_disksize < offset + size) - ip->i_disksize = offset + size; + if (ip->i_inode.i_size < offset + size) + i_size_write(&ip->i_inode, offset + size); ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; gfs2_dinode_out(ip, dibh->b_data); @@ -225,8 +225,8 @@ out: if (error) return error; - if (ip->i_disksize < offset + copied) - ip->i_disksize = offset + copied; + if (ip->i_inode.i_size < offset + copied) + i_size_write(&ip->i_inode, offset + copied); ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; gfs2_trans_add_bh(ip->i_gl, dibh, 1); @@ -275,12 +275,13 @@ static int gfs2_dir_read_data(struct gfs2_inode *ip, char *buf, u64 offset, unsigned int o; int copied = 0; int error = 0; + u64 disksize = i_size_read(&ip->i_inode); - if (offset >= ip->i_disksize) + if (offset >= disksize) return 0; - if (offset + size > ip->i_disksize) - size = ip->i_disksize - offset; + if (offset + size > disksize) + size = disksize - offset; if (!size) return 0; @@ -727,7 +728,7 @@ static struct gfs2_dirent *gfs2_dirent_search(struct inode *inode, unsigned hsize = 1 << ip->i_depth; unsigned index; u64 ln; - if (hsize * sizeof(u64) != ip->i_disksize) { + if (hsize * sizeof(u64) != i_size_read(inode)) { gfs2_consist_inode(ip); return ERR_PTR(-EIO); } @@ -879,7 +880,7 @@ static int dir_make_exhash(struct inode *inode) for (x = sdp->sd_hash_ptrs; x--; lp++) *lp = cpu_to_be64(bn); - dip->i_disksize = sdp->sd_sb.sb_bsize / 2; + i_size_write(inode, sdp->sd_sb.sb_bsize / 2); gfs2_add_inode_blocks(&dip->i_inode, 1); dip->i_diskflags |= GFS2_DIF_EXHASH; @@ -1057,11 +1058,12 @@ static int dir_double_exhash(struct gfs2_inode *dip) u64 *buf; u64 *from, *to; u64 block; + u64 disksize = i_size_read(&dip->i_inode); int x; int error = 0; hsize = 1 << dip->i_depth; - if (hsize * sizeof(u64) != dip->i_disksize) { + if (hsize * sizeof(u64) != disksize) { gfs2_consist_inode(dip); return -EIO; } @@ -1072,7 +1074,7 @@ static int dir_double_exhash(struct gfs2_inode *dip) if (!buf) return -ENOMEM; - for (block = dip->i_disksize >> sdp->sd_hash_bsize_shift; block--;) { + for (block = disksize >> sdp->sd_hash_bsize_shift; block--;) { error = gfs2_dir_read_data(dip, (char *)buf, block * sdp->sd_hash_bsize, sdp->sd_hash_bsize, 1); @@ -1370,7 +1372,7 @@ static int dir_e_read(struct inode *inode, u64 *offset, void *opaque, unsigned depth = 0; hsize = 1 << dip->i_depth; - if (hsize * sizeof(u64) != dip->i_disksize) { + if (hsize * sizeof(u64) != i_size_read(inode)) { gfs2_consist_inode(dip); return -EIO; } @@ -1784,7 +1786,7 @@ static int foreach_leaf(struct gfs2_inode *dip, leaf_call_t lc, void *data) int error = 0; hsize = 1 << dip->i_depth; - if (hsize * sizeof(u64) != dip->i_disksize) { + if (hsize * sizeof(u64) != i_size_read(&dip->i_inode)) { gfs2_consist_inode(dip); return -EIO; } diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index 4edd662..daadcd2 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -491,7 +491,7 @@ static int gfs2_open(struct inode *inode, struct file *file) goto fail; if (!(file->f_flags & O_LARGEFILE) && - ip->i_disksize > MAX_NON_LFS) { + i_size_read(inode) > MAX_NON_LFS) { error = -EOVERFLOW; goto fail_gunlock; } diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c index 49f97d3..621d80e 100644 --- a/fs/gfs2/glops.c +++ b/fs/gfs2/glops.c @@ -262,13 +262,12 @@ static int inode_go_dump(struct seq_file *seq, const struct gfs2_glock *gl) const struct gfs2_inode *ip = gl->gl_object; if (ip == NULL) return 0; - gfs2_print_dbg(seq, " I: n:%llu/%llu t:%u f:0x%02lx d:0x%08x s:%llu/%llu\n", + gfs2_print_dbg(seq, " I: n:%llu/%llu t:%u f:0x%02lx d:0x%08x s:%llu\n", (unsigned long long)ip->i_no_formal_ino, (unsigned long long)ip->i_no_addr, IF2DT(ip->i_inode.i_mode), ip->i_flags, (unsigned int)ip->i_diskflags, - (unsigned long long)ip->i_inode.i_size, - (unsigned long long)ip->i_disksize); + (unsigned long long)i_size_read(&ip->i_inode)); return 0; } diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index fdbf4b3..c119717 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -267,7 +267,6 @@ struct gfs2_inode { u64 i_no_formal_ino; u64 i_generation; u64 i_eattr; - loff_t i_disksize; unsigned long i_flags; /* GIF_... */ struct gfs2_glock *i_gl; /* Move into i_gh? */ struct gfs2_holder i_iopen_gh; diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 08140f1..06370f8 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -359,8 +359,7 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) * to do that. */ ip->i_inode.i_nlink = be32_to_cpu(str->di_nlink); - ip->i_disksize = be64_to_cpu(str->di_size); - i_size_write(&ip->i_inode, ip->i_disksize); + i_size_write(&ip->i_inode, be64_to_cpu(str->di_size)); gfs2_set_inode_blocks(&ip->i_inode, be64_to_cpu(str->di_blocks)); atime.tv_sec = be64_to_cpu(str->di_atime); atime.tv_nsec = be32_to_cpu(str->di_atime_nsec); @@ -1055,7 +1054,7 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf) str->di_uid = cpu_to_be32(ip->i_inode.i_uid); str->di_gid = cpu_to_be32(ip->i_inode.i_gid); str->di_nlink = cpu_to_be32(ip->i_inode.i_nlink); - str->di_size = cpu_to_be64(ip->i_disksize); + str->di_size = cpu_to_be64(i_size_read(&ip->i_inode)); str->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode)); str->di_atime = cpu_to_be64(ip->i_inode.i_atime.tv_sec); str->di_mtime = cpu_to_be64(ip->i_inode.i_mtime.tv_sec); @@ -1085,8 +1084,8 @@ void gfs2_dinode_print(const struct gfs2_inode *ip) (unsigned long long)ip->i_no_formal_ino); printk(KERN_INFO " no_addr = %llu\n", (unsigned long long)ip->i_no_addr); - printk(KERN_INFO " i_disksize = %llu\n", - (unsigned long long)ip->i_disksize); + printk(KERN_INFO " i_size = %llu\n", + (unsigned long long)i_size_read(&ip->i_inode)); printk(KERN_INFO " blocks = %llu\n", (unsigned long long)gfs2_get_inode_blocks(&ip->i_inode)); printk(KERN_INFO " i_goal = %llu\n", diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h index 300ada3..15ff4df 100644 --- a/fs/gfs2/inode.h +++ b/fs/gfs2/inode.h @@ -80,6 +80,19 @@ static inline void gfs2_inum_out(const struct gfs2_inode *ip, dent->de_inum.no_addr = cpu_to_be64(ip->i_no_addr); } +static inline int gfs2_check_internal_file_size(struct inode *inode, + u64 minsize, u64 maxsize) +{ + u64 size = i_size_read(inode); + if (size < minsize || size > maxsize) + goto err; + if (size & ((1 << inode->i_blkbits) - 1)) + goto err; + return 0; +err: + gfs2_consist_inode(GFS2_I(inode)); + return -EIO; +} extern void gfs2_set_iop(struct inode *inode); extern struct inode *gfs2_inode_lookup(struct super_block *sb, unsigned type, diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 4d4b1e8..5b5c87d 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -586,7 +586,7 @@ static int map_journal_extents(struct gfs2_sbd *sdp) prev_db = 0; - for (lb = 0; lb < ip->i_disksize >> sdp->sd_sb.sb_bsize_shift; lb++) { + for (lb = 0; lb < i_size_read(jd->jd_inode) >> sdp->sd_sb.sb_bsize_shift; lb++) { bh.b_state = 0; bh.b_blocknr = 0; bh.b_size = 1 << ip->i_inode.i_blkbits; diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 1d3f2fb..ee6ffd5 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -406,7 +406,6 @@ static int gfs2_symlink(struct inode *dir, struct dentry *dentry, ip = ghs[1].gh_gl->gl_object; - ip->i_disksize = size; i_size_write(inode, size); error = gfs2_meta_inode_buffer(ip, &dibh); @@ -461,7 +460,7 @@ static int gfs2_mkdir(struct inode *dir, struct dentry *dentry, int mode) ip = ghs[1].gh_gl->gl_object; ip->i_inode.i_nlink = 2; - ip->i_disksize = sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode); + i_size_write(inode, sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)); ip->i_diskflags |= GFS2_DIF_JDATA; ip->i_entries = 2; @@ -990,7 +989,7 @@ static void *gfs2_follow_link(struct dentry *dentry, struct nameidata *nd) struct gfs2_inode *ip = GFS2_I(dentry->d_inode); struct gfs2_holder i_gh; struct buffer_head *dibh; - unsigned int x; + unsigned int x, size; char *buf; int error; @@ -1002,7 +1001,8 @@ static void *gfs2_follow_link(struct dentry *dentry, struct nameidata *nd) return NULL; } - if (!ip->i_disksize) { + size = (unsigned int)i_size_read(&ip->i_inode); + if (size == 0) { gfs2_consist_inode(ip); buf = ERR_PTR(-EIO); goto out; @@ -1014,7 +1014,7 @@ static void *gfs2_follow_link(struct dentry *dentry, struct nameidata *nd) goto out; } - x = ip->i_disksize + 1; + x = size + 1; buf = kmalloc(x, GFP_NOFS); if (!buf) buf = ERR_PTR(-ENOMEM); diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 1bc6b56..9bc6dd9 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -735,10 +735,8 @@ get_a_page: goto out; size = loc + sizeof(struct gfs2_quota); - if (size > inode->i_size) { - ip->i_disksize = size; + if (size > inode->i_size) i_size_write(inode, size); - } inode->i_mtime = inode->i_atime = CURRENT_TIME; gfs2_trans_add_bh(ip->i_gl, dibh, 1); gfs2_dinode_out(ip, dibh->b_data); @@ -1190,18 +1188,17 @@ static void gfs2_quota_change_in(struct gfs2_quota_change_host *qc, const void * int gfs2_quota_init(struct gfs2_sbd *sdp) { struct gfs2_inode *ip = GFS2_I(sdp->sd_qc_inode); - unsigned int blocks = ip->i_disksize >> sdp->sd_sb.sb_bsize_shift; + u64 size = i_size_read(sdp->sd_qc_inode); + unsigned int blocks = size >> sdp->sd_sb.sb_bsize_shift; unsigned int x, slot = 0; unsigned int found = 0; u64 dblock; u32 extlen = 0; int error; - if (!ip->i_disksize || ip->i_disksize > (64 << 20) || - ip->i_disksize & (sdp->sd_sb.sb_bsize - 1)) { - gfs2_consist_inode(ip); + if (gfs2_check_internal_file_size(sdp->sd_qc_inode, 1, 64 << 20)) return -EIO; - } + sdp->sd_quota_slots = blocks * sdp->sd_qc_per_block; sdp->sd_quota_chunks = DIV_ROUND_UP(sdp->sd_quota_slots, 8 * PAGE_SIZE); diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 171a744..370c29b 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -500,7 +500,7 @@ u64 gfs2_ri_total(struct gfs2_sbd *sdp) for (rgrps = 0;; rgrps++) { loff_t pos = rgrps * sizeof(struct gfs2_rindex); - if (pos + sizeof(struct gfs2_rindex) >= ip->i_disksize) + if (pos + sizeof(struct gfs2_rindex) >= i_size_read(inode)) break; error = gfs2_internal_read(ip, &ra_state, buf, &pos, sizeof(struct gfs2_rindex)); @@ -588,7 +588,7 @@ static int gfs2_ri_update(struct gfs2_inode *ip) struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct inode *inode = &ip->i_inode; struct file_ra_state ra_state; - u64 rgrp_count = ip->i_disksize; + u64 rgrp_count = i_size_read(inode); int error; do_div(rgrp_count, sizeof(struct gfs2_rindex)); @@ -628,7 +628,7 @@ static int gfs2_ri_update_special(struct gfs2_inode *ip) for (sdp->sd_rgrps = 0;; sdp->sd_rgrps++) { /* Ignore partials */ if ((sdp->sd_rgrps + 1) * sizeof(struct gfs2_rindex) > - ip->i_disksize) + i_size_read(inode)) break; error = read_rindex_entry(ip, &ra_state); if (error) { diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index 77cb9f8..e031fa4 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -342,15 +342,14 @@ int gfs2_jdesc_check(struct gfs2_jdesc *jd) { struct gfs2_inode *ip = GFS2_I(jd->jd_inode); struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode); + u64 size = i_size_read(jd->jd_inode); - if (ip->i_disksize < (8 << 20) || ip->i_disksize > (1 << 30) || - (ip->i_disksize & (sdp->sd_sb.sb_bsize - 1))) { - gfs2_consist_inode(ip); + if (gfs2_check_internal_file_size(jd->jd_inode, 8 << 20, 1 << 30)) return -EIO; - } - jd->jd_blocks = ip->i_disksize >> sdp->sd_sb.sb_bsize_shift; - if (gfs2_write_alloc_required(ip, 0, ip->i_disksize)) { + jd->jd_blocks = size >> sdp->sd_sb.sb_bsize_shift; + + if (gfs2_write_alloc_required(ip, 0, size)) { gfs2_consist_inode(ip); return -EIO; } -- cgit v0.10.2 From 820969f353587281d645735c83c7f07d606e67ba Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 11 Aug 2010 09:53:47 +0100 Subject: GFS2: No longer experimental I think the time has arrvied to remove the experimental tag from GFS2. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/Kconfig b/fs/gfs2/Kconfig index cc96655..c465ae0 100644 --- a/fs/gfs2/Kconfig +++ b/fs/gfs2/Kconfig @@ -1,6 +1,6 @@ config GFS2_FS tristate "GFS2 file system support" - depends on EXPERIMENTAL && (64BIT || LBDAF) + depends on (64BIT || LBDAF) select DLM if GFS2_FS_LOCKING_DLM select CONFIGFS_FS if GFS2_FS_LOCKING_DLM select SYSFS if GFS2_FS_LOCKING_DLM -- cgit v0.10.2 From 9a3f236d40a99ea8dca3df40d8ef67631057cad6 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 23 Aug 2010 11:49:34 +0100 Subject: GFS2: Add a bug trap in allocation code This adds a check to ensure that if we reach the block allocator that we don't try and proceed if there is no alloc structure hanging off the inode. This should only happen if there is a bug in GFS2. The error return code is distinctive in order that it will be easily spotted. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 370c29b..66b6d4d 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1496,11 +1496,19 @@ int gfs2_alloc_block(struct gfs2_inode *ip, u64 *bn, unsigned int *n) struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct buffer_head *dibh; struct gfs2_alloc *al = ip->i_alloc; - struct gfs2_rgrpd *rgd = al->al_rgd; + struct gfs2_rgrpd *rgd; u32 goal, blk; u64 block; int error; + /* Only happens if there is a bug in gfs2, return something distinctive + * to ensure that it is noticed. + */ + if (al == NULL) + return -ECANCELED; + + rgd = al->al_rgd; + if (rgrp_contains_block(rgd, ip->i_goal)) goal = ip->i_goal - rgd->rd_data0; else -- cgit v0.10.2 From 3921120e757f9167f3fcd3a1781239824471b14d Mon Sep 17 00:00:00 2001 From: Benjamin Marzinski Date: Fri, 20 Aug 2010 00:21:02 -0500 Subject: GFS2: fallocate support This patch adds support for fallocate to gfs2. Since the gfs2 does not support uninitialized data blocks, it must write out zeros to all the blocks. However, since it does not need to lock any pages to read from, gfs2 can write out the zero blocks much more efficiently. On a moderately full filesystem, fallocate works around 5 times faster on average. The fallocate call also allows gfs2 to add blocks to the file without changing the filesize, which will make it possible for gfs2 to preallocate space for the rindex file, so that gfs2 can grow a completely full filesystem. Signed-off-by: Benjamin Marzinski Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index c92f36b..180ef8a 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -36,8 +36,8 @@ #include "glops.h" -static void gfs2_page_add_databufs(struct gfs2_inode *ip, struct page *page, - unsigned int from, unsigned int to) +void gfs2_page_add_databufs(struct gfs2_inode *ip, struct page *page, + unsigned int from, unsigned int to) { struct buffer_head *head = page_buffers(page); unsigned int bsize = head->b_size; diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index c119717..578234b 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -571,6 +571,7 @@ struct gfs2_sbd { struct list_head sd_rindex_mru_list; struct gfs2_rgrpd *sd_rindex_forward; unsigned int sd_rgrps; + unsigned int sd_max_rg_data; /* Journal index stuff */ diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h index 15ff4df..6720d7d 100644 --- a/fs/gfs2/inode.h +++ b/fs/gfs2/inode.h @@ -19,6 +19,8 @@ extern int gfs2_releasepage(struct page *page, gfp_t gfp_mask); extern int gfs2_internal_read(struct gfs2_inode *ip, struct file_ra_state *ra_state, char *buf, loff_t *pos, unsigned size); +extern void gfs2_page_add_databufs(struct gfs2_inode *ip, struct page *page, + unsigned int from, unsigned int to); extern void gfs2_set_aops(struct inode *inode); static inline int gfs2_is_stuffed(const struct gfs2_inode *ip) diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index ee6ffd5..f6da0d7 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -18,6 +18,8 @@ #include #include #include +#include +#include #include #include "gfs2.h" @@ -1277,6 +1279,257 @@ static int gfs2_removexattr(struct dentry *dentry, const char *name) return ret; } +static void empty_write_end(struct page *page, unsigned from, + unsigned to) +{ + struct gfs2_inode *ip = GFS2_I(page->mapping->host); + + page_zero_new_buffers(page, from, to); + flush_dcache_page(page); + mark_page_accessed(page); + + if (!gfs2_is_writeback(ip)) + gfs2_page_add_databufs(ip, page, from, to); + + block_commit_write(page, from, to); +} + + +static int write_empty_blocks(struct page *page, unsigned from, unsigned to) +{ + unsigned start, end, next; + struct buffer_head *bh, *head; + int error; + + if (!page_has_buffers(page)) { + error = block_prepare_write(page, from, to, gfs2_block_map); + if (unlikely(error)) + return error; + + empty_write_end(page, from, to); + return 0; + } + + bh = head = page_buffers(page); + next = end = 0; + while (next < from) { + next += bh->b_size; + bh = bh->b_this_page; + } + start = next; + do { + next += bh->b_size; + if (buffer_mapped(bh)) { + if (end) { + error = block_prepare_write(page, start, end, + gfs2_block_map); + if (unlikely(error)) + return error; + empty_write_end(page, start, end); + end = 0; + } + start = next; + } + else + end = next; + bh = bh->b_this_page; + } while (next < to); + + if (end) { + error = block_prepare_write(page, start, end, gfs2_block_map); + if (unlikely(error)) + return error; + empty_write_end(page, start, end); + } + + return 0; +} + +static int fallocate_chunk(struct inode *inode, loff_t offset, loff_t len, + int mode) +{ + struct gfs2_inode *ip = GFS2_I(inode); + struct buffer_head *dibh; + int error; + u64 start = offset >> PAGE_CACHE_SHIFT; + unsigned int start_offset = offset & ~PAGE_CACHE_MASK; + u64 end = (offset + len - 1) >> PAGE_CACHE_SHIFT; + pgoff_t curr; + struct page *page; + unsigned int end_offset = (offset + len) & ~PAGE_CACHE_MASK; + unsigned int from, to; + + if (!end_offset) + end_offset = PAGE_CACHE_SIZE; + + error = gfs2_meta_inode_buffer(ip, &dibh); + if (unlikely(error)) + goto out; + + gfs2_trans_add_bh(ip->i_gl, dibh, 1); + + if (gfs2_is_stuffed(ip)) { + error = gfs2_unstuff_dinode(ip, NULL); + if (unlikely(error)) + goto out; + } + + curr = start; + offset = start << PAGE_CACHE_SHIFT; + from = start_offset; + to = PAGE_CACHE_SIZE; + while (curr <= end) { + page = grab_cache_page_write_begin(inode->i_mapping, curr, + AOP_FLAG_NOFS); + if (unlikely(!page)) { + error = -ENOMEM; + goto out; + } + + if (curr == end) + to = end_offset; + error = write_empty_blocks(page, from, to); + if (!error && offset + to > inode->i_size && + !(mode & FALLOC_FL_KEEP_SIZE)) { + i_size_write(inode, offset + to); + } + unlock_page(page); + page_cache_release(page); + if (error) + goto out; + curr++; + offset += PAGE_CACHE_SIZE; + from = 0; + } + + gfs2_dinode_out(ip, dibh->b_data); + mark_inode_dirty(inode); + + brelse(dibh); + +out: + return error; +} + +static void calc_max_reserv(struct gfs2_inode *ip, loff_t max, loff_t *len, + unsigned int *data_blocks, unsigned int *ind_blocks) +{ + const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); + unsigned int max_blocks = ip->i_alloc->al_rgd->rd_free_clone; + unsigned int tmp, max_data = max_blocks - 3 * (sdp->sd_max_height - 1); + + for (tmp = max_data; tmp > sdp->sd_diptrs;) { + tmp = DIV_ROUND_UP(tmp, sdp->sd_inptrs); + max_data -= tmp; + } + /* This calculation isn't the exact reverse of gfs2_write_calc_reserve, + so it might end up with fewer data blocks */ + if (max_data <= *data_blocks) + return; + *data_blocks = max_data; + *ind_blocks = max_blocks - max_data; + *len = ((loff_t)max_data - 3) << sdp->sd_sb.sb_bsize_shift; + if (*len > max) { + *len = max; + gfs2_write_calc_reserv(ip, max, data_blocks, ind_blocks); + } +} + +static long gfs2_fallocate(struct inode *inode, int mode, loff_t offset, + loff_t len) +{ + struct gfs2_sbd *sdp = GFS2_SB(inode); + struct gfs2_inode *ip = GFS2_I(inode); + unsigned int data_blocks = 0, ind_blocks = 0, rblocks; + loff_t bytes, max_bytes; + struct gfs2_alloc *al; + int error; + loff_t next = (offset + len - 1) >> sdp->sd_sb.sb_bsize_shift; + next = (next + 1) << sdp->sd_sb.sb_bsize_shift; + + offset = (offset >> sdp->sd_sb.sb_bsize_shift) << + sdp->sd_sb.sb_bsize_shift; + + len = next - offset; + bytes = sdp->sd_max_rg_data * sdp->sd_sb.sb_bsize / 2; + if (!bytes) + bytes = UINT_MAX; + + gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &ip->i_gh); + error = gfs2_glock_nq(&ip->i_gh); + if (unlikely(error)) + goto out_uninit; + + if (!gfs2_write_alloc_required(ip, offset, len)) + goto out_unlock; + + while (len > 0) { + if (len < bytes) + bytes = len; + al = gfs2_alloc_get(ip); + if (!al) { + error = -ENOMEM; + goto out_unlock; + } + + error = gfs2_quota_lock_check(ip); + if (error) + goto out_alloc_put; + +retry: + gfs2_write_calc_reserv(ip, bytes, &data_blocks, &ind_blocks); + + al->al_requested = data_blocks + ind_blocks; + error = gfs2_inplace_reserve(ip); + if (error) { + if (error == -ENOSPC && bytes > sdp->sd_sb.sb_bsize) { + bytes >>= 1; + goto retry; + } + goto out_qunlock; + } + max_bytes = bytes; + calc_max_reserv(ip, len, &max_bytes, &data_blocks, &ind_blocks); + al->al_requested = data_blocks + ind_blocks; + + rblocks = RES_DINODE + ind_blocks + RES_STATFS + RES_QUOTA + + RES_RG_HDR + ip->i_alloc->al_rgd->rd_length; + if (gfs2_is_jdata(ip)) + rblocks += data_blocks ? data_blocks : 1; + + error = gfs2_trans_begin(sdp, rblocks, + PAGE_CACHE_SIZE/sdp->sd_sb.sb_bsize); + if (error) + goto out_trans_fail; + + error = fallocate_chunk(inode, offset, max_bytes, mode); + gfs2_trans_end(sdp); + + if (error) + goto out_trans_fail; + + len -= max_bytes; + offset += max_bytes; + gfs2_inplace_release(ip); + gfs2_quota_unlock(ip); + gfs2_alloc_put(ip); + } + goto out_unlock; + +out_trans_fail: + gfs2_inplace_release(ip); +out_qunlock: + gfs2_quota_unlock(ip); +out_alloc_put: + gfs2_alloc_put(ip); +out_unlock: + gfs2_glock_dq(&ip->i_gh); +out_uninit: + gfs2_holder_uninit(&ip->i_gh); + return error; +} + + static int gfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 start, u64 len) { @@ -1327,6 +1580,7 @@ const struct inode_operations gfs2_file_iops = { .getxattr = gfs2_getxattr, .listxattr = gfs2_listxattr, .removexattr = gfs2_removexattr, + .fallocate = gfs2_fallocate, .fiemap = gfs2_fiemap, }; diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 66b6d4d..f9ddcf4 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -589,6 +589,8 @@ static int gfs2_ri_update(struct gfs2_inode *ip) struct inode *inode = &ip->i_inode; struct file_ra_state ra_state; u64 rgrp_count = i_size_read(inode); + struct gfs2_rgrpd *rgd; + unsigned int max_data = 0; int error; do_div(rgrp_count, sizeof(struct gfs2_rindex)); @@ -603,6 +605,10 @@ static int gfs2_ri_update(struct gfs2_inode *ip) } } + list_for_each_entry(rgd, &sdp->sd_rindex_list, rd_list) + if (rgd->rd_data > max_data) + max_data = rgd->rd_data; + sdp->sd_max_rg_data = max_data; sdp->sd_rindex_uptodate = 1; return 0; } @@ -622,6 +628,8 @@ static int gfs2_ri_update_special(struct gfs2_inode *ip) struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct inode *inode = &ip->i_inode; struct file_ra_state ra_state; + struct gfs2_rgrpd *rgd; + unsigned int max_data = 0; int error; file_ra_state_init(&ra_state, inode->i_mapping); @@ -636,6 +644,10 @@ static int gfs2_ri_update_special(struct gfs2_inode *ip) return error; } } + list_for_each_entry(rgd, &sdp->sd_rindex_list, rd_list) + if (rgd->rd_data > max_data) + max_data = rgd->rd_data; + sdp->sd_max_rg_data = max_data; sdp->sd_rindex_uptodate = 1; return 0; diff --git a/fs/gfs2/trans.h b/fs/gfs2/trans.h index edf9d4b..b849eb7 100644 --- a/fs/gfs2/trans.h +++ b/fs/gfs2/trans.h @@ -20,6 +20,7 @@ struct gfs2_glock; #define RES_JDATA 1 #define RES_DATA 1 #define RES_LEAF 1 +#define RES_RG_HDR 1 #define RES_RG_BIT 2 #define RES_EATTR 1 #define RES_STATFS 1 -- cgit v0.10.2 From fe08d5a89726675a920b0e9bbbe849c46b27a6e5 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 23 Aug 2010 11:54:45 +0100 Subject: GFS2: Fix whitespace in previous patch Removes the offending space Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index f6da0d7..ce4f1df 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -1423,7 +1423,7 @@ static void calc_max_reserv(struct gfs2_inode *ip, loff_t max, loff_t *len, max_data -= tmp; } /* This calculation isn't the exact reverse of gfs2_write_calc_reserve, - so it might end up with fewer data blocks */ + so it might end up with fewer data blocks */ if (max_data <= *data_blocks) return; *data_blocks = max_data; -- cgit v0.10.2 From 7b5e3d5fcf0d6fce66050bd0313a7dc2ae4abc62 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 3 Sep 2010 09:39:20 +0100 Subject: GFS2: Don't enforce min hold time when two demotes occur in rapid succession Due to the design of the VFS, it is quite usual for operations on GFS2 to consist of a lookup (requiring a shared lock) followed by an operation requiring an exclusive lock. If a remote node has cached an exclusive lock, then it will receive two demote events in rapid succession firstly for a shared lock and then to unlocked. The existing min hold time code was triggering in this case, even if the node was otherwise idle since the state change time was being updated by the initial demote. This patch introduces logic to skip the min hold timer in the case that a "double demote" of this kind has occurred. The min hold timer will still be used in all other cases. A new glock flag is introduced which is used to keep track of whether there have been any newly queued holders since the last glock state change. The min hold time is only applied if the flag is set. Signed-off-by: Steven Whitehouse Tested-by: Abhijith Das diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 9adf8f9..8e478e2 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -441,6 +441,8 @@ static void state_change(struct gfs2_glock *gl, unsigned int new_state) else gfs2_glock_put_nolock(gl); } + if (held1 && held2 && list_empty(&gl->gl_holders)) + clear_bit(GLF_QUEUED, &gl->gl_flags); gl->gl_state = new_state; gl->gl_tchange = jiffies; @@ -1012,6 +1014,7 @@ fail: if (unlikely((gh->gh_flags & LM_FLAG_PRIORITY) && !insert_pt)) insert_pt = &gh2->gh_list; } + set_bit(GLF_QUEUED, &gl->gl_flags); if (likely(insert_pt == NULL)) { list_add_tail(&gh->gh_list, &gl->gl_holders); if (unlikely(gh->gh_flags & LM_FLAG_PRIORITY)) @@ -1310,10 +1313,12 @@ void gfs2_glock_cb(struct gfs2_glock *gl, unsigned int state) gfs2_glock_hold(gl); holdtime = gl->gl_tchange + gl->gl_ops->go_min_hold_time; - if (time_before(now, holdtime)) - delay = holdtime - now; - if (test_bit(GLF_REPLY_PENDING, &gl->gl_flags)) - delay = gl->gl_ops->go_min_hold_time; + if (test_bit(GLF_QUEUED, &gl->gl_flags)) { + if (time_before(now, holdtime)) + delay = holdtime - now; + if (test_bit(GLF_REPLY_PENDING, &gl->gl_flags)) + delay = gl->gl_ops->go_min_hold_time; + } spin_lock(&gl->gl_spin); handle_callback(gl, state, delay); @@ -1660,6 +1665,8 @@ static const char *gflags2str(char *buf, const unsigned long *gflags) *p++ = 'I'; if (test_bit(GLF_FROZEN, gflags)) *p++ = 'F'; + if (test_bit(GLF_QUEUED, gflags)) + *p++ = 'q'; *p = 0; return buf; } diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 578234b..b12ca10 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -196,6 +196,7 @@ enum { GLF_REPLY_PENDING = 9, GLF_INITIAL = 10, GLF_FROZEN = 11, + GLF_QUEUED = 12, }; struct gfs2_glock { diff --git a/fs/gfs2/trace_gfs2.h b/fs/gfs2/trace_gfs2.h index 148d55c..cedb0bb 100644 --- a/fs/gfs2/trace_gfs2.h +++ b/fs/gfs2/trace_gfs2.h @@ -39,7 +39,8 @@ {(1UL << GLF_INVALIDATE_IN_PROGRESS), "i" }, \ {(1UL << GLF_REPLY_PENDING), "r" }, \ {(1UL << GLF_INITIAL), "I" }, \ - {(1UL << GLF_FROZEN), "F" }) + {(1UL << GLF_FROZEN), "F" }, \ + {(1UL << GLF_QUEUED), "q" }) #ifndef NUMPTY #define NUMPTY -- cgit v0.10.2 From 1fea7c25a05d388c0cdbe02cbdaf3a2e70885581 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 8 Sep 2010 10:09:25 +0100 Subject: GFS2: Update handling of DLM return codes to match reality GFS2's idea of which return codes it needs to handle was based upon those listed in dlm.h. Those didn't cover all the possible codes and listed some which never happen. This updates GFS2 to handle all the codes which can actually be returned from the DLM under various circumstances. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/lock_dlm.c b/fs/gfs2/lock_dlm.c index 0e0470e..1c09425 100644 --- a/fs/gfs2/lock_dlm.c +++ b/fs/gfs2/lock_dlm.c @@ -42,9 +42,9 @@ static void gdlm_ast(void *arg) ret |= LM_OUT_CANCELED; goto out; case -EAGAIN: /* Try lock fails */ + case -EDEADLK: /* Deadlock detected */ goto out; - case -EINVAL: /* Invalid */ - case -ENOMEM: /* Out of memory */ + case -ETIMEDOUT: /* Canceled due to timeout */ ret |= LM_OUT_ERROR; goto out; case 0: /* Success */ -- cgit v0.10.2 From 9fa0ea9f26f64fbfc3dfd51d1dc2c230b65ffb19 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 13 Sep 2010 16:23:00 +0100 Subject: GFS2: Use new workqueue scheme The recovery workqueue can be freezable since we want it to finish what it is doing if the system is to be frozen (although why you'd want to freeze a cluster node is beyond me since it will result in it being ejected from the cluster). It does still make sense for single node GFS2 filesystems though. The glock workqueue will benefit from being able to run more work items concurrently. A test running postmark shows improved performance and multi-threaded workloads are likely to benefit even more. It needs to be high priority because the latency directly affects the latency of filesystem glock operations. The delete workqueue is similar to the recovery workqueue in that it must not get blocked by memory allocations, and may run for a long time. Potentially other GFS2 threads might also be converted to workqueues, but I'll leave that for a later patch. Signed-off-by: Steven Whitehouse Acked-by: Tejun Heo diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 8e478e2..c3f2a5c 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -1783,10 +1783,12 @@ int __init gfs2_glock_init(void) } #endif - glock_workqueue = create_workqueue("glock_workqueue"); + glock_workqueue = alloc_workqueue("glock_workqueue", WQ_RESCUER | + WQ_HIGHPRI | WQ_FREEZEABLE, 0); if (IS_ERR(glock_workqueue)) return PTR_ERR(glock_workqueue); - gfs2_delete_workqueue = create_workqueue("delete_workqueue"); + gfs2_delete_workqueue = alloc_workqueue("delete_workqueue", WQ_RESCUER | + WQ_FREEZEABLE, 0); if (IS_ERR(gfs2_delete_workqueue)) { destroy_workqueue(glock_workqueue); return PTR_ERR(gfs2_delete_workqueue); diff --git a/fs/gfs2/main.c b/fs/gfs2/main.c index b1e9630..1c8bbf2 100644 --- a/fs/gfs2/main.c +++ b/fs/gfs2/main.c @@ -140,7 +140,7 @@ static int __init init_gfs2_fs(void) error = -ENOMEM; gfs_recovery_wq = alloc_workqueue("gfs_recovery", - WQ_NON_REENTRANT | WQ_RESCUER, 0); + WQ_RESCUER | WQ_FREEZEABLE, 0); if (!gfs_recovery_wq) goto fail_wq; -- cgit v0.10.2 From 8d1235852b462cfb66aa036bd4a2686763c69ed4 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 17 Sep 2010 12:30:23 +0100 Subject: GFS2: Make . and .. qstrs constant Rather than calculating the qstrs for . and .. each time we need them, its better to keep a constant version of these and just refer to them when required. Signed-off-by: Steven Whitehouse Reviewed-by: Christoph Hellwig diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index c1042ae..5c356d0 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -79,6 +79,9 @@ #define gfs2_disk_hash2offset(h) (((u64)(h)) >> 1) #define gfs2_dir_offset2hash(p) ((u32)(((u64)(p)) << 1)) +struct qstr gfs2_qdot __read_mostly; +struct qstr gfs2_qdotdot __read_mostly; + typedef int (*leaf_call_t) (struct gfs2_inode *dip, u32 index, u32 len, u64 leaf_no, void *data); typedef int (*gfs2_dscan_t)(const struct gfs2_dirent *dent, diff --git a/fs/gfs2/dir.h b/fs/gfs2/dir.h index 4f91944..a98f644 100644 --- a/fs/gfs2/dir.h +++ b/fs/gfs2/dir.h @@ -17,23 +17,24 @@ struct inode; struct gfs2_inode; struct gfs2_inum; -struct inode *gfs2_dir_search(struct inode *dir, const struct qstr *filename); -int gfs2_dir_check(struct inode *dir, const struct qstr *filename, - const struct gfs2_inode *ip); -int gfs2_dir_add(struct inode *inode, const struct qstr *filename, - const struct gfs2_inode *ip, unsigned int type); -int gfs2_dir_del(struct gfs2_inode *dip, const struct qstr *filename); -int gfs2_dir_read(struct inode *inode, u64 *offset, void *opaque, - filldir_t filldir); -int gfs2_dir_mvino(struct gfs2_inode *dip, const struct qstr *filename, - const struct gfs2_inode *nip, unsigned int new_type); +extern struct inode *gfs2_dir_search(struct inode *dir, + const struct qstr *filename); +extern int gfs2_dir_check(struct inode *dir, const struct qstr *filename, + const struct gfs2_inode *ip); +extern int gfs2_dir_add(struct inode *inode, const struct qstr *filename, + const struct gfs2_inode *ip, unsigned int type); +extern int gfs2_dir_del(struct gfs2_inode *dip, const struct qstr *filename); +extern int gfs2_dir_read(struct inode *inode, u64 *offset, void *opaque, + filldir_t filldir); +extern int gfs2_dir_mvino(struct gfs2_inode *dip, const struct qstr *filename, + const struct gfs2_inode *nip, unsigned int new_type); -int gfs2_dir_exhash_dealloc(struct gfs2_inode *dip); +extern int gfs2_dir_exhash_dealloc(struct gfs2_inode *dip); -int gfs2_diradd_alloc_required(struct inode *dir, - const struct qstr *filename); -int gfs2_dir_get_new_buffer(struct gfs2_inode *ip, u64 block, - struct buffer_head **bhp); +extern int gfs2_diradd_alloc_required(struct inode *dir, + const struct qstr *filename); +extern int gfs2_dir_get_new_buffer(struct gfs2_inode *ip, u64 block, + struct buffer_head **bhp); static inline u32 gfs2_disk_hash(const char *data, int len) { @@ -61,4 +62,7 @@ static inline void gfs2_qstr2dirent(const struct qstr *name, u16 reclen, struct memcpy(dent + 1, name->name, name->len); } +extern struct qstr gfs2_qdot; +extern struct qstr gfs2_qdotdot; + #endif /* __DIR_DOT_H__ */ diff --git a/fs/gfs2/export.c b/fs/gfs2/export.c index dfe237a..06d5827 100644 --- a/fs/gfs2/export.c +++ b/fs/gfs2/export.c @@ -126,16 +126,9 @@ static int gfs2_get_name(struct dentry *parent, char *name, static struct dentry *gfs2_get_parent(struct dentry *child) { - struct qstr dotdot; struct dentry *dentry; - /* - * XXX(hch): it would be a good idea to keep this around as a - * static variable. - */ - gfs2_str2qstr(&dotdot, ".."); - - dentry = d_obtain_alias(gfs2_lookupi(child->d_inode, &dotdot, 1)); + dentry = d_obtain_alias(gfs2_lookupi(child->d_inode, &gfs2_qdotdot, 1)); if (!IS_ERR(dentry)) dentry->d_op = &gfs2_dops; return dentry; diff --git a/fs/gfs2/main.c b/fs/gfs2/main.c index 1c8bbf2..d7eb1e2 100644 --- a/fs/gfs2/main.c +++ b/fs/gfs2/main.c @@ -24,6 +24,7 @@ #include "glock.h" #include "quota.h" #include "recovery.h" +#include "dir.h" static struct shrinker qd_shrinker = { .shrink = gfs2_shrink_qd_memory, @@ -78,6 +79,9 @@ static int __init init_gfs2_fs(void) { int error; + gfs2_str2qstr(&gfs2_qdot, "."); + gfs2_str2qstr(&gfs2_qdotdot, ".."); + error = gfs2_sys_init(); if (error) return error; diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index ce4f1df..98a94cfc 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -471,18 +471,15 @@ static int gfs2_mkdir(struct inode *dir, struct dentry *dentry, int mode) if (!gfs2_assert_withdraw(sdp, !error)) { struct gfs2_dinode *di = (struct gfs2_dinode *)dibh->b_data; struct gfs2_dirent *dent = (struct gfs2_dirent *)(di+1); - struct qstr str; - gfs2_str2qstr(&str, "."); gfs2_trans_add_bh(ip->i_gl, dibh, 1); - gfs2_qstr2dirent(&str, GFS2_DIRENT_SIZE(str.len), dent); + gfs2_qstr2dirent(&gfs2_qdot, GFS2_DIRENT_SIZE(gfs2_qdot.len), dent); dent->de_inum = di->di_num; /* already GFS2 endian */ dent->de_type = cpu_to_be16(DT_DIR); di->di_entries = cpu_to_be32(1); - gfs2_str2qstr(&str, ".."); dent = (struct gfs2_dirent *)((char*)dent + GFS2_DIRENT_SIZE(1)); - gfs2_qstr2dirent(&str, dibh->b_size - GFS2_DIRENT_SIZE(1) - sizeof(struct gfs2_dinode), dent); + gfs2_qstr2dirent(&gfs2_qdotdot, dibh->b_size - GFS2_DIRENT_SIZE(1) - sizeof(struct gfs2_dinode), dent); gfs2_inum_out(dip, dent); dent->de_type = cpu_to_be16(DT_DIR); @@ -523,7 +520,6 @@ static int gfs2_mkdir(struct inode *dir, struct dentry *dentry, int mode) static int gfs2_rmdiri(struct gfs2_inode *dip, const struct qstr *name, struct gfs2_inode *ip) { - struct qstr dotname; int error; if (ip->i_entries != 2) { @@ -540,13 +536,11 @@ static int gfs2_rmdiri(struct gfs2_inode *dip, const struct qstr *name, if (error) return error; - gfs2_str2qstr(&dotname, "."); - error = gfs2_dir_del(ip, &dotname); + error = gfs2_dir_del(ip, &gfs2_qdot); if (error) return error; - gfs2_str2qstr(&dotname, ".."); - error = gfs2_dir_del(ip, &dotname); + error = gfs2_dir_del(ip, &gfs2_qdotdot); if (error) return error; @@ -695,11 +689,8 @@ static int gfs2_ok_to_move(struct gfs2_inode *this, struct gfs2_inode *to) struct inode *dir = &to->i_inode; struct super_block *sb = dir->i_sb; struct inode *tmp; - struct qstr dotdot; int error = 0; - gfs2_str2qstr(&dotdot, ".."); - igrab(dir); for (;;) { @@ -712,7 +703,7 @@ static int gfs2_ok_to_move(struct gfs2_inode *this, struct gfs2_inode *to) break; } - tmp = gfs2_lookupi(dir, &dotdot, 1); + tmp = gfs2_lookupi(dir, &gfs2_qdotdot, 1); if (IS_ERR(tmp)) { error = PTR_ERR(tmp); break; @@ -921,9 +912,6 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, } if (dir_rename) { - struct qstr name; - gfs2_str2qstr(&name, ".."); - error = gfs2_change_nlink(ndip, +1); if (error) goto out_end_trans; @@ -931,7 +919,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, if (error) goto out_end_trans; - error = gfs2_dir_mvino(ip, &name, ndip, DT_DIR); + error = gfs2_dir_mvino(ip, &gfs2_qdotdot, ndip, DT_DIR); if (error) goto out_end_trans; } else { -- cgit v0.10.2 From 23ac4ae827e6264e21b898f2cd3f601450aa02a6 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Fri, 17 Sep 2010 18:03:43 +0200 Subject: x86, k8: Rename k8.[ch] to amd_nb.[ch] and CONFIG_K8_NB to CONFIG_AMD_NB The file names are somehow misleading as the code is not specific to AMD K8 CPUs anymore. The files accomodate code for other AMD CPU northbridges as well. Same is true for the config option which is valid for AMD CPU northbridges in general and not specific to K8. Signed-off-by: Andreas Herrmann LKML-Reference: <20100917160343.GD4958@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..7fd41f0 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -670,7 +670,7 @@ config GART_IOMMU bool "GART IOMMU support" if EMBEDDED default y select SWIOTLB - depends on X86_64 && PCI && K8_NB + depends on X86_64 && PCI && AMD_NB ---help--- Support for full DMA access of devices with 32bit memory access only on systems with more than 3GB. This is usually needed for USB, @@ -2076,7 +2076,7 @@ config OLPC_OPENFIRMWARE endif # X86_32 -config K8_NB +config AMD_NB def_bool y depends on CPU_SUP_AMD && PCI diff --git a/arch/x86/include/asm/amd_nb.h b/arch/x86/include/asm/amd_nb.h new file mode 100644 index 0000000..c8517f8 --- /dev/null +++ b/arch/x86/include/asm/amd_nb.h @@ -0,0 +1,39 @@ +#ifndef _ASM_X86_AMD_NB_H +#define _ASM_X86_AMD_NB_H + +#include + +extern struct pci_device_id k8_nb_ids[]; +struct bootnode; + +extern int early_is_k8_nb(u32 value); +extern int cache_k8_northbridges(void); +extern void k8_flush_garts(void); +extern int k8_get_nodes(struct bootnode *nodes); +extern int k8_numa_init(unsigned long start_pfn, unsigned long end_pfn); +extern int k8_scan_nodes(void); + +struct k8_northbridge_info { + u16 num; + u8 gart_supported; + struct pci_dev **nb_misc; +}; +extern struct k8_northbridge_info k8_northbridges; + +#ifdef CONFIG_AMD_NB + +static inline struct pci_dev *node_to_k8_nb_misc(int node) +{ + return (node < k8_northbridges.num) ? k8_northbridges.nb_misc[node] : NULL; +} + +#else + +static inline struct pci_dev *node_to_k8_nb_misc(int node) +{ + return NULL; +} +#endif + + +#endif /* _ASM_X86_AMD_NB_H */ diff --git a/arch/x86/include/asm/k8.h b/arch/x86/include/asm/k8.h deleted file mode 100644 index 9cee145..0000000 --- a/arch/x86/include/asm/k8.h +++ /dev/null @@ -1,39 +0,0 @@ -#ifndef _ASM_X86_K8_H -#define _ASM_X86_K8_H - -#include - -extern struct pci_device_id k8_nb_ids[]; -struct bootnode; - -extern int early_is_k8_nb(u32 value); -extern int cache_k8_northbridges(void); -extern void k8_flush_garts(void); -extern int k8_get_nodes(struct bootnode *nodes); -extern int k8_numa_init(unsigned long start_pfn, unsigned long end_pfn); -extern int k8_scan_nodes(void); - -struct k8_northbridge_info { - u16 num; - u8 gart_supported; - struct pci_dev **nb_misc; -}; -extern struct k8_northbridge_info k8_northbridges; - -#ifdef CONFIG_K8_NB - -static inline struct pci_dev *node_to_k8_nb_misc(int node) -{ - return (node < k8_northbridges.num) ? k8_northbridges.nb_misc[node] : NULL; -} - -#else - -static inline struct pci_dev *node_to_k8_nb_misc(int node) -{ - return NULL; -} -#endif - - -#endif /* _ASM_X86_K8_H */ diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 0925676..25dc82d 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -87,7 +87,7 @@ obj-$(CONFIG_EARLY_PRINTK) += early_printk.o obj-$(CONFIG_HPET_TIMER) += hpet.o obj-$(CONFIG_APB_TIMER) += apb_timer.o -obj-$(CONFIG_K8_NB) += k8.o +obj-$(CONFIG_AMD_NB) += amd_nb.o obj-$(CONFIG_DEBUG_RODATA_TEST) += test_rodata.o obj-$(CONFIG_DEBUG_NX_TEST) += test_nx.o diff --git a/arch/x86/kernel/amd_nb.c b/arch/x86/kernel/amd_nb.c new file mode 100644 index 0000000..4ffc38d --- /dev/null +++ b/arch/x86/kernel/amd_nb.c @@ -0,0 +1,145 @@ +/* + * Shared support code for AMD K8 northbridges and derivates. + * Copyright 2006 Andi Kleen, SUSE Labs. Subject to GPLv2. + */ +#include +#include +#include +#include +#include +#include +#include + +static u32 *flush_words; + +struct pci_device_id k8_nb_ids[] = { + { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_K8_NB_MISC) }, + { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_10H_NB_MISC) }, + {} +}; +EXPORT_SYMBOL(k8_nb_ids); + +struct k8_northbridge_info k8_northbridges; +EXPORT_SYMBOL(k8_northbridges); + +static struct pci_dev *next_k8_northbridge(struct pci_dev *dev) +{ + do { + dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev); + if (!dev) + break; + } while (!pci_match_id(&k8_nb_ids[0], dev)); + return dev; +} + +int cache_k8_northbridges(void) +{ + int i; + struct pci_dev *dev; + + if (k8_northbridges.num) + return 0; + + dev = NULL; + while ((dev = next_k8_northbridge(dev)) != NULL) + k8_northbridges.num++; + + /* some CPU families (e.g. family 0x11) do not support GART */ + if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10) + k8_northbridges.gart_supported = 1; + + k8_northbridges.nb_misc = kmalloc((k8_northbridges.num + 1) * + sizeof(void *), GFP_KERNEL); + if (!k8_northbridges.nb_misc) + return -ENOMEM; + + if (!k8_northbridges.num) { + k8_northbridges.nb_misc[0] = NULL; + return 0; + } + + if (k8_northbridges.gart_supported) { + flush_words = kmalloc(k8_northbridges.num * sizeof(u32), + GFP_KERNEL); + if (!flush_words) { + kfree(k8_northbridges.nb_misc); + return -ENOMEM; + } + } + + dev = NULL; + i = 0; + while ((dev = next_k8_northbridge(dev)) != NULL) { + k8_northbridges.nb_misc[i] = dev; + if (k8_northbridges.gart_supported) + pci_read_config_dword(dev, 0x9c, &flush_words[i++]); + } + k8_northbridges.nb_misc[i] = NULL; + return 0; +} +EXPORT_SYMBOL_GPL(cache_k8_northbridges); + +/* Ignores subdevice/subvendor but as far as I can figure out + they're useless anyways */ +int __init early_is_k8_nb(u32 device) +{ + struct pci_device_id *id; + u32 vendor = device & 0xffff; + device >>= 16; + for (id = k8_nb_ids; id->vendor; id++) + if (vendor == id->vendor && device == id->device) + return 1; + return 0; +} + +void k8_flush_garts(void) +{ + int flushed, i; + unsigned long flags; + static DEFINE_SPINLOCK(gart_lock); + + if (!k8_northbridges.gart_supported) + return; + + /* Avoid races between AGP and IOMMU. In theory it's not needed + but I'm not sure if the hardware won't lose flush requests + when another is pending. This whole thing is so expensive anyways + that it doesn't matter to serialize more. -AK */ + spin_lock_irqsave(&gart_lock, flags); + flushed = 0; + for (i = 0; i < k8_northbridges.num; i++) { + pci_write_config_dword(k8_northbridges.nb_misc[i], 0x9c, + flush_words[i]|1); + flushed++; + } + for (i = 0; i < k8_northbridges.num; i++) { + u32 w; + /* Make sure the hardware actually executed the flush*/ + for (;;) { + pci_read_config_dword(k8_northbridges.nb_misc[i], + 0x9c, &w); + if (!(w & 1)) + break; + cpu_relax(); + } + } + spin_unlock_irqrestore(&gart_lock, flags); + if (!flushed) + printk("nothing to flush?\n"); +} +EXPORT_SYMBOL_GPL(k8_flush_garts); + +static __init int init_k8_nbs(void) +{ + int err = 0; + + err = cache_k8_northbridges(); + + if (err < 0) + printk(KERN_NOTICE "K8 NB: Cannot enumerate AMD northbridges.\n"); + + return err; +} + +/* This has to go after the PCI subsystem */ +fs_initcall(init_k8_nbs); diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c index a2e0caf..e91e042 100644 --- a/arch/x86/kernel/aperture_64.c +++ b/arch/x86/kernel/aperture_64.c @@ -27,7 +27,7 @@ #include #include #include -#include +#include #include int gart_iommu_aperture; diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 6fdfb0b2..12cd823 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -17,7 +17,7 @@ #include #include -#include +#include #include #define LVL_1_INST 1 @@ -306,7 +306,7 @@ struct _cache_attr { ssize_t (*store)(struct _cpuid4_info *, const char *, size_t count); }; -#ifdef CONFIG_K8_NB +#ifdef CONFIG_AMD_NB /* * L3 cache descriptors @@ -556,12 +556,12 @@ static struct _cache_attr cache_disable_0 = __ATTR(cache_disable_0, 0644, static struct _cache_attr cache_disable_1 = __ATTR(cache_disable_1, 0644, show_cache_disable_1, store_cache_disable_1); -#else /* CONFIG_K8_NB */ +#else /* CONFIG_AMD_NB */ static void __cpuinit amd_check_l3_disable(struct _cpuid4_info_regs *this_leaf, int index) { }; -#endif /* CONFIG_K8_NB */ +#endif /* CONFIG_AMD_NB */ static int __cpuinit cpuid4_cache_lookup_regs(int index, @@ -1000,7 +1000,7 @@ static struct attribute *default_attrs[] = { static struct attribute *default_l3_attrs[] = { DEFAULT_SYSFS_CACHE_ATTRS, -#ifdef CONFIG_K8_NB +#ifdef CONFIG_AMD_NB &cache_disable_0.attr, &cache_disable_1.attr, #endif diff --git a/arch/x86/kernel/k8.c b/arch/x86/kernel/k8.c deleted file mode 100644 index 5de1b6b..0000000 --- a/arch/x86/kernel/k8.c +++ /dev/null @@ -1,145 +0,0 @@ -/* - * Shared support code for AMD K8 northbridges and derivates. - * Copyright 2006 Andi Kleen, SUSE Labs. Subject to GPLv2. - */ -#include -#include -#include -#include -#include -#include -#include - -static u32 *flush_words; - -struct pci_device_id k8_nb_ids[] = { - { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_K8_NB_MISC) }, - { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_10H_NB_MISC) }, - {} -}; -EXPORT_SYMBOL(k8_nb_ids); - -struct k8_northbridge_info k8_northbridges; -EXPORT_SYMBOL(k8_northbridges); - -static struct pci_dev *next_k8_northbridge(struct pci_dev *dev) -{ - do { - dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev); - if (!dev) - break; - } while (!pci_match_id(&k8_nb_ids[0], dev)); - return dev; -} - -int cache_k8_northbridges(void) -{ - int i; - struct pci_dev *dev; - - if (k8_northbridges.num) - return 0; - - dev = NULL; - while ((dev = next_k8_northbridge(dev)) != NULL) - k8_northbridges.num++; - - /* some CPU families (e.g. family 0x11) do not support GART */ - if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10) - k8_northbridges.gart_supported = 1; - - k8_northbridges.nb_misc = kmalloc((k8_northbridges.num + 1) * - sizeof(void *), GFP_KERNEL); - if (!k8_northbridges.nb_misc) - return -ENOMEM; - - if (!k8_northbridges.num) { - k8_northbridges.nb_misc[0] = NULL; - return 0; - } - - if (k8_northbridges.gart_supported) { - flush_words = kmalloc(k8_northbridges.num * sizeof(u32), - GFP_KERNEL); - if (!flush_words) { - kfree(k8_northbridges.nb_misc); - return -ENOMEM; - } - } - - dev = NULL; - i = 0; - while ((dev = next_k8_northbridge(dev)) != NULL) { - k8_northbridges.nb_misc[i] = dev; - if (k8_northbridges.gart_supported) - pci_read_config_dword(dev, 0x9c, &flush_words[i++]); - } - k8_northbridges.nb_misc[i] = NULL; - return 0; -} -EXPORT_SYMBOL_GPL(cache_k8_northbridges); - -/* Ignores subdevice/subvendor but as far as I can figure out - they're useless anyways */ -int __init early_is_k8_nb(u32 device) -{ - struct pci_device_id *id; - u32 vendor = device & 0xffff; - device >>= 16; - for (id = k8_nb_ids; id->vendor; id++) - if (vendor == id->vendor && device == id->device) - return 1; - return 0; -} - -void k8_flush_garts(void) -{ - int flushed, i; - unsigned long flags; - static DEFINE_SPINLOCK(gart_lock); - - if (!k8_northbridges.gart_supported) - return; - - /* Avoid races between AGP and IOMMU. In theory it's not needed - but I'm not sure if the hardware won't lose flush requests - when another is pending. This whole thing is so expensive anyways - that it doesn't matter to serialize more. -AK */ - spin_lock_irqsave(&gart_lock, flags); - flushed = 0; - for (i = 0; i < k8_northbridges.num; i++) { - pci_write_config_dword(k8_northbridges.nb_misc[i], 0x9c, - flush_words[i]|1); - flushed++; - } - for (i = 0; i < k8_northbridges.num; i++) { - u32 w; - /* Make sure the hardware actually executed the flush*/ - for (;;) { - pci_read_config_dword(k8_northbridges.nb_misc[i], - 0x9c, &w); - if (!(w & 1)) - break; - cpu_relax(); - } - } - spin_unlock_irqrestore(&gart_lock, flags); - if (!flushed) - printk("nothing to flush?\n"); -} -EXPORT_SYMBOL_GPL(k8_flush_garts); - -static __init int init_k8_nbs(void) -{ - int err = 0; - - err = cache_k8_northbridges(); - - if (err < 0) - printk(KERN_NOTICE "K8 NB: Cannot enumerate AMD northbridges.\n"); - - return err; -} - -/* This has to go after the PCI subsystem */ -fs_initcall(init_k8_nbs); diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index 8f214a2..67e5665 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -39,7 +39,7 @@ #include #include #include -#include +#include #include static unsigned long iommu_bus_base; /* GART remapping area (physical) */ diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index c3a4fbb..77eea03 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -107,7 +107,7 @@ #include #include #include -#include +#include #ifdef CONFIG_X86_64 #include #endif diff --git a/arch/x86/mm/k8topology_64.c b/arch/x86/mm/k8topology_64.c index 970ed57..ab75b18 100644 --- a/arch/x86/mm/k8topology_64.c +++ b/arch/x86/mm/k8topology_64.c @@ -22,7 +22,7 @@ #include #include #include -#include +#include static struct bootnode __initdata nodes[8]; static nodemask_t __initdata nodes_parsed = NODE_MASK_NONE; diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c index a7bcc23..4962f1a 100644 --- a/arch/x86/mm/numa_64.c +++ b/arch/x86/mm/numa_64.c @@ -18,7 +18,7 @@ #include #include #include -#include +#include struct pglist_data *node_data[MAX_NUMNODES] __read_mostly; EXPORT_SYMBOL(node_data); diff --git a/drivers/char/agp/Kconfig b/drivers/char/agp/Kconfig index 4b66c69..5ddf67e 100644 --- a/drivers/char/agp/Kconfig +++ b/drivers/char/agp/Kconfig @@ -57,7 +57,7 @@ config AGP_AMD config AGP_AMD64 tristate "AMD Opteron/Athlon64 on-CPU GART support" - depends on AGP && X86 && K8_NB + depends on AGP && X86 && AMD_NB help This option gives you AGP support for the GLX component of X using the on-CPU northbridge of the AMD Athlon64/Opteron CPUs. diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c index bdf00d58..4d6087c 100644 --- a/drivers/char/agp/amd64-agp.c +++ b/drivers/char/agp/amd64-agp.c @@ -15,7 +15,7 @@ #include #include /* PAGE_SIZE */ #include -#include +#include #include #include "agp.h" diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index 70bb350..734e2e0 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -66,7 +66,7 @@ config EDAC_MCE config EDAC_AMD64 tristate "AMD64 (Opteron, Athlon64) K8, F10h, F11h" - depends on EDAC_MM_EDAC && K8_NB && X86_64 && PCI && EDAC_DECODE_MCE + depends on EDAC_MM_EDAC && AMD_NB && X86_64 && PCI && EDAC_DECODE_MCE help Support for error detection and correction on the AMD 64 Families of Memory Controllers (K8, F10h and F11h) diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c index 5babf6f..09fcc52 100644 --- a/drivers/edac/amd64_edac.c +++ b/drivers/edac/amd64_edac.c @@ -1,5 +1,5 @@ #include "amd64_edac.h" -#include +#include static struct edac_pci_ctl_info *amd64_ctl_pci; -- cgit v0.10.2 From ce5f68246bf2385d6174856708d0b746dc378f20 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 20 Sep 2010 13:04:45 -0700 Subject: x86, hotplug: In the MWAIT case of play_dead, CLFLUSH the cache line When we're using MWAIT for play_dead, explicitly CLFLUSH the cache line before executing MONITOR. This is a potential workaround for the Xeon 7400 erratum AAI65 after having a spurious wakeup and returning around the loop. "Potential" here because it is not certain that that erratum could actually trigger; however, the CLFLUSH should be harmless. Signed-off-by: H. Peter Anvin Acked-by: Venkatesh Pallipadi Cc: Asit Mallick Cc: Arjan van de Ven Cc: Len Brown diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 55c80ffb..fdccfe9 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1394,9 +1394,12 @@ static inline void mwait_play_dead(void) unsigned int highest_cstate = 0; unsigned int highest_subcstate = 0; int i; + void *mwait_ptr; if (!cpu_has(¤t_cpu_data, X86_FEATURE_MWAIT)) return; + if (!cpu_has(¤t_cpu_data, X86_FEATURE_CLFLSH)) + return; if (current_cpu_data.cpuid_level < CPUID_MWAIT_LEAF) return; @@ -1422,10 +1425,25 @@ static inline void mwait_play_dead(void) (highest_subcstate - 1); } + /* + * This should be a memory location in a cache line which is + * unlikely to be touched by other processors. The actual + * content is immaterial as it is not actually modified in any way. + */ + mwait_ptr = ¤t_thread_info()->flags; + wbinvd(); while (1) { - __monitor(¤t_thread_info()->flags, 0, 0); + /* + * The CLFLUSH is a workaround for erratum AAI65 for + * the Xeon 7400 series. It's not clear it is actually + * needed, but it should be harmless in either case. + * The WBINVD is insufficient due to the spurious-wakeup + * case where we return around the loop. + */ + clflush(mwait_ptr); + __monitor(mwait_ptr, 0, 0); mb(); __mwait(eax, 0); } -- cgit v0.10.2 From c2b9ff24a0df649d4d40947878b5b5ac39c7299e Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 20 Sep 2010 18:01:46 -0700 Subject: x86, cpu: Re-run get_cpu_cap() after adjusting the CPUID level At least on Intel, adjusting the max CPUID level can expose new CPUID features, so we need to re-run get_cpu_cap() after changing the CPUID level. Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 490dac6..f2f9ac7 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -545,7 +545,7 @@ void __cpuinit cpu_detect(struct cpuinfo_x86 *c) } } -static void __cpuinit get_cpu_cap(struct cpuinfo_x86 *c) +void __cpuinit get_cpu_cap(struct cpuinfo_x86 *c) { u32 tfms, xlvl; u32 ebx; diff --git a/arch/x86/kernel/cpu/cpu.h b/arch/x86/kernel/cpu/cpu.h index 3624e8a..c16456b 100644 --- a/arch/x86/kernel/cpu/cpu.h +++ b/arch/x86/kernel/cpu/cpu.h @@ -32,6 +32,7 @@ struct cpu_dev { extern const struct cpu_dev *const __x86_cpu_dev_start[], *const __x86_cpu_dev_end[]; +extern void get_cpu_cap(struct cpuinfo_x86 *c); extern void cpu_detect_cache_sizes(struct cpuinfo_x86 *c); #endif diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index 85f69cd..b438944 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -39,6 +39,7 @@ static void __cpuinit early_init_intel(struct cpuinfo_x86 *c) misc_enable &= ~MSR_IA32_MISC_ENABLE_LIMIT_CPUID; wrmsrl(MSR_IA32_MISC_ENABLE, misc_enable); c->cpuid_level = cpuid_eax(0); + get_cpu_cap(c); } } -- cgit v0.10.2 From d2520a426dc3033c00077e923a553fc6c98c7564 Mon Sep 17 00:00:00 2001 From: Kenneth Waters Date: Tue, 21 Sep 2010 00:58:23 -0700 Subject: Input: joydev - fix JSIOCSAXMAP ioctl Fixed JSIOCSAXMAP ioctl to update absmap, the map from hardware axis to event axis in addition to abspam. This fixes a regression introduced by 999b874f. Signed-off-by: Kenneth Waters Cc: stable@kernel.org Signed-off-by: Dmitry Torokhov diff --git a/drivers/input/joydev.c b/drivers/input/joydev.c index d85bd8a..22239e9 100644 --- a/drivers/input/joydev.c +++ b/drivers/input/joydev.c @@ -483,6 +483,9 @@ static int joydev_handle_JSIOCSAXMAP(struct joydev *joydev, memcpy(joydev->abspam, abspam, len); + for (i = 0; i < joydev->nabs; i++) + joydev->absmap[joydev->abspam[i]] = i; + out: kfree(abspam); return retval; -- cgit v0.10.2 From a800c7cc5380fbb6b4f2f3bd89f6776eb3da2242 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 21 Sep 2010 11:33:54 +0200 Subject: MAINTAINERS: Add IRQ subsystem I do not expect, that this will help anything, but at least it's going to remove the lame excuse about the missing maintainer entry. Signed-off-by: Thomas Gleixner diff --git a/MAINTAINERS b/MAINTAINERS index 50b8148..630a1f7 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3202,6 +3202,12 @@ F: drivers/net/irda/ F: include/net/irda/ F: net/irda/ +IRQ SUBSYSTEM +M: Thomas Gleixner +S: Maintained +T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git irq/core +F: kernel/irq/ + ISAPNP M: Jaroslav Kysela S: Maintained -- cgit v0.10.2 From 58b26c4c025778c09c7a1438ff185080e11b7d0a Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Fri, 10 Sep 2010 18:19:17 -0700 Subject: sched: Increment cache_nice_tries only on periodic lb scheduler uses cache_nice_tries as an indicator to do cache_hot and active load balance, when normal load balance fails. Currently, this value is changed on any failed load balance attempt. That ends up being not so nice to workloads that enter/exit idle often, as they do more frequent new_idle balance and that pretty soon results in cache hot tasks being pulled in. Making the cache_nice_tries ignore failed new_idle balance seems to make better sense. With that only the failed load balance in periodic load balance gets accounted and the rate of accumulation of cache_nice_tries will not depend on idle entry/exit (short running sleep-wakeup kind of tasks). This reduces movement of cache_hot tasks. schedstat diff (after-before) excerpt from a workload that has frequent and short wakeup-idle pattern (:2 in cpu col below refers to NEWIDLE idx) This snapshot was across ~400 seconds. Without this change: domainstats: domain0 cpu cnt bln fld imb gain hgain nobusyq nobusyg 0:2 306487 219575 73167 110069413 44583 19070 1172 218403 1:2 292139 194853 81421 120893383 50745 21902 1259 193594 2:2 283166 174607 91359 129699642 54931 23688 1287 173320 3:2 273998 161788 93991 132757146 57122 24351 1366 160422 4:2 289851 215692 62190 83398383 36377 13680 851 214841 5:2 316312 222146 77605 117582154 49948 20281 988 221158 6:2 297172 195596 83623 122133390 52801 21301 929 194667 7:2 283391 178078 86378 126622761 55122 22239 928 177150 8:2 297655 210359 72995 110246694 45798 19777 1125 209234 9:2 297357 202011 79363 119753474 50953 22088 1089 200922 10:2 278797 178703 83180 122514385 52969 22726 1128 177575 11:2 272661 167669 86978 127342327 55857 24342 1195 166474 12:2 293039 204031 73211 110282059 47285 19651 948 203083 13:2 289502 196762 76803 114712942 49339 20547 1016 195746 14:2 264446 169609 78292 115715605 50459 21017 982 168627 15:2 260968 163660 80142 116811793 51483 21281 1064 162596 With this change: domainstats: domain0 cpu cnt bln fld imb gain hgain nobusyq nobusyg 0:2 272347 187380 77455 105420270 24975 1 953 186427 1:2 267276 172360 86234 116242264 28087 6 1028 171332 2:2 259769 156777 93281 123243134 30555 1 1043 155734 3:2 250870 143129 97627 127370868 32026 6 1188 141941 4:2 248422 177116 64096 78261112 22202 2 757 176359 5:2 275595 180683 84950 116075022 29400 6 778 179905 6:2 262418 162609 88944 119256898 31056 4 817 161792 7:2 252204 147946 92646 122388300 32879 4 824 147122 8:2 262335 172239 81631 110477214 26599 4 864 171375 9:2 261563 164775 88016 117203621 28331 3 849 163926 10:2 243389 140949 93379 121353071 29585 2 909 140040 11:2 242795 134651 98310 124768957 30895 2 1016 133635 12:2 255234 166622 79843 104696912 26483 4 746 165876 13:2 244944 151595 83855 109808099 27787 3 801 150794 14:2 241301 140982 89935 116954383 30403 6 845 140137 15:2 232271 128564 92821 119185207 31207 4 1416 127148 Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1284167957-3675-1-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index a171138..aa16cf1 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -3031,7 +3031,14 @@ redo: if (!ld_moved) { schedstat_inc(sd, lb_failed[idle]); - sd->nr_balance_failed++; + /* + * Increment the failure counter only on periodic balance. + * We do not want newidle balance, which can be very + * frequent, pollute the failure counter causing + * excessive cache_hot migrations and active balances. + */ + if (idle != CPU_NEWLY_IDLE) + sd->nr_balance_failed++; if (need_active_balance(sd, sd_idle, idle, cpu_of(busiest), this_cpu)) { -- cgit v0.10.2 From 43fa5460fe60dea5c610490a1d263415419c60f6 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 20 Sep 2010 22:40:03 -0400 Subject: sched: Try not to migrate higher priority RT tasks When first working on the RT scheduler design, we concentrated on keeping all CPUs running RT tasks instead of having multiple RT tasks on a single CPU waiting for the migration thread to move them. Instead we take a more proactive stance and push or pull RT tasks from one CPU to another on wakeup or scheduling. When an RT task wakes up on a CPU that is running another RT task, instead of preempting it and killing the cache of the running RT task, we look to see if we can migrate the RT task that is waking up, even if the RT task waking up is of higher priority. This may sound a bit odd, but RT tasks should be limited in migration by the user anyway. But in practice, people do not do this, which causes high prio RT tasks to bounce around the CPUs. This becomes even worse when we have priority inheritance, because a high prio task can block on a lower prio task and boost its priority. When the lower prio task wakes up the high prio task, if it happens to be on the same CPU it will migrate off of it. But in reality, the above does not happen much either, because the wake up of the lower prio task, which has already been boosted, if it was on the same CPU as the higher prio task, it would then migrate off of it. But anyway, we do not want to migrate them either. To examine the scheduling, I created a test program and examined it under kernelshark. The test program created CPU * 2 threads, where each thread had a different priority. The program takes different options. The options used in this change log was to have priority inheritance mutexes or not. All threads did the following loop: static void grab_lock(long id, int iter, int l) { ftrace_write("thread %ld iter %d, taking lock %d\n", id, iter, l); pthread_mutex_lock(&locks[l]); ftrace_write("thread %ld iter %d, took lock %d\n", id, iter, l); busy_loop(nr_tasks - id); ftrace_write("thread %ld iter %d, unlock lock %d\n", id, iter, l); pthread_mutex_unlock(&locks[l]); } void *start_task(void *id) { [...] while (!done) { for (l = 0; l < nr_locks; l++) { grab_lock(id, i, l); ftrace_write("thread %ld iter %d sleeping\n", id, i); ms_sleep(id); } i++; } [...] } The busy_loop(ms) keeps the CPU spinning for ms milliseconds. The ms_sleep(ms) sleeps for ms milliseconds. The ftrace_write() writes to the ftrace buffer to help analyze via ftrace. The higher the id, the higher the prio, the shorter it does the busy loop, but the longer it spins. This is usually the case with RT tasks, the lower priority tasks usually run longer than higher priority tasks. At the end of the test, it records the number of loops each thread took, as well as the number of voluntary preemptions, non-voluntary preemptions, and number of migrations each thread took, taking the information from /proc/$$/sched and /proc/$$/status. Running this on a 4 CPU processor, the results without changes to the kernel looked like this: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 53 3220 1470 98 1: 562 773 724 98 2: 752 933 1375 98 3: 749 39 697 98 4: 758 5 515 98 5: 764 2 679 99 6: 761 2 535 99 7: 757 3 346 99 total: 5156 4977 6341 787 Each thread regardless of priority migrated a few hundred times. The higher priority tasks, were a little better but still took quite an impact. By letting higher priority tasks bump the lower prio task from the CPU, things changed a bit: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 37 2835 1937 98 1: 666 1821 1865 98 2: 654 1003 1385 98 3: 664 635 973 99 4: 698 197 352 99 5: 703 101 159 99 6: 708 1 75 99 7: 713 1 2 99 total: 4843 6594 6748 789 The total # of migrations did not change (several runs showed the difference all within the noise). But we now see a dramatic improvement to the higher priority tasks. (kernelshark showed that the watchdog timer bumped the highest priority task to give it the 2 count. This was actually consistent with every run). Notice that the # of iterations did not change either. The above was with priority inheritance mutexes. That is, when the higher prority task blocked on a lower priority task, the lower priority task would inherit the higher priority task (which shows why task 6 was bumped so many times). When not using priority inheritance mutexes, the current kernel shows this: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 56 3101 1892 95 1: 594 713 937 95 2: 625 188 618 95 3: 628 4 491 96 4: 640 7 468 96 5: 631 2 501 96 6: 641 1 466 96 7: 643 2 497 96 total: 4458 4018 5870 765 Not much changed with or without priority inheritance mutexes. But if we let the high priority task bump lower priority tasks on wakeup we see: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 115 3439 2782 98 1: 633 1354 1583 99 2: 652 919 1218 99 3: 645 713 934 99 4: 690 3 3 99 5: 694 1 4 99 6: 720 3 4 99 7: 747 0 1 100 Which shows a even bigger change. The big difference between task 3 and task 4 is because we have only 4 CPUs on the machine, causing the 4 highest prio tasks to always have preference. Although I did not measure cache misses, and I'm sure there would be little to measure since the test was not data intensive, I could imagine large improvements for higher priority tasks when dealing with lower priority tasks. Thus, I'm satisfied with making the change and agreeing with what Gregory Haskins argued a few years ago when we first had this discussion. One final note. All tasks in the above tests were RT tasks. Any RT task will always preempt a non RT task that is running on the CPU the RT task wants to run on. Signed-off-by: Steven Rostedt Signed-off-by: Peter Zijlstra Cc: Gregory Haskins LKML-Reference: <20100921024138.605460343@goodmis.org> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index d10c80e..6a02b38 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -960,18 +960,18 @@ select_task_rq_rt(struct rq *rq, struct task_struct *p, int sd_flag, int flags) * runqueue. Otherwise simply start this RT task * on its current runqueue. * - * We want to avoid overloading runqueues. Even if - * the RT task is of higher priority than the current RT task. - * RT tasks behave differently than other tasks. If - * one gets preempted, we try to push it off to another queue. - * So trying to keep a preempting RT task on the same - * cache hot CPU will force the running RT task to - * a cold CPU. So we waste all the cache for the lower - * RT task in hopes of saving some of a RT task - * that is just being woken and probably will have - * cold cache anyway. + * We want to avoid overloading runqueues. If the woken + * task is a higher priority, then it will stay on this CPU + * and the lower prio task should be moved to another CPU. + * Even though this will probably make the lower prio task + * lose its cache, we do not want to bounce a higher task + * around just because it gave up its CPU, perhaps for a + * lock? + * + * For equal prio tasks, we just let the scheduler sort it out. */ if (unlikely(rt_task(rq->curr)) && + rq->curr->prio < p->prio && (p->rt.nr_cpus_allowed > 1)) { int cpu = find_lowest_rq(p); @@ -1491,6 +1491,8 @@ static void task_woken_rt(struct rq *rq, struct task_struct *p) if (!task_running(rq, p) && !test_tsk_need_resched(rq->curr) && has_pushable_tasks(rq) && + rt_task(rq->curr) && + rq->curr->prio < p->prio && p->rt.nr_cpus_allowed > 1) push_rt_tasks(rq); } -- cgit v0.10.2 From b3bc211cfe7d5fe94b310480d78e00bea96fbf2a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 20 Sep 2010 22:40:04 -0400 Subject: sched: Give CPU bound RT tasks preference If a high priority task is waking up on a CPU that is running a lower priority task that is bound to a CPU, see if we can move the high RT task to another CPU first. Note, if all other CPUs are running higher priority tasks than the CPU bounded current task, then it will be preempted regardless. Signed-off-by: Steven Rostedt Signed-off-by: Peter Zijlstra Cc: Gregory Haskins LKML-Reference: <20100921024138.888922071@goodmis.org> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 6a02b38..baef30f 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -971,7 +971,8 @@ select_task_rq_rt(struct rq *rq, struct task_struct *p, int sd_flag, int flags) * For equal prio tasks, we just let the scheduler sort it out. */ if (unlikely(rt_task(rq->curr)) && - rq->curr->prio < p->prio && + (rq->curr->rt.nr_cpus_allowed < 2 || + rq->curr->prio < p->prio) && (p->rt.nr_cpus_allowed > 1)) { int cpu = find_lowest_rq(p); @@ -1491,9 +1492,10 @@ static void task_woken_rt(struct rq *rq, struct task_struct *p) if (!task_running(rq, p) && !test_tsk_need_resched(rq->curr) && has_pushable_tasks(rq) && + p->rt.nr_cpus_allowed > 1 && rt_task(rq->curr) && - rq->curr->prio < p->prio && - p->rt.nr_cpus_allowed > 1) + (rq->curr->rt.nr_cpus_allowed < 2 || + rq->curr->prio < p->prio)) push_rt_tasks(rq); } -- cgit v0.10.2 From a8027073eb127cd207070891374b5c54c2ce3d23 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 20 Sep 2010 15:13:34 -0400 Subject: tracing/sched: Add sched_pi_setprio tracepoint Add a tracepoint that shows the priority of a task being boosted via priority inheritance. Cc: Gregory Haskins Acked-by: Peter Zijlstra Signed-off-by: Steven Rostedt diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h index 9208c92..f633478 100644 --- a/include/trace/events/sched.h +++ b/include/trace/events/sched.h @@ -362,6 +362,35 @@ TRACE_EVENT(sched_stat_runtime, (unsigned long long)__entry->vruntime) ); +/* + * Tracepoint for showing priority inheritance modifying a tasks + * priority. + */ +TRACE_EVENT(sched_pi_setprio, + + TP_PROTO(struct task_struct *tsk, int newprio), + + TP_ARGS(tsk, newprio), + + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, oldprio ) + __field( int, newprio ) + ), + + TP_fast_assign( + memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN); + __entry->pid = tsk->pid; + __entry->oldprio = tsk->prio; + __entry->newprio = newprio; + ), + + TP_printk("comm=%s pid=%d oldprio=%d newprio=%d", + __entry->comm, __entry->pid, + __entry->oldprio, __entry->newprio) +); + #endif /* _TRACE_SCHED_H */ /* This part must be outside protection */ diff --git a/kernel/sched.c b/kernel/sched.c index 9ca8ad0..4ad4738 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4355,6 +4355,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) rq = task_rq_lock(p, &flags); + trace_sched_pi_setprio(p, prio); oldprio = p->prio; prev_class = p->sched_class; on_rq = p->se.on_rq; -- cgit v0.10.2 From 234bb549eea16ec7d5207ba747fb8dbf489e64c1 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 2 Sep 2010 13:46:34 +0100 Subject: x86, cleanups: Use clear_page/copy_page rather than memset/memcpy When operating on whole pages, use clear_page() and copy_page() in favor of memset() and memcpy(); after all that's what they are intended for. Signed-off-by: Jan Beulich LKML-Reference: <4C7FB8CA0200007800013F51@vpn.id2.novell.com> Cc: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c index 035c8c5..b3ea9db 100644 --- a/arch/x86/kernel/machine_kexec_64.c +++ b/arch/x86/kernel/machine_kexec_64.c @@ -36,7 +36,7 @@ static int init_one_level2_page(struct kimage *image, pgd_t *pgd, if (!page) goto out; pud = (pud_t *)page_address(page); - memset(pud, 0, PAGE_SIZE); + clear_page(pud); set_pgd(pgd, __pgd(__pa(pud) | _KERNPG_TABLE)); } pud = pud_offset(pgd, addr); @@ -45,7 +45,7 @@ static int init_one_level2_page(struct kimage *image, pgd_t *pgd, if (!page) goto out; pmd = (pmd_t *)page_address(page); - memset(pmd, 0, PAGE_SIZE); + clear_page(pmd); set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE)); } pmd = pmd_offset(pud, addr); diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 77d8c0f..22b06f7 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -1056,14 +1056,13 @@ int kvm_create_lapic(struct kvm_vcpu *vcpu) vcpu->arch.apic = apic; - apic->regs_page = alloc_page(GFP_KERNEL); + apic->regs_page = alloc_page(GFP_KERNEL|__GFP_ZERO); if (apic->regs_page == NULL) { printk(KERN_ERR "malloc apic regs error for vcpu %x\n", vcpu->vcpu_id); goto nomem_free_apic; } apic->regs = page_address(apic->regs_page); - memset(apic->regs, 0, PAGE_SIZE); apic->vcpu = vcpu; hrtimer_init(&apic->lapic_timer.timer, CLOCK_MONOTONIC, diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index bca7909..558f2d3 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -67,7 +67,7 @@ static __init void *alloc_low_page(void) panic("alloc_low_page: ran out of memory"); adr = __va(pfn * PAGE_SIZE); - memset(adr, 0, PAGE_SIZE); + clear_page(adr); return adr; } @@ -558,7 +558,7 @@ char swsusp_pg_dir[PAGE_SIZE] static inline void save_pg_dir(void) { - memcpy(swsusp_pg_dir, swapper_pg_dir, PAGE_SIZE); + copy_page(swsusp_pg_dir, swapper_pg_dir); } #else /* !CONFIG_ACPI_SLEEP */ static inline void save_pg_dir(void) diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 9a66746..7c48ad4 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -293,7 +293,7 @@ static __ref void *alloc_low_page(unsigned long *phys) panic("alloc_low_page: ran out of memory"); adr = early_memremap(pfn * PAGE_SIZE, PAGE_SIZE); - memset(adr, 0, PAGE_SIZE); + clear_page(adr); *phys = pfn * PAGE_SIZE; return adr; } -- cgit v0.10.2 From f57a024ed2ee00d62515ffd4040f5ded3e0eb853 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Thu, 23 Sep 2010 13:41:42 +0100 Subject: GFS2: Remove ignore_local_fs mount argument This is been a no-op for a very long time now. I'm pretty sure nobody uses it, but just in case we'll still accept it on the command line, but ignore it. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index b12ca10..c8a2db1 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -416,7 +416,6 @@ struct gfs2_args { char ar_locktable[GFS2_LOCKNAME_LEN]; /* Name of the Lock Table */ char ar_hostdata[GFS2_LOCKNAME_LEN]; /* Host specific data */ unsigned int ar_spectator:1; /* Don't get a journal */ - unsigned int ar_ignore_local_fs:1; /* Ignore optimisations */ unsigned int ar_localflocks:1; /* Let the VFS do flock|fcntl */ unsigned int ar_localcaching:1; /* Local caching */ unsigned int ar_debug:1; /* Oops on errors */ diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index e031fa4..06a4a7e 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -159,7 +159,7 @@ int gfs2_mount_args(struct gfs2_args *args, char *options) args->ar_spectator = 1; break; case Opt_ignore_local_fs: - args->ar_ignore_local_fs = 1; + /* Retained for backwards compat only */ break; case Opt_localflocks: args->ar_localflocks = 1; @@ -1128,7 +1128,6 @@ static int gfs2_remount_fs(struct super_block *sb, int *flags, char *data) /* Some flags must not be changed */ if (args_neq(&args, &sdp->sd_args, spectator) || - args_neq(&args, &sdp->sd_args, ignore_local_fs) || args_neq(&args, &sdp->sd_args, localflocks) || args_neq(&args, &sdp->sd_args, localcaching) || args_neq(&args, &sdp->sd_args, meta)) @@ -1233,8 +1232,6 @@ static int gfs2_show_options(struct seq_file *s, struct vfsmount *mnt) seq_printf(s, ",hostdata=%s", args->ar_hostdata); if (args->ar_spectator) seq_printf(s, ",spectator"); - if (args->ar_ignore_local_fs) - seq_printf(s, ",ignore_local_fs"); if (args->ar_localflocks) seq_printf(s, ",localflocks"); if (args->ar_localcaching) -- cgit v0.10.2 From c2048b003cfb840ad81bdc6eb55beb12a19a222e Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Thu, 23 Sep 2010 14:00:31 +0100 Subject: GFS2: Remove localcaching mount option This option defaulted to on for lock_nolock mounts and off otherwise. The only function was to avoid the revalidation of dentries. In the cluster case, that is entirely pointless and liable to cause coherency problems. The patch changes the revalidation to depend upon whether the fs is a local or cluster fs (i.e. it follows the existing default behaviour). I very much doubt anybody ever used this option as there is no reason to. Even so we will continue to accept it on the mount command line, but ignore it. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/dentry.c b/fs/gfs2/dentry.c index bb7907b..6798755 100644 --- a/fs/gfs2/dentry.c +++ b/fs/gfs2/dentry.c @@ -49,7 +49,7 @@ static int gfs2_drevalidate(struct dentry *dentry, struct nameidata *nd) ip = GFS2_I(inode); } - if (sdp->sd_args.ar_localcaching) + if (sdp->sd_lockstruct.ls_ops->lm_mount == NULL) goto valid; had_lock = (gfs2_glock_is_locked_by_me(dip->i_gl) != NULL); diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index c8a2db1..2990a0a 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -417,7 +417,6 @@ struct gfs2_args { char ar_hostdata[GFS2_LOCKNAME_LEN]; /* Host specific data */ unsigned int ar_spectator:1; /* Don't get a journal */ unsigned int ar_localflocks:1; /* Let the VFS do flock|fcntl */ - unsigned int ar_localcaching:1; /* Local caching */ unsigned int ar_debug:1; /* Oops on errors */ unsigned int ar_upgrade:1; /* Upgrade ondisk format */ unsigned int ar_posix_acl:1; /* Enable posix acls */ diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 5b5c87d..558bba4 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -1022,7 +1022,6 @@ static int gfs2_lm_mount(struct gfs2_sbd *sdp, int silent) if (!strcmp("lock_nolock", proto)) { lm = &nolock_ops; sdp->sd_args.ar_localflocks = 1; - sdp->sd_args.ar_localcaching = 1; #ifdef CONFIG_GFS2_FS_LOCKING_DLM } else if (!strcmp("lock_dlm", proto)) { lm = &gfs2_dlm_ops; diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index 06a4a7e..e78de8b 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -165,7 +165,7 @@ int gfs2_mount_args(struct gfs2_args *args, char *options) args->ar_localflocks = 1; break; case Opt_localcaching: - args->ar_localcaching = 1; + /* Retained for backwards compat only */ break; case Opt_debug: if (args->ar_errors == GFS2_ERRORS_PANIC) { @@ -1129,7 +1129,6 @@ static int gfs2_remount_fs(struct super_block *sb, int *flags, char *data) /* Some flags must not be changed */ if (args_neq(&args, &sdp->sd_args, spectator) || args_neq(&args, &sdp->sd_args, localflocks) || - args_neq(&args, &sdp->sd_args, localcaching) || args_neq(&args, &sdp->sd_args, meta)) return -EINVAL; @@ -1234,8 +1233,6 @@ static int gfs2_show_options(struct seq_file *s, struct vfsmount *mnt) seq_printf(s, ",spectator"); if (args->ar_localflocks) seq_printf(s, ",localflocks"); - if (args->ar_localcaching) - seq_printf(s, ",localcaching"); if (args->ar_debug) seq_printf(s, ",debug"); if (args->ar_upgrade) -- cgit v0.10.2 From 829f8ed2c963df7c23d1c644db6c4387eb1601fa Mon Sep 17 00:00:00 2001 From: Christian Dietrich Date: Mon, 6 Sep 2010 16:37:12 +0200 Subject: kernel: Remove undead ifdef CONFIG_DEBUG_LOCK_ALLOC The CONFIG_DEBUG_LOCK_ALLOC ifdef isn't necessary at this point, because it is checked in an outer ifdef level already and has no effect here. Signed-off-by: Christian Dietrich Signed-off-by: Paul E. McKenney diff --git a/kernel/srcu.c b/kernel/srcu.c index 2980da3..c71e075 100644 --- a/kernel/srcu.c +++ b/kernel/srcu.c @@ -46,11 +46,9 @@ static int init_srcu_struct_fields(struct srcu_struct *sp) int __init_srcu_struct(struct srcu_struct *sp, const char *name, struct lock_class_key *key) { -#ifdef CONFIG_DEBUG_LOCK_ALLOC /* Don't re-initialize a lock while it is held. */ debug_check_no_locks_freed((void *)sp, sizeof(*sp)); lockdep_init_map(&sp->dep_map, name, key, 0); -#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ return init_srcu_struct_fields(sp); } EXPORT_SYMBOL_GPL(__init_srcu_struct); -- cgit v0.10.2 From 53ecfba259f54b6967a35d19f4a564e3bc07997f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 13 Sep 2010 17:24:21 -0700 Subject: rcu: only one evaluation of arg in rcu_dereference_check() unless sparse The current version of the __rcu_access_pointer(), __rcu_dereference_check(), and __rcu_dereference_protected() macros evaluate their "p" argument three times, not counting typeof()s. This is bad news if that argument contains a side effect. This commit therefore evaluates this argument only once in normal kernel builds. However, the straightforward approach defeats sparse's RCU-pointer checking, so when __CHECKER__ is defined, the additional pair of evaluations of the "p" argument are performed in order to permit sparse to detect misuse of RCU-protected pointers. Signed-off-by: Paul E. McKenney Cc: Arnd Bergmann diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index 89414d6..03cda7b 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -310,24 +310,32 @@ extern int rcu_my_thread_group_empty(void); * (e.g., __rcu_bh, * __rcu_sched, and __srcu), should this make sense in * the future. */ + +#ifdef __CHECKER__ +#define rcu_dereference_sparse(p, space) \ + ((void)(((typeof(*p) space *)p) == p)) +#else /* #ifdef __CHECKER__ */ +#define rcu_dereference_sparse(p, space) +#endif /* #else #ifdef __CHECKER__ */ + #define __rcu_access_pointer(p, space) \ ({ \ typeof(*p) *_________p1 = (typeof(*p)*__force )ACCESS_ONCE(p); \ - (void) (((typeof (*p) space *)p) == p); \ + rcu_dereference_sparse(p, space); \ ((typeof(*p) __force __kernel *)(_________p1)); \ }) #define __rcu_dereference_check(p, c, space) \ ({ \ typeof(*p) *_________p1 = (typeof(*p)*__force )ACCESS_ONCE(p); \ rcu_lockdep_assert(c); \ - (void) (((typeof (*p) space *)p) == p); \ + rcu_dereference_sparse(p, space); \ smp_read_barrier_depends(); \ ((typeof(*p) __force __kernel *)(_________p1)); \ }) #define __rcu_dereference_protected(p, c, space) \ ({ \ rcu_lockdep_assert(c); \ - (void) (((typeof (*p) space *)p) == p); \ + rcu_dereference_sparse(p, space); \ ((typeof(*p) __force __kernel *)(p)); \ }) -- cgit v0.10.2 From 0ddea0ead2ced99eaaaddff2beb755381e5c89f8 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Sun, 19 Sep 2010 21:06:14 -0700 Subject: rcu: fix sparse errors in rcutorture.c Add the sparse __rcu address-space identifier and make a couple of variables static. Signed-off-by: Paul E. McKenney diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 7297102..9d8e8fb 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -120,7 +120,7 @@ struct rcu_torture { }; static LIST_HEAD(rcu_torture_freelist); -static struct rcu_torture *rcu_torture_current; +static struct rcu_torture __rcu *rcu_torture_current; static long rcu_torture_current_version; static struct rcu_torture rcu_tortures[10 * RCU_TORTURE_PIPE_LEN]; static DEFINE_SPINLOCK(rcu_torture_lock); @@ -153,8 +153,10 @@ int rcutorture_runnable = RCUTORTURE_RUNNABLE_INIT; #define FULLSTOP_SHUTDOWN 1 /* System shutdown with rcutorture running. */ #define FULLSTOP_RMMOD 2 /* Normal rmmod of rcutorture. */ static int fullstop = FULLSTOP_RMMOD; -DEFINE_MUTEX(fullstop_mutex); /* Protect fullstop transitions and spawning */ - /* of kthreads. */ +/* + * Protect fullstop transitions and spawning of kthreads. + */ +static DEFINE_MUTEX(fullstop_mutex); /* * Detect and respond to a system shutdown. @@ -737,7 +739,8 @@ rcu_torture_writer(void *arg) continue; rp->rtort_pipe_count = 0; udelay(rcu_random(&rand) & 0x3ff); - old_rp = rcu_torture_current; + old_rp = rcu_dereference_check(rcu_torture_current, + current == writer_task); rp->rtort_mbtest = 1; rcu_assign_pointer(rcu_torture_current, rp); smp_wmb(); /* Mods to old_rp must follow rcu_assign_pointer() */ -- cgit v0.10.2 From 269dcc1c2ec25864308ee03a3fa26ea819d9f5d0 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 7 Sep 2010 14:23:09 -0700 Subject: rcu: Add tracing data to support queueing models The current tracing data is not sufficient to deduce the average time that a callback spends waiting for a grace period to end. Add three per-CPU counters recording the number of callbacks invoked (ci), the number of callbacks orphaned (co), and the number of callbacks adopted (ca). Given the existing callback queue length (ql), the average wait time in absence of CPU hotplug operations is ql/ci. The units of wait time will be in terms of the duration over which ci was measured. In the presence of CPU hotplug operations, there is room for argument, but ql/(ci-co+ca) won't steer you too far wrong. Also fixes a typo called out by Lucas De Marchi . Signed-off-by: Paul E. McKenney diff --git a/Documentation/RCU/trace.txt b/Documentation/RCU/trace.txt index efd8cc9..a851118 100644 --- a/Documentation/RCU/trace.txt +++ b/Documentation/RCU/trace.txt @@ -125,6 +125,17 @@ o "b" is the batch limit for this CPU. If more than this number of RCU callbacks is ready to invoke, then the remainder will be deferred. +o "ci" is the number of RCU callbacks that have been invoked for + this CPU. Note that ci+ql is the number of callbacks that have + been registered in absence of CPU-hotplug activity. + +o "co" is the number of RCU callbacks that have been orphaned due to + this CPU going offline. + +o "ca" is the number of RCU callbacks that have been adopted due to + other CPUs going offline. Note that ci+co-ca+ql is the number of + RCU callbacks registered on this CPU. + There is also an rcu/rcudata.csv file with the same information in comma-separated-variable spreadsheet format. @@ -180,7 +191,7 @@ o "s" is the "signaled" state that drives force_quiescent_state()'s o "jfq" is the number of jiffies remaining for this grace period before force_quiescent_state() is invoked to help push things - along. Note that CPUs in dyntick-idle mode thoughout the grace + along. Note that CPUs in dyntick-idle mode throughout the grace period will not report on their own, but rather must be check by some other CPU via force_quiescent_state(). diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 42140a8..e750735 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1004,6 +1004,7 @@ static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp) for (i = 0; i < RCU_NEXT_SIZE; i++) rdp->nxttail[i] = &rdp->nxtlist; rsp->orphan_qlen += rdp->qlen; + rdp->n_cbs_orphaned += rdp->qlen; rdp->qlen = 0; raw_spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ } @@ -1025,6 +1026,7 @@ static void rcu_adopt_orphan_cbs(struct rcu_state *rsp) *rdp->nxttail[RCU_NEXT_TAIL] = rsp->orphan_cbs_list; rdp->nxttail[RCU_NEXT_TAIL] = rsp->orphan_cbs_tail; rdp->qlen += rsp->orphan_qlen; + rdp->n_cbs_adopted += rsp->orphan_qlen; rsp->orphan_cbs_list = NULL; rsp->orphan_cbs_tail = &rsp->orphan_cbs_list; rsp->orphan_qlen = 0; @@ -1156,6 +1158,7 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) /* Update count, and requeue any remaining callbacks. */ rdp->qlen -= count; + rdp->n_cbs_invoked += count; if (list != NULL) { *tail = rdp->nxtlist; rdp->nxtlist = list; diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 7918ba6..91d4170 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -202,6 +202,9 @@ struct rcu_data { long qlen; /* # of queued callbacks */ long qlen_last_fqs_check; /* qlen at last check for QS forcing */ + unsigned long n_cbs_invoked; /* count of RCU cbs invoked. */ + unsigned long n_cbs_orphaned; /* RCU cbs sent to orphanage. */ + unsigned long n_cbs_adopted; /* RCU cbs adopted from orphanage. */ unsigned long n_force_qs_snap; /* did other CPU force QS recently? */ long blimit; /* Upper limit on a processed batch */ diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c index 458e032..d15430b 100644 --- a/kernel/rcutree_trace.c +++ b/kernel/rcutree_trace.c @@ -64,7 +64,9 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp) rdp->dynticks_fqs); #endif /* #ifdef CONFIG_NO_HZ */ seq_printf(m, " of=%lu ri=%lu", rdp->offline_fqs, rdp->resched_ipi); - seq_printf(m, " ql=%ld b=%ld\n", rdp->qlen, rdp->blimit); + seq_printf(m, " ql=%ld b=%ld", rdp->qlen, rdp->blimit); + seq_printf(m, " ci=%lu co=%lu ca=%lu\n", + rdp->n_cbs_invoked, rdp->n_cbs_orphaned, rdp->n_cbs_adopted); } #define PRINT_RCU_DATA(name, func, m) \ @@ -119,7 +121,9 @@ static void print_one_rcu_data_csv(struct seq_file *m, struct rcu_data *rdp) rdp->dynticks_fqs); #endif /* #ifdef CONFIG_NO_HZ */ seq_printf(m, ",%lu,%lu", rdp->offline_fqs, rdp->resched_ipi); - seq_printf(m, ",%ld,%ld\n", rdp->qlen, rdp->blimit); + seq_printf(m, ",%ld,%ld", rdp->qlen, rdp->blimit); + seq_printf(m, ",%lu,%lu,%lu\n", + rdp->n_cbs_invoked, rdp->n_cbs_orphaned, rdp->n_cbs_adopted); } static int show_rcudata_csv(struct seq_file *m, void *unused) @@ -128,7 +132,7 @@ static int show_rcudata_csv(struct seq_file *m, void *unused) #ifdef CONFIG_NO_HZ seq_puts(m, "\"dt\",\"dt nesting\",\"dn\",\"df\","); #endif /* #ifdef CONFIG_NO_HZ */ - seq_puts(m, "\"of\",\"ri\",\"ql\",\"b\"\n"); + seq_puts(m, "\"of\",\"ri\",\"ql\",\"b\",\"ci\",\"co\",\"ca\"\n"); #ifdef CONFIG_TREE_PREEMPT_RCU seq_puts(m, "\"rcu_preempt:\"\n"); PRINT_RCU_DATA(rcu_preempt_data, print_one_rcu_data_csv, m); -- cgit v0.10.2 From 2dfbf4dfbe47a484bae20456c12b40763b9b6af7 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 15 Sep 2010 23:30:48 -0700 Subject: rcu: Add advice to PROVE_RCU_REPEATEDLY kernel config parameter The PROVE_RCU_REPEATEDLY has no "Say Y"/"Say N" advice, so this commit adds it. Reported-by: Johannes Berg Signed-off-by: Paul E. McKenney diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 52c2172..9886cf5 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -539,6 +539,10 @@ config PROVE_RCU_REPEATEDLY disabling, allowing multiple RCU-lockdep warnings to be printed on a single reboot. + Say Y to allow multiple RCU-lockdep warnings per boot. + + Say N if you are unsure. + config SPARSE_RCU_POINTER bool "RCU debugging: sparse-based checks for pointer usage" default n -- cgit v0.10.2 From d1ea13c6e2cce0106531852daaa93dd97aec9580 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 23 Sep 2010 18:40:07 +0200 Subject: genirq: Cleanup irq_chip->typename leftovers 3 years transition phase is enough. Cleanup the last users and remove the cruft. Signed-off-by: Thomas Gleixner Cc: Leo Chen Cc: Hirokazu Takata Cc: Chris Metcalf Cc: Jeff Dike Cc: Chris Zankel diff --git a/arch/arm/mach-bcmring/irq.c b/arch/arm/mach-bcmring/irq.c index dc1c493..e315263 100644 --- a/arch/arm/mach-bcmring/irq.c +++ b/arch/arm/mach-bcmring/irq.c @@ -67,21 +67,21 @@ static void bcmring_unmask_irq2(unsigned int irq) } static struct irq_chip bcmring_irq0_chip = { - .typename = "ARM-INTC0", + .name = "ARM-INTC0", .ack = bcmring_mask_irq0, .mask = bcmring_mask_irq0, /* mask a specific interrupt, blocking its delivery. */ .unmask = bcmring_unmask_irq0, /* unmaks an interrupt */ }; static struct irq_chip bcmring_irq1_chip = { - .typename = "ARM-INTC1", + .name = "ARM-INTC1", .ack = bcmring_mask_irq1, .mask = bcmring_mask_irq1, .unmask = bcmring_unmask_irq1, }; static struct irq_chip bcmring_irq2_chip = { - .typename = "ARM-SINTC", + .name = "ARM-SINTC", .ack = bcmring_mask_irq2, .mask = bcmring_mask_irq2, .unmask = bcmring_unmask_irq2, diff --git a/arch/m32r/kernel/irq.c b/arch/m32r/kernel/irq.c index 3c71f77..7db26f1 100644 --- a/arch/m32r/kernel/irq.c +++ b/arch/m32r/kernel/irq.c @@ -51,7 +51,7 @@ int show_interrupts(struct seq_file *p, void *v) for_each_online_cpu(j) seq_printf(p, "%10u ", kstat_irqs_cpu(i, j)); #endif - seq_printf(p, " %14s", irq_desc[i].chip->typename); + seq_printf(p, " %14s", irq_desc[i].chip->name); seq_printf(p, " %s", action->name); for (action=action->next; action; action = action->next) diff --git a/arch/m32r/platforms/m32104ut/setup.c b/arch/m32r/platforms/m32104ut/setup.c index 922fdfd..402a59d 100644 --- a/arch/m32r/platforms/m32104ut/setup.c +++ b/arch/m32r/platforms/m32104ut/setup.c @@ -65,7 +65,7 @@ static void shutdown_m32104ut_irq(unsigned int irq) static struct irq_chip m32104ut_irq_type = { - .typename = "M32104UT-IRQ", + .name = "M32104UT-IRQ", .startup = startup_m32104ut_irq, .shutdown = shutdown_m32104ut_irq, .enable = enable_m32104ut_irq, diff --git a/arch/m32r/platforms/m32700ut/setup.c b/arch/m32r/platforms/m32700ut/setup.c index 9c1bc74..80b1a02 100644 --- a/arch/m32r/platforms/m32700ut/setup.c +++ b/arch/m32r/platforms/m32700ut/setup.c @@ -71,7 +71,7 @@ static void shutdown_m32700ut_irq(unsigned int irq) static struct irq_chip m32700ut_irq_type = { - .typename = "M32700UT-IRQ", + .name = "M32700UT-IRQ", .startup = startup_m32700ut_irq, .shutdown = shutdown_m32700ut_irq, .enable = enable_m32700ut_irq, @@ -148,7 +148,7 @@ static void shutdown_m32700ut_pld_irq(unsigned int irq) static struct irq_chip m32700ut_pld_irq_type = { - .typename = "M32700UT-PLD-IRQ", + .name = "M32700UT-PLD-IRQ", .startup = startup_m32700ut_pld_irq, .shutdown = shutdown_m32700ut_pld_irq, .enable = enable_m32700ut_pld_irq, @@ -217,7 +217,7 @@ static void shutdown_m32700ut_lanpld_irq(unsigned int irq) static struct irq_chip m32700ut_lanpld_irq_type = { - .typename = "M32700UT-PLD-LAN-IRQ", + .name = "M32700UT-PLD-LAN-IRQ", .startup = startup_m32700ut_lanpld_irq, .shutdown = shutdown_m32700ut_lanpld_irq, .enable = enable_m32700ut_lanpld_irq, @@ -286,7 +286,7 @@ static void shutdown_m32700ut_lcdpld_irq(unsigned int irq) static struct irq_chip m32700ut_lcdpld_irq_type = { - .typename = "M32700UT-PLD-LCD-IRQ", + .name = "M32700UT-PLD-LCD-IRQ", .startup = startup_m32700ut_lcdpld_irq, .shutdown = shutdown_m32700ut_lcdpld_irq, .enable = enable_m32700ut_lcdpld_irq, diff --git a/arch/m32r/platforms/mappi/setup.c b/arch/m32r/platforms/mappi/setup.c index fb4b177..ea00c84 100644 --- a/arch/m32r/platforms/mappi/setup.c +++ b/arch/m32r/platforms/mappi/setup.c @@ -65,7 +65,7 @@ static void shutdown_mappi_irq(unsigned int irq) static struct irq_chip mappi_irq_type = { - .typename = "MAPPI-IRQ", + .name = "MAPPI-IRQ", .startup = startup_mappi_irq, .shutdown = shutdown_mappi_irq, .enable = enable_mappi_irq, diff --git a/arch/m32r/platforms/mappi2/setup.c b/arch/m32r/platforms/mappi2/setup.c index 6a65eda..c049376 100644 --- a/arch/m32r/platforms/mappi2/setup.c +++ b/arch/m32r/platforms/mappi2/setup.c @@ -72,7 +72,7 @@ static void shutdown_mappi2_irq(unsigned int irq) static struct irq_chip mappi2_irq_type = { - .typename = "MAPPI2-IRQ", + .name = "MAPPI2-IRQ", .startup = startup_mappi2_irq, .shutdown = shutdown_mappi2_irq, .enable = enable_mappi2_irq, diff --git a/arch/m32r/platforms/mappi3/setup.c b/arch/m32r/platforms/mappi3/setup.c index 9c337ae..882de25 100644 --- a/arch/m32r/platforms/mappi3/setup.c +++ b/arch/m32r/platforms/mappi3/setup.c @@ -72,7 +72,7 @@ static void shutdown_mappi3_irq(unsigned int irq) static struct irq_chip mappi3_irq_type = { - .typename = "MAPPI3-IRQ", + .name = "MAPPI3-IRQ", .startup = startup_mappi3_irq, .shutdown = shutdown_mappi3_irq, .enable = enable_mappi3_irq, diff --git a/arch/m32r/platforms/oaks32r/setup.c b/arch/m32r/platforms/oaks32r/setup.c index ed86574..d11d93b 100644 --- a/arch/m32r/platforms/oaks32r/setup.c +++ b/arch/m32r/platforms/oaks32r/setup.c @@ -63,7 +63,7 @@ static void shutdown_oaks32r_irq(unsigned int irq) static struct irq_chip oaks32r_irq_type = { - .typename = "OAKS32R-IRQ", + .name = "OAKS32R-IRQ", .startup = startup_oaks32r_irq, .shutdown = shutdown_oaks32r_irq, .enable = enable_oaks32r_irq, diff --git a/arch/m32r/platforms/opsput/setup.c b/arch/m32r/platforms/opsput/setup.c index 80d6806..5f3402a 100644 --- a/arch/m32r/platforms/opsput/setup.c +++ b/arch/m32r/platforms/opsput/setup.c @@ -72,7 +72,7 @@ static void shutdown_opsput_irq(unsigned int irq) static struct irq_chip opsput_irq_type = { - .typename = "OPSPUT-IRQ", + .name = "OPSPUT-IRQ", .startup = startup_opsput_irq, .shutdown = shutdown_opsput_irq, .enable = enable_opsput_irq, @@ -149,7 +149,7 @@ static void shutdown_opsput_pld_irq(unsigned int irq) static struct irq_chip opsput_pld_irq_type = { - .typename = "OPSPUT-PLD-IRQ", + .name = "OPSPUT-PLD-IRQ", .startup = startup_opsput_pld_irq, .shutdown = shutdown_opsput_pld_irq, .enable = enable_opsput_pld_irq, @@ -218,7 +218,7 @@ static void shutdown_opsput_lanpld_irq(unsigned int irq) static struct irq_chip opsput_lanpld_irq_type = { - .typename = "OPSPUT-PLD-LAN-IRQ", + .name = "OPSPUT-PLD-LAN-IRQ", .startup = startup_opsput_lanpld_irq, .shutdown = shutdown_opsput_lanpld_irq, .enable = enable_opsput_lanpld_irq, diff --git a/arch/m32r/platforms/usrv/setup.c b/arch/m32r/platforms/usrv/setup.c index 7573026..1beac7a 100644 --- a/arch/m32r/platforms/usrv/setup.c +++ b/arch/m32r/platforms/usrv/setup.c @@ -63,7 +63,7 @@ static void shutdown_mappi_irq(unsigned int irq) static struct irq_chip mappi_irq_type = { - .typename = "M32700-IRQ", + .name = "M32700-IRQ", .startup = startup_mappi_irq, .shutdown = shutdown_mappi_irq, .enable = enable_mappi_irq, @@ -136,7 +136,7 @@ static void shutdown_m32700ut_pld_irq(unsigned int irq) static struct irq_chip m32700ut_pld_irq_type = { - .typename = "USRV-PLD-IRQ", + .name = "USRV-PLD-IRQ", .startup = startup_m32700ut_pld_irq, .shutdown = shutdown_m32700ut_pld_irq, .enable = enable_m32700ut_pld_irq, diff --git a/arch/tile/kernel/irq.c b/arch/tile/kernel/irq.c index 596c600..9a27d56 100644 --- a/arch/tile/kernel/irq.c +++ b/arch/tile/kernel/irq.c @@ -208,7 +208,7 @@ static void tile_irq_chip_eoi(unsigned int irq) } static struct irq_chip tile_irq_chip = { - .typename = "tile_irq_chip", + .name = "tile_irq_chip", .ack = tile_irq_chip_ack, .eoi = tile_irq_chip_eoi, .mask = tile_irq_chip_mask, @@ -288,7 +288,7 @@ int show_interrupts(struct seq_file *p, void *v) for_each_online_cpu(j) seq_printf(p, "%10u ", kstat_irqs_cpu(i, j)); #endif - seq_printf(p, " %14s", irq_desc[i].chip->typename); + seq_printf(p, " %14s", irq_desc[i].chip->name); seq_printf(p, " %s", action->name); for (action = action->next; action; action = action->next) diff --git a/arch/um/kernel/irq.c b/arch/um/kernel/irq.c index a3f0b04..a746e30 100644 --- a/arch/um/kernel/irq.c +++ b/arch/um/kernel/irq.c @@ -46,7 +46,7 @@ int show_interrupts(struct seq_file *p, void *v) for_each_online_cpu(j) seq_printf(p, "%10u ", kstat_irqs_cpu(i, j)); #endif - seq_printf(p, " %14s", irq_desc[i].chip->typename); + seq_printf(p, " %14s", irq_desc[i].chip->name); seq_printf(p, " %s", action->name); for (action=action->next; action; action = action->next) @@ -369,7 +369,7 @@ static void dummy(unsigned int irq) /* This is used for everything else than the timer. */ static struct irq_chip normal_irq_type = { - .typename = "SIGIO", + .name = "SIGIO", .release = free_irq_by_irq_and_dev, .disable = dummy, .enable = dummy, @@ -378,7 +378,7 @@ static struct irq_chip normal_irq_type = { }; static struct irq_chip SIGVTALRM_irq_type = { - .typename = "SIGVTALRM", + .name = "SIGVTALRM", .release = free_irq_by_irq_and_dev, .shutdown = dummy, /* never called */ .disable = dummy, diff --git a/arch/xtensa/kernel/irq.c b/arch/xtensa/kernel/irq.c index c64a5d3..8750888 100644 --- a/arch/xtensa/kernel/irq.c +++ b/arch/xtensa/kernel/irq.c @@ -92,7 +92,7 @@ int show_interrupts(struct seq_file *p, void *v) for_each_online_cpu(j) seq_printf(p, "%10u ", kstat_irqs_cpu(i, j)); #endif - seq_printf(p, " %14s", irq_desc[i].chip->typename); + seq_printf(p, " %14s", irq_desc[i].chip->name); seq_printf(p, " %s", action->name); for (action=action->next; action; action = action->next) diff --git a/include/linux/irq.h b/include/linux/irq.h index c03243a..06273a2 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -106,7 +106,6 @@ struct msi_desc; * @bus_sync_unlock: function to sync and unlock slow bus (i2c) chips * * @release: release function solely used by UML - * @typename: obsoleted by name, kept as migration helper */ struct irq_chip { const char *name; @@ -135,11 +134,6 @@ struct irq_chip { #ifdef CONFIG_IRQ_RELEASE_METHOD void (*release)(unsigned int irq, void *dev_id); #endif - /* - * For compatibility, ->typename is copied into ->name. - * Will disappear. - */ - const char *typename; }; struct timer_rand_state; diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index b7091d5..4ea775c 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -344,8 +344,6 @@ void irq_chip_set_defaults(struct irq_chip *chip) if (!chip->shutdown) chip->shutdown = chip->disable != default_disable ? chip->disable : default_shutdown; - if (!chip->name) - chip->name = chip->typename; if (!chip->end) chip->end = dummy_irq_chip.end; } -- cgit v0.10.2 From 76fb657017588a0912f0d1d140cb807446e4ef05 Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Thu, 23 Sep 2010 17:28:04 +0100 Subject: x86, olpc: Only enable PCI configuration type override on XO-1 This configuration type override is for XO-1 only and must not happen on XO-1.5. Signed-off-by: Daniel Drake LKML-Reference: <20100923162805.0F6549D401B@zog.reactivated.net> Cc: Andres Solomon Cc: Grant Likely Signed-off-by: H. Peter Anvin diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..0ed4c9b 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1900,7 +1900,7 @@ config PCI_GODIRECT bool "Direct" config PCI_GOOLPC - bool "OLPC" + bool "OLPC XO-1" depends on OLPC config PCI_GOANY diff --git a/arch/x86/kernel/olpc.c b/arch/x86/kernel/olpc.c index 0e0cdde..635888c 100644 --- a/arch/x86/kernel/olpc.c +++ b/arch/x86/kernel/olpc.c @@ -242,8 +242,10 @@ static int __init olpc_init(void) (unsigned char *) &olpc_platform_info.ecver, 1); #ifdef CONFIG_PCI_OLPC - /* If the VSA exists let it emulate PCI, if not emulate in kernel */ - if (!cs5535_has_vsa2()) + /* If the VSA exists let it emulate PCI, if not emulate in kernel. + * XO-1 only. */ + if (olpc_platform_info.boardrev < olpc_board_pre(0xd0) && + !cs5535_has_vsa2()) x86_init.pci.arch_init = pci_olpc_init; #endif diff --git a/arch/x86/pci/olpc.c b/arch/x86/pci/olpc.c index b348154..13700ec 100644 --- a/arch/x86/pci/olpc.c +++ b/arch/x86/pci/olpc.c @@ -304,7 +304,7 @@ static struct pci_raw_ops pci_olpc_conf = { int __init pci_olpc_init(void) { - printk(KERN_INFO "PCI: Using configuration type OLPC\n"); + printk(KERN_INFO "PCI: Using configuration type OLPC XO-1\n"); raw_pci_ops = &pci_olpc_conf; is_lx = is_geode_lx(); return 0; -- cgit v0.10.2 From 3e3c486012a3dbb113c0ca15ee265d309d77aea9 Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Thu, 23 Sep 2010 17:28:46 +0100 Subject: x86, olpc: Rework BIOS signature check The XO-1.5 laptop is not currently detected as an OLPC machine because it fails this XO-1-centric check. Now that we have OLPC OFW support in the kernel, a more sensible check is to see if we found OFW during boot and check the architecture property. Also remove a now-meaningless codepath, as we're always going to have OFW support with OLPC. Signed-off-by: Daniel Drake LKML-Reference: <20100923162846.D8D409D401B@zog.reactivated.net> Cc: Andres Salomon Cc: Grant Likely Signed-off-by: H. Peter Anvin diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 0ed4c9b..c255255 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -2061,6 +2061,7 @@ config SCx200HR_TIMER config OLPC bool "One Laptop Per Child support" select GPIOLIB + select OLPC_OPENFIRMWARE ---help--- Add support for detecting the unique features of the OLPC XO hardware. @@ -2068,7 +2069,7 @@ config OLPC config OLPC_OPENFIRMWARE bool "Support for OLPC's Open Firmware" depends on !X86_64 && !X86_PAE - default y if OLPC + default n help This option adds support for the implementation of Open Firmware that is used on the OLPC XO-1 Children's Machine. diff --git a/arch/x86/include/asm/olpc_ofw.h b/arch/x86/include/asm/olpc_ofw.h index 08fde47..2a84781 100644 --- a/arch/x86/include/asm/olpc_ofw.h +++ b/arch/x86/include/asm/olpc_ofw.h @@ -21,10 +21,14 @@ extern void olpc_ofw_detect(void); /* install OFW's pde permanently into the kernel's pgtable */ extern void setup_olpc_ofw_pgd(void); +/* check if OFW was detected during boot */ +extern bool olpc_ofw_present(void); + #else /* !CONFIG_OLPC_OPENFIRMWARE */ static inline void olpc_ofw_detect(void) { } static inline void setup_olpc_ofw_pgd(void) { } +static inline bool olpc_ofw_present(void) { return false; } #endif /* !CONFIG_OLPC_OPENFIRMWARE */ diff --git a/arch/x86/kernel/olpc.c b/arch/x86/kernel/olpc.c index 635888c..37c49934 100644 --- a/arch/x86/kernel/olpc.c +++ b/arch/x86/kernel/olpc.c @@ -183,8 +183,21 @@ err: } EXPORT_SYMBOL_GPL(olpc_ec_cmd); -#ifdef CONFIG_OLPC_OPENFIRMWARE -static void __init platform_detect(void) +static bool __init check_ofw_architecture(void) +{ + size_t propsize; + char olpc_arch[5]; + const void *args[] = { NULL, "architecture", olpc_arch, (void *)5 }; + void *res[] = { &propsize }; + + if (olpc_ofw("getprop", args, res)) { + printk(KERN_ERR "ofw: getprop call failed!\n"); + return false; + } + return propsize == 5 && strncmp("OLPC", olpc_arch, 5) == 0; +} + +static u32 __init get_board_revision(void) { size_t propsize; __be32 rev; @@ -193,46 +206,27 @@ static void __init platform_detect(void) if (olpc_ofw("getprop", args, res) || propsize != 4) { printk(KERN_ERR "ofw: getprop call failed!\n"); - rev = cpu_to_be32(0); + return cpu_to_be32(0); } - olpc_platform_info.boardrev = be32_to_cpu(rev); + return be32_to_cpu(rev); } -#else -static void __init platform_detect(void) + +static bool __init platform_detect(void) { - /* stopgap until OFW support is added to the kernel */ - olpc_platform_info.boardrev = olpc_board(0xc2); + if (!check_ofw_architecture()) + return false; + olpc_platform_info.flags |= OLPC_F_PRESENT; + olpc_platform_info.boardrev = get_board_revision(); + return true; } -#endif static int __init olpc_init(void) { - unsigned char *romsig; - - /* The ioremap check is dangerous; limit what we run it on */ - if (!is_geode() || cs5535_has_vsa2()) + if (!olpc_ofw_present() || !platform_detect()) return 0; spin_lock_init(&ec_lock); - romsig = ioremap(0xffffffc0, 16); - if (!romsig) - return 0; - - if (strncmp(romsig, "CL1 Q", 7)) - goto unmap; - if (strncmp(romsig+6, romsig+13, 3)) { - printk(KERN_INFO "OLPC BIOS signature looks invalid. " - "Assuming not OLPC\n"); - goto unmap; - } - - printk(KERN_INFO "OLPC board with OpenFirmware %.16s\n", romsig); - olpc_platform_info.flags |= OLPC_F_PRESENT; - - /* get the platform revision */ - platform_detect(); - /* assume B1 and above models always have a DCON */ if (olpc_board_at_least(olpc_board(0xb1))) olpc_platform_info.flags |= OLPC_F_DCON; @@ -254,8 +248,6 @@ static int __init olpc_init(void) olpc_platform_info.boardrev >> 4, olpc_platform_info.ecver); -unmap: - iounmap(romsig); return 0; } diff --git a/arch/x86/kernel/olpc_ofw.c b/arch/x86/kernel/olpc_ofw.c index 3218aa7..7873204 100644 --- a/arch/x86/kernel/olpc_ofw.c +++ b/arch/x86/kernel/olpc_ofw.c @@ -74,6 +74,12 @@ int __olpc_ofw(const char *name, int nr_args, const void **args, int nr_res, } EXPORT_SYMBOL_GPL(__olpc_ofw); +bool olpc_ofw_present(void) +{ + return olpc_ofw_cif != NULL; +} +EXPORT_SYMBOL_GPL(olpc_ofw_present); + /* OFW cif _should_ be above this address */ #define OFW_MIN 0xff000000 -- cgit v0.10.2 From c80dbb58f9c3f84372e37dfe80b41810052ad62f Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 24 Sep 2010 09:55:07 +0100 Subject: GFS2: Remove upgrade mount option This option has never done anything useful. Also at the same time this cleans up the sb checks which are done at mount time. The debug option will be accepted, but ignored in future. Since it didn't do anything, there didn't seem much point in retaining it. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 2990a0a..6f6ff8a 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -418,7 +418,6 @@ struct gfs2_args { unsigned int ar_spectator:1; /* Don't get a journal */ unsigned int ar_localflocks:1; /* Let the VFS do flock|fcntl */ unsigned int ar_debug:1; /* Oops on errors */ - unsigned int ar_upgrade:1; /* Upgrade ondisk format */ unsigned int ar_posix_acl:1; /* Enable posix acls */ unsigned int ar_quota:2; /* off/account/on */ unsigned int ar_suiddir:1; /* suiddir support */ diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 558bba4..5b11f37 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -38,14 +38,6 @@ #define DO 0 #define UNDO 1 -static const u32 gfs2_old_fs_formats[] = { - 0 -}; - -static const u32 gfs2_old_multihost_formats[] = { - 0 -}; - /** * gfs2_tune_init - Fill a gfs2_tune structure with default values * @gt: tune @@ -135,8 +127,6 @@ static struct gfs2_sbd *init_sbd(struct super_block *sb) static int gfs2_check_sb(struct gfs2_sbd *sdp, struct gfs2_sb_host *sb, int silent) { - unsigned int x; - if (sb->sb_magic != GFS2_MAGIC || sb->sb_type != GFS2_METATYPE_SB) { if (!silent) @@ -150,55 +140,9 @@ static int gfs2_check_sb(struct gfs2_sbd *sdp, struct gfs2_sb_host *sb, int sile sb->sb_multihost_format == GFS2_FORMAT_MULTI) return 0; - if (sb->sb_fs_format != GFS2_FORMAT_FS) { - for (x = 0; gfs2_old_fs_formats[x]; x++) - if (gfs2_old_fs_formats[x] == sb->sb_fs_format) - break; - - if (!gfs2_old_fs_formats[x]) { - printk(KERN_WARNING - "GFS2: code version (%u, %u) is incompatible " - "with ondisk format (%u, %u)\n", - GFS2_FORMAT_FS, GFS2_FORMAT_MULTI, - sb->sb_fs_format, sb->sb_multihost_format); - printk(KERN_WARNING - "GFS2: I don't know how to upgrade this FS\n"); - return -EINVAL; - } - } - - if (sb->sb_multihost_format != GFS2_FORMAT_MULTI) { - for (x = 0; gfs2_old_multihost_formats[x]; x++) - if (gfs2_old_multihost_formats[x] == - sb->sb_multihost_format) - break; - - if (!gfs2_old_multihost_formats[x]) { - printk(KERN_WARNING - "GFS2: code version (%u, %u) is incompatible " - "with ondisk format (%u, %u)\n", - GFS2_FORMAT_FS, GFS2_FORMAT_MULTI, - sb->sb_fs_format, sb->sb_multihost_format); - printk(KERN_WARNING - "GFS2: I don't know how to upgrade this FS\n"); - return -EINVAL; - } - } + fs_warn(sdp, "Unknown on-disk format, unable to mount\n"); - if (!sdp->sd_args.ar_upgrade) { - printk(KERN_WARNING - "GFS2: code version (%u, %u) is incompatible " - "with ondisk format (%u, %u)\n", - GFS2_FORMAT_FS, GFS2_FORMAT_MULTI, - sb->sb_fs_format, sb->sb_multihost_format); - printk(KERN_INFO - "GFS2: Use the \"upgrade\" mount option to upgrade " - "the FS\n"); - printk(KERN_INFO "GFS2: See the manual for more details\n"); - return -EINVAL; - } - - return 0; + return -EINVAL; } static void end_bio_io_page(struct bio *bio, int error) diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index e78de8b..d85e0b7 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -179,7 +179,7 @@ int gfs2_mount_args(struct gfs2_args *args, char *options) args->ar_debug = 0; break; case Opt_upgrade: - args->ar_upgrade = 1; + /* Retained for backwards compat only */ break; case Opt_acl: args->ar_posix_acl = 1; @@ -1235,8 +1235,6 @@ static int gfs2_show_options(struct seq_file *s, struct vfsmount *mnt) seq_printf(s, ",localflocks"); if (args->ar_debug) seq_printf(s, ",debug"); - if (args->ar_upgrade) - seq_printf(s, ",upgrade"); if (args->ar_posix_acl) seq_printf(s, ",acl"); if (args->ar_quota != GFS2_QUOTA_DEFAULT) { -- cgit v0.10.2 From 3b4b682becdfa9f42321aa024d5cc84f71f06d8c Mon Sep 17 00:00:00 2001 From: Ma Ling Date: Fri, 17 Sep 2010 03:12:40 +0800 Subject: x86, mem: Optimize memmove for small size and unaligned cases movs instruction will combine data to accelerate moving data, however we need to concern two cases about it. 1. movs instruction need long lantency to startup, so here we use general mov instruction to copy data. 2. movs instruction is not good for unaligned case, even if src offset is 0x10, dest offset is 0x0, we avoid and handle the case by general mov instruction. Signed-off-by: Ma Ling LKML-Reference: <1284664360-6138-1-git-send-email-ling.ma@intel.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/lib/memcpy_32.c b/arch/x86/lib/memcpy_32.c index 81130d4..b908a59 100644 --- a/arch/x86/lib/memcpy_32.c +++ b/arch/x86/lib/memcpy_32.c @@ -22,36 +22,187 @@ EXPORT_SYMBOL(memset); void *memmove(void *dest, const void *src, size_t n) { - int d0, d1, d2; - - if (dest < src) { - if ((dest + n) < src) - return memcpy(dest, src, n); - else - __asm__ __volatile__( - "rep\n\t" - "movsb\n\t" - : "=&c" (d0), "=&S" (d1), "=&D" (d2) - :"0" (n), - "1" (src), - "2" (dest) - :"memory"); - } else { - if((src + n) < dest) - return memcpy(dest, src, n); - else - __asm__ __volatile__( - "std\n\t" - "rep\n\t" - "movsb\n\t" - "cld" - : "=&c" (d0), "=&S" (d1), "=&D" (d2) - :"0" (n), - "1" (n-1+src), - "2" (n-1+dest) - :"memory"); - } - - return dest; + int d0,d1,d2,d3,d4,d5; + char *ret = dest; + + __asm__ __volatile__( + /* Handle more 16bytes in loop */ + "cmp $0x10, %0\n\t" + "jb 1f\n\t" + + /* Decide forward/backward copy mode */ + "cmp %2, %1\n\t" + "jb 2f\n\t" + + /* + * movs instruction have many startup latency + * so we handle small size by general register. + */ + "cmp $680, %0\n\t" + "jb 3f\n\t" + /* + * movs instruction is only good for aligned case. + */ + "mov %1, %3\n\t" + "xor %2, %3\n\t" + "and $0xff, %3\n\t" + "jz 4f\n\t" + "3:\n\t" + "sub $0x10, %0\n\t" + + /* + * We gobble 16byts forward in each loop. + */ + "3:\n\t" + "sub $0x10, %0\n\t" + "mov 0*4(%1), %3\n\t" + "mov 1*4(%1), %4\n\t" + "mov %3, 0*4(%2)\n\t" + "mov %4, 1*4(%2)\n\t" + "mov 2*4(%1), %3\n\t" + "mov 3*4(%1), %4\n\t" + "mov %3, 2*4(%2)\n\t" + "mov %4, 3*4(%2)\n\t" + "lea 0x10(%1), %1\n\t" + "lea 0x10(%2), %2\n\t" + "jae 3b\n\t" + "add $0x10, %0\n\t" + "jmp 1f\n\t" + + /* + * Handle data forward by movs. + */ + ".p2align 4\n\t" + "4:\n\t" + "mov -4(%1, %0), %3\n\t" + "lea -4(%2, %0), %4\n\t" + "shr $2, %0\n\t" + "rep movsl\n\t" + "mov %3, (%4)\n\t" + "jmp 11f\n\t" + /* + * Handle data backward by movs. + */ + ".p2align 4\n\t" + "6:\n\t" + "mov (%1), %3\n\t" + "mov %2, %4\n\t" + "lea -4(%1, %0), %1\n\t" + "lea -4(%2, %0), %2\n\t" + "shr $2, %0\n\t" + "std\n\t" + "rep movsl\n\t" + "mov %3,(%4)\n\t" + "cld\n\t" + "jmp 11f\n\t" + + /* + * Start to prepare for backward copy. + */ + ".p2align 4\n\t" + "2:\n\t" + "cmp $680, %0\n\t" + "jb 5f\n\t" + "mov %1, %3\n\t" + "xor %2, %3\n\t" + "and $0xff, %3\n\t" + "jz 6b\n\t" + + /* + * Calculate copy position to tail. + */ + "5:\n\t" + "add %0, %1\n\t" + "add %0, %2\n\t" + "sub $0x10, %0\n\t" + + /* + * We gobble 16byts backward in each loop. + */ + "7:\n\t" + "sub $0x10, %0\n\t" + + "mov -1*4(%1), %3\n\t" + "mov -2*4(%1), %4\n\t" + "mov %3, -1*4(%2)\n\t" + "mov %4, -2*4(%2)\n\t" + "mov -3*4(%1), %3\n\t" + "mov -4*4(%1), %4\n\t" + "mov %3, -3*4(%2)\n\t" + "mov %4, -4*4(%2)\n\t" + "lea -0x10(%1), %1\n\t" + "lea -0x10(%2), %2\n\t" + "jae 7b\n\t" + /* + * Calculate copy position to head. + */ + "add $0x10, %0\n\t" + "sub %0, %1\n\t" + "sub %0, %2\n\t" + + /* + * Move data from 8 bytes to 15 bytes. + */ + ".p2align 4\n\t" + "1:\n\t" + "cmp $8, %0\n\t" + "jb 8f\n\t" + "mov 0*4(%1), %3\n\t" + "mov 1*4(%1), %4\n\t" + "mov -2*4(%1, %0), %5\n\t" + "mov -1*4(%1, %0), %1\n\t" + + "mov %3, 0*4(%2)\n\t" + "mov %4, 1*4(%2)\n\t" + "mov %5, -2*4(%2, %0)\n\t" + "mov %1, -1*4(%2, %0)\n\t" + "jmp 11f\n\t" + + /* + * Move data from 4 bytes to 7 bytes. + */ + ".p2align 4\n\t" + "8:\n\t" + "cmp $4, %0\n\t" + "jb 9f\n\t" + "mov 0*4(%1), %3\n\t" + "mov -1*4(%1, %0), %4\n\t" + "mov %3, 0*4(%2)\n\t" + "mov %4, -1*4(%2, %0)\n\t" + "jmp 11f\n\t" + + /* + * Move data from 2 bytes to 3 bytes. + */ + ".p2align 4\n\t" + "9:\n\t" + "cmp $2, %0\n\t" + "jb 10f\n\t" + "movw 0*2(%1), %%dx\n\t" + "movw -1*2(%1, %0), %%bx\n\t" + "movw %%dx, 0*2(%2)\n\t" + "movw %%bx, -1*2(%2, %0)\n\t" + "jmp 11f\n\t" + + /* + * Move data for 1 byte. + */ + ".p2align 4\n\t" + "10:\n\t" + "cmp $1, %0\n\t" + "jb 11f\n\t" + "movb (%1), %%cl\n\t" + "movb %%cl, (%2)\n\t" + ".p2align 4\n\t" + "11:" + : "=&c" (d0), "=&S" (d1), "=&D" (d2), + "=r" (d3),"=r" (d4), "=r"(d5) + :"0" (n), + "1" (src), + "2" (dest) + :"memory"); + + return ret; + } EXPORT_SYMBOL(memmove); diff --git a/arch/x86/lib/memmove_64.c b/arch/x86/lib/memmove_64.c index ecacc4b..6d0f0ec 100644 --- a/arch/x86/lib/memmove_64.c +++ b/arch/x86/lib/memmove_64.c @@ -8,50 +8,185 @@ #undef memmove void *memmove(void *dest, const void *src, size_t count) { - unsigned long d0, d1, d2, d3; - if (dest < src) { - if ((dest + count) < src) - return memcpy(dest, src, count); - else - __asm__ __volatile__( - "movq %0, %3\n\t" - "shr $3, %0\n\t" - "andq $7, %3\n\t" - "rep\n\t" - "movsq\n\t" - "movq %3, %0\n\t" - "rep\n\t" - "movsb" - : "=&c" (d0), "=&S" (d1), "=&D" (d2), "=r" (d3) - :"0" (count), - "1" (src), - "2" (dest) - :"memory"); - } else { - if((src + count) < dest) - return memcpy(dest, src, count); - else - __asm__ __volatile__( - "movq %0, %3\n\t" - "lea -8(%1, %0), %1\n\t" - "lea -8(%2, %0), %2\n\t" - "shr $3, %0\n\t" - "andq $7, %3\n\t" - "std\n\t" - "rep\n\t" - "movsq\n\t" - "lea 7(%1), %1\n\t" - "lea 7(%2), %2\n\t" - "movq %3, %0\n\t" - "rep\n\t" - "movsb\n\t" - "cld" - : "=&c" (d0), "=&S" (d1), "=&D" (d2), "=r" (d3) - :"0" (count), - "1" (src), - "2" (dest) - :"memory"); - } - return dest; + unsigned long d0,d1,d2,d3,d4,d5,d6,d7; + char *ret; + + __asm__ __volatile__( + /* Handle more 32bytes in loop */ + "mov %2, %3\n\t" + "cmp $0x20, %0\n\t" + "jb 1f\n\t" + + /* Decide forward/backward copy mode */ + "cmp %2, %1\n\t" + "jb 2f\n\t" + + /* + * movsq instruction have many startup latency + * so we handle small size by general register. + */ + "cmp $680, %0\n\t" + "jb 3f\n\t" + /* + * movsq instruction is only good for aligned case. + */ + "cmpb %%dil, %%sil\n\t" + "je 4f\n\t" + "3:\n\t" + "sub $0x20, %0\n\t" + /* + * We gobble 32byts forward in each loop. + */ + "5:\n\t" + "sub $0x20, %0\n\t" + "movq 0*8(%1), %4\n\t" + "movq 1*8(%1), %5\n\t" + "movq 2*8(%1), %6\n\t" + "movq 3*8(%1), %7\n\t" + "leaq 4*8(%1), %1\n\t" + + "movq %4, 0*8(%2)\n\t" + "movq %5, 1*8(%2)\n\t" + "movq %6, 2*8(%2)\n\t" + "movq %7, 3*8(%2)\n\t" + "leaq 4*8(%2), %2\n\t" + "jae 5b\n\t" + "addq $0x20, %0\n\t" + "jmp 1f\n\t" + /* + * Handle data forward by movsq. + */ + ".p2align 4\n\t" + "4:\n\t" + "movq %0, %8\n\t" + "movq -8(%1, %0), %4\n\t" + "lea -8(%2, %0), %5\n\t" + "shrq $3, %8\n\t" + "rep movsq\n\t" + "movq %4, (%5)\n\t" + "jmp 13f\n\t" + /* + * Handle data backward by movsq. + */ + ".p2align 4\n\t" + "7:\n\t" + "movq %0, %8\n\t" + "movq (%1), %4\n\t" + "movq %2, %5\n\t" + "leaq -8(%1, %0), %1\n\t" + "leaq -8(%2, %0), %2\n\t" + "shrq $3, %8\n\t" + "std\n\t" + "rep movsq\n\t" + "cld\n\t" + "movq %4, (%5)\n\t" + "jmp 13f\n\t" + + /* + * Start to prepare for backward copy. + */ + ".p2align 4\n\t" + "2:\n\t" + "cmp $680, %0\n\t" + "jb 6f \n\t" + "cmp %%dil, %%sil\n\t" + "je 7b \n\t" + "6:\n\t" + /* + * Calculate copy position to tail. + */ + "addq %0, %1\n\t" + "addq %0, %2\n\t" + "subq $0x20, %0\n\t" + /* + * We gobble 32byts backward in each loop. + */ + "8:\n\t" + "subq $0x20, %0\n\t" + "movq -1*8(%1), %4\n\t" + "movq -2*8(%1), %5\n\t" + "movq -3*8(%1), %6\n\t" + "movq -4*8(%1), %7\n\t" + "leaq -4*8(%1), %1\n\t" + + "movq %4, -1*8(%2)\n\t" + "movq %5, -2*8(%2)\n\t" + "movq %6, -3*8(%2)\n\t" + "movq %7, -4*8(%2)\n\t" + "leaq -4*8(%2), %2\n\t" + "jae 8b\n\t" + /* + * Calculate copy position to head. + */ + "addq $0x20, %0\n\t" + "subq %0, %1\n\t" + "subq %0, %2\n\t" + "1:\n\t" + "cmpq $16, %0\n\t" + "jb 9f\n\t" + /* + * Move data from 16 bytes to 31 bytes. + */ + "movq 0*8(%1), %4\n\t" + "movq 1*8(%1), %5\n\t" + "movq -2*8(%1, %0), %6\n\t" + "movq -1*8(%1, %0), %7\n\t" + "movq %4, 0*8(%2)\n\t" + "movq %5, 1*8(%2)\n\t" + "movq %6, -2*8(%2, %0)\n\t" + "movq %7, -1*8(%2, %0)\n\t" + "jmp 13f\n\t" + ".p2align 4\n\t" + "9:\n\t" + "cmpq $8, %0\n\t" + "jb 10f\n\t" + /* + * Move data from 8 bytes to 15 bytes. + */ + "movq 0*8(%1), %4\n\t" + "movq -1*8(%1, %0), %5\n\t" + "movq %4, 0*8(%2)\n\t" + "movq %5, -1*8(%2, %0)\n\t" + "jmp 13f\n\t" + "10:\n\t" + "cmpq $4, %0\n\t" + "jb 11f\n\t" + /* + * Move data from 4 bytes to 7 bytes. + */ + "movl (%1), %4d\n\t" + "movl -4(%1, %0), %5d\n\t" + "movl %4d, (%2)\n\t" + "movl %5d, -4(%2, %0)\n\t" + "jmp 13f\n\t" + "11:\n\t" + "cmp $2, %0\n\t" + "jb 12f\n\t" + /* + * Move data from 2 bytes to 3 bytes. + */ + "movw (%1), %4w\n\t" + "movw -2(%1, %0), %5w\n\t" + "movw %4w, (%2)\n\t" + "movw %5w, -2(%2, %0)\n\t" + "jmp 13f\n\t" + "12:\n\t" + "cmp $1, %0\n\t" + "jb 13f\n\t" + /* + * Move data for 1 byte. + */ + "movb (%1), %4b\n\t" + "movb %4b, (%2)\n\t" + "13:\n\t" + : "=&d" (d0), "=&S" (d1), "=&D" (d2), "=&a" (ret) , + "=r"(d3), "=r"(d4), "=r"(d5), "=r"(d6), "=&c" (d7) + :"0" (count), + "1" (src), + "2" (dest) + :"memory"); + + return ret; + } EXPORT_SYMBOL(memmove); -- cgit v0.10.2 From 4a109cc0511209276e58f1dbe53259e7a8f1508b Mon Sep 17 00:00:00 2001 From: Mark Brown Date: Fri, 24 Sep 2010 10:50:46 +0100 Subject: ASoC: Add Jassi Brar as Samsung maintainer Since Jassi is doing so much good work and can offer such good review for the Samsung CPU support add an entry for the Samsung ASoC drivers to MAINTAINERS listing him. Signed-off-by: Mark Brown Acked-by: Jassi Brar Acked-by: Liam Girdwood diff --git a/MAINTAINERS b/MAINTAINERS index b5b8baa..f51409b 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4965,6 +4965,12 @@ F: drivers/media/common/saa7146* F: drivers/media/video/*7146* F: include/media/*7146* +SAMSUNG AUDIO (ASoC) DRIVERS +M: Jassi Brar +L: alsa-devel@alsa-project.org (moderated for non-subscribers) +S: Supported +F: sound/soc/s3c24xx + TLG2300 VIDEO4LINUX-2 DRIVER M: Huang Shijie M: Kang Yong -- cgit v0.10.2 From d0795f912318f65b800c6b619d749c3bf7c930fb Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 27 Sep 2010 15:58:11 +0100 Subject: GFS2: Fix journal check for spectator mounts When checking journals for spectator mounts, we cannot rely on the journal being locked, whatever its jid might be. This patch ensures that we always get the journal locks when checking journals for a spectator mount. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/recovery.c b/fs/gfs2/recovery.c index f7f89a9..666548e 100644 --- a/fs/gfs2/recovery.c +++ b/fs/gfs2/recovery.c @@ -456,7 +456,8 @@ void gfs2_recover_func(struct work_struct *work) unsigned int pass; int error; - if (jd->jd_jid != sdp->sd_lockstruct.ls_jid) { + if (sdp->sd_args.ar_spectator || + (jd->jd_jid != sdp->sd_lockstruct.ls_jid)) { fs_info(sdp, "jid=%u: Trying to acquire journal lock...\n", jd->jd_jid); -- cgit v0.10.2 From fb0c484f72df9e336605b5659f39eb1ab5e8d946 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Thu, 19 Aug 2010 14:42:14 -0300 Subject: V4L/DVB: tm6000: depends on IR_CORE tm6000 uses IR interfaces, so it should depend on IR_CORE. ERROR: "get_rc_map" [drivers/staging/tm6000/tm6000.ko] undefined! ERROR: "ir_input_unregister" [drivers/staging/tm6000/tm6000.ko] undefined! ERROR: "__ir_input_register" [drivers/staging/tm6000/tm6000.ko] undefined! Signed-off-by: Randy Dunlap Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/staging/tm6000/Kconfig b/drivers/staging/tm6000/Kconfig index c725356..de7ebb9 100644 --- a/drivers/staging/tm6000/Kconfig +++ b/drivers/staging/tm6000/Kconfig @@ -1,6 +1,6 @@ config VIDEO_TM6000 tristate "TV Master TM5600/6000/6010 driver" - depends on VIDEO_DEV && I2C && INPUT && USB && EXPERIMENTAL + depends on VIDEO_DEV && I2C && INPUT && IR_CORE && USB && EXPERIMENTAL select VIDEO_TUNER select MEDIA_TUNER_XC2028 select MEDIA_TUNER_XC5000 -- cgit v0.10.2 From 0f63a14d357b3a80abb5203a574cbb3f8988c2f7 Mon Sep 17 00:00:00 2001 From: lawrence rust Date: Tue, 24 Aug 2010 06:50:48 -0300 Subject: V4L/DVB: cx88: Kconfig: Remove EXPERIMENTAL dependency from VIDEO_CX88_ALSA The cx88-alsa module has been around since January 2006 and has seen no significant changes since September 2007. It is stable in operation and so I believe that the 'experimental' tag is no longer warranted. Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/cx88/Kconfig b/drivers/media/video/cx88/Kconfig index 99dbae1..0fa85cb 100644 --- a/drivers/media/video/cx88/Kconfig +++ b/drivers/media/video/cx88/Kconfig @@ -17,7 +17,7 @@ config VIDEO_CX88 config VIDEO_CX88_ALSA tristate "Conexant 2388x DMA audio support" - depends on VIDEO_CX88 && SND && EXPERIMENTAL + depends on VIDEO_CX88 && SND select SND_PCM ---help--- This is a video4linux driver for direct (DMA) audio on -- cgit v0.10.2 From 590a58d18027b18db78f538aca592c2cdb249079 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 12 Aug 2010 04:41:58 -0300 Subject: V4L/DVB: unlock on error path If we return directly here then we miss out on some mutex_unlock()s Signed-off-by: Dan Carpenter Acked-by: Sylwester Nawrocki Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/s5p-fimc/fimc-core.c b/drivers/media/video/s5p-fimc/fimc-core.c index b151c7b..1beb226 100644 --- a/drivers/media/video/s5p-fimc/fimc-core.c +++ b/drivers/media/video/s5p-fimc/fimc-core.c @@ -822,7 +822,8 @@ static int fimc_m2m_s_fmt(struct file *file, void *priv, struct v4l2_format *f) } else { v4l2_err(&ctx->fimc_dev->m2m.v4l2_dev, "Wrong buffer/video queue type (%d)\n", f->type); - return -EINVAL; + ret = -EINVAL; + goto s_fmt_out; } pix = &f->fmt.pix; -- cgit v0.10.2 From 028816bc85db5ac6a562c2aff2113c7480d67919 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 12 Aug 2010 04:47:07 -0300 Subject: V4L/DVB: IR: ir-raw-event: null pointer dereference The original code dereferenced ir->raw after freeing it and setting it to NULL. Signed-off-by: Dan Carpenter Acked-by: Jarod Wilson Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/IR/ir-raw-event.c b/drivers/media/IR/ir-raw-event.c index 43094e7..8e0e1b1 100644 --- a/drivers/media/IR/ir-raw-event.c +++ b/drivers/media/IR/ir-raw-event.c @@ -279,9 +279,11 @@ int ir_raw_event_register(struct input_dev *input_dev) "rc%u", (unsigned int)ir->devno); if (IS_ERR(ir->raw->thread)) { + int ret = PTR_ERR(ir->raw->thread); + kfree(ir->raw); ir->raw = NULL; - return PTR_ERR(ir->raw->thread); + return ret; } mutex_lock(&ir_raw_handler_lock); -- cgit v0.10.2 From a8e07124500184f81541ddf3e9669000af9ac4bc Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 19 Aug 2010 06:47:50 -0300 Subject: V4L/DVB: opera1: remove unneeded NULL check "fw" is always a non-NULL pointer at this point, and anyway release_firmware() accepts NULL pointers. Signed-off-by: Dan Carpenter Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/dvb/dvb-usb/opera1.c b/drivers/media/dvb/dvb-usb/opera1.c index 6b22ec6..f896337 100644 --- a/drivers/media/dvb/dvb-usb/opera1.c +++ b/drivers/media/dvb/dvb-usb/opera1.c @@ -483,9 +483,7 @@ static int opera1_xilinx_load_firmware(struct usb_device *dev, } } kfree(p); - if (fw) { - release_firmware(fw); - } + release_firmware(fw); return ret; } -- cgit v0.10.2 From d5337966ce4639c775ff5edf92d78f5fad34ef0d Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 19 Aug 2010 06:50:04 -0300 Subject: V4L/DVB: pvrusb2: remove unneeded NULL checks We dereference "maskptr" unconditionally at the start of the function and also inside the call to parse_tlist() towards the end of the function. This function is called from store_val_any() and it always passes a non-NULL pointer. Signed-off-by: Dan Carpenter Acked-by: Mike Isely Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/pvrusb2/pvrusb2-ctrl.c b/drivers/media/video/pvrusb2/pvrusb2-ctrl.c index 1b992b8..55ea914 100644 --- a/drivers/media/video/pvrusb2/pvrusb2-ctrl.c +++ b/drivers/media/video/pvrusb2/pvrusb2-ctrl.c @@ -513,7 +513,7 @@ int pvr2_ctrl_sym_to_value(struct pvr2_ctrl *cptr, if (ret >= 0) { ret = pvr2_ctrl_range_check(cptr,*valptr); } - if (maskptr) *maskptr = ~0; + *maskptr = ~0; } else if (cptr->info->type == pvr2_ctl_bool) { ret = parse_token(ptr,len,valptr,boolNames, ARRAY_SIZE(boolNames)); @@ -522,7 +522,7 @@ int pvr2_ctrl_sym_to_value(struct pvr2_ctrl *cptr, } else if (ret == 0) { *valptr = (*valptr & 1) ? !0 : 0; } - if (maskptr) *maskptr = 1; + *maskptr = 1; } else if (cptr->info->type == pvr2_ctl_enum) { ret = parse_token( ptr,len,valptr, @@ -531,7 +531,7 @@ int pvr2_ctrl_sym_to_value(struct pvr2_ctrl *cptr, if (ret >= 0) { ret = pvr2_ctrl_range_check(cptr,*valptr); } - if (maskptr) *maskptr = ~0; + *maskptr = ~0; } else if (cptr->info->type == pvr2_ctl_bitmask) { ret = parse_tlist( ptr,len,maskptr,valptr, -- cgit v0.10.2 From 23e64d55859ea619c17832645e0e56faac84d750 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 19 Aug 2010 07:00:22 -0300 Subject: V4L/DVB: saa7164: move dereference under NULL check The original code dereferenced "port" before checking it for NULL. I moved the test down below the check. Also I changed the comparisons a little so people wouldn't get confused and think "port" and "buf" were ints instead of pointers. (Probably that's what lead to this issue in the first place.) There is only one caller for this function and it passes non-NULL pointers, so this is essentially a cleanup rather than a bugfix. Signed-off-by: Dan Carpenter Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/saa7164/saa7164-buffer.c b/drivers/media/video/saa7164/saa7164-buffer.c index 5713f3a..ddd25d3 100644 --- a/drivers/media/video/saa7164/saa7164-buffer.c +++ b/drivers/media/video/saa7164/saa7164-buffer.c @@ -136,10 +136,11 @@ ret: int saa7164_buffer_dealloc(struct saa7164_tsport *port, struct saa7164_buffer *buf) { - struct saa7164_dev *dev = port->dev; + struct saa7164_dev *dev; - if ((buf == 0) || (port == 0)) + if (!buf || !port) return SAA_ERR_BAD_PARAMETER; + dev = port->dev; dprintk(DBGLVL_BUF, "%s() deallocating buffer @ 0x%p\n", __func__, buf); -- cgit v0.10.2 From 29d834912bbcf0d095e6cd0e95b0d9790fdf7957 Mon Sep 17 00:00:00 2001 From: Dmitri Belimov Date: Mon, 23 Aug 2010 10:30:14 -0300 Subject: V4L/DVB: Fix regression for BeholdTV Columbus Some time a go our customers wrote me about problem with our TV card BeholdTV Columbus. It's PCMCIA TV card for notebook. As I understand v4l has some regression with autodetect address of tuners. I can set incorrect I2C address and had report about detect tuner. No any TV of course. When I set correct tuner type and I2C address of the tuners all works well. Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/saa7134/saa7134-cards.c b/drivers/media/video/saa7134/saa7134-cards.c index ec697fc..bb8d83d 100644 --- a/drivers/media/video/saa7134/saa7134-cards.c +++ b/drivers/media/video/saa7134/saa7134-cards.c @@ -4323,13 +4323,13 @@ struct saa7134_board saa7134_boards[] = { }, [SAA7134_BOARD_BEHOLD_COLUMBUS_TVFM] = { /* Beholder Intl. Ltd. 2008 */ - /*Dmitry Belimov */ - .name = "Beholder BeholdTV Columbus TVFM", + /* Dmitry Belimov */ + .name = "Beholder BeholdTV Columbus TV/FM", .audio_clock = 0x00187de7, .tuner_type = TUNER_ALPS_TSBE5_PAL, - .radio_type = UNSET, - .tuner_addr = ADDR_UNSET, - .radio_addr = ADDR_UNSET, + .radio_type = TUNER_TEA5767, + .tuner_addr = 0xc2 >> 1, + .radio_addr = 0xc0 >> 1, .tda9887_conf = TDA9887_PRESENT, .gpiomask = 0x000A8004, .inputs = {{ -- cgit v0.10.2 From 882787ff8fdeb0be790547ee9b22b281095e95da Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Fri, 3 Sep 2010 06:57:19 -0300 Subject: V4L/DVB: gspca - main: Fix a crash of some webcams on ARM arch MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When plugging some webcams on ARM, the system crashes. This is because we alloc buffer for an urb through usb_buffer_alloc, the alloced buffer is already in DMA coherent region, so we should set the flag of this urb to URB_NO_TRANSFER_DMA_MAP, otherwise when we submit this urb, the hcd core will handle this address as an non-DMA address and call dma_map_single/sg to map it. On arm architecture, dma_map_single a DMA coherent address will be catched by a BUG_ON(). Signed-off-by: Jason Wang Signed-off-by: Jean-François Moine Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/gspca/gspca.c b/drivers/media/video/gspca/gspca.c index b984610..78abc1c 100644 --- a/drivers/media/video/gspca/gspca.c +++ b/drivers/media/video/gspca/gspca.c @@ -223,6 +223,7 @@ static int alloc_and_submit_int_urb(struct gspca_dev *gspca_dev, usb_rcvintpipe(dev, ep->bEndpointAddress), buffer, buffer_len, int_irq, (void *)gspca_dev, interval); + urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP; gspca_dev->int_urb = urb; ret = usb_submit_urb(urb, GFP_KERNEL); if (ret < 0) { -- cgit v0.10.2 From c13df9cf6b1cad7fd088c9acceb98b6bffd9ca31 Mon Sep 17 00:00:00 2001 From: Jarod Wilson Date: Fri, 27 Aug 2010 18:21:14 -0300 Subject: V4L/DVB: mceusb: add two new ASUS device IDs Reported in lirc sf.net tracker and on lirc mailing list Signed-off-by: Jarod Wilson Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/IR/mceusb.c b/drivers/media/IR/mceusb.c index ac6bb2c..bc620e1 100644 --- a/drivers/media/IR/mceusb.c +++ b/drivers/media/IR/mceusb.c @@ -120,6 +120,10 @@ static struct usb_device_id mceusb_dev_table[] = { { USB_DEVICE(VENDOR_PHILIPS, 0x0613) }, /* Philips eHome Infrared Transceiver */ { USB_DEVICE(VENDOR_PHILIPS, 0x0815) }, + /* Philips/Spinel plus IR transceiver for ASUS */ + { USB_DEVICE(VENDOR_PHILIPS, 0x206c) }, + /* Philips/Spinel plus IR transceiver for ASUS */ + { USB_DEVICE(VENDOR_PHILIPS, 0x2088) }, /* Realtek MCE IR Receiver */ { USB_DEVICE(VENDOR_REALTEK, 0x0161) }, /* SMK/Toshiba G83C0004D410 */ -- cgit v0.10.2 From 3bfb317f97cfddbbec67bbe8e35ad38af3507397 Mon Sep 17 00:00:00 2001 From: Mauro Carvalho Chehab Date: Fri, 3 Sep 2010 10:50:24 -0300 Subject: V4L/DVB: Don't identify PV SBTVD Hybrid as a DibCom device As reported by Carlos, Prolink Pixelview SBTVD Hybrid is based on Conexant cx231xx + Fujitsu 86A20S demodulator. However, both shares the same USB ID. So, we need to use USB bcdDevice, in order to properly discover what's the board. We know for sure that bcd 0x100 is used for a dib0700 device, while bcd 0x4001 is used for a cx23102 device. This patch reserves two ranges, the first one from 0x0000-0x3f00 for dib0700, and the second from 0x4000-0x4fff for cx231xx devices. This may need fixes in the future, as we get access to other devices. Thanks-to: Carlos Americo Domiciano Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/dvb/dvb-usb/dib0700_devices.c b/drivers/media/dvb/dvb-usb/dib0700_devices.c index f634d2e..ce66c5a 100644 --- a/drivers/media/dvb/dvb-usb/dib0700_devices.c +++ b/drivers/media/dvb/dvb-usb/dib0700_devices.c @@ -1781,7 +1781,7 @@ struct usb_device_id dib0700_usb_id_table[] = { /* 60 */{ USB_DEVICE(USB_VID_TERRATEC, USB_PID_TERRATEC_CINERGY_T_XXS_2) }, { USB_DEVICE(USB_VID_DIBCOM, USB_PID_DIBCOM_STK807XPVR) }, { USB_DEVICE(USB_VID_DIBCOM, USB_PID_DIBCOM_STK807XP) }, - { USB_DEVICE(USB_VID_PIXELVIEW, USB_PID_PIXELVIEW_SBTVD) }, + { USB_DEVICE_VER(USB_VID_PIXELVIEW, USB_PID_PIXELVIEW_SBTVD, 0x000, 0x3f00) }, { USB_DEVICE(USB_VID_EVOLUTEPC, USB_PID_TVWAY_PLUS) }, /* 65 */{ USB_DEVICE(USB_VID_PINNACLE, USB_PID_PINNACLE_PCTV73ESE) }, { USB_DEVICE(USB_VID_PINNACLE, USB_PID_PINNACLE_PCTV282E) }, diff --git a/drivers/media/video/cx231xx/Makefile b/drivers/media/video/cx231xx/Makefile index 755dd0c..6f2b573 100644 --- a/drivers/media/video/cx231xx/Makefile +++ b/drivers/media/video/cx231xx/Makefile @@ -11,4 +11,5 @@ EXTRA_CFLAGS += -Idrivers/media/video EXTRA_CFLAGS += -Idrivers/media/common/tuners EXTRA_CFLAGS += -Idrivers/media/dvb/dvb-core EXTRA_CFLAGS += -Idrivers/media/dvb/frontends +EXTRA_CFLAGS += -Idrivers/media/dvb/dvb-usb diff --git a/drivers/media/video/cx231xx/cx231xx-cards.c b/drivers/media/video/cx231xx/cx231xx-cards.c index 6bdc0ef..5b7c9a9 100644 --- a/drivers/media/video/cx231xx/cx231xx-cards.c +++ b/drivers/media/video/cx231xx/cx231xx-cards.c @@ -32,6 +32,7 @@ #include #include +#include "dvb-usb-ids.h" #include "xc5000.h" #include "cx231xx.h" @@ -175,6 +176,8 @@ struct usb_device_id cx231xx_id_table[] = { .driver_info = CX231XX_BOARD_CNXT_RDE_250}, {USB_DEVICE(0x0572, 0x58A1), .driver_info = CX231XX_BOARD_CNXT_RDU_250}, + {USB_DEVICE_VER(USB_VID_PIXELVIEW, USB_PID_PIXELVIEW_SBTVD, 0x4000,0x4fff), + .driver_info = CX231XX_BOARD_UNKNOWN}, {}, }; -- cgit v0.10.2 From 04cab131ce2a267b6777a98d68fbc0cae44d4ba8 Mon Sep 17 00:00:00 2001 From: Mauro Carvalho Chehab Date: Wed, 8 Sep 2010 12:58:12 -0300 Subject: V4L/DVB: rc-core: increase repeat time As reported by Anton Blanchard , double IR events on 2.6.36-rc2 and a DViCO FusionHDTV DVB-T Dual Express are happening: [ 1351.032084] ir_keydown: i2c IR (FusionHDTV): key down event, key 0x0067, scancode 0x0051 [ 1351.281284] ir_keyup: keyup key 0x0067 ie one key down event and one key up event 250ms later. So, we need to increase the repeat timeout, to avoid this bug to hit. As we're doing it at core, this fix is not needed anymore at dib0700 driver. Thanks-to: Anton Blanchard Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/IR/ir-keytable.c b/drivers/media/IR/ir-keytable.c index 7e82a9d..d00ef19 100644 --- a/drivers/media/IR/ir-keytable.c +++ b/drivers/media/IR/ir-keytable.c @@ -510,6 +510,13 @@ int __ir_input_register(struct input_dev *input_dev, (ir_dev->props && ir_dev->props->driver_type == RC_DRIVER_IR_RAW) ? " in raw mode" : ""); + /* + * Default delay of 250ms is too short for some protocols, expecially + * since the timeout is currently set to 250ms. Increase it to 500ms, + * to avoid wrong repetition of the keycodes. + */ + input_dev->rep[REP_DELAY] = 500; + return 0; out_event: diff --git a/drivers/media/dvb/dvb-usb/dib0700_core.c b/drivers/media/dvb/dvb-usb/dib0700_core.c index fe81834..48397f1 100644 --- a/drivers/media/dvb/dvb-usb/dib0700_core.c +++ b/drivers/media/dvb/dvb-usb/dib0700_core.c @@ -673,9 +673,6 @@ static int dib0700_probe(struct usb_interface *intf, else dev->props.rc.core.bulk_mode = false; - /* Need a higher delay, to avoid wrong repeat */ - dev->rc_input_dev->rep[REP_DELAY] = 500; - dib0700_rc_setup(dev); return 0; -- cgit v0.10.2 From 67332ba83242f814490884161fee8676a9fd8a09 Mon Sep 17 00:00:00 2001 From: Maxim Levitsky Date: Mon, 6 Sep 2010 18:26:08 -0300 Subject: V4L/DVB: IR: fix duty cycle capability Due to typo lirc bridge enabled wrong capability. Signed-off-by: Maxim Levitsky Acked-by: Jarod Wilson Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/IR/ir-lirc-codec.c b/drivers/media/IR/ir-lirc-codec.c index 77b5946..e63f757 100644 --- a/drivers/media/IR/ir-lirc-codec.c +++ b/drivers/media/IR/ir-lirc-codec.c @@ -267,7 +267,7 @@ static int ir_lirc_register(struct input_dev *input_dev) features |= LIRC_CAN_SET_SEND_CARRIER; if (ir_dev->props->s_tx_duty_cycle) - features |= LIRC_CAN_SET_REC_DUTY_CYCLE; + features |= LIRC_CAN_SET_SEND_DUTY_CYCLE; } if (ir_dev->props->s_rx_carrier_range) -- cgit v0.10.2 From e0172fd373ab77a83ea952fd6a75c612e1b0bf9e Mon Sep 17 00:00:00 2001 From: Maxim Levitsky Date: Mon, 6 Sep 2010 18:26:09 -0300 Subject: V4L/DVB: IR: fix keys beeing stuck down forever The logic in ir_timer_keyup was inverted. In case that values aren't equal, the meaning of the time_is_after_eq_jiffies(ir->keyup_jiffies) is that ir->keyup_jiffies is after the the jiffies or equally that that jiffies are before the the ir->keyup_jiffies which is exactly the situation we want to avoid (that the timeout is in the future) Confusing Eh? Signed-off-by: Maxim Levitsky Acked-by: Jarod Wilson Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/IR/ir-keytable.c b/drivers/media/IR/ir-keytable.c index d00ef19..7961d59 100644 --- a/drivers/media/IR/ir-keytable.c +++ b/drivers/media/IR/ir-keytable.c @@ -319,7 +319,7 @@ static void ir_timer_keyup(unsigned long cookie) * a keyup event might follow immediately after the keydown. */ spin_lock_irqsave(&ir->keylock, flags); - if (time_is_after_eq_jiffies(ir->keyup_jiffies)) + if (time_is_before_eq_jiffies(ir->keyup_jiffies)) ir_keyup(ir); spin_unlock_irqrestore(&ir->keylock, flags); } -- cgit v0.10.2 From 00a220aa98133dc43c6f7016c218aaf3afd66e11 Mon Sep 17 00:00:00 2001 From: Maxim Levitsky Date: Mon, 6 Sep 2010 18:26:10 -0300 Subject: V4L/DVB: IR: extend MCE keymap These keys are found on remote bundled with Toshiba Qosmio F50-10q. Found and tested by, Sami R Signed-off-by: Maxim Levitsky Acked-by: Jarod Wilson Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/IR/keymaps/rc-rc6-mce.c b/drivers/media/IR/keymaps/rc-rc6-mce.c index 64264f7..39557ad 100644 --- a/drivers/media/IR/keymaps/rc-rc6-mce.c +++ b/drivers/media/IR/keymaps/rc-rc6-mce.c @@ -19,6 +19,7 @@ static struct ir_scancode rc6_mce[] = { { 0x800f0416, KEY_PLAY }, { 0x800f0418, KEY_PAUSE }, + { 0x800f046e, KEY_PLAYPAUSE }, { 0x800f0419, KEY_STOP }, { 0x800f0417, KEY_RECORD }, @@ -37,6 +38,8 @@ static struct ir_scancode rc6_mce[] = { { 0x800f0411, KEY_VOLUMEDOWN }, { 0x800f0412, KEY_CHANNELUP }, { 0x800f0413, KEY_CHANNELDOWN }, + { 0x800f043a, KEY_BRIGHTNESSUP }, + { 0x800f0480, KEY_BRIGHTNESSDOWN }, { 0x800f0401, KEY_NUMERIC_1 }, { 0x800f0402, KEY_NUMERIC_2 }, -- cgit v0.10.2 From 90e12cec707204930934acdb5efce5f94a163a5f Mon Sep 17 00:00:00 2001 From: Olivier Grenie Date: Tue, 7 Sep 2010 12:50:45 -0300 Subject: V4L/DVB: dib7770: enable the current mirror To improve performance on DiB7770-devices enabling the current mirror is needed. This patch adds an option to the dib7000p-driver to do that and it creates a separate device-entry in dib0700-device to use those changes on hardware which is using the DiB7770. Signed-off-by: Olivier Grenie Signed-off-by: Patrick Boettcher Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/dvb/dvb-usb/dib0700_devices.c b/drivers/media/dvb/dvb-usb/dib0700_devices.c index ce66c5a..385ce1c 100644 --- a/drivers/media/dvb/dvb-usb/dib0700_devices.c +++ b/drivers/media/dvb/dvb-usb/dib0700_devices.c @@ -940,6 +940,57 @@ static int stk7070p_frontend_attach(struct dvb_usb_adapter *adap) return adap->fe == NULL ? -ENODEV : 0; } +/* STK7770P */ +static struct dib7000p_config dib7770p_dib7000p_config = { + .output_mpeg2_in_188_bytes = 1, + + .agc_config_count = 1, + .agc = &dib7070_agc_config, + .bw = &dib7070_bw_config_12_mhz, + .tuner_is_baseband = 1, + .spur_protect = 1, + + .gpio_dir = DIB7000P_GPIO_DEFAULT_DIRECTIONS, + .gpio_val = DIB7000P_GPIO_DEFAULT_VALUES, + .gpio_pwm_pos = DIB7000P_GPIO_DEFAULT_PWM_POS, + + .hostbus_diversity = 1, + .enable_current_mirror = 1, +}; + +static int stk7770p_frontend_attach(struct dvb_usb_adapter *adap) +{ + struct usb_device_descriptor *p = &adap->dev->udev->descriptor; + if (p->idVendor == cpu_to_le16(USB_VID_PINNACLE) && + p->idProduct == cpu_to_le16(USB_PID_PINNACLE_PCTV72E)) + dib0700_set_gpio(adap->dev, GPIO6, GPIO_OUT, 0); + else + dib0700_set_gpio(adap->dev, GPIO6, GPIO_OUT, 1); + msleep(10); + dib0700_set_gpio(adap->dev, GPIO9, GPIO_OUT, 1); + dib0700_set_gpio(adap->dev, GPIO4, GPIO_OUT, 1); + dib0700_set_gpio(adap->dev, GPIO7, GPIO_OUT, 1); + dib0700_set_gpio(adap->dev, GPIO10, GPIO_OUT, 0); + + dib0700_ctrl_clock(adap->dev, 72, 1); + + msleep(10); + dib0700_set_gpio(adap->dev, GPIO10, GPIO_OUT, 1); + msleep(10); + dib0700_set_gpio(adap->dev, GPIO0, GPIO_OUT, 1); + + if (dib7000p_i2c_enumeration(&adap->dev->i2c_adap, 1, 18, + &dib7770p_dib7000p_config) != 0) { + err("%s: dib7000p_i2c_enumeration failed. Cannot continue\n", + __func__); + return -ENODEV; + } + + adap->fe = dvb_attach(dib7000p_attach, &adap->dev->i2c_adap, 0x80, + &dib7770p_dib7000p_config); + return adap->fe == NULL ? -ENODEV : 0; +} + /* DIB807x generic */ static struct dibx000_agc_config dib807x_agc_config[2] = { { @@ -2406,7 +2457,7 @@ struct dvb_usb_device_properties dib0700_devices[] = { .pid_filter_count = 32, .pid_filter = stk70x0p_pid_filter, .pid_filter_ctrl = stk70x0p_pid_filter_ctrl, - .frontend_attach = stk7070p_frontend_attach, + .frontend_attach = stk7770p_frontend_attach, .tuner_attach = dib7770p_tuner_attach, DIB0700_DEFAULT_STREAMING_CONFIG(0x02), diff --git a/drivers/media/dvb/frontends/dib7000p.c b/drivers/media/dvb/frontends/dib7000p.c index 2e28b97..73f59ab 100644 --- a/drivers/media/dvb/frontends/dib7000p.c +++ b/drivers/media/dvb/frontends/dib7000p.c @@ -260,6 +260,8 @@ static void dib7000p_set_adc_state(struct dib7000p_state *state, enum dibx000_ad // dprintk( "908: %x, 909: %x\n", reg_908, reg_909); + reg_908 |= (state->cfg.enable_current_mirror & 1) << 7; + dib7000p_write_word(state, 908, reg_908); dib7000p_write_word(state, 909, reg_909); } diff --git a/drivers/media/dvb/frontends/dib7000p.h b/drivers/media/dvb/frontends/dib7000p.h index 805dd13..04a7449 100644 --- a/drivers/media/dvb/frontends/dib7000p.h +++ b/drivers/media/dvb/frontends/dib7000p.h @@ -33,6 +33,9 @@ struct dib7000p_config { int (*agc_control) (struct dvb_frontend *, u8 before); u8 output_mode; + + u8 enable_current_mirror : 1; + }; #define DEFAULT_DIB7000P_I2C_ADDRESS 18 -- cgit v0.10.2 From 970d14c6cca8c71307a4d23fe373c5895175b2d7 Mon Sep 17 00:00:00 2001 From: Olivier Grenie Date: Tue, 7 Sep 2010 12:50:46 -0300 Subject: V4L/DVB: dib7000p: add disable sample and hold, and diversity delay parameter This patch improves the overall driver performance in diversity-reception scenarios. Signed-off-by: Olivier Grenie Signed-off-by: Patrick Boettcher Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/dvb/dvb-usb/dib0700_devices.c b/drivers/media/dvb/dvb-usb/dib0700_devices.c index 385ce1c..e06acd1 100644 --- a/drivers/media/dvb/dvb-usb/dib0700_devices.c +++ b/drivers/media/dvb/dvb-usb/dib0700_devices.c @@ -956,6 +956,7 @@ static struct dib7000p_config dib7770p_dib7000p_config = { .hostbus_diversity = 1, .enable_current_mirror = 1, + .disable_sample_and_hold = 0, }; static int stk7770p_frontend_attach(struct dvb_usb_adapter *adap) diff --git a/drivers/media/dvb/frontends/dib7000p.c b/drivers/media/dvb/frontends/dib7000p.c index 73f59ab..3aed0d4 100644 --- a/drivers/media/dvb/frontends/dib7000p.c +++ b/drivers/media/dvb/frontends/dib7000p.c @@ -260,6 +260,7 @@ static void dib7000p_set_adc_state(struct dib7000p_state *state, enum dibx000_ad // dprintk( "908: %x, 909: %x\n", reg_908, reg_909); + reg_909 |= (state->cfg.disable_sample_and_hold & 1) << 4; reg_908 |= (state->cfg.enable_current_mirror & 1) << 7; dib7000p_write_word(state, 908, reg_908); @@ -780,7 +781,10 @@ static void dib7000p_set_channel(struct dib7000p_state *state, struct dvb_fronte default: case GUARD_INTERVAL_1_32: value *= 1; break; } - state->div_sync_wait = (value * 3) / 2 + 32; // add 50% SFN margin + compensate for one DVSY-fifo TODO + if (state->cfg.diversity_delay == 0) + state->div_sync_wait = (value * 3) / 2 + 48; // add 50% SFN margin + compensate for one DVSY-fifo + else + state->div_sync_wait = (value * 3) / 2 + state->cfg.diversity_delay; // add 50% SFN margin + compensate for one DVSY-fifo /* deactive the possibility of diversity reception if extended interleaver */ state->div_force_off = !1 && ch->u.ofdm.transmission_mode != TRANSMISSION_MODE_8K; diff --git a/drivers/media/dvb/frontends/dib7000p.h b/drivers/media/dvb/frontends/dib7000p.h index 04a7449..da17345 100644 --- a/drivers/media/dvb/frontends/dib7000p.h +++ b/drivers/media/dvb/frontends/dib7000p.h @@ -33,8 +33,10 @@ struct dib7000p_config { int (*agc_control) (struct dvb_frontend *, u8 before); u8 output_mode; + u8 disable_sample_and_hold : 1; u8 enable_current_mirror : 1; + u8 diversity_delay; }; -- cgit v0.10.2 From 3cdadc50bbe8f04c1231c8af614cafd7ddd622bf Mon Sep 17 00:00:00 2001 From: Richard Zidlicky Date: Tue, 24 Aug 2010 09:52:36 -0300 Subject: V4L/DVB: dvb: fix smscore_getbuffer() logic Drivers shouldn't sleep while holding a spinlock. A previous workaround were to release the spinlock before callinc schedule(). This patch uses a different approach: it just waits for the siano hardware to answer. Signed-off-by: Richard Zidlicky Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/dvb/siano/smscoreapi.c b/drivers/media/dvb/siano/smscoreapi.c index d93468c..ff3b0fa 100644 --- a/drivers/media/dvb/siano/smscoreapi.c +++ b/drivers/media/dvb/siano/smscoreapi.c @@ -1098,33 +1098,26 @@ EXPORT_SYMBOL_GPL(smscore_onresponse); * * @return pointer to descriptor on success, NULL on error. */ -struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) + +struct smscore_buffer_t *get_entry(struct smscore_device_t *coredev) { struct smscore_buffer_t *cb = NULL; unsigned long flags; - DEFINE_WAIT(wait); - spin_lock_irqsave(&coredev->bufferslock, flags); - - /* This function must return a valid buffer, since the buffer list is - * finite, we check that there is an available buffer, if not, we wait - * until such buffer become available. - */ - - prepare_to_wait(&coredev->buffer_mng_waitq, &wait, TASK_INTERRUPTIBLE); - if (list_empty(&coredev->buffers)) { - spin_unlock_irqrestore(&coredev->bufferslock, flags); - schedule(); - spin_lock_irqsave(&coredev->bufferslock, flags); + if (!list_empty(&coredev->buffers)) { + cb = (struct smscore_buffer_t *) coredev->buffers.next; + list_del(&cb->entry); } + spin_unlock_irqrestore(&coredev->bufferslock, flags); + return cb; +} - finish_wait(&coredev->buffer_mng_waitq, &wait); - - cb = (struct smscore_buffer_t *) coredev->buffers.next; - list_del(&cb->entry); +struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) +{ + struct smscore_buffer_t *cb = NULL; - spin_unlock_irqrestore(&coredev->bufferslock, flags); + wait_event(coredev->buffer_mng_waitq, (cb = get_entry(coredev))); return cb; } -- cgit v0.10.2 From c10469c637602c2385e2993d8c730cc44fd47d23 Mon Sep 17 00:00:00 2001 From: Mauro Carvalho Chehab Date: Sat, 11 Sep 2010 11:37:51 -0300 Subject: V4L/DVB: cx231xx: Avoid an OOPS when card is unknown (card=0) As reported by: Carlos Americo Domiciano : [ 220.033500] cx231xx v4l2 driver loaded. [ 220.033571] cx231xx #0: New device Conexant Corporation Polaris AV Capturb @ 480 Mbps (1554:5010) with 6 interfaces [ 220.033577] cx231xx #0: registering interface 0 [ 220.033591] cx231xx #0: registering interface 1 [ 220.033654] cx231xx #0: registering interface 6 [ 220.033910] cx231xx #0: Identified as Unknown CX231xx video grabber (card=0) [ 220.033946] BUG: unable to handle kernel NULL pointer dereference at (null) [ 220.033955] IP: [] cx231xx_pre_card_setup+0x5d/0xb0 [cx231xx] Thanks-to: Carlos Americo Domiciano Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/cx231xx/cx231xx-cards.c b/drivers/media/video/cx231xx/cx231xx-cards.c index 5b7c9a9..f2a4900 100644 --- a/drivers/media/video/cx231xx/cx231xx-cards.c +++ b/drivers/media/video/cx231xx/cx231xx-cards.c @@ -229,14 +229,16 @@ void cx231xx_pre_card_setup(struct cx231xx *dev) dev->board.name, dev->model); /* set the direction for GPIO pins */ - cx231xx_set_gpio_direction(dev, dev->board.tuner_gpio->bit, 1); - cx231xx_set_gpio_value(dev, dev->board.tuner_gpio->bit, 1); - cx231xx_set_gpio_direction(dev, dev->board.tuner_sif_gpio, 1); + if (dev->board.tuner_gpio) { + cx231xx_set_gpio_direction(dev, dev->board.tuner_gpio->bit, 1); + cx231xx_set_gpio_value(dev, dev->board.tuner_gpio->bit, 1); + cx231xx_set_gpio_direction(dev, dev->board.tuner_sif_gpio, 1); - /* request some modules if any required */ + /* request some modules if any required */ - /* reset the Tuner */ - cx231xx_gpio_set(dev, dev->board.tuner_gpio); + /* reset the Tuner */ + cx231xx_gpio_set(dev, dev->board.tuner_gpio); + } /* set the mode to Analog mode initially */ cx231xx_set_mode(dev, CX231XX_ANALOG_MODE); -- cgit v0.10.2 From 2fc11536cf5c0b8eb4eb7e01a2a672a189e9280f Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Tue, 7 Sep 2010 06:10:45 -0300 Subject: V4L/DVB: videobuf-dma-sg: set correct size in last sg element This fixes a nasty memory corruption bug when using userptr I/O. The function videobuf_pages_to_sg() sets up the scatter-gather list for the DMA transfer to the userspace pages. The first transfer is setup correctly (the size is set to PAGE_SIZE - offset), but all other transfers have size PAGE_SIZE. This is wrong for the last transfer which may be less than PAGE_SIZE. Most, if not all, drivers will program the boards DMA engine correctly, i.e. even though the size in the last sg element is wrong, they will do their own size calculations and make sure the right amount is DMA-ed, and so seemingly prevent memory corruption. However, behind the scenes the dynamic DMA mapping support (in lib/swiotlb.c) may create bounce buffers if the memory pages are not in DMA-able memory. This happens for example on a 64-bit linux with a board that only supports 32-bit DMA. These bounce buffers DO use the information in the sg list to determine the size. So while the DMA engine transfers the correct amount of data, when the data is 'bounced' back too much is copied, causing buffer overwrites. The fix is simple: calculate and set the correct size for the last sg list element. Signed-off-by: Hans Verkuil Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/videobuf-dma-sg.c b/drivers/media/video/videobuf-dma-sg.c index 06f9a9c..2ad0bc2 100644 --- a/drivers/media/video/videobuf-dma-sg.c +++ b/drivers/media/video/videobuf-dma-sg.c @@ -94,7 +94,7 @@ err: * must free the memory. */ static struct scatterlist *videobuf_pages_to_sg(struct page **pages, - int nr_pages, int offset) + int nr_pages, int offset, size_t size) { struct scatterlist *sglist; int i; @@ -110,12 +110,14 @@ static struct scatterlist *videobuf_pages_to_sg(struct page **pages, /* DMA to highmem pages might not work */ goto highmem; sg_set_page(&sglist[0], pages[0], PAGE_SIZE - offset, offset); + size -= PAGE_SIZE - offset; for (i = 1; i < nr_pages; i++) { if (NULL == pages[i]) goto nopage; if (PageHighMem(pages[i])) goto highmem; - sg_set_page(&sglist[i], pages[i], PAGE_SIZE, 0); + sg_set_page(&sglist[i], pages[i], min(PAGE_SIZE, size), 0); + size -= min(PAGE_SIZE, size); } return sglist; @@ -170,7 +172,8 @@ static int videobuf_dma_init_user_locked(struct videobuf_dmabuf *dma, first = (data & PAGE_MASK) >> PAGE_SHIFT; last = ((data+size-1) & PAGE_MASK) >> PAGE_SHIFT; - dma->offset = data & ~PAGE_MASK; + dma->offset = data & ~PAGE_MASK; + dma->size = size; dma->nr_pages = last-first+1; dma->pages = kmalloc(dma->nr_pages * sizeof(struct page *), GFP_KERNEL); if (NULL == dma->pages) @@ -252,7 +255,7 @@ int videobuf_dma_map(struct device *dev, struct videobuf_dmabuf *dma) if (dma->pages) { dma->sglist = videobuf_pages_to_sg(dma->pages, dma->nr_pages, - dma->offset); + dma->offset, dma->size); } if (dma->vaddr) { dma->sglist = videobuf_vmalloc_to_sg(dma->vaddr, diff --git a/include/media/videobuf-dma-sg.h b/include/media/videobuf-dma-sg.h index 97e07f4..aa4ebb4 100644 --- a/include/media/videobuf-dma-sg.h +++ b/include/media/videobuf-dma-sg.h @@ -48,6 +48,7 @@ struct videobuf_dmabuf { /* for userland buffer */ int offset; + size_t size; struct page **pages; /* for kernel buffers */ -- cgit v0.10.2 From 04d174e99a6eca2f62b56c10ae1d7d0499d83e9d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jean-Fran=C3=A7ois=20Moine?= Date: Mon, 13 Sep 2010 05:22:37 -0300 Subject: V4L/DVB: gspca - sn9c20x: Bad transfer size of Bayer images MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As the end of frame marker does not contain any pixel, it must not be transferred. Reported-by: Ivo Jager Signed-off-by: Jean-François Moine Cc: stable@kernel.org Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/gspca/sn9c20x.c b/drivers/media/video/gspca/sn9c20x.c index 83a718f..9052d57 100644 --- a/drivers/media/video/gspca/sn9c20x.c +++ b/drivers/media/video/gspca/sn9c20x.c @@ -2357,8 +2357,7 @@ static void sd_pkt_scan(struct gspca_dev *gspca_dev, (data[33] << 10); avg_lum >>= 9; atomic_set(&sd->avg_lum, avg_lum); - gspca_frame_add(gspca_dev, LAST_PACKET, - data, len); + gspca_frame_add(gspca_dev, LAST_PACKET, NULL, 0); return; } if (gspca_dev->last_packet_type == LAST_PACKET) { -- cgit v0.10.2 From 1b376dac05058ac5c9616a507d95c19ea5ea2646 Mon Sep 17 00:00:00 2001 From: Stefan Ringel Date: Thu, 9 Sep 2010 14:45:22 -0300 Subject: V4L/DVB: tm6000: bugfix data handling Signed-off-by: Stefan Ringel Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/staging/tm6000/tm6000-input.c b/drivers/staging/tm6000/tm6000-input.c index 32f7a0a..54f7667 100644 --- a/drivers/staging/tm6000/tm6000-input.c +++ b/drivers/staging/tm6000/tm6000-input.c @@ -46,7 +46,7 @@ MODULE_PARM_DESC(enable_ir, "enable ir (default is enable"); } struct tm6000_ir_poll_result { - u8 rc_data[4]; + u16 rc_data; }; struct tm6000_IR { @@ -60,9 +60,9 @@ struct tm6000_IR { int polling; struct delayed_work work; u8 wait:1; + u8 key:1; struct urb *int_urb; u8 *urb_data; - u8 key:1; int (*get_key) (struct tm6000_IR *, struct tm6000_ir_poll_result *); @@ -122,13 +122,14 @@ static void tm6000_ir_urb_received(struct urb *urb) if (urb->status != 0) printk(KERN_INFO "not ready\n"); - else if (urb->actual_length > 0) + else if (urb->actual_length > 0) { memcpy(ir->urb_data, urb->transfer_buffer, urb->actual_length); - dprintk("data %02x %02x %02x %02x\n", ir->urb_data[0], - ir->urb_data[1], ir->urb_data[2], ir->urb_data[3]); + dprintk("data %02x %02x %02x %02x\n", ir->urb_data[0], + ir->urb_data[1], ir->urb_data[2], ir->urb_data[3]); - ir->key = 1; + ir->key = 1; + } rc = usb_submit_urb(urb, GFP_ATOMIC); } @@ -140,30 +141,47 @@ static int default_polling_getkey(struct tm6000_IR *ir, int rc; u8 buf[2]; - if (ir->wait && !&dev->int_in) { - poll_result->rc_data[0] = 0xff; + if (ir->wait && !&dev->int_in) return 0; - } if (&dev->int_in) { - poll_result->rc_data[0] = ir->urb_data[0]; - poll_result->rc_data[1] = ir->urb_data[1]; + if (ir->ir.ir_type == IR_TYPE_RC5) + poll_result->rc_data = ir->urb_data[0]; + else + poll_result->rc_data = ir->urb_data[0] | ir->urb_data[1] << 8; } else { tm6000_set_reg(dev, REQ_04_EN_DISABLE_MCU_INT, 2, 0); msleep(10); tm6000_set_reg(dev, REQ_04_EN_DISABLE_MCU_INT, 2, 1); msleep(10); - rc = tm6000_read_write_usb(dev, USB_DIR_IN | USB_TYPE_VENDOR | - USB_RECIP_DEVICE, REQ_02_GET_IR_CODE, 0, 0, buf, 1); + if (ir->ir.ir_type == IR_TYPE_RC5) { + rc = tm6000_read_write_usb(dev, USB_DIR_IN | + USB_TYPE_VENDOR | USB_RECIP_DEVICE, + REQ_02_GET_IR_CODE, 0, 0, buf, 1); - msleep(10); + msleep(10); - dprintk("read data=%02x\n", buf[0]); - if (rc < 0) - return rc; + dprintk("read data=%02x\n", buf[0]); + if (rc < 0) + return rc; - poll_result->rc_data[0] = buf[0]; + poll_result->rc_data = buf[0]; + } else { + rc = tm6000_read_write_usb(dev, USB_DIR_IN | + USB_TYPE_VENDOR | USB_RECIP_DEVICE, + REQ_02_GET_IR_CODE, 0, 0, buf, 2); + + msleep(10); + + dprintk("read data=%04x\n", buf[0] | buf[1] << 8); + if (rc < 0) + return rc; + + poll_result->rc_data = buf[0] | buf[1] << 8; + } + if ((poll_result->rc_data & 0x00ff) != 0xff) + ir->key = 1; } return 0; } @@ -180,12 +198,11 @@ static void tm6000_ir_handle_key(struct tm6000_IR *ir) return; } - dprintk("ir->get_key result data=%02x %02x\n", - poll_result.rc_data[0], poll_result.rc_data[1]); + dprintk("ir->get_key result data=%04x\n", poll_result.rc_data); - if (poll_result.rc_data[0] != 0xff && ir->key == 1) { + if (ir->key) { ir_input_keydown(ir->input->input_dev, &ir->ir, - poll_result.rc_data[0] | poll_result.rc_data[1] << 8); + (u32)poll_result.rc_data); ir_input_nokey(ir->input->input_dev, &ir->ir); ir->key = 0; -- cgit v0.10.2 From 5384a12b23160e11ff949a94172051476d308b66 Mon Sep 17 00:00:00 2001 From: Baruch Siach Date: Tue, 27 Jul 2010 09:06:07 -0300 Subject: V4L/DVB: mx2_camera: fix a race causing NULL dereference The mx25_camera_irq irq handler may get called after the camera has been deactivated (from mx2_camera_deactivate). Detect this situation, and bail out. Signed-off-by: Baruch Siach Signed-off-by: Guennadi Liakhovetski Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/mx2_camera.c b/drivers/media/video/mx2_camera.c index 66ff174..b6ea672 100644 --- a/drivers/media/video/mx2_camera.c +++ b/drivers/media/video/mx2_camera.c @@ -378,6 +378,9 @@ static void mx25_camera_frame_done(struct mx2_camera_dev *pcdev, int fb, spin_lock_irqsave(&pcdev->lock, flags); + if (*fb_active == NULL) + goto out; + vb = &(*fb_active)->vb; dev_dbg(pcdev->dev, "%s (vb=0x%p) 0x%08lx %d\n", __func__, vb, vb->baddr, vb->bsize); @@ -402,6 +405,7 @@ static void mx25_camera_frame_done(struct mx2_camera_dev *pcdev, int fb, *fb_active = buf; +out: spin_unlock_irqrestore(&pcdev->lock, flags); } -- cgit v0.10.2 From 6b6d33c746ace7bd0dbbdde674d3fb1100ab081d Mon Sep 17 00:00:00 2001 From: Michael Grzeschik Date: Tue, 3 Aug 2010 07:57:44 -0300 Subject: V4L/DVB: mt9m111: cropcap and s_crop check if type is VIDEO_CAPTURE Signed-off-by: Philipp Wiesner Signed-off-by: Michael Grzeschik Signed-off-by: Guennadi Liakhovetski Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/mt9m111.c b/drivers/media/video/mt9m111.c index 758a4db..28df175 100644 --- a/drivers/media/video/mt9m111.c +++ b/drivers/media/video/mt9m111.c @@ -447,6 +447,9 @@ static int mt9m111_s_crop(struct v4l2_subdev *sd, struct v4l2_crop *a) dev_dbg(&client->dev, "%s left=%d, top=%d, width=%d, height=%d\n", __func__, rect.left, rect.top, rect.width, rect.height); + if (a->type != V4L2_BUF_TYPE_VIDEO_CAPTURE) + return -EINVAL; + ret = mt9m111_make_rect(client, &rect); if (!ret) mt9m111->rect = rect; @@ -466,12 +469,14 @@ static int mt9m111_g_crop(struct v4l2_subdev *sd, struct v4l2_crop *a) static int mt9m111_cropcap(struct v4l2_subdev *sd, struct v4l2_cropcap *a) { + if (a->type != V4L2_BUF_TYPE_VIDEO_CAPTURE) + return -EINVAL; + a->bounds.left = MT9M111_MIN_DARK_COLS; a->bounds.top = MT9M111_MIN_DARK_ROWS; a->bounds.width = MT9M111_MAX_WIDTH; a->bounds.height = MT9M111_MAX_HEIGHT; a->defrect = a->bounds; - a->type = V4L2_BUF_TYPE_VIDEO_CAPTURE; a->pixelaspect.numerator = 1; a->pixelaspect.denominator = 1; -- cgit v0.10.2 From 01f5a394eac48b74c84434e95e74cd172b0682c3 Mon Sep 17 00:00:00 2001 From: Michael Grzeschik Date: Tue, 3 Aug 2010 07:57:45 -0300 Subject: V4L/DVB: mt9m111: added current colorspace at g_fmt Signed-off-by: Michael Grzeschik Signed-off-by: Guennadi Liakhovetski Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/mt9m111.c b/drivers/media/video/mt9m111.c index 28df175..c71af4e 100644 --- a/drivers/media/video/mt9m111.c +++ b/drivers/media/video/mt9m111.c @@ -492,6 +492,7 @@ static int mt9m111_g_fmt(struct v4l2_subdev *sd, mf->width = mt9m111->rect.width; mf->height = mt9m111->rect.height; mf->code = mt9m111->fmt->code; + mf->colorspace = mt9m111->fmt->colorspace; mf->field = V4L2_FIELD_NONE; return 0; -- cgit v0.10.2 From 23f0cb62ba3a1a11ab02cb0675f28befc56c89b9 Mon Sep 17 00:00:00 2001 From: Ionut Gabriel Popescu Date: Fri, 3 Sep 2010 18:20:41 -0300 Subject: V4L/DVB: mt9v022.c: Fixed compilation warning The drivers/media/video/mt9v022.c file, on line 405, tries a "case 0" o a v4l2_mbus_pixelcode enum which don't have an 0 value element, so I got a compile warning. That "case" is useless so it can be removed. Signed-off-by: Ionut Gabriel Popescu Signed-off-by: Guennadi Liakhovetski Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/mt9v022.c b/drivers/media/video/mt9v022.c index e7cd23c..b48473c 100644 --- a/drivers/media/video/mt9v022.c +++ b/drivers/media/video/mt9v022.c @@ -402,9 +402,6 @@ static int mt9v022_s_fmt(struct v4l2_subdev *sd, if (mt9v022->model != V4L2_IDENT_MT9V022IX7ATC) return -EINVAL; break; - case 0: - /* No format change, only geometry */ - break; default: return -EINVAL; } -- cgit v0.10.2 From b17a200f923070104a5b97085ed09ba0f682f637 Mon Sep 17 00:00:00 2001 From: Pawel Osciak Date: Fri, 10 Sep 2010 02:02:32 -0300 Subject: V4L/DVB: v4l: mem2mem_testdev: fix errorenous comparison Output buffer has to be at least the size of input buffer, not the other way around. Signed-off-by: Pawel Osciak Signed-off-by: Kyungmin Park Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/mem2mem_testdev.c b/drivers/media/video/mem2mem_testdev.c index 4525335..d31ec25 100644 --- a/drivers/media/video/mem2mem_testdev.c +++ b/drivers/media/video/mem2mem_testdev.c @@ -239,7 +239,7 @@ static int device_process(struct m2mtest_ctx *ctx, return -EFAULT; } - if (in_buf->vb.size < out_buf->vb.size) { + if (in_buf->vb.size > out_buf->vb.size) { v4l2_err(&dev->v4l2_dev, "Output buffer is too small\n"); return -EINVAL; } -- cgit v0.10.2 From 71088bad7426650e4ea5fb4182580ea8458442e7 Mon Sep 17 00:00:00 2001 From: Pawel Osciak Date: Fri, 10 Sep 2010 02:05:48 -0300 Subject: V4L/DVB: v4l: mem2mem_testdev: add missing release for video_device Video device was not being released on driver remove. Signed-off-by: Pawel Osciak Signed-off-by: Kyungmin Park Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/mem2mem_testdev.c b/drivers/media/video/mem2mem_testdev.c index d31ec25..a7210d9 100644 --- a/drivers/media/video/mem2mem_testdev.c +++ b/drivers/media/video/mem2mem_testdev.c @@ -1014,6 +1014,7 @@ static int m2mtest_remove(struct platform_device *pdev) v4l2_m2m_release(dev->m2m_dev); del_timer_sync(&dev->timer); video_unregister_device(dev->vfd); + video_device_release(dev->vfd); v4l2_device_unregister(&dev->v4l2_dev); kfree(dev); -- cgit v0.10.2 From fc00a1d90d7b64e44c2475833cc3021066f67ea7 Mon Sep 17 00:00:00 2001 From: Andy Walls Date: Thu, 16 Sep 2010 20:54:47 -0300 Subject: V4L/DVB: cx25840: Fix typo in volume control initialization: 65335 vs. 65535 The wrong value for the volume control limit, 65335 vs. 65535, prevents proper cx25840 v4l2_subdevice initialization. Reported-by: Igor M. liplianin Signed-off-by: Andy Walls Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/cx25840/cx25840-core.c b/drivers/media/video/cx25840/cx25840-core.c index 86ca8c2..f5a3e74 100644 --- a/drivers/media/video/cx25840/cx25840-core.c +++ b/drivers/media/video/cx25840/cx25840-core.c @@ -1996,7 +1996,7 @@ static int cx25840_probe(struct i2c_client *client, state->volume = v4l2_ctrl_new_std(&state->hdl, &cx25840_audio_ctrl_ops, V4L2_CID_AUDIO_VOLUME, - 0, 65335, 65535 / 100, default_volume); + 0, 65535, 65535 / 100, default_volume); state->mute = v4l2_ctrl_new_std(&state->hdl, &cx25840_audio_ctrl_ops, V4L2_CID_AUDIO_MUTE, 0, 1, 1, 0); -- cgit v0.10.2 From 405707985594169cfd0b1d97d29fcb4b4c6f2ac9 Mon Sep 17 00:00:00 2001 From: Dan Rosenberg Date: Wed, 15 Sep 2010 18:44:22 -0300 Subject: V4L/DVB: ivtvfb: prevent reading uninitialized stack memory The FBIOGET_VBLANK device ioctl allows unprivileged users to read 16 bytes of uninitialized stack memory, because the "reserved" member of the fb_vblank struct declared on the stack is not altered or zeroed before being copied back to the user. This patch takes care of it. Signed-off-by: Dan Rosenberg Signed-off-by: Andy Walls Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/ivtv/ivtvfb.c b/drivers/media/video/ivtv/ivtvfb.c index be03a71..f0316d0 100644 --- a/drivers/media/video/ivtv/ivtvfb.c +++ b/drivers/media/video/ivtv/ivtvfb.c @@ -466,6 +466,8 @@ static int ivtvfb_ioctl(struct fb_info *info, unsigned int cmd, unsigned long ar struct fb_vblank vblank; u32 trace; + memset(&vblank, 0, sizeof(struct fb_vblank)); + vblank.flags = FB_VBLANK_HAVE_COUNT |FB_VBLANK_HAVE_VCOUNT | FB_VBLANK_HAVE_VSYNC; trace = read_reg(IVTV_REG_DEC_LINE_FIELD) >> 16; -- cgit v0.10.2 From 9275b32bced808bd57bdf3884748ed1126ba90e0 Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Sun, 12 Sep 2010 08:31:29 -0300 Subject: V4L/DVB: uvcvideo: Fix support for Medion Akoya All-in-one PC integrated webcam The camera requires the STREAM_NO_FID quirk. Add a corresponding entry in the device IDs list. Signed-off-by: Laurent Pinchart Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/uvc/uvc_driver.c b/drivers/media/video/uvc/uvc_driver.c index 8bdd940..63756ca 100644 --- a/drivers/media/video/uvc/uvc_driver.c +++ b/drivers/media/video/uvc/uvc_driver.c @@ -2091,6 +2091,15 @@ static struct usb_device_id uvc_ids[] = { .bInterfaceProtocol = 0, .driver_info = UVC_QUIRK_PROBE_MINMAX | UVC_QUIRK_PROBE_DEF }, + /* IMC Networks (Medion Akoya) */ + { .match_flags = USB_DEVICE_ID_MATCH_DEVICE + | USB_DEVICE_ID_MATCH_INT_INFO, + .idVendor = 0x13d3, + .idProduct = 0x5103, + .bInterfaceClass = USB_CLASS_VIDEO, + .bInterfaceSubClass = 1, + .bInterfaceProtocol = 0, + .driver_info = UVC_QUIRK_STREAM_NO_FID }, /* Syntek (HP Spartan) */ { .match_flags = USB_DEVICE_ID_MATCH_DEVICE | USB_DEVICE_ID_MATCH_INT_INFO, -- cgit v0.10.2 From 86d8b6abbbc973a1dd7d3c8f9e3c3e06131703b2 Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Thu, 16 Sep 2010 12:37:26 -0300 Subject: V4L/DVB: uvcvideo: Restrict frame rates for Chicony CNF7129 webcam At all frame rates except 30fps and 5fps the camera produces very dark pictures. Auto-exposure is probably disabled by the camera at all frame rates except 30fps, making them pretty unusable. Work around the problem by introducing a new RESTRICT_FRAME_RATE quirk that disables all the frame rates except the default one. Signed-off-by: Laurent Pinchart Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/uvc/uvc_driver.c b/drivers/media/video/uvc/uvc_driver.c index 63756ca..2ac85d8 100644 --- a/drivers/media/video/uvc/uvc_driver.c +++ b/drivers/media/video/uvc/uvc_driver.c @@ -486,6 +486,12 @@ static int uvc_parse_format(struct uvc_device *dev, max(frame->dwFrameInterval[0], frame->dwDefaultFrameInterval)); + if (dev->quirks & UVC_QUIRK_RESTRICT_FRAME_RATE) { + frame->bFrameIntervalType = 1; + frame->dwFrameInterval[0] = + frame->dwDefaultFrameInterval; + } + uvc_trace(UVC_TRACE_DESCR, "- %ux%u (%u.%u fps)\n", frame->wWidth, frame->wHeight, 10000000/frame->dwDefaultFrameInterval, @@ -2026,6 +2032,15 @@ static struct usb_device_id uvc_ids[] = { .bInterfaceClass = USB_CLASS_VENDOR_SPEC, .bInterfaceSubClass = 1, .bInterfaceProtocol = 0 }, + /* Chicony CNF7129 (Asus EEE 100HE) */ + { .match_flags = USB_DEVICE_ID_MATCH_DEVICE + | USB_DEVICE_ID_MATCH_INT_INFO, + .idVendor = 0x04f2, + .idProduct = 0xb071, + .bInterfaceClass = USB_CLASS_VIDEO, + .bInterfaceSubClass = 1, + .bInterfaceProtocol = 0, + .driver_info = UVC_QUIRK_RESTRICT_FRAME_RATE }, /* Alcor Micro AU3820 (Future Boy PC USB Webcam) */ { .match_flags = USB_DEVICE_ID_MATCH_DEVICE | USB_DEVICE_ID_MATCH_INT_INFO, diff --git a/drivers/media/video/uvc/uvcvideo.h b/drivers/media/video/uvc/uvcvideo.h index bdacf3b..892e0e5 100644 --- a/drivers/media/video/uvc/uvcvideo.h +++ b/drivers/media/video/uvc/uvcvideo.h @@ -182,6 +182,7 @@ struct uvc_xu_control { #define UVC_QUIRK_IGNORE_SELECTOR_UNIT 0x00000020 #define UVC_QUIRK_FIX_BANDWIDTH 0x00000080 #define UVC_QUIRK_PROBE_DEF 0x00000100 +#define UVC_QUIRK_RESTRICT_FRAME_RATE 0x00000200 /* Format flags */ #define UVC_FMT_FLAG_COMPRESSED 0x00000001 -- cgit v0.10.2 From 81c69fc200bfa2f3a8b2fed614d1c6670c4a7e14 Mon Sep 17 00:00:00 2001 From: Pawel Osciak Date: Mon, 6 Sep 2010 03:53:43 -0300 Subject: V4L/DVB: v4l: s5p-fimc: Fix return value on probe() failure On failed create_workqueue() fimc_probe() was returning 0. Signed-off-by: Pawel Osciak Signed-off-by: Kyungmin Park Signed-off-by: Marek Szyprowski Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/s5p-fimc/fimc-core.c b/drivers/media/video/s5p-fimc/fimc-core.c index 1beb226..8cd1ef0 100644 --- a/drivers/media/video/s5p-fimc/fimc-core.c +++ b/drivers/media/video/s5p-fimc/fimc-core.c @@ -1415,8 +1415,10 @@ static int fimc_probe(struct platform_device *pdev) } fimc->work_queue = create_workqueue(dev_name(&fimc->pdev->dev)); - if (!fimc->work_queue) + if (!fimc->work_queue) { + ret = -ENOMEM; goto err_irq; + } ret = fimc_register_m2m_device(fimc); if (ret) -- cgit v0.10.2 From ddc79e0fdc16c05c3ff7f9b6ae9052bda0506108 Mon Sep 17 00:00:00 2001 From: Sylwester Nawrocki Date: Mon, 6 Sep 2010 03:53:44 -0300 Subject: V4L/DVB: v4l: s5p-fimc: Fix 3-planar formats handling and pixel offset error on S5PV210 SoCs Fix DMA engine pixel offset calculation for 3-planar YUV formats. On S5PV210 SoCs horizontal offset is applied as number of pixels, not bytes per line. [mchehab@redhat.com: CodingStyle cleanup] Signed-off-by: Sylwester Nawrocki Signed-off-by: Kyungmin Park Signed-off-by: Marek Szyprowski Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/s5p-fimc/fimc-core.c b/drivers/media/video/s5p-fimc/fimc-core.c index 8cd1ef0..6961c55 100644 --- a/drivers/media/video/s5p-fimc/fimc-core.c +++ b/drivers/media/video/s5p-fimc/fimc-core.c @@ -393,6 +393,37 @@ static void fimc_set_yuv_order(struct fimc_ctx *ctx) dbg("ctx->out_order_1p= %d", ctx->out_order_1p); } +static void fimc_prepare_dma_offset(struct fimc_ctx *ctx, struct fimc_frame *f) +{ + struct samsung_fimc_variant *variant = ctx->fimc_dev->variant; + + f->dma_offset.y_h = f->offs_h; + if (!variant->pix_hoff) + f->dma_offset.y_h *= (f->fmt->depth >> 3); + + f->dma_offset.y_v = f->offs_v; + + f->dma_offset.cb_h = f->offs_h; + f->dma_offset.cb_v = f->offs_v; + + f->dma_offset.cr_h = f->offs_h; + f->dma_offset.cr_v = f->offs_v; + + if (!variant->pix_hoff) { + if (f->fmt->planes_cnt == 3) { + f->dma_offset.cb_h >>= 1; + f->dma_offset.cr_h >>= 1; + } + if (f->fmt->color == S5P_FIMC_YCBCR420) { + f->dma_offset.cb_v >>= 1; + f->dma_offset.cr_v >>= 1; + } + } + + dbg("in_offset: color= %d, y_h= %d, y_v= %d", + f->fmt->color, f->dma_offset.y_h, f->dma_offset.y_v); +} + /** * fimc_prepare_config - check dimensions, operation and color mode * and pre-calculate offset and the scaling coefficients. @@ -406,7 +437,6 @@ static int fimc_prepare_config(struct fimc_ctx *ctx, u32 flags) { struct fimc_frame *s_frame, *d_frame; struct fimc_vid_buffer *buf = NULL; - struct samsung_fimc_variant *variant = ctx->fimc_dev->variant; int ret = 0; s_frame = &ctx->s_frame; @@ -419,61 +449,16 @@ static int fimc_prepare_config(struct fimc_ctx *ctx, u32 flags) swap(d_frame->width, d_frame->height); } - /* Prepare the output offset ratios for scaler. */ - d_frame->dma_offset.y_h = d_frame->offs_h; - if (!variant->pix_hoff) - d_frame->dma_offset.y_h *= (d_frame->fmt->depth >> 3); - - d_frame->dma_offset.y_v = d_frame->offs_v; + /* Prepare the DMA offset ratios for scaler. */ + fimc_prepare_dma_offset(ctx, &ctx->s_frame); + fimc_prepare_dma_offset(ctx, &ctx->d_frame); - d_frame->dma_offset.cb_h = d_frame->offs_h; - d_frame->dma_offset.cb_v = d_frame->offs_v; - - d_frame->dma_offset.cr_h = d_frame->offs_h; - d_frame->dma_offset.cr_v = d_frame->offs_v; - - if (!variant->pix_hoff && d_frame->fmt->planes_cnt == 3) { - d_frame->dma_offset.cb_h >>= 1; - d_frame->dma_offset.cb_v >>= 1; - d_frame->dma_offset.cr_h >>= 1; - d_frame->dma_offset.cr_v >>= 1; - } - - dbg("out offset: color= %d, y_h= %d, y_v= %d", - d_frame->fmt->color, - d_frame->dma_offset.y_h, d_frame->dma_offset.y_v); - - /* Prepare the input offset ratios for scaler. */ - s_frame->dma_offset.y_h = s_frame->offs_h; - if (!variant->pix_hoff) - s_frame->dma_offset.y_h *= (s_frame->fmt->depth >> 3); - s_frame->dma_offset.y_v = s_frame->offs_v; - - s_frame->dma_offset.cb_h = s_frame->offs_h; - s_frame->dma_offset.cb_v = s_frame->offs_v; - - s_frame->dma_offset.cr_h = s_frame->offs_h; - s_frame->dma_offset.cr_v = s_frame->offs_v; - - if (!variant->pix_hoff && s_frame->fmt->planes_cnt == 3) { - s_frame->dma_offset.cb_h >>= 1; - s_frame->dma_offset.cb_v >>= 1; - s_frame->dma_offset.cr_h >>= 1; - s_frame->dma_offset.cr_v >>= 1; - } - - dbg("in offset: color= %d, y_h= %d, y_v= %d", - s_frame->fmt->color, s_frame->dma_offset.y_h, - s_frame->dma_offset.y_v); - - fimc_set_yuv_order(ctx); - - /* Check against the scaler ratio. */ if (s_frame->height > (SCALER_MAX_VRATIO * d_frame->height) || s_frame->width > (SCALER_MAX_HRATIO * d_frame->width)) { err("out of scaler range"); return -EINVAL; } + fimc_set_yuv_order(ctx); } /* Input DMA mode is not allowed when the scaler is disabled. */ @@ -1495,6 +1480,7 @@ static struct samsung_fimc_variant fimc2_variant_s5p = { }; static struct samsung_fimc_variant fimc01_variant_s5pv210 = { + .pix_hoff = 1, .has_inp_rot = 1, .has_out_rot = 1, .min_inp_pixsize = 16, @@ -1509,6 +1495,7 @@ static struct samsung_fimc_variant fimc01_variant_s5pv210 = { }; static struct samsung_fimc_variant fimc2_variant_s5pv210 = { + .pix_hoff = 1, .min_inp_pixsize = 16, .min_out_pixsize = 32, -- cgit v0.10.2 From bf97b6734e027cc18abad420ab88f861f65d7816 Mon Sep 17 00:00:00 2001 From: Benjamin Marzinski Date: Mon, 27 Sep 2010 16:00:04 -0500 Subject: GFS2: reserve more blocks for transactions Some of the functions in GFS2 were not reserving space in the transaction for the resource group header and the resource groups bitblocks that get added when you do allocation. GFS2 now makes sure to reserve space for the resource group header and either all the bitblocks in the resource group, or one for each block that it may allocate, whichever is smaller using the new gfs2_rg_blocks() inline function. Signed-off-by: Benjamin Marzinski Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 180ef8a..1bf1788 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -663,6 +663,8 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping, rblocks += RES_STATFS + RES_QUOTA; if (&ip->i_inode == sdp->sd_rindex) rblocks += 2 * RES_STATFS; + if (alloc_required) + rblocks += gfs2_rg_blocks(al); error = gfs2_trans_begin(sdp, rblocks, PAGE_CACHE_SIZE/sdp->sd_sb.sb_bsize); diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 04513e9..5476c06 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -1151,7 +1151,7 @@ static int do_grow(struct inode *inode, u64 size) goto do_grow_qunlock; } - error = gfs2_trans_begin(sdp, RES_DINODE + 1, 0); + error = gfs2_trans_begin(sdp, RES_DINODE + RES_STATFS + RES_RG_BIT, 0); if (error) goto do_grow_release; diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index daadcd2..237ee6a 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -382,8 +382,10 @@ static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) rblocks = RES_DINODE + ind_blocks; if (gfs2_is_jdata(ip)) rblocks += data_blocks ? data_blocks : 1; - if (ind_blocks || data_blocks) + if (ind_blocks || data_blocks) { rblocks += RES_STATFS + RES_QUOTA; + rblocks += gfs2_rg_blocks(al); + } ret = gfs2_trans_begin(sdp, rblocks, 0); if (ret) goto out_trans_fail; diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 98a94cfc..fba0017 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -219,7 +219,7 @@ static int gfs2_link(struct dentry *old_dentry, struct inode *dir, goto out_gunlock_q; error = gfs2_trans_begin(sdp, sdp->sd_max_dirres + - al->al_rgd->rd_length + + gfs2_rg_blocks(al) + 2 * RES_DINODE + RES_STATFS + RES_QUOTA, 0); if (error) @@ -884,7 +884,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, goto out_gunlock_q; error = gfs2_trans_begin(sdp, sdp->sd_max_dirres + - al->al_rgd->rd_length + + gfs2_rg_blocks(al) + 4 * RES_DINODE + 4 * RES_LEAF + RES_STATFS + RES_QUOTA + 4, 0); if (error) @@ -1481,7 +1481,7 @@ retry: al->al_requested = data_blocks + ind_blocks; rblocks = RES_DINODE + ind_blocks + RES_STATFS + RES_QUOTA + - RES_RG_HDR + ip->i_alloc->al_rgd->rd_length; + RES_RG_HDR + gfs2_rg_blocks(al); if (gfs2_is_jdata(ip)) rblocks += data_blocks ? data_blocks : 1; diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 9bc6dd9..58a9b99 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -815,7 +815,7 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda) goto out_alloc; if (nalloc) - blocks += al->al_rgd->rd_length + nalloc * ind_blocks + RES_STATFS; + blocks += gfs2_rg_blocks(al) + nalloc * ind_blocks + RES_STATFS; error = gfs2_trans_begin(sdp, blocks, 0); if (error) @@ -1586,6 +1586,7 @@ static int gfs2_set_dqblk(struct super_block *sb, int type, qid_t id, error = gfs2_inplace_reserve(ip); if (error) goto out_alloc; + blocks += gfs2_rg_blocks(al); } error = gfs2_trans_begin(sdp, blocks + RES_DINODE + 1, 0); diff --git a/fs/gfs2/trans.h b/fs/gfs2/trans.h index b849eb7..fb56b78 100644 --- a/fs/gfs2/trans.h +++ b/fs/gfs2/trans.h @@ -26,6 +26,14 @@ struct gfs2_glock; #define RES_STATFS 1 #define RES_QUOTA 2 +/* reserve either the number of blocks to be allocated plus the rg header + * block, or all of the blocks in the rg, whichever is smaller */ +static inline unsigned int gfs2_rg_blocks(const struct gfs2_alloc *al) +{ + return (al->al_requested < al->al_rgd->rd_length)? + al->al_requested + 1 : al->al_rgd->rd_length; +} + int gfs2_trans_begin(struct gfs2_sbd *sdp, unsigned int blocks, unsigned int revokes); diff --git a/fs/gfs2/xattr.c b/fs/gfs2/xattr.c index 776af6e..30b58f07 100644 --- a/fs/gfs2/xattr.c +++ b/fs/gfs2/xattr.c @@ -734,7 +734,7 @@ static int ea_alloc_skeleton(struct gfs2_inode *ip, struct gfs2_ea_request *er, goto out_gunlock_q; error = gfs2_trans_begin(GFS2_SB(&ip->i_inode), - blks + al->al_rgd->rd_length + + blks + gfs2_rg_blocks(al) + RES_DINODE + RES_STATFS + RES_QUOTA, 0); if (error) goto out_ipres; -- cgit v0.10.2 From d594845106f34c079a6b05be01a37e4883c3bf4c Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 28 Sep 2010 10:17:47 +0100 Subject: GFS2: Fix compiler warning from previous patch This shouldn't really be required, but gcc can't tell that "al" is only accessed when initialised. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 1bf1788..6b24afb 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -615,7 +615,7 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping, unsigned int data_blocks = 0, ind_blocks = 0, rblocks; int alloc_required; int error = 0; - struct gfs2_alloc *al; + struct gfs2_alloc *al = NULL; pgoff_t index = pos >> PAGE_CACHE_SHIFT; unsigned from = pos & (PAGE_CACHE_SIZE - 1); unsigned to = from + len; -- cgit v0.10.2 From c741c4551237f9c1bdcd3b1b39b0883bd19a3723 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 29 Sep 2010 14:20:52 +0100 Subject: GFS2: Fix spectator umount issue The tests further down the recovery function relating to unlocking the journal need to be updated to match the intial test. Also, a test in the umount code which was surplus to requirements has been removed. Umounting spectator mounts now works correctly, as expected. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index c3f2a5c..8777885 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -1517,7 +1517,7 @@ static void clear_glock(struct gfs2_glock *gl) spin_unlock(&lru_lock); spin_lock(&gl->gl_spin); - if (find_first_holder(gl) == NULL && gl->gl_state != LM_ST_UNLOCKED) + if (gl->gl_state != LM_ST_UNLOCKED) handle_callback(gl, LM_ST_UNLOCKED, 0); spin_unlock(&gl->gl_spin); gfs2_glock_hold(gl); diff --git a/fs/gfs2/recovery.c b/fs/gfs2/recovery.c index 666548e..f2a02ed 100644 --- a/fs/gfs2/recovery.c +++ b/fs/gfs2/recovery.c @@ -455,12 +455,13 @@ void gfs2_recover_func(struct work_struct *work) int ro = 0; unsigned int pass; int error; + int jlocked = 0; if (sdp->sd_args.ar_spectator || (jd->jd_jid != sdp->sd_lockstruct.ls_jid)) { fs_info(sdp, "jid=%u: Trying to acquire journal lock...\n", jd->jd_jid); - + jlocked = 1; /* Acquire the journal lock so we can do recovery */ error = gfs2_glock_nq_num(sdp, jd->jd_jid, &gfs2_journal_glops, @@ -555,13 +556,12 @@ void gfs2_recover_func(struct work_struct *work) jd->jd_jid, t); } - if (jd->jd_jid != sdp->sd_lockstruct.ls_jid) - gfs2_glock_dq_uninit(&ji_gh); - gfs2_recovery_done(sdp, jd->jd_jid, LM_RD_SUCCESS); - if (jd->jd_jid != sdp->sd_lockstruct.ls_jid) + if (jlocked) { + gfs2_glock_dq_uninit(&ji_gh); gfs2_glock_dq_uninit(&j_gh); + } fs_info(sdp, "jid=%u: Done\n", jd->jd_jid); goto done; @@ -569,7 +569,7 @@ void gfs2_recover_func(struct work_struct *work) fail_gunlock_tr: gfs2_glock_dq_uninit(&t_gh); fail_gunlock_ji: - if (jd->jd_jid != sdp->sd_lockstruct.ls_jid) { + if (jlocked) { gfs2_glock_dq_uninit(&ji_gh); fail_gunlock_j: gfs2_glock_dq_uninit(&j_gh); -- cgit v0.10.2 From 43f74c199563a4273e528e2166d0650625a1e05f Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 29 Sep 2010 14:24:41 +0100 Subject: GFS2: Add "norecovery" mount option as a synonym for "spectator" XFS supports the "norecovery" mount option which is basically the same as the GFS2 spectator mode. This adds support for "norecovery" as a synonym for spectator mode, which is hopefully a more obvious description of what it actually does. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index d85e0b7..047d117 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -85,6 +85,7 @@ static const match_table_t tokens = { {Opt_locktable, "locktable=%s"}, {Opt_hostdata, "hostdata=%s"}, {Opt_spectator, "spectator"}, + {Opt_spectator, "norecovery"}, {Opt_ignore_local_fs, "ignore_local_fs"}, {Opt_localflocks, "localflocks"}, {Opt_localcaching, "localcaching"}, -- cgit v0.10.2 From feb47ca9314666d920855b8a235032dea2b2caa4 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 29 Sep 2010 15:04:18 +0100 Subject: GFS2: Improve journal allocation via sysfs Recently a feature was added to GFS2 to allow journal id allocation via sysfs. This patch builds upon that so that a negative journal id will be treated as an error code to be passed back as the return code from mount. This allows termination of the mount process if there is a failure. Also, the process has been updated so that the kernel will wait for a journal id, even in the "spectator" case. This is required in order to avoid mounting a filesystem in case there is an error while joining the cluster. In the spectator case, 0 is written into the file to indicate that all is well, and that mount should continue. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 6f6ff8a..764fbb4 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -494,7 +494,7 @@ struct gfs2_sb_host { */ struct lm_lockstruct { - unsigned int ls_jid; + int ls_jid; unsigned int ls_first; unsigned int ls_first_done; unsigned int ls_nodir; diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 5b11f37..aeafc23 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -1056,8 +1056,6 @@ static int gfs2_journalid_wait(void *word) static int wait_on_journal(struct gfs2_sbd *sdp) { - if (sdp->sd_args.ar_spectator) - return 0; if (sdp->sd_lockstruct.ls_ops->lm_mount == NULL) return 0; @@ -1160,6 +1158,20 @@ static int fill_super(struct super_block *sb, struct gfs2_args *args, int silent if (error) goto fail_sb; + /* + * If user space has failed to join the cluster or some similar + * failure has occurred, then the journal id will contain a + * negative (error) number. This will then be returned to the + * caller (of the mount syscall). We do this even for spectator + * mounts (which just write a jid of 0 to indicate "ok" even though + * the jid is unused in the spectator case) + */ + if (sdp->sd_lockstruct.ls_jid < 0) { + error = sdp->sd_lockstruct.ls_jid; + sdp->sd_lockstruct.ls_jid = 0; + goto fail_sb; + } + error = init_inodes(sdp, DO); if (error) goto fail_sb; diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c index ccacffd..64082a5 100644 --- a/fs/gfs2/sys.c +++ b/fs/gfs2/sys.c @@ -399,31 +399,32 @@ static ssize_t recover_status_show(struct gfs2_sbd *sdp, char *buf) static ssize_t jid_show(struct gfs2_sbd *sdp, char *buf) { - return sprintf(buf, "%u\n", sdp->sd_lockstruct.ls_jid); + return sprintf(buf, "%d\n", sdp->sd_lockstruct.ls_jid); } static ssize_t jid_store(struct gfs2_sbd *sdp, const char *buf, size_t len) { - unsigned jid; + int jid; int rv; - rv = sscanf(buf, "%u", &jid); + rv = sscanf(buf, "%d", &jid); if (rv != 1) return -EINVAL; spin_lock(&sdp->sd_jindex_spin); rv = -EINVAL; - if (sdp->sd_args.ar_spectator) - goto out; if (sdp->sd_lockstruct.ls_ops->lm_mount == NULL) goto out; rv = -EBUSY; - if (test_and_clear_bit(SDF_NOJOURNALID, &sdp->sd_flags) == 0) + if (test_bit(SDF_NOJOURNALID, &sdp->sd_flags) == 0) goto out; + rv = 0; + if (sdp->sd_args.ar_spectator && jid > 0) + rv = jid = -EINVAL; sdp->sd_lockstruct.ls_jid = jid; + clear_bit(SDF_NOJOURNALID, &sdp->sd_flags); smp_mb__after_clear_bit(); wake_up_bit(&sdp->sd_flags, SDF_NOJOURNALID); - rv = 0; out: spin_unlock(&sdp->sd_jindex_spin); return rv ? rv : len; @@ -617,7 +618,7 @@ static int gfs2_uevent(struct kset *kset, struct kobject *kobj, add_uevent_var(env, "LOCKTABLE=%s", sdp->sd_table_name); add_uevent_var(env, "LOCKPROTO=%s", sdp->sd_proto_name); if (!test_bit(SDF_NOJOURNALID, &sdp->sd_flags)) - add_uevent_var(env, "JOURNALID=%u", sdp->sd_lockstruct.ls_jid); + add_uevent_var(env, "JOURNALID=%d", sdp->sd_lockstruct.ls_jid); if (gfs2_uuid_valid(uuid)) add_uevent_var(env, "UUID=%pUB", uuid); return 0; -- cgit v0.10.2 From a08c7c68f702e2a2797a4035b6c0a756c4886c26 Mon Sep 17 00:00:00 2001 From: Brian Rogers Date: Wed, 22 Sep 2010 08:06:43 -0300 Subject: V4L/DVB: ir-core: Fix null dereferences in the protocols sysfs interface For some cards, ir_dev->props and ir_dev->raw are both NULL. These cards are using built-in IR decoding instead of raw, and can't easily be made to switch protocols. So upon reading /sys/class/rc/rc?/protocols on such a card, return 'builtin' as the supported and enabled protocol. Return -EINVAL on any attempts to change the protocol. And most important of all, don't crash. Signed-off-by: Brian Rogers Acked-by: Jarod Wilson Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/IR/ir-sysfs.c b/drivers/media/IR/ir-sysfs.c index 96dafc4..46d4246 100644 --- a/drivers/media/IR/ir-sysfs.c +++ b/drivers/media/IR/ir-sysfs.c @@ -67,13 +67,14 @@ static ssize_t show_protocols(struct device *d, char *tmp = buf; int i; - if (ir_dev->props->driver_type == RC_DRIVER_SCANCODE) { + if (ir_dev->props && ir_dev->props->driver_type == RC_DRIVER_SCANCODE) { enabled = ir_dev->rc_tab.ir_type; allowed = ir_dev->props->allowed_protos; - } else { + } else if (ir_dev->raw) { enabled = ir_dev->raw->enabled_protocols; allowed = ir_raw_get_allowed_protocols(); - } + } else + return sprintf(tmp, "[builtin]\n"); IR_dprintk(1, "allowed - 0x%llx, enabled - 0x%llx\n", (long long)allowed, @@ -121,10 +122,14 @@ static ssize_t store_protocols(struct device *d, int rc, i, count = 0; unsigned long flags; - if (ir_dev->props->driver_type == RC_DRIVER_SCANCODE) + if (ir_dev->props && ir_dev->props->driver_type == RC_DRIVER_SCANCODE) type = ir_dev->rc_tab.ir_type; - else + else if (ir_dev->raw) type = ir_dev->raw->enabled_protocols; + else { + IR_dprintk(1, "Protocol switching not supported\n"); + return -EINVAL; + } while ((tmp = strsep((char **) &data, " \n")) != NULL) { if (!*tmp) @@ -185,7 +190,7 @@ static ssize_t store_protocols(struct device *d, } } - if (ir_dev->props->driver_type == RC_DRIVER_SCANCODE) { + if (ir_dev->props && ir_dev->props->driver_type == RC_DRIVER_SCANCODE) { spin_lock_irqsave(&ir_dev->rc_tab.lock, flags); ir_dev->rc_tab.ir_type = type; spin_unlock_irqrestore(&ir_dev->rc_tab.lock, flags); -- cgit v0.10.2 From b2b476f53a9d24b00a313adf7f6ca92515a2af54 Mon Sep 17 00:00:00 2001 From: Pawel Osciak Date: Tue, 20 Jul 2010 13:49:16 -0300 Subject: V4L/DVB: v4l: videobuf: prevent passing a NULL to dma_free_coherent() When a driver that uses videobuf-dma-contig is used with the USERPTR memory access method a kernel oops might happen: a NULL address may be passed to dma_free_coherent(). This happens when an application calls REQBUFS and then exits without queuing any buffers. This patch fixes that bug. Signed-off-by: Pawel Osciak Signed-off-by: Kyungmin Park Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/video/videobuf-dma-contig.c b/drivers/media/video/videobuf-dma-contig.c index 372b87e..6ff9e4b 100644 --- a/drivers/media/video/videobuf-dma-contig.c +++ b/drivers/media/video/videobuf-dma-contig.c @@ -393,8 +393,10 @@ void videobuf_dma_contig_free(struct videobuf_queue *q, } /* read() method */ - dma_free_coherent(q->dev, mem->size, mem->vaddr, mem->dma_handle); - mem->vaddr = NULL; + if (mem->vaddr) { + dma_free_coherent(q->dev, mem->size, mem->vaddr, mem->dma_handle); + mem->vaddr = NULL; + } } EXPORT_SYMBOL_GPL(videobuf_dma_contig_free); -- cgit v0.10.2 From cc6e853c5e49ea3698e87415672d818c6d7a5ee9 Mon Sep 17 00:00:00 2001 From: Marek Szyprowski Date: Sun, 5 Sep 2010 02:42:33 -0300 Subject: V4L/DVB: v4l: radio: si470x: fix unneeded free_irq() call In case of error during probe() the driver calls free_irq() function on not yet allocated irq. This patches fixes the call sequence in case of the error. Signed-off-by: Marek Szyprowski Signed-off-by: Kyungmin Park Signed-off-by: Mauro Carvalho Chehab diff --git a/drivers/media/radio/si470x/radio-si470x-i2c.c b/drivers/media/radio/si470x/radio-si470x-i2c.c index 67a4ec8..4ce541a 100644 --- a/drivers/media/radio/si470x/radio-si470x-i2c.c +++ b/drivers/media/radio/si470x/radio-si470x-i2c.c @@ -395,7 +395,7 @@ static int __devinit si470x_i2c_probe(struct i2c_client *client, radio->registers[POWERCFG] = POWERCFG_ENABLE; if (si470x_set_register(radio, POWERCFG) < 0) { retval = -EIO; - goto err_all; + goto err_video; } msleep(110); -- cgit v0.10.2 From 59bfee6e0682635c269fb271422e2595fa441c21 Mon Sep 17 00:00:00 2001 From: Wolfram Sang Date: Thu, 30 Sep 2010 14:14:22 +0200 Subject: i2c: Remove obsolete cleanup for clientdata A few new i2c-drivers came into the kernel which clear the clientdata-pointer on exit. This is obsolete meanwhile, so fix it and hope the word will spread. Signed-off-by: Wolfram Sang Acked-by: Mark Brown Signed-off-by: Jean Delvare diff --git a/drivers/misc/bh1780gli.c b/drivers/misc/bh1780gli.c index 714c6b4..d5f3a3f 100644 --- a/drivers/misc/bh1780gli.c +++ b/drivers/misc/bh1780gli.c @@ -190,7 +190,6 @@ static int __devexit bh1780_remove(struct i2c_client *client) ddata = i2c_get_clientdata(client); sysfs_remove_group(&client->dev.kobj, &bh1780_attr_group); - i2c_set_clientdata(client, NULL); kfree(ddata); return 0; diff --git a/drivers/regulator/ad5398.c b/drivers/regulator/ad5398.c index df1fb53..a4be416 100644 --- a/drivers/regulator/ad5398.c +++ b/drivers/regulator/ad5398.c @@ -256,7 +256,6 @@ static int __devexit ad5398_remove(struct i2c_client *client) regulator_unregister(chip->rdev); kfree(chip); - i2c_set_clientdata(client, NULL); return 0; } diff --git a/drivers/regulator/isl6271a-regulator.c b/drivers/regulator/isl6271a-regulator.c index d61ecb8..b8cc638 100644 --- a/drivers/regulator/isl6271a-regulator.c +++ b/drivers/regulator/isl6271a-regulator.c @@ -191,8 +191,6 @@ static int __devexit isl6271a_remove(struct i2c_client *i2c) struct isl_pmic *pmic = i2c_get_clientdata(i2c); int i; - i2c_set_clientdata(i2c, NULL); - for (i = 0; i < 3; i++) regulator_unregister(pmic->rdev[i]); diff --git a/drivers/rtc/rtc-ds3232.c b/drivers/rtc/rtc-ds3232.c index 9daed8d..9de8516 100644 --- a/drivers/rtc/rtc-ds3232.c +++ b/drivers/rtc/rtc-ds3232.c @@ -268,7 +268,6 @@ out_irq: free_irq(client->irq, client); out_free: - i2c_set_clientdata(client, NULL); kfree(ds3232); return ret; } @@ -287,7 +286,6 @@ static int __devexit ds3232_remove(struct i2c_client *client) } rtc_device_unregister(ds3232->rtc); - i2c_set_clientdata(client, NULL); kfree(ds3232); return 0; } -- cgit v0.10.2 From 753419f59e10d7181e43f0b9cc5beff43ef3f7a4 Mon Sep 17 00:00:00 2001 From: Vishwanath BS Date: Thu, 30 Sep 2010 14:14:22 +0200 Subject: i2c: Fix for suspend/resume issue In current i2c core driver, call to pm_runtime_set_active from i2c_device_pm_resume will unconditionally enable i2c module and increment child count of the parent. Because of this, in CPU Idle path, i2c does not idle, preventing Core to enter retention. Also i2c module will not be suspended upon system suspend as pm_runtime_set_suspended is not called from i2c_device_pm_suspend. This issue is fixed by removing pm_runtime_set_active call from resume path which is not necessary. This fix has been tested on OMAP4430. Signed-off-by: Partha Basak Signed-off-by: Vishwanath BS Acked-by: Rafael J. Wysocki Cc: Kevin Hilman Cc: Ben Dooks Signed-off-by: Jean Delvare diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c index 6649176..13927d5 100644 --- a/drivers/i2c/i2c-core.c +++ b/drivers/i2c/i2c-core.c @@ -216,12 +216,6 @@ static int i2c_device_pm_resume(struct device *dev) else ret = i2c_legacy_resume(dev); - if (!ret) { - pm_runtime_disable(dev); - pm_runtime_set_active(dev); - pm_runtime_enable(dev); - } - return ret; } -- cgit v0.10.2 From 6abb930af064fb1cf4177d32e2c7bfb89eee0fe5 Mon Sep 17 00:00:00 2001 From: Yegor Yefremov Date: Thu, 30 Sep 2010 14:14:22 +0200 Subject: i2c-pca: Fix waitforcompletion() return value ret is still -1, if during the polling read_byte() returns at once with I2C_PCA_CON_SI set. So ret > 0 would lead *_waitforcompletion() to return 0, in spite of the proper behavior. The routine was rewritten, so that ret has always a proper value, before returning. Signed-off-by: Yegor Yefremov Reviewed-by: Wolfram Sang Cc: stable@kernel.org Signed-off-by: Jean Delvare diff --git a/drivers/i2c/busses/i2c-pca-isa.c b/drivers/i2c/busses/i2c-pca-isa.c index bbd7760..29933f8 100644 --- a/drivers/i2c/busses/i2c-pca-isa.c +++ b/drivers/i2c/busses/i2c-pca-isa.c @@ -71,8 +71,8 @@ static int pca_isa_readbyte(void *pd, int reg) static int pca_isa_waitforcompletion(void *pd) { - long ret = ~0; unsigned long timeout; + long ret; if (irq > -1) { ret = wait_event_timeout(pca_wait, @@ -81,11 +81,15 @@ static int pca_isa_waitforcompletion(void *pd) } else { /* Do polling */ timeout = jiffies + pca_isa_ops.timeout; - while (((pca_isa_readbyte(pd, I2C_PCA_CON) - & I2C_PCA_CON_SI) == 0) - && (ret = time_before(jiffies, timeout))) + do { + ret = time_before(jiffies, timeout); + if (pca_isa_readbyte(pd, I2C_PCA_CON) + & I2C_PCA_CON_SI) + break; udelay(100); + } while (ret); } + return ret > 0; } diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c index ef5c784..5f6d7f8 100644 --- a/drivers/i2c/busses/i2c-pca-platform.c +++ b/drivers/i2c/busses/i2c-pca-platform.c @@ -80,8 +80,8 @@ static void i2c_pca_pf_writebyte32(void *pd, int reg, int val) static int i2c_pca_pf_waitforcompletion(void *pd) { struct i2c_pca_pf_data *i2c = pd; - long ret = ~0; unsigned long timeout; + long ret; if (i2c->irq) { ret = wait_event_timeout(i2c->wait, @@ -90,10 +90,13 @@ static int i2c_pca_pf_waitforcompletion(void *pd) } else { /* Do polling */ timeout = jiffies + i2c->adap.timeout; - while (((i2c->algo_data.read_byte(i2c, I2C_PCA_CON) - & I2C_PCA_CON_SI) == 0) - && (ret = time_before(jiffies, timeout))) + do { + ret = time_before(jiffies, timeout); + if (i2c->algo_data.read_byte(i2c, I2C_PCA_CON) + & I2C_PCA_CON_SI) + break; udelay(100); + } while (ret); } return ret > 0; -- cgit v0.10.2 From 64b4782fc9e4bdc59ae90897e2258e4ec938690e Mon Sep 17 00:00:00 2001 From: Rajendra Nayak Date: Thu, 30 Sep 2010 14:14:22 +0200 Subject: i2c: Fix checks which cause legacy suspend to never get called For devices which are not adapted to runtime PM a call to pm_runtime_suspended always returns true. Hence the pm_runtime_suspended checks below prevent legacy suspend from getting called. So do a pm_runtime_suspended check only for devices with a dev_pm_ops populated (which hence do not rely on the legacy suspend.) Signed-off-by: Rajendra Nayak Acked-by: Rafael J. Wysocki Cc: Ben Dooks Cc: Mark Brown Cc: Kevin Hilman Signed-off-by: Jean Delvare diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c index 13927d5..9b3cac1 100644 --- a/drivers/i2c/i2c-core.c +++ b/drivers/i2c/i2c-core.c @@ -197,11 +197,12 @@ static int i2c_device_pm_suspend(struct device *dev) { const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL; - if (pm_runtime_suspended(dev)) - return 0; - - if (pm) - return pm->suspend ? pm->suspend(dev) : 0; + if (pm) { + if (pm_runtime_suspended(dev)) + return 0; + else + return pm->suspend ? pm->suspend(dev) : 0; + } return i2c_legacy_suspend(dev, PMSG_SUSPEND); } @@ -223,11 +224,12 @@ static int i2c_device_pm_freeze(struct device *dev) { const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL; - if (pm_runtime_suspended(dev)) - return 0; - - if (pm) - return pm->freeze ? pm->freeze(dev) : 0; + if (pm) { + if (pm_runtime_suspended(dev)) + return 0; + else + return pm->freeze ? pm->freeze(dev) : 0; + } return i2c_legacy_suspend(dev, PMSG_FREEZE); } @@ -236,11 +238,12 @@ static int i2c_device_pm_thaw(struct device *dev) { const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL; - if (pm_runtime_suspended(dev)) - return 0; - - if (pm) - return pm->thaw ? pm->thaw(dev) : 0; + if (pm) { + if (pm_runtime_suspended(dev)) + return 0; + else + return pm->thaw ? pm->thaw(dev) : 0; + } return i2c_legacy_resume(dev); } @@ -249,11 +252,12 @@ static int i2c_device_pm_poweroff(struct device *dev) { const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL; - if (pm_runtime_suspended(dev)) - return 0; - - if (pm) - return pm->poweroff ? pm->poweroff(dev) : 0; + if (pm) { + if (pm_runtime_suspended(dev)) + return 0; + else + return pm->poweroff ? pm->poweroff(dev) : 0; + } return i2c_legacy_suspend(dev, PMSG_HIBERNATE); } -- cgit v0.10.2 From 925bb9c649cf8d7200549b395f2ae291833dd494 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 30 Sep 2010 14:14:23 +0200 Subject: of/i2c: Fix module load order issue caused by of_i2c.c Commit 959e85f7, "i2c: add OF-style registration and binding" caused a module dependency loop where of_i2c.c calls functions in i2c-core, and i2c-core calls of_i2c_register_devices() in of_i2c. This means that when i2c support is built as a module when CONFIG_OF is set, then neither i2c_core nor of_i2c are able to be loaded. This patch fixes the problem by moving the of_i2c_register_devices() calls back into the device drivers. Device drivers already specifically request the core code to parse the device tree for devices anyway by setting the of_node pointer, so it isn't a big deal to also call the registration function. The drivers just become slightly more verbose. Signed-off-by: Grant Likely Signed-off-by: Jean Delvare diff --git a/drivers/i2c/busses/i2c-cpm.c b/drivers/i2c/busses/i2c-cpm.c index f7bd261..f2de3be 100644 --- a/drivers/i2c/busses/i2c-cpm.c +++ b/drivers/i2c/busses/i2c-cpm.c @@ -677,6 +677,11 @@ static int __devinit cpm_i2c_probe(struct platform_device *ofdev, dev_dbg(&ofdev->dev, "hw routines for %s registered.\n", cpm->adap.name); + /* + * register OF I2C devices + */ + of_i2c_register_devices(&cpm->adap); + return 0; out_shut: cpm_i2c_shutdown(cpm); diff --git a/drivers/i2c/busses/i2c-ibm_iic.c b/drivers/i2c/busses/i2c-ibm_iic.c index 43ca32f..89eedf4 100644 --- a/drivers/i2c/busses/i2c-ibm_iic.c +++ b/drivers/i2c/busses/i2c-ibm_iic.c @@ -761,6 +761,9 @@ static int __devinit iic_probe(struct platform_device *ofdev, dev_info(&ofdev->dev, "using %s mode\n", dev->fast_mode ? "fast (400 kHz)" : "standard (100 kHz)"); + /* Now register all the child nodes */ + of_i2c_register_devices(adap); + return 0; error_cleanup: diff --git a/drivers/i2c/busses/i2c-mpc.c b/drivers/i2c/busses/i2c-mpc.c index a1c419a..b74e6dc 100644 --- a/drivers/i2c/busses/i2c-mpc.c +++ b/drivers/i2c/busses/i2c-mpc.c @@ -632,6 +632,7 @@ static int __devinit fsl_i2c_probe(struct platform_device *op, dev_err(i2c->dev, "failed to add adapter\n"); goto fail_add; } + of_i2c_register_devices(&i2c->adap); return result; diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c index 9b3cac1..bea4c50 100644 --- a/drivers/i2c/i2c-core.c +++ b/drivers/i2c/i2c-core.c @@ -32,7 +32,6 @@ #include #include #include -#include #include #include #include @@ -874,9 +873,6 @@ static int i2c_register_adapter(struct i2c_adapter *adap) if (adap->nr < __i2c_first_dynamic_bus_num) i2c_scan_static_board_info(adap); - /* Register devices from the device tree */ - of_i2c_register_devices(adap); - /* Notify drivers */ mutex_lock(&core_lock); bus_for_each_drv(&i2c_bus_type, NULL, adap, __process_new_adapter); -- cgit v0.10.2 From 8c462b6047da80491b8cb6be878e8bf9313ac3e1 Mon Sep 17 00:00:00 2001 From: Mat Martineau Date: Tue, 24 Aug 2010 15:35:42 -0700 Subject: Bluetooth: Only enable L2CAP FCS for ERTM or streaming This fixes a bug which caused the FCS setting to show L2CAP_FCS_CRC16 with L2CAP modes other than ERTM or streaming. At present, this only affects the FCS value shown with getsockopt() for basic mode. Signed-off-by: Mat Martineau Signed-off-by: Gustavo F. Padovan diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index fadf26b..c784703 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -3071,6 +3071,17 @@ static inline int l2cap_connect_rsp(struct l2cap_conn *conn, struct l2cap_cmd_hd return 0; } +static inline void set_default_fcs(struct l2cap_pinfo *pi) +{ + /* FCS is enabled only in ERTM or streaming mode, if one or both + * sides request it. + */ + if (pi->mode != L2CAP_MODE_ERTM && pi->mode != L2CAP_MODE_STREAMING) + pi->fcs = L2CAP_FCS_NONE; + else if (!(pi->conf_state & L2CAP_CONF_NO_FCS_RECV)) + pi->fcs = L2CAP_FCS_CRC16; +} + static inline int l2cap_config_req(struct l2cap_conn *conn, struct l2cap_cmd_hdr *cmd, u16 cmd_len, u8 *data) { struct l2cap_conf_req *req = (struct l2cap_conf_req *) data; @@ -3135,9 +3146,7 @@ static inline int l2cap_config_req(struct l2cap_conn *conn, struct l2cap_cmd_hdr goto unlock; if (l2cap_pi(sk)->conf_state & L2CAP_CONF_INPUT_DONE) { - if (!(l2cap_pi(sk)->conf_state & L2CAP_CONF_NO_FCS_RECV) || - l2cap_pi(sk)->fcs != L2CAP_FCS_NONE) - l2cap_pi(sk)->fcs = L2CAP_FCS_CRC16; + set_default_fcs(l2cap_pi(sk)); sk->sk_state = BT_CONNECTED; @@ -3225,9 +3234,7 @@ static inline int l2cap_config_rsp(struct l2cap_conn *conn, struct l2cap_cmd_hdr l2cap_pi(sk)->conf_state |= L2CAP_CONF_INPUT_DONE; if (l2cap_pi(sk)->conf_state & L2CAP_CONF_OUTPUT_DONE) { - if (!(l2cap_pi(sk)->conf_state & L2CAP_CONF_NO_FCS_RECV) || - l2cap_pi(sk)->fcs != L2CAP_FCS_NONE) - l2cap_pi(sk)->fcs = L2CAP_FCS_CRC16; + set_default_fcs(l2cap_pi(sk)); sk->sk_state = BT_CONNECTED; l2cap_pi(sk)->next_tx_seq = 0; -- cgit v0.10.2 From 8183b775bc5b79b6b1e250019c9dd930554dfa94 Mon Sep 17 00:00:00 2001 From: Andrei Emeltchenko Date: Wed, 1 Sep 2010 15:17:25 +0300 Subject: Bluetooth: fix MTU L2CAP configuration parameter When receiving L2CAP negative configuration response with respect to MTU parameter we modify wrong field. MTU here means proposed value of MTU that the remote device intends to transmit. So for local L2CAP socket it is pi->imtu. Signed-off-by: Andrei Emeltchenko Acked-by: Ville Tervo Signed-off-by: Gustavo F. Padovan diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index c784703..9fad312 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -2771,10 +2771,10 @@ static int l2cap_parse_conf_rsp(struct sock *sk, void *rsp, int len, void *data, case L2CAP_CONF_MTU: if (val < L2CAP_DEFAULT_MIN_MTU) { *result = L2CAP_CONF_UNACCEPT; - pi->omtu = L2CAP_DEFAULT_MIN_MTU; + pi->imtu = L2CAP_DEFAULT_MIN_MTU; } else - pi->omtu = val; - l2cap_add_conf_opt(&ptr, L2CAP_CONF_MTU, 2, pi->omtu); + pi->imtu = val; + l2cap_add_conf_opt(&ptr, L2CAP_CONF_MTU, 2, pi->imtu); break; case L2CAP_CONF_FLUSH_TO: -- cgit v0.10.2 From ccbb84af28594e19fd4bf27ff2828c80d03b6081 Mon Sep 17 00:00:00 2001 From: "Gustavo F. Padovan" Date: Mon, 30 Aug 2010 18:44:44 -0300 Subject: Bluetooth: Simplify L2CAP Streaming mode sending As we don't have any error control on the Streaming mode, i.e., we don't need to keep a copy of the skb for later resending we don't need to call skb_clone() on it. Then we can go one further here, and dequeue the skb before sending it, that also means we don't need to look to sk->sk_send_head anymore. The patch saves memory and time when sending Streaming mode data, so it is good to mainline. Signed-off-by: Gustavo F. Padovan diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index 9fad312..f2062aa 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -1441,33 +1441,23 @@ static inline void l2cap_do_send(struct sock *sk, struct sk_buff *skb) static void l2cap_streaming_send(struct sock *sk) { - struct sk_buff *skb, *tx_skb; + struct sk_buff *skb; struct l2cap_pinfo *pi = l2cap_pi(sk); u16 control, fcs; - while ((skb = sk->sk_send_head)) { - tx_skb = skb_clone(skb, GFP_ATOMIC); - - control = get_unaligned_le16(tx_skb->data + L2CAP_HDR_SIZE); + while ((skb = skb_dequeue(TX_QUEUE(sk)))) { + control = get_unaligned_le16(skb->data + L2CAP_HDR_SIZE); control |= pi->next_tx_seq << L2CAP_CTRL_TXSEQ_SHIFT; - put_unaligned_le16(control, tx_skb->data + L2CAP_HDR_SIZE); + put_unaligned_le16(control, skb->data + L2CAP_HDR_SIZE); if (pi->fcs == L2CAP_FCS_CRC16) { - fcs = crc16(0, (u8 *)tx_skb->data, tx_skb->len - 2); - put_unaligned_le16(fcs, tx_skb->data + tx_skb->len - 2); + fcs = crc16(0, (u8 *)skb->data, skb->len - 2); + put_unaligned_le16(fcs, skb->data + skb->len - 2); } - l2cap_do_send(sk, tx_skb); + l2cap_do_send(sk, skb); pi->next_tx_seq = (pi->next_tx_seq + 1) % 64; - - if (skb_queue_is_last(TX_QUEUE(sk), skb)) - sk->sk_send_head = NULL; - else - sk->sk_send_head = skb_queue_next(TX_QUEUE(sk), skb); - - skb = skb_dequeue(TX_QUEUE(sk)); - kfree_skb(skb); } } -- cgit v0.10.2 From fad003b6c8e3d944d4453fd569b0702ef1af82b3 Mon Sep 17 00:00:00 2001 From: "Gustavo F. Padovan" Date: Sat, 14 Aug 2010 00:48:07 -0300 Subject: Bluetooth: Fix inconsistent lock state with RFCOMM When receiving a rfcomm connection with the old dund deamon a inconsistent lock state happens. That's because interrupts were already disabled by l2cap_conn_start() when rfcomm_sk_state_change() try to lock the spin_lock. As result we may have a inconsistent lock state for l2cap_conn_start() after rfcomm_sk_state_change() calls bh_lock_sock() and disable interrupts as well. [ 2833.151999] [ 2833.151999] ================================= [ 2833.151999] [ INFO: inconsistent lock state ] [ 2833.151999] 2.6.36-rc3 #2 [ 2833.151999] --------------------------------- [ 2833.151999] inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage. [ 2833.151999] krfcommd/2306 [HC0[0]:SC0[0]:HE1:SE1] takes: [ 2833.151999] (slock-AF_BLUETOOTH){+.?...}, at: [] rfcomm_sk_state_change+0x46/0x170 [rfcomm] [ 2833.151999] {IN-SOFTIRQ-W} state was registered at: [ 2833.151999] [] __lock_acquire+0x5b6/0x1560 [ 2833.151999] [] lock_acquire+0x5a/0x70 [ 2833.151999] [] _raw_spin_lock+0x2c/0x40 [ 2833.151999] [] l2cap_conn_start+0x92/0x640 [l2cap] [ 2833.151999] [] l2cap_sig_channel+0x6bf/0x1320 [l2cap] [ 2833.151999] [] l2cap_recv_frame+0x133/0x770 [l2cap] [ 2833.151999] [] l2cap_recv_acldata+0x1cb/0x390 [l2cap] [ 2833.151999] [] hci_rx_task+0x2ab/0x450 [bluetooth] [ 2833.151999] [] tasklet_action+0xcb/0xe0 [ 2833.151999] [] __do_softirq+0xae/0x150 [ 2833.151999] [] call_softirq+0x1c/0x30 [ 2833.151999] [] do_softirq+0x75/0xb0 [ 2833.151999] [] irq_exit+0x8d/0xa0 [ 2833.151999] [] smp_apic_timer_interrupt+0x6b/0xa0 [ 2833.151999] [] apic_timer_interrupt+0x13/0x20 [ 2833.151999] [] cpu_idle+0x5a/0xb0 [ 2833.151999] [] rest_init+0xad/0xc0 [ 2833.151999] [] start_kernel+0x2dd/0x2e8 [ 2833.151999] [] x86_64_start_reservations+0xf6/0xfa [ 2833.151999] [] x86_64_start_kernel+0xe4/0xeb [ 2833.151999] irq event stamp: 731 [ 2833.151999] hardirqs last enabled at (731): [] local_bh_enable_ip+0x82/0xe0 [ 2833.151999] hardirqs last disabled at (729): [] __do_softirq+0xce/0x150 [ 2833.151999] softirqs last enabled at (730): [] __do_softirq+0xfe/0x150 [ 2833.151999] softirqs last disabled at (711): [] call_softirq+0x1c/0x30 [ 2833.151999] [ 2833.151999] other info that might help us debug this: [ 2833.151999] 2 locks held by krfcommd/2306: [ 2833.151999] #0: (rfcomm_mutex){+.+.+.}, at: [] rfcomm_run+0x174/0xb20 [rfcomm] [ 2833.151999] #1: (&(&d->lock)->rlock){+.+...}, at: [] rfcomm_dlc_accept+0x53/0x100 [rfcomm] [ 2833.151999] [ 2833.151999] stack backtrace: [ 2833.151999] Pid: 2306, comm: krfcommd Tainted: G W 2.6.36-rc3 #2 [ 2833.151999] Call Trace: [ 2833.151999] [] print_usage_bug+0x171/0x180 [ 2833.151999] [] mark_lock+0x333/0x400 [ 2833.151999] [] __lock_acquire+0x63a/0x1560 [ 2833.151999] [] ? __lock_acquire+0xb25/0x1560 [ 2833.151999] [] lock_acquire+0x5a/0x70 [ 2833.151999] [] ? rfcomm_sk_state_change+0x46/0x170 [rfcomm] [ 2833.151999] [] _raw_spin_lock+0x2c/0x40 [ 2833.151999] [] ? rfcomm_sk_state_change+0x46/0x170 [rfcomm] [ 2833.151999] [] rfcomm_sk_state_change+0x46/0x170 [rfcomm] [ 2833.151999] [] rfcomm_dlc_accept+0x69/0x100 [rfcomm] [ 2833.151999] [] rfcomm_check_accept+0x59/0xd0 [rfcomm] [ 2833.151999] [] rfcomm_recv_frame+0x9fb/0x1320 [rfcomm] [ 2833.151999] [] ? _raw_spin_unlock_irqrestore+0x3b/0x60 [ 2833.151999] [] ? trace_hardirqs_on_caller+0x13d/0x180 [ 2833.151999] [] ? trace_hardirqs_on+0xd/0x10 [ 2833.151999] [] rfcomm_run+0x221/0xb20 [rfcomm] [ 2833.151999] [] ? schedule+0x287/0x780 [ 2833.151999] [] ? rfcomm_run+0x0/0xb20 [rfcomm] [ 2833.151999] [] kthread+0x96/0xa0 [ 2833.151999] [] kernel_thread_helper+0x4/0x10 [ 2833.151999] [] ? restore_args+0x0/0x30 [ 2833.151999] [] ? kthread+0x0/0xa0 [ 2833.151999] [] ? kernel_thread_helper+0x0/0x10 Signed-off-by: Gustavo F. Padovan diff --git a/net/bluetooth/rfcomm/sock.c b/net/bluetooth/rfcomm/sock.c index 44a6232..194b3a0 100644 --- a/net/bluetooth/rfcomm/sock.c +++ b/net/bluetooth/rfcomm/sock.c @@ -82,11 +82,14 @@ static void rfcomm_sk_data_ready(struct rfcomm_dlc *d, struct sk_buff *skb) static void rfcomm_sk_state_change(struct rfcomm_dlc *d, int err) { struct sock *sk = d->owner, *parent; + unsigned long flags; + if (!sk) return; BT_DBG("dlc %p state %ld err %d", d, d->state, err); + local_irq_save(flags); bh_lock_sock(sk); if (err) @@ -108,6 +111,7 @@ static void rfcomm_sk_state_change(struct rfcomm_dlc *d, int err) } bh_unlock_sock(sk); + local_irq_restore(flags); if (parent && sock_flag(sk, SOCK_ZAPPED)) { /* We have to drop DLC lock here, otherwise -- cgit v0.10.2 From b0239c80fe89d5832a68a0f3121a9d5ec9fb763e Mon Sep 17 00:00:00 2001 From: "Gustavo F. Padovan" Date: Wed, 8 Sep 2010 14:59:44 -0300 Subject: Revert "Bluetooth: Don't accept ConfigReq if we aren't in the BT_CONFIG state" This reverts commit 8cb8e6f1684be13b51f8429b15f39c140326b327. That commit introduced a regression with the Bluetooth Profile Tuning Suite(PTS), Reverting this make sure that L2CAP is in a qualificable state. Signed-off-by: Gustavo F. Padovan diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index f2062aa..44a8fb0 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -3089,14 +3089,8 @@ static inline int l2cap_config_req(struct l2cap_conn *conn, struct l2cap_cmd_hdr if (!sk) return -ENOENT; - if (sk->sk_state != BT_CONFIG) { - struct l2cap_cmd_rej rej; - - rej.reason = cpu_to_le16(0x0002); - l2cap_send_cmd(conn, cmd->ident, L2CAP_COMMAND_REJ, - sizeof(rej), &rej); + if (sk->sk_state == BT_DISCONN) goto unlock; - } /* Reject if config buffer is too small. */ len = cmd_len - sizeof(*req); -- cgit v0.10.2 From e454c844644683571617896ab2a4ce0109c1943e Mon Sep 17 00:00:00 2001 From: "Gustavo F. Padovan" Date: Tue, 21 Sep 2010 16:31:11 -0300 Subject: Bluetooth: Fix deadlock in the ERTM logic The Enhanced Retransmission Mode(ERTM) is a realiable mode of operation of the Bluetooth L2CAP layer. Think on it like a simplified version of TCP. The problem we were facing here was a deadlock. ERTM uses a backlog queue to queue incomimg packets while the user is helding the lock. At some moment the sk_sndbuf can be exceeded and we can't alloc new skbs then the code sleep with the lock to wait for memory, that stalls the ERTM connection once we can't read the acknowledgements packets in the backlog queue to free memory and make the allocation of outcoming skb successful. This patch actually affect all users of bt_skb_send_alloc(), i.e., all L2CAP modes and SCO. We are safe against socket states changes or channels deletion while the we are sleeping wait memory. Checking for the sk->sk_err and sk->sk_shutdown make the code safe, since any action that can leave the socket or the channel in a not usable state set one of the struct members at least. Then we can check both of them when getting the lock again and return with the proper error if something unexpected happens. Signed-off-by: Gustavo F. Padovan Signed-off-by: Ulisses Furquim diff --git a/include/net/bluetooth/bluetooth.h b/include/net/bluetooth/bluetooth.h index 27a902d..30fce01 100644 --- a/include/net/bluetooth/bluetooth.h +++ b/include/net/bluetooth/bluetooth.h @@ -161,12 +161,30 @@ static inline struct sk_buff *bt_skb_send_alloc(struct sock *sk, unsigned long l { struct sk_buff *skb; + release_sock(sk); if ((skb = sock_alloc_send_skb(sk, len + BT_SKB_RESERVE, nb, err))) { skb_reserve(skb, BT_SKB_RESERVE); bt_cb(skb)->incoming = 0; } + lock_sock(sk); + + if (!skb && *err) + return NULL; + + *err = sock_error(sk); + if (*err) + goto out; + + if (sk->sk_shutdown) { + *err = -ECONNRESET; + goto out; + } return skb; + +out: + kfree_skb(skb); + return NULL; } int bt_err(__u16 code); -- cgit v0.10.2 From 46290341cd649c2bfb69e5067c1804c0395c83a1 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Thu, 30 Sep 2010 10:34:00 -0400 Subject: GFS2 fatal: filesystem consistency error on rename This patch fixes a GFS2 problem whereby the first rename after a mount can result in a file system consistency error being flagged improperly and cause the file system to withdraw. The problem is that the rename code tries to run the rgrp list with function gfs2_blk2rgrpd before the rgrp list is guaranteed to be read in from disk. The patch makes the rename function hold the rindex glock (as the gfs2_unlink code does today) which reads in the rgrp list if need be. There were a total of three places in the rename code that improperly referenced the rgrp list without the rindex glock and this patch fixes all three. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index fba0017..0534510 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -736,7 +736,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, struct gfs2_inode *ip = GFS2_I(odentry->d_inode); struct gfs2_inode *nip = NULL; struct gfs2_sbd *sdp = GFS2_SB(odir); - struct gfs2_holder ghs[5], r_gh = { .gh_gl = NULL, }; + struct gfs2_holder ghs[5], r_gh = { .gh_gl = NULL, }, ri_gh; struct gfs2_rgrpd *nrgd; unsigned int num_gh; int dir_rename = 0; @@ -750,6 +750,9 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, return 0; } + error = gfs2_rindex_hold(sdp, &ri_gh); + if (error) + return error; if (odip != ndip) { error = gfs2_glock_nq_init(sdp->sd_rename_gl, LM_ST_EXCLUSIVE, @@ -879,7 +882,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, al->al_requested = sdp->sd_max_dirres; - error = gfs2_inplace_reserve(ndip); + error = gfs2_inplace_reserve_ri(ndip); if (error) goto out_gunlock_q; @@ -961,6 +964,7 @@ out_gunlock_r: if (r_gh.gh_gl) gfs2_glock_dq_uninit(&r_gh); out: + gfs2_glock_dq_uninit(&ri_gh); return error; } diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index f9ddcf4..fb67f593 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1200,7 +1200,8 @@ out: * Returns: errno */ -int gfs2_inplace_reserve_i(struct gfs2_inode *ip, char *file, unsigned int line) +int gfs2_inplace_reserve_i(struct gfs2_inode *ip, int hold_rindex, + char *file, unsigned int line) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_alloc *al = ip->i_alloc; @@ -1211,12 +1212,15 @@ int gfs2_inplace_reserve_i(struct gfs2_inode *ip, char *file, unsigned int line) return -EINVAL; try_again: - /* We need to hold the rindex unless the inode we're using is - the rindex itself, in which case it's already held. */ - if (ip != GFS2_I(sdp->sd_rindex)) - error = gfs2_rindex_hold(sdp, &al->al_ri_gh); - else if (!sdp->sd_rgrps) /* We may not have the rindex read in, so: */ - error = gfs2_ri_update_special(ip); + if (hold_rindex) { + /* We need to hold the rindex unless the inode we're using is + the rindex itself, in which case it's already held. */ + if (ip != GFS2_I(sdp->sd_rindex)) + error = gfs2_rindex_hold(sdp, &al->al_ri_gh); + else if (!sdp->sd_rgrps) /* We may not have the rindex read + in, so: */ + error = gfs2_ri_update_special(ip); + } if (error) return error; @@ -1227,7 +1231,7 @@ try_again: try to free it, and try the allocation again. */ error = get_local_rgrp(ip, &unlinked, &last_unlinked); if (error) { - if (ip != GFS2_I(sdp->sd_rindex)) + if (hold_rindex && ip != GFS2_I(sdp->sd_rindex)) gfs2_glock_dq_uninit(&al->al_ri_gh); if (error != -EAGAIN) return error; @@ -1269,7 +1273,7 @@ void gfs2_inplace_release(struct gfs2_inode *ip) al->al_rgd = NULL; if (al->al_rgd_gh.gh_gl) gfs2_glock_dq_uninit(&al->al_rgd_gh); - if (ip != GFS2_I(sdp->sd_rindex)) + if (ip != GFS2_I(sdp->sd_rindex) && al->al_ri_gh.gh_gl) gfs2_glock_dq_uninit(&al->al_ri_gh); } diff --git a/fs/gfs2/rgrp.h b/fs/gfs2/rgrp.h index f07119d..0e35c04 100644 --- a/fs/gfs2/rgrp.h +++ b/fs/gfs2/rgrp.h @@ -39,10 +39,12 @@ static inline void gfs2_alloc_put(struct gfs2_inode *ip) ip->i_alloc = NULL; } -extern int gfs2_inplace_reserve_i(struct gfs2_inode *ip, char *file, - unsigned int line); +extern int gfs2_inplace_reserve_i(struct gfs2_inode *ip, int hold_rindex, + char *file, unsigned int line); #define gfs2_inplace_reserve(ip) \ -gfs2_inplace_reserve_i((ip), __FILE__, __LINE__) + gfs2_inplace_reserve_i((ip), 1, __FILE__, __LINE__) +#define gfs2_inplace_reserve_ri(ip) \ + gfs2_inplace_reserve_i((ip), 0, __FILE__, __LINE__) extern void gfs2_inplace_release(struct gfs2_inode *ip); -- cgit v0.10.2 From 249e6353001e407edf5c9a74482ecfca90c8ff33 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:41:27 +0200 Subject: hfsplus: fix BKL leak in hfsplus_ioctl Currenly the HFSPLUS_IOC_EXT2_GETFLAGS case never unlocks the BKL, which can lead to easily reproduced lockups when doing multiple GETFLAGS ioctls. Fix this by only taking the BKL for the HFSPLUS_IOC_EXT2_SETFLAGS case as neither HFSPLUS_IOC_EXT2_GETFLAGS not the default error case needs it. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index ac405f0..59dc402 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -26,7 +26,6 @@ long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) struct inode *inode = filp->f_path.dentry->d_inode; unsigned int flags; - lock_kernel(); switch (cmd) { case HFSPLUS_IOC_EXT2_GETFLAGS: flags = 0; @@ -39,6 +38,8 @@ long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) return put_user(flags, (int __user *)arg); case HFSPLUS_IOC_EXT2_SETFLAGS: { int err = 0; + + lock_kernel(); err = mnt_want_write(filp->f_path.mnt); if (err) { unlock_kernel(); @@ -93,7 +94,6 @@ setflags_out: return err; } default: - unlock_kernel(); return -ENOTTY; } } -- cgit v0.10.2 From 94744567fef9602c3d8218a1d8f58c04cce354f6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:41:31 +0200 Subject: hfsplus: split hfsplus_ioctl Give each ioctl command a function of it's own. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index 59dc402..0e359c3 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -21,78 +21,93 @@ #include #include "hfsplus_fs.h" -long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) +static int hfsplus_ioctl_getflags(struct file *file, int __user *user_flags) { - struct inode *inode = filp->f_path.dentry->d_inode; + struct inode *inode = file->f_path.dentry->d_inode; + unsigned int flags = 0; + + if (HFSPLUS_I(inode).rootflags & HFSPLUS_FLG_IMMUTABLE) + flags |= FS_IMMUTABLE_FL; + if (HFSPLUS_I(inode).rootflags & HFSPLUS_FLG_APPEND) + flags |= FS_APPEND_FL; + if (HFSPLUS_I(inode).userflags & HFSPLUS_FLG_NODUMP) + flags |= FS_NODUMP_FL; + + return put_user(flags, user_flags); +} + +static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) +{ + struct inode *inode = file->f_path.dentry->d_inode; unsigned int flags; + int err = 0; - switch (cmd) { - case HFSPLUS_IOC_EXT2_GETFLAGS: - flags = 0; - if (HFSPLUS_I(inode).rootflags & HFSPLUS_FLG_IMMUTABLE) - flags |= FS_IMMUTABLE_FL; /* EXT2_IMMUTABLE_FL */ - if (HFSPLUS_I(inode).rootflags & HFSPLUS_FLG_APPEND) - flags |= FS_APPEND_FL; /* EXT2_APPEND_FL */ - if (HFSPLUS_I(inode).userflags & HFSPLUS_FLG_NODUMP) - flags |= FS_NODUMP_FL; /* EXT2_NODUMP_FL */ - return put_user(flags, (int __user *)arg); - case HFSPLUS_IOC_EXT2_SETFLAGS: { - int err = 0; - - lock_kernel(); - err = mnt_want_write(filp->f_path.mnt); - if (err) { - unlock_kernel(); - return err; - } + lock_kernel(); + err = mnt_want_write(file->f_path.mnt); + if (err) + goto out_unlock_kernel; - if (!is_owner_or_cap(inode)) { - err = -EACCES; - goto setflags_out; - } - if (get_user(flags, (int __user *)arg)) { - err = -EFAULT; - goto setflags_out; - } - if (flags & (FS_IMMUTABLE_FL|FS_APPEND_FL) || - HFSPLUS_I(inode).rootflags & (HFSPLUS_FLG_IMMUTABLE|HFSPLUS_FLG_APPEND)) { - if (!capable(CAP_LINUX_IMMUTABLE)) { - err = -EPERM; - goto setflags_out; - } - } + if (!is_owner_or_cap(inode)) { + err = -EACCES; + goto out_drop_write; + } - /* don't silently ignore unsupported ext2 flags */ - if (flags & ~(FS_IMMUTABLE_FL|FS_APPEND_FL|FS_NODUMP_FL)) { - err = -EOPNOTSUPP; - goto setflags_out; - } - if (flags & FS_IMMUTABLE_FL) { /* EXT2_IMMUTABLE_FL */ - inode->i_flags |= S_IMMUTABLE; - HFSPLUS_I(inode).rootflags |= HFSPLUS_FLG_IMMUTABLE; - } else { - inode->i_flags &= ~S_IMMUTABLE; - HFSPLUS_I(inode).rootflags &= ~HFSPLUS_FLG_IMMUTABLE; - } - if (flags & FS_APPEND_FL) { /* EXT2_APPEND_FL */ - inode->i_flags |= S_APPEND; - HFSPLUS_I(inode).rootflags |= HFSPLUS_FLG_APPEND; - } else { - inode->i_flags &= ~S_APPEND; - HFSPLUS_I(inode).rootflags &= ~HFSPLUS_FLG_APPEND; + if (get_user(flags, user_flags)) { + err = -EFAULT; + goto out_drop_write; + } + + if (flags & (FS_IMMUTABLE_FL|FS_APPEND_FL) || + HFSPLUS_I(inode).rootflags & (HFSPLUS_FLG_IMMUTABLE|HFSPLUS_FLG_APPEND)) { + if (!capable(CAP_LINUX_IMMUTABLE)) { + err = -EPERM; + goto out_drop_write; } - if (flags & FS_NODUMP_FL) /* EXT2_NODUMP_FL */ - HFSPLUS_I(inode).userflags |= HFSPLUS_FLG_NODUMP; - else - HFSPLUS_I(inode).userflags &= ~HFSPLUS_FLG_NODUMP; - - inode->i_ctime = CURRENT_TIME_SEC; - mark_inode_dirty(inode); -setflags_out: - mnt_drop_write(filp->f_path.mnt); - unlock_kernel(); - return err; } + + /* don't silently ignore unsupported ext2 flags */ + if (flags & ~(FS_IMMUTABLE_FL|FS_APPEND_FL|FS_NODUMP_FL)) { + err = -EOPNOTSUPP; + goto out_drop_write; + } + if (flags & FS_IMMUTABLE_FL) { + inode->i_flags |= S_IMMUTABLE; + HFSPLUS_I(inode).rootflags |= HFSPLUS_FLG_IMMUTABLE; + } else { + inode->i_flags &= ~S_IMMUTABLE; + HFSPLUS_I(inode).rootflags &= ~HFSPLUS_FLG_IMMUTABLE; + } + if (flags & FS_APPEND_FL) { + inode->i_flags |= S_APPEND; + HFSPLUS_I(inode).rootflags |= HFSPLUS_FLG_APPEND; + } else { + inode->i_flags &= ~S_APPEND; + HFSPLUS_I(inode).rootflags &= ~HFSPLUS_FLG_APPEND; + } + if (flags & FS_NODUMP_FL) + HFSPLUS_I(inode).userflags |= HFSPLUS_FLG_NODUMP; + else + HFSPLUS_I(inode).userflags &= ~HFSPLUS_FLG_NODUMP; + + inode->i_ctime = CURRENT_TIME_SEC; + mark_inode_dirty(inode); + +out_drop_write: + mnt_drop_write(file->f_path.mnt); +out_unlock_kernel: + unlock_kernel(); + return err; +} + +long hfsplus_ioctl(struct file *file, unsigned int cmd, unsigned long arg) +{ + void __user *argp = (void __user *)arg; + + switch (cmd) { + case HFSPLUS_IOC_EXT2_GETFLAGS: + return hfsplus_ioctl_getflags(file, argp); + case HFSPLUS_IOC_EXT2_SETFLAGS: + return hfsplus_ioctl_setflags(file, argp); default: return -ENOTTY; } -- cgit v0.10.2 From 6333816ade7e04a96ec0a34a8378c455e4f7c4dd Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:41:35 +0200 Subject: hfsplus: protect setflags using i_mutex Use i_mutex for protecting against concurrent setflags ioctls like in other filesystems and get rid of the BKL in hfsplus_ioctl. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index 0e359c3..906bd3dd 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -17,7 +17,6 @@ #include #include #include -#include #include #include "hfsplus_fs.h" @@ -42,10 +41,9 @@ static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) unsigned int flags; int err = 0; - lock_kernel(); err = mnt_want_write(file->f_path.mnt); if (err) - goto out_unlock_kernel; + goto out; if (!is_owner_or_cap(inode)) { err = -EACCES; @@ -57,18 +55,20 @@ static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) goto out_drop_write; } + mutex_lock(&inode->i_mutex); + if (flags & (FS_IMMUTABLE_FL|FS_APPEND_FL) || HFSPLUS_I(inode).rootflags & (HFSPLUS_FLG_IMMUTABLE|HFSPLUS_FLG_APPEND)) { if (!capable(CAP_LINUX_IMMUTABLE)) { err = -EPERM; - goto out_drop_write; + goto out_unlock_inode; } } /* don't silently ignore unsupported ext2 flags */ if (flags & ~(FS_IMMUTABLE_FL|FS_APPEND_FL|FS_NODUMP_FL)) { err = -EOPNOTSUPP; - goto out_drop_write; + goto out_unlock_inode; } if (flags & FS_IMMUTABLE_FL) { inode->i_flags |= S_IMMUTABLE; @@ -92,10 +92,11 @@ static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); +out_unlock_inode: + mutex_lock(&inode->i_mutex); out_drop_write: mnt_drop_write(file->f_path.mnt); -out_unlock_kernel: - unlock_kernel(); +out: return err; } -- cgit v0.10.2 From 40bf48afe92fcea61e7e164f0b2599fba8b88124 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:41:39 +0200 Subject: hfsplus: introduce alloc_mutex Use a new per-sb alloc_mutex instead of abusing i_mutex of the alloc_file to protect block allocations. This gets rid of lockdep nesting warnings and prepares for extending the scope of alloc_mutex. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/bitmap.c b/fs/hfsplus/bitmap.c index ea30afc..8a78176 100644 --- a/fs/hfsplus/bitmap.c +++ b/fs/hfsplus/bitmap.c @@ -29,7 +29,7 @@ int hfsplus_block_allocate(struct super_block *sb, u32 size, u32 offset, u32 *ma return size; dprint(DBG_BITMAP, "block_allocate: %u,%u,%u\n", size, offset, len); - mutex_lock(&HFSPLUS_SB(sb).alloc_file->i_mutex); + mutex_lock(&HFSPLUS_SB(sb).alloc_mutex); mapping = HFSPLUS_SB(sb).alloc_file->i_mapping; page = read_mapping_page(mapping, offset / PAGE_CACHE_BITS, NULL); if (IS_ERR(page)) { @@ -154,7 +154,7 @@ done: sb->s_dirt = 1; dprint(DBG_BITMAP, "-> %u,%u\n", start, *max); out: - mutex_unlock(&HFSPLUS_SB(sb).alloc_file->i_mutex); + mutex_unlock(&HFSPLUS_SB(sb).alloc_mutex); return start; } @@ -175,7 +175,7 @@ int hfsplus_block_free(struct super_block *sb, u32 offset, u32 count) if ((offset + count) > HFSPLUS_SB(sb).total_blocks) return -2; - mutex_lock(&HFSPLUS_SB(sb).alloc_file->i_mutex); + mutex_lock(&HFSPLUS_SB(sb).alloc_mutex); mapping = HFSPLUS_SB(sb).alloc_file->i_mapping; pnr = offset / PAGE_CACHE_BITS; page = read_mapping_page(mapping, pnr, NULL); @@ -226,7 +226,7 @@ out: kunmap(page); HFSPLUS_SB(sb).free_blocks += len; sb->s_dirt = 1; - mutex_unlock(&HFSPLUS_SB(sb).alloc_file->i_mutex); + mutex_unlock(&HFSPLUS_SB(sb).alloc_mutex); return 0; } diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index dc856be..df0a631 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -116,6 +116,9 @@ struct hfsplus_sb_info { struct inode *hidden_dir; struct nls_table *nls; + /* synchronize block allocations */ + struct mutex alloc_mutex; + /* Runtime variables */ u32 blockoffset; u32 sect_count; diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 3b55c05..3dc62aa 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -321,6 +321,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) sb->s_fs_info = sbi; INIT_HLIST_HEAD(&sbi->rsrc_inodes); + mutex_init(&sbi->alloc_mutex); hfsplus_fill_defaults(sbi); if (!hfsplus_parse_options(data, sbi)) { printk(KERN_ERR "hfs: unable to parse mount options\n"); -- cgit v0.10.2 From a9fdbf8c6070d49c482e209df7ee93d9ec41ea27 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:41:50 +0200 Subject: hfsplus: use alloc_mutex in hfsplus_sync_fs Use alloc_mutex to protect hfsplus_sync_fs against itself and concurrent allocations, which allows to get rid of lock_super in hfsplus. Note that most fields in the superblock still aren't protected against concurrent allocations, that will follow later. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 3dc62aa..4936642 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -162,7 +162,7 @@ int hfsplus_sync_fs(struct super_block *sb, int wait) dprint(DBG_SUPER, "hfsplus_write_super\n"); - lock_super(sb); + mutex_lock(&HFSPLUS_SB(sb).alloc_mutex); sb->s_dirt = 0; vhdr->free_blocks = cpu_to_be32(HFSPLUS_SB(sb).free_blocks); @@ -195,7 +195,7 @@ int hfsplus_sync_fs(struct super_block *sb, int wait) } HFSPLUS_SB(sb).flags &= ~HFSPLUS_SB_WRITEBACKUP; } - unlock_super(sb); + mutex_unlock(&HFSPLUS_SB(sb).alloc_mutex); return 0; } -- cgit v0.10.2 From e753a62156e952fd5a3c64f98454d9aeee3a2546 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:41:53 +0200 Subject: hfsplus: remove BKL from hfsplus_put_super Except for ->put_super the BKL is now gone from HFS, which means it's superflous there too as ->put_super is serialized by the VFS. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 4936642..a1d3fd9 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -12,7 +12,6 @@ #include #include #include -#include #include #include @@ -213,8 +212,6 @@ static void hfsplus_put_super(struct super_block *sb) if (!sb->s_fs_info) return; - lock_kernel(); - if (sb->s_dirt) hfsplus_write_super(sb); if (!(sb->s_flags & MS_RDONLY) && HFSPLUS_SB(sb).s_vhdr) { @@ -235,8 +232,6 @@ static void hfsplus_put_super(struct super_block *sb) unload_nls(HFSPLUS_SB(sb).nls); kfree(sb->s_fs_info); sb->s_fs_info = NULL; - - unlock_kernel(); } static int hfsplus_statfs(struct dentry *dentry, struct kstatfs *buf) -- cgit v0.10.2 From dd73a01a30d729e8fa6f829c4582650e258e36f9 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:42:59 +0200 Subject: hfsplus: fix HFSPLUS_SB calling convention HFSPLUS_SB doesn't return a pointer to the hfsplus-specific superblock information like all other FOO_SB macros, but dereference the pointer in a way that made it look like a direct struct derefence. This only works as long as the HFSPLUS_SB macro is used directly and prevents us from keepig a local hfsplus_sb_info pointer. Fix the calling convention and introduce a local sbi variable in all functions that use it constantly. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/bitmap.c b/fs/hfsplus/bitmap.c index 8a78176..ad57f59 100644 --- a/fs/hfsplus/bitmap.c +++ b/fs/hfsplus/bitmap.c @@ -17,6 +17,7 @@ int hfsplus_block_allocate(struct super_block *sb, u32 size, u32 offset, u32 *max) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); struct page *page; struct address_space *mapping; __be32 *pptr, *curr, *end; @@ -29,8 +30,8 @@ int hfsplus_block_allocate(struct super_block *sb, u32 size, u32 offset, u32 *ma return size; dprint(DBG_BITMAP, "block_allocate: %u,%u,%u\n", size, offset, len); - mutex_lock(&HFSPLUS_SB(sb).alloc_mutex); - mapping = HFSPLUS_SB(sb).alloc_file->i_mapping; + mutex_lock(&sbi->alloc_mutex); + mapping = sbi->alloc_file->i_mapping; page = read_mapping_page(mapping, offset / PAGE_CACHE_BITS, NULL); if (IS_ERR(page)) { start = size; @@ -150,16 +151,17 @@ done: set_page_dirty(page); kunmap(page); *max = offset + (curr - pptr) * 32 + i - start; - HFSPLUS_SB(sb).free_blocks -= *max; + sbi->free_blocks -= *max; sb->s_dirt = 1; dprint(DBG_BITMAP, "-> %u,%u\n", start, *max); out: - mutex_unlock(&HFSPLUS_SB(sb).alloc_mutex); + mutex_unlock(&sbi->alloc_mutex); return start; } int hfsplus_block_free(struct super_block *sb, u32 offset, u32 count) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); struct page *page; struct address_space *mapping; __be32 *pptr, *curr, *end; @@ -172,11 +174,11 @@ int hfsplus_block_free(struct super_block *sb, u32 offset, u32 count) dprint(DBG_BITMAP, "block_free: %u,%u\n", offset, count); /* are all of the bits in range? */ - if ((offset + count) > HFSPLUS_SB(sb).total_blocks) + if ((offset + count) > sbi->total_blocks) return -2; - mutex_lock(&HFSPLUS_SB(sb).alloc_mutex); - mapping = HFSPLUS_SB(sb).alloc_file->i_mapping; + mutex_lock(&sbi->alloc_mutex); + mapping = sbi->alloc_file->i_mapping; pnr = offset / PAGE_CACHE_BITS; page = read_mapping_page(mapping, pnr, NULL); pptr = kmap(page); @@ -224,9 +226,9 @@ done: out: set_page_dirty(page); kunmap(page); - HFSPLUS_SB(sb).free_blocks += len; + sbi->free_blocks += len; sb->s_dirt = 1; - mutex_unlock(&HFSPLUS_SB(sb).alloc_mutex); + mutex_unlock(&sbi->alloc_mutex); return 0; } diff --git a/fs/hfsplus/btree.c b/fs/hfsplus/btree.c index e49fcee..96e2128 100644 --- a/fs/hfsplus/btree.c +++ b/fs/hfsplus/btree.c @@ -61,12 +61,12 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) if (id == HFSPLUS_EXT_CNID) { tree->keycmp = hfsplus_ext_cmp_key; } else if (id == HFSPLUS_CAT_CNID) { - if ((HFSPLUS_SB(sb).flags & HFSPLUS_SB_HFSX) && + if ((HFSPLUS_SB(sb)->flags & HFSPLUS_SB_HFSX) && (head->key_type == HFSPLUS_KEY_BINARY)) tree->keycmp = hfsplus_cat_bin_cmp_key; else { tree->keycmp = hfsplus_cat_case_cmp_key; - HFSPLUS_SB(sb).flags |= HFSPLUS_SB_CASEFOLD; + HFSPLUS_SB(sb)->flags |= HFSPLUS_SB_CASEFOLD; } } else { printk(KERN_ERR "hfs: unknown B*Tree requested\n"); @@ -200,9 +200,9 @@ struct hfs_bnode *hfs_bmap_alloc(struct hfs_btree *tree) return ERR_PTR(res); HFSPLUS_I(inode).phys_size = inode->i_size = (loff_t)HFSPLUS_I(inode).alloc_blocks << - HFSPLUS_SB(tree->sb).alloc_blksz_shift; + HFSPLUS_SB(tree->sb)->alloc_blksz_shift; HFSPLUS_I(inode).fs_blocks = HFSPLUS_I(inode).alloc_blocks << - HFSPLUS_SB(tree->sb).fs_shift; + HFSPLUS_SB(tree->sb)->fs_shift; inode_set_bytes(inode, inode->i_size); count = inode->i_size >> tree->node_size_shift; tree->free_nodes = count - tree->node_count; diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index f6874ac..75ac1e4 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -86,6 +86,8 @@ static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct inode *inode) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(inode->i_sb); + if (S_ISDIR(inode->i_mode)) { struct hfsplus_cat_folder *folder; @@ -99,7 +101,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i folder->attribute_mod_date = folder->access_date = hfsp_now2mt(); hfsplus_set_perms(inode, &folder->permissions); - if (inode == HFSPLUS_SB(inode->i_sb).hidden_dir) + if (inode == sbi->hidden_dir) /* invisible and namelocked */ folder->user_info.frFlags = cpu_to_be16(0x5000); return sizeof(*folder); @@ -122,8 +124,8 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i file->user_info.fdType = cpu_to_be32(HFSP_SYMLINK_TYPE); file->user_info.fdCreator = cpu_to_be32(HFSP_SYMLINK_CREATOR); } else { - file->user_info.fdType = cpu_to_be32(HFSPLUS_SB(inode->i_sb).type); - file->user_info.fdCreator = cpu_to_be32(HFSPLUS_SB(inode->i_sb).creator); + file->user_info.fdType = cpu_to_be32(sbi->type); + file->user_info.fdCreator = cpu_to_be32(sbi->creator); } if ((file->permissions.rootflags | file->permissions.userflags) & HFSPLUS_FLG_IMMUTABLE) file->flags |= cpu_to_be16(HFSPLUS_FILE_LOCKED); @@ -131,7 +133,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i file->user_info.fdType = cpu_to_be32(HFSP_HARDLINK_TYPE); file->user_info.fdCreator = cpu_to_be32(HFSP_HFSPLUS_CREATOR); file->user_info.fdFlags = cpu_to_be16(0x100); - file->create_date = HFSPLUS_I(HFSPLUS_SB(inode->i_sb).hidden_dir).create_date; + file->create_date = HFSPLUS_I(sbi->hidden_dir).create_date; file->permissions.dev = cpu_to_be32(HFSPLUS_I(inode).dev); } return sizeof(*file); @@ -180,15 +182,14 @@ int hfsplus_find_cat(struct super_block *sb, u32 cnid, int hfsplus_create_cat(u32 cnid, struct inode *dir, struct qstr *str, struct inode *inode) { + struct super_block *sb = dir->i_sb; struct hfs_find_data fd; - struct super_block *sb; hfsplus_cat_entry entry; int entry_size; int err; dprint(DBG_CAT_MOD, "create_cat: %s,%u(%d)\n", str->name, cnid, inode->i_nlink); - sb = dir->i_sb; - hfs_find_init(HFSPLUS_SB(sb).cat_tree, &fd); + hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); hfsplus_cat_build_key(sb, fd.search_key, cnid, NULL); entry_size = hfsplus_fill_cat_thread(sb, &entry, S_ISDIR(inode->i_mode) ? @@ -234,7 +235,7 @@ err2: int hfsplus_delete_cat(u32 cnid, struct inode *dir, struct qstr *str) { - struct super_block *sb; + struct super_block *sb = dir->i_sb; struct hfs_find_data fd; struct hfsplus_fork_raw fork; struct list_head *pos; @@ -242,8 +243,7 @@ int hfsplus_delete_cat(u32 cnid, struct inode *dir, struct qstr *str) u16 type; dprint(DBG_CAT_MOD, "delete_cat: %s,%u\n", str ? str->name : NULL, cnid); - sb = dir->i_sb; - hfs_find_init(HFSPLUS_SB(sb).cat_tree, &fd); + hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); if (!str) { int len; @@ -312,7 +312,7 @@ int hfsplus_rename_cat(u32 cnid, struct inode *src_dir, struct qstr *src_name, struct inode *dst_dir, struct qstr *dst_name) { - struct super_block *sb; + struct super_block *sb = src_dir->i_sb; struct hfs_find_data src_fd, dst_fd; hfsplus_cat_entry entry; int entry_size, type; @@ -320,8 +320,7 @@ int hfsplus_rename_cat(u32 cnid, dprint(DBG_CAT_MOD, "rename_cat: %u - %lu,%s - %lu,%s\n", cnid, src_dir->i_ino, src_name->name, dst_dir->i_ino, dst_name->name); - sb = src_dir->i_sb; - hfs_find_init(HFSPLUS_SB(sb).cat_tree, &src_fd); + hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &src_fd); dst_fd = src_fd; /* find the old dir entry and read the data */ diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 764fd1b..584777d 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -39,7 +39,7 @@ static struct dentry *hfsplus_lookup(struct inode *dir, struct dentry *dentry, dentry->d_op = &hfsplus_dentry_operations; dentry->d_fsdata = NULL; - hfs_find_init(HFSPLUS_SB(sb).cat_tree, &fd); + hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); hfsplus_cat_build_key(sb, fd.search_key, dir->i_ino, &dentry->d_name); again: err = hfs_brec_read(&fd, &entry, sizeof(entry)); @@ -68,9 +68,9 @@ again: cnid = be32_to_cpu(entry.file.id); if (entry.file.user_info.fdType == cpu_to_be32(HFSP_HARDLINK_TYPE) && entry.file.user_info.fdCreator == cpu_to_be32(HFSP_HFSPLUS_CREATOR) && - (entry.file.create_date == HFSPLUS_I(HFSPLUS_SB(sb).hidden_dir).create_date || + (entry.file.create_date == HFSPLUS_I(HFSPLUS_SB(sb)->hidden_dir).create_date || entry.file.create_date == HFSPLUS_I(sb->s_root->d_inode).create_date) && - HFSPLUS_SB(sb).hidden_dir) { + HFSPLUS_SB(sb)->hidden_dir) { struct qstr str; char name[32]; @@ -86,7 +86,8 @@ again: linkid = be32_to_cpu(entry.file.permissions.dev); str.len = sprintf(name, "iNode%d", linkid); str.name = name; - hfsplus_cat_build_key(sb, fd.search_key, HFSPLUS_SB(sb).hidden_dir->i_ino, &str); + hfsplus_cat_build_key(sb, fd.search_key, + HFSPLUS_SB(sb)->hidden_dir->i_ino, &str); goto again; } } else if (!dentry->d_fsdata) @@ -124,7 +125,7 @@ static int hfsplus_readdir(struct file *filp, void *dirent, filldir_t filldir) if (filp->f_pos >= inode->i_size) return 0; - hfs_find_init(HFSPLUS_SB(sb).cat_tree, &fd); + hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); hfsplus_cat_build_key(sb, fd.search_key, inode->i_ino, NULL); err = hfs_brec_find(&fd); if (err) @@ -180,8 +181,9 @@ static int hfsplus_readdir(struct file *filp, void *dirent, filldir_t filldir) err = -EIO; goto out; } - if (HFSPLUS_SB(sb).hidden_dir && - HFSPLUS_SB(sb).hidden_dir->i_ino == be32_to_cpu(entry.folder.id)) + if (HFSPLUS_SB(sb)->hidden_dir && + HFSPLUS_SB(sb)->hidden_dir->i_ino == + be32_to_cpu(entry.folder.id)) goto next; if (filldir(dirent, strbuf, len, filp->f_pos, be32_to_cpu(entry.folder.id), DT_DIR)) @@ -260,7 +262,7 @@ static int hfsplus_create(struct inode *dir, struct dentry *dentry, int mode, static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, struct dentry *dst_dentry) { - struct super_block *sb = dst_dir->i_sb; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(dst_dir->i_sb); struct inode *inode = src_dentry->d_inode; struct inode *src_dir = src_dentry->d_parent->d_inode; struct qstr str; @@ -279,22 +281,22 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, str.len = sprintf(name, "iNode%d", id); res = hfsplus_rename_cat(inode->i_ino, src_dir, &src_dentry->d_name, - HFSPLUS_SB(sb).hidden_dir, &str); + sbi->hidden_dir, &str); if (!res) break; if (res != -EEXIST) return res; } HFSPLUS_I(inode).dev = id; - cnid = HFSPLUS_SB(sb).next_cnid++; + cnid = sbi->next_cnid++; src_dentry->d_fsdata = (void *)(unsigned long)cnid; res = hfsplus_create_cat(cnid, src_dir, &src_dentry->d_name, inode); if (res) /* panic? */ return res; - HFSPLUS_SB(sb).file_count++; + sbi->file_count++; } - cnid = HFSPLUS_SB(sb).next_cnid++; + cnid = sbi->next_cnid++; res = hfsplus_create_cat(cnid, dst_dir, &dst_dentry->d_name, inode); if (res) return res; @@ -304,15 +306,15 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, atomic_inc(&inode->i_count); inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); - HFSPLUS_SB(sb).file_count++; - sb->s_dirt = 1; + sbi->file_count++; + dst_dir->i_sb->s_dirt = 1; return 0; } static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) { - struct super_block *sb = dir->i_sb; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(dir->i_sb); struct inode *inode = dentry->d_inode; struct qstr str; char name[32]; @@ -329,7 +331,7 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) str.len = sprintf(name, "temp%lu", inode->i_ino); res = hfsplus_rename_cat(inode->i_ino, dir, &dentry->d_name, - HFSPLUS_SB(sb).hidden_dir, &str); + sbi->hidden_dir, &str); if (!res) inode->i_flags |= S_DEAD; return res; @@ -344,10 +346,10 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) clear_nlink(inode); if (!inode->i_nlink) { if (inode->i_ino != cnid) { - HFSPLUS_SB(sb).file_count--; + sbi->file_count--; if (!atomic_read(&HFSPLUS_I(inode).opencnt)) { res = hfsplus_delete_cat(inode->i_ino, - HFSPLUS_SB(sb).hidden_dir, + sbi->hidden_dir, NULL); if (!res) hfsplus_delete_inode(inode); @@ -356,7 +358,7 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) } else hfsplus_delete_inode(inode); } else - HFSPLUS_SB(sb).file_count--; + sbi->file_count--; inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); diff --git a/fs/hfsplus/extents.c b/fs/hfsplus/extents.c index 0022eec..1819698 100644 --- a/fs/hfsplus/extents.c +++ b/fs/hfsplus/extents.c @@ -108,7 +108,7 @@ void hfsplus_ext_write_extent(struct inode *inode) if (HFSPLUS_I(inode).flags & HFSPLUS_FLG_EXT_DIRTY) { struct hfs_find_data fd; - hfs_find_init(HFSPLUS_SB(inode->i_sb).ext_tree, &fd); + hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); __hfsplus_ext_write_extent(inode, &fd); hfs_find_exit(&fd); } @@ -162,7 +162,7 @@ static int hfsplus_ext_read_extent(struct inode *inode, u32 block) block < HFSPLUS_I(inode).cached_start + HFSPLUS_I(inode).cached_blocks) return 0; - hfs_find_init(HFSPLUS_SB(inode->i_sb).ext_tree, &fd); + hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); res = __hfsplus_ext_cache_extent(&fd, inode, block); hfs_find_exit(&fd); return res; @@ -172,16 +172,15 @@ static int hfsplus_ext_read_extent(struct inode *inode, u32 block) int hfsplus_get_block(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create) { - struct super_block *sb; + struct super_block *sb = inode->i_sb; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); int res = -EIO; u32 ablock, dblock, mask; int shift; - sb = inode->i_sb; - /* Convert inode block to disk allocation block */ - shift = HFSPLUS_SB(sb).alloc_blksz_shift - sb->s_blocksize_bits; - ablock = iblock >> HFSPLUS_SB(sb).fs_shift; + shift = sbi->alloc_blksz_shift - sb->s_blocksize_bits; + ablock = iblock >> sbi->fs_shift; if (iblock >= HFSPLUS_I(inode).fs_blocks) { if (iblock > HFSPLUS_I(inode).fs_blocks || !create) @@ -215,8 +214,8 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock, done: dprint(DBG_EXTENT, "get_block(%lu): %llu - %u\n", inode->i_ino, (long long)iblock, dblock); - mask = (1 << HFSPLUS_SB(sb).fs_shift) - 1; - map_bh(bh_result, sb, (dblock << HFSPLUS_SB(sb).fs_shift) + HFSPLUS_SB(sb).blockoffset + (iblock & mask)); + mask = (1 << sbi->fs_shift) - 1; + map_bh(bh_result, sb, (dblock << sbi->fs_shift) + sbi->blockoffset + (iblock & mask)); if (create) { set_buffer_new(bh_result); HFSPLUS_I(inode).phys_size += sb->s_blocksize; @@ -327,7 +326,7 @@ int hfsplus_free_fork(struct super_block *sb, u32 cnid, struct hfsplus_fork_raw if (total_blocks == blocks) return 0; - hfs_find_init(HFSPLUS_SB(sb).ext_tree, &fd); + hfs_find_init(HFSPLUS_SB(sb)->ext_tree, &fd); do { res = __hfsplus_ext_read_extent(&fd, ext_entry, cnid, total_blocks, type); @@ -348,13 +347,16 @@ int hfsplus_free_fork(struct super_block *sb, u32 cnid, struct hfsplus_fork_raw int hfsplus_file_extend(struct inode *inode) { struct super_block *sb = inode->i_sb; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); u32 start, len, goal; int res; - if (HFSPLUS_SB(sb).alloc_file->i_size * 8 < HFSPLUS_SB(sb).total_blocks - HFSPLUS_SB(sb).free_blocks + 8) { + if (sbi->alloc_file->i_size * 8 < + sbi->total_blocks - sbi->free_blocks + 8) { // extend alloc file - printk(KERN_ERR "hfs: extend alloc file! (%Lu,%u,%u)\n", HFSPLUS_SB(sb).alloc_file->i_size * 8, - HFSPLUS_SB(sb).total_blocks, HFSPLUS_SB(sb).free_blocks); + printk(KERN_ERR "hfs: extend alloc file! (%Lu,%u,%u)\n", + sbi->alloc_file->i_size * 8, + sbi->total_blocks, sbi->free_blocks); return -ENOSPC; } @@ -369,8 +371,8 @@ int hfsplus_file_extend(struct inode *inode) } len = HFSPLUS_I(inode).clump_blocks; - start = hfsplus_block_allocate(sb, HFSPLUS_SB(sb).total_blocks, goal, &len); - if (start >= HFSPLUS_SB(sb).total_blocks) { + start = hfsplus_block_allocate(sb, sbi->total_blocks, goal, &len); + if (start >= sbi->total_blocks) { start = hfsplus_block_allocate(sb, goal, 0, &len); if (start >= goal) { res = -ENOSPC; @@ -463,13 +465,14 @@ void hfsplus_file_truncate(struct inode *inode) } else if (inode->i_size == HFSPLUS_I(inode).phys_size) return; - blk_cnt = (inode->i_size + HFSPLUS_SB(sb).alloc_blksz - 1) >> HFSPLUS_SB(sb).alloc_blksz_shift; + blk_cnt = (inode->i_size + HFSPLUS_SB(sb)->alloc_blksz - 1) >> + HFSPLUS_SB(sb)->alloc_blksz_shift; alloc_cnt = HFSPLUS_I(inode).alloc_blocks; if (blk_cnt == alloc_cnt) goto out; mutex_lock(&HFSPLUS_I(inode).extents_lock); - hfs_find_init(HFSPLUS_SB(sb).ext_tree, &fd); + hfs_find_init(HFSPLUS_SB(sb)->ext_tree, &fd); while (1) { if (alloc_cnt == HFSPLUS_I(inode).first_blocks) { hfsplus_free_extents(sb, HFSPLUS_I(inode).first_extents, diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index df0a631..55f42b4 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -375,17 +375,16 @@ int hfsplus_read_wrapper(struct super_block *); int hfs_part_find(struct super_block *, sector_t *, sector_t *); /* access macros */ -/* static inline struct hfsplus_sb_info *HFSPLUS_SB(struct super_block *sb) { return sb->s_fs_info; } +/* static inline struct hfsplus_inode_info *HFSPLUS_I(struct inode *inode) { return list_entry(inode, struct hfsplus_inode_info, vfs_inode); } */ -#define HFSPLUS_SB(super) (*(struct hfsplus_sb_info *)(super)->s_fs_info) #define HFSPLUS_I(inode) (*list_entry(inode, struct hfsplus_inode_info, vfs_inode)) #if 1 diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index c5a979d..050eee6 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -62,13 +62,13 @@ static int hfsplus_releasepage(struct page *page, gfp_t mask) switch (inode->i_ino) { case HFSPLUS_EXT_CNID: - tree = HFSPLUS_SB(sb).ext_tree; + tree = HFSPLUS_SB(sb)->ext_tree; break; case HFSPLUS_CAT_CNID: - tree = HFSPLUS_SB(sb).cat_tree; + tree = HFSPLUS_SB(sb)->cat_tree; break; case HFSPLUS_ATTR_CNID: - tree = HFSPLUS_SB(sb).attr_tree; + tree = HFSPLUS_SB(sb)->attr_tree; break; default: BUG(); @@ -190,7 +190,7 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, struct dentry *dent mutex_init(&HFSPLUS_I(inode).extents_lock); HFSPLUS_I(inode).flags = HFSPLUS_FLG_RSRC; - hfs_find_init(HFSPLUS_SB(sb).cat_tree, &fd); + hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); err = hfsplus_find_cat(sb, dir->i_ino, &fd); if (!err) err = hfsplus_cat_read_inode(inode, &fd); @@ -202,7 +202,7 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, struct dentry *dent HFSPLUS_I(inode).rsrc_inode = dir; HFSPLUS_I(dir).rsrc_inode = inode; igrab(dir); - hlist_add_head(&inode->i_hash, &HFSPLUS_SB(sb).rsrc_inodes); + hlist_add_head(&inode->i_hash, &HFSPLUS_SB(sb)->rsrc_inodes); mark_inode_dirty(inode); out: d_add(dentry, inode); @@ -211,26 +211,24 @@ out: static void hfsplus_get_perms(struct inode *inode, struct hfsplus_perm *perms, int dir) { - struct super_block *sb = inode->i_sb; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(inode->i_sb); u16 mode; mode = be16_to_cpu(perms->mode); inode->i_uid = be32_to_cpu(perms->owner); if (!inode->i_uid && !mode) - inode->i_uid = HFSPLUS_SB(sb).uid; + inode->i_uid = sbi->uid; inode->i_gid = be32_to_cpu(perms->group); if (!inode->i_gid && !mode) - inode->i_gid = HFSPLUS_SB(sb).gid; + inode->i_gid = sbi->gid; if (dir) { - mode = mode ? (mode & S_IALLUGO) : - (S_IRWXUGO & ~(HFSPLUS_SB(sb).umask)); + mode = mode ? (mode & S_IALLUGO) : (S_IRWXUGO & ~(sbi->umask)); mode |= S_IFDIR; } else if (!mode) - mode = S_IFREG | ((S_IRUGO|S_IWUGO) & - ~(HFSPLUS_SB(sb).umask)); + mode = S_IFREG | ((S_IRUGO|S_IWUGO) & ~(sbi->umask)); inode->i_mode = mode; HFSPLUS_I(inode).rootflags = perms->rootflags; @@ -282,7 +280,8 @@ static int hfsplus_file_release(struct inode *inode, struct file *file) mutex_lock(&inode->i_mutex); hfsplus_file_truncate(inode); if (inode->i_flags & S_DEAD) { - hfsplus_delete_cat(inode->i_ino, HFSPLUS_SB(sb).hidden_dir, NULL); + hfsplus_delete_cat(inode->i_ino, + HFSPLUS_SB(sb)->hidden_dir, NULL); hfsplus_delete_inode(inode); } mutex_unlock(&inode->i_mutex); @@ -361,11 +360,13 @@ static const struct file_operations hfsplus_file_operations = { struct inode *hfsplus_new_inode(struct super_block *sb, int mode) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); struct inode *inode = new_inode(sb); + if (!inode) return NULL; - inode->i_ino = HFSPLUS_SB(sb).next_cnid++; + inode->i_ino = sbi->next_cnid++; inode->i_mode = mode; inode->i_uid = current_fsuid(); inode->i_gid = current_fsgid(); @@ -386,22 +387,22 @@ struct inode *hfsplus_new_inode(struct super_block *sb, int mode) HFSPLUS_I(inode).rsrc_inode = NULL; if (S_ISDIR(inode->i_mode)) { inode->i_size = 2; - HFSPLUS_SB(sb).folder_count++; + sbi->folder_count++; inode->i_op = &hfsplus_dir_inode_operations; inode->i_fop = &hfsplus_dir_operations; } else if (S_ISREG(inode->i_mode)) { - HFSPLUS_SB(sb).file_count++; + sbi->file_count++; inode->i_op = &hfsplus_file_inode_operations; inode->i_fop = &hfsplus_file_operations; inode->i_mapping->a_ops = &hfsplus_aops; - HFSPLUS_I(inode).clump_blocks = HFSPLUS_SB(sb).data_clump_blocks; + HFSPLUS_I(inode).clump_blocks = sbi->data_clump_blocks; } else if (S_ISLNK(inode->i_mode)) { - HFSPLUS_SB(sb).file_count++; + sbi->file_count++; inode->i_op = &page_symlink_inode_operations; inode->i_mapping->a_ops = &hfsplus_aops; HFSPLUS_I(inode).clump_blocks = 1; } else - HFSPLUS_SB(sb).file_count++; + sbi->file_count++; insert_inode_hash(inode); mark_inode_dirty(inode); sb->s_dirt = 1; @@ -414,11 +415,11 @@ void hfsplus_delete_inode(struct inode *inode) struct super_block *sb = inode->i_sb; if (S_ISDIR(inode->i_mode)) { - HFSPLUS_SB(sb).folder_count--; + HFSPLUS_SB(sb)->folder_count--; sb->s_dirt = 1; return; } - HFSPLUS_SB(sb).file_count--; + HFSPLUS_SB(sb)->file_count--; if (S_ISREG(inode->i_mode)) { if (!inode->i_nlink) { inode->i_size = 0; @@ -434,6 +435,7 @@ void hfsplus_delete_inode(struct inode *inode) void hfsplus_inode_read_fork(struct inode *inode, struct hfsplus_fork_raw *fork) { struct super_block *sb = inode->i_sb; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); u32 count; int i; @@ -450,10 +452,13 @@ void hfsplus_inode_read_fork(struct inode *inode, struct hfsplus_fork_raw *fork) inode->i_size = HFSPLUS_I(inode).phys_size = be64_to_cpu(fork->total_size); HFSPLUS_I(inode).fs_blocks = (inode->i_size + sb->s_blocksize - 1) >> sb->s_blocksize_bits; inode_set_bytes(inode, HFSPLUS_I(inode).fs_blocks << sb->s_blocksize_bits); - HFSPLUS_I(inode).clump_blocks = be32_to_cpu(fork->clump_size) >> HFSPLUS_SB(sb).alloc_blksz_shift; - if (!HFSPLUS_I(inode).clump_blocks) - HFSPLUS_I(inode).clump_blocks = HFSPLUS_IS_RSRC(inode) ? HFSPLUS_SB(sb).rsrc_clump_blocks : - HFSPLUS_SB(sb).data_clump_blocks; + HFSPLUS_I(inode).clump_blocks = + be32_to_cpu(fork->clump_size) >> sbi->alloc_blksz_shift; + if (!HFSPLUS_I(inode).clump_blocks) { + HFSPLUS_I(inode).clump_blocks = HFSPLUS_IS_RSRC(inode) ? + sbi->rsrc_clump_blocks : + sbi->data_clump_blocks; + } } void hfsplus_inode_write_fork(struct inode *inode, struct hfsplus_fork_raw *fork) @@ -538,7 +543,7 @@ int hfsplus_cat_write_inode(struct inode *inode) if (!main_inode->i_nlink) return 0; - if (hfs_find_init(HFSPLUS_SB(main_inode->i_sb).cat_tree, &fd)) + if (hfs_find_init(HFSPLUS_SB(main_inode->i_sb)->cat_tree, &fd)) /* panic? */ return -EIO; diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index 906bd3dd..66dd64b 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -126,7 +126,7 @@ int hfsplus_setxattr(struct dentry *dentry, const char *name, if (!S_ISREG(inode->i_mode) || HFSPLUS_IS_RSRC(inode)) return -EOPNOTSUPP; - res = hfs_find_init(HFSPLUS_SB(inode->i_sb).cat_tree, &fd); + res = hfs_find_init(HFSPLUS_SB(inode->i_sb)->cat_tree, &fd); if (res) return res; res = hfsplus_find_cat(inode->i_sb, inode->i_ino, &fd); @@ -169,7 +169,7 @@ ssize_t hfsplus_getxattr(struct dentry *dentry, const char *name, return -EOPNOTSUPP; if (size) { - res = hfs_find_init(HFSPLUS_SB(inode->i_sb).cat_tree, &fd); + res = hfs_find_init(HFSPLUS_SB(inode->i_sb)->cat_tree, &fd); if (res) return res; res = hfsplus_find_cat(inode->i_sb, inode->i_ino, &fd); diff --git a/fs/hfsplus/options.c b/fs/hfsplus/options.c index 572628b..f2598ae 100644 --- a/fs/hfsplus/options.c +++ b/fs/hfsplus/options.c @@ -171,7 +171,7 @@ done: int hfsplus_show_options(struct seq_file *seq, struct vfsmount *mnt) { - struct hfsplus_sb_info *sbi = &HFSPLUS_SB(mnt->mnt_sb); + struct hfsplus_sb_info *sbi = HFSPLUS_SB(mnt->mnt_sb); if (sbi->creator != HFSPLUS_DEF_CR_TYPE) seq_printf(seq, ",creator=%.4s", (char *)&sbi->creator); diff --git a/fs/hfsplus/part_tbl.c b/fs/hfsplus/part_tbl.c index 1528a6f..208b16c 100644 --- a/fs/hfsplus/part_tbl.c +++ b/fs/hfsplus/part_tbl.c @@ -74,6 +74,7 @@ struct old_pmap { int hfs_part_find(struct super_block *sb, sector_t *part_start, sector_t *part_size) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); struct buffer_head *bh; __be16 *data; int i, size, res; @@ -95,7 +96,7 @@ int hfs_part_find(struct super_block *sb, for (i = 0; i < size; p++, i++) { if (p->pdStart && p->pdSize && p->pdFSID == cpu_to_be32(0x54465331)/*"TFS1"*/ && - (HFSPLUS_SB(sb).part < 0 || HFSPLUS_SB(sb).part == i)) { + (sbi->part < 0 || sbi->part == i)) { *part_start += be32_to_cpu(p->pdStart); *part_size = be32_to_cpu(p->pdSize); res = 0; @@ -111,7 +112,7 @@ int hfs_part_find(struct super_block *sb, size = be32_to_cpu(pm->pmMapBlkCnt); for (i = 0; i < size;) { if (!memcmp(pm->pmPartType,"Apple_HFS", 9) && - (HFSPLUS_SB(sb).part < 0 || HFSPLUS_SB(sb).part == i)) { + (sbi->part < 0 || sbi->part == i)) { *part_start += be32_to_cpu(pm->pmPyPartStart); *part_size = be32_to_cpu(pm->pmPartBlkCnt); res = 0; diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index a1d3fd9..1bf00b9 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -41,7 +41,7 @@ struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID) { read_inode: - hfs_find_init(HFSPLUS_SB(inode->i_sb).cat_tree, &fd); + hfs_find_init(HFSPLUS_SB(inode->i_sb)->cat_tree, &fd); err = hfsplus_find_cat(inode->i_sb, inode->i_ino, &fd); if (!err) err = hfsplus_cat_read_inode(inode, &fd); @@ -50,7 +50,7 @@ struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) goto bad_inode; goto done; } - vhdr = HFSPLUS_SB(inode->i_sb).s_vhdr; + vhdr = HFSPLUS_SB(inode->i_sb)->s_vhdr; switch(inode->i_ino) { case HFSPLUS_ROOT_CNID: goto read_inode; @@ -89,6 +89,7 @@ bad_inode: static int hfsplus_write_inode(struct inode *inode, struct writeback_control *wbc) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(inode->i_sb); struct hfsplus_vh *vhdr; int ret = 0; @@ -97,48 +98,48 @@ static int hfsplus_write_inode(struct inode *inode, if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID) { return hfsplus_cat_write_inode(inode); } - vhdr = HFSPLUS_SB(inode->i_sb).s_vhdr; + vhdr = sbi->s_vhdr; switch (inode->i_ino) { case HFSPLUS_ROOT_CNID: ret = hfsplus_cat_write_inode(inode); break; case HFSPLUS_EXT_CNID: if (vhdr->ext_file.total_size != cpu_to_be64(inode->i_size)) { - HFSPLUS_SB(inode->i_sb).flags |= HFSPLUS_SB_WRITEBACKUP; + sbi->flags |= HFSPLUS_SB_WRITEBACKUP; inode->i_sb->s_dirt = 1; } hfsplus_inode_write_fork(inode, &vhdr->ext_file); - hfs_btree_write(HFSPLUS_SB(inode->i_sb).ext_tree); + hfs_btree_write(sbi->ext_tree); break; case HFSPLUS_CAT_CNID: if (vhdr->cat_file.total_size != cpu_to_be64(inode->i_size)) { - HFSPLUS_SB(inode->i_sb).flags |= HFSPLUS_SB_WRITEBACKUP; + sbi->flags |= HFSPLUS_SB_WRITEBACKUP; inode->i_sb->s_dirt = 1; } hfsplus_inode_write_fork(inode, &vhdr->cat_file); - hfs_btree_write(HFSPLUS_SB(inode->i_sb).cat_tree); + hfs_btree_write(sbi->cat_tree); break; case HFSPLUS_ALLOC_CNID: if (vhdr->alloc_file.total_size != cpu_to_be64(inode->i_size)) { - HFSPLUS_SB(inode->i_sb).flags |= HFSPLUS_SB_WRITEBACKUP; + sbi->flags |= HFSPLUS_SB_WRITEBACKUP; inode->i_sb->s_dirt = 1; } hfsplus_inode_write_fork(inode, &vhdr->alloc_file); break; case HFSPLUS_START_CNID: if (vhdr->start_file.total_size != cpu_to_be64(inode->i_size)) { - HFSPLUS_SB(inode->i_sb).flags |= HFSPLUS_SB_WRITEBACKUP; + sbi->flags |= HFSPLUS_SB_WRITEBACKUP; inode->i_sb->s_dirt = 1; } hfsplus_inode_write_fork(inode, &vhdr->start_file); break; case HFSPLUS_ATTR_CNID: if (vhdr->attr_file.total_size != cpu_to_be64(inode->i_size)) { - HFSPLUS_SB(inode->i_sb).flags |= HFSPLUS_SB_WRITEBACKUP; + sbi->flags |= HFSPLUS_SB_WRITEBACKUP; inode->i_sb->s_dirt = 1; } hfsplus_inode_write_fork(inode, &vhdr->attr_file); - hfs_btree_write(HFSPLUS_SB(inode->i_sb).attr_tree); + hfs_btree_write(sbi->attr_tree); break; } return ret; @@ -157,44 +158,46 @@ static void hfsplus_evict_inode(struct inode *inode) int hfsplus_sync_fs(struct super_block *sb, int wait) { - struct hfsplus_vh *vhdr = HFSPLUS_SB(sb).s_vhdr; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); + struct hfsplus_vh *vhdr = sbi->s_vhdr; dprint(DBG_SUPER, "hfsplus_write_super\n"); - mutex_lock(&HFSPLUS_SB(sb).alloc_mutex); + mutex_lock(&sbi->alloc_mutex); sb->s_dirt = 0; - vhdr->free_blocks = cpu_to_be32(HFSPLUS_SB(sb).free_blocks); - vhdr->next_alloc = cpu_to_be32(HFSPLUS_SB(sb).next_alloc); - vhdr->next_cnid = cpu_to_be32(HFSPLUS_SB(sb).next_cnid); - vhdr->folder_count = cpu_to_be32(HFSPLUS_SB(sb).folder_count); - vhdr->file_count = cpu_to_be32(HFSPLUS_SB(sb).file_count); + vhdr->free_blocks = cpu_to_be32(sbi->free_blocks); + vhdr->next_alloc = cpu_to_be32(sbi->next_alloc); + vhdr->next_cnid = cpu_to_be32(sbi->next_cnid); + vhdr->folder_count = cpu_to_be32(sbi->folder_count); + vhdr->file_count = cpu_to_be32(sbi->file_count); - mark_buffer_dirty(HFSPLUS_SB(sb).s_vhbh); - if (HFSPLUS_SB(sb).flags & HFSPLUS_SB_WRITEBACKUP) { - if (HFSPLUS_SB(sb).sect_count) { + mark_buffer_dirty(sbi->s_vhbh); + if (sbi->flags & HFSPLUS_SB_WRITEBACKUP) { + if (sbi->sect_count) { struct buffer_head *bh; u32 block, offset; - block = HFSPLUS_SB(sb).blockoffset; - block += (HFSPLUS_SB(sb).sect_count - 2) >> (sb->s_blocksize_bits - 9); - offset = ((HFSPLUS_SB(sb).sect_count - 2) << 9) & (sb->s_blocksize - 1); - printk(KERN_DEBUG "hfs: backup: %u,%u,%u,%u\n", HFSPLUS_SB(sb).blockoffset, - HFSPLUS_SB(sb).sect_count, block, offset); + block = sbi->blockoffset; + block += (sbi->sect_count - 2) >> (sb->s_blocksize_bits - 9); + offset = ((sbi->sect_count - 2) << 9) & (sb->s_blocksize - 1); + printk(KERN_DEBUG "hfs: backup: %u,%u,%u,%u\n", + sbi->blockoffset, sbi->sect_count, + block, offset); bh = sb_bread(sb, block); if (bh) { vhdr = (struct hfsplus_vh *)(bh->b_data + offset); if (be16_to_cpu(vhdr->signature) == HFSPLUS_VOLHEAD_SIG) { - memcpy(vhdr, HFSPLUS_SB(sb).s_vhdr, sizeof(*vhdr)); + memcpy(vhdr, sbi->s_vhdr, sizeof(*vhdr)); mark_buffer_dirty(bh); brelse(bh); } else printk(KERN_WARNING "hfs: backup not found!\n"); } } - HFSPLUS_SB(sb).flags &= ~HFSPLUS_SB_WRITEBACKUP; + sbi->flags &= ~HFSPLUS_SB_WRITEBACKUP; } - mutex_unlock(&HFSPLUS_SB(sb).alloc_mutex); + mutex_unlock(&sbi->alloc_mutex); return 0; } @@ -208,28 +211,31 @@ static void hfsplus_write_super(struct super_block *sb) static void hfsplus_put_super(struct super_block *sb) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); + dprint(DBG_SUPER, "hfsplus_put_super\n"); + if (!sb->s_fs_info) return; if (sb->s_dirt) hfsplus_write_super(sb); - if (!(sb->s_flags & MS_RDONLY) && HFSPLUS_SB(sb).s_vhdr) { - struct hfsplus_vh *vhdr = HFSPLUS_SB(sb).s_vhdr; + if (!(sb->s_flags & MS_RDONLY) && sbi->s_vhdr) { + struct hfsplus_vh *vhdr = sbi->s_vhdr; vhdr->modify_date = hfsp_now2mt(); vhdr->attributes |= cpu_to_be32(HFSPLUS_VOL_UNMNT); vhdr->attributes &= cpu_to_be32(~HFSPLUS_VOL_INCNSTNT); - mark_buffer_dirty(HFSPLUS_SB(sb).s_vhbh); - sync_dirty_buffer(HFSPLUS_SB(sb).s_vhbh); + mark_buffer_dirty(sbi->s_vhbh); + sync_dirty_buffer(sbi->s_vhbh); } - hfs_btree_close(HFSPLUS_SB(sb).cat_tree); - hfs_btree_close(HFSPLUS_SB(sb).ext_tree); - iput(HFSPLUS_SB(sb).alloc_file); - iput(HFSPLUS_SB(sb).hidden_dir); - brelse(HFSPLUS_SB(sb).s_vhbh); - unload_nls(HFSPLUS_SB(sb).nls); + hfs_btree_close(sbi->cat_tree); + hfs_btree_close(sbi->ext_tree); + iput(sbi->alloc_file); + iput(sbi->hidden_dir); + brelse(sbi->s_vhbh); + unload_nls(sbi->nls); kfree(sb->s_fs_info); sb->s_fs_info = NULL; } @@ -237,15 +243,16 @@ static void hfsplus_put_super(struct super_block *sb) static int hfsplus_statfs(struct dentry *dentry, struct kstatfs *buf) { struct super_block *sb = dentry->d_sb; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); u64 id = huge_encode_dev(sb->s_bdev->bd_dev); buf->f_type = HFSPLUS_SUPER_MAGIC; buf->f_bsize = sb->s_blocksize; - buf->f_blocks = HFSPLUS_SB(sb).total_blocks << HFSPLUS_SB(sb).fs_shift; - buf->f_bfree = HFSPLUS_SB(sb).free_blocks << HFSPLUS_SB(sb).fs_shift; + buf->f_blocks = sbi->total_blocks << sbi->fs_shift; + buf->f_bfree = sbi->free_blocks << sbi->fs_shift; buf->f_bavail = buf->f_bfree; buf->f_files = 0xFFFFFFFF; - buf->f_ffree = 0xFFFFFFFF - HFSPLUS_SB(sb).next_cnid; + buf->f_ffree = 0xFFFFFFFF - sbi->next_cnid; buf->f_fsid.val[0] = (u32)id; buf->f_fsid.val[1] = (u32)(id >> 32); buf->f_namelen = HFSPLUS_MAX_STRLEN; @@ -258,11 +265,11 @@ static int hfsplus_remount(struct super_block *sb, int *flags, char *data) if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) return 0; if (!(*flags & MS_RDONLY)) { - struct hfsplus_vh *vhdr = HFSPLUS_SB(sb).s_vhdr; + struct hfsplus_vh *vhdr = HFSPLUS_SB(sb)->s_vhdr; struct hfsplus_sb_info sbi; memset(&sbi, 0, sizeof(struct hfsplus_sb_info)); - sbi.nls = HFSPLUS_SB(sb).nls; + sbi.nls = HFSPLUS_SB(sb)->nls; if (!hfsplus_parse_options(data, &sbi)) return -EINVAL; @@ -340,7 +347,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) err = -EINVAL; goto cleanup; } - vhdr = HFSPLUS_SB(sb).s_vhdr; + vhdr = sbi->s_vhdr; /* Copy parts of the volume header into the superblock */ sb->s_magic = HFSPLUS_VOLHEAD_SIG; @@ -349,18 +356,20 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) printk(KERN_ERR "hfs: wrong filesystem version\n"); goto cleanup; } - HFSPLUS_SB(sb).total_blocks = be32_to_cpu(vhdr->total_blocks); - HFSPLUS_SB(sb).free_blocks = be32_to_cpu(vhdr->free_blocks); - HFSPLUS_SB(sb).next_alloc = be32_to_cpu(vhdr->next_alloc); - HFSPLUS_SB(sb).next_cnid = be32_to_cpu(vhdr->next_cnid); - HFSPLUS_SB(sb).file_count = be32_to_cpu(vhdr->file_count); - HFSPLUS_SB(sb).folder_count = be32_to_cpu(vhdr->folder_count); - HFSPLUS_SB(sb).data_clump_blocks = be32_to_cpu(vhdr->data_clump_sz) >> HFSPLUS_SB(sb).alloc_blksz_shift; - if (!HFSPLUS_SB(sb).data_clump_blocks) - HFSPLUS_SB(sb).data_clump_blocks = 1; - HFSPLUS_SB(sb).rsrc_clump_blocks = be32_to_cpu(vhdr->rsrc_clump_sz) >> HFSPLUS_SB(sb).alloc_blksz_shift; - if (!HFSPLUS_SB(sb).rsrc_clump_blocks) - HFSPLUS_SB(sb).rsrc_clump_blocks = 1; + sbi->total_blocks = be32_to_cpu(vhdr->total_blocks); + sbi->free_blocks = be32_to_cpu(vhdr->free_blocks); + sbi->next_alloc = be32_to_cpu(vhdr->next_alloc); + sbi->next_cnid = be32_to_cpu(vhdr->next_cnid); + sbi->file_count = be32_to_cpu(vhdr->file_count); + sbi->folder_count = be32_to_cpu(vhdr->folder_count); + sbi->data_clump_blocks = + be32_to_cpu(vhdr->data_clump_sz) >> sbi->alloc_blksz_shift; + if (!sbi->data_clump_blocks) + sbi->data_clump_blocks = 1; + sbi->rsrc_clump_blocks = + be32_to_cpu(vhdr->rsrc_clump_sz) >> sbi->alloc_blksz_shift; + if (!sbi->rsrc_clump_blocks) + sbi->rsrc_clump_blocks = 1; /* Set up operations so we can load metadata */ sb->s_op = &hfsplus_sops; @@ -383,13 +392,13 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) sbi->flags &= ~HFSPLUS_SB_FORCE; /* Load metadata objects (B*Trees) */ - HFSPLUS_SB(sb).ext_tree = hfs_btree_open(sb, HFSPLUS_EXT_CNID); - if (!HFSPLUS_SB(sb).ext_tree) { + sbi->ext_tree = hfs_btree_open(sb, HFSPLUS_EXT_CNID); + if (!sbi->ext_tree) { printk(KERN_ERR "hfs: failed to load extents file\n"); goto cleanup; } - HFSPLUS_SB(sb).cat_tree = hfs_btree_open(sb, HFSPLUS_CAT_CNID); - if (!HFSPLUS_SB(sb).cat_tree) { + sbi->cat_tree = hfs_btree_open(sb, HFSPLUS_CAT_CNID); + if (!sbi->cat_tree) { printk(KERN_ERR "hfs: failed to load catalog file\n"); goto cleanup; } @@ -400,7 +409,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) err = PTR_ERR(inode); goto cleanup; } - HFSPLUS_SB(sb).alloc_file = inode; + sbi->alloc_file = inode; /* Load the root directory */ root = hfsplus_iget(sb, HFSPLUS_ROOT_CNID); @@ -419,7 +428,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) str.len = sizeof(HFSP_HIDDENDIR_NAME) - 1; str.name = HFSP_HIDDENDIR_NAME; - hfs_find_init(HFSPLUS_SB(sb).cat_tree, &fd); + hfs_find_init(sbi->cat_tree, &fd); hfsplus_cat_build_key(sb, fd.search_key, HFSPLUS_ROOT_CNID, &str); if (!hfs_brec_read(&fd, &entry, sizeof(entry))) { hfs_find_exit(&fd); @@ -430,7 +439,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) err = PTR_ERR(inode); goto cleanup; } - HFSPLUS_SB(sb).hidden_dir = inode; + sbi->hidden_dir = inode; } else hfs_find_exit(&fd); @@ -445,15 +454,15 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) be32_add_cpu(&vhdr->write_count, 1); vhdr->attributes &= cpu_to_be32(~HFSPLUS_VOL_UNMNT); vhdr->attributes |= cpu_to_be32(HFSPLUS_VOL_INCNSTNT); - mark_buffer_dirty(HFSPLUS_SB(sb).s_vhbh); - sync_dirty_buffer(HFSPLUS_SB(sb).s_vhbh); + mark_buffer_dirty(sbi->s_vhbh); + sync_dirty_buffer(sbi->s_vhbh); - if (!HFSPLUS_SB(sb).hidden_dir) { + if (!sbi->hidden_dir) { printk(KERN_DEBUG "hfs: create hidden dir...\n"); - HFSPLUS_SB(sb).hidden_dir = hfsplus_new_inode(sb, S_IFDIR); - hfsplus_create_cat(HFSPLUS_SB(sb).hidden_dir->i_ino, sb->s_root->d_inode, - &str, HFSPLUS_SB(sb).hidden_dir); - mark_inode_dirty(HFSPLUS_SB(sb).hidden_dir); + sbi->hidden_dir = hfsplus_new_inode(sb, S_IFDIR); + hfsplus_create_cat(sbi->hidden_dir->i_ino, sb->s_root->d_inode, + &str, sbi->hidden_dir); + mark_inode_dirty(sbi->hidden_dir); } out: unload_nls(sbi->nls); diff --git a/fs/hfsplus/unicode.c b/fs/hfsplus/unicode.c index 628ccf6..a15fcff 100644 --- a/fs/hfsplus/unicode.c +++ b/fs/hfsplus/unicode.c @@ -121,7 +121,7 @@ static u16 *hfsplus_compose_lookup(u16 *p, u16 cc) int hfsplus_uni2asc(struct super_block *sb, const struct hfsplus_unistr *ustr, char *astr, int *len_p) { const hfsplus_unichr *ip; - struct nls_table *nls = HFSPLUS_SB(sb).nls; + struct nls_table *nls = HFSPLUS_SB(sb)->nls; u8 *op; u16 cc, c0, c1; u16 *ce1, *ce2; @@ -132,7 +132,7 @@ int hfsplus_uni2asc(struct super_block *sb, const struct hfsplus_unistr *ustr, c ustrlen = be16_to_cpu(ustr->length); len = *len_p; ce1 = NULL; - compose = !(HFSPLUS_SB(sb).flags & HFSPLUS_SB_NODECOMPOSE); + compose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); while (ustrlen > 0) { c0 = be16_to_cpu(*ip++); @@ -246,7 +246,7 @@ out: static inline int asc2unichar(struct super_block *sb, const char *astr, int len, wchar_t *uc) { - int size = HFSPLUS_SB(sb).nls->char2uni(astr, len, uc); + int size = HFSPLUS_SB(sb)->nls->char2uni(astr, len, uc); if (size <= 0) { *uc = '?'; size = 1; @@ -293,7 +293,7 @@ int hfsplus_asc2uni(struct super_block *sb, struct hfsplus_unistr *ustr, u16 *dstr, outlen = 0; wchar_t c; - decompose = !(HFSPLUS_SB(sb).flags & HFSPLUS_SB_NODECOMPOSE); + decompose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); while (outlen < HFSPLUS_MAX_STRLEN && len > 0) { size = asc2unichar(sb, astr, len, &c); @@ -330,8 +330,8 @@ int hfsplus_hash_dentry(struct dentry *dentry, struct qstr *str) wchar_t c; u16 c2; - casefold = (HFSPLUS_SB(sb).flags & HFSPLUS_SB_CASEFOLD); - decompose = !(HFSPLUS_SB(sb).flags & HFSPLUS_SB_NODECOMPOSE); + casefold = (HFSPLUS_SB(sb)->flags & HFSPLUS_SB_CASEFOLD); + decompose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); hash = init_name_hash(); astr = str->name; len = str->len; @@ -373,8 +373,8 @@ int hfsplus_compare_dentry(struct dentry *dentry, struct qstr *s1, struct qstr * u16 c1, c2; wchar_t c; - casefold = (HFSPLUS_SB(sb).flags & HFSPLUS_SB_CASEFOLD); - decompose = !(HFSPLUS_SB(sb).flags & HFSPLUS_SB_NODECOMPOSE); + casefold = (HFSPLUS_SB(sb)->flags & HFSPLUS_SB_CASEFOLD); + decompose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); astr1 = s1->name; len1 = s1->len; astr2 = s2->name; diff --git a/fs/hfsplus/wrapper.c b/fs/hfsplus/wrapper.c index bed78ac8..c89e198 100644 --- a/fs/hfsplus/wrapper.c +++ b/fs/hfsplus/wrapper.c @@ -65,8 +65,8 @@ static int hfsplus_get_last_session(struct super_block *sb, *start = 0; *size = sb->s_bdev->bd_inode->i_size >> 9; - if (HFSPLUS_SB(sb).session >= 0) { - te.cdte_track = HFSPLUS_SB(sb).session; + if (HFSPLUS_SB(sb)->session >= 0) { + te.cdte_track = HFSPLUS_SB(sb)->session; te.cdte_format = CDROM_LBA; res = ioctl_by_bdev(sb->s_bdev, CDROMREADTOCENTRY, (unsigned long)&te); if (!res && (te.cdte_ctrl & CDROM_DATA_TRACK) == 4) { @@ -87,6 +87,7 @@ static int hfsplus_get_last_session(struct super_block *sb, /* Takes in super block, returns true if good data read */ int hfsplus_read_wrapper(struct super_block *sb) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); struct buffer_head *bh; struct hfsplus_vh *vhdr; struct hfsplus_wd wd; @@ -122,7 +123,7 @@ int hfsplus_read_wrapper(struct super_block *sb) if (vhdr->signature == cpu_to_be16(HFSPLUS_VOLHEAD_SIG)) break; if (vhdr->signature == cpu_to_be16(HFSPLUS_VOLHEAD_SIGX)) { - HFSPLUS_SB(sb).flags |= HFSPLUS_SB_HFSX; + sbi->flags |= HFSPLUS_SB_HFSX; break; } brelse(bh); @@ -143,11 +144,11 @@ int hfsplus_read_wrapper(struct super_block *sb) if (blocksize < HFSPLUS_SECTOR_SIZE || ((blocksize - 1) & blocksize)) return -EINVAL; - HFSPLUS_SB(sb).alloc_blksz = blocksize; - HFSPLUS_SB(sb).alloc_blksz_shift = 0; + sbi->alloc_blksz = blocksize; + sbi->alloc_blksz_shift = 0; while ((blocksize >>= 1) != 0) - HFSPLUS_SB(sb).alloc_blksz_shift++; - blocksize = min(HFSPLUS_SB(sb).alloc_blksz, (u32)PAGE_SIZE); + sbi->alloc_blksz_shift++; + blocksize = min(sbi->alloc_blksz, (u32)PAGE_SIZE); /* align block size to block offset */ while (part_start & ((blocksize >> HFSPLUS_SECTOR_SHIFT) - 1)) @@ -158,23 +159,22 @@ int hfsplus_read_wrapper(struct super_block *sb) return -EINVAL; } - HFSPLUS_SB(sb).blockoffset = part_start >> - (sb->s_blocksize_bits - HFSPLUS_SECTOR_SHIFT); - HFSPLUS_SB(sb).sect_count = part_size; - HFSPLUS_SB(sb).fs_shift = HFSPLUS_SB(sb).alloc_blksz_shift - - sb->s_blocksize_bits; + sbi->blockoffset = + part_start >> (sb->s_blocksize_bits - HFSPLUS_SECTOR_SHIFT); + sbi->sect_count = part_size; + sbi->fs_shift = sbi->alloc_blksz_shift - sb->s_blocksize_bits; bh = sb_bread512(sb, part_start + HFSPLUS_VOLHEAD_SECTOR, vhdr); if (!bh) return -EIO; /* should still be the same... */ - if (vhdr->signature != (HFSPLUS_SB(sb).flags & HFSPLUS_SB_HFSX ? + if (vhdr->signature != (sbi->flags & HFSPLUS_SB_HFSX ? cpu_to_be16(HFSPLUS_VOLHEAD_SIGX) : cpu_to_be16(HFSPLUS_VOLHEAD_SIG))) goto error; - HFSPLUS_SB(sb).s_vhbh = bh; - HFSPLUS_SB(sb).s_vhdr = vhdr; + sbi->s_vhbh = bh; + sbi->s_vhdr = vhdr; return 0; error: -- cgit v0.10.2 From 6af502de224c3742936d54eee7e3690c09822934 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:43:31 +0200 Subject: hfsplus: fix HFSPLUS_I calling convention HFSPLUS_I doesn't return a pointer to the hfsplus-specific inode information like all other FOO_I macros, but dereference the pointer in a way that made it look like a direct struct derefence. This only works as long as the HFSPLUS_I macro is used directly and prevents us from keepig a local hfsplus_inode_info pointer. Fix the calling convention and introduce a local hip variable in all functions that use it constantly. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/btree.c b/fs/hfsplus/btree.c index 96e2128..f75cf22 100644 --- a/fs/hfsplus/btree.c +++ b/fs/hfsplus/btree.c @@ -192,17 +192,18 @@ struct hfs_bnode *hfs_bmap_alloc(struct hfs_btree *tree) while (!tree->free_nodes) { struct inode *inode = tree->inode; + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); u32 count; int res; res = hfsplus_file_extend(inode); if (res) return ERR_PTR(res); - HFSPLUS_I(inode).phys_size = inode->i_size = - (loff_t)HFSPLUS_I(inode).alloc_blocks << + hip->phys_size = inode->i_size = + (loff_t)hip->alloc_blocks << HFSPLUS_SB(tree->sb)->alloc_blksz_shift; - HFSPLUS_I(inode).fs_blocks = HFSPLUS_I(inode).alloc_blocks << - HFSPLUS_SB(tree->sb)->fs_shift; + hip->fs_blocks = + hip->alloc_blocks << HFSPLUS_SB(tree->sb)->fs_shift; inode_set_bytes(inode, inode->i_size); count = inode->i_size >> tree->node_size_shift; tree->free_nodes = count - tree->node_count; diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index 75ac1e4..48979c4 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -77,8 +77,8 @@ static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) perms->rootflags |= HFSPLUS_FLG_APPEND; else perms->rootflags &= ~HFSPLUS_FLG_APPEND; - HFSPLUS_I(inode).rootflags = perms->rootflags; - HFSPLUS_I(inode).userflags = perms->userflags; + HFSPLUS_I(inode)->rootflags = perms->rootflags; + HFSPLUS_I(inode)->userflags = perms->userflags; perms->mode = cpu_to_be16(inode->i_mode); perms->owner = cpu_to_be32(inode->i_uid); perms->group = cpu_to_be32(inode->i_gid); @@ -95,7 +95,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i memset(folder, 0, sizeof(*folder)); folder->type = cpu_to_be16(HFSPLUS_FOLDER); folder->id = cpu_to_be32(inode->i_ino); - HFSPLUS_I(inode).create_date = + HFSPLUS_I(inode)->create_date = folder->create_date = folder->content_mod_date = folder->attribute_mod_date = @@ -113,7 +113,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i file->type = cpu_to_be16(HFSPLUS_FILE); file->flags = cpu_to_be16(HFSPLUS_FILE_THREAD_EXISTS); file->id = cpu_to_be32(cnid); - HFSPLUS_I(inode).create_date = + HFSPLUS_I(inode)->create_date = file->create_date = file->content_mod_date = file->attribute_mod_date = @@ -133,8 +133,8 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i file->user_info.fdType = cpu_to_be32(HFSP_HARDLINK_TYPE); file->user_info.fdCreator = cpu_to_be32(HFSP_HFSPLUS_CREATOR); file->user_info.fdFlags = cpu_to_be16(0x100); - file->create_date = HFSPLUS_I(sbi->hidden_dir).create_date; - file->permissions.dev = cpu_to_be32(HFSPLUS_I(inode).dev); + file->create_date = HFSPLUS_I(sbi->hidden_dir)->create_date; + file->permissions.dev = cpu_to_be32(HFSPLUS_I(inode)->dev); } return sizeof(*file); } @@ -279,7 +279,7 @@ int hfsplus_delete_cat(u32 cnid, struct inode *dir, struct qstr *str) hfsplus_free_fork(sb, cnid, &fork, HFSPLUS_TYPE_RSRC); } - list_for_each(pos, &HFSPLUS_I(dir).open_dir_list) { + list_for_each(pos, &HFSPLUS_I(dir)->open_dir_list) { struct hfsplus_readdir_data *rd = list_entry(pos, struct hfsplus_readdir_data, list); if (fd.tree->keycmp(fd.search_key, (void *)&rd->key) < 0) diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 584777d..5ca8308 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -68,8 +68,8 @@ again: cnid = be32_to_cpu(entry.file.id); if (entry.file.user_info.fdType == cpu_to_be32(HFSP_HARDLINK_TYPE) && entry.file.user_info.fdCreator == cpu_to_be32(HFSP_HFSPLUS_CREATOR) && - (entry.file.create_date == HFSPLUS_I(HFSPLUS_SB(sb)->hidden_dir).create_date || - entry.file.create_date == HFSPLUS_I(sb->s_root->d_inode).create_date) && + (entry.file.create_date == HFSPLUS_I(HFSPLUS_SB(sb)->hidden_dir)->create_date || + entry.file.create_date == HFSPLUS_I(sb->s_root->d_inode)->create_date) && HFSPLUS_SB(sb)->hidden_dir) { struct qstr str; char name[32]; @@ -102,7 +102,7 @@ again: if (IS_ERR(inode)) return ERR_CAST(inode); if (S_ISREG(inode->i_mode)) - HFSPLUS_I(inode).dev = linkid; + HFSPLUS_I(inode)->dev = linkid; out: d_add(dentry, inode); return NULL; @@ -219,7 +219,7 @@ static int hfsplus_readdir(struct file *filp, void *dirent, filldir_t filldir) } filp->private_data = rd; rd->file = filp; - list_add(&rd->list, &HFSPLUS_I(inode).open_dir_list); + list_add(&rd->list, &HFSPLUS_I(inode)->open_dir_list); } memcpy(&rd->key, fd.key, sizeof(struct hfsplus_cat_key)); out: @@ -287,7 +287,7 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, if (res != -EEXIST) return res; } - HFSPLUS_I(inode).dev = id; + HFSPLUS_I(inode)->dev = id; cnid = sbi->next_cnid++; src_dentry->d_fsdata = (void *)(unsigned long)cnid; res = hfsplus_create_cat(cnid, src_dir, &src_dentry->d_name, inode); @@ -326,7 +326,7 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) cnid = (u32)(unsigned long)dentry->d_fsdata; if (inode->i_ino == cnid && - atomic_read(&HFSPLUS_I(inode).opencnt)) { + atomic_read(&HFSPLUS_I(inode)->opencnt)) { str.name = name; str.len = sprintf(name, "temp%lu", inode->i_ino); res = hfsplus_rename_cat(inode->i_ino, @@ -347,7 +347,7 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) if (!inode->i_nlink) { if (inode->i_ino != cnid) { sbi->file_count--; - if (!atomic_read(&HFSPLUS_I(inode).opencnt)) { + if (!atomic_read(&HFSPLUS_I(inode)->opencnt)) { res = hfsplus_delete_cat(inode->i_ino, sbi->hidden_dir, NULL); diff --git a/fs/hfsplus/extents.c b/fs/hfsplus/extents.c index 1819698..b1017ea 100644 --- a/fs/hfsplus/extents.c +++ b/fs/hfsplus/extents.c @@ -85,27 +85,32 @@ static u32 hfsplus_ext_lastblock(struct hfsplus_extent *ext) static void __hfsplus_ext_write_extent(struct inode *inode, struct hfs_find_data *fd) { + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); int res; - hfsplus_ext_build_key(fd->search_key, inode->i_ino, HFSPLUS_I(inode).cached_start, - HFSPLUS_IS_RSRC(inode) ? HFSPLUS_TYPE_RSRC : HFSPLUS_TYPE_DATA); + hfsplus_ext_build_key(fd->search_key, inode->i_ino, hip->cached_start, + HFSPLUS_IS_RSRC(inode) ? + HFSPLUS_TYPE_RSRC : HFSPLUS_TYPE_DATA); + res = hfs_brec_find(fd); - if (HFSPLUS_I(inode).flags & HFSPLUS_FLG_EXT_NEW) { + if (hip->flags & HFSPLUS_FLG_EXT_NEW) { if (res != -ENOENT) return; - hfs_brec_insert(fd, HFSPLUS_I(inode).cached_extents, sizeof(hfsplus_extent_rec)); - HFSPLUS_I(inode).flags &= ~(HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW); + hfs_brec_insert(fd, hip->cached_extents, + sizeof(hfsplus_extent_rec)); + hip->flags &= ~(HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW); } else { if (res) return; - hfs_bnode_write(fd->bnode, HFSPLUS_I(inode).cached_extents, fd->entryoffset, fd->entrylength); - HFSPLUS_I(inode).flags &= ~HFSPLUS_FLG_EXT_DIRTY; + hfs_bnode_write(fd->bnode, hip->cached_extents, + fd->entryoffset, fd->entrylength); + hip->flags &= ~HFSPLUS_FLG_EXT_DIRTY; } } void hfsplus_ext_write_extent(struct inode *inode) { - if (HFSPLUS_I(inode).flags & HFSPLUS_FLG_EXT_DIRTY) { + if (HFSPLUS_I(inode)->flags & HFSPLUS_FLG_EXT_DIRTY) { struct hfs_find_data fd; hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); @@ -136,30 +141,34 @@ static inline int __hfsplus_ext_read_extent(struct hfs_find_data *fd, static inline int __hfsplus_ext_cache_extent(struct hfs_find_data *fd, struct inode *inode, u32 block) { + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); int res; - if (HFSPLUS_I(inode).flags & HFSPLUS_FLG_EXT_DIRTY) + if (hip->flags & HFSPLUS_FLG_EXT_DIRTY) __hfsplus_ext_write_extent(inode, fd); - res = __hfsplus_ext_read_extent(fd, HFSPLUS_I(inode).cached_extents, inode->i_ino, - block, HFSPLUS_IS_RSRC(inode) ? HFSPLUS_TYPE_RSRC : HFSPLUS_TYPE_DATA); + res = __hfsplus_ext_read_extent(fd, hip->cached_extents, inode->i_ino, + block, HFSPLUS_IS_RSRC(inode) ? + HFSPLUS_TYPE_RSRC : + HFSPLUS_TYPE_DATA); if (!res) { - HFSPLUS_I(inode).cached_start = be32_to_cpu(fd->key->ext.start_block); - HFSPLUS_I(inode).cached_blocks = hfsplus_ext_block_count(HFSPLUS_I(inode).cached_extents); + hip->cached_start = be32_to_cpu(fd->key->ext.start_block); + hip->cached_blocks = hfsplus_ext_block_count(hip->cached_extents); } else { - HFSPLUS_I(inode).cached_start = HFSPLUS_I(inode).cached_blocks = 0; - HFSPLUS_I(inode).flags &= ~(HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW); + hip->cached_start = hip->cached_blocks = 0; + hip->flags &= ~(HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW); } return res; } static int hfsplus_ext_read_extent(struct inode *inode, u32 block) { + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); struct hfs_find_data fd; int res; - if (block >= HFSPLUS_I(inode).cached_start && - block < HFSPLUS_I(inode).cached_start + HFSPLUS_I(inode).cached_blocks) + if (block >= hip->cached_start && + block < hip->cached_start + hip->cached_blocks) return 0; hfs_find_init(HFSPLUS_SB(inode->i_sb)->ext_tree, &fd); @@ -174,6 +183,7 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock, { struct super_block *sb = inode->i_sb; struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); int res = -EIO; u32 ablock, dblock, mask; int shift; @@ -182,10 +192,10 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock, shift = sbi->alloc_blksz_shift - sb->s_blocksize_bits; ablock = iblock >> sbi->fs_shift; - if (iblock >= HFSPLUS_I(inode).fs_blocks) { - if (iblock > HFSPLUS_I(inode).fs_blocks || !create) + if (iblock >= hip->fs_blocks) { + if (iblock > hip->fs_blocks || !create) return -EIO; - if (ablock >= HFSPLUS_I(inode).alloc_blocks) { + if (ablock >= hip->alloc_blocks) { res = hfsplus_file_extend(inode); if (res) return res; @@ -193,24 +203,24 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock, } else create = 0; - if (ablock < HFSPLUS_I(inode).first_blocks) { - dblock = hfsplus_ext_find_block(HFSPLUS_I(inode).first_extents, ablock); + if (ablock < hip->first_blocks) { + dblock = hfsplus_ext_find_block(hip->first_extents, ablock); goto done; } if (inode->i_ino == HFSPLUS_EXT_CNID) return -EIO; - mutex_lock(&HFSPLUS_I(inode).extents_lock); + mutex_lock(&hip->extents_lock); res = hfsplus_ext_read_extent(inode, ablock); if (!res) { - dblock = hfsplus_ext_find_block(HFSPLUS_I(inode).cached_extents, ablock - - HFSPLUS_I(inode).cached_start); + dblock = hfsplus_ext_find_block(hip->cached_extents, + ablock - hip->cached_start); } else { - mutex_unlock(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&hip->extents_lock); return -EIO; } - mutex_unlock(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&hip->extents_lock); done: dprint(DBG_EXTENT, "get_block(%lu): %llu - %u\n", inode->i_ino, (long long)iblock, dblock); @@ -218,8 +228,8 @@ done: map_bh(bh_result, sb, (dblock << sbi->fs_shift) + sbi->blockoffset + (iblock & mask)); if (create) { set_buffer_new(bh_result); - HFSPLUS_I(inode).phys_size += sb->s_blocksize; - HFSPLUS_I(inode).fs_blocks++; + hip->phys_size += sb->s_blocksize; + hip->fs_blocks++; inode_add_bytes(inode, sb->s_blocksize); mark_inode_dirty(inode); } @@ -348,6 +358,7 @@ int hfsplus_file_extend(struct inode *inode) { struct super_block *sb = inode->i_sb; struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); u32 start, len, goal; int res; @@ -360,17 +371,17 @@ int hfsplus_file_extend(struct inode *inode) return -ENOSPC; } - mutex_lock(&HFSPLUS_I(inode).extents_lock); - if (HFSPLUS_I(inode).alloc_blocks == HFSPLUS_I(inode).first_blocks) - goal = hfsplus_ext_lastblock(HFSPLUS_I(inode).first_extents); + mutex_lock(&hip->extents_lock); + if (hip->alloc_blocks == hip->first_blocks) + goal = hfsplus_ext_lastblock(hip->first_extents); else { - res = hfsplus_ext_read_extent(inode, HFSPLUS_I(inode).alloc_blocks); + res = hfsplus_ext_read_extent(inode, hip->alloc_blocks); if (res) goto out; - goal = hfsplus_ext_lastblock(HFSPLUS_I(inode).cached_extents); + goal = hfsplus_ext_lastblock(hip->cached_extents); } - len = HFSPLUS_I(inode).clump_blocks; + len = hip->clump_blocks; start = hfsplus_block_allocate(sb, sbi->total_blocks, goal, &len); if (start >= sbi->total_blocks) { start = hfsplus_block_allocate(sb, goal, 0, &len); @@ -381,41 +392,41 @@ int hfsplus_file_extend(struct inode *inode) } dprint(DBG_EXTENT, "extend %lu: %u,%u\n", inode->i_ino, start, len); - if (HFSPLUS_I(inode).alloc_blocks <= HFSPLUS_I(inode).first_blocks) { - if (!HFSPLUS_I(inode).first_blocks) { + + if (hip->alloc_blocks <= hip->first_blocks) { + if (!hip->first_blocks) { dprint(DBG_EXTENT, "first extents\n"); /* no extents yet */ - HFSPLUS_I(inode).first_extents[0].start_block = cpu_to_be32(start); - HFSPLUS_I(inode).first_extents[0].block_count = cpu_to_be32(len); + hip->first_extents[0].start_block = cpu_to_be32(start); + hip->first_extents[0].block_count = cpu_to_be32(len); res = 0; } else { /* try to append to extents in inode */ - res = hfsplus_add_extent(HFSPLUS_I(inode).first_extents, - HFSPLUS_I(inode).alloc_blocks, + res = hfsplus_add_extent(hip->first_extents, + hip->alloc_blocks, start, len); if (res == -ENOSPC) goto insert_extent; } if (!res) { - hfsplus_dump_extent(HFSPLUS_I(inode).first_extents); - HFSPLUS_I(inode).first_blocks += len; + hfsplus_dump_extent(hip->first_extents); + hip->first_blocks += len; } } else { - res = hfsplus_add_extent(HFSPLUS_I(inode).cached_extents, - HFSPLUS_I(inode).alloc_blocks - - HFSPLUS_I(inode).cached_start, + res = hfsplus_add_extent(hip->cached_extents, + hip->alloc_blocks - hip->cached_start, start, len); if (!res) { - hfsplus_dump_extent(HFSPLUS_I(inode).cached_extents); - HFSPLUS_I(inode).flags |= HFSPLUS_FLG_EXT_DIRTY; - HFSPLUS_I(inode).cached_blocks += len; + hfsplus_dump_extent(hip->cached_extents); + hip->flags |= HFSPLUS_FLG_EXT_DIRTY; + hip->cached_blocks += len; } else if (res == -ENOSPC) goto insert_extent; } out: - mutex_unlock(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&hip->extents_lock); if (!res) { - HFSPLUS_I(inode).alloc_blocks += len; + hip->alloc_blocks += len; mark_inode_dirty(inode); } return res; @@ -424,13 +435,13 @@ insert_extent: dprint(DBG_EXTENT, "insert new extent\n"); hfsplus_ext_write_extent(inode); - memset(HFSPLUS_I(inode).cached_extents, 0, sizeof(hfsplus_extent_rec)); - HFSPLUS_I(inode).cached_extents[0].start_block = cpu_to_be32(start); - HFSPLUS_I(inode).cached_extents[0].block_count = cpu_to_be32(len); - hfsplus_dump_extent(HFSPLUS_I(inode).cached_extents); - HFSPLUS_I(inode).flags |= HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW; - HFSPLUS_I(inode).cached_start = HFSPLUS_I(inode).alloc_blocks; - HFSPLUS_I(inode).cached_blocks = len; + memset(hip->cached_extents, 0, sizeof(hfsplus_extent_rec)); + hip->cached_extents[0].start_block = cpu_to_be32(start); + hip->cached_extents[0].block_count = cpu_to_be32(len); + hfsplus_dump_extent(hip->cached_extents); + hip->flags |= HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW; + hip->cached_start = hip->alloc_blocks; + hip->cached_blocks = len; res = 0; goto out; @@ -439,13 +450,15 @@ insert_extent: void hfsplus_file_truncate(struct inode *inode) { struct super_block *sb = inode->i_sb; + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); struct hfs_find_data fd; u32 alloc_cnt, blk_cnt, start; int res; - dprint(DBG_INODE, "truncate: %lu, %Lu -> %Lu\n", inode->i_ino, - (long long)HFSPLUS_I(inode).phys_size, inode->i_size); - if (inode->i_size > HFSPLUS_I(inode).phys_size) { + dprint(DBG_INODE, "truncate: %lu, %Lu -> %Lu\n", + inode->i_ino, (long long)hip->phys_size, inode->i_size); + + if (inode->i_size > hip->phys_size) { struct address_space *mapping = inode->i_mapping; struct page *page; void *fsdata; @@ -462,48 +475,48 @@ void hfsplus_file_truncate(struct inode *inode) return; mark_inode_dirty(inode); return; - } else if (inode->i_size == HFSPLUS_I(inode).phys_size) + } else if (inode->i_size == hip->phys_size) return; blk_cnt = (inode->i_size + HFSPLUS_SB(sb)->alloc_blksz - 1) >> HFSPLUS_SB(sb)->alloc_blksz_shift; - alloc_cnt = HFSPLUS_I(inode).alloc_blocks; + alloc_cnt = hip->alloc_blocks; if (blk_cnt == alloc_cnt) goto out; - mutex_lock(&HFSPLUS_I(inode).extents_lock); + mutex_lock(&hip->extents_lock); hfs_find_init(HFSPLUS_SB(sb)->ext_tree, &fd); while (1) { - if (alloc_cnt == HFSPLUS_I(inode).first_blocks) { - hfsplus_free_extents(sb, HFSPLUS_I(inode).first_extents, + if (alloc_cnt == hip->first_blocks) { + hfsplus_free_extents(sb, hip->first_extents, alloc_cnt, alloc_cnt - blk_cnt); - hfsplus_dump_extent(HFSPLUS_I(inode).first_extents); - HFSPLUS_I(inode).first_blocks = blk_cnt; + hfsplus_dump_extent(hip->first_extents); + hip->first_blocks = blk_cnt; break; } res = __hfsplus_ext_cache_extent(&fd, inode, alloc_cnt); if (res) break; - start = HFSPLUS_I(inode).cached_start; - hfsplus_free_extents(sb, HFSPLUS_I(inode).cached_extents, + start = hip->cached_start; + hfsplus_free_extents(sb, hip->cached_extents, alloc_cnt - start, alloc_cnt - blk_cnt); - hfsplus_dump_extent(HFSPLUS_I(inode).cached_extents); + hfsplus_dump_extent(hip->cached_extents); if (blk_cnt > start) { - HFSPLUS_I(inode).flags |= HFSPLUS_FLG_EXT_DIRTY; + hip->flags |= HFSPLUS_FLG_EXT_DIRTY; break; } alloc_cnt = start; - HFSPLUS_I(inode).cached_start = HFSPLUS_I(inode).cached_blocks = 0; - HFSPLUS_I(inode).flags &= ~(HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW); + hip->cached_start = hip->cached_blocks = 0; + hip->flags &= ~(HFSPLUS_FLG_EXT_DIRTY | HFSPLUS_FLG_EXT_NEW); hfs_brec_remove(&fd); } hfs_find_exit(&fd); - mutex_unlock(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&hip->extents_lock); - HFSPLUS_I(inode).alloc_blocks = blk_cnt; + hip->alloc_blocks = blk_cnt; out: - HFSPLUS_I(inode).phys_size = inode->i_size; - HFSPLUS_I(inode).fs_blocks = (inode->i_size + sb->s_blocksize - 1) >> sb->s_blocksize_bits; - inode_set_bytes(inode, HFSPLUS_I(inode).fs_blocks << sb->s_blocksize_bits); + hip->phys_size = inode->i_size; + hip->fs_blocks = (inode->i_size + sb->s_blocksize - 1) >> sb->s_blocksize_bits; + inode_set_bytes(inode, hip->fs_blocks << sb->s_blocksize_bits); mark_inode_dirty(inode); } diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 55f42b4..169cef9 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -187,8 +187,8 @@ struct hfsplus_inode_info { #define HFSPLUS_FLG_EXT_DIRTY 0x0002 #define HFSPLUS_FLG_EXT_NEW 0x0004 -#define HFSPLUS_IS_DATA(inode) (!(HFSPLUS_I(inode).flags & HFSPLUS_FLG_RSRC)) -#define HFSPLUS_IS_RSRC(inode) (HFSPLUS_I(inode).flags & HFSPLUS_FLG_RSRC) +#define HFSPLUS_IS_DATA(inode) (!(HFSPLUS_I(inode)->flags & HFSPLUS_FLG_RSRC)) +#define HFSPLUS_IS_RSRC(inode) (HFSPLUS_I(inode)->flags & HFSPLUS_FLG_RSRC) struct hfs_find_data { /* filled by caller */ @@ -379,13 +379,11 @@ static inline struct hfsplus_sb_info *HFSPLUS_SB(struct super_block *sb) { return sb->s_fs_info; } -/* + static inline struct hfsplus_inode_info *HFSPLUS_I(struct inode *inode) { return list_entry(inode, struct hfsplus_inode_info, vfs_inode); } -*/ -#define HFSPLUS_I(inode) (*list_entry(inode, struct hfsplus_inode_info, vfs_inode)) #if 1 #define hfsplus_kmap(p) ({ struct page *__p = (p); kmap(__p); }) diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index 050eee6..309defb 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -36,7 +36,7 @@ static int hfsplus_write_begin(struct file *file, struct address_space *mapping, *pagep = NULL; ret = cont_write_begin(file, mapping, pos, len, flags, pagep, fsdata, hfsplus_get_block, - &HFSPLUS_I(mapping->host).phys_size); + &HFSPLUS_I(mapping->host)->phys_size); if (unlikely(ret)) { loff_t isize = mapping->host->i_size; if (pos + len > isize) @@ -172,12 +172,13 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, struct dentry *dent struct hfs_find_data fd; struct super_block *sb = dir->i_sb; struct inode *inode = NULL; + struct hfsplus_inode_info *hip; int err; if (HFSPLUS_IS_RSRC(dir) || strcmp(dentry->d_name.name, "rsrc")) goto out; - inode = HFSPLUS_I(dir).rsrc_inode; + inode = HFSPLUS_I(dir)->rsrc_inode; if (inode) goto out; @@ -185,10 +186,11 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, struct dentry *dent if (!inode) return ERR_PTR(-ENOMEM); + hip = HFSPLUS_I(inode); inode->i_ino = dir->i_ino; - INIT_LIST_HEAD(&HFSPLUS_I(inode).open_dir_list); - mutex_init(&HFSPLUS_I(inode).extents_lock); - HFSPLUS_I(inode).flags = HFSPLUS_FLG_RSRC; + INIT_LIST_HEAD(&hip->open_dir_list); + mutex_init(&hip->extents_lock); + hip->flags = HFSPLUS_FLG_RSRC; hfs_find_init(HFSPLUS_SB(sb)->cat_tree, &fd); err = hfsplus_find_cat(sb, dir->i_ino, &fd); @@ -199,8 +201,8 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, struct dentry *dent iput(inode); return ERR_PTR(err); } - HFSPLUS_I(inode).rsrc_inode = dir; - HFSPLUS_I(dir).rsrc_inode = inode; + hip->rsrc_inode = dir; + HFSPLUS_I(dir)->rsrc_inode = inode; igrab(dir); hlist_add_head(&inode->i_hash, &HFSPLUS_SB(sb)->rsrc_inodes); mark_inode_dirty(inode); @@ -231,8 +233,8 @@ static void hfsplus_get_perms(struct inode *inode, struct hfsplus_perm *perms, i mode = S_IFREG | ((S_IRUGO|S_IWUGO) & ~(sbi->umask)); inode->i_mode = mode; - HFSPLUS_I(inode).rootflags = perms->rootflags; - HFSPLUS_I(inode).userflags = perms->userflags; + HFSPLUS_I(inode)->rootflags = perms->rootflags; + HFSPLUS_I(inode)->userflags = perms->userflags; if (perms->rootflags & HFSPLUS_FLG_IMMUTABLE) inode->i_flags |= S_IMMUTABLE; else @@ -253,20 +255,20 @@ static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) perms->rootflags |= HFSPLUS_FLG_APPEND; else perms->rootflags &= ~HFSPLUS_FLG_APPEND; - perms->userflags = HFSPLUS_I(inode).userflags; + perms->userflags = HFSPLUS_I(inode)->userflags; perms->mode = cpu_to_be16(inode->i_mode); perms->owner = cpu_to_be32(inode->i_uid); perms->group = cpu_to_be32(inode->i_gid); - perms->dev = cpu_to_be32(HFSPLUS_I(inode).dev); + perms->dev = cpu_to_be32(HFSPLUS_I(inode)->dev); } static int hfsplus_file_open(struct inode *inode, struct file *file) { if (HFSPLUS_IS_RSRC(inode)) - inode = HFSPLUS_I(inode).rsrc_inode; + inode = HFSPLUS_I(inode)->rsrc_inode; if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS) return -EOVERFLOW; - atomic_inc(&HFSPLUS_I(inode).opencnt); + atomic_inc(&HFSPLUS_I(inode)->opencnt); return 0; } @@ -275,8 +277,8 @@ static int hfsplus_file_release(struct inode *inode, struct file *file) struct super_block *sb = inode->i_sb; if (HFSPLUS_IS_RSRC(inode)) - inode = HFSPLUS_I(inode).rsrc_inode; - if (atomic_dec_and_test(&HFSPLUS_I(inode).opencnt)) { + inode = HFSPLUS_I(inode)->rsrc_inode; + if (atomic_dec_and_test(&HFSPLUS_I(inode)->opencnt)) { mutex_lock(&inode->i_mutex); hfsplus_file_truncate(inode); if (inode->i_flags & S_DEAD) { @@ -362,6 +364,7 @@ struct inode *hfsplus_new_inode(struct super_block *sb, int mode) { struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); struct inode *inode = new_inode(sb); + struct hfsplus_inode_info *hip; if (!inode) return NULL; @@ -372,19 +375,21 @@ struct inode *hfsplus_new_inode(struct super_block *sb, int mode) inode->i_gid = current_fsgid(); inode->i_nlink = 1; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; - INIT_LIST_HEAD(&HFSPLUS_I(inode).open_dir_list); - mutex_init(&HFSPLUS_I(inode).extents_lock); - atomic_set(&HFSPLUS_I(inode).opencnt, 0); - HFSPLUS_I(inode).flags = 0; - memset(HFSPLUS_I(inode).first_extents, 0, sizeof(hfsplus_extent_rec)); - memset(HFSPLUS_I(inode).cached_extents, 0, sizeof(hfsplus_extent_rec)); - HFSPLUS_I(inode).alloc_blocks = 0; - HFSPLUS_I(inode).first_blocks = 0; - HFSPLUS_I(inode).cached_start = 0; - HFSPLUS_I(inode).cached_blocks = 0; - HFSPLUS_I(inode).phys_size = 0; - HFSPLUS_I(inode).fs_blocks = 0; - HFSPLUS_I(inode).rsrc_inode = NULL; + + hip = HFSPLUS_I(inode); + INIT_LIST_HEAD(&hip->open_dir_list); + mutex_init(&hip->extents_lock); + atomic_set(&hip->opencnt, 0); + hip->flags = 0; + memset(hip->first_extents, 0, sizeof(hfsplus_extent_rec)); + memset(hip->cached_extents, 0, sizeof(hfsplus_extent_rec)); + hip->alloc_blocks = 0; + hip->first_blocks = 0; + hip->cached_start = 0; + hip->cached_blocks = 0; + hip->phys_size = 0; + hip->fs_blocks = 0; + hip->rsrc_inode = NULL; if (S_ISDIR(inode->i_mode)) { inode->i_size = 2; sbi->folder_count++; @@ -395,12 +400,12 @@ struct inode *hfsplus_new_inode(struct super_block *sb, int mode) inode->i_op = &hfsplus_file_inode_operations; inode->i_fop = &hfsplus_file_operations; inode->i_mapping->a_ops = &hfsplus_aops; - HFSPLUS_I(inode).clump_blocks = sbi->data_clump_blocks; + hip->clump_blocks = sbi->data_clump_blocks; } else if (S_ISLNK(inode->i_mode)) { sbi->file_count++; inode->i_op = &page_symlink_inode_operations; inode->i_mapping->a_ops = &hfsplus_aops; - HFSPLUS_I(inode).clump_blocks = 1; + hip->clump_blocks = 1; } else sbi->file_count++; insert_inode_hash(inode); @@ -436,26 +441,27 @@ void hfsplus_inode_read_fork(struct inode *inode, struct hfsplus_fork_raw *fork) { struct super_block *sb = inode->i_sb; struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb); + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); u32 count; int i; - memcpy(&HFSPLUS_I(inode).first_extents, &fork->extents, - sizeof(hfsplus_extent_rec)); + memcpy(&hip->first_extents, &fork->extents, sizeof(hfsplus_extent_rec)); for (count = 0, i = 0; i < 8; i++) count += be32_to_cpu(fork->extents[i].block_count); - HFSPLUS_I(inode).first_blocks = count; - memset(HFSPLUS_I(inode).cached_extents, 0, sizeof(hfsplus_extent_rec)); - HFSPLUS_I(inode).cached_start = 0; - HFSPLUS_I(inode).cached_blocks = 0; - - HFSPLUS_I(inode).alloc_blocks = be32_to_cpu(fork->total_blocks); - inode->i_size = HFSPLUS_I(inode).phys_size = be64_to_cpu(fork->total_size); - HFSPLUS_I(inode).fs_blocks = (inode->i_size + sb->s_blocksize - 1) >> sb->s_blocksize_bits; - inode_set_bytes(inode, HFSPLUS_I(inode).fs_blocks << sb->s_blocksize_bits); - HFSPLUS_I(inode).clump_blocks = + hip->first_blocks = count; + memset(hip->cached_extents, 0, sizeof(hfsplus_extent_rec)); + hip->cached_start = 0; + hip->cached_blocks = 0; + + hip->alloc_blocks = be32_to_cpu(fork->total_blocks); + hip->phys_size = inode->i_size = be64_to_cpu(fork->total_size); + hip->fs_blocks = + (inode->i_size + sb->s_blocksize - 1) >> sb->s_blocksize_bits; + inode_set_bytes(inode, hip->fs_blocks << sb->s_blocksize_bits); + hip->clump_blocks = be32_to_cpu(fork->clump_size) >> sbi->alloc_blksz_shift; - if (!HFSPLUS_I(inode).clump_blocks) { - HFSPLUS_I(inode).clump_blocks = HFSPLUS_IS_RSRC(inode) ? + if (!hip->clump_blocks) { + hip->clump_blocks = HFSPLUS_IS_RSRC(inode) ? sbi->rsrc_clump_blocks : sbi->data_clump_blocks; } @@ -463,10 +469,10 @@ void hfsplus_inode_read_fork(struct inode *inode, struct hfsplus_fork_raw *fork) void hfsplus_inode_write_fork(struct inode *inode, struct hfsplus_fork_raw *fork) { - memcpy(&fork->extents, &HFSPLUS_I(inode).first_extents, + memcpy(&fork->extents, &HFSPLUS_I(inode)->first_extents, sizeof(hfsplus_extent_rec)); fork->total_size = cpu_to_be64(inode->i_size); - fork->total_blocks = cpu_to_be32(HFSPLUS_I(inode).alloc_blocks); + fork->total_blocks = cpu_to_be32(HFSPLUS_I(inode)->alloc_blocks); } int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) @@ -477,7 +483,7 @@ int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) type = hfs_bnode_read_u16(fd->bnode, fd->entryoffset); - HFSPLUS_I(inode).dev = 0; + HFSPLUS_I(inode)->dev = 0; if (type == HFSPLUS_FOLDER) { struct hfsplus_cat_folder *folder = &entry.folder; @@ -491,8 +497,8 @@ int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) inode->i_atime = hfsp_mt2ut(folder->access_date); inode->i_mtime = hfsp_mt2ut(folder->content_mod_date); inode->i_ctime = hfsp_mt2ut(folder->attribute_mod_date); - HFSPLUS_I(inode).create_date = folder->create_date; - HFSPLUS_I(inode).fs_blocks = 0; + HFSPLUS_I(inode)->create_date = folder->create_date; + HFSPLUS_I(inode)->fs_blocks = 0; inode->i_op = &hfsplus_dir_inode_operations; inode->i_fop = &hfsplus_dir_operations; } else if (type == HFSPLUS_FILE) { @@ -523,7 +529,7 @@ int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) inode->i_atime = hfsp_mt2ut(file->access_date); inode->i_mtime = hfsp_mt2ut(file->content_mod_date); inode->i_ctime = hfsp_mt2ut(file->attribute_mod_date); - HFSPLUS_I(inode).create_date = file->create_date; + HFSPLUS_I(inode)->create_date = file->create_date; } else { printk(KERN_ERR "hfs: bad catalog entry used to create inode\n"); res = -EIO; @@ -538,7 +544,7 @@ int hfsplus_cat_write_inode(struct inode *inode) hfsplus_cat_entry entry; if (HFSPLUS_IS_RSRC(inode)) - main_inode = HFSPLUS_I(inode).rsrc_inode; + main_inode = HFSPLUS_I(inode)->rsrc_inode; if (!main_inode->i_nlink) return 0; @@ -582,9 +588,9 @@ int hfsplus_cat_write_inode(struct inode *inode) sizeof(struct hfsplus_cat_file)); hfsplus_inode_write_fork(inode, &file->data_fork); if (S_ISREG(inode->i_mode)) - HFSPLUS_I(inode).dev = inode->i_nlink; + HFSPLUS_I(inode)->dev = inode->i_nlink; if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) - HFSPLUS_I(inode).dev = kdev_t_to_nr(inode->i_rdev); + HFSPLUS_I(inode)->dev = kdev_t_to_nr(inode->i_rdev); hfsplus_set_perms(inode, &file->permissions); if ((file->permissions.rootflags | file->permissions.userflags) & HFSPLUS_FLG_IMMUTABLE) file->flags |= cpu_to_be16(HFSPLUS_FILE_LOCKED); diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index 66dd64b..c9ac443 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -23,13 +23,14 @@ static int hfsplus_ioctl_getflags(struct file *file, int __user *user_flags) { struct inode *inode = file->f_path.dentry->d_inode; + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); unsigned int flags = 0; - if (HFSPLUS_I(inode).rootflags & HFSPLUS_FLG_IMMUTABLE) + if (hip->rootflags & HFSPLUS_FLG_IMMUTABLE) flags |= FS_IMMUTABLE_FL; - if (HFSPLUS_I(inode).rootflags & HFSPLUS_FLG_APPEND) + if (hip->rootflags & HFSPLUS_FLG_APPEND) flags |= FS_APPEND_FL; - if (HFSPLUS_I(inode).userflags & HFSPLUS_FLG_NODUMP) + if (hip->userflags & HFSPLUS_FLG_NODUMP) flags |= FS_NODUMP_FL; return put_user(flags, user_flags); @@ -38,6 +39,7 @@ static int hfsplus_ioctl_getflags(struct file *file, int __user *user_flags) static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) { struct inode *inode = file->f_path.dentry->d_inode; + struct hfsplus_inode_info *hip = HFSPLUS_I(inode); unsigned int flags; int err = 0; @@ -58,7 +60,7 @@ static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) mutex_lock(&inode->i_mutex); if (flags & (FS_IMMUTABLE_FL|FS_APPEND_FL) || - HFSPLUS_I(inode).rootflags & (HFSPLUS_FLG_IMMUTABLE|HFSPLUS_FLG_APPEND)) { + hip->rootflags & (HFSPLUS_FLG_IMMUTABLE|HFSPLUS_FLG_APPEND)) { if (!capable(CAP_LINUX_IMMUTABLE)) { err = -EPERM; goto out_unlock_inode; @@ -72,22 +74,22 @@ static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) } if (flags & FS_IMMUTABLE_FL) { inode->i_flags |= S_IMMUTABLE; - HFSPLUS_I(inode).rootflags |= HFSPLUS_FLG_IMMUTABLE; + hip->rootflags |= HFSPLUS_FLG_IMMUTABLE; } else { inode->i_flags &= ~S_IMMUTABLE; - HFSPLUS_I(inode).rootflags &= ~HFSPLUS_FLG_IMMUTABLE; + hip->rootflags &= ~HFSPLUS_FLG_IMMUTABLE; } if (flags & FS_APPEND_FL) { inode->i_flags |= S_APPEND; - HFSPLUS_I(inode).rootflags |= HFSPLUS_FLG_APPEND; + hip->rootflags |= HFSPLUS_FLG_APPEND; } else { inode->i_flags &= ~S_APPEND; - HFSPLUS_I(inode).rootflags &= ~HFSPLUS_FLG_APPEND; + hip->rootflags &= ~HFSPLUS_FLG_APPEND; } if (flags & FS_NODUMP_FL) - HFSPLUS_I(inode).userflags |= HFSPLUS_FLG_NODUMP; + hip->userflags |= HFSPLUS_FLG_NODUMP; else - HFSPLUS_I(inode).userflags &= ~HFSPLUS_FLG_NODUMP; + hip->userflags &= ~HFSPLUS_FLG_NODUMP; inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 1bf00b9..a7bf89e 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -33,11 +33,11 @@ struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) if (!(inode->i_state & I_NEW)) return inode; - INIT_LIST_HEAD(&HFSPLUS_I(inode).open_dir_list); - mutex_init(&HFSPLUS_I(inode).extents_lock); - HFSPLUS_I(inode).flags = 0; - HFSPLUS_I(inode).rsrc_inode = NULL; - atomic_set(&HFSPLUS_I(inode).opencnt, 0); + INIT_LIST_HEAD(&HFSPLUS_I(inode)->open_dir_list); + mutex_init(&HFSPLUS_I(inode)->extents_lock); + HFSPLUS_I(inode)->flags = 0; + HFSPLUS_I(inode)->rsrc_inode = NULL; + atomic_set(&HFSPLUS_I(inode)->opencnt, 0); if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID) { read_inode: @@ -151,8 +151,8 @@ static void hfsplus_evict_inode(struct inode *inode) truncate_inode_pages(&inode->i_data, 0); end_writeback(inode); if (HFSPLUS_IS_RSRC(inode)) { - HFSPLUS_I(HFSPLUS_I(inode).rsrc_inode).rsrc_inode = NULL; - iput(HFSPLUS_I(inode).rsrc_inode); + HFSPLUS_I(HFSPLUS_I(inode)->rsrc_inode)->rsrc_inode = NULL; + iput(HFSPLUS_I(inode)->rsrc_inode); } } @@ -491,7 +491,7 @@ static struct inode *hfsplus_alloc_inode(struct super_block *sb) static void hfsplus_destroy_inode(struct inode *inode) { - kmem_cache_free(hfsplus_inode_cachep, &HFSPLUS_I(inode)); + kmem_cache_free(hfsplus_inode_cachep, HFSPLUS_I(inode)); } #define HFSPLUS_INODE_SIZE sizeof(struct hfsplus_inode_info) -- cgit v0.10.2 From fc4fff82104fa096eada73943fe5249500acd5fa Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:43:41 +0200 Subject: hfsplus: clean up hfsplus_iget Add a new hfsplus_system_read_inode for reading the special system inodes and streamline the fastpath iget code. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index a7bf89e..bd09ea2 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -20,12 +20,42 @@ static void hfsplus_destroy_inode(struct inode *inode); #include "hfsplus_fs.h" +static int hfsplus_system_read_inode(struct inode *inode) +{ + struct hfsplus_vh *vhdr = HFSPLUS_SB(inode->i_sb)->s_vhdr; + + switch (inode->i_ino) { + case HFSPLUS_EXT_CNID: + hfsplus_inode_read_fork(inode, &vhdr->ext_file); + inode->i_mapping->a_ops = &hfsplus_btree_aops; + break; + case HFSPLUS_CAT_CNID: + hfsplus_inode_read_fork(inode, &vhdr->cat_file); + inode->i_mapping->a_ops = &hfsplus_btree_aops; + break; + case HFSPLUS_ALLOC_CNID: + hfsplus_inode_read_fork(inode, &vhdr->alloc_file); + inode->i_mapping->a_ops = &hfsplus_aops; + break; + case HFSPLUS_START_CNID: + hfsplus_inode_read_fork(inode, &vhdr->start_file); + break; + case HFSPLUS_ATTR_CNID: + hfsplus_inode_read_fork(inode, &vhdr->attr_file); + inode->i_mapping->a_ops = &hfsplus_btree_aops; + break; + default: + return -EIO; + } + + return 0; +} + struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) { struct hfs_find_data fd; - struct hfsplus_vh *vhdr; struct inode *inode; - long err = -EIO; + int err; inode = iget_locked(sb, ino); if (!inode) @@ -39,51 +69,24 @@ struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) HFSPLUS_I(inode)->rsrc_inode = NULL; atomic_set(&HFSPLUS_I(inode)->opencnt, 0); - if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID) { - read_inode: + if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID || + inode->i_ino == HFSPLUS_ROOT_CNID) { hfs_find_init(HFSPLUS_SB(inode->i_sb)->cat_tree, &fd); err = hfsplus_find_cat(inode->i_sb, inode->i_ino, &fd); if (!err) err = hfsplus_cat_read_inode(inode, &fd); hfs_find_exit(&fd); - if (err) - goto bad_inode; - goto done; + } else { + err = hfsplus_system_read_inode(inode); } - vhdr = HFSPLUS_SB(inode->i_sb)->s_vhdr; - switch(inode->i_ino) { - case HFSPLUS_ROOT_CNID: - goto read_inode; - case HFSPLUS_EXT_CNID: - hfsplus_inode_read_fork(inode, &vhdr->ext_file); - inode->i_mapping->a_ops = &hfsplus_btree_aops; - break; - case HFSPLUS_CAT_CNID: - hfsplus_inode_read_fork(inode, &vhdr->cat_file); - inode->i_mapping->a_ops = &hfsplus_btree_aops; - break; - case HFSPLUS_ALLOC_CNID: - hfsplus_inode_read_fork(inode, &vhdr->alloc_file); - inode->i_mapping->a_ops = &hfsplus_aops; - break; - case HFSPLUS_START_CNID: - hfsplus_inode_read_fork(inode, &vhdr->start_file); - break; - case HFSPLUS_ATTR_CNID: - hfsplus_inode_read_fork(inode, &vhdr->attr_file); - inode->i_mapping->a_ops = &hfsplus_btree_aops; - break; - default: - goto bad_inode; + + if (err) { + iget_failed(inode); + return ERR_PTR(err); } -done: unlock_new_inode(inode); return inode; - -bad_inode: - iget_failed(inode); - return ERR_PTR(err); } static int hfsplus_write_inode(struct inode *inode, -- cgit v0.10.2 From b5080f77ed2de3c8ac67a63044f8a781c75207d9 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:43:43 +0200 Subject: hfsplus: clean up hfsplus_write_inode Add a new hfsplus_system_write_inode for writing the special system inodes and streamline the fastpath write_inode code. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index bd09ea2..e485a38 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -89,63 +89,57 @@ struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) return inode; } -static int hfsplus_write_inode(struct inode *inode, - struct writeback_control *wbc) +static int hfsplus_system_write_inode(struct inode *inode) { struct hfsplus_sb_info *sbi = HFSPLUS_SB(inode->i_sb); - struct hfsplus_vh *vhdr; - int ret = 0; + struct hfsplus_vh *vhdr = sbi->s_vhdr; + struct hfsplus_fork_raw *fork; + struct hfs_btree *tree = NULL; - dprint(DBG_INODE, "hfsplus_write_inode: %lu\n", inode->i_ino); - hfsplus_ext_write_extent(inode); - if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID) { - return hfsplus_cat_write_inode(inode); - } - vhdr = sbi->s_vhdr; switch (inode->i_ino) { - case HFSPLUS_ROOT_CNID: - ret = hfsplus_cat_write_inode(inode); - break; case HFSPLUS_EXT_CNID: - if (vhdr->ext_file.total_size != cpu_to_be64(inode->i_size)) { - sbi->flags |= HFSPLUS_SB_WRITEBACKUP; - inode->i_sb->s_dirt = 1; - } - hfsplus_inode_write_fork(inode, &vhdr->ext_file); - hfs_btree_write(sbi->ext_tree); + fork = &vhdr->ext_file; + tree = sbi->ext_tree; break; case HFSPLUS_CAT_CNID: - if (vhdr->cat_file.total_size != cpu_to_be64(inode->i_size)) { - sbi->flags |= HFSPLUS_SB_WRITEBACKUP; - inode->i_sb->s_dirt = 1; - } - hfsplus_inode_write_fork(inode, &vhdr->cat_file); - hfs_btree_write(sbi->cat_tree); + fork = &vhdr->cat_file; + tree = sbi->cat_tree; break; case HFSPLUS_ALLOC_CNID: - if (vhdr->alloc_file.total_size != cpu_to_be64(inode->i_size)) { - sbi->flags |= HFSPLUS_SB_WRITEBACKUP; - inode->i_sb->s_dirt = 1; - } - hfsplus_inode_write_fork(inode, &vhdr->alloc_file); + fork = &vhdr->alloc_file; break; case HFSPLUS_START_CNID: - if (vhdr->start_file.total_size != cpu_to_be64(inode->i_size)) { - sbi->flags |= HFSPLUS_SB_WRITEBACKUP; - inode->i_sb->s_dirt = 1; - } - hfsplus_inode_write_fork(inode, &vhdr->start_file); + fork = &vhdr->start_file; break; case HFSPLUS_ATTR_CNID: - if (vhdr->attr_file.total_size != cpu_to_be64(inode->i_size)) { - sbi->flags |= HFSPLUS_SB_WRITEBACKUP; - inode->i_sb->s_dirt = 1; - } - hfsplus_inode_write_fork(inode, &vhdr->attr_file); - hfs_btree_write(sbi->attr_tree); - break; + fork = &vhdr->attr_file; + tree = sbi->attr_tree; + default: + return -EIO; + } + + if (fork->total_size != cpu_to_be64(inode->i_size)) { + sbi->flags |= HFSPLUS_SB_WRITEBACKUP; + inode->i_sb->s_dirt = 1; } - return ret; + hfsplus_inode_write_fork(inode, fork); + if (tree) + hfs_btree_write(tree); + return 0; +} + +static int hfsplus_write_inode(struct inode *inode, + struct writeback_control *wbc) +{ + dprint(DBG_INODE, "hfsplus_write_inode: %lu\n", inode->i_ino); + + hfsplus_ext_write_extent(inode); + + if (inode->i_ino >= HFSPLUS_FIRSTUSER_CNID || + inode->i_ino == HFSPLUS_ROOT_CNID) + return hfsplus_cat_write_inode(inode); + else + return hfsplus_system_write_inode(inode); } static void hfsplus_evict_inode(struct inode *inode) -- cgit v0.10.2 From 30d3abbec730a5a9c954a6342271f7a7db155b08 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:43:50 +0200 Subject: hfsplus: merge mknod/mkdir/creat Make hfsplus_mkdir and hfsplus_create call hfsplus_mknod instead of duplicating the code. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 5ca8308..7efcf75 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -237,28 +237,6 @@ static int hfsplus_dir_release(struct inode *inode, struct file *file) return 0; } -static int hfsplus_create(struct inode *dir, struct dentry *dentry, int mode, - struct nameidata *nd) -{ - struct inode *inode; - int res; - - inode = hfsplus_new_inode(dir->i_sb, mode); - if (!inode) - return -ENOSPC; - - res = hfsplus_create_cat(inode->i_ino, dir, &dentry->d_name, inode); - if (res) { - inode->i_nlink = 0; - hfsplus_delete_inode(inode); - iput(inode); - return res; - } - hfsplus_instantiate(dentry, inode, inode->i_ino); - mark_inode_dirty(inode); - return 0; -} - static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, struct dentry *dst_dentry) { @@ -365,27 +343,6 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) return res; } -static int hfsplus_mkdir(struct inode *dir, struct dentry *dentry, int mode) -{ - struct inode *inode; - int res; - - inode = hfsplus_new_inode(dir->i_sb, S_IFDIR | mode); - if (!inode) - return -ENOSPC; - - res = hfsplus_create_cat(inode->i_ino, dir, &dentry->d_name, inode); - if (res) { - inode->i_nlink = 0; - hfsplus_delete_inode(inode); - iput(inode); - return res; - } - hfsplus_instantiate(dentry, inode, inode->i_ino); - mark_inode_dirty(inode); - return 0; -} - static int hfsplus_rmdir(struct inode *dir, struct dentry *dentry) { struct inode *inode; @@ -438,12 +395,10 @@ static int hfsplus_symlink(struct inode *dir, struct dentry *dentry, static int hfsplus_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev) { - struct super_block *sb; struct inode *inode; int res; - sb = dir->i_sb; - inode = hfsplus_new_inode(sb, mode); + inode = hfsplus_new_inode(dir->i_sb, mode); if (!inode) return -ENOSPC; @@ -454,13 +409,26 @@ static int hfsplus_mknod(struct inode *dir, struct dentry *dentry, iput(inode); return res; } - init_special_inode(inode, mode, rdev); + + if (S_ISBLK(mode) || S_ISCHR(mode) || S_ISFIFO(mode) || S_ISSOCK(mode)) + init_special_inode(inode, mode, rdev); + hfsplus_instantiate(dentry, inode, inode->i_ino); mark_inode_dirty(inode); - return 0; } +static int hfsplus_create(struct inode *dir, struct dentry *dentry, int mode, + struct nameidata *nd) +{ + return hfsplus_mknod(dir, dentry, mode, 0); +} + +static int hfsplus_mkdir(struct inode *dir, struct dentry *dentry, int mode) +{ + return hfsplus_mknod(dir, dentry, mode | S_IFDIR, 0); +} + static int hfsplus_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry) { -- cgit v0.10.2 From f17c89bfcc9cccd405098eac3ec1ebfddf03279e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:43:54 +0200 Subject: hfsplus: fix error handling in hfsplus_symlink We need to free the inode again on a hfsplus_create_cat failure. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 7efcf75..f8ae468 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -364,31 +364,29 @@ static int hfsplus_rmdir(struct inode *dir, struct dentry *dentry) static int hfsplus_symlink(struct inode *dir, struct dentry *dentry, const char *symname) { - struct super_block *sb; struct inode *inode; int res; - sb = dir->i_sb; - inode = hfsplus_new_inode(sb, S_IFLNK | S_IRWXUGO); + inode = hfsplus_new_inode(dir->i_sb, S_IFLNK | S_IRWXUGO); if (!inode) return -ENOSPC; res = page_symlink(inode, symname, strlen(symname) + 1); - if (res) { - inode->i_nlink = 0; - hfsplus_delete_inode(inode); - iput(inode); - return res; - } + if (res) + goto out_err; - mark_inode_dirty(inode); res = hfsplus_create_cat(inode->i_ino, dir, &dentry->d_name, inode); + if (res) + goto out_err; - if (!res) { - hfsplus_instantiate(dentry, inode, inode->i_ino); - mark_inode_dirty(inode); - } + hfsplus_instantiate(dentry, inode, inode->i_ino); + mark_inode_dirty(inode); + return 0; +out_err: + inode->i_nlink = 0; + hfsplus_delete_inode(inode); + iput(inode); return res; } -- cgit v0.10.2 From 66e5db05bb6670f314d90aba5998e6a033e4d563 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:43:58 +0200 Subject: hfsplus: do not cache and write next_alloc We never look at it, nor change the next_alloc field in the superblock. So don't bother caching it or writing it out in hfsplus_sync_fs. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 169cef9..5e2418a 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -129,7 +129,6 @@ struct hfsplus_sb_info { int alloc_blksz_shift; u32 total_blocks; u32 free_blocks; - u32 next_alloc; u32 next_cnid; u32 file_count; u32 folder_count; diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index e485a38..f310a1f 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -164,7 +164,6 @@ int hfsplus_sync_fs(struct super_block *sb, int wait) sb->s_dirt = 0; vhdr->free_blocks = cpu_to_be32(sbi->free_blocks); - vhdr->next_alloc = cpu_to_be32(sbi->next_alloc); vhdr->next_cnid = cpu_to_be32(sbi->next_cnid); vhdr->folder_count = cpu_to_be32(sbi->folder_count); vhdr->file_count = cpu_to_be32(sbi->file_count); @@ -355,7 +354,6 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) } sbi->total_blocks = be32_to_cpu(vhdr->total_blocks); sbi->free_blocks = be32_to_cpu(vhdr->free_blocks); - sbi->next_alloc = be32_to_cpu(vhdr->next_alloc); sbi->next_cnid = be32_to_cpu(vhdr->next_cnid); sbi->file_count = be32_to_cpu(vhdr->file_count); sbi->folder_count = be32_to_cpu(vhdr->folder_count); -- cgit v0.10.2 From 58a818f532e83f337689358c102ba2048d1b37f5 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:44:02 +0200 Subject: hfsplus: remove the rsrc_inodes list We never walk the list - the only reason for it is to make the resource fork inodes appear hashed to the writeback code. Borrow a trick from JFS to do that without needing a list head. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 5e2418a..0cd9ba0 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -145,8 +145,6 @@ struct hfsplus_sb_info { int part, session; unsigned long flags; - - struct hlist_head rsrc_inodes; }; #define HFSPLUS_SB_WRITEBACKUP 0x0001 diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index 309defb..a05b3af 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -204,7 +204,15 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, struct dentry *dent hip->rsrc_inode = dir; HFSPLUS_I(dir)->rsrc_inode = inode; igrab(dir); - hlist_add_head(&inode->i_hash, &HFSPLUS_SB(sb)->rsrc_inodes); + + /* + * __mark_inode_dirty expects inodes to be hashed. Since we don't + * want resource fork inodes in the regular inode space, we make them + * appear hashed, but do not put on any lists. hlist_del() + * will work fine and require no locking. + */ + inode->i_hash.pprev = &inode->i_hash.next; + mark_inode_dirty(inode); out: d_add(dentry, inode); diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index f310a1f..923f385 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -318,7 +318,6 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) return -ENOMEM; sb->s_fs_info = sbi; - INIT_HLIST_HEAD(&sbi->rsrc_inodes); mutex_init(&sbi->alloc_mutex); hfsplus_fill_defaults(sbi); if (!hfsplus_parse_options(data, sbi)) { -- cgit v0.10.2 From 7ac9fb9c2a50963b699b3548e6f00698c1554dc6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:45:08 +0200 Subject: hfsplus: add per-superblock lock for volume header updates Lock updates to the mutal fields in the volume header, and document the locing in the hfsplus_sb_info structure. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index f8ae468..1c81eed 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -251,6 +251,7 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, if (HFSPLUS_IS_RSRC(inode)) return -EPERM; + mutex_lock(&sbi->vh_mutex); if (inode->i_ino == (u32)(unsigned long)src_dentry->d_fsdata) { for (;;) { get_random_bytes(&id, sizeof(cnid)); @@ -263,7 +264,7 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, if (!res) break; if (res != -EEXIST) - return res; + goto out; } HFSPLUS_I(inode)->dev = id; cnid = sbi->next_cnid++; @@ -271,13 +272,13 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, res = hfsplus_create_cat(cnid, src_dir, &src_dentry->d_name, inode); if (res) /* panic? */ - return res; + goto out; sbi->file_count++; } cnid = sbi->next_cnid++; res = hfsplus_create_cat(cnid, dst_dir, &dst_dentry->d_name, inode); if (res) - return res; + goto out; inc_nlink(inode); hfsplus_instantiate(dst_dentry, inode, cnid); @@ -286,8 +287,9 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, mark_inode_dirty(inode); sbi->file_count++; dst_dir->i_sb->s_dirt = 1; - - return 0; +out: + mutex_unlock(&sbi->vh_mutex); + return res; } static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) @@ -302,6 +304,7 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) if (HFSPLUS_IS_RSRC(inode)) return -EPERM; + mutex_lock(&sbi->vh_mutex); cnid = (u32)(unsigned long)dentry->d_fsdata; if (inode->i_ino == cnid && atomic_read(&HFSPLUS_I(inode)->opencnt)) { @@ -312,11 +315,11 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) sbi->hidden_dir, &str); if (!res) inode->i_flags |= S_DEAD; - return res; + goto out; } res = hfsplus_delete_cat(cnid, dir, &dentry->d_name); if (res) - return res; + goto out; if (inode->i_nlink > 0) drop_nlink(inode); @@ -339,37 +342,44 @@ static int hfsplus_unlink(struct inode *dir, struct dentry *dentry) sbi->file_count--; inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); - +out: + mutex_unlock(&sbi->vh_mutex); return res; } static int hfsplus_rmdir(struct inode *dir, struct dentry *dentry) { - struct inode *inode; + struct hfsplus_sb_info *sbi = HFSPLUS_SB(dir->i_sb); + struct inode *inode = dentry->d_inode; int res; - inode = dentry->d_inode; if (inode->i_size != 2) return -ENOTEMPTY; + + mutex_lock(&sbi->vh_mutex); res = hfsplus_delete_cat(inode->i_ino, dir, &dentry->d_name); if (res) - return res; + goto out; clear_nlink(inode); inode->i_ctime = CURRENT_TIME_SEC; hfsplus_delete_inode(inode); mark_inode_dirty(inode); - return 0; +out: + mutex_unlock(&sbi->vh_mutex); + return res; } static int hfsplus_symlink(struct inode *dir, struct dentry *dentry, const char *symname) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(dir->i_sb); struct inode *inode; - int res; + int res = -ENOSPC; + mutex_lock(&sbi->vh_mutex); inode = hfsplus_new_inode(dir->i_sb, S_IFLNK | S_IRWXUGO); if (!inode) - return -ENOSPC; + goto out; res = page_symlink(inode, symname, strlen(symname) + 1); if (res) @@ -381,31 +391,35 @@ static int hfsplus_symlink(struct inode *dir, struct dentry *dentry, hfsplus_instantiate(dentry, inode, inode->i_ino); mark_inode_dirty(inode); - return 0; + goto out; out_err: inode->i_nlink = 0; hfsplus_delete_inode(inode); iput(inode); +out: + mutex_unlock(&sbi->vh_mutex); return res; } static int hfsplus_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev) { + struct hfsplus_sb_info *sbi = HFSPLUS_SB(dir->i_sb); struct inode *inode; - int res; + int res = -ENOSPC; + mutex_lock(&sbi->vh_mutex); inode = hfsplus_new_inode(dir->i_sb, mode); if (!inode) - return -ENOSPC; + goto out; res = hfsplus_create_cat(inode->i_ino, dir, &dentry->d_name, inode); if (res) { inode->i_nlink = 0; hfsplus_delete_inode(inode); iput(inode); - return res; + goto out; } if (S_ISBLK(mode) || S_ISCHR(mode) || S_ISFIFO(mode) || S_ISSOCK(mode)) @@ -413,7 +427,9 @@ static int hfsplus_mknod(struct inode *dir, struct dentry *dentry, hfsplus_instantiate(dentry, inode, inode->i_ino); mark_inode_dirty(inode); - return 0; +out: + mutex_unlock(&sbi->vh_mutex); + return res; } static int hfsplus_create(struct inode *dir, struct dentry *dentry, int mode, diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 0cd9ba0..08865ed 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -116,23 +116,26 @@ struct hfsplus_sb_info { struct inode *hidden_dir; struct nls_table *nls; - /* synchronize block allocations */ - struct mutex alloc_mutex; - /* Runtime variables */ u32 blockoffset; u32 sect_count; int fs_shift; - /* Stuff in host order from Vol Header */ + /* immutable data from the volume header */ u32 alloc_blksz; int alloc_blksz_shift; u32 total_blocks; + u32 data_clump_blocks, rsrc_clump_blocks; + + /* mutable data from the volume header, protected by alloc_mutex */ u32 free_blocks; + struct mutex alloc_mutex; + + /* mutable data from the volume header, protected by vh_mutex */ u32 next_cnid; u32 file_count; u32 folder_count; - u32 data_clump_blocks, rsrc_clump_blocks; + struct mutex vh_mutex; /* Config options */ u32 creator; diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 923f385..b766c17 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -160,6 +160,7 @@ int hfsplus_sync_fs(struct super_block *sb, int wait) dprint(DBG_SUPER, "hfsplus_write_super\n"); + mutex_lock(&sbi->vh_mutex); mutex_lock(&sbi->alloc_mutex); sb->s_dirt = 0; @@ -194,6 +195,7 @@ int hfsplus_sync_fs(struct super_block *sb, int wait) sbi->flags &= ~HFSPLUS_SB_WRITEBACKUP; } mutex_unlock(&sbi->alloc_mutex); + mutex_unlock(&sbi->vh_mutex); return 0; } @@ -319,6 +321,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) sb->s_fs_info = sbi; mutex_init(&sbi->alloc_mutex); + mutex_init(&sbi->vh_mutex); hfsplus_fill_defaults(sbi); if (!hfsplus_parse_options(data, sbi)) { printk(KERN_ERR "hfs: unable to parse mount options\n"); @@ -453,9 +456,13 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) if (!sbi->hidden_dir) { printk(KERN_DEBUG "hfs: create hidden dir...\n"); + + mutex_lock(&sbi->vh_mutex); sbi->hidden_dir = hfsplus_new_inode(sb, S_IFDIR); hfsplus_create_cat(sbi->hidden_dir->i_ino, sb->s_root->d_inode, &str, sbi->hidden_dir); + mutex_unlock(&sbi->vh_mutex); + mark_inode_dirty(sbi->hidden_dir); } out: -- cgit v0.10.2 From 84adede31267af37141da2b2b04293c5ea8af7ae Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:45:20 +0200 Subject: hfsplus: use atomic bitops for the superblock flags The flags in the HFS+-specific superlock do get modified during runtime, use atomic bitops to make the modifications SMP safe. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/btree.c b/fs/hfsplus/btree.c index f75cf22..27504f5 100644 --- a/fs/hfsplus/btree.c +++ b/fs/hfsplus/btree.c @@ -61,12 +61,12 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) if (id == HFSPLUS_EXT_CNID) { tree->keycmp = hfsplus_ext_cmp_key; } else if (id == HFSPLUS_CAT_CNID) { - if ((HFSPLUS_SB(sb)->flags & HFSPLUS_SB_HFSX) && + if (test_bit(HFSPLUS_SB_HFSX, &HFSPLUS_SB(sb)->flags) && (head->key_type == HFSPLUS_KEY_BINARY)) tree->keycmp = hfsplus_cat_bin_cmp_key; else { tree->keycmp = hfsplus_cat_case_cmp_key; - HFSPLUS_SB(sb)->flags |= HFSPLUS_SB_CASEFOLD; + set_bit(HFSPLUS_SB_CASEFOLD, &HFSPLUS_SB(sb)->flags); } } else { printk(KERN_ERR "hfs: unknown B*Tree requested\n"); diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 08865ed..c521d44 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -150,11 +150,11 @@ struct hfsplus_sb_info { unsigned long flags; }; -#define HFSPLUS_SB_WRITEBACKUP 0x0001 -#define HFSPLUS_SB_NODECOMPOSE 0x0002 -#define HFSPLUS_SB_FORCE 0x0004 -#define HFSPLUS_SB_HFSX 0x0008 -#define HFSPLUS_SB_CASEFOLD 0x0010 +#define HFSPLUS_SB_WRITEBACKUP 0 +#define HFSPLUS_SB_NODECOMPOSE 1 +#define HFSPLUS_SB_FORCE 2 +#define HFSPLUS_SB_HFSX 3 +#define HFSPLUS_SB_CASEFOLD 4 struct hfsplus_inode_info { diff --git a/fs/hfsplus/options.c b/fs/hfsplus/options.c index f2598ae..f9ab276 100644 --- a/fs/hfsplus/options.c +++ b/fs/hfsplus/options.c @@ -143,13 +143,13 @@ int hfsplus_parse_options(char *input, struct hfsplus_sb_info *sbi) kfree(p); break; case opt_decompose: - sbi->flags &= ~HFSPLUS_SB_NODECOMPOSE; + clear_bit(HFSPLUS_SB_NODECOMPOSE, &sbi->flags); break; case opt_nodecompose: - sbi->flags |= HFSPLUS_SB_NODECOMPOSE; + set_bit(HFSPLUS_SB_NODECOMPOSE, &sbi->flags); break; case opt_force: - sbi->flags |= HFSPLUS_SB_FORCE; + set_bit(HFSPLUS_SB_FORCE, &sbi->flags); break; default: return 0; @@ -184,7 +184,7 @@ int hfsplus_show_options(struct seq_file *seq, struct vfsmount *mnt) seq_printf(seq, ",session=%u", sbi->session); if (sbi->nls) seq_printf(seq, ",nls=%s", sbi->nls->charset); - if (sbi->flags & HFSPLUS_SB_NODECOMPOSE) + if (test_bit(HFSPLUS_SB_NODECOMPOSE, &sbi->flags)) seq_printf(seq, ",nodecompose"); return 0; } diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index b766c17..9a88d75 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -119,7 +119,7 @@ static int hfsplus_system_write_inode(struct inode *inode) } if (fork->total_size != cpu_to_be64(inode->i_size)) { - sbi->flags |= HFSPLUS_SB_WRITEBACKUP; + set_bit(HFSPLUS_SB_WRITEBACKUP, &sbi->flags); inode->i_sb->s_dirt = 1; } hfsplus_inode_write_fork(inode, fork); @@ -170,7 +170,7 @@ int hfsplus_sync_fs(struct super_block *sb, int wait) vhdr->file_count = cpu_to_be32(sbi->file_count); mark_buffer_dirty(sbi->s_vhbh); - if (sbi->flags & HFSPLUS_SB_WRITEBACKUP) { + if (test_and_clear_bit(HFSPLUS_SB_WRITEBACKUP, &sbi->flags)) { if (sbi->sect_count) { struct buffer_head *bh; u32 block, offset; @@ -192,7 +192,6 @@ int hfsplus_sync_fs(struct super_block *sb, int wait) printk(KERN_WARNING "hfs: backup not found!\n"); } } - sbi->flags &= ~HFSPLUS_SB_WRITEBACKUP; } mutex_unlock(&sbi->alloc_mutex); mutex_unlock(&sbi->vh_mutex); @@ -276,7 +275,7 @@ static int hfsplus_remount(struct super_block *sb, int *flags, char *data) "running fsck.hfsplus is recommended. leaving read-only.\n"); sb->s_flags |= MS_RDONLY; *flags |= MS_RDONLY; - } else if (sbi.flags & HFSPLUS_SB_FORCE) { + } else if (test_bit(HFSPLUS_SB_FORCE, &sbi.flags)) { /* nothing */ } else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_SOFTLOCK)) { printk(KERN_WARNING "hfs: filesystem is marked locked, leaving read-only.\n"); @@ -376,7 +375,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) printk(KERN_WARNING "hfs: Filesystem was not cleanly unmounted, " "running fsck.hfsplus is recommended. mounting read-only.\n"); sb->s_flags |= MS_RDONLY; - } else if (sbi->flags & HFSPLUS_SB_FORCE) { + } else if (test_and_clear_bit(HFSPLUS_SB_FORCE, &sbi->flags)) { /* nothing */ } else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_SOFTLOCK)) { printk(KERN_WARNING "hfs: Filesystem is marked locked, mounting read-only.\n"); @@ -386,7 +385,6 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) "use the force option at your own risk, mounting read-only.\n"); sb->s_flags |= MS_RDONLY; } - sbi->flags &= ~HFSPLUS_SB_FORCE; /* Load metadata objects (B*Trees) */ sbi->ext_tree = hfs_btree_open(sb, HFSPLUS_EXT_CNID); diff --git a/fs/hfsplus/unicode.c b/fs/hfsplus/unicode.c index a15fcff..b66d67d 100644 --- a/fs/hfsplus/unicode.c +++ b/fs/hfsplus/unicode.c @@ -132,7 +132,7 @@ int hfsplus_uni2asc(struct super_block *sb, const struct hfsplus_unistr *ustr, c ustrlen = be16_to_cpu(ustr->length); len = *len_p; ce1 = NULL; - compose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); + compose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags); while (ustrlen > 0) { c0 = be16_to_cpu(*ip++); @@ -293,7 +293,7 @@ int hfsplus_asc2uni(struct super_block *sb, struct hfsplus_unistr *ustr, u16 *dstr, outlen = 0; wchar_t c; - decompose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); + decompose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags); while (outlen < HFSPLUS_MAX_STRLEN && len > 0) { size = asc2unichar(sb, astr, len, &c); @@ -330,8 +330,8 @@ int hfsplus_hash_dentry(struct dentry *dentry, struct qstr *str) wchar_t c; u16 c2; - casefold = (HFSPLUS_SB(sb)->flags & HFSPLUS_SB_CASEFOLD); - decompose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); + casefold = test_bit(HFSPLUS_SB_CASEFOLD, &HFSPLUS_SB(sb)->flags); + decompose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags); hash = init_name_hash(); astr = str->name; len = str->len; @@ -373,8 +373,8 @@ int hfsplus_compare_dentry(struct dentry *dentry, struct qstr *s1, struct qstr * u16 c1, c2; wchar_t c; - casefold = (HFSPLUS_SB(sb)->flags & HFSPLUS_SB_CASEFOLD); - decompose = !(HFSPLUS_SB(sb)->flags & HFSPLUS_SB_NODECOMPOSE); + casefold = test_bit(HFSPLUS_SB_CASEFOLD, &HFSPLUS_SB(sb)->flags); + decompose = !test_bit(HFSPLUS_SB_NODECOMPOSE, &HFSPLUS_SB(sb)->flags); astr1 = s1->name; len1 = s1->len; astr2 = s2->name; diff --git a/fs/hfsplus/wrapper.c b/fs/hfsplus/wrapper.c index c89e198..8972c20 100644 --- a/fs/hfsplus/wrapper.c +++ b/fs/hfsplus/wrapper.c @@ -123,7 +123,7 @@ int hfsplus_read_wrapper(struct super_block *sb) if (vhdr->signature == cpu_to_be16(HFSPLUS_VOLHEAD_SIG)) break; if (vhdr->signature == cpu_to_be16(HFSPLUS_VOLHEAD_SIGX)) { - sbi->flags |= HFSPLUS_SB_HFSX; + set_bit(HFSPLUS_SB_HFSX, &sbi->flags); break; } brelse(bh); @@ -169,10 +169,14 @@ int hfsplus_read_wrapper(struct super_block *sb) return -EIO; /* should still be the same... */ - if (vhdr->signature != (sbi->flags & HFSPLUS_SB_HFSX ? - cpu_to_be16(HFSPLUS_VOLHEAD_SIGX) : - cpu_to_be16(HFSPLUS_VOLHEAD_SIG))) - goto error; + if (test_bit(HFSPLUS_SB_HFSX, &sbi->flags)) { + if (vhdr->signature != cpu_to_be16(HFSPLUS_VOLHEAD_SIGX)) + goto error; + } else { + if (vhdr->signature != cpu_to_be16(HFSPLUS_VOLHEAD_SIG)) + goto error; + } + sbi->s_vhbh = bh; sbi->s_vhdr = vhdr; -- cgit v0.10.2 From 89755dcace09b44b3aa024bf302d9b19b4c24cad Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:45:25 +0200 Subject: hfsplus: protect readdir against removals from open_dir_list We already have i_mutex for readdir and the namespace operations that add entries to open_dir_list, the only thing that was missing was the removal in hfsplus_dir_release. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 1c81eed..93fa45c 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -231,7 +231,9 @@ static int hfsplus_dir_release(struct inode *inode, struct file *file) { struct hfsplus_readdir_data *rd = file->private_data; if (rd) { + mutex_lock(&inode->i_mutex); list_del(&rd->list); + mutex_unlock(&inode->i_mutex); kfree(rd); } return 0; -- cgit v0.10.2 From 7fcc99f4f2ddb1c39abc05fbb9b32f05b03c7f8f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 05:46:31 +0200 Subject: hfsplus: add missing extent locking in hfsplus_write_inode Most of the extent handling code already does proper SMP locking, but hfsplus_write_inode was calling into hfsplus_ext_write_extent without taking the extents_lock. Fix this by splitting hfsplus_ext_write_extent into an internal helper that expects the lock, and a public interface that first acquires it. Also add a few locking asserts and document the locking rules in hfsplus_fs.h. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/extents.c b/fs/hfsplus/extents.c index b1017ea..0c9cb18 100644 --- a/fs/hfsplus/extents.c +++ b/fs/hfsplus/extents.c @@ -88,6 +88,8 @@ static void __hfsplus_ext_write_extent(struct inode *inode, struct hfs_find_data struct hfsplus_inode_info *hip = HFSPLUS_I(inode); int res; + WARN_ON(!mutex_is_locked(&hip->extents_lock)); + hfsplus_ext_build_key(fd->search_key, inode->i_ino, hip->cached_start, HFSPLUS_IS_RSRC(inode) ? HFSPLUS_TYPE_RSRC : HFSPLUS_TYPE_DATA); @@ -108,7 +110,7 @@ static void __hfsplus_ext_write_extent(struct inode *inode, struct hfs_find_data } } -void hfsplus_ext_write_extent(struct inode *inode) +static void hfsplus_ext_write_extent_locked(struct inode *inode) { if (HFSPLUS_I(inode)->flags & HFSPLUS_FLG_EXT_DIRTY) { struct hfs_find_data fd; @@ -119,6 +121,13 @@ void hfsplus_ext_write_extent(struct inode *inode) } } +void hfsplus_ext_write_extent(struct inode *inode) +{ + mutex_lock(&HFSPLUS_I(inode)->extents_lock); + hfsplus_ext_write_extent_locked(inode); + mutex_unlock(&HFSPLUS_I(inode)->extents_lock); +} + static inline int __hfsplus_ext_read_extent(struct hfs_find_data *fd, struct hfsplus_extent *extent, u32 cnid, u32 block, u8 type) @@ -144,6 +153,8 @@ static inline int __hfsplus_ext_cache_extent(struct hfs_find_data *fd, struct in struct hfsplus_inode_info *hip = HFSPLUS_I(inode); int res; + WARN_ON(!mutex_is_locked(&hip->extents_lock)); + if (hip->flags & HFSPLUS_FLG_EXT_DIRTY) __hfsplus_ext_write_extent(inode, fd); @@ -433,7 +444,7 @@ out: insert_extent: dprint(DBG_EXTENT, "insert new extent\n"); - hfsplus_ext_write_extent(inode); + hfsplus_ext_write_extent_locked(inode); memset(hip->cached_extents, 0, sizeof(hfsplus_extent_rec)); hip->cached_extents[0].start_block = cpu_to_be32(start); diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index c521d44..c007cc2 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -158,28 +158,36 @@ struct hfsplus_sb_info { struct hfsplus_inode_info { - struct mutex extents_lock; - u32 clump_blocks, alloc_blocks; - sector_t fs_blocks; - /* Allocation extents from catalog record or volume header */ - hfsplus_extent_rec first_extents; - u32 first_blocks; - hfsplus_extent_rec cached_extents; - u32 cached_start, cached_blocks; atomic_t opencnt; - struct inode *rsrc_inode; + /* + * Extent allocation information, protected by extents_lock. + */ + u32 first_blocks; + u32 clump_blocks; + u32 alloc_blocks; + u32 cached_start; + u32 cached_blocks; + hfsplus_extent_rec first_extents; + hfsplus_extent_rec cached_extents; unsigned long flags; + struct mutex extents_lock; + /* + * Immutable data. + */ + struct inode *rsrc_inode; __be32 create_date; - /* Device number in hfsplus_permissions in catalog */ u32 dev; - /* BSD system and user file flags */ - u8 rootflags; - u8 userflags; + /* + * Protected by i_mutex. + */ + sector_t fs_blocks; + u8 rootflags, userflags; /* BSD system and user file flags */ struct list_head open_dir_list; loff_t phys_size; + struct inode vfs_inode; }; -- cgit v0.10.2 From 467c3d9cd541eef284ff8118069b088e015b8d6a Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 1 Oct 2010 05:46:52 +0200 Subject: hfsplus: convert tree_lock to mutex tree_lock is used as mutex so make it a mutex. Signed-off-by: Thomas Gleixner Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/bfind.c b/fs/hfsplus/bfind.c index 5007a41..68c7983 100644 --- a/fs/hfsplus/bfind.c +++ b/fs/hfsplus/bfind.c @@ -23,7 +23,7 @@ int hfs_find_init(struct hfs_btree *tree, struct hfs_find_data *fd) fd->search_key = ptr; fd->key = ptr + tree->max_key_len + 2; dprint(DBG_BNODE_REFS, "find_init: %d (%p)\n", tree->cnid, __builtin_return_address(0)); - down(&tree->tree_lock); + mutex_lock(&tree->tree_lock); return 0; } @@ -32,7 +32,7 @@ void hfs_find_exit(struct hfs_find_data *fd) hfs_bnode_put(fd->bnode); kfree(fd->search_key); dprint(DBG_BNODE_REFS, "find_exit: %d (%p)\n", fd->tree->cnid, __builtin_return_address(0)); - up(&fd->tree->tree_lock); + mutex_unlock(&fd->tree->tree_lock); fd->tree = NULL; } diff --git a/fs/hfsplus/btree.c b/fs/hfsplus/btree.c index 27504f5..8306479 100644 --- a/fs/hfsplus/btree.c +++ b/fs/hfsplus/btree.c @@ -30,7 +30,7 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) if (!tree) return NULL; - init_MUTEX(&tree->tree_lock); + mutex_init(&tree->tree_lock); spin_lock_init(&tree->hash_lock); tree->sb = sb; tree->cnid = id; diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index c007cc2..5cda963 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -62,7 +62,7 @@ struct hfs_btree { unsigned int depth; //unsigned int map1_size, map_size; - struct semaphore tree_lock; + struct mutex tree_lock; unsigned int pages_per_bnode; spinlock_t hash_lock; -- cgit v0.10.2 From c1e0ddbf0a97e1704d7f13b4934f9acca002402d Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 17 Sep 2010 23:26:24 -0700 Subject: ACPI: Handle ACPI0007 Device in acpi_early_set_pdc After | commit d8191fa4a33fdc817277da4f2b7f771ff605a41c | Author: Alex Chiang | Date: Mon Feb 22 12:11:39 2010 -0700 | | ACPI: processor: driver doesn't need to evaluate _PDC | | Now that the early _PDC evaluation path knows how to correctly | evaluate _PDC on only physically present processors, there's no | need for the processor driver to evaluate it later when it loads. | | To cover the hotplug case, push _PDC evaluation down into the | hotplug paths. only cpu with Processor Statement get processed with _PDC If bios is using Device object instead of Processor statement. SSDTs for Pstate/Cstate/Tstate can not be loaded dynamically. Need to try to scan ACPI0007 in addition to Processor. That commit is between 2.6.34-rc1 and 2.6.34-rc2, so stable tree for 2.6.34+ need this patch. Signed-off-by: Yinghai Lu Reviewed-by: Bjorn Helgaas Signed-off-by: Len Brown diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c index e9699aae..df6e167 100644 --- a/drivers/acpi/processor_core.c +++ b/drivers/acpi/processor_core.c @@ -352,4 +352,5 @@ void __init acpi_early_processor_set_pdc(void) acpi_walk_namespace(ACPI_TYPE_PROCESSOR, ACPI_ROOT_OBJECT, ACPI_UINT32_MAX, early_init_pdc, NULL, NULL, NULL); + acpi_get_devices("ACPI0007", early_init_pdc, NULL, NULL); } -- cgit v0.10.2 From 40de9a7cebc4e0b23cd6863c84c2279f0ccadebb Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 1 Oct 2010 09:12:08 +0200 Subject: hfsplus: fix rename over directories When renaming over a directory we need to use hfsplus_rmdir instead of hfsplus_unlink to evict the victim. This makes sure we properly error out on non-empty directory as required by Posix (BZ #16571), and it also makes sure we do the right thing in case i_nlink will every be set correctly for directories on hfsplus. Reported-by: Vlado Plaga Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 93fa45c..33aab21 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -452,7 +452,10 @@ static int hfsplus_rename(struct inode *old_dir, struct dentry *old_dentry, /* Unlink destination if it already exists */ if (new_dentry->d_inode) { - res = hfsplus_unlink(new_dir, new_dentry); + if (S_ISDIR(new_dentry->d_inode->i_mode)) + res = hfsplus_rmdir(new_dir, new_dentry); + else + res = hfsplus_unlink(new_dir, new_dentry); if (res) return res; } -- cgit v0.10.2 From 3fdbf004c1706480a7c7fac3c9d836fa6df20d7d Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Thu, 30 Sep 2010 14:32:35 +0200 Subject: x86, mtrr: Assume SYS_CFG[Tom2ForceMemTypeWB] exists on all future AMD CPUs Instead of adapting the CPU family check in amd_special_default_mtrr() for each new CPU family assume that all new AMD CPUs support the necessary bits in SYS_CFG MSR. Tom2Enabled is architectural (defined in APM Vol.2). Tom2ForceMemTypeWB is defined in all BKDGs starting with K8 NPT. In pre K8-NPT BKDG this bit is reserved (read as zero). W/o this adaption Linux would unnecessarily complain about bad MTRR settings on every new AMD CPU family, e.g. [ 0.000000] WARNING: BIOS bug: CPU MTRRs don't cover all of memory, losing 4863MB of RAM. Cc: stable@kernel.org # .32.x, .35.x Signed-off-by: Andreas Herrmann LKML-Reference: <20100930123235.GB20545@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/mtrr/cleanup.c b/arch/x86/kernel/cpu/mtrr/cleanup.c index c5f59d0..ac140c7 100644 --- a/arch/x86/kernel/cpu/mtrr/cleanup.c +++ b/arch/x86/kernel/cpu/mtrr/cleanup.c @@ -827,7 +827,7 @@ int __init amd_special_default_mtrr(void) if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) return 0; - if (boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) + if (boot_cpu_data.x86 < 0xf) return 0; /* In case some hypervisor doesn't pass SYSCFG through: */ if (rdmsr_safe(MSR_K8_SYSCFG, &l, &h) < 0) -- cgit v0.10.2 From 420b13b60a3e5c5dcc6ec290e131cf5fbc603d94 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Thu, 30 Sep 2010 14:33:58 +0200 Subject: x86, nmi: Support NMI watchdog on newer AMD CPU families CPU families 0x12, 0x14 and 0x15 support this functionality. Signed-off-by: Andreas Herrmann LKML-Reference: <20100930123357.GC20545@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c index fb329e9..d9f4ff8 100644 --- a/arch/x86/kernel/cpu/perfctr-watchdog.c +++ b/arch/x86/kernel/cpu/perfctr-watchdog.c @@ -700,11 +700,10 @@ static void probe_nmi_watchdog(void) { switch (boot_cpu_data.x86_vendor) { case X86_VENDOR_AMD: - if (boot_cpu_data.x86 != 6 && boot_cpu_data.x86 != 15 && - boot_cpu_data.x86 != 16 && boot_cpu_data.x86 != 17) - return; - wd_ops = &k7_wd_ops; - break; + if (boot_cpu_data.x86 == 6 || + (boot_cpu_data.x86 >= 0xf && boot_cpu_data.x86 <= 0x15)) + wd_ops = &k7_wd_ops; + return; case X86_VENDOR_INTEL: /* Work around where perfctr1 doesn't have a working enable * bit as described in the following errata: -- cgit v0.10.2 From 23588c38a84c9175c6668789b64ffba4651e5c6a Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Thu, 30 Sep 2010 14:36:28 +0200 Subject: x86, amd: Add support for CPUID topology extension of AMD CPUs Node information (ID, number of internal nodes) is provided via CPUID Fn8000_001e_ECX. See AMD CPUID Specification (Publication # 25481, Revision 2.34, September 2010). Signed-off-by: Andreas Herrmann LKML-Reference: <20100930123628.GD20545@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 0f0ace5d..7e6a37d 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -253,37 +253,41 @@ static int __cpuinit nearby_node(int apicid) #endif /* - * Fixup core topology information for AMD multi-node processors. - * Assumption: Number of cores in each internal node is the same. + * Fixup core topology information for + * (1) AMD multi-node processors + * Assumption: Number of cores in each internal node is the same. */ #ifdef CONFIG_X86_HT -static void __cpuinit amd_fixup_dcm(struct cpuinfo_x86 *c) +static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c) { - unsigned long long value; u32 nodes, cores_per_node; + u8 node_id; + unsigned long long value; int cpu = smp_processor_id(); - if (!cpu_has(c, X86_FEATURE_NODEID_MSR)) - return; - - /* fixup topology information only once for a core */ - if (cpu_has(c, X86_FEATURE_AMD_DCM)) + /* get information required for multi-node processors */ + if (cpu_has(c, X86_FEATURE_TOPOEXT)) { + value = cpuid_ecx(0x8000001e); + nodes = ((value >> 8) & 7) + 1; + node_id = value & 7; + } else if (cpu_has(c, X86_FEATURE_NODEID_MSR)) { + rdmsrl(MSR_FAM10H_NODE_ID, value); + nodes = ((value >> 3) & 7) + 1; + node_id = value & 7; + } else return; - rdmsrl(MSR_FAM10H_NODE_ID, value); + /* fixup multi-node processor information */ + if (nodes > 1) { + set_cpu_cap(c, X86_FEATURE_AMD_DCM); + cores_per_node = c->x86_max_cores / nodes; - nodes = ((value >> 3) & 7) + 1; - if (nodes == 1) - return; - - set_cpu_cap(c, X86_FEATURE_AMD_DCM); - cores_per_node = c->x86_max_cores / nodes; + /* store NodeID, use llc_shared_map to store sibling info */ + per_cpu(cpu_llc_id, cpu) = node_id; - /* store NodeID, use llc_shared_map to store sibling info */ - per_cpu(cpu_llc_id, cpu) = value & 7; - - /* fixup core id to be in range from 0 to (cores_per_node - 1) */ - c->cpu_core_id = c->cpu_core_id % cores_per_node; + /* core id to be in range from 0 to (cores_per_node - 1) */ + c->cpu_core_id = c->cpu_core_id % cores_per_node; + } } #endif @@ -304,9 +308,7 @@ static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c) c->phys_proc_id = c->initial_apicid >> bits; /* use socket ID also for last level cache */ per_cpu(cpu_llc_id, cpu) = c->phys_proc_id; - /* fixup topology information on multi-node processors */ - if ((c->x86 == 0x10) && (c->x86_model == 9)) - amd_fixup_dcm(c); + amd_get_topology(c); #endif } -- cgit v0.10.2 From 6057b4d331f19a3ea51aec463ea7839c128b3227 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Thu, 30 Sep 2010 14:38:57 +0200 Subject: x86, amd: Extract compute unit information for AMD CPUs Get compute unit information from CPUID Fn8000_001E_EBX. (See AMD CPUID Specification - publication # 25481, revision 2.34, September 2010.) Note that each core on a compute unit still has a core_id of its own. Signed-off-by: Andreas Herrmann LKML-Reference: <20100930123857.GE20545@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index 325b7bd..69e80c2 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -110,6 +110,8 @@ struct cpuinfo_x86 { u16 phys_proc_id; /* Core id: */ u16 cpu_core_id; + /* Compute unit id */ + u8 compute_unit_id; /* Index into per_cpu list: */ u16 cpu_index; #endif diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 7e6a37d..70168ab 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -256,21 +256,29 @@ static int __cpuinit nearby_node(int apicid) * Fixup core topology information for * (1) AMD multi-node processors * Assumption: Number of cores in each internal node is the same. + * (2) AMD processors supporting compute units */ #ifdef CONFIG_X86_HT static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c) { - u32 nodes, cores_per_node; + u32 nodes; u8 node_id; - unsigned long long value; int cpu = smp_processor_id(); /* get information required for multi-node processors */ if (cpu_has(c, X86_FEATURE_TOPOEXT)) { - value = cpuid_ecx(0x8000001e); - nodes = ((value >> 8) & 7) + 1; - node_id = value & 7; + u32 eax, ebx, ecx, edx; + + cpuid(0x8000001e, &eax, &ebx, &ecx, &edx); + nodes = ((ecx >> 8) & 7) + 1; + node_id = ecx & 7; + + /* get compute unit information */ + smp_num_siblings = ((ebx >> 8) & 3) + 1; + c->compute_unit_id = ebx & 0xff; } else if (cpu_has(c, X86_FEATURE_NODEID_MSR)) { + u64 value; + rdmsrl(MSR_FAM10H_NODE_ID, value); nodes = ((value >> 3) & 7) + 1; node_id = value & 7; @@ -279,6 +287,8 @@ static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c) /* fixup multi-node processor information */ if (nodes > 1) { + u32 cores_per_node; + set_cpu_cap(c, X86_FEATURE_AMD_DCM); cores_per_node = c->x86_max_cores / nodes; -- cgit v0.10.2 From d4fbe4f03557e1fd4d9bbb3a1957aad560f39e96 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Thu, 30 Sep 2010 14:41:56 +0200 Subject: x86, amd: Use compute unit information to determine thread siblings This information is vital for different load balancing policies. Signed-off-by: Andreas Herrmann LKML-Reference: <20100930124156.GF20545@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 8b3bfc4..bc2cc44 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -397,6 +397,19 @@ void __cpuinit smp_store_cpu_info(int id) identify_secondary_cpu(c); } +static void __cpuinit link_thread_siblings(int cpu1, int cpu2) +{ + struct cpuinfo_x86 *c1 = &cpu_data(cpu1); + struct cpuinfo_x86 *c2 = &cpu_data(cpu2); + + cpumask_set_cpu(cpu1, cpu_sibling_mask(cpu2)); + cpumask_set_cpu(cpu2, cpu_sibling_mask(cpu1)); + cpumask_set_cpu(cpu1, cpu_core_mask(cpu2)); + cpumask_set_cpu(cpu2, cpu_core_mask(cpu1)); + cpumask_set_cpu(cpu1, c2->llc_shared_map); + cpumask_set_cpu(cpu2, c1->llc_shared_map); +} + void __cpuinit set_cpu_sibling_map(int cpu) { @@ -409,14 +422,13 @@ void __cpuinit set_cpu_sibling_map(int cpu) for_each_cpu(i, cpu_sibling_setup_mask) { struct cpuinfo_x86 *o = &cpu_data(i); - if (c->phys_proc_id == o->phys_proc_id && - c->cpu_core_id == o->cpu_core_id) { - cpumask_set_cpu(i, cpu_sibling_mask(cpu)); - cpumask_set_cpu(cpu, cpu_sibling_mask(i)); - cpumask_set_cpu(i, cpu_core_mask(cpu)); - cpumask_set_cpu(cpu, cpu_core_mask(i)); - cpumask_set_cpu(i, c->llc_shared_map); - cpumask_set_cpu(cpu, o->llc_shared_map); + if (cpu_has(c, X86_FEATURE_TOPOEXT)) { + if (c->phys_proc_id == o->phys_proc_id && + c->compute_unit_id == o->compute_unit_id) + link_thread_siblings(cpu, i); + } else if (c->phys_proc_id == o->phys_proc_id && + c->cpu_core_id == o->cpu_core_id) { + link_thread_siblings(cpu, i); } } } else { -- cgit v0.10.2 From 5c80cc78de46aef6cd5e714208da05c3f7f548f8 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Thu, 30 Sep 2010 14:43:16 +0200 Subject: x86, amd_nb: Enable GART support for AMD family 0x15 CPUs AMD CPU family 0x15 still supports GART for compatibility reasons. Signed-off-by: Andreas Herrmann LKML-Reference: <20100930124316.GG20545@loge.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/amd_nb.c b/arch/x86/kernel/amd_nb.c index 4ffc38d..8f6463d 100644 --- a/arch/x86/kernel/amd_nb.c +++ b/arch/x86/kernel/amd_nb.c @@ -15,6 +15,7 @@ static u32 *flush_words; struct pci_device_id k8_nb_ids[] = { { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_K8_NB_MISC) }, { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_10H_NB_MISC) }, + { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_15H_NB_MISC) }, {} }; EXPORT_SYMBOL(k8_nb_ids); @@ -45,7 +46,8 @@ int cache_k8_northbridges(void) k8_northbridges.num++; /* some CPU families (e.g. family 0x11) do not support GART */ - if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10) + if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10 || + boot_cpu_data.x86 == 0x15) k8_northbridges.gart_supported = 1; k8_northbridges.nb_misc = kmalloc((k8_northbridges.num + 1) * diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index 10d3330..edc0279 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -514,6 +514,7 @@ #define PCI_DEVICE_ID_AMD_11H_NB_DRAM 0x1302 #define PCI_DEVICE_ID_AMD_11H_NB_MISC 0x1303 #define PCI_DEVICE_ID_AMD_11H_NB_LINK 0x1304 +#define PCI_DEVICE_ID_AMD_15H_NB_MISC 0x1603 #define PCI_DEVICE_ID_AMD_LANCE 0x2000 #define PCI_DEVICE_ID_AMD_LANCE_HOME 0x2001 #define PCI_DEVICE_ID_AMD_SCSI 0x2020 -- cgit v0.10.2 From 3deb11ef16632fb76daead0db281f6f7d19332f2 Mon Sep 17 00:00:00 2001 From: "Lee, Chun-Yi" Date: Fri, 1 Oct 2010 16:28:29 +0800 Subject: ACPI: add DMI to disable AML Vista compatibility on MSI GX723 Notebook The brightness control hotkey don't work with Vista compatibility because the MSI GX723 includes an infinite while loop in DSDT when brightness control hotkey pressed. The MSI GX723 uses Nvidia video. Perhaps the loop is specific to the Nvidia Vista driver... This patch should be reverted once nouveau grows support to call the ACPI NVIF method. Signed-off-by: Lee, Chun-Yi Signed-off-by: Len Brown diff --git a/drivers/acpi/blacklist.c b/drivers/acpi/blacklist.c index 2bb28b9..47cb4cb 100644 --- a/drivers/acpi/blacklist.c +++ b/drivers/acpi/blacklist.c @@ -202,6 +202,23 @@ static struct dmi_system_id acpi_osi_dmi_table[] __initdata = { }, }, { + /* + * There have a NVIF method in MSI GX723 DSDT need call by Nvidia + * driver (e.g. nouveau) when user press brightness hotkey. + * Currently, nouveau driver didn't do the job and it causes there + * have a infinite while loop in DSDT when user press hotkey. + * We add MSI GX723's dmi information to this table for workaround + * this issue. + * Will remove MSI GX723 from the table after nouveau grows support. + */ + .callback = dmi_disable_osi_vista, + .ident = "MSI GX723", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Micro-Star International"), + DMI_MATCH(DMI_PRODUCT_NAME, "GX723"), + }, + }, + { .callback = dmi_disable_osi_vista, .ident = "Sony VGN-NS10J_S", .matches = { -- cgit v0.10.2 From ae878ae280bea286ff2b1e1cb6e609dd8cb4501d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Maciej=20=C5=BBenczykowski?= Date: Sun, 3 Oct 2010 14:49:00 -0700 Subject: net: Fix IPv6 PMTU disc. w/ asymmetric routes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Maciej Å»enczykowski Signed-off-by: David S. Miller diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 8323136..a275c6e 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -1556,14 +1556,13 @@ out: * i.e. Path MTU discovery */ -void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr, - struct net_device *dev, u32 pmtu) +static void rt6_do_pmtu_disc(struct in6_addr *daddr, struct in6_addr *saddr, + struct net *net, u32 pmtu, int ifindex) { struct rt6_info *rt, *nrt; - struct net *net = dev_net(dev); int allfrag = 0; - rt = rt6_lookup(net, daddr, saddr, dev->ifindex, 0); + rt = rt6_lookup(net, daddr, saddr, ifindex, 0); if (rt == NULL) return; @@ -1631,6 +1630,27 @@ out: dst_release(&rt->dst); } +void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr, + struct net_device *dev, u32 pmtu) +{ + struct net *net = dev_net(dev); + + /* + * RFC 1981 states that a node "MUST reduce the size of the packets it + * is sending along the path" that caused the Packet Too Big message. + * Since it's not possible in the general case to determine which + * interface was used to send the original packet, we update the MTU + * on the interface that will be used to send future packets. We also + * update the MTU on the interface that received the Packet Too Big in + * case the original packet was forced out that interface with + * SO_BINDTODEVICE or similar. This is the next best thing to the + * correct behaviour, which would be to update the MTU on all + * interfaces. + */ + rt6_do_pmtu_disc(daddr, saddr, net, pmtu, 0); + rt6_do_pmtu_disc(daddr, saddr, net, pmtu, dev->ifindex); +} + /* * Misc support functions */ -- cgit v0.10.2 From 482964e56e1320cb7952faa1932d8ecf59c4bf75 Mon Sep 17 00:00:00 2001 From: Nagendra Tomar Date: Sat, 2 Oct 2010 23:45:06 +0000 Subject: net: Fix the condition passed to sk_wait_event() This patch fixes the condition (3rd arg) passed to sk_wait_event() in sk_stream_wait_memory(). The incorrect check in sk_stream_wait_memory() causes the following soft lockup in tcp_sendmsg() when the global tcp memory pool has exhausted. >>> snip <<< localhost kernel: BUG: soft lockup - CPU#3 stuck for 11s! [sshd:6429] localhost kernel: CPU 3: localhost kernel: RIP: 0010:[sk_stream_wait_memory+0xcd/0x200] [sk_stream_wait_memory+0xcd/0x200] sk_stream_wait_memory+0xcd/0x200 localhost kernel: localhost kernel: Call Trace: localhost kernel: [sk_stream_wait_memory+0x1b1/0x200] sk_stream_wait_memory+0x1b1/0x200 localhost kernel: [] autoremove_wake_function+0x0/0x40 localhost kernel: [ipv6:tcp_sendmsg+0x6e6/0xe90] tcp_sendmsg+0x6e6/0xce0 localhost kernel: [sock_aio_write+0x126/0x140] sock_aio_write+0x126/0x140 localhost kernel: [xfs:do_sync_write+0xf1/0x130] do_sync_write+0xf1/0x130 localhost kernel: [] autoremove_wake_function+0x0/0x40 localhost kernel: [hrtimer_start+0xe3/0x170] hrtimer_start+0xe3/0x170 localhost kernel: [vfs_write+0x185/0x190] vfs_write+0x185/0x190 localhost kernel: [sys_write+0x50/0x90] sys_write+0x50/0x90 localhost kernel: [system_call+0x7e/0x83] system_call+0x7e/0x83 >>> snip <<< What is happening is, that the sk_wait_event() condition passed from sk_stream_wait_memory() evaluates to true for the case of tcp global memory exhaustion. This is because both sk_stream_memory_free() and vm_wait are true which causes sk_wait_event() to *not* call schedule_timeout(). Hence sk_stream_wait_memory() returns immediately to the caller w/o sleeping. This causes the caller to again try allocation, which again fails and again calls sk_stream_wait_memory(), and so on. [ Bug introduced by commit c1cbe4b7ad0bc4b1d98ea708a3fecb7362aa4088 ("[NET]: Avoid atomic xchg() for non-error case") -DaveM ] Signed-off-by: Nagendra Singh Tomar Signed-off-by: David S. Miller diff --git a/net/core/stream.c b/net/core/stream.c index d959e0f..f5df85d 100644 --- a/net/core/stream.c +++ b/net/core/stream.c @@ -141,10 +141,10 @@ int sk_stream_wait_memory(struct sock *sk, long *timeo_p) set_bit(SOCK_NOSPACE, &sk->sk_socket->flags); sk->sk_write_pending++; - sk_wait_event(sk, ¤t_timeo, !sk->sk_err && - !(sk->sk_shutdown & SEND_SHUTDOWN) && - sk_stream_memory_free(sk) && - vm_wait); + sk_wait_event(sk, ¤t_timeo, sk->sk_err || + (sk->sk_shutdown & SEND_SHUTDOWN) || + (sk_stream_memory_free(sk) && + !vm_wait)); sk->sk_write_pending--; if (vm_wait) { -- cgit v0.10.2 From c5d3557103f8bef81d7a150ab9cc970099cd58a2 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Sun, 3 Oct 2010 15:37:42 +0000 Subject: Revert "ipv4: Make INET_LRO a bool instead of tristate." This reverts commit e81963b180ac502fda0326edf059b1e29cdef1a2. LRO is now deprecated in favour of GRO, and only a few drivers use it, so it is desirable to build it as a module in distribution kernels. The original change to prevent building it as a module was made in an attempt to avoid the case where some dependents are set to y and some to m, and INET_LRO can be set to m rather than y. However, the Kconfig system will reliably set INET_LRO=y in this case. Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig index 72380a3..7cd7760 100644 --- a/net/ipv4/Kconfig +++ b/net/ipv4/Kconfig @@ -413,7 +413,7 @@ config INET_XFRM_MODE_BEET If unsure, say Y. config INET_LRO - bool "Large Receive Offload (ipv4/tcp)" + tristate "Large Receive Offload (ipv4/tcp)" default y ---help--- Support for Large Receive Offload (ipv4/tcp). -- cgit v0.10.2 From 10ccff62bd3de7a64cf98f4c37ec0414b8affd4f Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Sun, 3 Oct 2010 15:42:05 +0000 Subject: netdev: Depend on INET before selecting INET_LRO Since 'select' ignores dependencies, drivers that select INET_LRO must depend on INET. This fixes the broken configuration reported in . Reported-by: Subrata Modak Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig index 2cc81a5..5db667c 100644 --- a/drivers/net/Kconfig +++ b/drivers/net/Kconfig @@ -2428,7 +2428,7 @@ config UGETH_TX_ON_DEMAND config MV643XX_ETH tristate "Marvell Discovery (643XX) and Orion ethernet support" - depends on MV64X60 || PPC32 || PLAT_ORION + depends on (MV64X60 || PPC32 || PLAT_ORION) && INET select INET_LRO select PHYLIB help @@ -2803,7 +2803,7 @@ config NIU config PASEMI_MAC tristate "PA Semi 1/10Gbit MAC" - depends on PPC_PASEMI && PCI + depends on PPC_PASEMI && PCI && INET select PHYLIB select INET_LRO help -- cgit v0.10.2 From 5b7c84066733c5dfb0e4016d939757b38de189e4 Mon Sep 17 00:00:00 2001 From: David Stevens Date: Thu, 30 Sep 2010 14:29:40 +0000 Subject: ipv4: correct IGMP behavior on v3 query during v2-compatibility mode A recent patch to allow IGMPv2 responses to IGMPv3 queries bypasses length checks for valid query lengths, incorrectly resets the v2_seen timer, and does not support IGMPv1. The following patch responds with a v2 report as required by IGMPv2 while correcting the other problems introduced by the patch. Signed-Off-By: David L Stevens Signed-off-by: David S. Miller diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index 1fdcacd..2a4bb76 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -834,7 +834,7 @@ static void igmp_heard_query(struct in_device *in_dev, struct sk_buff *skb, int mark = 0; - if (len == 8 || IGMP_V2_SEEN(in_dev)) { + if (len == 8) { if (ih->code == 0) { /* Alas, old v1 router presents here. */ @@ -856,6 +856,18 @@ static void igmp_heard_query(struct in_device *in_dev, struct sk_buff *skb, igmpv3_clear_delrec(in_dev); } else if (len < 12) { return; /* ignore bogus packet; freed by caller */ + } else if (IGMP_V1_SEEN(in_dev)) { + /* This is a v3 query with v1 queriers present */ + max_delay = IGMP_Query_Response_Interval; + group = 0; + } else if (IGMP_V2_SEEN(in_dev)) { + /* this is a v3 query with v2 queriers present; + * Interpretation of the max_delay code is problematic here. + * A real v2 host would use ih_code directly, while v3 has a + * different encoding. We use the v3 encoding as more likely + * to be intended in a v3 query. + */ + max_delay = IGMPV3_MRC(ih3->code)*(HZ/IGMP_TIMER_SCALE); } else { /* v3 */ if (!pskb_may_pull(skb, sizeof(struct igmpv3_query))) return; -- cgit v0.10.2 From d7e0d19aa0fdd22819d35db551bd54c1bcf9c2aa Mon Sep 17 00:00:00 2001 From: Dan Rosenberg Date: Fri, 1 Oct 2010 11:16:58 +0000 Subject: sctp: prevent reading out-of-bounds memory Two user-controlled allocations in SCTP are subsequently dereferenced as sockaddr structs, without checking if the dereferenced struct members fall beyond the end of the allocated chunk. There doesn't appear to be any information leakage here based on how these members are used and additional checking, but it's still worth fixing. [akpm@linux-foundation.org: remove unfashionable newlines, fix gmail tab->space conversion] Signed-off-by: Dan Rosenberg Acked-by: Vlad Yasevich Cc: David Miller Signed-off-by: Andrew Morton Signed-off-by: David S. Miller diff --git a/net/sctp/socket.c b/net/sctp/socket.c index ca44917..fbb7077 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -916,6 +916,11 @@ SCTP_STATIC int sctp_setsockopt_bindx(struct sock* sk, /* Walk through the addrs buffer and count the number of addresses. */ addr_buf = kaddrs; while (walk_size < addrs_size) { + if (walk_size + sizeof(sa_family_t) > addrs_size) { + kfree(kaddrs); + return -EINVAL; + } + sa_addr = (struct sockaddr *)addr_buf; af = sctp_get_af_specific(sa_addr->sa_family); @@ -1002,9 +1007,13 @@ static int __sctp_connect(struct sock* sk, /* Walk through the addrs buffer and count the number of addresses. */ addr_buf = kaddrs; while (walk_size < addrs_size) { + if (walk_size + sizeof(sa_family_t) > addrs_size) { + err = -EINVAL; + goto out_free; + } + sa_addr = (union sctp_addr *)addr_buf; af = sctp_get_af_specific(sa_addr->sa.sa_family); - port = ntohs(sa_addr->v4.sin_port); /* If the address family is not supported or if this address * causes the address buffer to overflow return EINVAL. @@ -1014,6 +1023,8 @@ static int __sctp_connect(struct sock* sk, goto out_free; } + port = ntohs(sa_addr->v4.sin_port); + /* Save current address so we can work with it */ memcpy(&to, sa_addr, af->sockaddr_len); -- cgit v0.10.2 From 51e97a12bef19b7e43199fc153cf9bd5f2140362 Mon Sep 17 00:00:00 2001 From: Dan Rosenberg Date: Fri, 1 Oct 2010 11:51:47 +0000 Subject: sctp: Fix out-of-bounds reading in sctp_asoc_get_hmac() The sctp_asoc_get_hmac() function iterates through a peer's hmac_ids array and attempts to ensure that only a supported hmac entry is returned. The current code fails to do this properly - if the last id in the array is out of range (greater than SCTP_AUTH_HMAC_ID_MAX), the id integer remains set after exiting the loop, and the address of an out-of-bounds entry will be returned and subsequently used in the parent function, causing potentially ugly memory corruption. This patch resets the id integer to 0 on encountering an invalid id so that NULL will be returned after finishing the loop if no valid ids are found. Signed-off-by: Dan Rosenberg Acked-by: Vlad Yasevich Signed-off-by: David S. Miller diff --git a/net/sctp/auth.c b/net/sctp/auth.c index 8636639..ddbbf7c 100644 --- a/net/sctp/auth.c +++ b/net/sctp/auth.c @@ -543,16 +543,20 @@ struct sctp_hmac *sctp_auth_asoc_get_hmac(const struct sctp_association *asoc) id = ntohs(hmacs->hmac_ids[i]); /* Check the id is in the supported range */ - if (id > SCTP_AUTH_HMAC_ID_MAX) + if (id > SCTP_AUTH_HMAC_ID_MAX) { + id = 0; continue; + } /* See is we support the id. Supported IDs have name and * length fields set, so that we can allocated and use * them. We can safely just check for name, for without the * name, we can't allocate the TFM. */ - if (!sctp_hmac_list[id].hmac_name) + if (!sctp_hmac_list[id].hmac_name) { + id = 0; continue; + } break; } -- cgit v0.10.2 From 35a27e8e6565910e5202f1af83ea85305fbf5ceb Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 1 Oct 2010 16:35:59 +0200 Subject: mfd: Rename struct irq_data to struct sih_irq_data struct irq_data is the preferred name for the data associated to an interrupt in the core code. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/drivers/mfd/twl4030-irq.c b/drivers/mfd/twl4030-irq.c index 097f24d..61c8664 100644 --- a/drivers/mfd/twl4030-irq.c +++ b/drivers/mfd/twl4030-irq.c @@ -78,7 +78,7 @@ struct sih { u8 irq_lines; /* number of supported irq lines */ /* SIR ignored -- set interrupt, for testing only */ - struct irq_data { + struct sih_irq_data { u8 isr_offset; u8 imr_offset; } mask[2]; -- cgit v0.10.2 From e7ffec1eb178db349c7fbfd647725e9d06619962 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 3 Oct 2010 20:08:13 +0200 Subject: isdn: Free irq_data namespace The irq_data namespace is the preference for the generic irq layer. Rename the union typedef in drivers/isdn/act2000/act2000.h Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/drivers/isdn/act2000/act2000.h b/drivers/isdn/act2000/act2000.h index d4c5051..88c9423 100644 --- a/drivers/isdn/act2000/act2000.h +++ b/drivers/isdn/act2000/act2000.h @@ -141,9 +141,9 @@ typedef struct irq_data_isa { __u8 rcvhdr[8]; } irq_data_isa; -typedef union irq_data { +typedef union act2000_irq_data { irq_data_isa isa; -} irq_data; +} act2000_irq_data; /* * Per card driver data @@ -176,7 +176,7 @@ typedef struct act2000_card { char *status_buf_read; char *status_buf_write; char *status_buf_end; - irq_data idat; /* Data used for IRQ handler */ + act2000_irq_data idat; /* Data used for IRQ handler */ isdn_if interface; /* Interface to upper layer */ char regname[35]; /* Name used for request_region */ } act2000_card; -- cgit v0.10.2 From d9817ebeeef16e01487549312c68540ca8f1561b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:45:59 +0000 Subject: genirq: Provide Kconfig The generic irq Kconfig options are copied around all archs. Provide a generic Kconfig file which can be included. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121843.217333624@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/init/Kconfig b/init/Kconfig index 2de5b1c..1df1a87 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -332,6 +332,8 @@ config AUDIT_TREE depends on AUDITSYSCALL select FSNOTIFY +source "kernel/irq/Kconfig" + menu "RCU Subsystem" choice diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig new file mode 100644 index 0000000..e0fc6cd --- /dev/null +++ b/kernel/irq/Kconfig @@ -0,0 +1,54 @@ +config HAVE_GENERIC_HARDIRQS + def_bool n + +if HAVE_GENERIC_HARDIRQS +menu "IRQ subsystem" +# +# Interrupt subsystem related configuration options +# +config GENERIC_HARDIRQS + def_bool y + +config GENERIC_HARDIRQS_NO__DO_IRQ + def_bool y + +# Options selectable by the architecture code +config HAVE_SPARSE_IRQ + def_bool n + +config GENERIC_IRQ_PROBE + def_bool n + +config GENERIC_PENDING_IRQ + def_bool n + +if SPARSE_IRQ && NUMA +config NUMA_IRQ_DESC + def_bool n +endif + +config AUTO_IRQ_AFFINITY + def_bool n + +config IRQ_PER_CPU + def_bool n + +config HARDIRQS_SW_RESEND + def_bool n + +config SPARSE_IRQ + bool "Support sparse irq numbering" + depends on HAVE_SPARSE_IRQ + ---help--- + + Sparse irq numbering is useful for distro kernels that want + to define a high CONFIG_NR_CPUS value but still want to have + low kernel memory footprint on smaller machines. + + ( Sparse irqs can also be beneficial on NUMA boxes, as they spread + out the interrupt descriptors in a more NUMA-friendly way. ) + + If you don't know what to do here, say N. + +endmenu +endif -- cgit v0.10.2 From 3bb9808e99bcc36eecb8e082bf70efb2a0bcdcb7 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:46:02 +0000 Subject: x86: Use genirq Kconfig Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121843.314600915@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..3ec657f 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -59,6 +59,11 @@ config X86 select ANON_INODES select HAVE_ARCH_KMEMCHECK select HAVE_USER_RETURN_NOTIFIER + select HAVE_GENERIC_HARDIRQS + select HAVE_SPARSE_IRQ + select NUMA_IRQ_DESC if (SPARSE_IRQ && NUMA) + select GENERIC_IRQ_PROBE + select GENERIC_PENDING_IRQ if SMP config INSTRUCTION_DECODER def_bool (KPROBES || PERF_EVENTS) @@ -200,20 +205,6 @@ config HAVE_INTEL_TXT def_bool y depends on EXPERIMENTAL && DMAR && ACPI -# Use the generic interrupt handling code in kernel/irq/: -config GENERIC_HARDIRQS - def_bool y - -config GENERIC_HARDIRQS_NO__DO_IRQ - def_bool y - -config GENERIC_IRQ_PROBE - def_bool y - -config GENERIC_PENDING_IRQ - def_bool y - depends on GENERIC_HARDIRQS && SMP - config USE_GENERIC_SMP_HELPERS def_bool y depends on SMP @@ -296,23 +287,6 @@ config X86_X2APIC If you don't know what to do here, say N. -config SPARSE_IRQ - bool "Support sparse irq numbering" - depends on PCI_MSI || HT_IRQ - ---help--- - This enables support for sparse irqs. This is useful for distro - kernels that want to define a high CONFIG_NR_CPUS value but still - want to have low kernel memory footprint on smaller machines. - - ( Sparse IRQs can also be beneficial on NUMA boxes, as they spread - out the irq_desc[] array in a more NUMA-friendly way. ) - - If you don't know what to do here, say N. - -config NUMA_IRQ_DESC - def_bool y - depends on SPARSE_IRQ && NUMA - config X86_MPPARSE bool "Enable MPS table" if ACPI default y -- cgit v0.10.2 From ff7dcd44dd446db2c3e13bdedf2d52b8e0127f16 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:25 +0000 Subject: genirq: Create irq_data Low level chip functions need access to irq_desc->handler_data, irq_desc->chip_data and irq_desc->msi_desc. We hand down the irq number to the low level functions, so they need to lookup irq_desc. With sparse irq this means a radix tree lookup. We could hand down irq_desc itself, but low level chip functions have no need to fiddle with it directly and we want to restrict access to irq_desc further. Preparatory patch for new chip functions. Note, that the ugly anon union/struct is there to avoid a full tree wide clean up for now. This is not going to last 3 years like __do_IRQ() Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121841.645542300@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/include/linux/irq.h b/include/linux/irq.h index 06273a2..363c76f 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -84,6 +84,37 @@ struct proc_dir_entry; struct msi_desc; /** + * struct irq_data - per irq and irq chip data passed down to chip functions + * @irq: interrupt number + * @node: node index useful for balancing + * @chip: low level interrupt hardware access + * @handler_data: per-IRQ data for the irq_chip methods + * @chip_data: platform-specific per-chip private data for the chip + * methods, to allow shared chip implementations + * @msi_desc: MSI descriptor + * @affinity: IRQ affinity on SMP + * @irq_2_iommu: iommu with this irq + * + * The fields here need to overlay the ones in irq_desc until we + * cleaned up the direct references and switched everything over to + * irq_data. + */ +struct irq_data { + unsigned int irq; + unsigned int node; + struct irq_chip *chip; + void *handler_data; + void *chip_data; + struct msi_desc *msi_desc; +#ifdef CONFIG_SMP + cpumask_var_t affinity; +#endif +#ifdef CONFIG_INTR_REMAP + struct irq_2_iommu *irq_2_iommu; +#endif +}; + +/** * struct irq_chip - hardware interrupt chip descriptor * * @name: name for /proc/interrupts @@ -140,16 +171,10 @@ struct timer_rand_state; struct irq_2_iommu; /** * struct irq_desc - interrupt descriptor - * @irq: interrupt number for this descriptor + * @irq_data: per irq and chip data passed down to chip functions * @timer_rand_state: pointer to timer rand state struct * @kstat_irqs: irq stats per cpu - * @irq_2_iommu: iommu with this irq * @handle_irq: highlevel irq-events handler [if NULL, __do_IRQ()] - * @chip: low level interrupt hardware access - * @msi_desc: MSI descriptor - * @handler_data: per-IRQ data for the irq_chip methods - * @chip_data: platform-specific per-chip private data for the chip - * methods, to allow shared chip implementations * @action: the irq action chain * @status: status information * @depth: disable-depth, for nested irq_disable() calls @@ -158,8 +183,6 @@ struct irq_2_iommu; * @last_unhandled: aging timer for unhandled count * @irqs_unhandled: stats field for spurious unhandled interrupts * @lock: locking for SMP - * @affinity: IRQ affinity on SMP - * @node: node index useful for balancing * @pending_mask: pending rebalanced interrupts * @threads_active: number of irqaction threads currently running * @wait_for_threads: wait queue for sync_irq to wait for threaded handlers @@ -167,17 +190,32 @@ struct irq_2_iommu; * @name: flow handler name for /proc/interrupts output */ struct irq_desc { - unsigned int irq; - struct timer_rand_state *timer_rand_state; - unsigned int *kstat_irqs; + + /* + * This union will go away, once we fixed the direct access to + * irq_desc all over the place. The direct fields are a 1:1 + * overlay of irq_data. + */ + union { + struct irq_data irq_data; + struct { + unsigned int irq; + unsigned int node; + struct irq_chip *chip; + void *handler_data; + void *chip_data; + struct msi_desc *msi_desc; +#ifdef CONFIG_SMP + cpumask_var_t affinity; +#endif #ifdef CONFIG_INTR_REMAP - struct irq_2_iommu *irq_2_iommu; + struct irq_2_iommu *irq_2_iommu; #endif + }; + }; + struct timer_rand_state *timer_rand_state; + unsigned int *kstat_irqs; irq_flow_handler_t handle_irq; - struct irq_chip *chip; - struct msi_desc *msi_desc; - void *handler_data; - void *chip_data; struct irqaction *action; /* IRQ action list */ unsigned int status; /* IRQ status */ @@ -188,9 +226,7 @@ struct irq_desc { unsigned int irqs_unhandled; raw_spinlock_t lock; #ifdef CONFIG_SMP - cpumask_var_t affinity; const struct cpumask *affinity_hint; - unsigned int node; #ifdef CONFIG_GENERIC_PENDING_IRQ cpumask_var_t pending_mask; #endif @@ -406,15 +442,15 @@ extern int set_irq_chip_data(unsigned int irq, void *data); extern int set_irq_type(unsigned int irq, unsigned int type); extern int set_irq_msi(unsigned int irq, struct msi_desc *entry); -#define get_irq_chip(irq) (irq_to_desc(irq)->chip) -#define get_irq_chip_data(irq) (irq_to_desc(irq)->chip_data) -#define get_irq_data(irq) (irq_to_desc(irq)->handler_data) -#define get_irq_msi(irq) (irq_to_desc(irq)->msi_desc) +#define get_irq_chip(irq) (irq_to_desc(irq)->irq_data.chip) +#define get_irq_chip_data(irq) (irq_to_desc(irq)->irq_data.chip_data) +#define get_irq_data(irq) (irq_to_desc(irq)->irq_data.handler_data) +#define get_irq_msi(irq) (irq_to_desc(irq)->irq_data.msi_desc) -#define get_irq_desc_chip(desc) ((desc)->chip) -#define get_irq_desc_chip_data(desc) ((desc)->chip_data) -#define get_irq_desc_data(desc) ((desc)->handler_data) -#define get_irq_desc_msi(desc) ((desc)->msi_desc) +#define get_irq_desc_chip(desc) ((desc)->irq_data.chip) +#define get_irq_desc_chip_data(desc) ((desc)->irq_data.chip_data) +#define get_irq_desc_data(desc) ((desc)->irq_data.handler_data) +#define get_irq_desc_msi(desc) ((desc)->irq_data.msi_desc) #endif /* CONFIG_GENERIC_HARDIRQS */ diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 27e5c69..099d4fc 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -75,12 +75,10 @@ EXPORT_SYMBOL_GPL(nr_irqs); #ifdef CONFIG_SPARSE_IRQ static struct irq_desc irq_desc_init = { - .irq = -1, - .status = IRQ_DISABLED, - .chip = &no_irq_chip, - .handle_irq = handle_bad_irq, - .depth = 1, - .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc_init.lock), + .status = IRQ_DISABLED, + .handle_irq = handle_bad_irq, + .depth = 1, + .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc_init.lock), }; void __ref init_kstat_irqs(struct irq_desc *desc, int node, int nr) @@ -105,7 +103,7 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int node) memcpy(desc, &irq_desc_init, sizeof(struct irq_desc)); raw_spin_lock_init(&desc->lock); - desc->irq = irq; + desc->irq_data.irq = irq; #ifdef CONFIG_SMP desc->node = node; #endif @@ -151,12 +149,10 @@ void replace_irq_desc(unsigned int irq, struct irq_desc *desc) static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_smp = { [0 ... NR_IRQS_LEGACY-1] = { - .irq = -1, - .status = IRQ_DISABLED, - .chip = &no_irq_chip, - .handle_irq = handle_bad_irq, - .depth = 1, - .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc_init.lock), + .status = IRQ_DISABLED, + .handle_irq = handle_bad_irq, + .depth = 1, + .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc_init.lock), } }; @@ -183,8 +179,11 @@ int __init early_irq_init(void) kstat_irqs_legacy = kzalloc_node(NR_IRQS_LEGACY * nr_cpu_ids * sizeof(int), GFP_NOWAIT, node); + irq_desc_init.irq_data.chip = &no_irq_chip; + for (i = 0; i < legacy_count; i++) { - desc[i].irq = i; + desc[i].irq_data.irq = i; + desc[i].irq_data.chip = &no_irq_chip; #ifdef CONFIG_SMP desc[i].node = node; #endif @@ -241,11 +240,10 @@ out_unlock: struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { [0 ... NR_IRQS-1] = { - .status = IRQ_DISABLED, - .chip = &no_irq_chip, - .handle_irq = handle_bad_irq, - .depth = 1, - .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock), + .status = IRQ_DISABLED, + .handle_irq = handle_bad_irq, + .depth = 1, + .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock), } }; @@ -264,7 +262,8 @@ int __init early_irq_init(void) count = ARRAY_SIZE(irq_desc); for (i = 0; i < count; i++) { - desc[i].irq = i; + desc[i].irq_data.irq = i; + desc[i].irq_data.chip = &no_irq_chip; alloc_desc_masks(&desc[i], 0, true); init_desc_masks(&desc[i]); desc[i].kstat_irqs = kstat_irqs_all[i]; -- cgit v0.10.2 From 6b8ff3120c758340505dddf08ad685ebb841d5d5 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 1 Oct 2010 12:58:38 +0200 Subject: genirq: Convert core code to irq_data Convert all references in the core code to orq, chip, handler_data, chip_data, msi_desc, affinity to irq_data.* Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/include/linux/irq.h b/include/linux/irq.h index 363c76f..002351d 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -475,12 +475,12 @@ static inline bool alloc_desc_masks(struct irq_desc *desc, int node, gfp = GFP_NOWAIT; #ifdef CONFIG_CPUMASK_OFFSTACK - if (!alloc_cpumask_var_node(&desc->affinity, gfp, node)) + if (!alloc_cpumask_var_node(&desc->irq_data.affinity, gfp, node)) return false; #ifdef CONFIG_GENERIC_PENDING_IRQ if (!alloc_cpumask_var_node(&desc->pending_mask, gfp, node)) { - free_cpumask_var(desc->affinity); + free_cpumask_var(desc->irq_data.affinity); return false; } #endif @@ -490,7 +490,7 @@ static inline bool alloc_desc_masks(struct irq_desc *desc, int node, static inline void init_desc_masks(struct irq_desc *desc) { - cpumask_setall(desc->affinity); + cpumask_setall(desc->irq_data.affinity); #ifdef CONFIG_GENERIC_PENDING_IRQ cpumask_clear(desc->pending_mask); #endif @@ -510,7 +510,7 @@ static inline void init_copy_desc_masks(struct irq_desc *old_desc, struct irq_desc *new_desc) { #ifdef CONFIG_CPUMASK_OFFSTACK - cpumask_copy(new_desc->affinity, old_desc->affinity); + cpumask_copy(new_desc->irq_data.affinity, old_desc->irq_data.affinity); #ifdef CONFIG_GENERIC_PENDING_IRQ cpumask_copy(new_desc->pending_mask, old_desc->pending_mask); @@ -521,7 +521,7 @@ static inline void init_copy_desc_masks(struct irq_desc *old_desc, static inline void free_desc_masks(struct irq_desc *old_desc, struct irq_desc *new_desc) { - free_cpumask_var(old_desc->affinity); + free_cpumask_var(old_desc->irq_data.affinity); #ifdef CONFIG_GENERIC_PENDING_IRQ free_cpumask_var(old_desc->pending_mask); diff --git a/kernel/irq/autoprobe.c b/kernel/irq/autoprobe.c index 2295a31..f9bf9b2 100644 --- a/kernel/irq/autoprobe.c +++ b/kernel/irq/autoprobe.c @@ -57,9 +57,9 @@ unsigned long probe_irq_on(void) * Some chips need to know about probing in * progress: */ - if (desc->chip->set_type) - desc->chip->set_type(i, IRQ_TYPE_PROBE); - desc->chip->startup(i); + if (desc->irq_data.chip->set_type) + desc->irq_data.chip->set_type(i, IRQ_TYPE_PROBE); + desc->irq_data.chip->startup(i); } raw_spin_unlock_irq(&desc->lock); } @@ -76,7 +76,7 @@ unsigned long probe_irq_on(void) raw_spin_lock_irq(&desc->lock); if (!desc->action && !(desc->status & IRQ_NOPROBE)) { desc->status |= IRQ_AUTODETECT | IRQ_WAITING; - if (desc->chip->startup(i)) + if (desc->irq_data.chip->startup(i)) desc->status |= IRQ_PENDING; } raw_spin_unlock_irq(&desc->lock); @@ -98,7 +98,7 @@ unsigned long probe_irq_on(void) /* It triggered already - consider it spurious. */ if (!(status & IRQ_WAITING)) { desc->status = status & ~IRQ_AUTODETECT; - desc->chip->shutdown(i); + desc->irq_data.chip->shutdown(i); } else if (i < 32) mask |= 1 << i; @@ -137,7 +137,7 @@ unsigned int probe_irq_mask(unsigned long val) mask |= 1 << i; desc->status = status & ~IRQ_AUTODETECT; - desc->chip->shutdown(i); + desc->irq_data.chip->shutdown(i); } raw_spin_unlock_irq(&desc->lock); } @@ -181,7 +181,7 @@ int probe_irq_off(unsigned long val) nr_of_irqs++; } desc->status = status & ~IRQ_AUTODETECT; - desc->chip->shutdown(i); + desc->irq_data.chip->shutdown(i); } raw_spin_unlock_irq(&desc->lock); } diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 4ea775c..e0e93ff 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -32,18 +32,18 @@ static void dynamic_irq_init_x(unsigned int irq, bool keep_chip_data) /* Ensure we don't have left over values from a previous use of this irq */ raw_spin_lock_irqsave(&desc->lock, flags); desc->status = IRQ_DISABLED; - desc->chip = &no_irq_chip; + desc->irq_data.chip = &no_irq_chip; desc->handle_irq = handle_bad_irq; desc->depth = 1; - desc->msi_desc = NULL; - desc->handler_data = NULL; + desc->irq_data.msi_desc = NULL; + desc->irq_data.handler_data = NULL; if (!keep_chip_data) - desc->chip_data = NULL; + desc->irq_data.chip_data = NULL; desc->action = NULL; desc->irq_count = 0; desc->irqs_unhandled = 0; #ifdef CONFIG_SMP - cpumask_setall(desc->affinity); + cpumask_setall(desc->irq_data.affinity); #ifdef CONFIG_GENERIC_PENDING_IRQ cpumask_clear(desc->pending_mask); #endif @@ -64,7 +64,7 @@ void dynamic_irq_init(unsigned int irq) * dynamic_irq_init_keep_chip_data - initialize a dynamically allocated irq * @irq: irq number to initialize * - * does not set irq_to_desc(irq)->chip_data to NULL + * does not set irq_to_desc(irq)->irq_data.chip_data to NULL */ void dynamic_irq_init_keep_chip_data(unsigned int irq) { @@ -88,12 +88,12 @@ static void dynamic_irq_cleanup_x(unsigned int irq, bool keep_chip_data) irq); return; } - desc->msi_desc = NULL; - desc->handler_data = NULL; + desc->irq_data.msi_desc = NULL; + desc->irq_data.handler_data = NULL; if (!keep_chip_data) - desc->chip_data = NULL; + desc->irq_data.chip_data = NULL; desc->handle_irq = handle_bad_irq; - desc->chip = &no_irq_chip; + desc->irq_data.chip = &no_irq_chip; desc->name = NULL; clear_kstat_irqs(desc); raw_spin_unlock_irqrestore(&desc->lock, flags); @@ -112,7 +112,7 @@ void dynamic_irq_cleanup(unsigned int irq) * dynamic_irq_cleanup_keep_chip_data - cleanup a dynamically allocated irq * @irq: irq number to initialize * - * does not set irq_to_desc(irq)->chip_data to NULL + * does not set irq_to_desc(irq)->irq_data.chip_data to NULL */ void dynamic_irq_cleanup_keep_chip_data(unsigned int irq) { @@ -140,7 +140,7 @@ int set_irq_chip(unsigned int irq, struct irq_chip *chip) raw_spin_lock_irqsave(&desc->lock, flags); irq_chip_set_defaults(chip); - desc->chip = chip; + desc->irq_data.chip = chip; raw_spin_unlock_irqrestore(&desc->lock, flags); return 0; @@ -193,7 +193,7 @@ int set_irq_data(unsigned int irq, void *data) } raw_spin_lock_irqsave(&desc->lock, flags); - desc->handler_data = data; + desc->irq_data.handler_data = data; raw_spin_unlock_irqrestore(&desc->lock, flags); return 0; } @@ -218,7 +218,7 @@ int set_irq_msi(unsigned int irq, struct msi_desc *entry) } raw_spin_lock_irqsave(&desc->lock, flags); - desc->msi_desc = entry; + desc->irq_data.msi_desc = entry; if (entry) entry->irq = irq; raw_spin_unlock_irqrestore(&desc->lock, flags); @@ -243,13 +243,13 @@ int set_irq_chip_data(unsigned int irq, void *data) return -EINVAL; } - if (!desc->chip) { + if (!desc->irq_data.chip) { printk(KERN_ERR "BUG: bad set_irq_chip_data(IRQ#%d)\n", irq); return -EINVAL; } raw_spin_lock_irqsave(&desc->lock, flags); - desc->chip_data = data; + desc->irq_data.chip_data = data; raw_spin_unlock_irqrestore(&desc->lock, flags); return 0; @@ -291,7 +291,7 @@ static void default_enable(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - desc->chip->unmask(irq); + desc->irq_data.chip->unmask(irq); desc->status &= ~IRQ_MASKED; } @@ -309,7 +309,7 @@ static unsigned int default_startup(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - desc->chip->enable(irq); + desc->irq_data.chip->enable(irq); return 0; } @@ -320,7 +320,7 @@ static void default_shutdown(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - desc->chip->mask(irq); + desc->irq_data.chip->mask(irq); desc->status |= IRQ_MASKED; } @@ -350,28 +350,28 @@ void irq_chip_set_defaults(struct irq_chip *chip) static inline void mask_ack_irq(struct irq_desc *desc, int irq) { - if (desc->chip->mask_ack) - desc->chip->mask_ack(irq); + if (desc->irq_data.chip->mask_ack) + desc->irq_data.chip->mask_ack(irq); else { - desc->chip->mask(irq); - if (desc->chip->ack) - desc->chip->ack(irq); + desc->irq_data.chip->mask(irq); + if (desc->irq_data.chip->ack) + desc->irq_data.chip->ack(irq); } desc->status |= IRQ_MASKED; } static inline void mask_irq(struct irq_desc *desc, int irq) { - if (desc->chip->mask) { - desc->chip->mask(irq); + if (desc->irq_data.chip->mask) { + desc->irq_data.chip->mask(irq); desc->status |= IRQ_MASKED; } } static inline void unmask_irq(struct irq_desc *desc, int irq) { - if (desc->chip->unmask) { - desc->chip->unmask(irq); + if (desc->irq_data.chip->unmask) { + desc->irq_data.chip->unmask(irq); desc->status &= ~IRQ_MASKED; } } @@ -552,7 +552,7 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc) raw_spin_lock(&desc->lock); desc->status &= ~IRQ_INPROGRESS; out: - desc->chip->eoi(irq); + desc->irq_data.chip->eoi(irq); raw_spin_unlock(&desc->lock); } @@ -594,8 +594,8 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) kstat_incr_irqs_this_cpu(irq, desc); /* Start handling the irq */ - if (desc->chip->ack) - desc->chip->ack(irq); + if (desc->irq_data.chip->ack) + desc->irq_data.chip->ack(irq); /* Mark the IRQ currently in progress.*/ desc->status |= IRQ_INPROGRESS; @@ -648,15 +648,15 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc) kstat_incr_irqs_this_cpu(irq, desc); - if (desc->chip->ack) - desc->chip->ack(irq); + if (desc->irq_data.chip->ack) + desc->irq_data.chip->ack(irq); action_ret = handle_IRQ_event(irq, desc->action); if (!noirqdebug) note_interrupt(irq, desc, action_ret); - if (desc->chip->eoi) - desc->chip->eoi(irq); + if (desc->irq_data.chip->eoi) + desc->irq_data.chip->eoi(irq); } void @@ -674,7 +674,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, if (!handle) handle = handle_bad_irq; - else if (desc->chip == &no_irq_chip) { + else if (desc->irq_data.chip == &no_irq_chip) { printk(KERN_WARNING "Trying to install %sinterrupt handler " "for IRQ%d\n", is_chained ? "chained " : "", irq); /* @@ -684,7 +684,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, * prevent us to setup the interrupt at all. Switch it to * dummy_irq_chip for easy transition. */ - desc->chip = &dummy_irq_chip; + desc->irq_data.chip = &dummy_irq_chip; } chip_bus_lock(irq, desc); @@ -692,7 +692,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, /* Uninstall? */ if (handle == handle_bad_irq) { - if (desc->chip != &no_irq_chip) + if (desc->irq_data.chip != &no_irq_chip) mask_ack_irq(desc, irq); desc->status |= IRQ_DISABLED; desc->depth = 1; @@ -704,7 +704,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, desc->status &= ~IRQ_DISABLED; desc->status |= IRQ_NOREQUEST | IRQ_NOPROBE; desc->depth = 0; - desc->chip->startup(irq); + desc->irq_data.chip->startup(irq); } raw_spin_unlock_irqrestore(&desc->lock, flags); chip_bus_sync_unlock(irq, desc); diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 099d4fc..fc27d76 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -105,7 +105,7 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int node) raw_spin_lock_init(&desc->lock); desc->irq_data.irq = irq; #ifdef CONFIG_SMP - desc->node = node; + desc->irq_data.node = node; #endif lockdep_set_class(&desc->lock, &irq_desc_lock_class); init_kstat_irqs(desc, node, nr_cpu_ids); @@ -185,7 +185,7 @@ int __init early_irq_init(void) desc[i].irq_data.irq = i; desc[i].irq_data.chip = &no_irq_chip; #ifdef CONFIG_SMP - desc[i].node = node; + desc[i].irq_data.node = node; #endif desc[i].kstat_irqs = kstat_irqs_legacy + i * nr_cpu_ids; lockdep_set_class(&desc[i].lock, &irq_desc_lock_class); @@ -456,20 +456,20 @@ unsigned int __do_IRQ(unsigned int irq) /* * No locking required for CPU-local interrupts: */ - if (desc->chip->ack) - desc->chip->ack(irq); + if (desc->irq_data.chip->ack) + desc->irq_data.chip->ack(irq); if (likely(!(desc->status & IRQ_DISABLED))) { action_ret = handle_IRQ_event(irq, desc->action); if (!noirqdebug) note_interrupt(irq, desc, action_ret); } - desc->chip->end(irq); + desc->irq_data.chip->end(irq); return 1; } raw_spin_lock(&desc->lock); - if (desc->chip->ack) - desc->chip->ack(irq); + if (desc->irq_data.chip->ack) + desc->irq_data.chip->ack(irq); /* * REPLAY is when Linux resends an IRQ that was dropped earlier * WAITING is used by probe to mark irqs that are being tested @@ -529,7 +529,7 @@ out: * The ->end() handler has to deal with interrupts which got * disabled while the handler was running. */ - desc->chip->end(irq); + desc->irq_data.chip->end(irq); raw_spin_unlock(&desc->lock); return 1; diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index c63f3bc..a805a00 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -43,14 +43,14 @@ extern void irq_set_thread_affinity(struct irq_desc *desc); /* Inline functions for support of irq chips on slow busses */ static inline void chip_bus_lock(unsigned int irq, struct irq_desc *desc) { - if (unlikely(desc->chip->bus_lock)) - desc->chip->bus_lock(irq); + if (unlikely(desc->irq_data.chip->bus_lock)) + desc->irq_data.chip->bus_lock(irq); } static inline void chip_bus_sync_unlock(unsigned int irq, struct irq_desc *desc) { - if (unlikely(desc->chip->bus_sync_unlock)) - desc->chip->bus_sync_unlock(irq); + if (unlikely(desc->irq_data.chip->bus_sync_unlock)) + desc->irq_data.chip->bus_sync_unlock(irq); } /* @@ -67,8 +67,8 @@ static inline void print_irq_desc(unsigned int irq, struct irq_desc *desc) irq, desc, desc->depth, desc->irq_count, desc->irqs_unhandled); printk("->handle_irq(): %p, ", desc->handle_irq); print_symbol("%s\n", (unsigned long)desc->handle_irq); - printk("->chip(): %p, ", desc->chip); - print_symbol("%s\n", (unsigned long)desc->chip); + printk("->irq_data.chip(): %p, ", desc->irq_data.chip); + print_symbol("%s\n", (unsigned long)desc->irq_data.chip); printk("->action(): %p\n", desc->action); if (desc->action) { printk("->action->handler(): %p, ", desc->action->handler); diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index c3003e9..4dfb195 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -73,8 +73,8 @@ int irq_can_set_affinity(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - if (CHECK_IRQ_PER_CPU(desc->status) || !desc->chip || - !desc->chip->set_affinity) + if (CHECK_IRQ_PER_CPU(desc->status) || !desc->irq_data.chip || + !desc->irq_data.chip->set_affinity) return 0; return 1; @@ -111,15 +111,15 @@ int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) struct irq_desc *desc = irq_to_desc(irq); unsigned long flags; - if (!desc->chip->set_affinity) + if (!desc->irq_data.chip->set_affinity) return -EINVAL; raw_spin_lock_irqsave(&desc->lock, flags); #ifdef CONFIG_GENERIC_PENDING_IRQ if (desc->status & IRQ_MOVE_PCNTXT) { - if (!desc->chip->set_affinity(irq, cpumask)) { - cpumask_copy(desc->affinity, cpumask); + if (!desc->irq_data.chip->set_affinity(irq, cpumask)) { + cpumask_copy(desc->irq_data.affinity, cpumask); irq_set_thread_affinity(desc); } } @@ -128,8 +128,8 @@ int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) cpumask_copy(desc->pending_mask, cpumask); } #else - if (!desc->chip->set_affinity(irq, cpumask)) { - cpumask_copy(desc->affinity, cpumask); + if (!desc->irq_data.chip->set_affinity(irq, cpumask)) { + cpumask_copy(desc->irq_data.affinity, cpumask); irq_set_thread_affinity(desc); } #endif @@ -168,16 +168,16 @@ static int setup_affinity(unsigned int irq, struct irq_desc *desc) * one of the targets is online. */ if (desc->status & (IRQ_AFFINITY_SET | IRQ_NO_BALANCING)) { - if (cpumask_any_and(desc->affinity, cpu_online_mask) + if (cpumask_any_and(desc->irq_data.affinity, cpu_online_mask) < nr_cpu_ids) goto set_affinity; else desc->status &= ~IRQ_AFFINITY_SET; } - cpumask_and(desc->affinity, cpu_online_mask, irq_default_affinity); + cpumask_and(desc->irq_data.affinity, cpu_online_mask, irq_default_affinity); set_affinity: - desc->chip->set_affinity(irq, desc->affinity); + desc->irq_data.chip->set_affinity(irq, desc->irq_data.affinity); return 0; } @@ -223,7 +223,7 @@ void __disable_irq(struct irq_desc *desc, unsigned int irq, bool suspend) if (!desc->depth++) { desc->status |= IRQ_DISABLED; - desc->chip->disable(irq); + desc->irq_data.chip->disable(irq); } } @@ -313,7 +313,7 @@ void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume) * IRQ line is re-enabled. * * This function may be called from IRQ context only when - * desc->chip->bus_lock and desc->chip->bus_sync_unlock are NULL ! + * desc->irq_data.chip->bus_lock and desc->chip->bus_sync_unlock are NULL ! */ void enable_irq(unsigned int irq) { @@ -336,8 +336,8 @@ static int set_irq_wake_real(unsigned int irq, unsigned int on) struct irq_desc *desc = irq_to_desc(irq); int ret = -ENXIO; - if (desc->chip->set_wake) - ret = desc->chip->set_wake(irq, on); + if (desc->irq_data.chip->set_wake) + ret = desc->irq_data.chip->set_wake(irq, on); return ret; } @@ -432,7 +432,7 @@ int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, unsigned long flags) { int ret; - struct irq_chip *chip = desc->chip; + struct irq_chip *chip = desc->irq_data.chip; if (!chip || !chip->set_type) { /* @@ -457,8 +457,8 @@ int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, desc->status &= ~(IRQ_LEVEL | IRQ_TYPE_SENSE_MASK); desc->status |= flags; - if (chip != desc->chip) - irq_chip_set_defaults(desc->chip); + if (chip != desc->irq_data.chip) + irq_chip_set_defaults(desc->irq_data.chip); } return ret; @@ -528,7 +528,7 @@ again: if (!(desc->status & IRQ_DISABLED) && (desc->status & IRQ_MASKED)) { desc->status &= ~IRQ_MASKED; - desc->chip->unmask(irq); + desc->irq_data.chip->unmask(irq); } raw_spin_unlock_irq(&desc->lock); chip_bus_sync_unlock(irq, desc); @@ -556,7 +556,7 @@ irq_thread_check_affinity(struct irq_desc *desc, struct irqaction *action) } raw_spin_lock_irq(&desc->lock); - cpumask_copy(mask, desc->affinity); + cpumask_copy(mask, desc->irq_data.affinity); raw_spin_unlock_irq(&desc->lock); set_cpus_allowed_ptr(current, mask); @@ -657,7 +657,7 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) if (!desc) return -EINVAL; - if (desc->chip == &no_irq_chip) + if (desc->irq_data.chip == &no_irq_chip) return -ENOSYS; /* * Some drivers like serial.c use request_irq() heavily, @@ -752,7 +752,7 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) } if (!shared) { - irq_chip_set_defaults(desc->chip); + irq_chip_set_defaults(desc->irq_data.chip); init_waitqueue_head(&desc->wait_for_threads); @@ -779,7 +779,7 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) if (!(desc->status & IRQ_NOAUTOEN)) { desc->depth = 0; desc->status &= ~IRQ_DISABLED; - desc->chip->startup(irq); + desc->irq_data.chip->startup(irq); } else /* Undo nested disables: */ desc->depth = 1; @@ -912,17 +912,17 @@ static struct irqaction *__free_irq(unsigned int irq, void *dev_id) /* Currently used only by UML, might disappear one day: */ #ifdef CONFIG_IRQ_RELEASE_METHOD - if (desc->chip->release) - desc->chip->release(irq, dev_id); + if (desc->irq_data.chip->release) + desc->irq_data.chip->release(irq, dev_id); #endif /* If this was the last handler, shut down the IRQ line: */ if (!desc->action) { desc->status |= IRQ_DISABLED; - if (desc->chip->shutdown) - desc->chip->shutdown(irq); + if (desc->irq_data.chip->shutdown) + desc->irq_data.chip->shutdown(irq); else - desc->chip->disable(irq); + desc->irq_data.chip->disable(irq); } #ifdef CONFIG_SMP diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index 2419622..f923c37 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -24,7 +24,7 @@ void move_masked_irq(int irq) if (unlikely(cpumask_empty(desc->pending_mask))) return; - if (!desc->chip->set_affinity) + if (!desc->irq_data.chip->set_affinity) return; assert_raw_spin_locked(&desc->lock); @@ -43,8 +43,8 @@ void move_masked_irq(int irq) */ if (likely(cpumask_any_and(desc->pending_mask, cpu_online_mask) < nr_cpu_ids)) - if (!desc->chip->set_affinity(irq, desc->pending_mask)) { - cpumask_copy(desc->affinity, desc->pending_mask); + if (!desc->irq_data.chip->set_affinity(irq, desc->pending_mask)) { + cpumask_copy(desc->irq_data.affinity, desc->pending_mask); irq_set_thread_affinity(desc); } @@ -61,8 +61,8 @@ void move_native_irq(int irq) if (unlikely(desc->status & IRQ_DISABLED)) return; - desc->chip->mask(irq); + desc->irq_data.chip->mask(irq); move_masked_irq(irq); - desc->chip->unmask(irq); + desc->irq_data.chip->unmask(irq); } diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index 65d3845..e7f1f16 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -44,7 +44,7 @@ static bool init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, return false; } raw_spin_lock_init(&desc->lock); - desc->node = node; + desc->irq_data.node = node; lockdep_set_class(&desc->lock, &irq_desc_lock_class); init_copy_kstat_irqs(old_desc, desc, node, nr_cpu_ids); init_copy_desc_masks(old_desc, desc); @@ -66,7 +66,7 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, unsigned int irq; unsigned long flags; - irq = old_desc->irq; + irq = old_desc->irq_data.irq; raw_spin_lock_irqsave(&sparse_irq_lock, flags); @@ -109,10 +109,10 @@ out_unlock: struct irq_desc *move_irq_desc(struct irq_desc *desc, int node) { /* those static or target node is -1, do not move them */ - if (desc->irq < NR_IRQS_LEGACY || node == -1) + if (desc->irq_data.irq < NR_IRQS_LEGACY || node == -1) return desc; - if (desc->node != node) + if (desc->irq_data.node != node) desc = __real_move_irq_desc(desc, node); return desc; diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index 09a2ee5..9b0da94 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -21,7 +21,7 @@ static struct proc_dir_entry *root_irq_dir; static int irq_affinity_proc_show(struct seq_file *m, void *v) { struct irq_desc *desc = irq_to_desc((long)m->private); - const struct cpumask *mask = desc->affinity; + const struct cpumask *mask = desc->irq_data.affinity; #ifdef CONFIG_GENERIC_PENDING_IRQ if (desc->status & IRQ_MOVE_PENDING) @@ -65,7 +65,7 @@ static ssize_t irq_affinity_proc_write(struct file *file, cpumask_var_t new_value; int err; - if (!irq_to_desc(irq)->chip->set_affinity || no_irq_affinity || + if (!irq_to_desc(irq)->irq_data.chip->set_affinity || no_irq_affinity || irq_balancing_disabled(irq)) return -EIO; @@ -185,7 +185,7 @@ static int irq_node_proc_show(struct seq_file *m, void *v) { struct irq_desc *desc = irq_to_desc((long) m->private); - seq_printf(m, "%d\n", desc->node); + seq_printf(m, "%d\n", desc->irq_data.node); return 0; } @@ -269,7 +269,7 @@ void register_irq_proc(unsigned int irq, struct irq_desc *desc) { char name [MAX_NAMELEN]; - if (!root_irq_dir || (desc->chip == &no_irq_chip) || desc->dir) + if (!root_irq_dir || (desc->irq_data.chip == &no_irq_chip) || desc->dir) return; memset(name, 0, MAX_NAMELEN); diff --git a/kernel/irq/resend.c b/kernel/irq/resend.c index 090c376..47c56a0 100644 --- a/kernel/irq/resend.c +++ b/kernel/irq/resend.c @@ -60,7 +60,7 @@ void check_irq_resend(struct irq_desc *desc, unsigned int irq) /* * Make sure the interrupt is enabled, before resending it: */ - desc->chip->enable(irq); + desc->irq_data.chip->enable(irq); /* * We do not resend level type interrupts. Level type @@ -70,7 +70,8 @@ void check_irq_resend(struct irq_desc *desc, unsigned int irq) if ((status & (IRQ_LEVEL | IRQ_PENDING | IRQ_REPLAY)) == IRQ_PENDING) { desc->status = (status & ~IRQ_PENDING) | IRQ_REPLAY; - if (!desc->chip->retrigger || !desc->chip->retrigger(irq)) { + if (!desc->irq_data.chip->retrigger || + !desc->irq_data.chip->retrigger(irq)) { #ifdef CONFIG_HARDIRQS_SW_RESEND /* Set it pending and activate the softirq: */ set_bit(irq, irqs_resend); diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c index 89fb90a..36c2c92 100644 --- a/kernel/irq/spurious.c +++ b/kernel/irq/spurious.c @@ -78,8 +78,8 @@ static int try_one_irq(int irq, struct irq_desc *desc) * If we did actual work for the real IRQ line we must let the * IRQ controller clean up too */ - if (work && desc->chip && desc->chip->end) - desc->chip->end(irq); + if (work && desc->irq_data.chip && desc->irq_data.chip->end) + desc->irq_data.chip->end(irq); raw_spin_unlock(&desc->lock); return ok; @@ -254,7 +254,7 @@ void note_interrupt(unsigned int irq, struct irq_desc *desc, printk(KERN_EMERG "Disabling IRQ #%d\n", irq); desc->status |= IRQ_DISABLED | IRQ_SPURIOUS_DISABLED; desc->depth++; - desc->chip->disable(irq); + desc->irq_data.chip->disable(irq); mod_timer(&poll_spurious_irq_timer, jiffies + POLL_SPURIOUS_IRQ_INTERVAL); -- cgit v0.10.2 From f8822657e799b02c55556c99a601261e207a299d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:32 +0000 Subject: genirq: Provide advanced irq chip functions The low level irq chip functions want access to irq_desc->irq_data. Provide new functions which hand down irq_data instead of the irq number so these functions avoid to call irq_to_desc() which is a radix tree lookup in case of sparse irq. This provides all the old functions except one: end(). end() is a relict of __do_IRQ() and will just go away with the __do_IRQ() code. The replacement for set_affinity() has an extra argument "bool force". The reason for this is to notify the low level code, that the move has to be done right away and cannot be delayed until the next interrupt happens. That's necessary to handle the irq fixup on cpu unplug in the generic code. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121841.742126604@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/include/linux/irq.h b/include/linux/irq.h index 002351d..0c83cbd 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -118,23 +118,38 @@ struct irq_data { * struct irq_chip - hardware interrupt chip descriptor * * @name: name for /proc/interrupts - * @startup: start up the interrupt (defaults to ->enable if NULL) - * @shutdown: shut down the interrupt (defaults to ->disable if NULL) - * @enable: enable the interrupt (defaults to chip->unmask if NULL) - * @disable: disable the interrupt - * @ack: start of a new interrupt - * @mask: mask an interrupt source - * @mask_ack: ack and mask an interrupt source - * @unmask: unmask an interrupt source - * @eoi: end of interrupt - chip level - * @end: end of interrupt - flow level - * @set_affinity: set the CPU affinity on SMP machines - * @retrigger: resend an IRQ to the CPU - * @set_type: set the flow type (IRQ_TYPE_LEVEL/etc.) of an IRQ - * @set_wake: enable/disable power-management wake-on of an IRQ + * @startup: deprecated, replaced by irq_startup + * @shutdown: deprecated, replaced by irq_shutdown + * @enable: deprecated, replaced by irq_enable + * @disable: deprecated, replaced by irq_disable + * @ack: deprecated, replaced by irq_ack + * @mask: deprecated, replaced by irq_mask + * @mask_ack: deprecated, replaced by irq_mask_ack + * @unmask: deprecated, replaced by irq_unmask + * @eoi: deprecated, replaced by irq_eoi + * @end: deprecated, will go away with __do_IRQ() + * @set_affinity: deprecated, replaced by irq_set_affinity + * @retrigger: deprecated, replaced by irq_retrigger + * @set_type: deprecated, replaced by irq_set_type + * @set_wake: deprecated, replaced by irq_wake + * @bus_lock: deprecated, replaced by irq_bus_lock + * @bus_sync_unlock: deprecated, replaced by irq_bus_sync_unlock * - * @bus_lock: function to lock access to slow bus (i2c) chips - * @bus_sync_unlock: function to sync and unlock slow bus (i2c) chips + * @irq_startup: start up the interrupt (defaults to ->enable if NULL) + * @irq_shutdown: shut down the interrupt (defaults to ->disable if NULL) + * @irq_enable: enable the interrupt (defaults to chip->unmask if NULL) + * @irq_disable: disable the interrupt + * @irq_ack: start of a new interrupt + * @irq_mask: mask an interrupt source + * @irq_mask_ack: ack and mask an interrupt source + * @irq_unmask: unmask an interrupt source + * @irq_eoi: end of interrupt + * @irq_set_affinity: set the CPU affinity on SMP machines + * @irq_retrigger: resend an IRQ to the CPU + * @irq_set_type: set the flow type (IRQ_TYPE_LEVEL/etc.) of an IRQ + * @irq_set_wake: enable/disable power-management wake-on of an IRQ + * @irq_bus_lock: function to lock access to slow bus (i2c) chips + * @irq_bus_sync_unlock:function to sync and unlock slow bus (i2c) chips * * @release: release function solely used by UML */ @@ -161,6 +176,25 @@ struct irq_chip { void (*bus_lock)(unsigned int irq); void (*bus_sync_unlock)(unsigned int irq); + unsigned int (*irq_startup)(struct irq_data *data); + void (*irq_shutdown)(struct irq_data *data); + void (*irq_enable)(struct irq_data *data); + void (*irq_disable)(struct irq_data *data); + + void (*irq_ack)(struct irq_data *data); + void (*irq_mask)(struct irq_data *data); + void (*irq_mask_ack)(struct irq_data *data); + void (*irq_unmask)(struct irq_data *data); + void (*irq_eoi)(struct irq_data *data); + + int (*irq_set_affinity)(struct irq_data *data, const struct cpumask *dest, bool force); + int (*irq_retrigger)(struct irq_data *data); + int (*irq_set_type)(struct irq_data *data, unsigned int flow_type); + int (*irq_set_wake)(struct irq_data *data, unsigned int on); + + void (*irq_bus_lock)(struct irq_data *data); + void (*irq_bus_sync_unlock)(struct irq_data *data); + /* Currently used only by UML, might disappear one day.*/ #ifdef CONFIG_IRQ_RELEASE_METHOD void (*release)(unsigned int irq, void *dev_id); -- cgit v0.10.2 From a77c4635915021c646cc017f22239e66d1aab4d5 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 1 Oct 2010 14:44:58 +0200 Subject: genirq: Add new functions to dummy chips The compat functions go away when the core code is converted. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index fc27d76..adca5b4 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -291,7 +291,15 @@ void clear_kstat_irqs(struct irq_desc *desc) * What should we do if we get a hw irq event on an illegal vector? * Each architecture has to answer this themself. */ -static void ack_bad(unsigned int irq) +static void ack_bad(struct irq_data *data) +{ + struct irq_desc *desc = irq_data_to_desc(data); + + print_irq_desc(data->irq, desc); + ack_bad_irq(data->irq); +} + +static void compat_ack_bad(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); @@ -302,11 +310,16 @@ static void ack_bad(unsigned int irq) /* * NOP functions */ -static void noop(unsigned int irq) +static void noop(struct irq_data *data) { } + +static unsigned int noop_ret(struct irq_data *data) { + return 0; } -static unsigned int noop_ret(unsigned int irq) +static void compat_noop(unsigned int irq) { } + +static unsigned int compat_noop_ret(unsigned int irq) { return 0; } @@ -316,12 +329,17 @@ static unsigned int noop_ret(unsigned int irq) */ struct irq_chip no_irq_chip = { .name = "none", - .startup = noop_ret, - .shutdown = noop, - .enable = noop, - .disable = noop, - .ack = ack_bad, - .end = noop, + .irq_startup = noop_ret, + .irq_shutdown = noop, + .irq_enable = noop, + .irq_disable = noop, + .irq_ack = ack_bad, + .startup = compat_noop_ret, + .shutdown = compat_noop, + .enable = compat_noop, + .disable = compat_noop, + .ack = compat_ack_bad, + .end = compat_noop, }; /* @@ -330,14 +348,21 @@ struct irq_chip no_irq_chip = { */ struct irq_chip dummy_irq_chip = { .name = "dummy", - .startup = noop_ret, - .shutdown = noop, - .enable = noop, - .disable = noop, - .ack = noop, - .mask = noop, - .unmask = noop, - .end = noop, + .irq_startup = noop_ret, + .irq_shutdown = noop, + .irq_enable = noop, + .irq_disable = noop, + .irq_ack = noop, + .irq_mask = noop, + .irq_unmask = noop, + .startup = compat_noop_ret, + .shutdown = compat_noop, + .enable = compat_noop, + .disable = compat_noop, + .ack = compat_noop, + .mask = compat_noop, + .unmask = compat_noop, + .end = compat_noop, }; /* diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index a805a00..562fc7e 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -4,6 +4,8 @@ extern int noirqdebug; +#define irq_data_to_desc(data) container_of(data, struct irq_desc, irq_data) + /* Set default functions for irq_chip structures: */ extern void irq_chip_set_defaults(struct irq_chip *chip); -- cgit v0.10.2 From 3876ec9ef3775d062345b3760d3271ecb8cd3fea Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:35 +0000 Subject: genirq: Provide compat handling for bus_lock/bus_sync_unlock Wrap the old chip functions for bus_lock/bus_sync_unlock until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121841.842536121@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index e0e93ff..77e551d 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -324,6 +324,17 @@ static void default_shutdown(unsigned int irq) desc->status |= IRQ_MASKED; } +/* Temporary migration helpers */ +static void compat_bus_lock(struct irq_data *data) +{ + data->chip->bus_lock(data->irq); +} + +static void compat_bus_sync_unlock(struct irq_data *data) +{ + data->chip->bus_sync_unlock(data->irq); +} + /* * Fixup enable/disable function pointers */ @@ -346,6 +357,11 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->disable : default_shutdown; if (!chip->end) chip->end = dummy_irq_chip.end; + + if (chip->bus_lock) + chip->irq_bus_lock = compat_bus_lock; + if (chip->bus_sync_unlock) + chip->irq_bus_sync_unlock = compat_bus_sync_unlock; } static inline void mask_ack_irq(struct irq_desc *desc, int irq) @@ -687,7 +703,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, desc->irq_data.chip = &dummy_irq_chip; } - chip_bus_lock(irq, desc); + chip_bus_lock(desc); raw_spin_lock_irqsave(&desc->lock, flags); /* Uninstall? */ @@ -707,7 +723,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, desc->irq_data.chip->startup(irq); } raw_spin_unlock_irqrestore(&desc->lock, flags); - chip_bus_sync_unlock(irq, desc); + chip_bus_sync_unlock(desc); } EXPORT_SYMBOL_GPL(__set_irq_handler); diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index 562fc7e..ecafbfe 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -43,16 +43,16 @@ extern int irq_select_affinity_usr(unsigned int irq); extern void irq_set_thread_affinity(struct irq_desc *desc); /* Inline functions for support of irq chips on slow busses */ -static inline void chip_bus_lock(unsigned int irq, struct irq_desc *desc) +static inline void chip_bus_lock(struct irq_desc *desc) { - if (unlikely(desc->irq_data.chip->bus_lock)) - desc->irq_data.chip->bus_lock(irq); + if (unlikely(desc->irq_data.chip->irq_bus_lock)) + desc->irq_data.chip->irq_bus_lock(&desc->irq_data); } -static inline void chip_bus_sync_unlock(unsigned int irq, struct irq_desc *desc) +static inline void chip_bus_sync_unlock(struct irq_desc *desc) { - if (unlikely(desc->irq_data.chip->bus_sync_unlock)) - desc->irq_data.chip->bus_sync_unlock(irq); + if (unlikely(desc->irq_data.chip->irq_bus_sync_unlock)) + desc->irq_data.chip->irq_bus_sync_unlock(&desc->irq_data); } /* diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 4dfb195..dfb02ff 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -246,11 +246,11 @@ void disable_irq_nosync(unsigned int irq) if (!desc) return; - chip_bus_lock(irq, desc); + chip_bus_lock(desc); raw_spin_lock_irqsave(&desc->lock, flags); __disable_irq(desc, irq, false); raw_spin_unlock_irqrestore(&desc->lock, flags); - chip_bus_sync_unlock(irq, desc); + chip_bus_sync_unlock(desc); } EXPORT_SYMBOL(disable_irq_nosync); @@ -323,11 +323,11 @@ void enable_irq(unsigned int irq) if (!desc) return; - chip_bus_lock(irq, desc); + chip_bus_lock(desc); raw_spin_lock_irqsave(&desc->lock, flags); __enable_irq(desc, irq, false); raw_spin_unlock_irqrestore(&desc->lock, flags); - chip_bus_sync_unlock(irq, desc); + chip_bus_sync_unlock(desc); } EXPORT_SYMBOL(enable_irq); @@ -507,7 +507,7 @@ static int irq_wait_for_interrupt(struct irqaction *action) static void irq_finalize_oneshot(unsigned int irq, struct irq_desc *desc) { again: - chip_bus_lock(irq, desc); + chip_bus_lock(desc); raw_spin_lock_irq(&desc->lock); /* @@ -521,7 +521,7 @@ again: */ if (unlikely(desc->status & IRQ_INPROGRESS)) { raw_spin_unlock_irq(&desc->lock); - chip_bus_sync_unlock(irq, desc); + chip_bus_sync_unlock(desc); cpu_relax(); goto again; } @@ -531,7 +531,7 @@ again: desc->irq_data.chip->unmask(irq); } raw_spin_unlock_irq(&desc->lock); - chip_bus_sync_unlock(irq, desc); + chip_bus_sync_unlock(desc); } #ifdef CONFIG_SMP @@ -997,9 +997,9 @@ void free_irq(unsigned int irq, void *dev_id) if (!desc) return; - chip_bus_lock(irq, desc); + chip_bus_lock(desc); kfree(__free_irq(irq, dev_id)); - chip_bus_sync_unlock(irq, desc); + chip_bus_sync_unlock(desc); } EXPORT_SYMBOL(free_irq); @@ -1086,9 +1086,9 @@ int request_threaded_irq(unsigned int irq, irq_handler_t handler, action->name = devname; action->dev_id = dev_id; - chip_bus_lock(irq, desc); + chip_bus_lock(desc); retval = __setup_irq(irq, desc, action); - chip_bus_sync_unlock(irq, desc); + chip_bus_sync_unlock(desc); if (retval) kfree(action); -- cgit v0.10.2 From e2c0f8ff0fc26959952fbfa89f732fef928df77f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:42 +0000 Subject: genirq: Provide compat handling for chip->mask() Wrap the old chip function mask() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121841.940355859@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 77e551d..c041270 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -320,11 +320,16 @@ static void default_shutdown(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - desc->irq_data.chip->mask(irq); + desc->irq_data.chip->irq_mask(&desc->irq_data); desc->status |= IRQ_MASKED; } /* Temporary migration helpers */ +static void compat_irq_mask(struct irq_data *data) +{ + data->chip->mask(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -362,6 +367,9 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_bus_lock = compat_bus_lock; if (chip->bus_sync_unlock) chip->irq_bus_sync_unlock = compat_bus_sync_unlock; + + if (chip->mask) + chip->irq_mask = compat_irq_mask; } static inline void mask_ack_irq(struct irq_desc *desc, int irq) @@ -369,17 +377,17 @@ static inline void mask_ack_irq(struct irq_desc *desc, int irq) if (desc->irq_data.chip->mask_ack) desc->irq_data.chip->mask_ack(irq); else { - desc->irq_data.chip->mask(irq); + desc->irq_data.chip->irq_mask(&desc->irq_data); if (desc->irq_data.chip->ack) desc->irq_data.chip->ack(irq); } desc->status |= IRQ_MASKED; } -static inline void mask_irq(struct irq_desc *desc, int irq) +static inline void mask_irq(struct irq_desc *desc) { - if (desc->irq_data.chip->mask) { - desc->irq_data.chip->mask(irq); + if (desc->irq_data.chip->irq_mask) { + desc->irq_data.chip->irq_mask(&desc->irq_data); desc->status |= IRQ_MASKED; } } @@ -553,7 +561,7 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc) action = desc->action; if (unlikely(!action || (desc->status & IRQ_DISABLED))) { desc->status |= IRQ_PENDING; - mask_irq(desc, irq); + mask_irq(desc); goto out; } @@ -621,7 +629,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) irqreturn_t action_ret; if (unlikely(!action)) { - mask_irq(desc, irq); + mask_irq(desc); goto out_unlock; } diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index adca5b4..3b160ac 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -360,7 +360,6 @@ struct irq_chip dummy_irq_chip = { .enable = compat_noop, .disable = compat_noop, .ack = compat_noop, - .mask = compat_noop, .unmask = compat_noop, .end = compat_noop, }; diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index f923c37..b165ec2 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -61,7 +61,7 @@ void move_native_irq(int irq) if (unlikely(desc->status & IRQ_DISABLED)) return; - desc->irq_data.chip->mask(irq); + desc->irq_data.chip->irq_mask(&desc->irq_data); move_masked_irq(irq); desc->irq_data.chip->unmask(irq); } -- cgit v0.10.2 From 0eda58b7f3a30c9a13d83db1cfaab00e1c452055 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:44 +0000 Subject: genirq: Provide compat handling for chip->unmask() Wrap the old chip function unmask() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.043608928@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index c041270..dbdb59a 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -291,7 +291,7 @@ static void default_enable(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - desc->irq_data.chip->unmask(irq); + desc->irq_data.chip->irq_unmask(&desc->irq_data); desc->status &= ~IRQ_MASKED; } @@ -330,6 +330,11 @@ static void compat_irq_mask(struct irq_data *data) data->chip->mask(data->irq); } +static void compat_irq_unmask(struct irq_data *data) +{ + data->chip->unmask(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -370,6 +375,9 @@ void irq_chip_set_defaults(struct irq_chip *chip) if (chip->mask) chip->irq_mask = compat_irq_mask; + + if (chip->unmask) + chip->irq_unmask = compat_irq_unmask; } static inline void mask_ack_irq(struct irq_desc *desc, int irq) @@ -392,10 +400,10 @@ static inline void mask_irq(struct irq_desc *desc) } } -static inline void unmask_irq(struct irq_desc *desc, int irq) +static inline void unmask_irq(struct irq_desc *desc) { - if (desc->irq_data.chip->unmask) { - desc->irq_data.chip->unmask(irq); + if (desc->irq_data.chip->irq_unmask) { + desc->irq_data.chip->irq_unmask(&desc->irq_data); desc->status &= ~IRQ_MASKED; } } @@ -524,7 +532,7 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc) desc->status &= ~IRQ_INPROGRESS; if (!(desc->status & (IRQ_DISABLED | IRQ_ONESHOT))) - unmask_irq(desc, irq); + unmask_irq(desc); out_unlock: raw_spin_unlock(&desc->lock); } @@ -641,7 +649,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) if (unlikely((desc->status & (IRQ_PENDING | IRQ_MASKED | IRQ_DISABLED)) == (IRQ_PENDING | IRQ_MASKED))) { - unmask_irq(desc, irq); + unmask_irq(desc); } desc->status &= ~IRQ_PENDING; diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 3b160ac..f334c8c 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -360,7 +360,6 @@ struct irq_chip dummy_irq_chip = { .enable = compat_noop, .disable = compat_noop, .ack = compat_noop, - .unmask = compat_noop, .end = compat_noop, }; diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index dfb02ff..b3986bc 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -528,7 +528,7 @@ again: if (!(desc->status & IRQ_DISABLED) && (desc->status & IRQ_MASKED)) { desc->status &= ~IRQ_MASKED; - desc->irq_data.chip->unmask(irq); + desc->irq_data.chip->irq_unmask(&desc->irq_data); } raw_spin_unlock_irq(&desc->lock); chip_bus_sync_unlock(desc); diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index b165ec2..7888e5d 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -63,6 +63,6 @@ void move_native_irq(int irq) desc->irq_data.chip->irq_mask(&desc->irq_data); move_masked_irq(irq); - desc->irq_data.chip->unmask(irq); + desc->irq_data.chip->irq_unmask(&desc->irq_data); } -- cgit v0.10.2 From 22a49163e90dd7088f801dd54e25b04e1f337e9b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:47 +0000 Subject: genirq: Provide compat handling for chip->ack() Wrap the old chip function ack() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.142624725@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index dbdb59a..864c3ab 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -335,6 +335,11 @@ static void compat_irq_unmask(struct irq_data *data) data->chip->unmask(data->irq); } +static void compat_irq_ack(struct irq_data *data) +{ + data->chip->ack(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -372,12 +377,12 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_bus_lock = compat_bus_lock; if (chip->bus_sync_unlock) chip->irq_bus_sync_unlock = compat_bus_sync_unlock; - if (chip->mask) chip->irq_mask = compat_irq_mask; - if (chip->unmask) chip->irq_unmask = compat_irq_unmask; + if (chip->ack) + chip->irq_ack = compat_irq_ack; } static inline void mask_ack_irq(struct irq_desc *desc, int irq) @@ -386,8 +391,8 @@ static inline void mask_ack_irq(struct irq_desc *desc, int irq) desc->irq_data.chip->mask_ack(irq); else { desc->irq_data.chip->irq_mask(&desc->irq_data); - if (desc->irq_data.chip->ack) - desc->irq_data.chip->ack(irq); + if (desc->irq_data.chip->irq_ack) + desc->irq_data.chip->irq_ack(&desc->irq_data); } desc->status |= IRQ_MASKED; } @@ -626,8 +631,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) kstat_incr_irqs_this_cpu(irq, desc); /* Start handling the irq */ - if (desc->irq_data.chip->ack) - desc->irq_data.chip->ack(irq); + desc->irq_data.chip->irq_ack(&desc->irq_data); /* Mark the IRQ currently in progress.*/ desc->status |= IRQ_INPROGRESS; @@ -680,8 +684,8 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc) kstat_incr_irqs_this_cpu(irq, desc); - if (desc->irq_data.chip->ack) - desc->irq_data.chip->ack(irq); + if (desc->irq_data.chip->irq_ack) + desc->irq_data.chip->irq_ack(&desc->irq_data); action_ret = handle_IRQ_event(irq, desc->action); if (!noirqdebug) diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index f334c8c..9ba7aec 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -299,14 +299,6 @@ static void ack_bad(struct irq_data *data) ack_bad_irq(data->irq); } -static void compat_ack_bad(unsigned int irq) -{ - struct irq_desc *desc = irq_to_desc(irq); - - print_irq_desc(irq, desc); - ack_bad_irq(irq); -} - /* * NOP functions */ @@ -338,7 +330,6 @@ struct irq_chip no_irq_chip = { .shutdown = compat_noop, .enable = compat_noop, .disable = compat_noop, - .ack = compat_ack_bad, .end = compat_noop, }; @@ -359,7 +350,6 @@ struct irq_chip dummy_irq_chip = { .shutdown = compat_noop, .enable = compat_noop, .disable = compat_noop, - .ack = compat_noop, .end = compat_noop, }; -- cgit v0.10.2 From 9205e31d1af0f725e71bbae10d199c6b9e8d6dd8 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:50 +0000 Subject: genirq: Provide compat handling for chip->mask_ack() Wrap the old chip function mask_ack() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.240806983@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 864c3ab..09c1a44 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -340,6 +340,11 @@ static void compat_irq_ack(struct irq_data *data) data->chip->ack(data->irq); } +static void compat_irq_mask_ack(struct irq_data *data) +{ + data->chip->mask_ack(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -383,12 +388,14 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_unmask = compat_irq_unmask; if (chip->ack) chip->irq_ack = compat_irq_ack; + if (chip->mask_ack) + chip->irq_mask_ack = compat_irq_mask_ack; } -static inline void mask_ack_irq(struct irq_desc *desc, int irq) +static inline void mask_ack_irq(struct irq_desc *desc) { - if (desc->irq_data.chip->mask_ack) - desc->irq_data.chip->mask_ack(irq); + if (desc->irq_data.chip->irq_mask_ack) + desc->irq_data.chip->irq_mask_ack(&desc->irq_data); else { desc->irq_data.chip->irq_mask(&desc->irq_data); if (desc->irq_data.chip->irq_ack) @@ -511,7 +518,7 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc) irqreturn_t action_ret; raw_spin_lock(&desc->lock); - mask_ack_irq(desc, irq); + mask_ack_irq(desc); if (unlikely(desc->status & IRQ_INPROGRESS)) goto out_unlock; @@ -625,7 +632,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) if (unlikely((desc->status & (IRQ_INPROGRESS | IRQ_DISABLED)) || !desc->action)) { desc->status |= (IRQ_PENDING | IRQ_MASKED); - mask_ack_irq(desc, irq); + mask_ack_irq(desc); goto out_unlock; } kstat_incr_irqs_this_cpu(irq, desc); @@ -729,7 +736,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, /* Uninstall? */ if (handle == handle_bad_irq) { if (desc->irq_data.chip != &no_irq_chip) - mask_ack_irq(desc, irq); + mask_ack_irq(desc); desc->status |= IRQ_DISABLED; desc->depth = 1; } -- cgit v0.10.2 From 0c5c15572ac096001f52d37b416f2a4be9aebb80 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:53 +0000 Subject: genirq: Provide compat handling for chip->eoi() Wrap the old chip function eoi() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.339657617@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 09c1a44..c8648a8 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -345,6 +345,11 @@ static void compat_irq_mask_ack(struct irq_data *data) data->chip->mask_ack(data->irq); } +static void compat_irq_eoi(struct irq_data *data) +{ + data->chip->eoi(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -390,6 +395,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_ack = compat_irq_ack; if (chip->mask_ack) chip->irq_mask_ack = compat_irq_mask_ack; + if (chip->eoi) + chip->irq_eoi = compat_irq_eoi; } static inline void mask_ack_irq(struct irq_desc *desc) @@ -596,7 +603,7 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc) raw_spin_lock(&desc->lock); desc->status &= ~IRQ_INPROGRESS; out: - desc->irq_data.chip->eoi(irq); + desc->irq_data.chip->irq_eoi(&desc->irq_data); raw_spin_unlock(&desc->lock); } @@ -698,8 +705,8 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc) if (!noirqdebug) note_interrupt(irq, desc, action_ret); - if (desc->irq_data.chip->eoi) - desc->irq_data.chip->eoi(irq); + if (desc->irq_data.chip->irq_eoi) + desc->irq_data.chip->irq_eoi(&desc->irq_data); } void -- cgit v0.10.2 From c5f756344c390f629243b4a28c2bd198fdfd7ee9 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:44:56 +0000 Subject: genirq: Provide compat handling for chip->enable() Wrap the old chip function enable() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.437159182@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index c8648a8..a95b478 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -287,9 +287,9 @@ EXPORT_SYMBOL_GPL(set_irq_nested_thread); /* * default enable function */ -static void default_enable(unsigned int irq) +static void default_enable(struct irq_data *data) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_desc *desc = irq_data_to_desc(data); desc->irq_data.chip->irq_unmask(&desc->irq_data); desc->status &= ~IRQ_MASKED; @@ -309,7 +309,7 @@ static unsigned int default_startup(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - desc->irq_data.chip->enable(irq); + desc->irq_data.chip->irq_enable(&desc->irq_data); return 0; } @@ -350,6 +350,11 @@ static void compat_irq_eoi(struct irq_data *data) data->chip->eoi(data->irq); } +static void compat_irq_enable(struct irq_data *data) +{ + data->chip->enable(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -365,8 +370,18 @@ static void compat_bus_sync_unlock(struct irq_data *data) */ void irq_chip_set_defaults(struct irq_chip *chip) { - if (!chip->enable) - chip->enable = default_enable; + /* + * Compat fixup functions need to be before we set the + * defaults for enable/disable/startup/shutdown + */ + if (chip->enable) + chip->irq_enable = compat_irq_enable; + + /* + * The real defaults + */ + if (!chip->irq_enable) + chip->irq_enable = default_enable; if (!chip->disable) chip->disable = default_disable; if (!chip->startup) diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 9ba7aec..ac8c749 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -328,7 +328,6 @@ struct irq_chip no_irq_chip = { .irq_ack = ack_bad, .startup = compat_noop_ret, .shutdown = compat_noop, - .enable = compat_noop, .disable = compat_noop, .end = compat_noop, }; @@ -348,7 +347,6 @@ struct irq_chip dummy_irq_chip = { .irq_unmask = noop, .startup = compat_noop_ret, .shutdown = compat_noop, - .enable = compat_noop, .disable = compat_noop, .end = compat_noop, }; diff --git a/kernel/irq/resend.c b/kernel/irq/resend.c index 47c56a0..a798a23 100644 --- a/kernel/irq/resend.c +++ b/kernel/irq/resend.c @@ -60,7 +60,7 @@ void check_irq_resend(struct irq_desc *desc, unsigned int irq) /* * Make sure the interrupt is enabled, before resending it: */ - desc->irq_data.chip->enable(irq); + desc->irq_data.chip->irq_enable(&desc->irq_data); /* * We do not resend level type interrupts. Level type -- cgit v0.10.2 From bc310dda41be6439364c8f3b9fe7c9d743d22b1c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:45:02 +0000 Subject: genirq: Provide compat handling for chip->disable()/shutdown() Wrap the old chip functions disable() and shutdown() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.532070631@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/autoprobe.c b/kernel/irq/autoprobe.c index f9bf9b2..95806a4 100644 --- a/kernel/irq/autoprobe.c +++ b/kernel/irq/autoprobe.c @@ -98,7 +98,7 @@ unsigned long probe_irq_on(void) /* It triggered already - consider it spurious. */ if (!(status & IRQ_WAITING)) { desc->status = status & ~IRQ_AUTODETECT; - desc->irq_data.chip->shutdown(i); + desc->irq_data.chip->irq_shutdown(&desc->irq_data); } else if (i < 32) mask |= 1 << i; @@ -137,7 +137,7 @@ unsigned int probe_irq_mask(unsigned long val) mask |= 1 << i; desc->status = status & ~IRQ_AUTODETECT; - desc->irq_data.chip->shutdown(i); + desc->irq_data.chip->irq_shutdown(&desc->irq_data); } raw_spin_unlock_irq(&desc->lock); } @@ -181,7 +181,7 @@ int probe_irq_off(unsigned long val) nr_of_irqs++; } desc->status = status & ~IRQ_AUTODETECT; - desc->irq_data.chip->shutdown(i); + desc->irq_data.chip->irq_shutdown(&desc->irq_data); } raw_spin_unlock_irq(&desc->lock); } diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index a95b478..b8a47f0 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -298,7 +298,7 @@ static void default_enable(struct irq_data *data) /* * default disable function */ -static void default_disable(unsigned int irq) +static void default_disable(struct irq_data *data) { } @@ -316,9 +316,9 @@ static unsigned int default_startup(unsigned int irq) /* * default shutdown function */ -static void default_shutdown(unsigned int irq) +static void default_shutdown(struct irq_data *data) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_desc *desc = irq_data_to_desc(data); desc->irq_data.chip->irq_mask(&desc->irq_data); desc->status |= IRQ_MASKED; @@ -355,6 +355,16 @@ static void compat_irq_enable(struct irq_data *data) data->chip->enable(data->irq); } +static void compat_irq_disable(struct irq_data *data) +{ + data->chip->disable(data->irq); +} + +static void compat_irq_shutdown(struct irq_data *data) +{ + data->chip->shutdown(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -376,28 +386,35 @@ void irq_chip_set_defaults(struct irq_chip *chip) */ if (chip->enable) chip->irq_enable = compat_irq_enable; + if (chip->disable) + chip->irq_disable = compat_irq_disable; + if (chip->shutdown) + chip->irq_shutdown = compat_irq_shutdown; /* * The real defaults */ if (!chip->irq_enable) chip->irq_enable = default_enable; - if (!chip->disable) - chip->disable = default_disable; + if (!chip->irq_disable) + chip->irq_disable = default_disable; if (!chip->startup) chip->startup = default_startup; /* - * We use chip->disable, when the user provided its own. When - * we have default_disable set for chip->disable, then we need + * We use chip->irq_disable, when the user provided its own. When + * we have default_disable set for chip->irq_disable, then we need * to use default_shutdown, otherwise the irq line is not * disabled on free_irq(): */ - if (!chip->shutdown) - chip->shutdown = chip->disable != default_disable ? - chip->disable : default_shutdown; + if (!chip->irq_shutdown) + chip->irq_shutdown = chip->irq_disable != default_disable ? + chip->irq_disable : default_shutdown; if (!chip->end) chip->end = dummy_irq_chip.end; + /* + * Now fix up the remaining compat handlers + */ if (chip->bus_lock) chip->irq_bus_lock = compat_bus_lock; if (chip->bus_sync_unlock) diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index ac8c749..60e25c4 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -327,8 +327,6 @@ struct irq_chip no_irq_chip = { .irq_disable = noop, .irq_ack = ack_bad, .startup = compat_noop_ret, - .shutdown = compat_noop, - .disable = compat_noop, .end = compat_noop, }; @@ -346,8 +344,6 @@ struct irq_chip dummy_irq_chip = { .irq_mask = noop, .irq_unmask = noop, .startup = compat_noop_ret, - .shutdown = compat_noop, - .disable = compat_noop, .end = compat_noop, }; diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index b3986bc..f3f36f6 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -223,7 +223,7 @@ void __disable_irq(struct irq_desc *desc, unsigned int irq, bool suspend) if (!desc->depth++) { desc->status |= IRQ_DISABLED; - desc->irq_data.chip->disable(irq); + desc->irq_data.chip->irq_disable(&desc->irq_data); } } @@ -919,10 +919,10 @@ static struct irqaction *__free_irq(unsigned int irq, void *dev_id) /* If this was the last handler, shut down the IRQ line: */ if (!desc->action) { desc->status |= IRQ_DISABLED; - if (desc->irq_data.chip->shutdown) - desc->irq_data.chip->shutdown(irq); + if (desc->irq_data.chip->irq_shutdown) + desc->irq_data.chip->irq_shutdown(&desc->irq_data); else - desc->irq_data.chip->disable(irq); + desc->irq_data.chip->irq_disable(&desc->irq_data); } #ifdef CONFIG_SMP diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c index 36c2c92..9ee704d 100644 --- a/kernel/irq/spurious.c +++ b/kernel/irq/spurious.c @@ -254,7 +254,7 @@ void note_interrupt(unsigned int irq, struct irq_desc *desc, printk(KERN_EMERG "Disabling IRQ #%d\n", irq); desc->status |= IRQ_DISABLED | IRQ_SPURIOUS_DISABLED; desc->depth++; - desc->irq_data.chip->disable(irq); + desc->irq_data.chip->irq_disable(&desc->irq_data); mod_timer(&poll_spurious_irq_timer, jiffies + POLL_SPURIOUS_IRQ_INTERVAL); -- cgit v0.10.2 From 37e12df709f09eac17314d79a52190ac46746e33 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:45:38 +0000 Subject: genirq: Provide compat handling for chip->startup() Wrap the old chip function startup() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.635152961@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/autoprobe.c b/kernel/irq/autoprobe.c index 95806a4..7a46825 100644 --- a/kernel/irq/autoprobe.c +++ b/kernel/irq/autoprobe.c @@ -59,7 +59,7 @@ unsigned long probe_irq_on(void) */ if (desc->irq_data.chip->set_type) desc->irq_data.chip->set_type(i, IRQ_TYPE_PROBE); - desc->irq_data.chip->startup(i); + desc->irq_data.chip->irq_startup(&desc->irq_data); } raw_spin_unlock_irq(&desc->lock); } @@ -76,7 +76,7 @@ unsigned long probe_irq_on(void) raw_spin_lock_irq(&desc->lock); if (!desc->action && !(desc->status & IRQ_NOPROBE)) { desc->status |= IRQ_AUTODETECT | IRQ_WAITING; - if (desc->irq_data.chip->startup(i)) + if (desc->irq_data.chip->irq_startup(&desc->irq_data)) desc->status |= IRQ_PENDING; } raw_spin_unlock_irq(&desc->lock); diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index b8a47f0..cce85f0 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -305,11 +305,11 @@ static void default_disable(struct irq_data *data) /* * default startup function */ -static unsigned int default_startup(unsigned int irq) +static unsigned int default_startup(struct irq_data *data) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_desc *desc = irq_data_to_desc(data); - desc->irq_data.chip->irq_enable(&desc->irq_data); + desc->irq_data.chip->irq_enable(data); return 0; } @@ -365,6 +365,11 @@ static void compat_irq_shutdown(struct irq_data *data) data->chip->shutdown(data->irq); } +static unsigned int compat_irq_startup(struct irq_data *data) +{ + return data->chip->startup(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -390,6 +395,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_disable = compat_irq_disable; if (chip->shutdown) chip->irq_shutdown = compat_irq_shutdown; + if (chip->startup) + chip->irq_startup = compat_irq_startup; /* * The real defaults @@ -398,8 +405,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_enable = default_enable; if (!chip->irq_disable) chip->irq_disable = default_disable; - if (!chip->startup) - chip->startup = default_startup; + if (!chip->irq_startup) + chip->irq_startup = default_startup; /* * We use chip->irq_disable, when the user provided its own. When * we have default_disable set for chip->irq_disable, then we need @@ -786,7 +793,7 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, desc->status &= ~IRQ_DISABLED; desc->status |= IRQ_NOREQUEST | IRQ_NOPROBE; desc->depth = 0; - desc->irq_data.chip->startup(irq); + desc->irq_data.chip->irq_startup(&desc->irq_data); } raw_spin_unlock_irqrestore(&desc->lock, flags); chip_bus_sync_unlock(desc); diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 60e25c4..8d0697f 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -311,11 +311,6 @@ static unsigned int noop_ret(struct irq_data *data) static void compat_noop(unsigned int irq) { } -static unsigned int compat_noop_ret(unsigned int irq) -{ - return 0; -} - /* * Generic no controller implementation */ @@ -326,7 +321,6 @@ struct irq_chip no_irq_chip = { .irq_enable = noop, .irq_disable = noop, .irq_ack = ack_bad, - .startup = compat_noop_ret, .end = compat_noop, }; @@ -343,7 +337,6 @@ struct irq_chip dummy_irq_chip = { .irq_ack = noop, .irq_mask = noop, .irq_unmask = noop, - .startup = compat_noop_ret, .end = compat_noop, }; diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index f3f36f6..31d7678 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -779,7 +779,7 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) if (!(desc->status & IRQ_NOAUTOEN)) { desc->depth = 0; desc->status &= ~IRQ_DISABLED; - desc->irq_data.chip->startup(irq); + desc->irq_data.chip->irq_startup(&desc->irq_data); } else /* Undo nested disables: */ desc->depth = 1; -- cgit v0.10.2 From c96b3b3c448592a0b87ef20306deb8b1fb4878c7 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:45:41 +0000 Subject: genirq: Provide compat handling for chip->set_affinity() Wrap the old chip function set_affinity() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.732894108@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index cce85f0..df51792 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -370,6 +370,12 @@ static unsigned int compat_irq_startup(struct irq_data *data) return data->chip->startup(data->irq); } +static int compat_irq_set_affinity(struct irq_data *data, + const struct cpumask *dest, bool force) +{ + return data->chip->set_affinity(data->irq, dest); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -436,6 +442,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_mask_ack = compat_irq_mask_ack; if (chip->eoi) chip->irq_eoi = compat_irq_eoi; + if (chip->set_affinity) + chip->irq_set_affinity = compat_irq_set_affinity; } static inline void mask_ack_irq(struct irq_desc *desc) diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 31d7678..305a60f 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -74,7 +74,7 @@ int irq_can_set_affinity(unsigned int irq) struct irq_desc *desc = irq_to_desc(irq); if (CHECK_IRQ_PER_CPU(desc->status) || !desc->irq_data.chip || - !desc->irq_data.chip->set_affinity) + !desc->irq_data.chip->irq_set_affinity) return 0; return 1; @@ -109,16 +109,17 @@ void irq_set_thread_affinity(struct irq_desc *desc) int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) { struct irq_desc *desc = irq_to_desc(irq); + struct irq_chip *chip = desc->irq_data.chip; unsigned long flags; - if (!desc->irq_data.chip->set_affinity) + if (!chip->irq_set_affinity) return -EINVAL; raw_spin_lock_irqsave(&desc->lock, flags); #ifdef CONFIG_GENERIC_PENDING_IRQ if (desc->status & IRQ_MOVE_PCNTXT) { - if (!desc->irq_data.chip->set_affinity(irq, cpumask)) { + if (!chip->irq_set_affinity(&desc->irq_data, cpumask, false)) { cpumask_copy(desc->irq_data.affinity, cpumask); irq_set_thread_affinity(desc); } @@ -128,7 +129,7 @@ int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) cpumask_copy(desc->pending_mask, cpumask); } #else - if (!desc->irq_data.chip->set_affinity(irq, cpumask)) { + if (!chip->irq_set_affinity(&desc->irq_data, cpumask, false)) { cpumask_copy(desc->irq_data.affinity, cpumask); irq_set_thread_affinity(desc); } @@ -177,7 +178,7 @@ static int setup_affinity(unsigned int irq, struct irq_desc *desc) cpumask_and(desc->irq_data.affinity, cpu_online_mask, irq_default_affinity); set_affinity: - desc->irq_data.chip->set_affinity(irq, desc->irq_data.affinity); + desc->irq_data.chip->irq_set_affinity(&desc->irq_data, desc->irq_data.affinity, false); return 0; } diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index 7888e5d..1d25419 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -7,6 +7,7 @@ void move_masked_irq(int irq) { struct irq_desc *desc = irq_to_desc(irq); + struct irq_chip *chip = desc->irq_data.chip; if (likely(!(desc->status & IRQ_MOVE_PENDING))) return; @@ -24,7 +25,7 @@ void move_masked_irq(int irq) if (unlikely(cpumask_empty(desc->pending_mask))) return; - if (!desc->irq_data.chip->set_affinity) + if (!chip->irq_set_affinity) return; assert_raw_spin_locked(&desc->lock); @@ -43,7 +44,8 @@ void move_masked_irq(int irq) */ if (likely(cpumask_any_and(desc->pending_mask, cpu_online_mask) < nr_cpu_ids)) - if (!desc->irq_data.chip->set_affinity(irq, desc->pending_mask)) { + if (!chip->irq_set_affinity(&desc->irq_data, + desc->pending_mask, false)) { cpumask_copy(desc->irq_data.affinity, desc->pending_mask); irq_set_thread_affinity(desc); } diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index 9b0da94..d9fddf9 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -65,7 +65,7 @@ static ssize_t irq_affinity_proc_write(struct file *file, cpumask_var_t new_value; int err; - if (!irq_to_desc(irq)->irq_data.chip->set_affinity || no_irq_affinity || + if (!irq_to_desc(irq)->irq_data.chip->irq_set_affinity || no_irq_affinity || irq_balancing_disabled(irq)) return -EIO; -- cgit v0.10.2 From b2ba2c30033c10cca2454f8b44bf98f5249e61c6 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:45:47 +0000 Subject: genirq: Provide compat handling for chip->set_type() Wrap the old chip function set_type() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.832261548@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/autoprobe.c b/kernel/irq/autoprobe.c index 7a46825..505798f 100644 --- a/kernel/irq/autoprobe.c +++ b/kernel/irq/autoprobe.c @@ -57,8 +57,9 @@ unsigned long probe_irq_on(void) * Some chips need to know about probing in * progress: */ - if (desc->irq_data.chip->set_type) - desc->irq_data.chip->set_type(i, IRQ_TYPE_PROBE); + if (desc->irq_data.chip->irq_set_type) + desc->irq_data.chip->irq_set_type(&desc->irq_data, + IRQ_TYPE_PROBE); desc->irq_data.chip->irq_startup(&desc->irq_data); } raw_spin_unlock_irq(&desc->lock); diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index df51792..b7dd02a 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -376,6 +376,11 @@ static int compat_irq_set_affinity(struct irq_data *data, return data->chip->set_affinity(data->irq, dest); } +static int compat_irq_set_type(struct irq_data *data, unsigned int type) +{ + return data->chip->set_type(data->irq, type); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -444,6 +449,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_eoi = compat_irq_eoi; if (chip->set_affinity) chip->irq_set_affinity = compat_irq_set_affinity; + if (chip->set_type) + chip->irq_set_type = compat_irq_set_type; } static inline void mask_ack_irq(struct irq_desc *desc) diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 305a60f..3618362 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -430,12 +430,12 @@ void compat_irq_chip_set_default_handler(struct irq_desc *desc) } int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, - unsigned long flags) + unsigned long flags) { int ret; struct irq_chip *chip = desc->irq_data.chip; - if (!chip || !chip->set_type) { + if (!chip || !chip->irq_set_type) { /* * IRQF_TRIGGER_* but the PIC does not support multiple * flow-types? @@ -446,11 +446,11 @@ int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, } /* caller masked out all except trigger mode flags */ - ret = chip->set_type(irq, flags); + ret = chip->irq_set_type(&desc->irq_data, flags); if (ret) - pr_err("setting trigger mode %d for irq %u failed (%pF)\n", - (int)flags, irq, chip->set_type); + pr_err("setting trigger mode %lu for irq %u failed (%pF)\n", + flags, irq, chip->irq_set_type); else { if (flags & (IRQ_TYPE_LEVEL_LOW | IRQ_TYPE_LEVEL_HIGH)) flags |= IRQ_LEVEL; -- cgit v0.10.2 From 2f7e99bb9be6a2d8d7b808dc86037710cc8b7bf1 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:45:50 +0000 Subject: genirq: Provide compat handling for chip->set_wake() Wrap the old chip function set_wake() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121842.927527393@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index b7dd02a..8775dd3 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -381,6 +381,11 @@ static int compat_irq_set_type(struct irq_data *data, unsigned int type) return data->chip->set_type(data->irq, type); } +static int compat_irq_set_wake(struct irq_data *data, unsigned int on) +{ + return data->chip->set_wake(data->irq, on); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -451,6 +456,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_set_affinity = compat_irq_set_affinity; if (chip->set_type) chip->irq_set_type = compat_irq_set_type; + if (chip->set_wake) + chip->irq_set_wake = compat_irq_set_wake; } static inline void mask_ack_irq(struct irq_desc *desc) diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 3618362..644e8d5 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -337,8 +337,8 @@ static int set_irq_wake_real(unsigned int irq, unsigned int on) struct irq_desc *desc = irq_to_desc(irq); int ret = -ENXIO; - if (desc->irq_data.chip->set_wake) - ret = desc->irq_data.chip->set_wake(irq, on); + if (desc->irq_data.chip->irq_set_wake) + ret = desc->irq_data.chip->irq_set_wake(&desc->irq_data, on); return ret; } -- cgit v0.10.2 From 21e2b8c62cca8f7dbec0c8c131ca1637e4a5670f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 12:45:53 +0000 Subject: genirq: Provide compat handling for chip->retrigger() Wrap the old chip function retrigger() until the migration is complete and the old chip functions are removed. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra LKML-Reference: <20100927121843.025801092@linutronix.de> Reviewed-by: H. Peter Anvin Reviewed-by: Ingo Molnar diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 8775dd3..f2c4d28 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -386,6 +386,11 @@ static int compat_irq_set_wake(struct irq_data *data, unsigned int on) return data->chip->set_wake(data->irq, on); } +static int compat_irq_retrigger(struct irq_data *data) +{ + return data->chip->retrigger(data->irq); +} + static void compat_bus_lock(struct irq_data *data) { data->chip->bus_lock(data->irq); @@ -458,6 +463,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_set_type = compat_irq_set_type; if (chip->set_wake) chip->irq_set_wake = compat_irq_set_wake; + if (chip->retrigger) + chip->irq_retrigger = compat_irq_retrigger; } static inline void mask_ack_irq(struct irq_desc *desc) diff --git a/kernel/irq/resend.c b/kernel/irq/resend.c index a798a23..891115a 100644 --- a/kernel/irq/resend.c +++ b/kernel/irq/resend.c @@ -70,8 +70,8 @@ void check_irq_resend(struct irq_desc *desc, unsigned int irq) if ((status & (IRQ_LEVEL | IRQ_PENDING | IRQ_REPLAY)) == IRQ_PENDING) { desc->status = (status & ~IRQ_PENDING) | IRQ_REPLAY; - if (!desc->irq_data.chip->retrigger || - !desc->irq_data.chip->retrigger(irq)) { + if (!desc->irq_data.chip->irq_retrigger || + !desc->irq_data.chip->irq_retrigger(&desc->irq_data)) { #ifdef CONFIG_HARDIRQS_SW_RESEND /* Set it pending and activate the softirq: */ set_bit(irq, irqs_resend); -- cgit v0.10.2 From bd151412263a67b5321e9dd1d5b4bf6d96fdebf3 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 1 Oct 2010 15:17:14 +0200 Subject: genirq: Provide config option to disable deprecated code This option covers now the old chip functions and the irq_desc data fields which are moving to struct irq_data. More stuff will follow. Pretty handy for testing a conversion, whether something broke or not. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/include/linux/irq.h b/include/linux/irq.h index 0c83cbd..82ed823 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -155,6 +155,7 @@ struct irq_data { */ struct irq_chip { const char *name; +#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED unsigned int (*startup)(unsigned int irq); void (*shutdown)(unsigned int irq); void (*enable)(unsigned int irq); @@ -175,7 +176,7 @@ struct irq_chip { void (*bus_lock)(unsigned int irq); void (*bus_sync_unlock)(unsigned int irq); - +#endif unsigned int (*irq_startup)(struct irq_data *data); void (*irq_shutdown)(struct irq_data *data); void (*irq_enable)(struct irq_data *data); @@ -225,6 +226,9 @@ struct irq_2_iommu; */ struct irq_desc { +#ifdef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED + struct irq_data irq_data; +#else /* * This union will go away, once we fixed the direct access to * irq_desc all over the place. The direct fields are a 1:1 @@ -247,6 +251,8 @@ struct irq_desc { #endif }; }; +#endif + struct timer_rand_state *timer_rand_state; unsigned int *kstat_irqs; irq_flow_handler_t handle_irq; diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig index e0fc6cd..a42c019 100644 --- a/kernel/irq/Kconfig +++ b/kernel/irq/Kconfig @@ -12,6 +12,10 @@ config GENERIC_HARDIRQS config GENERIC_HARDIRQS_NO__DO_IRQ def_bool y +# Select this to disable the deprecated stuff +config GENERIC_HARDIRQS_NO_DEPRECATED + def_bool n + # Options selectable by the architecture code config HAVE_SPARSE_IRQ def_bool n diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index f2c4d28..3235479 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -324,6 +324,7 @@ static void default_shutdown(struct irq_data *data) desc->status |= IRQ_MASKED; } +#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED /* Temporary migration helpers */ static void compat_irq_mask(struct irq_data *data) { @@ -400,12 +401,14 @@ static void compat_bus_sync_unlock(struct irq_data *data) { data->chip->bus_sync_unlock(data->irq); } +#endif /* * Fixup enable/disable function pointers */ void irq_chip_set_defaults(struct irq_chip *chip) { +#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED /* * Compat fixup functions need to be before we set the * defaults for enable/disable/startup/shutdown @@ -418,7 +421,7 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_shutdown = compat_irq_shutdown; if (chip->startup) chip->irq_startup = compat_irq_startup; - +#endif /* * The real defaults */ @@ -437,6 +440,8 @@ void irq_chip_set_defaults(struct irq_chip *chip) if (!chip->irq_shutdown) chip->irq_shutdown = chip->irq_disable != default_disable ? chip->irq_disable : default_shutdown; + +#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED if (!chip->end) chip->end = dummy_irq_chip.end; @@ -465,6 +470,7 @@ void irq_chip_set_defaults(struct irq_chip *chip) chip->irq_set_wake = compat_irq_set_wake; if (chip->retrigger) chip->irq_retrigger = compat_irq_retrigger; +#endif } static inline void mask_ack_irq(struct irq_desc *desc) diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 8d0697f..3fcef37 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -309,7 +309,12 @@ static unsigned int noop_ret(struct irq_data *data) return 0; } +#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED static void compat_noop(unsigned int irq) { } +#define END_INIT .end = compat_noop +#else +#define END_INIT +#endif /* * Generic no controller implementation @@ -321,7 +326,7 @@ struct irq_chip no_irq_chip = { .irq_enable = noop, .irq_disable = noop, .irq_ack = ack_bad, - .end = compat_noop, + END_INIT }; /* @@ -337,7 +342,7 @@ struct irq_chip dummy_irq_chip = { .irq_ack = noop, .irq_mask = noop, .irq_unmask = noop, - .end = compat_noop, + END_INIT }; /* diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index ecafbfe..b905f0a 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -42,6 +42,16 @@ extern int irq_select_affinity_usr(unsigned int irq); extern void irq_set_thread_affinity(struct irq_desc *desc); +#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED +static inline void irq_end(unsigned int irq, struct irq_desc *desc) +{ + if (desc->irq_data.chip && desc->irq_data.chip->end) + desc->irq_data.chip->end(irq); +} +#else +static inline void irq_end(unsigned int irq, struct irq_desc *desc) { } +#endif + /* Inline functions for support of irq chips on slow busses */ static inline void chip_bus_lock(struct irq_desc *desc) { diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c index 9ee704d..3089d3b 100644 --- a/kernel/irq/spurious.c +++ b/kernel/irq/spurious.c @@ -14,6 +14,8 @@ #include #include +#include "internals.h" + static int irqfixup __read_mostly; #define POLL_SPURIOUS_IRQ_INTERVAL (HZ/10) @@ -78,8 +80,8 @@ static int try_one_irq(int irq, struct irq_desc *desc) * If we did actual work for the real IRQ line we must let the * IRQ controller clean up too */ - if (work && desc->irq_data.chip && desc->irq_data.chip->end) - desc->irq_data.chip->end(irq); + if (work) + irq_end(irq, desc); raw_spin_unlock(&desc->lock); return ok; -- cgit v0.10.2 From ff5ff6060bf880aac233e68dd666cbe9e39ec620 Mon Sep 17 00:00:00 2001 From: Arnaud Lacombe Date: Sun, 26 Sep 2010 16:22:03 -0400 Subject: kconfig: delay symbol direct dependency initialization This fixes the use-after-free and associated crash in kconfig introduced in commit 246cf9c26bf11f2bffbecea6e5bd222eee7b1df8. Signed-off-by: Arnaud Lacombe Acked-by: Catalin Marinas Signed-off-by: Michal Marek diff --git a/scripts/kconfig/expr.h b/scripts/kconfig/expr.h index 6ee2e4f..170459c 100644 --- a/scripts/kconfig/expr.h +++ b/scripts/kconfig/expr.h @@ -165,7 +165,6 @@ struct menu { struct symbol *sym; struct property *prompt; struct expr *dep; - struct expr *dir_dep; unsigned int flags; char *help; struct file *file; diff --git a/scripts/kconfig/menu.c b/scripts/kconfig/menu.c index 4fb5902..edda8b4 100644 --- a/scripts/kconfig/menu.c +++ b/scripts/kconfig/menu.c @@ -107,7 +107,6 @@ static struct expr *menu_check_dep(struct expr *e) void menu_add_dep(struct expr *dep) { current_entry->dep = expr_alloc_and(current_entry->dep, menu_check_dep(dep)); - current_entry->dir_dep = current_entry->dep; } void menu_set_type(int type) @@ -291,10 +290,6 @@ void menu_finalize(struct menu *parent) for (menu = parent->list; menu; menu = menu->next) menu_finalize(menu); } else if (sym) { - /* ignore inherited dependencies for dir_dep */ - sym->dir_dep.expr = expr_transform(expr_copy(parent->dir_dep)); - sym->dir_dep.expr = expr_eliminate_dups(sym->dir_dep.expr); - basedep = parent->prompt ? parent->prompt->visible.expr : NULL; basedep = expr_trans_compare(basedep, E_UNEQUAL, &symbol_no); basedep = expr_eliminate_dups(expr_transform(basedep)); @@ -325,6 +320,8 @@ void menu_finalize(struct menu *parent) parent->next = last_menu->next; last_menu->next = NULL; } + + sym->dir_dep.expr = parent->dep; } for (menu = parent->list; menu; menu = menu->next) { if (sym && sym_is_choice(sym) && -- cgit v0.10.2 From 475d92fc6e72cd123dc5dbb9e70cdb80b0cfdf2d Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Tue, 28 Sep 2010 14:02:02 +0100 Subject: ARM: 6416/1: errata: faulty hazard checking in the Store Buffer may lead to data corruption On the r2p0, r2p1 and r2p2 versions of the Cortex-A9, data corruption can occur under very rare conditions due to a store buffer optimisation. This workaround sets a bit in the diagnostic register of the Cortex-A9, disabling the optimisation and preventing the problem from occurring. Acked-by: Catalin Marinas Signed-off-by: Will Deacon Signed-off-by: Russell King diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 88c97bc..9c26ba7 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -1101,6 +1101,20 @@ config ARM_ERRATA_720789 invalidated are not, resulting in an incoherency in the system page tables. The workaround changes the TLB flushing routines to invalidate entries regardless of the ASID. + +config ARM_ERRATA_743622 + bool "ARM errata: Faulty hazard checking in the Store Buffer may lead to data corruption" + depends on CPU_V7 + help + This option enables the workaround for the 743622 Cortex-A9 + (r2p0..r2p2) erratum. Under very rare conditions, a faulty + optimisation in the Cortex-A9 Store Buffer may lead to data + corruption. This workaround sets a specific bit in the diagnostic + register of the Cortex-A9 which disables the Store Buffer + optimisation, preventing the defect from occurring. This has no + visible impact on the overall performance or power consumption of the + processor. + endmenu source "arch/arm/common/Kconfig" diff --git a/arch/arm/mm/proc-v7.S b/arch/arm/mm/proc-v7.S index 7563ff0..75619c5 100644 --- a/arch/arm/mm/proc-v7.S +++ b/arch/arm/mm/proc-v7.S @@ -253,6 +253,14 @@ __v7_setup: orreq r10, r10, #1 << 22 @ set bit #22 mcreq p15, 0, r10, c15, c0, 1 @ write diagnostic register #endif +#ifdef CONFIG_ARM_ERRATA_743622 + teq r6, #0x20 @ present in r2p0 + teqne r6, #0x21 @ present in r2p1 + teqne r6, #0x22 @ present in r2p2 + mrceq p15, 0, r10, c15, c0, 1 @ read diagnostic register + orreq r10, r10, #1 << 6 @ set bit #6 + mcreq p15, 0, r10, c15, c0, 1 @ write diagnostic register +#endif 3: mov r10, #0 #ifdef HARVARD_CACHE -- cgit v0.10.2 From 7f58217bb6d39edac68ae0988cdb96a520ad916f Mon Sep 17 00:00:00 2001 From: Santosh Shilimkar Date: Thu, 30 Sep 2010 09:02:17 +0100 Subject: ARM: 6419/1: mmu: Fix MT_MEMORY and MT_MEMORY_NONCACHED pte flags The commit f1a2481c0 sets up the default flags for MT_MEMORY and MT_MEMORY_NONCACHED memory types. L_PTE_USER flag is wrongly set as default for these entries so remove it. Also adding the 'L_PTE_WRITE' flag so that these pages become read-write instead of just being read-only [this stops them being exposed to userspace, which is the main concern here --rmk] Reported-by: Catalin Marinas Signed-off-by: Santosh Shilimkar Acked-by: Catalin Marinas Signed-off-by: Russell King diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c index 6a3a2d0..e8ed9dc 100644 --- a/arch/arm/mm/mmu.c +++ b/arch/arm/mm/mmu.c @@ -248,7 +248,7 @@ static struct mem_type mem_types[] = { }, [MT_MEMORY] = { .prot_pte = L_PTE_PRESENT | L_PTE_YOUNG | L_PTE_DIRTY | - L_PTE_USER | L_PTE_EXEC, + L_PTE_WRITE | L_PTE_EXEC, .prot_l1 = PMD_TYPE_TABLE, .prot_sect = PMD_TYPE_SECT | PMD_SECT_AP_WRITE, .domain = DOMAIN_KERNEL, @@ -259,7 +259,7 @@ static struct mem_type mem_types[] = { }, [MT_MEMORY_NONCACHED] = { .prot_pte = L_PTE_PRESENT | L_PTE_YOUNG | L_PTE_DIRTY | - L_PTE_USER | L_PTE_EXEC | L_PTE_MT_BUFFERABLE, + L_PTE_WRITE | L_PTE_EXEC | L_PTE_MT_BUFFERABLE, .prot_l1 = PMD_TYPE_TABLE, .prot_sect = PMD_TYPE_SECT | PMD_SECT_AP_WRITE, .domain = DOMAIN_KERNEL, -- cgit v0.10.2 From ccdf2e1bca8a45eaf89eb142dbed3551886413fe Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Mon, 27 Sep 2010 18:12:12 +0100 Subject: ARM: 6412/1: kprobes-decode: add support for MOVW instruction The MOVW instruction moves a 16-bit immediate into the bottom halfword of the destination register. This patch ensures that kprobes leaves the 16-bit immediate intact, rather than assume a 12-bit immediate and mask out the upper 4 bits. Acked-by: Nicolas Pitre Signed-off-by: Will Deacon Signed-off-by: Russell King diff --git a/arch/arm/kernel/kprobes-decode.c b/arch/arm/kernel/kprobes-decode.c index 8bccbfa..2c1f005 100644 --- a/arch/arm/kernel/kprobes-decode.c +++ b/arch/arm/kernel/kprobes-decode.c @@ -1162,11 +1162,12 @@ space_cccc_001x(kprobe_opcode_t insn, struct arch_specific_insn *asi) { /* * MSR : cccc 0011 0x10 xxxx xxxx xxxx xxxx xxxx - * Undef : cccc 0011 0x00 xxxx xxxx xxxx xxxx xxxx + * Undef : cccc 0011 0100 xxxx xxxx xxxx xxxx xxxx * ALU op with S bit and Rd == 15 : * cccc 001x xxx1 xxxx 1111 xxxx xxxx xxxx */ - if ((insn & 0x0f900000) == 0x03200000 || /* MSR & Undef */ + if ((insn & 0x0fb00000) == 0x03200000 || /* MSR */ + (insn & 0x0ff00000) == 0x03400000 || /* Undef */ (insn & 0x0e10f000) == 0x0210f000) /* ALU s-bit, R15 */ return INSN_REJECTED; @@ -1177,7 +1178,7 @@ space_cccc_001x(kprobe_opcode_t insn, struct arch_specific_insn *asi) * *S (bit 20) updates condition codes * ADC/SBC/RSC reads the C flag */ - insn &= 0xfff00fff; /* Rn = r0, Rd = r0 */ + insn &= 0xffff0fff; /* Rd = r0 */ asi->insn[0] = insn; asi->insn_handler = (insn & (1 << 20)) ? /* S-bit */ emulate_alu_imm_rwflags : emulate_alu_imm_rflags; -- cgit v0.10.2 From 366d4a43b1a7a861c356a0e407c4f03f454d42ea Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Mon, 4 Oct 2010 09:31:27 +0200 Subject: x86, cpu: Fix X86_FEATURE_NOPL ba0593bf553c450a03dbc5f8c1f0ff58b778a0c8 cleared the aforementioned cpuid bit only on 32-bit due to various problems with Virtual PC. This somehow got lost during the 32- + 64-bit merge so restore the feature bit on 64-bit. For that, set it explicitly for non-constant arguments of cpu_has(). Update comment for future reference. Signed-off-by: Borislav Petkov LKML-Reference: <20101004073127.GA20305@liondog.tnic> Cc: Ryan O'Neill Cc: Linus Torvalds Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index f2f9ac7..927e630 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -704,16 +704,21 @@ void __init early_cpu_init(void) } /* - * The NOPL instruction is supposed to exist on all CPUs with - * family >= 6; unfortunately, that's not true in practice because - * of early VIA chips and (more importantly) broken virtualizers that - * are not easy to detect. In the latter case it doesn't even *fail* - * reliably, so probing for it doesn't even work. Disable it completely + * The NOPL instruction is supposed to exist on all CPUs of family >= 6; + * unfortunately, that's not true in practice because of early VIA + * chips and (more importantly) broken virtualizers that are not easy + * to detect. In the latter case it doesn't even *fail* reliably, so + * probing for it doesn't even work. Disable it completely on 32-bit * unless we can find a reliable way to detect all the broken cases. + * Enable it explicitly on 64-bit for non-constant inputs of cpu_has(). */ static void __cpuinit detect_nopl(struct cpuinfo_x86 *c) { +#ifdef CONFIG_X86_32 clear_cpu_cap(c, X86_FEATURE_NOPL); +#else + set_cpu_cap(c, X86_FEATURE_NOPL); +#endif } static void __cpuinit generic_identify(struct cpuinfo_x86 *c) -- cgit v0.10.2 From eaa71b318c5ed0cd1ac3182a533471dc5edf372d Mon Sep 17 00:00:00 2001 From: "Gustavo F. Padovan" Date: Mon, 4 Oct 2010 19:28:52 -0300 Subject: Bluetooth: Disallow to change L2CAP_OPTIONS values when connected L2CAP doesn't permit change like MTU, FCS, TxWindow values while the connection is alive, we can only set that before the connection/configuration process. That can lead to bugs in the L2CAP operation. Signed-off-by: Gustavo F. Padovan diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index 44a8fb0..0b54b7d 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -1950,6 +1950,11 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, char __us switch (optname) { case L2CAP_OPTIONS: + if (sk->sk_state == BT_CONNECTED) { + err = -EINVAL; + break; + } + opts.imtu = l2cap_pi(sk)->imtu; opts.omtu = l2cap_pi(sk)->omtu; opts.flush_to = l2cap_pi(sk)->flush_to; -- cgit v0.10.2 From f6cd378372bff06093d72f978c0150eeed3ea201 Mon Sep 17 00:00:00 2001 From: Dmitry Torokhov Date: Mon, 4 Oct 2010 21:46:11 -0700 Subject: Input: wacom - fix runtime PM related deadlock When runtime PM is enabled by default for input devices, X hangs in wacom open: [] mutex_lock+0x1a/0x40 [] wacom_resume+0x3b/0x90 [wacom] [] usb_resume_interface+0xd2/0x190 [] usb_resume_both+0x6d/0x110 [] usb_runtime_resume+0x24/0x40 [] __pm_runtime_resume+0x26f/0x450 [] __pm_runtime_resume+0x1da/0x450 [] pm_runtime_resume+0x2a/0x50 [] usb_autopm_get_interface+0x26/0x60 [] wacom_open+0x36/0x90 [wacom] wacom_open() takes wacom->lock and calls usb_autopm_get_interface(), which in turn calls wacom_resume() which tries to acquire the lock again. The fix is to call usb_autopm_get_interface() first, before we take the lock. Since we do not do usb_autopm_put_interface() until wacom_close() is called runtime PM is effectively disabled for the driver, however changing it now would risk regressions so the complete fix will have to wait till the next merge window. Reported-by: Jiri Slaby Acked-by: Oliver Neukum Signed-off-by: Dmitry Torokhov diff --git a/drivers/input/tablet/wacom_sys.c b/drivers/input/tablet/wacom_sys.c index 42ba369..b35876e 100644 --- a/drivers/input/tablet/wacom_sys.c +++ b/drivers/input/tablet/wacom_sys.c @@ -103,27 +103,26 @@ static void wacom_sys_irq(struct urb *urb) static int wacom_open(struct input_dev *dev) { struct wacom *wacom = input_get_drvdata(dev); + int retval = 0; - mutex_lock(&wacom->lock); - - wacom->irq->dev = wacom->usbdev; - - if (usb_autopm_get_interface(wacom->intf) < 0) { - mutex_unlock(&wacom->lock); + if (usb_autopm_get_interface(wacom->intf) < 0) return -EIO; - } + + mutex_lock(&wacom->lock); if (usb_submit_urb(wacom->irq, GFP_KERNEL)) { - usb_autopm_put_interface(wacom->intf); - mutex_unlock(&wacom->lock); - return -EIO; + retval = -EIO; + goto out; } wacom->open = true; wacom->intf->needs_remote_wakeup = 1; +out: mutex_unlock(&wacom->lock); - return 0; + if (retval) + usb_autopm_put_interface(wacom->intf); + return retval; } static void wacom_close(struct input_dev *dev) @@ -135,6 +134,8 @@ static void wacom_close(struct input_dev *dev) wacom->open = false; wacom->intf->needs_remote_wakeup = 0; mutex_unlock(&wacom->lock); + + usb_autopm_put_interface(wacom->intf); } static int wacom_parse_hid(struct usb_interface *intf, struct hid_descriptor *hid_desc, -- cgit v0.10.2 From 4e18b3edf71f5d4ad653e3c2ff6560878e965f96 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Mon, 4 Oct 2010 02:28:36 +0000 Subject: cls_u32: signedness bug skb_headroom() is unsigned so "skb_headroom(skb) + toff" is also unsigned and can't be less than zero. This test was added in 66d50d25: "u32: negative offset fix" It was supposed to fix a regression. Signed-off-by: Dan Carpenter Signed-off-by: David S. Miller diff --git a/net/sched/cls_u32.c b/net/sched/cls_u32.c index 7416a5c..b0c2a82 100644 --- a/net/sched/cls_u32.c +++ b/net/sched/cls_u32.c @@ -137,7 +137,7 @@ next_knode: int toff = off + key->off + (off2 & key->offmask); __be32 *data, _data; - if (skb_headroom(skb) + toff < 0) + if (skb_headroom(skb) + toff > INT_MAX) goto out; data = skb_header_pointer(skb, toff, 4, &_data); -- cgit v0.10.2 From cdaf9a2f280b25dd2fb1e04da5d3899411766e1b Mon Sep 17 00:00:00 2001 From: Russell King Date: Tue, 5 Oct 2010 11:29:28 +0100 Subject: ARM: fix section mismatch warnings in Versatile Express WARNING: vmlinux.o(.text+0xbf30): Section mismatch in reference from the function v2m_timer_init() to the function .init.text:sp804_clocksource_init() The function v2m_timer_init() references the function __init sp804_clocksource_init(). This is often because v2m_timer_init lacks a __init annotation or the annotation of sp804_clocksource_init is wrong. WARNING: vmlinux.o(.text+0xbf3c): Section mismatch in reference from the function v2m_timer_init() to the function .init.text:sp804_clockevents_init() The function v2m_timer_init() references the function __init sp804_clockevents_init(). This is often because v2m_timer_init lacks a __init annotation or the annotation of sp804_clockevents_init is wrong. WARNING: vmlinux.o(.text+0xc524): Section mismatch in reference from the function ct_ca9x4_init() to the function .init.text:l2x0_init() The function ct_ca9x4_init() references the function __init l2x0_init(). This is often because ct_ca9x4_init lacks a __init annotation or the annotation of l2x0_init is wrong. WARNING: vmlinux.o(.text+0xc530): Section mismatch in reference from the function ct_ca9x4_init() to the function .init.text:clkdev_add_table() The function ct_ca9x4_init() references the function __init clkdev_add_table(). This is often because ct_ca9x4_init lacks a __init annotation or the annotation of clkdev_add_table is wrong. WARNING: vmlinux.o(.text+0xc578): Section mismatch in reference from the function ct_ca9x4_init() to the (unknown reference) .init.data:(unknown) The function ct_ca9x4_init() references the (unknown reference) __initdata (unknown). This is often because ct_ca9x4_init lacks a __initdata annotation or the annotation of (unknown) is wrong. Fix these by making ct_ca9x4_init() and v2m_timer_init() both __init. Signed-off-by: Russell King diff --git a/arch/arm/mach-vexpress/ct-ca9x4.c b/arch/arm/mach-vexpress/ct-ca9x4.c index efb1270..71fb173 100644 --- a/arch/arm/mach-vexpress/ct-ca9x4.c +++ b/arch/arm/mach-vexpress/ct-ca9x4.c @@ -68,7 +68,7 @@ static void __init ct_ca9x4_init_irq(void) } #if 0 -static void ct_ca9x4_timer_init(void) +static void __init ct_ca9x4_timer_init(void) { writel(0, MMIO_P2V(CT_CA9X4_TIMER0) + TIMER_CTRL); writel(0, MMIO_P2V(CT_CA9X4_TIMER1) + TIMER_CTRL); @@ -222,7 +222,7 @@ static struct platform_device pmu_device = { .resource = pmu_resources, }; -static void ct_ca9x4_init(void) +static void __init ct_ca9x4_init(void) { int i; diff --git a/arch/arm/mach-vexpress/v2m.c b/arch/arm/mach-vexpress/v2m.c index 817f0ad..7eaa232 100644 --- a/arch/arm/mach-vexpress/v2m.c +++ b/arch/arm/mach-vexpress/v2m.c @@ -48,7 +48,7 @@ void __init v2m_map_io(struct map_desc *tile, size_t num) } -static void v2m_timer_init(void) +static void __init v2m_timer_init(void) { writel(0, MMIO_P2V(V2M_TIMER0) + TIMER_CTRL); writel(0, MMIO_P2V(V2M_TIMER1) + TIMER_CTRL); -- cgit v0.10.2 From 0fa035e52373386381fcdfbd506e32eac0c662f0 Mon Sep 17 00:00:00 2001 From: Hari Kanigeri Date: Fri, 20 Aug 2010 13:50:18 +0000 Subject: omap: iommu-load cam register before flushing the entry The flush_iotlb_page is not loading the cam register before flushing the cam entry. This causes wrong entry to be flushed out from the TLB, and if the entry happens to be a locked TLB entry it would lead to MMU faults. The fix is to load the cam register with the address to be flushed before flushing the TLB entry. Signed-off-by: Hari Kanigeri Acked-by: Hiroshi DOYU Signed-off-by: Tony Lindgren diff --git a/arch/arm/plat-omap/iommu.c b/arch/arm/plat-omap/iommu.c index a202a2c..6cd151b 100644 --- a/arch/arm/plat-omap/iommu.c +++ b/arch/arm/plat-omap/iommu.c @@ -320,6 +320,7 @@ void flush_iotlb_page(struct iommu *obj, u32 da) if ((start <= da) && (da < start + bytes)) { dev_dbg(obj->dev, "%s: %08x<=%08x(%x)\n", __func__, start, da, bytes); + iotlb_load_cr(obj, &cr); iommu_write_reg(obj, 1, MMU_FLUSH_ENTRY); } } -- cgit v0.10.2 From 6230d18cc7c4c68b7a38ea73bf5910e7652e5b21 Mon Sep 17 00:00:00 2001 From: minskey guo Date: Fri, 17 Sep 2010 14:02:37 +0800 Subject: seqno mask of THM_ITV register is 16bit The mask of sequence number in THM_ITV register is 16bit width instead of 8bit. Signed-off-by: minskey guo Acked-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 9024480..fb31700 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -230,7 +230,7 @@ #define THM_TC2 0xac #define THM_DTV 0xb0 #define THM_ITV 0xd8 -#define ITV_ME_SEQNO_MASK 0x000f0000 /* ME should update every ~200ms */ +#define ITV_ME_SEQNO_MASK 0x00ff0000 /* ME should update every ~200ms */ #define ITV_ME_SEQNO_SHIFT (16) #define ITV_MCH_TEMP_MASK 0x0000ff00 #define ITV_MCH_TEMP_SHIFT (8) -- cgit v0.10.2 From c21eae4f7c38db0e4693fb4cb24fb42fb83d8c15 Mon Sep 17 00:00:00 2001 From: minskey guo Date: Fri, 17 Sep 2010 14:03:01 +0800 Subject: old_cpu_power is wrongly divided by 65535 in ips_monitor() The variable old_cpu_power is used to save the value of THM_CEC register. In get_cpu_power(), it will be divided by 65535. Signed-off-by: minskey guo Acked-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index fb31700..3c7b25c 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -948,7 +948,7 @@ static int ips_monitor(void *data) ITV_ME_SEQNO_SHIFT; seqno_timestamp = get_jiffies_64(); - old_cpu_power = thm_readl(THM_CEC) / 65535; + old_cpu_power = thm_readl(THM_CEC); schedule_timeout_interruptible(msecs_to_jiffies(IPS_SAMPLE_PERIOD)); /* Collect an initial average */ -- cgit v0.10.2 From fed522f7ea780d195d5d3e55df95fee520136e17 Mon Sep 17 00:00:00 2001 From: minskey guo Date: Fri, 17 Sep 2010 14:03:15 +0800 Subject: Release symbol on error-handling path of ips_get_i915_syms() In ips_get_i915_syms(), the symbol i915_gpu_busy() is not released when error occurs. Signed-off-by: minskey guo Acked-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 3c7b25c..71dcc41 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -1390,7 +1390,7 @@ static bool ips_get_i915_syms(struct ips_driver *ips) return true; out_put_busy: - symbol_put(i915_gpu_turbo_disable); + symbol_put(i915_gpu_busy); out_put_lower: symbol_put(i915_gpu_lower); out_put_raise: -- cgit v0.10.2 From a7abda8d721359363d679c5f2de964f29419568c Mon Sep 17 00:00:00 2001 From: minskey guo Date: Fri, 17 Sep 2010 14:03:27 +0800 Subject: NULL pointer might be used in ips_monitor() The patch is to create ips_adjust thread before ips_monitor begins to run because the latter will kthread_stop() or wake up the former via ips->adjust pointer. Without this change, it is possible that ips->adjust is NULL when kthread_stop() or wake_up_process() is called in ips_monitor(). Signed-off-by: minskey guo Acked-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 71dcc41..bfa9c72 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -940,7 +940,6 @@ static int ips_monitor(void *data) kfree(mch_samples); kfree(cpu_samples); kfree(mchp_samples); - kthread_stop(ips->adjust); return -ENOMEM; } @@ -1535,19 +1534,24 @@ static int ips_probe(struct pci_dev *dev, const struct pci_device_id *id) ips_enable_cpu_turbo(ips); ips->cpu_turbo_enabled = true; - /* Set up the work queue and monitor/adjust threads */ - ips->monitor = kthread_run(ips_monitor, ips, "ips-monitor"); - if (IS_ERR(ips->monitor)) { + /* Create thermal adjust thread */ + ips->adjust = kthread_create(ips_adjust, ips, "ips-adjust"); + if (IS_ERR(ips->adjust)) { dev_err(&dev->dev, - "failed to create thermal monitor thread, aborting\n"); + "failed to create thermal adjust thread, aborting\n"); ret = -ENOMEM; goto error_free_irq; + } - ips->adjust = kthread_create(ips_adjust, ips, "ips-adjust"); - if (IS_ERR(ips->adjust)) { + /* + * Set up the work queue and monitor thread. The monitor thread + * will wake up ips_adjust thread. + */ + ips->monitor = kthread_run(ips_monitor, ips, "ips-monitor"); + if (IS_ERR(ips->monitor)) { dev_err(&dev->dev, - "failed to create thermal adjust thread, aborting\n"); + "failed to create thermal monitor thread, aborting\n"); ret = -ENOMEM; goto error_thread_cleanup; } @@ -1566,7 +1570,7 @@ static int ips_probe(struct pci_dev *dev, const struct pci_device_id *id) return ret; error_thread_cleanup: - kthread_stop(ips->monitor); + kthread_stop(ips->adjust); error_free_irq: free_irq(ips->dev->irq, ips); error_unmap: -- cgit v0.10.2 From 354aeeb1ca8f82ea133ede21987034addc75057a Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Thu, 23 Sep 2010 23:49:28 +0200 Subject: IPS driver: don't toggle CPU turbo on unsupported CPUs If the CPU doesn't support turbo, don't try to enable/disable it. http://bugzilla.kernel.org/show_bug.cgi?id=18742 Signed-off-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index bfa9c72..71d04ef 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -51,7 +51,6 @@ * TODO: * - handle CPU hotplug * - provide turbo enable/disable api - * - make sure we can write turbo enable/disable reg based on MISC_EN * * Related documents: * - CDI 403777, 403778 - Auburndale EDS vol 1 & 2 @@ -325,6 +324,7 @@ struct ips_driver { bool gpu_preferred; bool poll_turbo_status; bool second_cpu; + bool turbo_toggle_allowed; struct ips_mcp_limits *limits; /* Optional MCH interfaces for if i915 is in use */ @@ -461,7 +461,8 @@ static void ips_enable_cpu_turbo(struct ips_driver *ips) if (ips->__cpu_turbo_on) return; - on_each_cpu(do_enable_cpu_turbo, ips, 1); + if (ips->turbo_toggle_allowed) + on_each_cpu(do_enable_cpu_turbo, ips, 1); ips->__cpu_turbo_on = true; } @@ -498,7 +499,8 @@ static void ips_disable_cpu_turbo(struct ips_driver *ips) if (!ips->__cpu_turbo_on) return; - on_each_cpu(do_disable_cpu_turbo, ips, 1); + if (ips->turbo_toggle_allowed) + on_each_cpu(do_disable_cpu_turbo, ips, 1); ips->__cpu_turbo_on = false; } @@ -1332,8 +1334,10 @@ static struct ips_mcp_limits *ips_detect_cpu(struct ips_driver *ips) * turbo manually or we'll get an illegal MSR access, even though * turbo will still be available. */ - if (!(misc_en & IA32_MISC_TURBO_EN)) - ; /* add turbo MSR write allowed flag if necessary */ + if (misc_en & IA32_MISC_TURBO_EN) + ips->turbo_toggle_allowed = true; + else + ips->turbo_toggle_allowed = false; if (strstr(boot_cpu_data.x86_model_id, "CPU M")) limits = &ips_sv_limits; -- cgit v0.10.2 From eceab272fb895148f6293b5c0644fc2dd36d3aff Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Thu, 23 Sep 2010 23:49:29 +0200 Subject: IPS driver: verify BIOS provided limits They're optional. If not present or sane, we should use the CPU defaults. Signed-off-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 71d04ef..c402cc4 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -665,6 +665,27 @@ static bool mch_exceeded(struct ips_driver *ips) } /** + * verify_limits - verify BIOS provided limits + * @ips: IPS structure + * + * BIOS can optionally provide non-default limits for power and temp. Check + * them here and use the defaults if the BIOS values are not provided or + * are otherwise unusable. + */ +static void verify_limits(struct ips_driver *ips) +{ + if (ips->mcp_power_limit < ips->limits->mcp_power_limit || + ips->mcp_power_limit > 35000) + ips->mcp_power_limit = ips->limits->mcp_power_limit; + + if (ips->mcp_temp_limit < ips->limits->core_temp_limit || + ips->mcp_temp_limit < ips->limits->mch_temp_limit || + ips->mcp_temp_limit > 150) + ips->mcp_temp_limit = min(ips->limits->core_temp_limit, + ips->limits->mch_temp_limit); +} + +/** * update_turbo_limits - get various limits & settings from regs * @ips: IPS driver struct * @@ -688,6 +709,7 @@ static void update_turbo_limits(struct ips_driver *ips) ips->mcp_temp_limit = thm_readw(THM_PTL); ips->mcp_power_limit = thm_readw(THM_MPPC); + verify_limits(ips); /* Ignore BIOS CPU vs GPU pref */ } @@ -1156,6 +1178,7 @@ static irqreturn_t ips_irq_handler(int irq, void *arg) STS_PTL_SHIFT; ips->mcp_power_limit = (tc1 & STS_PPL_MASK) >> STS_PPL_SHIFT; + verify_limits(ips); spin_unlock(&ips->turbo_status_lock); thm_writeb(THM_SEC, SEC_ACK); -- cgit v0.10.2 From a8c096adbd2b55942ff13c8bbc573a7551768003 Mon Sep 17 00:00:00 2001 From: Tim Gardner Date: Tue, 28 Sep 2010 14:58:15 -0600 Subject: intel_ips: Print MCP limit exceeded values. Print some interesting values when MCP limits are exceeded. Signed-off-by: Tim Gardner Cc: Matthew Garrett Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index c402cc4..07a1a4a 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -600,17 +600,29 @@ static bool mcp_exceeded(struct ips_driver *ips) { unsigned long flags; bool ret = false; + u32 temp_limit; + u32 avg_power; + const char *msg = "MCP limit exceeded: "; spin_lock_irqsave(&ips->turbo_status_lock, flags); - if (ips->mcp_avg_temp > (ips->mcp_temp_limit * 100)) - ret = true; - if (ips->cpu_avg_power + ips->mch_avg_power > ips->mcp_power_limit) + + temp_limit = ips->mcp_temp_limit * 100; + if (ips->mcp_avg_temp > temp_limit) { + dev_info(&ips->dev->dev, + "%sAvg temp %u, limit %u\n", msg, ips->mcp_avg_temp, + temp_limit); ret = true; - spin_unlock_irqrestore(&ips->turbo_status_lock, flags); + } - if (ret) + avg_power = ips->cpu_avg_power + ips->mch_avg_power; + if (avg_power > ips->mcp_power_limit) { dev_info(&ips->dev->dev, - "MCP power or thermal limit exceeded\n"); + "%sAvg power %u, limit %u\n", msg, avg_power, + ips->mcp_power_limit); + ret = true; + } + + spin_unlock_irqrestore(&ips->turbo_status_lock, flags); return ret; } -- cgit v0.10.2 From 070c0ee1ef9f5550cac9247190f0289349f28c01 Mon Sep 17 00:00:00 2001 From: Andy Whitcroft Date: Tue, 5 Oct 2010 09:48:42 +0100 Subject: intel_ips -- ensure we do not enable gpu turbo mode without driver linkage Both when polling the current turbo status (in poll_turbo_status mode) and when handling thermal events (in ips_irq_handler) the current status of GPU turbo is updated to match the hardware status. However if during driver initialisation we were unable aquire linkage to the i915 driver enabling GPU turbo will lead to an oops on the first attempt to determine GPU busy status. Ensure that we do not enable GPU turbo unless we have driver linkage. BugLink: http://bugs.launchpad.net/bugs/632430 Cc: stable@kernel.org Signed-off-by: Andy Whitcroft Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 07a1a4a..76e41dc 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -715,7 +715,8 @@ static void update_turbo_limits(struct ips_driver *ips) u32 hts = thm_readl(THM_HTS); ips->cpu_turbo_enabled = !(hts & HTS_PCTD_DIS); - ips->gpu_turbo_enabled = !(hts & HTS_GTD_DIS); + if (ips->gpu_busy) + ips->gpu_turbo_enabled = !(hts & HTS_GTD_DIS); ips->core_power_limit = thm_readw(THM_MPCPC); ips->mch_power_limit = thm_readw(THM_MMGPC); ips->mcp_temp_limit = thm_readw(THM_PTL); @@ -1185,7 +1186,8 @@ static irqreturn_t ips_irq_handler(int irq, void *arg) STS_GPL_SHIFT; /* ignore EC CPU vs GPU pref */ ips->cpu_turbo_enabled = !(sts & STS_PCTD_DIS); - ips->gpu_turbo_enabled = !(sts & STS_GTD_DIS); + if (ips->gpu_busy) + ips->gpu_turbo_enabled = !(sts & STS_GTD_DIS); ips->mcp_temp_limit = (sts & STS_PTL_MASK) >> STS_PTL_SHIFT; ips->mcp_power_limit = (tc1 & STS_PPL_MASK) >> -- cgit v0.10.2 From 4fd07ac00d87b942cc8d8f30a27192fea2fc4ab2 Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Tue, 5 Oct 2010 11:26:22 -0700 Subject: IPS driver: apply BIOS provided CPU limit if different from default The BIOS may hand us a lower CPU power limit than the default for a given SKU. We should use it in case the platform isn't designed to dissapate the full TDP of a given part. Signed-off-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 76e41dc..50ab1c9 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -1391,9 +1391,10 @@ static struct ips_mcp_limits *ips_detect_cpu(struct ips_driver *ips) tdp = turbo_power & TURBO_TDP_MASK; /* Sanity check TDP against CPU */ - if (limits->mcp_power_limit != (tdp / 8) * 1000) { - dev_warn(&ips->dev->dev, "Warning: CPU TDP doesn't match expected value (found %d, expected %d)\n", - tdp / 8, limits->mcp_power_limit / 1000); + if (limits->core_power_limit != (tdp / 8) * 1000) { + dev_info(&ips->dev->dev, "CPU TDP doesn't match expected value (found %d, expected %d)\n", + tdp / 8, limits->core_power_limit / 1000); + limits->core_power_limit = (tdp / 8) * 1000; } out: -- cgit v0.10.2 From 96f3823f537088c13735cfdfbf284436c802352a Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Tue, 5 Oct 2010 14:50:59 -0400 Subject: [PATCH 2/2] IPS driver: disable CPU turbo The undocumented interface we're using for reading CPU power seems to be overreporting power. Until we figure out how to correct it, disable CPU turbo and power reporting to be safe. This will keep the CPU within default limits and still allow us to increase GPU frequency as needed. Signed-off-by: Jesse Barnes Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 50ab1c9..4b3ecde 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -715,8 +715,15 @@ static void update_turbo_limits(struct ips_driver *ips) u32 hts = thm_readl(THM_HTS); ips->cpu_turbo_enabled = !(hts & HTS_PCTD_DIS); + /* + * Disable turbo for now, until we can figure out why the power figures + * are wrong + */ + ips->cpu_turbo_enabled = false; + if (ips->gpu_busy) ips->gpu_turbo_enabled = !(hts & HTS_GTD_DIS); + ips->core_power_limit = thm_readw(THM_MPCPC); ips->mch_power_limit = thm_readw(THM_MMGPC); ips->mcp_temp_limit = thm_readw(THM_PTL); @@ -895,7 +902,7 @@ static u32 get_cpu_power(struct ips_driver *ips, u32 *last, int period) ret = (ret * 1000) / 65535; *last = val; - return ret; + return 0; } static const u16 temp_decay_factor = 2; @@ -1186,6 +1193,11 @@ static irqreturn_t ips_irq_handler(int irq, void *arg) STS_GPL_SHIFT; /* ignore EC CPU vs GPU pref */ ips->cpu_turbo_enabled = !(sts & STS_PCTD_DIS); + /* + * Disable turbo for now, until we can figure + * out why the power figures are wrong + */ + ips->cpu_turbo_enabled = false; if (ips->gpu_busy) ips->gpu_turbo_enabled = !(sts & STS_GTD_DIS); ips->mcp_temp_limit = (sts & STS_PTL_MASK) >> @@ -1573,8 +1585,8 @@ static int ips_probe(struct pci_dev *dev, const struct pci_device_id *id) /* Save turbo limits & ratios */ rdmsrl(TURBO_POWER_CURRENT_LIMIT, ips->orig_turbo_limit); - ips_enable_cpu_turbo(ips); - ips->cpu_turbo_enabled = true; + ips_disable_cpu_turbo(ips); + ips->cpu_turbo_enabled = false; /* Create thermal adjust thread */ ips->adjust = kthread_create(ips_adjust, ips, "ips-adjust"); -- cgit v0.10.2 From d24a9da573444ab4aff38af2f4a0da07408ff491 Mon Sep 17 00:00:00 2001 From: Matthew Garrett Date: Tue, 5 Oct 2010 14:54:06 -0400 Subject: IPS driver: Fix limit clamping when reducing CPU power Values here are in internal units rather than Watts, so we shouldn't perform any conversion. Signed-off-by: Matthew Garrett diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index 4b3ecde..c44a5e8 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -415,7 +415,7 @@ static void ips_cpu_lower(struct ips_driver *ips) new_limit = cur_limit - 8; /* 1W decrease */ /* Clamp to SKU TDP limit */ - if (((new_limit * 10) / 8) < (ips->orig_turbo_limit & TURBO_TDP_MASK)) + if (new_limit < (ips->orig_turbo_limit & TURBO_TDP_MASK)) new_limit = ips->orig_turbo_limit & TURBO_TDP_MASK; thm_writew(THM_MPCPC, (new_limit * 10) / 8); -- cgit v0.10.2 From c4b8c01112a16f3cb8a4bffb11b89f2df079f904 Mon Sep 17 00:00:00 2001 From: Jeff Kirsher Date: Tue, 5 Oct 2010 01:16:44 +0000 Subject: ixgbevf.txt: Update ixgbevf documentation Update the documentation for the ixgbevf (ixgbe virtual function driver). Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/Documentation/networking/ixgbevf.txt b/Documentation/networking/ixgbevf.txt old mode 100755 new mode 100644 index 19015de..21dd5d1 --- a/Documentation/networking/ixgbevf.txt +++ b/Documentation/networking/ixgbevf.txt @@ -1,19 +1,16 @@ Linux* Base Driver for Intel(R) Network Connection ================================================== -November 24, 2009 +Intel Gigabit Linux driver. +Copyright(c) 1999 - 2010 Intel Corporation. Contents ======== -- In This Release - Identifying Your Adapter - Known Issues/Troubleshooting - Support -In This Release -=============== - This file describes the ixgbevf Linux* Base Driver for Intel Network Connection. @@ -33,7 +30,7 @@ Identifying Your Adapter For more information on how to identify your adapter, go to the Adapter & Driver ID Guide at: - http://support.intel.com/support/network/sb/CS-008441.htm + http://support.intel.com/support/go/network/adapter/idguide.htm Known Issues/Troubleshooting ============================ @@ -57,34 +54,3 @@ or the Intel Wired Networking project hosted by Sourceforge at: If an issue is identified with the released source code on the supported kernel with a supported adapter, email the specific information related to the issue to e1000-devel@lists.sf.net - -License -======= - -Intel 10 Gigabit Linux driver. -Copyright(c) 1999 - 2009 Intel Corporation. - -This program is free software; you can redistribute it and/or modify it -under the terms and conditions of the GNU General Public License, -version 2, as published by the Free Software Foundation. - -This program is distributed in the hope it will be useful, but WITHOUT -ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or -FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for -more details. - -You should have received a copy of the GNU General Public License along with -this program; if not, write to the Free Software Foundation, Inc., -51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. - -The full GNU General Public License is included in this distribution in -the file called "COPYING". - -Trademarks -========== - -Intel, Itanium, and Pentium are trademarks or registered trademarks of -Intel Corporation or its subsidiaries in the United States and other -countries. - -* Other names and brands may be claimed as the property of others. -- cgit v0.10.2 From 2bff89c3f340776398bfaf6c94404ffcd09f6e77 Mon Sep 17 00:00:00 2001 From: Jeff Kirsher Date: Tue, 5 Oct 2010 01:17:05 +0000 Subject: e1000.txt: Update e1000 documentation Updated the e1000 networking driver documentation. Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/Documentation/networking/e1000.txt b/Documentation/networking/e1000.txt index 2df7186..d9271e7 100644 --- a/Documentation/networking/e1000.txt +++ b/Documentation/networking/e1000.txt @@ -1,82 +1,35 @@ Linux* Base Driver for the Intel(R) PRO/1000 Family of Adapters =============================================================== -September 26, 2006 - +Intel Gigabit Linux driver. +Copyright(c) 1999 - 2010 Intel Corporation. Contents ======== -- In This Release - Identifying Your Adapter -- Building and Installation - Command Line Parameters - Speed and Duplex Configuration - Additional Configurations -- Known Issues - Support - -In This Release -=============== - -This file describes the Linux* Base Driver for the Intel(R) PRO/1000 Family -of Adapters. This driver includes support for Itanium(R)2-based systems. - -For questions related to hardware requirements, refer to the documentation -supplied with your Intel PRO/1000 adapter. All hardware requirements listed -apply to use with Linux. - -The following features are now available in supported kernels: - - Native VLANs - - Channel Bonding (teaming) - - SNMP - -Channel Bonding documentation can be found in the Linux kernel source: -/Documentation/networking/bonding.txt - -The driver information previously displayed in the /proc filesystem is not -supported in this release. Alternatively, you can use ethtool (version 1.6 -or later), lspci, and ifconfig to obtain the same information. - -Instructions on updating ethtool can be found in the section "Additional -Configurations" later in this document. - -NOTE: The Intel(R) 82562v 10/100 Network Connection only provides 10/100 -support. - - Identifying Your Adapter ======================== For more information on how to identify your adapter, go to the Adapter & Driver ID Guide at: - http://support.intel.com/support/network/adapter/pro100/21397.htm + http://support.intel.com/support/go/network/adapter/idguide.htm For the latest Intel network drivers for Linux, refer to the following website. In the search field, enter your adapter name or type, or use the networking link on the left to search for your adapter: - http://downloadfinder.intel.com/scripts-df/support_intel.asp - + http://support.intel.com/support/go/network/adapter/home.htm Command Line Parameters ======================= -If the driver is built as a module, the following optional parameters -are used by entering them on the command line with the modprobe command -using this syntax: - - modprobe e1000 [ @@ -217,6 +222,7 @@ handle_level_irq handle_edge_irq + handle_fasteoi_irq handle_simple_irq handle_percpu_irq @@ -233,33 +239,33 @@ are used by the default flow implementations. The following helper functions are implemented (simplified excerpt): -default_enable(irq) +default_enable(struct irq_data *data) { - desc->chip->unmask(irq); + desc->chip->irq_unmask(data); } -default_disable(irq) +default_disable(struct irq_data *data) { - if (!delay_disable(irq)) - desc->chip->mask(irq); + if (!delay_disable(data)) + desc->chip->irq_mask(data); } -default_ack(irq) +default_ack(struct irq_data *data) { - chip->ack(irq); + chip->irq_ack(data); } -default_mask_ack(irq) +default_mask_ack(struct irq_data *data) { - if (chip->mask_ack) { - chip->mask_ack(irq); + if (chip->irq_mask_ack) { + chip->irq_mask_ack(data); } else { - chip->mask(irq); - chip->ack(irq); + chip->irq_mask(data); + chip->irq_ack(data); } } -noop(irq) +noop(struct irq_data *data)) { } @@ -278,12 +284,27 @@ noop(irq) The following control flow is implemented (simplified excerpt): -desc->chip->start(); +desc->chip->irq_mask(); handle_IRQ_event(desc->action); -desc->chip->end(); +desc->chip->irq_unmask(); - + + + Default Fast EOI IRQ flow handler + + handle_fasteoi_irq provides a generic implementation + for interrupts, which only need an EOI at the end of + the handler + + + The following control flow is implemented (simplified excerpt): + +handle_IRQ_event(desc->action); +desc->chip->irq_eoi(); + + + Default Edge IRQ flow handler @@ -294,20 +315,19 @@ desc->chip->end(); The following control flow is implemented (simplified excerpt): if (desc->status & running) { - desc->chip->hold(); + desc->chip->irq_mask(); desc->status |= pending | masked; return; } -desc->chip->start(); +desc->chip->irq_ack(); desc->status |= running; do { if (desc->status & masked) - desc->chip->enable(); + desc->chip->irq_unmask(); desc->status &= ~pending; handle_IRQ_event(desc->action); } while (status & pending); desc->status &= ~running; -desc->chip->end(); @@ -342,9 +362,9 @@ handle_IRQ_event(desc->action); The following control flow is implemented (simplified excerpt): -desc->chip->start(); handle_IRQ_event(desc->action); -desc->chip->end(); +if (desc->chip->irq_eoi) + desc->chip->irq_eoi(); @@ -375,8 +395,7 @@ desc->chip->end(); mechanism. (It's necessary to enable CONFIG_HARDIRQS_SW_RESEND when you want to use the delayed interrupt disable feature and your hardware is not capable of retriggering an interrupt.) - The delayed interrupt disable can be runtime enabled, per interrupt, - by setting the IRQ_DELAYED_DISABLE flag in the irq_desc status field. + The delayed interrupt disable is not configurable. @@ -387,13 +406,13 @@ desc->chip->end(); contains all the direct chip relevant functions, which can be utilized by the irq flow implementations. - ack() - mask_ack() - Optional, recommended for performance - mask() - unmask() - retrigger() - Optional - set_type() - Optional - set_wake() - Optional + irq_ack() + irq_mask_ack() - Optional, recommended for performance + irq_mask() + irq_unmask() + irq_retrigger() - Optional + irq_set_type() - Optional + irq_set_wake() - Optional These primitives are strictly intended to mean what they say: ack means ACK, masking means masking of an IRQ line, etc. It is up to the flow @@ -458,6 +477,7 @@ desc->chip->end(); This chapter contains the autogenerated documentation of the internal functions. +!Ikernel/irq/irqdesc.c !Ikernel/irq/handle.c !Ikernel/irq/chip.c -- cgit v0.10.2 From aa99ec0f3f26bf2bcd0fa5176de93598427f1e5e Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 20:02:56 +0200 Subject: genirq: Use sane sparse allocator Make irq_to_desc_alloc_node() a wrapper around the new allocator. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index 35d9052..7cbe4f9 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -51,7 +51,7 @@ static int alloc_masks(struct irq_desc *desc, gfp_t gfp, int node) static void desc_smp_init(struct irq_desc *desc, int node) { - desc->node = node; + desc->irq_data.node = node; cpumask_copy(desc->irq_data.affinity, irq_default_affinity); } @@ -84,13 +84,6 @@ static DECLARE_BITMAP(allocated_irqs, NR_IRQS); #ifdef CONFIG_SPARSE_IRQ -static struct irq_desc irq_desc_init = { - .status = IRQ_DEFAULT_INIT_FLAGS, - .handle_irq = handle_bad_irq, - .depth = 1, - .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc_init.lock), -}; - void __ref init_kstat_irqs(struct irq_desc *desc, int node, int nr) { void *ptr; @@ -108,29 +101,6 @@ void __ref init_kstat_irqs(struct irq_desc *desc, int node, int nr) } } -static void init_one_irq_desc(int irq, struct irq_desc *desc, int node) -{ - memcpy(desc, &irq_desc_init, sizeof(struct irq_desc)); - - raw_spin_lock_init(&desc->lock); - desc->irq_data.irq = irq; -#ifdef CONFIG_SMP - desc->irq_data.node = node; -#endif - lockdep_set_class(&desc->lock, &irq_desc_lock_class); - init_kstat_irqs(desc, node, nr_cpu_ids); - if (!desc->kstat_irqs) { - printk(KERN_ERR "can not alloc kstat_irqs\n"); - BUG_ON(1); - } - if (!alloc_desc_masks(desc, node, false)) { - printk(KERN_ERR "can not alloc irq_desc cpumasks\n"); - BUG_ON(1); - } - init_desc_masks(desc); - arch_init_chip_data(desc, node); -} - static RADIX_TREE(irq_desc_tree, GFP_ATOMIC); static void irq_insert_desc(unsigned int irq, struct irq_desc *desc) @@ -171,8 +141,9 @@ static inline void free_masks(struct irq_desc *desc) { } static struct irq_desc *alloc_desc(int irq, int node) { + /* Temporary hack until we can switch to GFP_KERNEL */ + gfp_t gfp = gfp_allowed_mask == GFP_BOOT_MASK ? GFP_NOWAIT : GFP_ATOMIC; struct irq_desc *desc; - gfp_t gfp = GFP_KERNEL; desc = kzalloc_node(sizeof(*desc), gfp, node); if (!desc) @@ -226,6 +197,8 @@ static int alloc_descs(unsigned int start, unsigned int cnt, int node) desc = alloc_desc(start + i, node); if (!desc) goto err; + /* temporary until I fixed x86 madness */ + arch_init_chip_data(desc, node); raw_spin_lock_irqsave(&sparse_irq_lock, flags); irq_insert_desc(start + i, desc); raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); @@ -242,23 +215,19 @@ err: return -ENOMEM; } -static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_smp = { - [0 ... NR_IRQS_LEGACY-1] = { - .status = IRQ_DEFAULT_INIT_FLAGS, - .handle_irq = handle_bad_irq, - .depth = 1, - .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc_init.lock), - } -}; +struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) +{ + int res = irq_alloc_descs(irq, irq, 1, node); -static unsigned int *kstat_irqs_legacy; + if (res == -EEXIST || res == irq) + return irq_to_desc(irq); + return NULL; +} int __init early_irq_init(void) { + int i, node = first_online_node; struct irq_desc *desc; - int legacy_count; - int node; - int i; init_irq_default_affinity(); @@ -266,71 +235,14 @@ int __init early_irq_init(void) arch_probe_nr_irqs(); printk(KERN_INFO "NR_IRQS:%d nr_irqs:%d\n", NR_IRQS, nr_irqs); - desc = irq_desc_legacy; - legacy_count = ARRAY_SIZE(irq_desc_legacy); - node = first_online_node; - - /* allocate based on nr_cpu_ids */ - kstat_irqs_legacy = kzalloc_node(NR_IRQS_LEGACY * nr_cpu_ids * - sizeof(int), GFP_NOWAIT, node); - - irq_desc_init.irq_data.chip = &no_irq_chip; - - for (i = 0; i < legacy_count; i++) { - desc[i].irq_data.irq = i; - desc[i].irq_data.chip = &no_irq_chip; -#ifdef CONFIG_SMP - desc[i].irq_data.node = node; -#endif - desc[i].kstat_irqs = kstat_irqs_legacy + i * nr_cpu_ids; - lockdep_set_class(&desc[i].lock, &irq_desc_lock_class); - alloc_desc_masks(&desc[i], node, true); - init_desc_masks(&desc[i]); - irq_insert_desc(i, &desc[i]); + for (i = 0; i < NR_IRQS_LEGACY; i++) { + desc = alloc_desc(i, node); + set_bit(i, allocated_irqs); + irq_insert_desc(i, desc); } - return arch_early_irq_init(); } -struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) -{ - struct irq_desc *desc; - unsigned long flags; - - if (irq >= nr_irqs) { - WARN(1, "irq (%d) >= nr_irqs (%d) in irq_to_desc_alloc\n", - irq, nr_irqs); - return NULL; - } - - desc = irq_to_desc(irq); - if (desc) - return desc; - - raw_spin_lock_irqsave(&sparse_irq_lock, flags); - - /* We have to check it to avoid races with another CPU */ - desc = irq_to_desc(irq); - if (desc) - goto out_unlock; - - desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); - - printk(KERN_DEBUG " alloc irq_desc for %d on node %d\n", irq, node); - if (!desc) { - printk(KERN_ERR "can not alloc irq_desc\n"); - BUG_ON(1); - } - init_one_irq_desc(irq, desc, node); - - irq_insert_desc(irq, desc); - -out_unlock: - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); - - return desc; -} - #else /* !CONFIG_SPARSE_IRQ */ struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { @@ -345,9 +257,8 @@ struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { static unsigned int kstat_irqs_all[NR_IRQS][NR_CPUS]; int __init early_irq_init(void) { + int count, i, node = first_online_node; struct irq_desc *desc; - int count; - int i; init_irq_default_affinity(); @@ -359,9 +270,9 @@ int __init early_irq_init(void) for (i = 0; i < count; i++) { desc[i].irq_data.irq = i; desc[i].irq_data.chip = &no_irq_chip; - alloc_desc_masks(&desc[i], 0, true); - init_desc_masks(&desc[i]); desc[i].kstat_irqs = kstat_irqs_all[i]; + alloc_masks(desc + i, GFP_KERNEL, node); + desc_smp_init(desc + i, node); lockdep_set_class(&desc[i].lock, &irq_desc_lock_class); } return arch_early_irq_init(); -- cgit v0.10.2 From b683de2b3cb17bb10fa6fd4af614dc75b5749fe0 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 27 Sep 2010 20:55:03 +0200 Subject: genirq: Query arch for number of early descriptors sparse irq sets up NR_IRQS_LEGACY irq descriptors and archs then go ahead and allocate more. Use the unused return value of arch_probe_nr_irqs() to let the architecture return the number of early allocations. Fix up all users. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c index c0d5c3b..5456d11 100644 --- a/arch/arm/kernel/irq.c +++ b/arch/arm/kernel/irq.c @@ -157,10 +157,8 @@ void __init init_IRQ(void) struct irq_desc *desc; int irq; - for (irq = 0; irq < nr_irqs; irq++) { - desc = irq_to_desc_alloc_node(irq, 0); + for (irq = 0; irq < nr_irqs; irq++) desc->status |= IRQ_NOREQUEST | IRQ_NOPROBE; - } init_arch_irq(); } @@ -169,7 +167,7 @@ void __init init_IRQ(void) int __init arch_probe_nr_irqs(void) { nr_irqs = arch_nr_irqs ? arch_nr_irqs : NR_IRQS; - return 0; + return nr_irqs; } #endif diff --git a/arch/sh/kernel/irq.c b/arch/sh/kernel/irq.c index 257de1f..ae5bac3 100644 --- a/arch/sh/kernel/irq.c +++ b/arch/sh/kernel/irq.c @@ -290,7 +290,7 @@ void __init init_IRQ(void) int __init arch_probe_nr_irqs(void) { nr_irqs = sh_mv.mv_nr_irqs; - return 0; + return NR_IRQS_LEGACY; } #endif diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index f1efeba..5aee1d1 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3880,7 +3880,7 @@ int __init arch_probe_nr_irqs(void) if (nr < nr_irqs) nr_irqs = nr; - return 0; + return NR_IRQS_LEGACY; } #endif diff --git a/include/linux/irq.h b/include/linux/irq.h index 096b74d..ef87882 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -214,6 +214,10 @@ struct irq_chip { */ #include +#ifndef NR_IRQS_LEGACY +# define NR_IRQS_LEGACY 0 +#endif + #ifndef ARCH_IRQ_INIT_FLAGS # define ARCH_IRQ_INIT_FLAGS 0 #endif diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index 7cbe4f9..a1fbd1d 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -226,16 +226,16 @@ struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) int __init early_irq_init(void) { - int i, node = first_online_node; + int i, initcnt, node = first_online_node; struct irq_desc *desc; init_irq_default_affinity(); - /* initialize nr_irqs based on nr_cpu_ids */ - arch_probe_nr_irqs(); - printk(KERN_INFO "NR_IRQS:%d nr_irqs:%d\n", NR_IRQS, nr_irqs); + /* Let arch update nr_irqs and return the nr of preallocated irqs */ + initcnt = arch_probe_nr_irqs(); + printk(KERN_INFO "NR_IRQS:%d nr_irqs:%d %d\n", NR_IRQS, nr_irqs, initcnt); - for (i = 0; i < NR_IRQS_LEGACY; i++) { + for (i = 0; i < initcnt; i++) { desc = alloc_desc(i, node); set_bit(i, allocated_irqs); irq_insert_desc(i, desc); diff --git a/kernel/softirq.c b/kernel/softirq.c index 07b4f1b..14a7b80 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -886,9 +886,10 @@ int __init __weak early_irq_init(void) return 0; } +#ifdef CONFIG_GENERIC_HARDIRQS int __init __weak arch_probe_nr_irqs(void) { - return 0; + return NR_IRQS_LEGACY; } int __init __weak arch_early_irq_init(void) @@ -900,3 +901,4 @@ int __weak arch_init_chip_data(struct irq_desc *desc, int node) { return 0; } +#endif -- cgit v0.10.2 From 1c9db52534a2c0e9776788cd34ccc193289fc18c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 16:46:51 +0200 Subject: pci: Convert msi to new irq_chip functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Jesse Barnes Cc: Benjamin Herrenschmidt Cc: "David S. Miller" Cc: Tony Luck Cc: Russell King diff --git a/arch/arm/mach-iop13xx/msi.c b/arch/arm/mach-iop13xx/msi.c index f34b0ed..7149fcc 100644 --- a/arch/arm/mach-iop13xx/msi.c +++ b/arch/arm/mach-iop13xx/msi.c @@ -164,10 +164,10 @@ static void iop13xx_msi_nop(unsigned int irq) static struct irq_chip iop13xx_msi_chip = { .name = "PCI-MSI", .ack = iop13xx_msi_nop, - .enable = unmask_msi_irq, - .disable = mask_msi_irq, - .mask = mask_msi_irq, - .unmask = unmask_msi_irq, + .irq_enable = unmask_msi_irq, + .irq_disable = mask_msi_irq, + .irq_mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, }; int arch_setup_msi_irq(struct pci_dev *pdev, struct msi_desc *desc) diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c index 4a746ea..dbb43424 100644 --- a/arch/ia64/kernel/msi_ia64.c +++ b/arch/ia64/kernel/msi_ia64.c @@ -104,8 +104,8 @@ static int ia64_msi_retrigger_irq(unsigned int irq) */ static struct irq_chip ia64_msi_chip = { .name = "PCI-MSI", - .mask = mask_msi_irq, - .unmask = unmask_msi_irq, + .irq_mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, .ack = ia64_ack_msi_irq, #ifdef CONFIG_SMP .set_affinity = ia64_set_msi_irq_affinity, diff --git a/arch/ia64/sn/kernel/msi_sn.c b/arch/ia64/sn/kernel/msi_sn.c index 0c72dd4..a5e500f 100644 --- a/arch/ia64/sn/kernel/msi_sn.c +++ b/arch/ia64/sn/kernel/msi_sn.c @@ -228,8 +228,8 @@ static int sn_msi_retrigger_irq(unsigned int irq) static struct irq_chip sn_msi_chip = { .name = "PCI-MSI", - .mask = mask_msi_irq, - .unmask = unmask_msi_irq, + .irq_mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, .ack = sn_ack_msi_irq, #ifdef CONFIG_SMP .set_affinity = sn_set_msi_irq_affinity, diff --git a/arch/powerpc/platforms/cell/axon_msi.c b/arch/powerpc/platforms/cell/axon_msi.c index 9708553..e3e379c 100644 --- a/arch/powerpc/platforms/cell/axon_msi.c +++ b/arch/powerpc/platforms/cell/axon_msi.c @@ -310,9 +310,9 @@ static void axon_msi_teardown_msi_irqs(struct pci_dev *dev) } static struct irq_chip msic_irq_chip = { - .mask = mask_msi_irq, - .unmask = unmask_msi_irq, - .shutdown = unmask_msi_irq, + .irq_mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, + .irq_shutdown = mask_msi_irq, .name = "AXON-MSI", }; diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c index 93834b0..67e2c4b 100644 --- a/arch/powerpc/platforms/pseries/xics.c +++ b/arch/powerpc/platforms/pseries/xics.c @@ -243,7 +243,7 @@ static unsigned int xics_startup(unsigned int virq) * at that level, so we do it here by hand. */ if (irq_to_desc(virq)->msi_desc) - unmask_msi_irq(virq); + unmask_msi_irq(irq_get_irq_data(virq)); /* unmask it */ xics_unmask_irq(virq); diff --git a/arch/powerpc/sysdev/fsl_msi.c b/arch/powerpc/sysdev/fsl_msi.c index 87991d3..bdbd896 100644 --- a/arch/powerpc/sysdev/fsl_msi.c +++ b/arch/powerpc/sysdev/fsl_msi.c @@ -51,8 +51,8 @@ static void fsl_msi_end_irq(unsigned int virq) } static struct irq_chip fsl_msi_chip = { - .mask = mask_msi_irq, - .unmask = unmask_msi_irq, + .irq_mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, .ack = fsl_msi_end_irq, .name = "FSL-MSI", }; diff --git a/arch/powerpc/sysdev/mpic_pasemi_msi.c b/arch/powerpc/sysdev/mpic_pasemi_msi.c index 3b6a9a4..320ad5a 100644 --- a/arch/powerpc/sysdev/mpic_pasemi_msi.c +++ b/arch/powerpc/sysdev/mpic_pasemi_msi.c @@ -39,24 +39,24 @@ static struct mpic *msi_mpic; -static void mpic_pasemi_msi_mask_irq(unsigned int irq) +static void mpic_pasemi_msi_mask_irq(struct irq_data *data) { - pr_debug("mpic_pasemi_msi_mask_irq %d\n", irq); - mask_msi_irq(irq); - mpic_mask_irq(irq); + pr_debug("mpic_pasemi_msi_mask_irq %d\n", data->irq); + mask_msi_irq(data); + mpic_mask_irq(data->irq); } -static void mpic_pasemi_msi_unmask_irq(unsigned int irq) +static void mpic_pasemi_msi_unmask_irq(struct irq_data *data) { - pr_debug("mpic_pasemi_msi_unmask_irq %d\n", irq); - mpic_unmask_irq(irq); - unmask_msi_irq(irq); + pr_debug("mpic_pasemi_msi_unmask_irq %d\n", data->irq); + mpic_unmask_irq(data->irq); + unmask_msi_irq(data); } static struct irq_chip mpic_pasemi_msi_chip = { - .shutdown = mpic_pasemi_msi_mask_irq, - .mask = mpic_pasemi_msi_mask_irq, - .unmask = mpic_pasemi_msi_unmask_irq, + .irq_shutdown = mpic_pasemi_msi_mask_irq, + .irq_mask = mpic_pasemi_msi_mask_irq, + .irq_unmask = mpic_pasemi_msi_unmask_irq, .eoi = mpic_end_irq, .set_type = mpic_set_irq_type, .set_affinity = mpic_set_affinity, diff --git a/arch/powerpc/sysdev/mpic_u3msi.c b/arch/powerpc/sysdev/mpic_u3msi.c index bcbfe79..a2b028b 100644 --- a/arch/powerpc/sysdev/mpic_u3msi.c +++ b/arch/powerpc/sysdev/mpic_u3msi.c @@ -23,22 +23,22 @@ /* A bit ugly, can we get this from the pci_dev somehow? */ static struct mpic *msi_mpic; -static void mpic_u3msi_mask_irq(unsigned int irq) +static void mpic_u3msi_mask_irq(struct irq_data *data) { - mask_msi_irq(irq); - mpic_mask_irq(irq); + mask_msi_irq(data); + mpic_mask_irq(data->irq); } -static void mpic_u3msi_unmask_irq(unsigned int irq) +static void mpic_u3msi_unmask_irq(struct irq_data *data) { - mpic_unmask_irq(irq); - unmask_msi_irq(irq); + mpic_unmask_irq(data->irq); + unmask_msi_irq(data); } static struct irq_chip mpic_u3msi_chip = { - .shutdown = mpic_u3msi_mask_irq, - .mask = mpic_u3msi_mask_irq, - .unmask = mpic_u3msi_unmask_irq, + .irq_shutdown = mpic_u3msi_mask_irq, + .irq_mask = mpic_u3msi_mask_irq, + .irq_unmask = mpic_u3msi_unmask_irq, .eoi = mpic_end_irq, .set_type = mpic_set_irq_type, .set_affinity = mpic_set_affinity, diff --git a/arch/sparc/kernel/pci_msi.c b/arch/sparc/kernel/pci_msi.c index 548b8ca..b210416 100644 --- a/arch/sparc/kernel/pci_msi.c +++ b/arch/sparc/kernel/pci_msi.c @@ -114,10 +114,10 @@ static void free_msi(struct pci_pbm_info *pbm, int msi_num) static struct irq_chip msi_irq = { .name = "PCI-MSI", - .mask = mask_msi_irq, - .unmask = unmask_msi_irq, - .enable = unmask_msi_irq, - .disable = mask_msi_irq, + .irq_mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, + .irq_enable = unmask_msi_irq, + .irq_disable = mask_msi_irq, /* XXX affinity XXX */ }; diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 7556eb7..b79938f 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3441,8 +3441,8 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) */ static struct irq_chip msi_chip = { .name = "PCI-MSI", - .unmask = unmask_msi_irq, - .mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, + .irq_mask = mask_msi_irq, .ack = ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = set_msi_irq_affinity, @@ -3452,8 +3452,8 @@ static struct irq_chip msi_chip = { static struct irq_chip msi_ir_chip = { .name = "IR-PCI-MSI", - .unmask = unmask_msi_irq, - .mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, + .irq_mask = mask_msi_irq, #ifdef CONFIG_INTR_REMAP .ack = ir_ack_apic_edge, #ifdef CONFIG_SMP diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c index 69b7be3..55e0f93 100644 --- a/drivers/pci/msi.c +++ b/drivers/pci/msi.c @@ -170,27 +170,27 @@ static void msix_mask_irq(struct msi_desc *desc, u32 flag) desc->masked = __msix_mask_irq(desc, flag); } -static void msi_set_mask_bit(unsigned irq, u32 flag) +static void msi_set_mask_bit(struct irq_data *data, u32 flag) { - struct msi_desc *desc = get_irq_msi(irq); + struct msi_desc *desc = irq_data_get_msi(data); if (desc->msi_attrib.is_msix) { msix_mask_irq(desc, flag); readl(desc->mask_base); /* Flush write to device */ } else { - unsigned offset = irq - desc->dev->irq; + unsigned offset = data->irq - desc->dev->irq; msi_mask_irq(desc, 1 << offset, flag << offset); } } -void mask_msi_irq(unsigned int irq) +void mask_msi_irq(struct irq_data *data) { - msi_set_mask_bit(irq, 1); + msi_set_mask_bit(data, 1); } -void unmask_msi_irq(unsigned int irq) +void unmask_msi_irq(struct irq_data *data) { - msi_set_mask_bit(irq, 0); + msi_set_mask_bit(data, 0); } void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) diff --git a/include/linux/msi.h b/include/linux/msi.h index 91b05c1..329d17c 100644 --- a/include/linux/msi.h +++ b/include/linux/msi.h @@ -11,8 +11,9 @@ struct msi_msg { /* Helper functions */ struct irq_desc; -extern void mask_msi_irq(unsigned int irq); -extern void unmask_msi_irq(unsigned int irq); +struct irq_data; +extern void mask_msi_irq(struct irq_data *data); +extern void unmask_msi_irq(struct irq_data *data); extern void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); extern void get_cached_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); extern void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); -- cgit v0.10.2 From 39431acb1a4c464e62471cb3058b8ffffb9244db Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 19:09:51 +0200 Subject: pci: Cleanup the irq_desc mess in msi Handing down irq_desc to msi just so that msi can access irq_desc.irq_data.msi_desc is a pretty stupid idea. The calling code can hand down a pointer to msi_desc so msi code does not need to know about the irq descriptor at all. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Jesse Barnes diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index b79938f..74bb027 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3383,14 +3383,14 @@ static int set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) cfg = desc->chip_data; - get_cached_msi_msg_desc(desc, &msg); + __get_cached_msi_msg(desc->irq_data.msi_desc, &msg); msg.data &= ~MSI_DATA_VECTOR_MASK; msg.data |= MSI_DATA_VECTOR(cfg->vector); msg.address_lo &= ~MSI_ADDR_DEST_ID_MASK; msg.address_lo |= MSI_ADDR_DEST_ID(dest); - write_msi_msg_desc(desc, &msg); + __write_msi_msg(desc->irq_data.msi_desc, &msg); return 0; } diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c index 55e0f93..5fcf5ae 100644 --- a/drivers/pci/msi.c +++ b/drivers/pci/msi.c @@ -193,10 +193,8 @@ void unmask_msi_irq(struct irq_data *data) msi_set_mask_bit(data, 0); } -void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) +void __read_msi_msg(struct msi_desc *entry, struct msi_msg *msg) { - struct msi_desc *entry = get_irq_desc_msi(desc); - BUG_ON(entry->dev->current_state != PCI_D0); if (entry->msi_attrib.is_msix) { @@ -227,15 +225,13 @@ void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) void read_msi_msg(unsigned int irq, struct msi_msg *msg) { - struct irq_desc *desc = irq_to_desc(irq); + struct msi_desc *entry = get_irq_msi(irq); - read_msi_msg_desc(desc, msg); + __read_msi_msg(entry, msg); } -void get_cached_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) +void __get_cached_msi_msg(struct msi_desc *entry, struct msi_msg *msg) { - struct msi_desc *entry = get_irq_desc_msi(desc); - /* Assert that the cache is valid, assuming that * valid messages are not all-zeroes. */ BUG_ON(!(entry->msg.address_hi | entry->msg.address_lo | @@ -246,15 +242,13 @@ void get_cached_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) void get_cached_msi_msg(unsigned int irq, struct msi_msg *msg) { - struct irq_desc *desc = irq_to_desc(irq); + struct msi_desc *entry = get_irq_msi(irq); - get_cached_msi_msg_desc(desc, msg); + __get_cached_msi_msg(entry, msg); } -void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) +void __write_msi_msg(struct msi_desc *entry, struct msi_msg *msg) { - struct msi_desc *entry = get_irq_desc_msi(desc); - if (entry->dev->current_state != PCI_D0) { /* Don't touch the hardware now */ } else if (entry->msi_attrib.is_msix) { @@ -292,9 +286,9 @@ void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) void write_msi_msg(unsigned int irq, struct msi_msg *msg) { - struct irq_desc *desc = irq_to_desc(irq); + struct msi_desc *entry = get_irq_msi(irq); - write_msi_msg_desc(desc, msg); + __write_msi_msg(entry, msg); } static void free_msi_irqs(struct pci_dev *dev) diff --git a/include/linux/msi.h b/include/linux/msi.h index 329d17c..05acced 100644 --- a/include/linux/msi.h +++ b/include/linux/msi.h @@ -10,13 +10,13 @@ struct msi_msg { }; /* Helper functions */ -struct irq_desc; struct irq_data; +struct msi_desc; extern void mask_msi_irq(struct irq_data *data); extern void unmask_msi_irq(struct irq_data *data); -extern void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); -extern void get_cached_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); -extern void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); +extern void __read_msi_msg(struct msi_desc *entry, struct msi_msg *msg); +extern void __get_cached_msi_msg(struct msi_desc *entry, struct msi_msg *msg); +extern void __write_msi_msg(struct msi_desc *entry, struct msi_msg *msg); extern void read_msi_msg(unsigned int irq, struct msi_msg *msg); extern void get_cached_msi_msg(unsigned int irq, struct msi_msg *msg); extern void write_msi_msg(unsigned int irq, struct msi_msg *msg); -- cgit v0.10.2 From 011d578fdadb64bcc1deedbb02216bfee6a9b4dc Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 00:15:31 +0200 Subject: x86: Remove useless reinitialization of irq descriptors The descriptors are already initialized in exactly this way. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c index 990ae7c..a91ab50 100644 --- a/arch/x86/kernel/irqinit.c +++ b/arch/x86/kernel/irqinit.c @@ -100,6 +100,8 @@ int vector_used_by_percpu_irq(unsigned int vector) void __init init_ISA_irqs(void) { + struct irq_chip *chip = legacy_pic->chip; + const char *name = chip->name; int i; #if defined(CONFIG_X86_64) || defined(CONFIG_X86_LOCAL_APIC) @@ -107,19 +109,8 @@ void __init init_ISA_irqs(void) #endif legacy_pic->init(0); - /* - * 16 old-style INTA-cycle interrupts: - */ - for (i = 0; i < legacy_pic->nr_legacy_irqs; i++) { - struct irq_desc *desc = irq_to_desc(i); - - desc->status = IRQ_DISABLED; - desc->action = NULL; - desc->depth = 1; - - set_irq_chip_and_handler_name(i, &i8259A_chip, - handle_level_irq, "XT"); - } + for (i = 0; i < legacy_pic->nr_legacy_irqs; i++) + set_irq_chip_and_handler_name(i, chip, handle_level_irq, name); } void __init init_IRQ(void) -- cgit v0.10.2 From a3c08e5d80c54e32423efbba113b02942c91f726 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 20:24:58 +0200 Subject: x86: Convert irq_chip access to new functions Before moving the irq chips to the new functions, fixup direct callers. The cpu offline irq fixup code needs to become generic and archs need to honour the "force" flag as an indicator, but that's for later. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index 91fd0c7..d765bdc 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -159,7 +159,7 @@ int show_interrupts(struct seq_file *p, void *v) seq_printf(p, "%*d: ", prec, i); for_each_online_cpu(j) seq_printf(p, "%10u ", kstat_irqs_cpu(i, j)); - seq_printf(p, " %8s", desc->chip->name); + seq_printf(p, " %8s", desc->irq_data.chip->name); seq_printf(p, "-%-8s", desc->name); if (action) { @@ -282,6 +282,7 @@ void fixup_irqs(void) unsigned int irq, vector; static int warned; struct irq_desc *desc; + struct irq_data *data; for_each_irq_desc(irq, desc) { int break_affinity = 0; @@ -296,7 +297,8 @@ void fixup_irqs(void) /* interrupt's are disabled at this point */ raw_spin_lock(&desc->lock); - affinity = desc->affinity; + data = &desc->irq_data; + affinity = data->affinity; if (!irq_has_action(irq) || cpumask_equal(affinity, cpu_online_mask)) { raw_spin_unlock(&desc->lock); @@ -315,16 +317,16 @@ void fixup_irqs(void) affinity = cpu_all_mask; } - if (!(desc->status & IRQ_MOVE_PCNTXT) && desc->chip->mask) - desc->chip->mask(irq); + if (!(desc->status & IRQ_MOVE_PCNTXT) && data->chip->irq_mask) + data->chip->irq_mask(data); - if (desc->chip->set_affinity) - desc->chip->set_affinity(irq, affinity); + if (data->chip->irq_set_affinity) + data->chip->irq_set_affinity(data, affinity, true); else if (!(warned++)) set_affinity = 0; - if (!(desc->status & IRQ_MOVE_PCNTXT) && desc->chip->unmask) - desc->chip->unmask(irq); + if (!(desc->status & IRQ_MOVE_PCNTXT) && data->chip->irq_unmask) + data->chip->irq_unmask(data); raw_spin_unlock(&desc->lock); @@ -355,10 +357,10 @@ void fixup_irqs(void) if (irr & (1 << (vector % 32))) { irq = __get_cpu_var(vector_irq)[vector]; - desc = irq_to_desc(irq); + data = irq_get_irq_data(irq); raw_spin_lock(&desc->lock); - if (desc->chip->retrigger) - desc->chip->retrigger(irq); + if (data->chip->irq_retrigger) + data->chip->irq_retrigger(data); raw_spin_unlock(&desc->lock); } } -- cgit v0.10.2 From a5ef2e70405c8a9ee380b5ff33a008c75454791f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 11:11:10 +0200 Subject: x86: Sanitize apb timer interrupt handling Disable the interrupt in CPU_DEAD where it belongs. Remove the open coded irq_desc manipulation. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Jacob Pan diff --git a/arch/x86/kernel/apb_timer.c b/arch/x86/kernel/apb_timer.c index 08f75fb..42a70a2 100644 --- a/arch/x86/kernel/apb_timer.c +++ b/arch/x86/kernel/apb_timer.c @@ -231,34 +231,6 @@ static void apbt_restart_clocksource(struct clocksource *cs) apbt_start_counter(phy_cs_timer_id); } -/* Setup IRQ routing via IOAPIC */ -#ifdef CONFIG_SMP -static void apbt_setup_irq(struct apbt_dev *adev) -{ - struct irq_chip *chip; - struct irq_desc *desc; - - /* timer0 irq has been setup early */ - if (adev->irq == 0) - return; - desc = irq_to_desc(adev->irq); - chip = get_irq_chip(adev->irq); - disable_irq(adev->irq); - desc->status |= IRQ_MOVE_PCNTXT; - irq_set_affinity(adev->irq, cpumask_of(adev->cpu)); - /* APB timer irqs are set up as mp_irqs, timer is edge triggerred */ - set_irq_chip_and_handler_name(adev->irq, chip, handle_edge_irq, "edge"); - enable_irq(adev->irq); - if (system_state == SYSTEM_BOOTING) - if (request_irq(adev->irq, apbt_interrupt_handler, - IRQF_TIMER | IRQF_DISABLED | IRQF_NOBALANCING, - adev->name, adev)) { - printk(KERN_ERR "Failed request IRQ for APBT%d\n", - adev->num); - } -} -#endif - static void apbt_enable_int(int n) { unsigned long ctrl = apbt_readl(n, APBTMR_N_CONTROL); @@ -334,6 +306,27 @@ static int __init apbt_clockevent_register(void) } #ifdef CONFIG_SMP + +static void apbt_setup_irq(struct apbt_dev *adev) +{ + /* timer0 irq has been setup early */ + if (adev->irq == 0) + return; + + if (system_state == SYSTEM_BOOTING) { + irq_modify_status(adev->irq, 0, IRQ_MOVE_PCNTXT); + /* APB timer irqs are set up as mp_irqs, timer is edge type */ + __set_irq_handler(adev->irq, handle_edge_irq, 0, "edge"); + if (request_irq(adev->irq, apbt_interrupt_handler, + IRQF_TIMER | IRQF_DISABLED | IRQF_NOBALANCING, + adev->name, adev)) { + printk(KERN_ERR "Failed request IRQ for APBT%d\n", + adev->num); + } + } else + enable_irq(adev->irq); +} + /* Should be called with per cpu */ void apbt_setup_secondary_clock(void) { @@ -389,10 +382,11 @@ static int apbt_cpuhp_notify(struct notifier_block *n, switch (action & 0xf) { case CPU_DEAD: + disable_irq(adev->irq); apbt_disable_int(cpu); - if (system_state == SYSTEM_RUNNING) + if (system_state == SYSTEM_RUNNING) { pr_debug("skipping APBT CPU %lu offline\n", cpu); - else if (adev) { + } else if (adev) { pr_debug("APBT clockevent for cpu %lu offline\n", cpu); free_irq(adev->irq, adev); } -- cgit v0.10.2 From fe25c7fc2e036e1569faac8715a8aa5496cda78d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 14:57:24 +0200 Subject: x86: lguest: Convert to new irq chip functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Rusty Russell diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c index 9d5f558..2d4e6fc 100644 --- a/arch/x86/lguest/boot.c +++ b/arch/x86/lguest/boot.c @@ -791,22 +791,22 @@ static void lguest_flush_tlb_kernel(void) * simple as setting a bit. We don't actually "ack" interrupts as such, we * just mask and unmask them. I wonder if we should be cleverer? */ -static void disable_lguest_irq(unsigned int irq) +static void disable_lguest_irq(struct irq_data *data) { - set_bit(irq, lguest_data.blocked_interrupts); + set_bit(data->irq, lguest_data.blocked_interrupts); } -static void enable_lguest_irq(unsigned int irq) +static void enable_lguest_irq(struct irq_data *data) { - clear_bit(irq, lguest_data.blocked_interrupts); + clear_bit(data->irq, lguest_data.blocked_interrupts); } /* This structure describes the lguest IRQ controller. */ static struct irq_chip lguest_irq_controller = { .name = "lguest", - .mask = disable_lguest_irq, - .mask_ack = disable_lguest_irq, - .unmask = enable_lguest_irq, + .irq_mask = disable_lguest_irq, + .irq_mask_ack = disable_lguest_irq, + .irq_unmask = enable_lguest_irq, }; /* -- cgit v0.10.2 From 020dd984d7c0792e8234f2e4b4fb0534fe750f9d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 14:59:58 +0200 Subject: x86: Cleanup visws interrupt handling Remove the open coded access to irq_desc and convert to the new irq chip functions. Change the mask function of piix4_virtual_irq_type so we can use the generic irq handling function for the virtual interrupt instead of open coding it. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/visws_quirks.c b/arch/x86/kernel/visws_quirks.c index e680ea5..3371bd0 100644 --- a/arch/x86/kernel/visws_quirks.c +++ b/arch/x86/kernel/visws_quirks.c @@ -66,10 +66,7 @@ static void __init visws_time_init(void) } /* Replaces the default init_ISA_irqs in the generic setup */ -static void __init visws_pre_intr_init(void) -{ - init_VISWS_APIC_irqs(); -} +static void __init visws_pre_intr_init(void); /* Quirk for machine specific memory setup. */ @@ -429,67 +426,34 @@ static int is_co_apic(unsigned int irq) /* * This is the SGI Cobalt (IO-)APIC: */ - -static void enable_cobalt_irq(unsigned int irq) +static void enable_cobalt_irq(struct irq_data *data) { - co_apic_set(is_co_apic(irq), irq); + co_apic_set(is_co_apic(data->irq), data->irq); } -static void disable_cobalt_irq(unsigned int irq) +static void disable_cobalt_irq(struct irq_data *data) { - int entry = is_co_apic(irq); + int entry = is_co_apic(data->irq); co_apic_write(CO_APIC_LO(entry), CO_APIC_MASK); co_apic_read(CO_APIC_LO(entry)); } -/* - * "irq" really just serves to identify the device. Here is where we - * map this to the Cobalt APIC entry where it's physically wired. - * This is called via request_irq -> setup_irq -> irq_desc->startup() - */ -static unsigned int startup_cobalt_irq(unsigned int irq) +static void ack_cobalt_irq(struct irq_data *data) { unsigned long flags; - struct irq_desc *desc = irq_to_desc(irq); spin_lock_irqsave(&cobalt_lock, flags); - if ((desc->status & (IRQ_DISABLED | IRQ_INPROGRESS | IRQ_WAITING))) - desc->status &= ~(IRQ_DISABLED | IRQ_INPROGRESS | IRQ_WAITING); - enable_cobalt_irq(irq); - spin_unlock_irqrestore(&cobalt_lock, flags); - return 0; -} - -static void ack_cobalt_irq(unsigned int irq) -{ - unsigned long flags; - - spin_lock_irqsave(&cobalt_lock, flags); - disable_cobalt_irq(irq); + disable_cobalt_irq(data); apic_write(APIC_EOI, APIC_EIO_ACK); spin_unlock_irqrestore(&cobalt_lock, flags); } -static void end_cobalt_irq(unsigned int irq) -{ - unsigned long flags; - struct irq_desc *desc = irq_to_desc(irq); - - spin_lock_irqsave(&cobalt_lock, flags); - if (!(desc->status & (IRQ_DISABLED | IRQ_INPROGRESS))) - enable_cobalt_irq(irq); - spin_unlock_irqrestore(&cobalt_lock, flags); -} - static struct irq_chip cobalt_irq_type = { - .name = "Cobalt-APIC", - .startup = startup_cobalt_irq, - .shutdown = disable_cobalt_irq, - .enable = enable_cobalt_irq, - .disable = disable_cobalt_irq, - .ack = ack_cobalt_irq, - .end = end_cobalt_irq, + .name = "Cobalt-APIC", + .irq_enable = enable_cobalt_irq, + .irq_disable = disable_cobalt_irq, + .irq_ack = ack_cobalt_irq, }; @@ -503,35 +467,34 @@ static struct irq_chip cobalt_irq_type = { * interrupt controller type, and through a special virtual interrupt- * controller. Device drivers only see the virtual interrupt sources. */ -static unsigned int startup_piix4_master_irq(unsigned int irq) +static unsigned int startup_piix4_master_irq(struct irq_data *data) { legacy_pic->init(0); - - return startup_cobalt_irq(irq); + enable_cobalt_irq(data); } -static void end_piix4_master_irq(unsigned int irq) +static void end_piix4_master_irq(struct irq_data *data) { unsigned long flags; spin_lock_irqsave(&cobalt_lock, flags); - enable_cobalt_irq(irq); + enable_cobalt_irq(data); spin_unlock_irqrestore(&cobalt_lock, flags); } static struct irq_chip piix4_master_irq_type = { - .name = "PIIX4-master", - .startup = startup_piix4_master_irq, - .ack = ack_cobalt_irq, - .end = end_piix4_master_irq, + .name = "PIIX4-master", + .irq_startup = startup_piix4_master_irq, + .irq_ack = ack_cobalt_irq, }; +static void pii4_mask(struct irq_data *data) { } static struct irq_chip piix4_virtual_irq_type = { - .name = "PIIX4-virtual", + .name = "PIIX4-virtual", + .mask = pii4_mask, }; - /* * PIIX4-8259 master/virtual functions to handle interrupt requests * from legacy devices: floppy, parallel, serial, rtc. @@ -549,9 +512,8 @@ static struct irq_chip piix4_virtual_irq_type = { */ static irqreturn_t piix4_master_intr(int irq, void *dev_id) { - int realirq; - struct irq_desc *desc; unsigned long flags; + int realirq; raw_spin_lock_irqsave(&i8259A_lock, flags); @@ -592,18 +554,10 @@ static irqreturn_t piix4_master_intr(int irq, void *dev_id) raw_spin_unlock_irqrestore(&i8259A_lock, flags); - desc = irq_to_desc(realirq); - /* * handle this 'virtual interrupt' as a Cobalt one now. */ - kstat_incr_irqs_this_cpu(realirq, desc); - - if (likely(desc->action != NULL)) - handle_IRQ_event(realirq, desc->action); - - if (!(desc->status & IRQ_DISABLED)) - legacy_pic->chip->unmask(realirq); + generic_handle_irq(realirq); return IRQ_HANDLED; @@ -624,41 +578,35 @@ static struct irqaction cascade_action = { static inline void set_piix4_virtual_irq_type(void) { - piix4_virtual_irq_type.shutdown = i8259A_chip.mask; piix4_virtual_irq_type.enable = i8259A_chip.unmask; piix4_virtual_irq_type.disable = i8259A_chip.mask; + piix4_virtual_irq_type.unmask = i8259A_chip.unmask; } -void init_VISWS_APIC_irqs(void) +static void __init visws_pre_intr_init(void) { int i; - for (i = 0; i < CO_IRQ_APIC0 + CO_APIC_LAST + 1; i++) { - struct irq_desc *desc = irq_to_desc(i); - - desc->status = IRQ_DISABLED; - desc->action = 0; - desc->depth = 1; + set_piix4_virtual_irq_type(); - if (i == 0) { - desc->chip = &cobalt_irq_type; - } - else if (i == CO_IRQ_IDE0) { - desc->chip = &cobalt_irq_type; - } - else if (i == CO_IRQ_IDE1) { - desc->chip = &cobalt_irq_type; - } - else if (i == CO_IRQ_8259) { - desc->chip = &piix4_master_irq_type; - } - else if (i < CO_IRQ_APIC0) { - set_piix4_virtual_irq_type(); - desc->chip = &piix4_virtual_irq_type; - } - else if (IS_CO_APIC(i)) { - desc->chip = &cobalt_irq_type; - } + for (i = 0; i < CO_IRQ_APIC0 + CO_APIC_LAST + 1; i++) { + struct irq_chip *chip = NULL; + + if (i == 0) + chip = &cobalt_irq_type; + else if (i == CO_IRQ_IDE0) + chip = &cobalt_irq_type; + else if (i == CO_IRQ_IDE1) + >chip = &cobalt_irq_type; + else if (i == CO_IRQ_8259) + chip = &piix4_master_irq_type; + else if (i < CO_IRQ_APIC0) + chip = &piix4_virtual_irq_type; + else if (IS_CO_APIC(i)) + chip = &cobalt_irq_type; + + if (chip) + set_irq_chip(i, chip); } setup_irq(CO_IRQ_8259, &master_action); -- cgit v0.10.2 From 4305df947ca1fd52867c8d56837a4e6b1e33167c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 15:01:33 +0200 Subject: x86: i8259: Convert to new irq_chip functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/include/asm/i8259.h b/arch/x86/include/asm/i8259.h index 1655147..a203659 100644 --- a/arch/x86/include/asm/i8259.h +++ b/arch/x86/include/asm/i8259.h @@ -55,6 +55,8 @@ extern struct irq_chip i8259A_chip; struct legacy_pic { int nr_legacy_irqs; struct irq_chip *chip; + void (*mask)(unsigned int irq); + void (*unmask)(unsigned int irq); void (*mask_all)(void); void (*restore_mask)(void); void (*init)(int auto_eoi); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 74bb027..e5ae2a2 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1459,7 +1459,7 @@ static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq ioapic_register_intr(irq, desc, trigger); if (irq < legacy_pic->nr_legacy_irqs) - legacy_pic->chip->mask(irq); + legacy_pic->mask(irq); ioapic_write_entry(apic_id, pin, entry); } @@ -2233,7 +2233,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) raw_spin_lock_irqsave(&ioapic_lock, flags); if (irq < legacy_pic->nr_legacy_irqs) { - legacy_pic->chip->mask(irq); + legacy_pic->mask(irq); if (legacy_pic->irq_pending(irq)) was_pending = 1; } @@ -2928,7 +2928,7 @@ static inline void __init check_timer(void) /* * get/set the timer IRQ vector: */ - legacy_pic->chip->mask(0); + legacy_pic->mask(0); assign_irq_vector(0, cfg, apic->target_cpus()); /* @@ -3000,7 +3000,7 @@ static inline void __init check_timer(void) if (timer_irq_works()) { if (nmi_watchdog == NMI_IO_APIC) { setup_nmi(); - legacy_pic->chip->unmask(0); + legacy_pic->unmask(0); } if (disable_timer_pin_1 > 0) clear_IO_APIC_pin(0, pin1); @@ -3023,14 +3023,14 @@ static inline void __init check_timer(void) */ replace_pin_at_irq_node(cfg, node, apic1, pin1, apic2, pin2); setup_timer_IRQ0_pin(apic2, pin2, cfg->vector); - legacy_pic->chip->unmask(0); + legacy_pic->unmask(0); if (timer_irq_works()) { apic_printk(APIC_QUIET, KERN_INFO "....... works.\n"); timer_through_8259 = 1; if (nmi_watchdog == NMI_IO_APIC) { - legacy_pic->chip->mask(0); + legacy_pic->mask(0); setup_nmi(); - legacy_pic->chip->unmask(0); + legacy_pic->unmask(0); } goto out; } @@ -3038,7 +3038,7 @@ static inline void __init check_timer(void) * Cleanup, just in case ... */ local_irq_disable(); - legacy_pic->chip->mask(0); + legacy_pic->mask(0); clear_IO_APIC_pin(apic2, pin2); apic_printk(APIC_QUIET, KERN_INFO "....... failed.\n"); } @@ -3057,14 +3057,14 @@ static inline void __init check_timer(void) lapic_register_intr(0, desc); apic_write(APIC_LVT0, APIC_DM_FIXED | cfg->vector); /* Fixed mode */ - legacy_pic->chip->unmask(0); + legacy_pic->unmask(0); if (timer_irq_works()) { apic_printk(APIC_QUIET, KERN_INFO "..... works.\n"); goto out; } local_irq_disable(); - legacy_pic->chip->mask(0); + legacy_pic->mask(0); apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_FIXED | cfg->vector); apic_printk(APIC_QUIET, KERN_INFO "..... failed.\n"); diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c index a43f71c..c90041c 100644 --- a/arch/x86/kernel/apic/nmi.c +++ b/arch/x86/kernel/apic/nmi.c @@ -178,7 +178,7 @@ int __init check_nmi_watchdog(void) error: if (nmi_watchdog == NMI_IO_APIC) { if (!timer_through_8259) - legacy_pic->chip->mask(0); + legacy_pic->mask(0); on_each_cpu(__acpi_nmi_disable, NULL, 1); } diff --git a/arch/x86/kernel/i8259.c b/arch/x86/kernel/i8259.c index cafa7c8..20757cb 100644 --- a/arch/x86/kernel/i8259.c +++ b/arch/x86/kernel/i8259.c @@ -29,24 +29,10 @@ * plus some generic x86 specific things if generic specifics makes * any sense at all. */ +static void init_8259A(int auto_eoi); static int i8259A_auto_eoi; DEFINE_RAW_SPINLOCK(i8259A_lock); -static void mask_and_ack_8259A(unsigned int); -static void mask_8259A(void); -static void unmask_8259A(void); -static void disable_8259A_irq(unsigned int irq); -static void enable_8259A_irq(unsigned int irq); -static void init_8259A(int auto_eoi); -static int i8259A_irq_pending(unsigned int irq); - -struct irq_chip i8259A_chip = { - .name = "XT-PIC", - .mask = disable_8259A_irq, - .disable = disable_8259A_irq, - .unmask = enable_8259A_irq, - .mask_ack = mask_and_ack_8259A, -}; /* * 8259A PIC functions to handle ISA devices: @@ -68,7 +54,7 @@ unsigned int cached_irq_mask = 0xffff; */ unsigned long io_apic_irqs; -static void disable_8259A_irq(unsigned int irq) +static void mask_8259A_irq(unsigned int irq) { unsigned int mask = 1 << irq; unsigned long flags; @@ -82,7 +68,12 @@ static void disable_8259A_irq(unsigned int irq) raw_spin_unlock_irqrestore(&i8259A_lock, flags); } -static void enable_8259A_irq(unsigned int irq) +static void disable_8259A_irq(struct irq_data *data) +{ + mask_8259A_irq(data->irq); +} + +static void unmask_8259A_irq(unsigned int irq) { unsigned int mask = ~(1 << irq); unsigned long flags; @@ -96,6 +87,11 @@ static void enable_8259A_irq(unsigned int irq) raw_spin_unlock_irqrestore(&i8259A_lock, flags); } +static void enable_8259A_irq(struct irq_data *data) +{ + unmask_8259A_irq(data->irq); +} + static int i8259A_irq_pending(unsigned int irq) { unsigned int mask = 1<irq; unsigned int irqmask = 1 << irq; unsigned long flags; @@ -223,6 +220,14 @@ spurious_8259A_irq: } } +struct irq_chip i8259A_chip = { + .name = "XT-PIC", + .irq_mask = disable_8259A_irq, + .irq_disable = disable_8259A_irq, + .irq_unmask = enable_8259A_irq, + .irq_mask_ack = mask_and_ack_8259A, +}; + static char irq_trigger[2]; /** * ELCR registers (0x4d0, 0x4d1) control edge/level of IRQ @@ -342,9 +347,9 @@ static void init_8259A(int auto_eoi) * In AEOI mode we just have to mask the interrupt * when acking. */ - i8259A_chip.mask_ack = disable_8259A_irq; + i8259A_chip.irq_mask_ack = disable_8259A_irq; else - i8259A_chip.mask_ack = mask_and_ack_8259A; + i8259A_chip.irq_mask_ack = mask_and_ack_8259A; udelay(100); /* wait for 8259A to initialize */ @@ -363,14 +368,6 @@ static void init_8259A(int auto_eoi) static void legacy_pic_noop(void) { }; static void legacy_pic_uint_noop(unsigned int unused) { }; static void legacy_pic_int_noop(int unused) { }; - -static struct irq_chip dummy_pic_chip = { - .name = "dummy pic", - .mask = legacy_pic_uint_noop, - .unmask = legacy_pic_uint_noop, - .disable = legacy_pic_uint_noop, - .mask_ack = legacy_pic_uint_noop, -}; static int legacy_pic_irq_pending_noop(unsigned int irq) { return 0; @@ -378,7 +375,9 @@ static int legacy_pic_irq_pending_noop(unsigned int irq) struct legacy_pic null_legacy_pic = { .nr_legacy_irqs = 0, - .chip = &dummy_pic_chip, + .chip = &dummy_irq_chip, + .mask = legacy_pic_uint_noop, + .unmask = legacy_pic_uint_noop, .mask_all = legacy_pic_noop, .restore_mask = legacy_pic_noop, .init = legacy_pic_int_noop, @@ -389,7 +388,9 @@ struct legacy_pic null_legacy_pic = { struct legacy_pic default_legacy_pic = { .nr_legacy_irqs = NR_IRQS_LEGACY, .chip = &i8259A_chip, - .mask_all = mask_8259A, + .mask = mask_8259A_irq, + .unmask = unmask_8259A_irq, + .mask_all = mask_8259A, .restore_mask = unmask_8259A, .init = init_8259A, .irq_pending = i8259A_irq_pending, diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 87a8c6b..864b386 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -324,9 +324,9 @@ notrace static void __cpuinit start_secondary(void *unused) check_tsc_sync_target(); if (nmi_watchdog == NMI_IO_APIC) { - legacy_pic->chip->mask(0); + legacy_pic->mask(0); enable_NMI_through_LVT0(); - legacy_pic->chip->unmask(0); + legacy_pic->unmask(0); } /* This must be done before setting cpu_online_mask */ -- cgit v0.10.2 From d4eba29770244e7cc5e60c0977d73d84148a3d6d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 24 Sep 2010 12:26:18 +0200 Subject: x86: Cleanup access to irq_data Fixup the open coded access to irq_desc->[handler_data|chip_data|msi-desc] Use the macros and inline functions for it. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index e5ae2a2..fa0d92a 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -190,7 +190,7 @@ struct irq_cfg *irq_cfg(unsigned int irq) desc = irq_to_desc(irq); if (desc) - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); return cfg; } @@ -219,10 +219,11 @@ int arch_init_chip_data(struct irq_desc *desc, int node) { struct irq_cfg *cfg; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); if (!cfg) { - desc->chip_data = get_one_free_irq_cfg(node); - if (!desc->chip_data) { + cfg = get_one_free_irq_cfg(node); + desc->chip_data = cfg; + if (!cfg) { printk(KERN_ERR "can not alloc irq_cfg\n"); BUG_ON(1); } @@ -325,8 +326,8 @@ void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) { struct irq_cfg *old_cfg, *cfg; - old_cfg = old_desc->chip_data; - cfg = desc->chip_data; + old_cfg = get_irq_desc_chip_data(old_desc); + cfg = get_irq_desc_chip_data(desc); if (old_cfg == cfg) return; @@ -594,7 +595,7 @@ static void __mask_IO_APIC_irq(struct irq_cfg *cfg) static void mask_IO_APIC_irq_desc(struct irq_desc *desc) { - struct irq_cfg *cfg = desc->chip_data; + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); unsigned long flags; BUG_ON(!cfg); @@ -606,7 +607,7 @@ static void mask_IO_APIC_irq_desc(struct irq_desc *desc) static void unmask_IO_APIC_irq_desc(struct irq_desc *desc) { - struct irq_cfg *cfg = desc->chip_data; + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); unsigned long flags; raw_spin_lock_irqsave(&ioapic_lock, flags); @@ -1269,7 +1270,7 @@ void __setup_vector_irq(int cpu) raw_spin_lock(&vector_lock); /* Mark the inuse vectors */ for_each_irq_desc(irq, desc) { - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); /* * If it is a legacy IRQ handled by the legacy PIC, this cpu @@ -1427,7 +1428,7 @@ static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq if (!IO_APIC_IRQ(irq)) return; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); /* * For legacy irqs, cfg->domain starts with cpu 0 for legacy @@ -1516,7 +1517,7 @@ static void __init setup_IO_APIC_irqs(void) printk(KERN_INFO "can not get irq_desc for %d\n", irq); continue; } - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); add_pin_to_irq_node(cfg, node, apic_id, pin); /* * don't mark it in pin_programmed, so later acpi could @@ -1567,7 +1568,7 @@ void setup_IO_APIC_irq_extra(u32 gsi) return; } - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); add_pin_to_irq_node(cfg, node, apic_id, pin); if (test_bit(pin, mp_ioapic_routing[apic_id].pin_programmed)) { @@ -1718,7 +1719,7 @@ __apicdebuginit(void) print_IO_APIC(void) for_each_irq_desc(irq, desc) { struct irq_pin_list *entry; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); if (!cfg) continue; entry = cfg->irq_2_pin; @@ -2323,7 +2324,7 @@ set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask, return -1; irq = desc->irq; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); if (assign_irq_vector(irq, cfg, mask)) return -1; @@ -2343,7 +2344,7 @@ set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) int ret = -1; irq = desc->irq; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); raw_spin_lock_irqsave(&ioapic_lock, flags); ret = set_desc_affinity(desc, mask, &dest); @@ -2396,7 +2397,7 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) if (get_irte(irq, &irte)) return ret; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); if (assign_irq_vector(irq, cfg, mask)) return ret; @@ -2500,7 +2501,7 @@ unlock: static void __irq_complete_move(struct irq_desc **descp, unsigned vector) { struct irq_desc *desc = *descp; - struct irq_cfg *cfg = desc->chip_data; + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); unsigned me; if (likely(!cfg->move_in_progress)) @@ -2520,7 +2521,7 @@ static void irq_complete_move(struct irq_desc **descp) void irq_force_complete_move(int irq) { struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg = desc->chip_data; + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); if (!cfg) return; @@ -2588,7 +2589,7 @@ static void eoi_ioapic_irq(struct irq_desc *desc) unsigned int irq; irq = desc->irq; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); raw_spin_lock_irqsave(&ioapic_lock, flags); __eoi_ioapic_irq(irq, cfg); @@ -2644,7 +2645,7 @@ static void ack_apic_level(unsigned int irq) * we use the above logic (mask+edge followed by unmask+level) from * Manfred Spraul to clear the remote IRR. */ - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); i = cfg->vector; v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1)); @@ -2695,7 +2696,7 @@ static void ack_apic_level(unsigned int irq) * accurate and is causing problems then it is a hardware bug * and you can go talk to the chipset vendor about it. */ - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); if (!io_apic_level_ack_pending(cfg)) move_masked_irq(irq); unmask_IO_APIC_irq_desc(desc); @@ -2763,7 +2764,7 @@ static inline void init_IO_APIC_traps(void) * 0x80, because int 0x80 is hm, kind of importantish. ;) */ for_each_irq_desc(irq, desc) { - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); if (IO_APIC_IRQ(irq) && cfg && !cfg->vector) { /* * Hmm.. We don't have an entry for this, @@ -2917,7 +2918,7 @@ int timer_through_8259 __initdata; static inline void __init check_timer(void) { struct irq_desc *desc = irq_to_desc(0); - struct irq_cfg *cfg = desc->chip_data; + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); int node = cpu_to_node(0); int apic1, pin1, apic2, pin2; unsigned long flags; @@ -3250,13 +3251,13 @@ unsigned int create_irq_nr(unsigned int irq_want, int node) printk(KERN_INFO "can not get irq_desc for %d\n", new); continue; } - cfg_new = desc_new->chip_data; + cfg_new = get_irq_desc_chip_data(desc_new); if (cfg_new->vector != 0) continue; desc_new = move_irq_desc(desc_new, node); - cfg_new = desc_new->chip_data; + cfg_new = get_irq_desc_chip_data(desc_new); if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0) irq = new; @@ -3381,7 +3382,7 @@ static int set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) if (set_desc_affinity(desc, mask, &dest)) return -1; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); __get_cached_msi_msg(desc->irq_data.msi_desc, &msg); @@ -3403,7 +3404,7 @@ static int ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg = desc->chip_data; + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); unsigned int dest; struct irte irte; @@ -3595,7 +3596,7 @@ static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) if (set_desc_affinity(desc, mask, &dest)) return -1; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); dmar_msi_read(irq, &msg); @@ -3650,7 +3651,7 @@ static int hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) if (set_desc_affinity(desc, mask, &dest)) return -1; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); hpet_msi_read(irq, &msg); @@ -3756,7 +3757,7 @@ static int set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) if (set_desc_affinity(desc, mask, &dest)) return -1; - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); target_ht_irq(irq, dest, cfg->vector); @@ -3903,7 +3904,7 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq, * IRQs < 16 are already in the irq_2_pin[] map */ if (irq >= legacy_pic->nr_legacy_irqs) { - cfg = desc->chip_data; + cfg = get_irq_desc_chip_data(desc); if (add_pin_to_irq_node_nopanic(cfg, node, ioapic, pin)) { printk(KERN_INFO "can not add pin %d for irq %d\n", pin, irq); diff --git a/arch/x86/kernel/uv_irq.c b/arch/x86/kernel/uv_irq.c index 1132129..2233a42 100644 --- a/arch/x86/kernel/uv_irq.c +++ b/arch/x86/kernel/uv_irq.c @@ -209,7 +209,7 @@ static void arch_disable_uv_irq(int mmr_pnode, unsigned long mmr_offset) static int uv_set_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg = desc->chip_data; + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); unsigned int dest; unsigned long mmr_value; struct uv_IO_APIC_route_entry *entry; -- cgit v0.10.2 From dd5f15e5cf104c9170b76ae3274f35b42a3e4161 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 15:18:35 +0200 Subject: x86: Cleanup io_apic Sanitize functions. Remove irq_desc pointer magic. Preparatory patch for further cleanups. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index fa0d92a..3c4dee8 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -572,11 +572,6 @@ static void __unmask_and_level_IO_APIC_irq(struct irq_pin_list *entry) IO_APIC_REDIR_LEVEL_TRIGGER, NULL); } -static void __unmask_IO_APIC_irq(struct irq_cfg *cfg) -{ - io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, 0, NULL); -} - static void io_apic_sync(struct irq_pin_list *entry) { /* @@ -588,44 +583,41 @@ static void io_apic_sync(struct irq_pin_list *entry) readl(&io_apic->data); } -static void __mask_IO_APIC_irq(struct irq_cfg *cfg) +static void mask_ioapic(struct irq_cfg *cfg) { + unsigned long flags; + + raw_spin_lock_irqsave(&ioapic_lock, flags); io_apic_modify_irq(cfg, ~0, IO_APIC_REDIR_MASKED, &io_apic_sync); + raw_spin_unlock_irqrestore(&ioapic_lock, flags); } -static void mask_IO_APIC_irq_desc(struct irq_desc *desc) +static void mask_ioapic_irq(unsigned int irq) { - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); - unsigned long flags; + struct irq_cfg *cfg = get_irq_chip_data(irq); - BUG_ON(!cfg); + mask_ioapic(cfg); +} - raw_spin_lock_irqsave(&ioapic_lock, flags); - __mask_IO_APIC_irq(cfg); - raw_spin_unlock_irqrestore(&ioapic_lock, flags); +static void __unmask_ioapic(struct irq_cfg *cfg) +{ + io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, 0, NULL); } -static void unmask_IO_APIC_irq_desc(struct irq_desc *desc) +static void unmask_ioapic(struct irq_cfg *cfg) { - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); unsigned long flags; raw_spin_lock_irqsave(&ioapic_lock, flags); - __unmask_IO_APIC_irq(cfg); + __unmask_ioapic(cfg); raw_spin_unlock_irqrestore(&ioapic_lock, flags); } -static void mask_IO_APIC_irq(unsigned int irq) +static void unmask_ioapic_irq(unsigned int irq) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_cfg *cfg = get_irq_chip_data(irq); - mask_IO_APIC_irq_desc(desc); -} -static void unmask_IO_APIC_irq(unsigned int irq) -{ - struct irq_desc *desc = irq_to_desc(irq); - - unmask_IO_APIC_irq_desc(desc); + unmask_ioapic(cfg); } static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin) @@ -2239,7 +2231,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) was_pending = 1; } cfg = irq_cfg(irq); - __unmask_IO_APIC_irq(cfg); + __unmask_ioapic(cfg); raw_spin_unlock_irqrestore(&ioapic_lock, flags); return was_pending; @@ -2498,10 +2490,8 @@ unlock: irq_exit(); } -static void __irq_complete_move(struct irq_desc **descp, unsigned vector) +static void __irq_complete_move(struct irq_cfg *cfg, unsigned vector) { - struct irq_desc *desc = *descp; - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); unsigned me; if (likely(!cfg->move_in_progress)) @@ -2513,30 +2503,29 @@ static void __irq_complete_move(struct irq_desc **descp, unsigned vector) send_cleanup_vector(cfg); } -static void irq_complete_move(struct irq_desc **descp) +static void irq_complete_move(struct irq_cfg *cfg) { - __irq_complete_move(descp, ~get_irq_regs()->orig_ax); + __irq_complete_move(cfg, ~get_irq_regs()->orig_ax); } void irq_force_complete_move(int irq) { - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); + struct irq_cfg *cfg = get_irq_chip_data(irq); if (!cfg) return; - __irq_complete_move(&desc, cfg->vector); + __irq_complete_move(cfg, cfg->vector); } #else -static inline void irq_complete_move(struct irq_desc **descp) {} +static inline void irq_complete_move(struct irq_cfg *cfg) { } #endif static void ack_apic_edge(unsigned int irq) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_cfg *cfg = get_irq_chip_data(irq); - irq_complete_move(&desc); + irq_complete_move(cfg); move_native_irq(irq); ack_APIC_irq(); } @@ -2559,10 +2548,12 @@ atomic_t irq_mis_count; * Otherwise, we simulate the EOI message manually by changing the trigger * mode to edge and then back to level, with RTE being masked during this. */ -static void __eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) +static void eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) { struct irq_pin_list *entry; + unsigned long flags; + raw_spin_lock_irqsave(&ioapic_lock, flags); for_each_irq_pin(entry, cfg->irq_2_pin) { if (mp_ioapics[entry->apic].apicver >= 0x20) { /* @@ -2580,36 +2571,22 @@ static void __eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) __unmask_and_level_IO_APIC_irq(entry); } } -} - -static void eoi_ioapic_irq(struct irq_desc *desc) -{ - struct irq_cfg *cfg; - unsigned long flags; - unsigned int irq; - - irq = desc->irq; - cfg = get_irq_desc_chip_data(desc); - - raw_spin_lock_irqsave(&ioapic_lock, flags); - __eoi_ioapic_irq(irq, cfg); raw_spin_unlock_irqrestore(&ioapic_lock, flags); } static void ack_apic_level(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); + struct irq_cfg *cfg = get_irq_desc_chip_data(desc); + int i, do_unmask_irq = 0; unsigned long v; - int i; - struct irq_cfg *cfg; - int do_unmask_irq = 0; - irq_complete_move(&desc); + irq_complete_move(cfg); #ifdef CONFIG_GENERIC_PENDING_IRQ /* If we are moving the irq we need to mask it */ if (unlikely(desc->status & IRQ_MOVE_PENDING)) { do_unmask_irq = 1; - mask_IO_APIC_irq_desc(desc); + mask_ioapic(cfg); } #endif @@ -2645,7 +2622,6 @@ static void ack_apic_level(unsigned int irq) * we use the above logic (mask+edge followed by unmask+level) from * Manfred Spraul to clear the remote IRR. */ - cfg = get_irq_desc_chip_data(desc); i = cfg->vector; v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1)); @@ -2665,7 +2641,7 @@ static void ack_apic_level(unsigned int irq) if (!(v & (1 << (i & 0x1f)))) { atomic_inc(&irq_mis_count); - eoi_ioapic_irq(desc); + eoi_ioapic_irq(irq, cfg); } /* Now we can move and renable the irq */ @@ -2696,10 +2672,9 @@ static void ack_apic_level(unsigned int irq) * accurate and is causing problems then it is a hardware bug * and you can go talk to the chipset vendor about it. */ - cfg = get_irq_desc_chip_data(desc); if (!io_apic_level_ack_pending(cfg)) move_masked_irq(irq); - unmask_IO_APIC_irq_desc(desc); + unmask_ioapic(cfg); } } @@ -2711,18 +2686,18 @@ static void ir_ack_apic_edge(unsigned int irq) static void ir_ack_apic_level(unsigned int irq) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_cfg *cfg = get_irq_chip_data(irq); ack_APIC_irq(); - eoi_ioapic_irq(desc); + eoi_ioapic_irq(irq, cfg); } #endif /* CONFIG_INTR_REMAP */ static struct irq_chip ioapic_chip __read_mostly = { .name = "IO-APIC", .startup = startup_ioapic_irq, - .mask = mask_IO_APIC_irq, - .unmask = unmask_IO_APIC_irq, + .mask = mask_ioapic_irq, + .unmask = unmask_ioapic_irq, .ack = ack_apic_edge, .eoi = ack_apic_level, #ifdef CONFIG_SMP @@ -2734,8 +2709,8 @@ static struct irq_chip ioapic_chip __read_mostly = { static struct irq_chip ir_ioapic_chip __read_mostly = { .name = "IR-IO-APIC", .startup = startup_ioapic_irq, - .mask = mask_IO_APIC_irq, - .unmask = unmask_IO_APIC_irq, + .mask = mask_ioapic_irq, + .unmask = unmask_ioapic_irq, #ifdef CONFIG_INTR_REMAP .ack = ir_ack_apic_edge, .eoi = ir_ack_apic_level, @@ -2996,7 +2971,7 @@ static inline void __init check_timer(void) int idx; idx = find_irq_entry(apic1, pin1, mp_INT); if (idx != -1 && irq_trigger(idx)) - unmask_IO_APIC_irq_desc(desc); + unmask_ioapic(cfg); } if (timer_irq_works()) { if (nmi_watchdog == NMI_IO_APIC) { -- cgit v0.10.2 From 61a38ce3f59cdb4654e9444329195bd57c3baf74 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 16:00:34 +0200 Subject: x86: io_apic: Convert startup to new irq_chip function Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 3c4dee8..5ced690 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -2218,11 +2218,10 @@ static int __init timer_irq_works(void) * an edge even if it isn't on the 8259A... */ -static unsigned int startup_ioapic_irq(unsigned int irq) +static unsigned int startup_ioapic_irq(struct irq_data *data) { - int was_pending = 0; + int was_pending = 0, irq = data->irq; unsigned long flags; - struct irq_cfg *cfg; raw_spin_lock_irqsave(&ioapic_lock, flags); if (irq < legacy_pic->nr_legacy_irqs) { @@ -2230,8 +2229,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) if (legacy_pic->irq_pending(irq)) was_pending = 1; } - cfg = irq_cfg(irq); - __unmask_ioapic(cfg); + __unmask_ioapic(data->chip_data); raw_spin_unlock_irqrestore(&ioapic_lock, flags); return was_pending; @@ -2695,7 +2693,7 @@ static void ir_ack_apic_level(unsigned int irq) static struct irq_chip ioapic_chip __read_mostly = { .name = "IO-APIC", - .startup = startup_ioapic_irq, + .irq_startup = startup_ioapic_irq, .mask = mask_ioapic_irq, .unmask = unmask_ioapic_irq, .ack = ack_apic_edge, @@ -2708,7 +2706,7 @@ static struct irq_chip ioapic_chip __read_mostly = { static struct irq_chip ir_ioapic_chip __read_mostly = { .name = "IR-IO-APIC", - .startup = startup_ioapic_irq, + .irq_startup = startup_ioapic_irq, .mask = mask_ioapic_irq, .unmask = unmask_ioapic_irq, #ifdef CONFIG_INTR_REMAP -- cgit v0.10.2 From 90297c5fe71d32a2a0ead38bd8f6b1112a2e5ac0 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 16:03:54 +0200 Subject: x86: ioapic: Convert mask to new irq_chip function Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 5ced690..b8b013f 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -592,11 +592,9 @@ static void mask_ioapic(struct irq_cfg *cfg) raw_spin_unlock_irqrestore(&ioapic_lock, flags); } -static void mask_ioapic_irq(unsigned int irq) +static void mask_ioapic_irq(struct irq_data *data) { - struct irq_cfg *cfg = get_irq_chip_data(irq); - - mask_ioapic(cfg); + mask_ioapic(data->chip_data); } static void __unmask_ioapic(struct irq_cfg *cfg) @@ -613,11 +611,9 @@ static void unmask_ioapic(struct irq_cfg *cfg) raw_spin_unlock_irqrestore(&ioapic_lock, flags); } -static void unmask_ioapic_irq(unsigned int irq) +static void unmask_ioapic_irq(struct irq_data *data) { - struct irq_cfg *cfg = get_irq_chip_data(irq); - - unmask_ioapic(cfg); + unmask_ioapic(data->chip_data); } static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin) @@ -2235,10 +2231,9 @@ static unsigned int startup_ioapic_irq(struct irq_data *data) return was_pending; } -static int ioapic_retrigger_irq(unsigned int irq) +static int ioapic_retrigger_irq(struct irq_data *data) { - - struct irq_cfg *cfg = irq_cfg(irq); + struct irq_cfg *cfg = data->chip_data; unsigned long flags; raw_spin_lock_irqsave(&vector_lock, flags); @@ -2519,12 +2514,10 @@ void irq_force_complete_move(int irq) static inline void irq_complete_move(struct irq_cfg *cfg) { } #endif -static void ack_apic_edge(unsigned int irq) +static void ack_apic_edge(struct irq_data *data) { - struct irq_cfg *cfg = get_irq_chip_data(irq); - - irq_complete_move(cfg); - move_native_irq(irq); + irq_complete_move(data->chip_data); + move_native_irq(data->irq); ack_APIC_irq(); } @@ -2572,11 +2565,11 @@ static void eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) raw_spin_unlock_irqrestore(&ioapic_lock, flags); } -static void ack_apic_level(unsigned int irq) +static void ack_apic_level(struct irq_data *data) { + struct irq_cfg *cfg = data->chip_data; + int i, do_unmask_irq = 0, irq = data->irq; struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); - int i, do_unmask_irq = 0; unsigned long v; irq_complete_move(cfg); @@ -2677,46 +2670,44 @@ static void ack_apic_level(unsigned int irq) } #ifdef CONFIG_INTR_REMAP -static void ir_ack_apic_edge(unsigned int irq) +static void ir_ack_apic_edge(struct irq_data *data) { ack_APIC_irq(); } -static void ir_ack_apic_level(unsigned int irq) +static void ir_ack_apic_level(struct irq_data *data) { - struct irq_cfg *cfg = get_irq_chip_data(irq); - ack_APIC_irq(); - eoi_ioapic_irq(irq, cfg); + eoi_ioapic_irq(data->irq, data->chip_data); } #endif /* CONFIG_INTR_REMAP */ static struct irq_chip ioapic_chip __read_mostly = { .name = "IO-APIC", .irq_startup = startup_ioapic_irq, - .mask = mask_ioapic_irq, - .unmask = unmask_ioapic_irq, - .ack = ack_apic_edge, - .eoi = ack_apic_level, + .irq_mask = mask_ioapic_irq, + .irq_unmask = unmask_ioapic_irq, + .irq_ack = ack_apic_edge, + .irq_eoi = ack_apic_level, #ifdef CONFIG_SMP .set_affinity = set_ioapic_affinity_irq, #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static struct irq_chip ir_ioapic_chip __read_mostly = { .name = "IR-IO-APIC", .irq_startup = startup_ioapic_irq, - .mask = mask_ioapic_irq, - .unmask = unmask_ioapic_irq, + .irq_mask = mask_ioapic_irq, + .irq_unmask = unmask_ioapic_irq, #ifdef CONFIG_INTR_REMAP - .ack = ir_ack_apic_edge, - .eoi = ir_ack_apic_level, + .irq_ack = ir_ack_apic_edge, + .irq_eoi = ir_ack_apic_level, #ifdef CONFIG_SMP .set_affinity = set_ir_ioapic_affinity_irq, #endif #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static inline void init_IO_APIC_traps(void) @@ -2757,7 +2748,7 @@ static inline void init_IO_APIC_traps(void) * The local APIC irq-chip implementation: */ -static void mask_lapic_irq(unsigned int irq) +static void mask_lapic_irq(struct irq_data *data) { unsigned long v; @@ -2765,7 +2756,7 @@ static void mask_lapic_irq(unsigned int irq) apic_write(APIC_LVT0, v | APIC_LVT_MASKED); } -static void unmask_lapic_irq(unsigned int irq) +static void unmask_lapic_irq(struct irq_data *data) { unsigned long v; @@ -2773,16 +2764,16 @@ static void unmask_lapic_irq(unsigned int irq) apic_write(APIC_LVT0, v & ~APIC_LVT_MASKED); } -static void ack_lapic_irq(unsigned int irq) +static void ack_lapic_irq(struct irq_data *data) { ack_APIC_irq(); } static struct irq_chip lapic_chip __read_mostly = { .name = "local-APIC", - .mask = mask_lapic_irq, - .unmask = unmask_lapic_irq, - .ack = ack_lapic_irq, + .irq_mask = mask_lapic_irq, + .irq_unmask = unmask_lapic_irq, + .irq_ack = ack_lapic_irq, }; static void lapic_register_intr(int irq, struct irq_desc *desc) @@ -3417,11 +3408,11 @@ static struct irq_chip msi_chip = { .name = "PCI-MSI", .irq_unmask = unmask_msi_irq, .irq_mask = mask_msi_irq, - .ack = ack_apic_edge, + .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = set_msi_irq_affinity, #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static struct irq_chip msi_ir_chip = { @@ -3429,12 +3420,12 @@ static struct irq_chip msi_ir_chip = { .irq_unmask = unmask_msi_irq, .irq_mask = mask_msi_irq, #ifdef CONFIG_INTR_REMAP - .ack = ir_ack_apic_edge, + .irq_ack = ir_ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = ir_set_msi_irq_affinity, #endif #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; /* @@ -3589,11 +3580,11 @@ static struct irq_chip dmar_msi_type = { .name = "DMAR_MSI", .unmask = dmar_msi_unmask, .mask = dmar_msi_mask, - .ack = ack_apic_edge, + .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = dmar_msi_set_affinity, #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; int arch_setup_dmar_msi(unsigned int irq) @@ -3645,23 +3636,23 @@ static struct irq_chip ir_hpet_msi_type = { .unmask = hpet_msi_unmask, .mask = hpet_msi_mask, #ifdef CONFIG_INTR_REMAP - .ack = ir_ack_apic_edge, + .irq_ack = ir_ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = ir_set_msi_irq_affinity, #endif #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static struct irq_chip hpet_msi_type = { .name = "HPET_MSI", .unmask = hpet_msi_unmask, .mask = hpet_msi_mask, - .ack = ack_apic_edge, + .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = hpet_msi_set_affinity, #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; int arch_setup_hpet_msi(unsigned int irq, unsigned int id) @@ -3743,11 +3734,11 @@ static struct irq_chip ht_irq_chip = { .name = "PCI-HT", .mask = mask_ht_irq, .unmask = unmask_ht_irq, - .ack = ack_apic_edge, + .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = set_ht_irq_affinity, #endif - .retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) -- cgit v0.10.2 From d0fbca8f9304d1760fdc906b35b06e89fbdbb51f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 16:18:39 +0200 Subject: x86: ioapic/hpet: Convert to new chip functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/include/asm/hpet.h b/arch/x86/include/asm/hpet.h index 1d5c08a..2c392d6 100644 --- a/arch/x86/include/asm/hpet.h +++ b/arch/x86/include/asm/hpet.h @@ -74,10 +74,12 @@ extern void hpet_disable(void); extern unsigned int hpet_readl(unsigned int a); extern void force_hpet_resume(void); -extern void hpet_msi_unmask(unsigned int irq); -extern void hpet_msi_mask(unsigned int irq); -extern void hpet_msi_write(unsigned int irq, struct msi_msg *msg); -extern void hpet_msi_read(unsigned int irq, struct msi_msg *msg); +struct irq_data; +extern void hpet_msi_unmask(struct irq_data *data); +extern void hpet_msi_mask(struct irq_data *data); +struct hpet_dev; +extern void hpet_msi_write(struct hpet_dev *hdev, struct msi_msg *msg); +extern void hpet_msi_read(struct hpet_dev *hdev, struct msi_msg *msg); #ifdef CONFIG_PCI_MSI extern int arch_setup_hpet_msi(unsigned int irq, unsigned int id); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index b8b013f..49aa857 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3605,26 +3605,25 @@ int arch_setup_dmar_msi(unsigned int irq) #ifdef CONFIG_HPET_TIMER #ifdef CONFIG_SMP -static int hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) +static int hpet_msi_set_affinity(struct irq_data *data, + const struct cpumask *mask, bool force) { - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg; + struct irq_desc *desc = irq_to_desc(data->irq); + struct irq_cfg *cfg = data->chip_data; struct msi_msg msg; unsigned int dest; if (set_desc_affinity(desc, mask, &dest)) return -1; - cfg = get_irq_desc_chip_data(desc); - - hpet_msi_read(irq, &msg); + hpet_msi_read(data->handler_data, &msg); msg.data &= ~MSI_DATA_VECTOR_MASK; msg.data |= MSI_DATA_VECTOR(cfg->vector); msg.address_lo &= ~MSI_ADDR_DEST_ID_MASK; msg.address_lo |= MSI_ADDR_DEST_ID(dest); - hpet_msi_write(irq, &msg); + hpet_msi_write(data->handler_data, &msg); return 0; } @@ -3633,8 +3632,8 @@ static int hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) static struct irq_chip ir_hpet_msi_type = { .name = "IR-HPET_MSI", - .unmask = hpet_msi_unmask, - .mask = hpet_msi_mask, + .irq_unmask = hpet_msi_unmask, + .irq_mask = hpet_msi_mask, #ifdef CONFIG_INTR_REMAP .irq_ack = ir_ack_apic_edge, #ifdef CONFIG_SMP @@ -3646,20 +3645,19 @@ static struct irq_chip ir_hpet_msi_type = { static struct irq_chip hpet_msi_type = { .name = "HPET_MSI", - .unmask = hpet_msi_unmask, - .mask = hpet_msi_mask, + .irq_unmask = hpet_msi_unmask, + .irq_mask = hpet_msi_mask, .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP - .set_affinity = hpet_msi_set_affinity, + .irq_set_affinity = hpet_msi_set_affinity, #endif .irq_retrigger = ioapic_retrigger_irq, }; int arch_setup_hpet_msi(unsigned int irq, unsigned int id) { - int ret; struct msi_msg msg; - struct irq_desc *desc = irq_to_desc(irq); + int ret; if (intr_remapping_enabled) { struct intel_iommu *iommu = map_hpet_to_ir(id); @@ -3677,8 +3675,8 @@ int arch_setup_hpet_msi(unsigned int irq, unsigned int id) if (ret < 0) return ret; - hpet_msi_write(irq, &msg); - desc->status |= IRQ_MOVE_PCNTXT; + hpet_msi_write(get_irq_data(irq), &msg); + irq_set_status_flags(irq,IRQ_MOVE_PCNTXT); if (irq_remapped(irq)) set_irq_chip_and_handler_name(irq, &ir_hpet_msi_type, handle_edge_irq, "edge"); diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 7494999..efaf906 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -440,9 +440,9 @@ static int hpet_legacy_next_event(unsigned long delta, static DEFINE_PER_CPU(struct hpet_dev *, cpu_hpet_dev); static struct hpet_dev *hpet_devs; -void hpet_msi_unmask(unsigned int irq) +void hpet_msi_unmask(struct irq_data *data) { - struct hpet_dev *hdev = get_irq_data(irq); + struct hpet_dev *hdev = data->handler_data; unsigned int cfg; /* unmask it */ @@ -451,10 +451,10 @@ void hpet_msi_unmask(unsigned int irq) hpet_writel(cfg, HPET_Tn_CFG(hdev->num)); } -void hpet_msi_mask(unsigned int irq) +void hpet_msi_mask(struct irq_data *data) { + struct hpet_dev *hdev = data->handler_data; unsigned int cfg; - struct hpet_dev *hdev = get_irq_data(irq); /* mask it */ cfg = hpet_readl(HPET_Tn_CFG(hdev->num)); @@ -462,18 +462,14 @@ void hpet_msi_mask(unsigned int irq) hpet_writel(cfg, HPET_Tn_CFG(hdev->num)); } -void hpet_msi_write(unsigned int irq, struct msi_msg *msg) +void hpet_msi_write(struct hpet_dev *hdev, struct msi_msg *msg) { - struct hpet_dev *hdev = get_irq_data(irq); - hpet_writel(msg->data, HPET_Tn_ROUTE(hdev->num)); hpet_writel(msg->address_lo, HPET_Tn_ROUTE(hdev->num) + 4); } -void hpet_msi_read(unsigned int irq, struct msi_msg *msg) +void hpet_msi_read(struct hpet_dev *hdev, struct msi_msg *msg) { - struct hpet_dev *hdev = get_irq_data(irq); - msg->data = hpet_readl(HPET_Tn_ROUTE(hdev->num)); msg->address_lo = hpet_readl(HPET_Tn_ROUTE(hdev->num) + 4); msg->address_hi = 0; -- cgit v0.10.2 From 5c2837fbaa609e615ef9a1c58a4cd26ce90be35b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 17:15:11 +0200 Subject: dmar: Convert to new irq chip functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: David Woodhouse diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c index dbb43424..00b19a4 100644 --- a/arch/ia64/kernel/msi_ia64.c +++ b/arch/ia64/kernel/msi_ia64.c @@ -160,8 +160,8 @@ static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) static struct irq_chip dmar_msi_type = { .name = "DMAR_MSI", - .unmask = dmar_msi_unmask, - .mask = dmar_msi_mask, + .irq_unmask = dmar_msi_unmask, + .irq_mask = dmar_msi_mask, .ack = ia64_ack_msi_irq, #ifdef CONFIG_SMP .set_affinity = dmar_msi_set_affinity, diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 49aa857..72b253e 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3578,8 +3578,8 @@ static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) static struct irq_chip dmar_msi_type = { .name = "DMAR_MSI", - .unmask = dmar_msi_unmask, - .mask = dmar_msi_mask, + .irq_unmask = dmar_msi_unmask, + .irq_mask = dmar_msi_mask, .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = dmar_msi_set_affinity, diff --git a/drivers/pci/dmar.c b/drivers/pci/dmar.c index 0a19708..3de3a43 100644 --- a/drivers/pci/dmar.c +++ b/drivers/pci/dmar.c @@ -1221,9 +1221,9 @@ const char *dmar_get_fault_reason(u8 fault_reason, int *fault_type) } } -void dmar_msi_unmask(unsigned int irq) +void dmar_msi_unmask(struct irq_data *data) { - struct intel_iommu *iommu = get_irq_data(irq); + struct intel_iommu *iommu = irq_data_get_irq_data(data); unsigned long flag; /* unmask it */ @@ -1234,10 +1234,10 @@ void dmar_msi_unmask(unsigned int irq) spin_unlock_irqrestore(&iommu->register_lock, flag); } -void dmar_msi_mask(unsigned int irq) +void dmar_msi_mask(struct irq_data *data) { unsigned long flag; - struct intel_iommu *iommu = get_irq_data(irq); + struct intel_iommu *iommu = irq_data_get_irq_data(data); /* mask it */ spin_lock_irqsave(&iommu->register_lock, flag); diff --git a/include/linux/dmar.h b/include/linux/dmar.h index d7cecc9..cb86aa1 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -187,8 +187,9 @@ static inline int set_msi_sid(struct irte *irte, struct pci_dev *dev) /* Can't use the common MSI interrupt functions * since DMAR is not a pci device */ -extern void dmar_msi_unmask(unsigned int irq); -extern void dmar_msi_mask(unsigned int irq); +struct irq_data; +extern void dmar_msi_unmask(struct irq_data *data); +extern void dmar_msi_mask(struct irq_data *data); extern void dmar_msi_read(int irq, struct msi_msg *msg); extern void dmar_msi_write(int irq, struct msi_msg *msg); extern int dmar_set_interrupt(struct intel_iommu *iommu); -- cgit v0.10.2 From e9f7ac664bfc36685a8eb3315ec21c067d0cee36 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 17:22:09 +0200 Subject: ht: Convert to new irq_chip functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Jesse Barnes diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 72b253e..5579f3f 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3730,8 +3730,8 @@ static int set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) static struct irq_chip ht_irq_chip = { .name = "PCI-HT", - .mask = mask_ht_irq, - .unmask = unmask_ht_irq, + .irq_mask = mask_ht_irq, + .irq_unmask = unmask_ht_irq, .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP .set_affinity = set_ht_irq_affinity, diff --git a/drivers/pci/htirq.c b/drivers/pci/htirq.c index 98abf8b..834842a 100644 --- a/drivers/pci/htirq.c +++ b/drivers/pci/htirq.c @@ -57,28 +57,22 @@ void fetch_ht_irq_msg(unsigned int irq, struct ht_irq_msg *msg) *msg = cfg->msg; } -void mask_ht_irq(unsigned int irq) +void mask_ht_irq(struct irq_data *data) { - struct ht_irq_cfg *cfg; - struct ht_irq_msg msg; - - cfg = get_irq_data(irq); + struct ht_irq_cfg *cfg = irq_data_get_irq_data(data); + struct ht_irq_msg msg = cfg->msg; - msg = cfg->msg; msg.address_lo |= 1; - write_ht_irq_msg(irq, &msg); + write_ht_irq_msg(data->irq, &msg); } -void unmask_ht_irq(unsigned int irq) +void unmask_ht_irq(struct irq_data *data) { - struct ht_irq_cfg *cfg; - struct ht_irq_msg msg; - - cfg = get_irq_data(irq); + struct ht_irq_cfg *cfg = irq_data_get_irq_data(data); + struct ht_irq_msg msg = cfg->msg; - msg = cfg->msg; msg.address_lo &= ~1; - write_ht_irq_msg(irq, &msg); + write_ht_irq_msg(data->irq, &msg); } /** diff --git a/include/linux/htirq.h b/include/linux/htirq.h index c96ea46..70a1dbb 100644 --- a/include/linux/htirq.h +++ b/include/linux/htirq.h @@ -9,8 +9,9 @@ struct ht_irq_msg { /* Helper functions.. */ void fetch_ht_irq_msg(unsigned int irq, struct ht_irq_msg *msg); void write_ht_irq_msg(unsigned int irq, struct ht_irq_msg *msg); -void mask_ht_irq(unsigned int irq); -void unmask_ht_irq(unsigned int irq); +struct irq_data; +void mask_ht_irq(struct irq_data *data); +void unmask_ht_irq(struct irq_data *data); /* The arch hook for getting things started */ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev); -- cgit v0.10.2 From 60c69948e5b6357ac0d5ef2a2d0ce31c173c3c64 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 17:28:38 +0200 Subject: x86: ioapic: Clean up the direct access to irq_desc Most of it is useless pseudo optimization. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 5579f3f..82c3c66 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -150,10 +150,7 @@ static struct irq_cfg irq_cfgx[NR_IRQS]; int __init arch_early_irq_init(void) { struct irq_cfg *cfg; - struct irq_desc *desc; - int count; - int node; - int i; + int count, node, i; if (!legacy_pic->nr_legacy_irqs) { nr_irqs_gsi = 0; @@ -165,8 +162,7 @@ int __init arch_early_irq_init(void) node = cpu_to_node(0); for (i = 0; i < count; i++) { - desc = irq_to_desc(i); - desc->chip_data = &cfg[i]; + set_irq_chip_data(i, &cfg[i]); zalloc_cpumask_var_node(&cfg[i].domain, GFP_NOWAIT, node); zalloc_cpumask_var_node(&cfg[i].old_domain, GFP_NOWAIT, node); /* @@ -185,14 +181,7 @@ int __init arch_early_irq_init(void) #ifdef CONFIG_SPARSE_IRQ struct irq_cfg *irq_cfg(unsigned int irq) { - struct irq_cfg *cfg = NULL; - struct irq_desc *desc; - - desc = irq_to_desc(irq); - if (desc) - cfg = get_irq_desc_chip_data(desc); - - return cfg; + return get_irq_chip_data(irq); } static struct irq_cfg *get_one_free_irq_cfg(int node) @@ -1316,17 +1305,17 @@ static inline int IO_APIC_irq_trigger(int irq) } #endif -static void ioapic_register_intr(int irq, struct irq_desc *desc, unsigned long trigger) +static void ioapic_register_intr(unsigned int irq, unsigned long trigger) { if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) || trigger == IOAPIC_LEVEL) - desc->status |= IRQ_LEVEL; + irq_set_status_flags(irq, IRQ_LEVEL); else - desc->status &= ~IRQ_LEVEL; + irq_clear_status_flags(irq, IRQ_LEVEL); if (irq_remapped(irq)) { - desc->status |= IRQ_MOVE_PCNTXT; + irq_set_status_flags(irq, IRQ_MOVE_PCNTXT); if (trigger) set_irq_chip_and_handler_name(irq, &ir_ioapic_chip, handle_fasteoi_irq, @@ -1406,18 +1395,14 @@ int setup_ioapic_entry(int apic_id, int irq, return 0; } -static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq_desc *desc, - int trigger, int polarity) +static void setup_ioapic_irq(int apic_id, int pin, unsigned int irq, + struct irq_cfg *cfg, int trigger, int polarity) { - struct irq_cfg *cfg; struct IO_APIC_route_entry entry; unsigned int dest; if (!IO_APIC_IRQ(irq)) return; - - cfg = get_irq_desc_chip_data(desc); - /* * For legacy irqs, cfg->domain starts with cpu 0 for legacy * controllers like 8259. Now that IO-APIC can handle this irq, update @@ -1446,7 +1431,7 @@ static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq return; } - ioapic_register_intr(irq, desc, trigger); + ioapic_register_intr(irq, trigger); if (irq < legacy_pic->nr_legacy_irqs) legacy_pic->mask(irq); @@ -1511,8 +1496,8 @@ static void __init setup_IO_APIC_irqs(void) * don't mark it in pin_programmed, so later acpi could * set it correctly when irq < 16 */ - setup_IO_APIC_irq(apic_id, pin, irq, desc, - irq_trigger(idx), irq_polarity(idx)); + setup_ioapic_irq(apic_id, pin, irq, cfg, irq_trigger(idx), + irq_polarity(idx)); } if (notcon) @@ -1566,7 +1551,7 @@ void setup_IO_APIC_irq_extra(u32 gsi) } set_bit(pin, mp_ioapic_routing[apic_id].pin_programmed); - setup_IO_APIC_irq(apic_id, pin, irq, desc, + setup_ioapic_irq(apic_id, pin, irq, cfg, irq_trigger(idx), irq_polarity(idx)); } @@ -2776,9 +2761,9 @@ static struct irq_chip lapic_chip __read_mostly = { .irq_ack = ack_lapic_irq, }; -static void lapic_register_intr(int irq, struct irq_desc *desc) +static void lapic_register_intr(int irq) { - desc->status &= ~IRQ_LEVEL; + irq_clear_status_flags(irq, IRQ_LEVEL); set_irq_chip_and_handler_name(irq, &lapic_chip, handle_edge_irq, "edge"); } @@ -2881,8 +2866,7 @@ int timer_through_8259 __initdata; */ static inline void __init check_timer(void) { - struct irq_desc *desc = irq_to_desc(0); - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); + struct irq_cfg *cfg = get_irq_chip_data(0); int node = cpu_to_node(0); int apic1, pin1, apic2, pin2; unsigned long flags; @@ -2952,7 +2936,7 @@ static inline void __init check_timer(void) add_pin_to_irq_node(cfg, node, apic1, pin1); setup_timer_IRQ0_pin(apic1, pin1, cfg->vector); } else { - /* for edge trigger, setup_IO_APIC_irq already + /* for edge trigger, setup_ioapic_irq already * leave it unmasked. * so only need to unmask if it is level-trigger * do we really have level trigger timer? @@ -3020,7 +3004,7 @@ static inline void __init check_timer(void) apic_printk(APIC_QUIET, KERN_INFO "...trying to set up timer as Virtual Wire IRQ...\n"); - lapic_register_intr(0, desc); + lapic_register_intr(0); apic_write(APIC_LVT0, APIC_DM_FIXED | cfg->vector); /* Fixed mode */ legacy_pic->unmask(0); @@ -3457,8 +3441,8 @@ static int msi_alloc_irte(struct pci_dev *dev, int irq, int nvec) static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc, int irq) { - int ret; struct msi_msg msg; + int ret; ret = msi_compose_msg(dev, irq, &msg, -1); if (ret < 0) @@ -3468,11 +3452,7 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc, int irq) write_msi_msg(irq, &msg); if (irq_remapped(irq)) { - struct irq_desc *desc = irq_to_desc(irq); - /* - * irq migration in process context - */ - desc->status |= IRQ_MOVE_PCNTXT; + irq_set_status_flags(irq, IRQ_MOVE_PCNTXT); set_irq_chip_and_handler_name(irq, &msi_ir_chip, handle_edge_irq, "edge"); } else set_irq_chip_and_handler_name(irq, &msi_chip, handle_edge_irq, "edge"); @@ -3484,13 +3464,10 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc, int irq) int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) { - unsigned int irq; - int ret, sub_handle; + int node, ret, sub_handle, index = 0; + unsigned int irq, irq_want; struct msi_desc *msidesc; - unsigned int irq_want; struct intel_iommu *iommu = NULL; - int index = 0; - int node; /* x86 doesn't support multiple MSI yet */ if (type == PCI_CAP_ID_MSI && nvec > 1) @@ -3676,7 +3653,7 @@ int arch_setup_hpet_msi(unsigned int irq, unsigned int id) return ret; hpet_msi_write(get_irq_data(irq), &msg); - irq_set_status_flags(irq,IRQ_MOVE_PCNTXT); + irq_set_status_flags(irq, IRQ_MOVE_PCNTXT); if (irq_remapped(irq)) set_irq_chip_and_handler_name(irq, &ir_hpet_msi_type, handle_edge_irq, "edge"); @@ -3862,11 +3839,12 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq, trigger = irq_attr->trigger; polarity = irq_attr->polarity; + cfg = get_irq_desc_chip_data(desc); + /* * IRQs < 16 are already in the irq_2_pin[] map */ if (irq >= legacy_pic->nr_legacy_irqs) { - cfg = get_irq_desc_chip_data(desc); if (add_pin_to_irq_node_nopanic(cfg, node, ioapic, pin)) { printk(KERN_INFO "can not add pin %d for irq %d\n", pin, irq); @@ -3874,7 +3852,7 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq, } } - setup_IO_APIC_irq(ioapic, pin, irq, desc, trigger, polarity); + setup_ioapic_irq(ioapic, pin, irq, cfg, trigger, polarity); return 0; } @@ -4258,13 +4236,12 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base) void __init pre_init_apic_IRQ0(void) { struct irq_cfg *cfg; - struct irq_desc *desc; printk(KERN_INFO "Early APIC setup for system timer0\n"); #ifndef CONFIG_SMP phys_cpu_present_map = physid_mask_of_physid(boot_cpu_physical_apicid); #endif - desc = irq_to_desc_alloc_node(0, 0); + irq_to_desc_alloc_node(0, 0); setup_local_APIC(); @@ -4272,5 +4249,5 @@ void __init pre_init_apic_IRQ0(void) add_pin_to_irq_node(cfg, 0, 0, 0); set_irq_chip_and_handler_name(0, &ioapic_chip, handle_edge_irq, "edge"); - setup_IO_APIC_irq(0, 0, 0, desc, 0, 0); + setup_ioapic_irq(0, 0, 0, cfg, 0, 0); } -- cgit v0.10.2 From f7e909eae444ff733ecc5628af76d89c363ab480 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 21:40:23 +0200 Subject: x86: Prepare the affinity common functions for taking struct irq_data * While at it rename it to sensible function names and fix the return value from unsigned to int for __ioapic_set_affinity (set_desc_affinity). Returning -1 in a function returning unsigned int is somewhat strange. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index 46c0fe0..76848f2 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -95,9 +95,9 @@ extern struct irq_cfg *irq_cfg(unsigned int); extern int assign_irq_vector(int, struct irq_cfg *, const struct cpumask *); extern void send_cleanup_vector(struct irq_cfg *); -struct irq_desc; -extern unsigned int set_desc_affinity(struct irq_desc *, const struct cpumask *, - unsigned int *dest_id); +struct irq_data; +int __ioapic_set_affinity(struct irq_data *, const struct cpumask *, + unsigned int *dest_id); extern int IO_APIC_get_PCI_irq_vector(int bus, int devfn, int pin, struct io_apic_irq_attr *irq_attr); extern void setup_ioapic_dest(void); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 82c3c66..60ca9a4 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -2279,65 +2279,46 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq } /* - * Either sets desc->affinity to a valid value, and returns + * Either sets data->affinity to a valid value, and returns * ->cpu_mask_to_apicid of that in dest_id, or returns -1 and - * leaves desc->affinity untouched. + * leaves data->affinity untouched. */ -unsigned int -set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask, - unsigned int *dest_id) +int __ioapic_set_affinity(struct irq_data *data, const struct cpumask *mask, + unsigned int *dest_id) { - struct irq_cfg *cfg; - unsigned int irq; + struct irq_cfg *cfg = data->chip_data; if (!cpumask_intersects(mask, cpu_online_mask)) return -1; - irq = desc->irq; - cfg = get_irq_desc_chip_data(desc); - if (assign_irq_vector(irq, cfg, mask)) + if (assign_irq_vector(data->irq, data->chip_data, mask)) return -1; - cpumask_copy(desc->affinity, mask); + cpumask_copy(data->affinity, mask); - *dest_id = apic->cpu_mask_to_apicid_and(desc->affinity, cfg->domain); + *dest_id = apic->cpu_mask_to_apicid_and(mask, cfg->domain); return 0; } static int -set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) +ioapic_set_affinity(struct irq_data *data, const struct cpumask *mask, + bool force) { - struct irq_cfg *cfg; + unsigned int dest, irq = data->irq; unsigned long flags; - unsigned int dest; - unsigned int irq; - int ret = -1; - - irq = desc->irq; - cfg = get_irq_desc_chip_data(desc); + int ret; raw_spin_lock_irqsave(&ioapic_lock, flags); - ret = set_desc_affinity(desc, mask, &dest); + ret = __ioapic_set_affinity(data, mask, &dest); if (!ret) { /* Only the high 8 bits are valid. */ dest = SET_APIC_LOGICAL_ID(dest); - __target_IO_APIC_irq(irq, dest, cfg); + __target_IO_APIC_irq(irq, dest, data->chip_data); } raw_spin_unlock_irqrestore(&ioapic_lock, flags); - return ret; } -static int -set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) -{ - struct irq_desc *desc; - - desc = irq_to_desc(irq); - - return set_ioapic_affinity_irq_desc(desc, mask); -} - #ifdef CONFIG_INTR_REMAP /* @@ -2668,16 +2649,16 @@ static void ir_ack_apic_level(struct irq_data *data) #endif /* CONFIG_INTR_REMAP */ static struct irq_chip ioapic_chip __read_mostly = { - .name = "IO-APIC", - .irq_startup = startup_ioapic_irq, - .irq_mask = mask_ioapic_irq, - .irq_unmask = unmask_ioapic_irq, - .irq_ack = ack_apic_edge, - .irq_eoi = ack_apic_level, + .name = "IO-APIC", + .irq_startup = startup_ioapic_irq, + .irq_mask = mask_ioapic_irq, + .irq_unmask = unmask_ioapic_irq, + .irq_ack = ack_apic_edge, + .irq_eoi = ack_apic_level, #ifdef CONFIG_SMP - .set_affinity = set_ioapic_affinity_irq, + .irq_set_affinity = ioapic_set_affinity, #endif - .irq_retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static struct irq_chip ir_ioapic_chip __read_mostly = { @@ -3327,7 +3308,7 @@ static int set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) struct msi_msg msg; unsigned int dest; - if (set_desc_affinity(desc, mask, &dest)) + if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) return -1; cfg = get_irq_desc_chip_data(desc); @@ -3359,7 +3340,7 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) if (get_irte(irq, &irte)) return -1; - if (set_desc_affinity(desc, mask, &dest)) + if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) return -1; irte.vector = cfg->vector; @@ -3534,7 +3515,7 @@ static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) struct msi_msg msg; unsigned int dest; - if (set_desc_affinity(desc, mask, &dest)) + if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) return -1; cfg = get_irq_desc_chip_data(desc); @@ -3590,7 +3571,7 @@ static int hpet_msi_set_affinity(struct irq_data *data, struct msi_msg msg; unsigned int dest; - if (set_desc_affinity(desc, mask, &dest)) + if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) return -1; hpet_msi_read(data->handler_data, &msg); @@ -3693,7 +3674,7 @@ static int set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) struct irq_cfg *cfg; unsigned int dest; - if (set_desc_affinity(desc, mask, &dest)) + if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) return -1; cfg = get_irq_desc_chip_data(desc); @@ -4045,14 +4026,14 @@ void __init setup_ioapic_dest(void) */ if (desc->status & (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) - mask = desc->affinity; + mask = desc->irq_data.affinity; else mask = apic->target_cpus(); if (intr_remapping_enabled) set_ir_ioapic_affinity_irq_desc(desc, mask); else - set_ioapic_affinity_irq_desc(desc, mask); + ioapic_set_affinity(&desc->irq_data, mask, false); } } diff --git a/arch/x86/kernel/uv_irq.c b/arch/x86/kernel/uv_irq.c index 2233a42..56a42e1 100644 --- a/arch/x86/kernel/uv_irq.c +++ b/arch/x86/kernel/uv_irq.c @@ -216,7 +216,7 @@ static int uv_set_irq_affinity(unsigned int irq, const struct cpumask *mask) unsigned long mmr_offset; int mmr_pnode; - if (set_desc_affinity(desc, mask, &dest)) + if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) return -1; mmr_value = 0; -- cgit v0.10.2 From 5346b2a78fa3b900da672928978475acdd4632dc Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 21:49:03 +0200 Subject: x86: Convert msi affinity setting to new chip functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Jesse Barnes diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 60ca9a4..268c545 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3301,26 +3301,24 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, } #ifdef CONFIG_SMP -static int set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) +static int +msi_set_affinity(struct irq_data *data, const struct cpumask *mask, bool force) { - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg; + struct irq_cfg *cfg = data->chip_data; struct msi_msg msg; unsigned int dest; - if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) + if (__ioapic_set_affinity(data, mask, &dest)) return -1; - cfg = get_irq_desc_chip_data(desc); - - __get_cached_msi_msg(desc->irq_data.msi_desc, &msg); + __get_cached_msi_msg(data->msi_desc, &msg); msg.data &= ~MSI_DATA_VECTOR_MASK; msg.data |= MSI_DATA_VECTOR(cfg->vector); msg.address_lo &= ~MSI_ADDR_DEST_ID_MASK; msg.address_lo |= MSI_ADDR_DEST_ID(dest); - __write_msi_msg(desc->irq_data.msi_desc, &msg); + __write_msi_msg(data->msi_desc, &msg); return 0; } @@ -3370,14 +3368,14 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) * which implement the MSI or MSI-X Capability Structure. */ static struct irq_chip msi_chip = { - .name = "PCI-MSI", - .irq_unmask = unmask_msi_irq, - .irq_mask = mask_msi_irq, - .irq_ack = ack_apic_edge, + .name = "PCI-MSI", + .irq_unmask = unmask_msi_irq, + .irq_mask = mask_msi_irq, + .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP - .set_affinity = set_msi_irq_affinity, + .irq_set_affinity = msi_set_affinity, #endif - .irq_retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static struct irq_chip msi_ir_chip = { -- cgit v0.10.2 From f19f5ecc920215decfea54f26e3eb14064506675 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 21:50:22 +0200 Subject: x86: Convert remapped ioapic affinity setting to new irq chip function Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Suresh Siddha diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 268c545..49cc27d 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -2333,24 +2333,21 @@ ioapic_set_affinity(struct irq_data *data, const struct cpumask *mask, * the interrupt-remapping table entry. */ static int -migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) +ir_ioapic_set_affinity(struct irq_data *data, const struct cpumask *mask, + bool force) { - struct irq_cfg *cfg; + struct irq_cfg *cfg = data->chip_data; + unsigned int dest, irq = data->irq; struct irte irte; - unsigned int dest; - unsigned int irq; - int ret = -1; if (!cpumask_intersects(mask, cpu_online_mask)) - return ret; + return -EINVAL; - irq = desc->irq; if (get_irte(irq, &irte)) - return ret; + return -EBUSY; - cfg = get_irq_desc_chip_data(desc); if (assign_irq_vector(irq, cfg, mask)) - return ret; + return -EBUSY; dest = apic->cpu_mask_to_apicid_and(cfg->domain, mask); @@ -2365,29 +2362,14 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) if (cfg->move_in_progress) send_cleanup_vector(cfg); - cpumask_copy(desc->affinity, mask); - + cpumask_copy(data->affinity, mask); return 0; } -/* - * Migrates the IRQ destination in the process context. - */ -static int set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, - const struct cpumask *mask) -{ - return migrate_ioapic_irq_desc(desc, mask); -} -static int set_ir_ioapic_affinity_irq(unsigned int irq, - const struct cpumask *mask) -{ - struct irq_desc *desc = irq_to_desc(irq); - - return set_ir_ioapic_affinity_irq_desc(desc, mask); -} #else -static inline int set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, - const struct cpumask *mask) +static inline int +ir_ioapic_set_affinity(struct irq_data *data, const struct cpumask *mask, + bool force) { return 0; } @@ -2662,18 +2644,18 @@ static struct irq_chip ioapic_chip __read_mostly = { }; static struct irq_chip ir_ioapic_chip __read_mostly = { - .name = "IR-IO-APIC", - .irq_startup = startup_ioapic_irq, - .irq_mask = mask_ioapic_irq, - .irq_unmask = unmask_ioapic_irq, + .name = "IR-IO-APIC", + .irq_startup = startup_ioapic_irq, + .irq_mask = mask_ioapic_irq, + .irq_unmask = unmask_ioapic_irq, #ifdef CONFIG_INTR_REMAP - .irq_ack = ir_ack_apic_edge, - .irq_eoi = ir_ack_apic_level, + .irq_ack = ir_ack_apic_edge, + .irq_eoi = ir_ack_apic_level, #ifdef CONFIG_SMP - .set_affinity = set_ir_ioapic_affinity_irq, + .irq_set_affinity = ir_ioapic_set_affinity, #endif #endif - .irq_retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static inline void init_IO_APIC_traps(void) @@ -4029,7 +4011,7 @@ void __init setup_ioapic_dest(void) mask = apic->target_cpus(); if (intr_remapping_enabled) - set_ir_ioapic_affinity_irq_desc(desc, mask); + ir_ioapic_set_affinity(&desc->irq_data, mask, false); else ioapic_set_affinity(&desc->irq_data, mask, false); } -- cgit v0.10.2 From b5d1c465794f521c352d9c1a33159750c9c3fa84 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 22:15:49 +0200 Subject: x86: Convert remapped msi to new chip.irq_set_affinity function Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Suresh Siddha Cc: Jesse Barnes diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 49cc27d..13f8e28 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3310,17 +3310,17 @@ msi_set_affinity(struct irq_data *data, const struct cpumask *mask, bool force) * done in the process context using interrupt-remapping hardware. */ static int -ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) +ir_msi_set_affinity(struct irq_data *data, const struct cpumask *mask, + bool force) { - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); - unsigned int dest; + struct irq_cfg *cfg = data->chip_data; + unsigned int dest, irq = data->irq; struct irte irte; if (get_irte(irq, &irte)) return -1; - if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) + if (__ioapic_set_affinity(data, mask, &dest)) return -1; irte.vector = cfg->vector; @@ -3361,16 +3361,16 @@ static struct irq_chip msi_chip = { }; static struct irq_chip msi_ir_chip = { - .name = "IR-PCI-MSI", - .irq_unmask = unmask_msi_irq, - .irq_mask = mask_msi_irq, + .name = "IR-PCI-MSI", + .irq_unmask = unmask_msi_irq, + .irq_mask = mask_msi_irq, #ifdef CONFIG_INTR_REMAP - .irq_ack = ir_ack_apic_edge, + .irq_ack = ir_ack_apic_edge, #ifdef CONFIG_SMP - .set_affinity = ir_set_msi_irq_affinity, + .irq_set_affinity = ir_msi_set_affinity, #endif #endif - .irq_retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; /* @@ -3569,16 +3569,16 @@ static int hpet_msi_set_affinity(struct irq_data *data, #endif /* CONFIG_SMP */ static struct irq_chip ir_hpet_msi_type = { - .name = "IR-HPET_MSI", - .irq_unmask = hpet_msi_unmask, - .irq_mask = hpet_msi_mask, + .name = "IR-HPET_MSI", + .irq_unmask = hpet_msi_unmask, + .irq_mask = hpet_msi_mask, #ifdef CONFIG_INTR_REMAP - .irq_ack = ir_ack_apic_edge, + .irq_ack = ir_ack_apic_edge, #ifdef CONFIG_SMP - .set_affinity = ir_set_msi_irq_affinity, + .irq_set_affinity = ir_msi_set_affinity, #endif #endif - .irq_retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; static struct irq_chip hpet_msi_type = { -- cgit v0.10.2 From fe52b2d25919eaa01c51651a664f4f2ba6bd2a11 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 22:19:29 +0200 Subject: x86: Convert dmar affinity setting to new chip function Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: David Woodhouse diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 13f8e28..6f8ac4c 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3488,18 +3488,17 @@ void arch_teardown_msi_irq(unsigned int irq) #if defined (CONFIG_DMAR) || defined (CONFIG_INTR_REMAP) #ifdef CONFIG_SMP -static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) +static int +dmar_msi_set_affinity(struct irq_data *data, const struct cpumask *mask, + bool force) { - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg; + struct irq_cfg *cfg = data->chip_data; + unsigned int dest, irq = data->irq; struct msi_msg msg; - unsigned int dest; - if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) + if (__ioapic_set_affinity(data, mask, &dest)) return -1; - cfg = get_irq_desc_chip_data(desc); - dmar_msi_read(irq, &msg); msg.data &= ~MSI_DATA_VECTOR_MASK; @@ -3515,14 +3514,14 @@ static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) #endif /* CONFIG_SMP */ static struct irq_chip dmar_msi_type = { - .name = "DMAR_MSI", - .irq_unmask = dmar_msi_unmask, - .irq_mask = dmar_msi_mask, - .irq_ack = ack_apic_edge, + .name = "DMAR_MSI", + .irq_unmask = dmar_msi_unmask, + .irq_mask = dmar_msi_mask, + .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP - .set_affinity = dmar_msi_set_affinity, + .irq_set_affinity = dmar_msi_set_affinity, #endif - .irq_retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; int arch_setup_dmar_msi(unsigned int irq) -- cgit v0.10.2 From 0e09ddf2d71aeff92ff8055ac7600b85c255ee85 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 22:21:26 +0200 Subject: x86: Cleanup hpet affinity setting Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 6f8ac4c..0b9ec3c 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3545,12 +3545,11 @@ int arch_setup_dmar_msi(unsigned int irq) static int hpet_msi_set_affinity(struct irq_data *data, const struct cpumask *mask, bool force) { - struct irq_desc *desc = irq_to_desc(data->irq); struct irq_cfg *cfg = data->chip_data; struct msi_msg msg; unsigned int dest; - if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) + if (__ioapic_set_affinity(data, mask, &dest)) return -1; hpet_msi_read(data->handler_data, &msg); -- cgit v0.10.2 From be5b7bf73802a9391158d9fcc0bc6b07670c73a5 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 22:31:46 +0200 Subject: x86: Convert ht set_affinity to new chip function Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Jesse Barnes diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 0b9ec3c..b144f7a 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3646,33 +3646,30 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) write_ht_irq_msg(irq, &msg); } -static int set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) +static int +ht_set_affinity(struct irq_data *data, const struct cpumask *mask, bool force) { - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg; + struct irq_cfg *cfg = data->chip_data; unsigned int dest; - if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) + if (__ioapic_set_affinity(data, mask, &dest)) return -1; - cfg = get_irq_desc_chip_data(desc); - - target_ht_irq(irq, dest, cfg->vector); - + target_ht_irq(data->irq, dest, cfg->vector); return 0; } #endif static struct irq_chip ht_irq_chip = { - .name = "PCI-HT", - .irq_mask = mask_ht_irq, - .irq_unmask = unmask_ht_irq, - .irq_ack = ack_apic_edge, + .name = "PCI-HT", + .irq_mask = mask_ht_irq, + .irq_unmask = unmask_ht_irq, + .irq_ack = ack_apic_edge, #ifdef CONFIG_SMP - .set_affinity = set_ht_irq_affinity, + .irq_set_affinity = ht_set_affinity, #endif - .irq_retrigger = ioapic_retrigger_irq, + .irq_retrigger = ioapic_retrigger_irq, }; int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) -- cgit v0.10.2 From 7e495529b62cf462eb2d8875fe15ca446b8e1f94 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 23:31:50 +0200 Subject: x86: ioapic: Cleanup some more Cleanup after the irq_chip conversion a bit. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index b144f7a..4303099 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -131,13 +131,9 @@ struct irq_pin_list { struct irq_pin_list *next; }; -static struct irq_pin_list *get_one_free_irq_2_pin(int node) +static struct irq_pin_list *alloc_irq_pin_list(int node) { - struct irq_pin_list *pin; - - pin = kzalloc_node(sizeof(*pin), GFP_ATOMIC, node); - - return pin; + return kzalloc_node(sizeof(struct irq_pin_list), GFP_ATOMIC, node); } /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */ @@ -232,7 +228,7 @@ init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int node) if (!old_entry) return; - entry = get_one_free_irq_2_pin(node); + entry = alloc_irq_pin_list(node); if (!entry) return; @@ -242,7 +238,7 @@ init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int node) tail = entry; old_entry = old_entry->next; while (old_entry) { - entry = get_one_free_irq_2_pin(node); + entry = alloc_irq_pin_list(node); if (!entry) { entry = head; while (entry) { @@ -471,7 +467,7 @@ static void ioapic_mask_entry(int apic, int pin) * fast in the common case, and fast for shared ISA-space IRQs. */ static int -add_pin_to_irq_node_nopanic(struct irq_cfg *cfg, int node, int apic, int pin) +__add_pin_to_irq_node(struct irq_cfg *cfg, int node, int apic, int pin) { struct irq_pin_list **last, *entry; @@ -483,7 +479,7 @@ add_pin_to_irq_node_nopanic(struct irq_cfg *cfg, int node, int apic, int pin) last = &entry->next; } - entry = get_one_free_irq_2_pin(node); + entry = alloc_irq_pin_list(node); if (!entry) { printk(KERN_ERR "can not alloc irq_pin_list (%d,%d,%d)\n", node, apic, pin); @@ -498,7 +494,7 @@ add_pin_to_irq_node_nopanic(struct irq_cfg *cfg, int node, int apic, int pin) static void add_pin_to_irq_node(struct irq_cfg *cfg, int node, int apic, int pin) { - if (add_pin_to_irq_node_nopanic(cfg, node, apic, pin)) + if (__add_pin_to_irq_node(cfg, node, apic, pin)) panic("IO-APIC: failed to add irq-pin. Can not proceed\n"); } @@ -3801,7 +3797,7 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq, * IRQs < 16 are already in the irq_2_pin[] map */ if (irq >= legacy_pic->nr_legacy_irqs) { - if (add_pin_to_irq_node_nopanic(cfg, node, ioapic, pin)) { + if (__add_pin_to_irq_node(cfg, node, ioapic, pin)) { printk(KERN_INFO "can not add pin %d for irq %d\n", pin, irq); return 0; -- cgit v0.10.2 From 6e2fff50a5bd72a3f9e6f3ef6e9137efddb2d580 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 6 Oct 2010 22:07:03 +0200 Subject: x86: ioapic: Cleanup get_one_free_irq_cfg() Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 4303099..452f781 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -185,19 +185,18 @@ static struct irq_cfg *get_one_free_irq_cfg(int node) struct irq_cfg *cfg; cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); - if (cfg) { - if (!zalloc_cpumask_var_node(&cfg->domain, GFP_ATOMIC, node)) { - kfree(cfg); - cfg = NULL; - } else if (!zalloc_cpumask_var_node(&cfg->old_domain, - GFP_ATOMIC, node)) { - free_cpumask_var(cfg->domain); - kfree(cfg); - cfg = NULL; - } - } - + if (!cfg) + return NULL; + if (!zalloc_cpumask_var_node(&cfg->domain, GFP_ATOMIC, node)) + goto out_cfg; + if (!zalloc_cpumask_var_node(&cfg->old_domain, GFP_ATOMIC, node)) + goto out_domain; return cfg; +out_domain: + free_cpumask_var(cfg->domain); +out_cfg: + kfree(cfg); + return NULL; } int arch_init_chip_data(struct irq_desc *desc, int node) -- cgit v0.10.2 From 08c33db6d044d9dc74ddf8d9ee3cb1fa3eca262b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 6 Oct 2010 22:14:21 +0200 Subject: x86: Implement new allocator functions Implement new allocator functions which make use of the core changes. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 452f781..065c5dc 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -199,6 +199,13 @@ out_cfg: return NULL; } +static void free_irq_cfg(struct irq_cfg *cfg) +{ + free_cpumask_var(cfg->domain); + free_cpumask_var(cfg->old_domain); + kfree(cfg); +} + int arch_init_chip_data(struct irq_desc *desc, int node) { struct irq_cfg *cfg; @@ -299,13 +306,6 @@ void arch_init_copy_chip_data(struct irq_desc *old_desc, init_copy_irq_2_pin(old_cfg, cfg, node); } -static void free_irq_cfg(struct irq_cfg *cfg) -{ - free_cpumask_var(cfg->domain); - free_cpumask_var(cfg->old_domain); - kfree(cfg); -} - void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) { struct irq_cfg *old_cfg, *cfg; @@ -325,13 +325,53 @@ void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) /* end for move_irq_desc */ #else + struct irq_cfg *irq_cfg(unsigned int irq) { return irq < nr_irqs ? irq_cfgx + irq : NULL; } +static struct irq_cfg *get_one_free_irq_cfg(unsigned int irq, int node) +{ + return irq_cfgx + irq; +} + +static inline void free_irq_cfg(struct irq_cfg *cfg) { } + #endif +static struct irq_cfg *alloc_irq_and_cfg_at(unsigned int at, int node) +{ + int res = irq_alloc_desc_at(at, node); + struct irq_cfg *cfg; + + if (res < 0) { + if (res != -EEXIST) + return NULL; + cfg = get_irq_chip_data(at); + if (cfg) + return cfg; + } + + cfg = get_one_free_irq_cfg(node); + if (cfg) + set_irq_chip_data(at, cfg); + else + irq_free_desc(at); + return cfg; +} + +static int alloc_irq_from(unsigned int from, int node) +{ + return irq_alloc_desc_from(from, node); +} + +static void free_irq_at(unsigned int at, struct irq_cfg *cfg) +{ + free_irq_cfg(cfg); + irq_free_desc(at); +} + struct io_apic { unsigned int index; unsigned int unused[3]; -- cgit v0.10.2 From f981a3dc1941035a108da1276c448de6b10ddac9 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 10:44:21 +0200 Subject: x86: io_apic: Prepare alloc/free_irq_cfg() Rename the grossly misnamed get_one_free_irq_cfg() to alloc_irq_cfg(). Add a (not yet used) irq number argument to free_irq_cfg() Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 065c5dc..06da8fe 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -180,7 +180,7 @@ struct irq_cfg *irq_cfg(unsigned int irq) return get_irq_chip_data(irq); } -static struct irq_cfg *get_one_free_irq_cfg(int node) +static struct irq_cfg *alloc_irq_cfg(unsigned int irq, int node) { struct irq_cfg *cfg; @@ -199,7 +199,7 @@ out_cfg: return NULL; } -static void free_irq_cfg(struct irq_cfg *cfg) +static void free_irq_cfg(unsigned int at, struct irq_cfg *cfg) { free_cpumask_var(cfg->domain); free_cpumask_var(cfg->old_domain); @@ -212,7 +212,7 @@ int arch_init_chip_data(struct irq_desc *desc, int node) cfg = get_irq_desc_chip_data(desc); if (!cfg) { - cfg = get_one_free_irq_cfg(node); + cfg = alloc_irq_cfg(desc->irq, node); desc->chip_data = cfg; if (!cfg) { printk(KERN_ERR "can not alloc irq_cfg\n"); @@ -289,7 +289,7 @@ void arch_init_copy_chip_data(struct irq_desc *old_desc, struct irq_cfg *cfg; struct irq_cfg *old_cfg; - cfg = get_one_free_irq_cfg(node); + cfg = alloc_irq_cfg(desc->irq, node); if (!cfg) return; @@ -318,7 +318,7 @@ void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) if (old_cfg) { free_irq_2_pin(old_cfg, cfg); - free_irq_cfg(old_cfg); + free_irq_cfg(old_desc->irq, old_cfg); old_desc->chip_data = NULL; } } @@ -331,12 +331,12 @@ struct irq_cfg *irq_cfg(unsigned int irq) return irq < nr_irqs ? irq_cfgx + irq : NULL; } -static struct irq_cfg *get_one_free_irq_cfg(unsigned int irq, int node) +static struct irq_cfg *alloc_irq_cfg(unsigned int irq, int node) { return irq_cfgx + irq; } -static inline void free_irq_cfg(struct irq_cfg *cfg) { } +static inline void free_irq_cfg(unsigned int at, struct irq_cfg *cfg) { } #endif @@ -353,7 +353,7 @@ static struct irq_cfg *alloc_irq_and_cfg_at(unsigned int at, int node) return cfg; } - cfg = get_one_free_irq_cfg(node); + cfg = alloc_irq_cfg(at, node); if (cfg) set_irq_chip_data(at, cfg); else @@ -368,7 +368,7 @@ static int alloc_irq_from(unsigned int from, int node) static void free_irq_at(unsigned int at, struct irq_cfg *cfg) { - free_irq_cfg(cfg); + free_irq_cfg(at, cfg); irq_free_desc(at); } -- cgit v0.10.2 From fe6dab4e79e82ec35879bfe1a8968b7d15ac0d91 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 8 Oct 2010 22:44:02 -0700 Subject: x86: Don't setup ioapic irq for sci twice The sparseirq rework triggered a warning in the iommu code, which was caused by setting up ioapic for ACPI irq 9 twice. This function is solely to handle interrupts which are on a secondary ioapic and outside the legacy irq range. Replace the sparse irq_to_desc check with a non ifdeffed version. [ tglx: Moved it before the ioapic sparse conversion and simplified the inverse logic ] Signed-off-by: Yinghai Lu LKML-Reference: <4CB00122.3030301@kernel.org> Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 06da8fe..5aae718 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1565,11 +1565,11 @@ void setup_IO_APIC_irq_extra(u32 gsi) return; irq = pin_2_irq(idx, apic_id, pin); -#ifdef CONFIG_SPARSE_IRQ - desc = irq_to_desc(irq); - if (desc) + + /* Only handle the non legacy irqs on secondary ioapics */ + if (apic_id == 0 || irq < NR_IRQS_LEGACY) return; -#endif + desc = irq_to_desc_alloc_node(irq, node); if (!desc) { printk(KERN_INFO "can not get irq_desc for %d\n", irq); -- cgit v0.10.2 From fbc6bff04a095e049be290ff6f6ac68839166bd6 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Sep 2010 20:34:53 +0200 Subject: x86: ioapic: Cleanup sparse irq code Switch over to the new allocator and remove all the magic which was caused by the unability to destroy irq descriptors. Get rid of the create_irq_nr() loop for sparse and non sparse irq. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 3ec657f..8cc5108 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -61,7 +61,6 @@ config X86 select HAVE_USER_RETURN_NOTIFIER select HAVE_GENERIC_HARDIRQS select HAVE_SPARSE_IRQ - select NUMA_IRQ_DESC if (SPARSE_IRQ && NUMA) select GENERIC_IRQ_PROBE select GENERIC_PENDING_IRQ if SMP diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 5aae718..ed34029 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -157,6 +157,9 @@ int __init arch_early_irq_init(void) count = ARRAY_SIZE(irq_cfgx); node = cpu_to_node(0); + /* Make sure the legacy interrupts are marked in the bitmap */ + irq_reserve_irqs(0, legacy_pic->nr_legacy_irqs); + for (i = 0; i < count; i++) { set_irq_chip_data(i, &cfg[i]); zalloc_cpumask_var_node(&cfg[i].domain, GFP_NOWAIT, node); @@ -201,11 +204,15 @@ out_cfg: static void free_irq_cfg(unsigned int at, struct irq_cfg *cfg) { + if (!cfg) + return; + set_irq_chip_data(at, NULL); free_cpumask_var(cfg->domain); free_cpumask_var(cfg->old_domain); kfree(cfg); } +#if 0 int arch_init_chip_data(struct irq_desc *desc, int node) { struct irq_cfg *cfg; @@ -323,6 +330,7 @@ void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) } } /* end for move_irq_desc */ +#endif #else @@ -1479,11 +1487,9 @@ static struct { static void __init setup_IO_APIC_irqs(void) { - int apic_id, pin, idx, irq; - int notcon = 0; - struct irq_desc *desc; - struct irq_cfg *cfg; + int apic_id, pin, idx, irq, notcon = 0; int node = cpu_to_node(0); + struct irq_cfg *cfg; apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); @@ -1520,12 +1526,10 @@ static void __init setup_IO_APIC_irqs(void) apic->multi_timer_check(apic_id, irq)) continue; - desc = irq_to_desc_alloc_node(irq, node); - if (!desc) { - printk(KERN_INFO "can not get irq_desc for %d\n", irq); + cfg = alloc_irq_and_cfg_at(irq, node); + if (!cfg) continue; - } - cfg = get_irq_desc_chip_data(desc); + add_pin_to_irq_node(cfg, node, apic_id, pin); /* * don't mark it in pin_programmed, so later acpi could @@ -1547,9 +1551,7 @@ static void __init setup_IO_APIC_irqs(void) */ void setup_IO_APIC_irq_extra(u32 gsi) { - int apic_id = 0, pin, idx, irq; - int node = cpu_to_node(0); - struct irq_desc *desc; + int apic_id = 0, pin, idx, irq, node = cpu_to_node(0); struct irq_cfg *cfg; /* @@ -1570,13 +1572,10 @@ void setup_IO_APIC_irq_extra(u32 gsi) if (apic_id == 0 || irq < NR_IRQS_LEGACY) return; - desc = irq_to_desc_alloc_node(irq, node); - if (!desc) { - printk(KERN_INFO "can not get irq_desc for %d\n", irq); + cfg = alloc_irq_and_cfg_at(irq, node); + if (!cfg) return; - } - cfg = get_irq_desc_chip_data(desc); add_pin_to_irq_node(cfg, node, apic_id, pin); if (test_bit(pin, mp_ioapic_routing[apic_id].pin_programmed)) { @@ -3177,44 +3176,37 @@ device_initcall(ioapic_init_sysfs); /* * Dynamic irq allocate and deallocation */ -unsigned int create_irq_nr(unsigned int irq_want, int node) +unsigned int create_irq_nr(unsigned int from, int node) { - /* Allocate an unused irq */ - unsigned int irq; - unsigned int new; + struct irq_cfg *cfg; unsigned long flags; - struct irq_cfg *cfg_new = NULL; - struct irq_desc *desc_new = NULL; - - irq = 0; - if (irq_want < nr_irqs_gsi) - irq_want = nr_irqs_gsi; - - raw_spin_lock_irqsave(&vector_lock, flags); - for (new = irq_want; new < nr_irqs; new++) { - desc_new = irq_to_desc_alloc_node(new, node); - if (!desc_new) { - printk(KERN_INFO "can not get irq_desc for %d\n", new); - continue; - } - cfg_new = get_irq_desc_chip_data(desc_new); - - if (cfg_new->vector != 0) - continue; + unsigned int ret = 0; + int irq; - desc_new = move_irq_desc(desc_new, node); - cfg_new = get_irq_desc_chip_data(desc_new); + if (from < nr_irqs_gsi) + from = nr_irqs_gsi; - if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0) - irq = new; - break; + irq = alloc_irq_from(from, node); + if (irq < 0) + return 0; + cfg = alloc_irq_cfg(irq, node); + if (!cfg) { + free_irq_at(irq, NULL); + return 0; } - raw_spin_unlock_irqrestore(&vector_lock, flags); - if (irq > 0) - dynamic_irq_init_keep_chip_data(irq); + raw_spin_lock_irqsave(&vector_lock, flags); + if (!__assign_irq_vector(irq, cfg, apic->target_cpus())) + ret = irq; + raw_spin_unlock_irqrestore(&vector_lock, flags); - return irq; + if (ret) { + set_irq_chip_data(irq, cfg); + irq_clear_status_flags(irq, IRQ_NOREQUEST); + } else { + free_irq_at(irq, cfg); + } + return ret; } int create_irq(void) @@ -3234,14 +3226,16 @@ int create_irq(void) void destroy_irq(unsigned int irq) { + struct irq_cfg *cfg = get_irq_chip_data(irq); unsigned long flags; - dynamic_irq_cleanup_keep_chip_data(irq); + irq_set_status_flags(irq, IRQ_NOREQUEST|IRQ_NOPROBE); free_irte(irq); raw_spin_lock_irqsave(&vector_lock, flags); - __clear_irq_vector(irq, get_irq_chip_data(irq)); + __clear_irq_vector(irq, cfg); raw_spin_unlock_irqrestore(&vector_lock, flags); + free_irq_at(irq, cfg); } /* @@ -3802,7 +3796,6 @@ int __init arch_probe_nr_irqs(void) static int __io_apic_set_pci_routing(struct device *dev, int irq, struct io_apic_irq_attr *irq_attr) { - struct irq_desc *desc; struct irq_cfg *cfg; int node; int ioapic, pin; @@ -3820,18 +3813,14 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq, else node = cpu_to_node(0); - desc = irq_to_desc_alloc_node(irq, node); - if (!desc) { - printk(KERN_INFO "can not get irq_desc %d\n", irq); + cfg = alloc_irq_and_cfg_at(irq, node); + if (!cfg) return 0; - } pin = irq_attr->ioapic_pin; trigger = irq_attr->trigger; polarity = irq_attr->polarity; - cfg = get_irq_desc_chip_data(desc); - /* * IRQs < 16 are already in the irq_2_pin[] map */ @@ -4232,11 +4221,11 @@ void __init pre_init_apic_IRQ0(void) #ifndef CONFIG_SMP phys_cpu_present_map = physid_mask_of_physid(boot_cpu_physical_apicid); #endif - irq_to_desc_alloc_node(0, 0); + /* Make sure the irq descriptor is set up */ + cfg = alloc_irq_and_cfg_at(0, 0); setup_local_APIC(); - cfg = irq_cfg(0); add_pin_to_irq_node(cfg, 0, 0, 0); set_irq_chip_and_handler_name(0, &ioapic_chip, handle_edge_irq, "edge"); -- cgit v0.10.2 From bc5fdf9f3aad37406b3c8d635a7940cd65de0c12 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 10:40:53 +0200 Subject: x86: io_apic: Remove the now unused sparse_irq arch_* functions Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index ed34029..6ff6bb8 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -212,126 +212,6 @@ static void free_irq_cfg(unsigned int at, struct irq_cfg *cfg) kfree(cfg); } -#if 0 -int arch_init_chip_data(struct irq_desc *desc, int node) -{ - struct irq_cfg *cfg; - - cfg = get_irq_desc_chip_data(desc); - if (!cfg) { - cfg = alloc_irq_cfg(desc->irq, node); - desc->chip_data = cfg; - if (!cfg) { - printk(KERN_ERR "can not alloc irq_cfg\n"); - BUG_ON(1); - } - } - - return 0; -} - -/* for move_irq_desc */ -static void -init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int node) -{ - struct irq_pin_list *old_entry, *head, *tail, *entry; - - cfg->irq_2_pin = NULL; - old_entry = old_cfg->irq_2_pin; - if (!old_entry) - return; - - entry = alloc_irq_pin_list(node); - if (!entry) - return; - - entry->apic = old_entry->apic; - entry->pin = old_entry->pin; - head = entry; - tail = entry; - old_entry = old_entry->next; - while (old_entry) { - entry = alloc_irq_pin_list(node); - if (!entry) { - entry = head; - while (entry) { - head = entry->next; - kfree(entry); - entry = head; - } - /* still use the old one */ - return; - } - entry->apic = old_entry->apic; - entry->pin = old_entry->pin; - tail->next = entry; - tail = entry; - old_entry = old_entry->next; - } - - tail->next = NULL; - cfg->irq_2_pin = head; -} - -static void free_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg) -{ - struct irq_pin_list *entry, *next; - - if (old_cfg->irq_2_pin == cfg->irq_2_pin) - return; - - entry = old_cfg->irq_2_pin; - - while (entry) { - next = entry->next; - kfree(entry); - entry = next; - } - old_cfg->irq_2_pin = NULL; -} - -void arch_init_copy_chip_data(struct irq_desc *old_desc, - struct irq_desc *desc, int node) -{ - struct irq_cfg *cfg; - struct irq_cfg *old_cfg; - - cfg = alloc_irq_cfg(desc->irq, node); - - if (!cfg) - return; - - desc->chip_data = cfg; - - old_cfg = old_desc->chip_data; - - cfg->vector = old_cfg->vector; - cfg->move_in_progress = old_cfg->move_in_progress; - cpumask_copy(cfg->domain, old_cfg->domain); - cpumask_copy(cfg->old_domain, old_cfg->old_domain); - - init_copy_irq_2_pin(old_cfg, cfg, node); -} - -void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) -{ - struct irq_cfg *old_cfg, *cfg; - - old_cfg = get_irq_desc_chip_data(old_desc); - cfg = get_irq_desc_chip_data(desc); - - if (old_cfg == cfg) - return; - - if (old_cfg) { - free_irq_2_pin(old_cfg, cfg); - free_irq_cfg(old_desc->irq, old_cfg); - old_desc->chip_data = NULL; - } -} -/* end for move_irq_desc */ -#endif - #else struct irq_cfg *irq_cfg(unsigned int irq) -- cgit v0.10.2 From 0e1e367aab873becf3e21d9b0cf46d11154ebc3f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 4 Oct 2010 16:20:16 +0200 Subject: pci: intr-remap: Free irte memory if SPARSE_IRQ=y With SPARSE_IRQ=y the irte descriptors are dynamically allocated, but not freed in free_irte(). That was ok as long as the sparse irq core was not freeing irq descriptors on destroy_irq(). Now we leak the irte descriptor. Free it in free_irte(). Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index fd1d286..87ec390 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -95,6 +95,15 @@ static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) return desc->irq_2_iommu; } +static void irq_2_iommu_free(unsigned int irq) +{ + struct irq_data *d = irq_get_irq_data(irq); + struct irq_2_iommu *p = d->irq_2_iommu; + + d->irq_2_iommu = NULL; + kfree(p); +} + #else /* !CONFIG_SPARSE_IRQ */ static struct irq_2_iommu irq_2_iommuX[NR_IRQS]; @@ -110,6 +119,9 @@ static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) { return irq_2_iommu(irq); } + +static void irq_2_iommu_free(unsigned int irq) { } + #endif static DEFINE_SPINLOCK(irq_2_ir_lock); @@ -440,6 +452,8 @@ int free_irte(int irq) spin_unlock_irqrestore(&irq_2_ir_lock, flags); + irq_2_iommu_free(irq); + return rc; } -- cgit v0.10.2 From a8ef54aeb4308d3730ff31807c495390b98de16a Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 4 Oct 2010 16:51:27 +0200 Subject: pci: intr_remap: Use irq_data No need to dereference irq_desc. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index 87ec390..cb62529 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -54,45 +54,22 @@ struct irq_2_iommu { }; #ifdef CONFIG_GENERIC_HARDIRQS -static struct irq_2_iommu *get_one_free_irq_2_iommu(int node) -{ - struct irq_2_iommu *iommu; - - iommu = kzalloc_node(sizeof(*iommu), GFP_ATOMIC, node); - printk(KERN_DEBUG "alloc irq_2_iommu on node %d\n", node); - - return iommu; -} - static struct irq_2_iommu *irq_2_iommu(unsigned int irq) { - struct irq_desc *desc; - - desc = irq_to_desc(irq); - - if (WARN_ON_ONCE(!desc)) - return NULL; - - return desc->irq_2_iommu; + return get_irq_iommu(irq); } static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) { - struct irq_desc *desc; - struct irq_2_iommu *irq_iommu; + struct irq_data *data = irq_get_irq_data(irq); - desc = irq_to_desc(irq); - if (!desc) { - printk(KERN_INFO "can not get irq_desc for %d\n", irq); - return NULL; - } - - irq_iommu = desc->irq_2_iommu; - - if (!irq_iommu) - desc->irq_2_iommu = get_one_free_irq_2_iommu(irq_node(irq)); + if (WARN_ONCE(data->irq_2_iommu, + KERN_DEBUG "irq_2_iommu!=NULL irq %u\n", irq)) + return data->irq_2_iommu; - return desc->irq_2_iommu; + data->irq_2_iommu = kzalloc_node(sizeof(*data->irq_2_iommu), + GFP_ATOMIC, data->node); + return data->irq_2_iommu; } static void irq_2_iommu_free(unsigned int irq) -- cgit v0.10.2 From d0ad63927c6d4d511e172c78ba4a623539ef6901 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 4 Oct 2010 18:41:37 +0200 Subject: pci: intr_remap: Remove unused functions No users. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index cb62529..343f729 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -275,28 +275,6 @@ int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, u16 subhandle) return 0; } -int clear_irte_irq(int irq, struct intel_iommu *iommu, u16 index) -{ - struct irq_2_iommu *irq_iommu; - unsigned long flags; - - spin_lock_irqsave(&irq_2_ir_lock, flags); - irq_iommu = valid_irq_2_iommu(irq); - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); - return -1; - } - - irq_iommu->iommu = NULL; - irq_iommu->irte_index = 0; - irq_iommu->sub_handle = 0; - irq_2_iommu(irq)->irte_mask = 0; - - spin_unlock_irqrestore(&irq_2_ir_lock, flags); - - return 0; -} - int modify_irte(int irq, struct irte *irte_modified) { int rc; @@ -328,31 +306,6 @@ int modify_irte(int irq, struct irte *irte_modified) return rc; } -int flush_irte(int irq) -{ - int rc; - int index; - struct intel_iommu *iommu; - struct irq_2_iommu *irq_iommu; - unsigned long flags; - - spin_lock_irqsave(&irq_2_ir_lock, flags); - irq_iommu = valid_irq_2_iommu(irq); - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); - return -1; - } - - iommu = irq_iommu->iommu; - - index = irq_iommu->irte_index + irq_iommu->sub_handle; - - rc = qi_flush_iec(iommu, index, irq_iommu->irte_mask); - spin_unlock_irqrestore(&irq_2_ir_lock, flags); - - return rc; -} - struct intel_iommu *map_hpet_to_ir(u8 hpet_id) { int i; diff --git a/include/linux/dmar.h b/include/linux/dmar.h index cb86aa1..200439e 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -119,8 +119,6 @@ extern int alloc_irte(struct intel_iommu *iommu, int irq, u16 count); extern int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, u16 sub_handle); extern int map_irq_to_irte_handle(int irq, u16 *sub_handle); -extern int clear_irte_irq(int irq, struct intel_iommu *iommu, u16 index); -extern int flush_irte(int irq); extern int free_irte(int irq); extern int irq_remapped(int irq); -- cgit v0.10.2 From 423f085952fd7253407cb92984cc2d495a564481 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 10 Oct 2010 11:39:09 +0200 Subject: x86: Embedd irq_2_iommu into irq_cfg That interrupt remapping code is x86 specific and tied to the io_apic code. No need for separate allocator functions in the interrupt remapping code. This allows to simplify the code and irq_2_iommu is small (13 bytes on 64bit) so it's not a real problem even if interrupt remapping is runtime disabled. If it's compile time disabled the impact is zero. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index 76848f2..e756c4b 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -78,6 +78,13 @@ static inline void set_io_apic_irq_attr(struct io_apic_irq_attr *irq_attr, irq_attr->polarity = polarity; } +struct irq_2_iommu { + struct intel_iommu *iommu; + u16 irte_index; + u16 sub_handle; + u8 irte_mask; +}; + /* * This is performance-critical, we want to do it O(1) * @@ -89,6 +96,9 @@ struct irq_cfg { cpumask_var_t old_domain; u8 vector; u8 move_in_progress : 1; +#ifdef CONFIG_INTR_REMAP + struct irq_2_iommu irq_2_iommu; +#endif }; extern struct irq_cfg *irq_cfg(unsigned int); diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index 343f729..0f4691c 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -46,13 +46,6 @@ static __init int setup_intremap(char *str) } early_param("intremap", setup_intremap); -struct irq_2_iommu { - struct intel_iommu *iommu; - u16 irte_index; - u16 sub_handle; - u8 irte_mask; -}; - #ifdef CONFIG_GENERIC_HARDIRQS static struct irq_2_iommu *irq_2_iommu(unsigned int irq) { diff --git a/include/linux/dmar.h b/include/linux/dmar.h index 200439e..4475f8c 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -106,6 +106,7 @@ struct irte { __u64 high; }; }; + #ifdef CONFIG_INTR_REMAP extern int intr_remapping_enabled; extern int intr_remapping_supported(void); -- cgit v0.10.2 From 349d67673c08cbc299a69086b0f5447cf1094e9b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 10 Oct 2010 12:29:27 +0200 Subject: intr_remap: Use irq_2_iommu in struct irq_cfg Switch the intr_remapping code to use the irq_2_iommu struct in irg_cfg. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index 0f4691c..ea46935 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -46,54 +46,21 @@ static __init int setup_intremap(char *str) } early_param("intremap", setup_intremap); -#ifdef CONFIG_GENERIC_HARDIRQS static struct irq_2_iommu *irq_2_iommu(unsigned int irq) { - return get_irq_iommu(irq); + struct irq_cfg *cfg = get_irq_chip_data(irq); + return cfg ? &cfg->irq_2_iommu : NULL; } static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) { - struct irq_data *data = irq_get_irq_data(irq); - - if (WARN_ONCE(data->irq_2_iommu, - KERN_DEBUG "irq_2_iommu!=NULL irq %u\n", irq)) - return data->irq_2_iommu; - - data->irq_2_iommu = kzalloc_node(sizeof(*data->irq_2_iommu), - GFP_ATOMIC, data->node); - return data->irq_2_iommu; + return irq_2_iommu(irq); } static void irq_2_iommu_free(unsigned int irq) { - struct irq_data *d = irq_get_irq_data(irq); - struct irq_2_iommu *p = d->irq_2_iommu; - - d->irq_2_iommu = NULL; - kfree(p); } -#else /* !CONFIG_SPARSE_IRQ */ - -static struct irq_2_iommu irq_2_iommuX[NR_IRQS]; - -static struct irq_2_iommu *irq_2_iommu(unsigned int irq) -{ - if (irq < nr_irqs) - return &irq_2_iommuX[irq]; - - return NULL; -} -static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) -{ - return irq_2_iommu(irq); -} - -static void irq_2_iommu_free(unsigned int irq) { } - -#endif - static DEFINE_SPINLOCK(irq_2_ir_lock); static struct irq_2_iommu *valid_irq_2_iommu(unsigned int irq) -- cgit v0.10.2 From d585d060b42bd36f6f0b23ff327d3b91f80c7139 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 10 Oct 2010 12:34:27 +0200 Subject: intr_remap: Simplify the code further Having irq_2_iommu in struct irq_cfg allows further simplifications. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index ea46935..a620b8b 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -46,58 +46,31 @@ static __init int setup_intremap(char *str) } early_param("intremap", setup_intremap); +static DEFINE_SPINLOCK(irq_2_ir_lock); + static struct irq_2_iommu *irq_2_iommu(unsigned int irq) { struct irq_cfg *cfg = get_irq_chip_data(irq); return cfg ? &cfg->irq_2_iommu : NULL; } -static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) -{ - return irq_2_iommu(irq); -} - -static void irq_2_iommu_free(unsigned int irq) -{ -} - -static DEFINE_SPINLOCK(irq_2_ir_lock); - -static struct irq_2_iommu *valid_irq_2_iommu(unsigned int irq) -{ - struct irq_2_iommu *irq_iommu; - - irq_iommu = irq_2_iommu(irq); - - if (!irq_iommu) - return NULL; - - if (!irq_iommu->iommu) - return NULL; - - return irq_iommu; -} - int irq_remapped(int irq) { - return valid_irq_2_iommu(irq) != NULL; + struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); + + return irq_iommu ? irq_iommu->iommu != NULL : 0; } int get_irte(int irq, struct irte *entry) { - int index; - struct irq_2_iommu *irq_iommu; + struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); unsigned long flags; + int index; - if (!entry) + if (!entry || !irq_iommu) return -1; spin_lock_irqsave(&irq_2_ir_lock, flags); - irq_iommu = valid_irq_2_iommu(irq); - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); - return -1; - } index = irq_iommu->irte_index + irq_iommu->sub_handle; *entry = *(irq_iommu->iommu->ir_table->base + index); @@ -109,20 +82,14 @@ int get_irte(int irq, struct irte *entry) int alloc_irte(struct intel_iommu *iommu, int irq, u16 count) { struct ir_table *table = iommu->ir_table; - struct irq_2_iommu *irq_iommu; + struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); u16 index, start_index; unsigned int mask = 0; unsigned long flags; int i; - if (!count) - return -1; - -#ifndef CONFIG_SPARSE_IRQ - /* protect irq_2_iommu_alloc later */ - if (irq >= nr_irqs) + if (!count || !irq_iommu) return -1; -#endif /* * start the IRTE search from index 0. @@ -163,13 +130,6 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count) for (i = index; i < index + count; i++) table->base[i].present = 1; - irq_iommu = irq_2_iommu_alloc(irq); - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); - printk(KERN_ERR "can't allocate irq_2_iommu\n"); - return -1; - } - irq_iommu->iommu = iommu; irq_iommu->irte_index = index; irq_iommu->sub_handle = 0; @@ -193,17 +153,14 @@ static int qi_flush_iec(struct intel_iommu *iommu, int index, int mask) int map_irq_to_irte_handle(int irq, u16 *sub_handle) { - int index; - struct irq_2_iommu *irq_iommu; + struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); unsigned long flags; + int index; - spin_lock_irqsave(&irq_2_ir_lock, flags); - irq_iommu = valid_irq_2_iommu(irq); - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); + if (!irq_iommu) return -1; - } + spin_lock_irqsave(&irq_2_ir_lock, flags); *sub_handle = irq_iommu->sub_handle; index = irq_iommu->irte_index; spin_unlock_irqrestore(&irq_2_ir_lock, flags); @@ -212,18 +169,13 @@ int map_irq_to_irte_handle(int irq, u16 *sub_handle) int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, u16 subhandle) { - struct irq_2_iommu *irq_iommu; + struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); unsigned long flags; - spin_lock_irqsave(&irq_2_ir_lock, flags); - - irq_iommu = irq_2_iommu_alloc(irq); - - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); - printk(KERN_ERR "can't allocate irq_2_iommu\n"); + if (!irq_iommu) return -1; - } + + spin_lock_irqsave(&irq_2_ir_lock, flags); irq_iommu->iommu = iommu; irq_iommu->irte_index = index; @@ -237,19 +189,16 @@ int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, u16 subhandle) int modify_irte(int irq, struct irte *irte_modified) { - int rc; - int index; - struct irte *irte; + struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); struct intel_iommu *iommu; - struct irq_2_iommu *irq_iommu; unsigned long flags; + struct irte *irte; + int rc, index; - spin_lock_irqsave(&irq_2_ir_lock, flags); - irq_iommu = valid_irq_2_iommu(irq); - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); + if (!irq_iommu) return -1; - } + + spin_lock_irqsave(&irq_2_ir_lock, flags); iommu = irq_iommu->iommu; @@ -322,16 +271,14 @@ static int clear_entries(struct irq_2_iommu *irq_iommu) int free_irte(int irq) { - int rc = 0; - struct irq_2_iommu *irq_iommu; + struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); unsigned long flags; + int rc; - spin_lock_irqsave(&irq_2_ir_lock, flags); - irq_iommu = valid_irq_2_iommu(irq); - if (!irq_iommu) { - spin_unlock_irqrestore(&irq_2_ir_lock, flags); + if (!irq_iommu) return -1; - } + + spin_lock_irqsave(&irq_2_ir_lock, flags); rc = clear_entries(irq_iommu); @@ -342,8 +289,6 @@ int free_irte(int irq) spin_unlock_irqrestore(&irq_2_ir_lock, flags); - irq_2_iommu_free(irq); - return rc; } -- cgit v0.10.2 From 1a0730d6649113c820217387a011a17dd4aff3ad Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 11 Oct 2010 11:55:37 +0200 Subject: x86: Speed up the irq_remapped check in hot pathes irq_2_iommu is in struct irq_cfg, so we can do the irq_remapped check based on irq_cfg instead of going through a lookup function. That's especially interesting in the eoi_ioapic_irq() hotpath. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/arch/x86/include/asm/irq_remapping.h b/arch/x86/include/asm/irq_remapping.h index 8d84150..1c23360 100644 --- a/arch/x86/include/asm/irq_remapping.h +++ b/arch/x86/include/asm/irq_remapping.h @@ -24,10 +24,18 @@ static inline void prepare_irte(struct irte *irte, int vector, irte->dest_id = IRTE_DEST(dest); irte->redir_hint = 1; } +static inline bool irq_remapped(struct irq_cfg *cfg) +{ + return cfg->irq_2_iommu.iommu != NULL; +} #else static void prepare_irte(struct irte *irte, int vector, unsigned int dest) { } +static inline bool irq_remapped(struct irq_cfg *cfg) +{ + return false; +} #endif #endif /* _ASM_X86_IRQ_REMAPPING_H */ diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 6ff6bb8..1b8e8a1 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1237,7 +1237,7 @@ static void ioapic_register_intr(unsigned int irq, unsigned long trigger) else irq_clear_status_flags(irq, IRQ_LEVEL); - if (irq_remapped(irq)) { + if (irq_remapped(get_irq_chip_data(irq))) { irq_set_status_flags(irq, IRQ_MOVE_PCNTXT); if (trigger) set_irq_chip_and_handler_name(irq, &ir_ioapic_chip, @@ -2183,7 +2183,7 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq * With interrupt-remapping, destination information comes * from interrupt-remapping table entry. */ - if (!irq_remapped(irq)) + if (!irq_remapped(cfg)) io_apic_write(apic, 0x11 + pin*2, dest); reg = io_apic_read(apic, 0x10 + pin*2); reg &= ~IO_APIC_REDIR_VECTOR_MASK; @@ -2415,7 +2415,7 @@ static void eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) * intr-remapping table entry. Hence for the io-apic * EOI we use the pin number. */ - if (irq_remapped(irq)) + if (irq_remapped(cfg)) io_apic_eoi(entry->apic, entry->pin); else io_apic_eoi(entry->apic, cfg->vector); @@ -3139,7 +3139,7 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, dest = apic->cpu_mask_to_apicid_and(cfg->domain, apic->target_cpus()); - if (irq_remapped(irq)) { + if (irq_remapped(get_irq_chip_data(irq))) { struct irte irte; int ir_index; u16 sub_handle; @@ -3321,7 +3321,7 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc, int irq) set_irq_msi(irq, msidesc); write_msi_msg(irq, &msg); - if (irq_remapped(irq)) { + if (irq_remapped(get_irq_chip_data(irq))) { irq_set_status_flags(irq, IRQ_MOVE_PCNTXT); set_irq_chip_and_handler_name(irq, &msi_ir_chip, handle_edge_irq, "edge"); } else @@ -3522,7 +3522,7 @@ int arch_setup_hpet_msi(unsigned int irq, unsigned int id) hpet_msi_write(get_irq_data(irq), &msg); irq_set_status_flags(irq, IRQ_MOVE_PCNTXT); - if (irq_remapped(irq)) + if (irq_remapped(get_irq_chip_data(irq))) set_irq_chip_and_handler_name(irq, &ir_hpet_msi_type, handle_edge_irq, "edge"); else diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index a620b8b..ec87cd6 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -54,13 +54,6 @@ static struct irq_2_iommu *irq_2_iommu(unsigned int irq) return cfg ? &cfg->irq_2_iommu : NULL; } -int irq_remapped(int irq) -{ - struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); - - return irq_iommu ? irq_iommu->iommu != NULL : 0; -} - int get_irte(int irq, struct irte *entry) { struct irq_2_iommu *irq_iommu = irq_2_iommu(irq); diff --git a/include/linux/dmar.h b/include/linux/dmar.h index 4475f8c..51651b7 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -122,7 +122,6 @@ extern int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, extern int map_irq_to_irte_handle(int irq, u16 *sub_handle); extern int free_irte(int irq); -extern int irq_remapped(int irq); extern struct intel_iommu *map_dev_to_ir(struct pci_dev *dev); extern struct intel_iommu *map_ioapic_to_ir(int apic); extern struct intel_iommu *map_hpet_to_ir(u8 id); @@ -176,7 +175,6 @@ static inline int set_msi_sid(struct irte *irte, struct pci_dev *dev) return 0; } -#define irq_remapped(irq) (0) #define enable_intr_remapping(mode) (-1) #define disable_intr_remapping() (0) #define reenable_intr_remapping(mode) (0) -- cgit v0.10.2 From 10ba1e0eeef6a3c9453d96364e28cb4d911e1ac3 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 11 Oct 2010 12:21:18 +0200 Subject: genirq: Remove irq_2_iommu irq_2_iommu is now in the x86 code where it belongs. Remove all leftovers. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Acked-by: Suresh Siddha Cc: David Woodhouse Cc: Jesse Barnes diff --git a/include/linux/irq.h b/include/linux/irq.h index ef87882..49702b2 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -85,7 +85,6 @@ typedef void (*irq_flow_handler_t)(unsigned int irq, #endif struct msi_desc; -struct irq_2_iommu; /** * struct irq_data - per irq and irq chip data passed down to chip functions @@ -97,7 +96,6 @@ struct irq_2_iommu; * methods, to allow shared chip implementations * @msi_desc: MSI descriptor * @affinity: IRQ affinity on SMP - * @irq_2_iommu: iommu with this irq * * The fields here need to overlay the ones in irq_desc until we * cleaned up the direct references and switched everything over to @@ -113,9 +111,6 @@ struct irq_data { #ifdef CONFIG_SMP cpumask_var_t affinity; #endif -#ifdef CONFIG_INTR_REMAP - struct irq_2_iommu *irq_2_iommu; -#endif }; /** @@ -389,19 +384,6 @@ static inline struct msi_desc *irq_data_get_msi(struct irq_data *d) return d->msi_desc; } -#ifdef CONFIG_INTR_REMAP -static inline struct irq_2_iommu *get_irq_iommu(unsigned int irq) -{ - struct irq_data *d = irq_get_irq_data(irq); - return d ? d->irq_2_iommu : NULL; -} - -static inline struct irq_2_iommu *irq_data_get_iommu(struct irq_data *d) -{ - return d->irq_2_iommu; -} -#endif - int irq_alloc_descs(int irq, unsigned int from, unsigned int cnt, int node); void irq_free_descs(unsigned int irq, unsigned int cnt); int irq_reserve_irqs(unsigned int from, unsigned int cnt); diff --git a/include/linux/irqdesc.h b/include/linux/irqdesc.h index 22e426f..f77dc56 100644 --- a/include/linux/irqdesc.h +++ b/include/linux/irqdesc.h @@ -10,7 +10,6 @@ struct proc_dir_entry; struct timer_rand_state; -struct irq_2_iommu; /** * struct irq_desc - interrupt descriptor * @irq_data: per irq and chip data passed down to chip functions @@ -53,9 +52,6 @@ struct irq_desc { #ifdef CONFIG_SMP cpumask_var_t affinity; #endif -#ifdef CONFIG_INTR_REMAP - struct irq_2_iommu *irq_2_iommu; -#endif }; }; #endif -- cgit v0.10.2 From 1a8ce7ff68d777195da2d340561bda610e533a64 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 4 Oct 2010 21:08:56 +0200 Subject: x86: Make io_apic.c local functions static No users outside of io_apic.c Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 9cb2edb..c8be456 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -170,12 +170,6 @@ extern int restore_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries); extern void probe_nr_irqs_gsi(void); -extern int setup_ioapic_entry(int apic, int irq, - struct IO_APIC_route_entry *entry, - unsigned int destination, int trigger, - int polarity, int vector, int pin); -extern void ioapic_write_entry(int apic, int pin, - struct IO_APIC_route_entry e); extern void setup_ioapic_ids_from_mpc(void); struct mp_ioapic_gsi{ diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 1b8e8a1..0102543 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -364,7 +364,7 @@ __ioapic_write_entry(int apic, int pin, struct IO_APIC_route_entry e) io_apic_write(apic, 0x10 + 2*pin, eu.w1); } -void ioapic_write_entry(int apic, int pin, struct IO_APIC_route_entry e) +static void ioapic_write_entry(int apic, int pin, struct IO_APIC_route_entry e) { unsigned long flags; raw_spin_lock_irqsave(&ioapic_lock, flags); @@ -1259,10 +1259,10 @@ static void ioapic_register_intr(unsigned int irq, unsigned long trigger) handle_edge_irq, "edge"); } -int setup_ioapic_entry(int apic_id, int irq, - struct IO_APIC_route_entry *entry, - unsigned int destination, int trigger, - int polarity, int vector, int pin) +static int setup_ioapic_entry(int apic_id, int irq, + struct IO_APIC_route_entry *entry, + unsigned int destination, int trigger, + int polarity, int vector, int pin) { /* * add it to the IO-APIC irq-routing table: -- cgit v0.10.2 From 48b2650196364e4ef124efb841b63c2326e4ccb2 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 30 Sep 2010 11:43:08 +0200 Subject: x86: uv: Clean up the direct access to irq_desc Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index e756c4b..d5905fd 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -101,7 +101,6 @@ struct irq_cfg { #endif }; -extern struct irq_cfg *irq_cfg(unsigned int); extern int assign_irq_vector(int, struct irq_cfg *, const struct cpumask *); extern void send_cleanup_vector(struct irq_cfg *); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 0102543..5193f20 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -178,7 +178,7 @@ int __init arch_early_irq_init(void) } #ifdef CONFIG_SPARSE_IRQ -struct irq_cfg *irq_cfg(unsigned int irq) +static struct irq_cfg *irq_cfg(unsigned int irq) { return get_irq_chip_data(irq); } diff --git a/arch/x86/kernel/uv_irq.c b/arch/x86/kernel/uv_irq.c index 56a42e1..7b24460 100644 --- a/arch/x86/kernel/uv_irq.c +++ b/arch/x86/kernel/uv_irq.c @@ -28,34 +28,21 @@ struct uv_irq_2_mmr_pnode{ static spinlock_t uv_irq_lock; static struct rb_root uv_irq_root; -static int uv_set_irq_affinity(unsigned int, const struct cpumask *); +static int uv_set_irq_affinity(struct irq_data *, const struct cpumask *, bool); -static void uv_noop(unsigned int irq) -{ -} - -static unsigned int uv_noop_ret(unsigned int irq) -{ - return 0; -} +static void uv_noop(struct irq_data *data) { } -static void uv_ack_apic(unsigned int irq) +static void uv_ack_apic(struct irq_data *data) { ack_APIC_irq(); } static struct irq_chip uv_irq_chip = { - .name = "UV-CORE", - .startup = uv_noop_ret, - .shutdown = uv_noop, - .enable = uv_noop, - .disable = uv_noop, - .ack = uv_noop, - .mask = uv_noop, - .unmask = uv_noop, - .eoi = uv_ack_apic, - .end = uv_noop, - .set_affinity = uv_set_irq_affinity, + .name = "UV-CORE", + .irq_mask = uv_noop, + .irq_unmask = uv_noop, + .irq_eoi = uv_ack_apic, + .irq_set_affinity = uv_set_irq_affinity, }; /* @@ -144,26 +131,22 @@ arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, unsigned long mmr_offset, int limit) { const struct cpumask *eligible_cpu = cpumask_of(cpu); - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg; - int mmr_pnode; + struct irq_cfg *cfg = get_irq_chip_data(irq); unsigned long mmr_value; struct uv_IO_APIC_route_entry *entry; - int err; + int mmr_pnode, err; BUILD_BUG_ON(sizeof(struct uv_IO_APIC_route_entry) != sizeof(unsigned long)); - cfg = irq_cfg(irq); - err = assign_irq_vector(irq, cfg, eligible_cpu); if (err != 0) return err; if (limit == UV_AFFINITY_CPU) - desc->status |= IRQ_NO_BALANCING; + irq_set_status_flags(irq, IRQ_NO_BALANCING); else - desc->status |= IRQ_MOVE_PCNTXT; + irq_set_status_flags(irq, IRQ_MOVE_PCNTXT); set_irq_chip_and_handler_name(irq, &uv_irq_chip, handle_percpu_irq, irq_name); @@ -206,17 +189,17 @@ static void arch_disable_uv_irq(int mmr_pnode, unsigned long mmr_offset) uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value); } -static int uv_set_irq_affinity(unsigned int irq, const struct cpumask *mask) +static int +uv_set_irq_affinity(struct irq_data *data, const struct cpumask *mask, + bool force) { - struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg = get_irq_desc_chip_data(desc); + struct irq_cfg *cfg = data->chip_data; unsigned int dest; - unsigned long mmr_value; + unsigned long mmr_value, mmr_offset; struct uv_IO_APIC_route_entry *entry; - unsigned long mmr_offset; int mmr_pnode; - if (__ioapic_set_affinity(&desc->irq_data, mask, &dest)) + if (__ioapic_set_affinity(data, mask, &dest)) return -1; mmr_value = 0; @@ -231,7 +214,7 @@ static int uv_set_irq_affinity(unsigned int irq, const struct cpumask *mask) entry->dest = dest; /* Get previously stored MMR and pnode of hub sourcing interrupts */ - if (uv_irq_2_mmr_info(irq, &mmr_offset, &mmr_pnode)) + if (uv_irq_2_mmr_info(data->irq, &mmr_offset, &mmr_pnode)) return -1; uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value); -- cgit v0.10.2 From ad9f43340f48c5f7a0a5ef7656986e23d06bf996 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 30 Sep 2010 11:26:43 +0200 Subject: x86: Use sane enumeration Instead of looping through all interrupts, use the bitmap lookup to find the next. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 5193f20..057b0e1 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1160,7 +1160,6 @@ void __setup_vector_irq(int cpu) /* Initialize vector_irq on a new cpu */ int irq, vector; struct irq_cfg *cfg; - struct irq_desc *desc; /* * vector_lock will make sure that we don't run into irq vector @@ -1169,9 +1168,10 @@ void __setup_vector_irq(int cpu) */ raw_spin_lock(&vector_lock); /* Mark the inuse vectors */ - for_each_irq_desc(irq, desc) { - cfg = get_irq_desc_chip_data(desc); - + for_each_active_irq(irq) { + cfg = get_irq_chip_data(irq); + if (!cfg) + continue; /* * If it is a legacy IRQ handled by the legacy PIC, this cpu * will be part of the irq_cfg's domain. @@ -1516,7 +1516,6 @@ __apicdebuginit(void) print_IO_APIC(void) union IO_APIC_reg_03 reg_03; unsigned long flags; struct irq_cfg *cfg; - struct irq_desc *desc; unsigned int irq; printk(KERN_DEBUG "number of MP IRQ sources: %d.\n", mp_irq_entries); @@ -1603,10 +1602,10 @@ __apicdebuginit(void) print_IO_APIC(void) } } printk(KERN_DEBUG "IRQ to pin mappings:\n"); - for_each_irq_desc(irq, desc) { + for_each_active_irq(irq) { struct irq_pin_list *entry; - cfg = get_irq_desc_chip_data(desc); + cfg = get_irq_chip_data(irq); if (!cfg) continue; entry = cfg->irq_2_pin; @@ -2574,9 +2573,8 @@ static struct irq_chip ir_ioapic_chip __read_mostly = { static inline void init_IO_APIC_traps(void) { - int irq; - struct irq_desc *desc; struct irq_cfg *cfg; + unsigned int irq; /* * NOTE! The local APIC isn't very good at handling @@ -2589,8 +2587,8 @@ static inline void init_IO_APIC_traps(void) * Also, we've got to be careful not to trash gate * 0x80, because int 0x80 is hm, kind of importantish. ;) */ - for_each_irq_desc(irq, desc) { - cfg = get_irq_desc_chip_data(desc); + for_each_active_irq(irq) { + cfg = get_irq_chip_data(irq); if (IO_APIC_IRQ(irq) && cfg && !cfg->vector) { /* * Hmm.. We don't have an entry for this, @@ -2601,7 +2599,7 @@ static inline void init_IO_APIC_traps(void) legacy_pic->make_irq(irq); else /* Strange. Oh, well.. */ - desc->chip = &no_irq_chip; + set_irq_chip(irq, &no_irq_chip); } } } -- cgit v0.10.2 From 77dff1c755c3218691e95e7e38ee14323b35dbdb Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 29 Sep 2010 17:37:10 +0200 Subject: x86: xen: Sanitise sparse_irq handling There seems to be more cleanups possible, but that's left to the xen experts :) Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Jeremy Fitzhardinge diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 13365ba..7d24b0d 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -338,30 +338,29 @@ static void unmask_evtchn(int port) static int find_unbound_irq(void) { - int irq; - struct irq_desc *desc; + struct irq_data *data; + int irq, res; for (irq = 0; irq < nr_irqs; irq++) { - desc = irq_to_desc(irq); + data = irq_get_irq_data(irq); /* only 0->15 have init'd desc; handle irq > 16 */ - if (desc == NULL) + if (!data) break; - if (desc->chip == &no_irq_chip) + if (data->chip == &no_irq_chip) break; - if (desc->chip != &xen_dynamic_chip) + if (data->chip != &xen_dynamic_chip) continue; if (irq_info[irq].type == IRQT_UNBOUND) - break; + return irq; } if (irq == nr_irqs) panic("No available IRQ to bind to: increase nr_irqs!\n"); - desc = irq_to_desc_alloc_node(irq, 0); - if (WARN_ON(desc == NULL)) - return -1; + res = irq_alloc_desc_at(irq, 0); - dynamic_irq_init_keep_chip_data(irq); + if (WARN_ON(res != irq)) + return -1; return irq; } @@ -495,7 +494,7 @@ static void unbind_from_irq(unsigned int irq) if (irq_info[irq].type != IRQT_UNBOUND) { irq_info[irq] = mk_unbound_info(); - dynamic_irq_cleanup(irq); + irq_free_desc(irq); } spin_unlock(&irq_mapping_update_lock); -- cgit v0.10.2 From b7d0d8258a9f71949b810e0f82a3d75088f4d364 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 29 Sep 2010 18:44:23 +0200 Subject: genirq: Remove arch_init_chip_data() This function should have not been there in the first place. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index a0384a4..1998898 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -641,11 +641,8 @@ static inline void init_irq_proc(void) struct seq_file; int show_interrupts(struct seq_file *p, void *v); -struct irq_desc; - extern int early_irq_init(void); extern int arch_probe_nr_irqs(void); extern int arch_early_irq_init(void); -extern int arch_init_chip_data(struct irq_desc *desc, int node); #endif diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index a1fbd1d..6c71f8e 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -197,8 +197,6 @@ static int alloc_descs(unsigned int start, unsigned int cnt, int node) desc = alloc_desc(start + i, node); if (!desc) goto err; - /* temporary until I fixed x86 madness */ - arch_init_chip_data(desc, node); raw_spin_lock_irqsave(&sparse_irq_lock, flags); irq_insert_desc(start + i, desc); raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); diff --git a/kernel/softirq.c b/kernel/softirq.c index 14a7b80..d19b1c9 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -896,9 +896,4 @@ int __init __weak arch_early_irq_init(void) { return 0; } - -int __weak arch_init_chip_data(struct irq_desc *desc, int node) -{ - return 0; -} #endif -- cgit v0.10.2 From b7b29338dc7111ed8bd4d6555d84afae13ebe752 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 29 Sep 2010 18:46:55 +0200 Subject: genirq: Sanitize dynamic irq handling Use the cleanup functions of the dynamic allocator. No need to have separate implementations. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/include/linux/irq.h b/include/linux/irq.h index 49702b2..e963911 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -326,11 +326,15 @@ extern unsigned int create_irq_nr(unsigned int irq_want, int node); extern int create_irq(void); extern void destroy_irq(unsigned int irq); -/* Dynamic irq helper functions */ -extern void dynamic_irq_init(unsigned int irq); -void dynamic_irq_init_keep_chip_data(unsigned int irq); +/* + * Dynamic irq helper functions. Obsolete. Use irq_alloc_desc* and + * irq_free_desc instead. + */ extern void dynamic_irq_cleanup(unsigned int irq); -void dynamic_irq_cleanup_keep_chip_data(unsigned int irq); +static inline void dynamic_irq_init(unsigned int irq) +{ + dynamic_irq_cleanup(irq); +} /* Set/get chip/data for an IRQ: */ extern int set_irq_chip(unsigned int irq, struct irq_chip *chip); diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 3405761..baa5c4a 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -18,108 +18,6 @@ #include "internals.h" -static void dynamic_irq_init_x(unsigned int irq, bool keep_chip_data) -{ - struct irq_desc *desc; - unsigned long flags; - - desc = irq_to_desc(irq); - if (!desc) { - WARN(1, KERN_ERR "Trying to initialize invalid IRQ%d\n", irq); - return; - } - - /* Ensure we don't have left over values from a previous use of this irq */ - raw_spin_lock_irqsave(&desc->lock, flags); - desc->status = IRQ_DEFAULT_INIT_FLAGS; - desc->irq_data.chip = &no_irq_chip; - desc->handle_irq = handle_bad_irq; - desc->depth = 1; - desc->irq_data.msi_desc = NULL; - desc->irq_data.handler_data = NULL; - if (!keep_chip_data) - desc->irq_data.chip_data = NULL; - desc->action = NULL; - desc->irq_count = 0; - desc->irqs_unhandled = 0; -#ifdef CONFIG_SMP - cpumask_setall(desc->irq_data.affinity); -#ifdef CONFIG_GENERIC_PENDING_IRQ - cpumask_clear(desc->pending_mask); -#endif -#endif - raw_spin_unlock_irqrestore(&desc->lock, flags); -} - -/** - * dynamic_irq_init - initialize a dynamically allocated irq - * @irq: irq number to initialize - */ -void dynamic_irq_init(unsigned int irq) -{ - dynamic_irq_init_x(irq, false); -} - -/** - * dynamic_irq_init_keep_chip_data - initialize a dynamically allocated irq - * @irq: irq number to initialize - * - * does not set irq_to_desc(irq)->irq_data.chip_data to NULL - */ -void dynamic_irq_init_keep_chip_data(unsigned int irq) -{ - dynamic_irq_init_x(irq, true); -} - -static void dynamic_irq_cleanup_x(unsigned int irq, bool keep_chip_data) -{ - struct irq_desc *desc = irq_to_desc(irq); - unsigned long flags; - - if (!desc) { - WARN(1, KERN_ERR "Trying to cleanup invalid IRQ%d\n", irq); - return; - } - - raw_spin_lock_irqsave(&desc->lock, flags); - if (desc->action) { - raw_spin_unlock_irqrestore(&desc->lock, flags); - WARN(1, KERN_ERR "Destroying IRQ%d without calling free_irq\n", - irq); - return; - } - desc->irq_data.msi_desc = NULL; - desc->irq_data.handler_data = NULL; - if (!keep_chip_data) - desc->irq_data.chip_data = NULL; - desc->handle_irq = handle_bad_irq; - desc->irq_data.chip = &no_irq_chip; - desc->name = NULL; - clear_kstat_irqs(desc); - raw_spin_unlock_irqrestore(&desc->lock, flags); -} - -/** - * dynamic_irq_cleanup - cleanup a dynamically allocated irq - * @irq: irq number to initialize - */ -void dynamic_irq_cleanup(unsigned int irq) -{ - dynamic_irq_cleanup_x(irq, false); -} - -/** - * dynamic_irq_cleanup_keep_chip_data - cleanup a dynamically allocated irq - * @irq: irq number to initialize - * - * does not set irq_to_desc(irq)->irq_data.chip_data to NULL - */ -void dynamic_irq_cleanup_keep_chip_data(unsigned int irq) -{ - dynamic_irq_cleanup_x(irq, true); -} - - /** * set_irq_chip - set the irq chip for an irq * @irq: irq number diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index 8eb01e3..f444203 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -20,7 +20,6 @@ extern void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume); extern struct lock_class_key irq_desc_lock_class; extern void init_kstat_irqs(struct irq_desc *desc, int node, int nr); -extern void clear_kstat_irqs(struct irq_desc *desc); extern raw_spinlock_t sparse_irq_lock; /* Resending of interrupts :*/ diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index 6c71f8e..c9d5a1c 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -53,12 +53,21 @@ static void desc_smp_init(struct irq_desc *desc, int node) { desc->irq_data.node = node; cpumask_copy(desc->irq_data.affinity, irq_default_affinity); +#ifdef CONFIG_GENERIC_PENDING_IRQ + cpumask_clear(desc->pending_mask); +#endif +} + +static inline int desc_node(struct irq_desc *desc) +{ + return desc->irq_data.node; } #else static inline int alloc_masks(struct irq_desc *desc, gfp_t gfp, int node) { return 0; } static inline void desc_smp_init(struct irq_desc *desc, int node) { } +static inline int desc_node(struct irq_desc *desc) { return 0; } #endif static void desc_set_defaults(unsigned int irq, struct irq_desc *desc, int node) @@ -71,6 +80,8 @@ static void desc_set_defaults(unsigned int irq, struct irq_desc *desc, int node) desc->status = IRQ_DEFAULT_INIT_FLAGS; desc->handle_irq = handle_bad_irq; desc->depth = 1; + desc->irq_count = 0; + desc->irqs_unhandled = 0; desc->name = NULL; memset(desc->kstat_irqs, 0, nr_cpu_ids * sizeof(*(desc->kstat_irqs))); desc_smp_init(desc, node); @@ -286,23 +297,9 @@ struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node) return irq_to_desc(irq); } -#ifdef CONFIG_SMP -static inline int desc_node(struct irq_desc *desc) -{ - return desc->irq_data.node; -} -#else -static inline int desc_node(struct irq_desc *desc) { return 0; } -#endif - static void free_desc(unsigned int irq) { - struct irq_desc *desc = irq_to_desc(irq); - unsigned long flags; - - raw_spin_lock_irqsave(&desc->lock, flags); - desc_set_defaults(irq, desc, desc_node(desc)); - raw_spin_unlock_irqrestore(&desc->lock, flags); + dynamic_irq_cleanup(irq); } static inline int alloc_descs(unsigned int start, unsigned int cnt, int node) @@ -409,10 +406,18 @@ unsigned int irq_get_next_irq(unsigned int offset) return find_next_bit(allocated_irqs, nr_irqs, offset); } -/* Statistics access */ -void clear_kstat_irqs(struct irq_desc *desc) +/** + * dynamic_irq_cleanup - cleanup a dynamically allocated irq + * @irq: irq number to initialize + */ +void dynamic_irq_cleanup(unsigned int irq) { - memset(desc->kstat_irqs, 0, nr_cpu_ids * sizeof(*(desc->kstat_irqs))); + struct irq_desc *desc = irq_to_desc(irq); + unsigned long flags; + + raw_spin_lock_irqsave(&desc->lock, flags); + desc_set_defaults(irq, desc, desc_node(desc)); + raw_spin_unlock_irqrestore(&desc->lock, flags); } unsigned int kstat_irqs_cpu(unsigned int irq, int cpu) -- cgit v0.10.2 From 78f90d91f395cd0dc1ef3f21e0c5cd6fd50d202c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 29 Sep 2010 17:18:47 +0200 Subject: genirq: Remove the now unused sparse irq leftovers The move_irq_desc() function was only used due to the problem that the allocator did not free the old descriptors. So the descriptors had to be moved in create_irq_nr(). That's history. The code would have never been able to move active interrupt descriptors on affinity settings. That can be done in a completely different way w/o all this horror. Remove all of it. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/include/linux/irqdesc.h b/include/linux/irqdesc.h index f77dc56..979c68c 100644 --- a/include/linux/irqdesc.h +++ b/include/linux/irqdesc.h @@ -82,24 +82,16 @@ struct irq_desc { const char *name; } ____cacheline_internodealigned_in_smp; -extern void arch_init_copy_chip_data(struct irq_desc *old_desc, - struct irq_desc *desc, int node); -extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc); - #ifndef CONFIG_SPARSE_IRQ extern struct irq_desc irq_desc[NR_IRQS]; #endif -#ifdef CONFIG_NUMA_IRQ_DESC -extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int node); -#else +/* Will be removed once the last users in power and sh are gone */ +extern struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node); static inline struct irq_desc *move_irq_desc(struct irq_desc *desc, int node) { return desc; } -#endif - -extern struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node); #ifdef CONFIG_GENERIC_HARDIRQS diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig index a42c019..31d766b 100644 --- a/kernel/irq/Kconfig +++ b/kernel/irq/Kconfig @@ -26,11 +26,6 @@ config GENERIC_IRQ_PROBE config GENERIC_PENDING_IRQ def_bool n -if SPARSE_IRQ && NUMA -config NUMA_IRQ_DESC - def_bool n -endif - config AUTO_IRQ_AFFINITY def_bool n diff --git a/kernel/irq/Makefile b/kernel/irq/Makefile index 1eaab0da..54329cd 100644 --- a/kernel/irq/Makefile +++ b/kernel/irq/Makefile @@ -3,5 +3,4 @@ obj-y := irqdesc.o handle.o manage.o spurious.o resend.o chip.o dummychip.o devr obj-$(CONFIG_GENERIC_IRQ_PROBE) += autoprobe.o obj-$(CONFIG_PROC_FS) += proc.o obj-$(CONFIG_GENERIC_PENDING_IRQ) += migration.o -obj-$(CONFIG_NUMA_IRQ_DESC) += numa_migrate.o obj-$(CONFIG_PM_SLEEP) += pm.o diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index f444203..4571ae7 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -18,17 +18,11 @@ extern int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, extern void __disable_irq(struct irq_desc *desc, unsigned int irq, bool susp); extern void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume); -extern struct lock_class_key irq_desc_lock_class; extern void init_kstat_irqs(struct irq_desc *desc, int node, int nr); -extern raw_spinlock_t sparse_irq_lock; /* Resending of interrupts :*/ void check_irq_resend(struct irq_desc *desc, unsigned int irq); -#ifdef CONFIG_SPARSE_IRQ -void replace_irq_desc(unsigned int irq, struct irq_desc *desc); -#endif - #ifdef CONFIG_PROC_FS extern void register_irq_proc(unsigned int irq, struct irq_desc *desc); extern void unregister_irq_proc(unsigned int irq, struct irq_desc *desc); @@ -110,99 +104,3 @@ static inline void print_irq_desc(unsigned int irq, struct irq_desc *desc) #undef P -/* Stuff below will be cleaned up after the sparse allocator is done */ - -#ifdef CONFIG_SMP -/** - * alloc_desc_masks - allocate cpumasks for irq_desc - * @desc: pointer to irq_desc struct - * @node: node which will be handling the cpumasks - * @boot: true if need bootmem - * - * Allocates affinity and pending_mask cpumask if required. - * Returns true if successful (or not required). - */ -static inline bool alloc_desc_masks(struct irq_desc *desc, int node, - bool boot) -{ - gfp_t gfp = GFP_ATOMIC; - - if (boot) - gfp = GFP_NOWAIT; - -#ifdef CONFIG_CPUMASK_OFFSTACK - if (!alloc_cpumask_var_node(&desc->irq_data.affinity, gfp, node)) - return false; - -#ifdef CONFIG_GENERIC_PENDING_IRQ - if (!alloc_cpumask_var_node(&desc->pending_mask, gfp, node)) { - free_cpumask_var(desc->irq_data.affinity); - return false; - } -#endif -#endif - return true; -} - -static inline void init_desc_masks(struct irq_desc *desc) -{ - cpumask_setall(desc->irq_data.affinity); -#ifdef CONFIG_GENERIC_PENDING_IRQ - cpumask_clear(desc->pending_mask); -#endif -} - -/** - * init_copy_desc_masks - copy cpumasks for irq_desc - * @old_desc: pointer to old irq_desc struct - * @new_desc: pointer to new irq_desc struct - * - * Insures affinity and pending_masks are copied to new irq_desc. - * If !CONFIG_CPUMASKS_OFFSTACK the cpumasks are embedded in the - * irq_desc struct so the copy is redundant. - */ - -static inline void init_copy_desc_masks(struct irq_desc *old_desc, - struct irq_desc *new_desc) -{ -#ifdef CONFIG_CPUMASK_OFFSTACK - cpumask_copy(new_desc->irq_data.affinity, old_desc->irq_data.affinity); - -#ifdef CONFIG_GENERIC_PENDING_IRQ - cpumask_copy(new_desc->pending_mask, old_desc->pending_mask); -#endif -#endif -} - -static inline void free_desc_masks(struct irq_desc *old_desc, - struct irq_desc *new_desc) -{ - free_cpumask_var(old_desc->irq_data.affinity); - -#ifdef CONFIG_GENERIC_PENDING_IRQ - free_cpumask_var(old_desc->pending_mask); -#endif -} - -#else /* !CONFIG_SMP */ - -static inline bool alloc_desc_masks(struct irq_desc *desc, int node, - bool boot) -{ - return true; -} - -static inline void init_desc_masks(struct irq_desc *desc) -{ -} - -static inline void init_copy_desc_masks(struct irq_desc *old_desc, - struct irq_desc *new_desc) -{ -} - -static inline void free_desc_masks(struct irq_desc *old_desc, - struct irq_desc *new_desc) -{ -} -#endif /* CONFIG_SMP */ diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index c9d5a1c..4f0b9c9 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -20,7 +20,7 @@ /* * lockdep: we want to handle all irq_desc locks as a single lock-class: */ -struct lock_class_key irq_desc_lock_class; +static struct lock_class_key irq_desc_lock_class; #if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS) static void __init init_irq_default_affinity(void) @@ -90,28 +90,11 @@ static void desc_set_defaults(unsigned int irq, struct irq_desc *desc, int node) int nr_irqs = NR_IRQS; EXPORT_SYMBOL_GPL(nr_irqs); -DEFINE_RAW_SPINLOCK(sparse_irq_lock); +static DEFINE_RAW_SPINLOCK(sparse_irq_lock); static DECLARE_BITMAP(allocated_irqs, NR_IRQS); #ifdef CONFIG_SPARSE_IRQ -void __ref init_kstat_irqs(struct irq_desc *desc, int node, int nr) -{ - void *ptr; - - ptr = kzalloc_node(nr * sizeof(*desc->kstat_irqs), - GFP_ATOMIC, node); - - /* - * don't overwite if can not get new one - * init_copy_kstat_irqs() could still use old one - */ - if (ptr) { - printk(KERN_DEBUG " alloc kstat_irqs on node %d\n", node); - desc->kstat_irqs = ptr; - } -} - static RADIX_TREE(irq_desc_tree, GFP_ATOMIC); static void irq_insert_desc(unsigned int irq, struct irq_desc *desc) @@ -124,15 +107,6 @@ struct irq_desc *irq_to_desc(unsigned int irq) return radix_tree_lookup(&irq_desc_tree, irq); } -void replace_irq_desc(unsigned int irq, struct irq_desc *desc) -{ - void **ptr; - - ptr = radix_tree_lookup_slot(&irq_desc_tree, irq); - if (ptr) - radix_tree_replace_slot(ptr, desc); -} - static void delete_irq_desc(unsigned int irq) { radix_tree_delete(&irq_desc_tree, irq); diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c deleted file mode 100644 index e7f1f16..0000000 --- a/kernel/irq/numa_migrate.c +++ /dev/null @@ -1,120 +0,0 @@ -/* - * NUMA irq-desc migration code - * - * Migrate IRQ data structures (irq_desc, chip_data, etc.) over to - * the new "home node" of the IRQ. - */ - -#include -#include -#include -#include -#include -#include - -#include "internals.h" - -static void init_copy_kstat_irqs(struct irq_desc *old_desc, - struct irq_desc *desc, - int node, int nr) -{ - init_kstat_irqs(desc, node, nr); - - if (desc->kstat_irqs != old_desc->kstat_irqs) - memcpy(desc->kstat_irqs, old_desc->kstat_irqs, - nr * sizeof(*desc->kstat_irqs)); -} - -static void free_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc) -{ - if (old_desc->kstat_irqs == desc->kstat_irqs) - return; - - kfree(old_desc->kstat_irqs); - old_desc->kstat_irqs = NULL; -} - -static bool init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, - struct irq_desc *desc, int node) -{ - memcpy(desc, old_desc, sizeof(struct irq_desc)); - if (!alloc_desc_masks(desc, node, false)) { - printk(KERN_ERR "irq %d: can not get new irq_desc cpumask " - "for migration.\n", irq); - return false; - } - raw_spin_lock_init(&desc->lock); - desc->irq_data.node = node; - lockdep_set_class(&desc->lock, &irq_desc_lock_class); - init_copy_kstat_irqs(old_desc, desc, node, nr_cpu_ids); - init_copy_desc_masks(old_desc, desc); - arch_init_copy_chip_data(old_desc, desc, node); - return true; -} - -static void free_one_irq_desc(struct irq_desc *old_desc, struct irq_desc *desc) -{ - free_kstat_irqs(old_desc, desc); - free_desc_masks(old_desc, desc); - arch_free_chip_data(old_desc, desc); -} - -static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, - int node) -{ - struct irq_desc *desc; - unsigned int irq; - unsigned long flags; - - irq = old_desc->irq_data.irq; - - raw_spin_lock_irqsave(&sparse_irq_lock, flags); - - /* We have to check it to avoid races with another CPU */ - desc = irq_to_desc(irq); - - if (desc && old_desc != desc) - goto out_unlock; - - desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); - if (!desc) { - printk(KERN_ERR "irq %d: can not get new irq_desc " - "for migration.\n", irq); - /* still use old one */ - desc = old_desc; - goto out_unlock; - } - if (!init_copy_one_irq_desc(irq, old_desc, desc, node)) { - /* still use old one */ - kfree(desc); - desc = old_desc; - goto out_unlock; - } - - replace_irq_desc(irq, desc); - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); - - /* free the old one */ - free_one_irq_desc(old_desc, desc); - kfree(old_desc); - - return desc; - -out_unlock: - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); - - return desc; -} - -struct irq_desc *move_irq_desc(struct irq_desc *desc, int node) -{ - /* those static or target node is -1, do not move them */ - if (desc->irq_data.irq < NR_IRQS_LEGACY || node == -1) - return desc; - - if (desc->irq_data.node != node) - desc = __real_move_irq_desc(desc, node); - - return desc; -} - -- cgit v0.10.2 From c2f31c37b7303150ffcce53f6c6ed4ac62952b34 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 30 Sep 2010 12:19:03 +0200 Subject: x86: lguest: Use new irq allocator Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar Cc: Rusty Russell diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c index 2d4e6fc..73b1e1a 100644 --- a/arch/x86/lguest/boot.c +++ b/arch/x86/lguest/boot.c @@ -838,12 +838,12 @@ static void __init lguest_init_IRQ(void) * rather than set them in lguest_init_IRQ we are called here every time an * lguest device needs an interrupt. * - * FIXME: irq_to_desc_alloc_node() can fail due to lack of memory, we should + * FIXME: irq_alloc_desc_at() can fail due to lack of memory, we should * pass that up! */ void lguest_setup_irq(unsigned int irq) { - irq_to_desc_alloc_node(irq, 0); + irq_alloc_desc_at(irq, 0); set_irq_chip_and_handler_name(irq, &lguest_irq_controller, handle_level_irq, "level"); } -- cgit v0.10.2 From a05a900a51c7622ebd8ddb41f14f8bf9db599d8d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 8 Oct 2010 12:47:53 +0200 Subject: genirq: Make sparse_lock a mutex No callers from atomic regions. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index 4f0b9c9..0e302f9 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -90,7 +90,7 @@ static void desc_set_defaults(unsigned int irq, struct irq_desc *desc, int node) int nr_irqs = NR_IRQS; EXPORT_SYMBOL_GPL(nr_irqs); -static DEFINE_RAW_SPINLOCK(sparse_irq_lock); +static DEFINE_MUTEX(sparse_irq_lock); static DECLARE_BITMAP(allocated_irqs, NR_IRQS); #ifdef CONFIG_SPARSE_IRQ @@ -159,13 +159,12 @@ err_desc: static void free_desc(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); - unsigned long flags; unregister_irq_proc(irq, desc); - raw_spin_lock_irqsave(&sparse_irq_lock, flags); + mutex_lock(&sparse_irq_lock); delete_irq_desc(irq); - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); + mutex_unlock(&sparse_irq_lock); free_masks(desc); kfree(desc->kstat_irqs); @@ -175,16 +174,15 @@ static void free_desc(unsigned int irq) static int alloc_descs(unsigned int start, unsigned int cnt, int node) { struct irq_desc *desc; - unsigned long flags; int i; for (i = 0; i < cnt; i++) { desc = alloc_desc(start + i, node); if (!desc) goto err; - raw_spin_lock_irqsave(&sparse_irq_lock, flags); + mutex_lock(&sparse_irq_lock); irq_insert_desc(start + i, desc); - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); + mutex_unlock(&sparse_irq_lock); } return start; @@ -192,9 +190,9 @@ err: for (i--; i >= 0; i--) free_desc(start + i); - raw_spin_lock_irqsave(&sparse_irq_lock, flags); + mutex_lock(&sparse_irq_lock); bitmap_clear(allocated_irqs, start, cnt); - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); + mutex_unlock(&sparse_irq_lock); return -ENOMEM; } @@ -291,7 +289,6 @@ static inline int alloc_descs(unsigned int start, unsigned int cnt, int node) */ void irq_free_descs(unsigned int from, unsigned int cnt) { - unsigned long flags; int i; if (from >= nr_irqs || (from + cnt) > nr_irqs) @@ -300,9 +297,9 @@ void irq_free_descs(unsigned int from, unsigned int cnt) for (i = 0; i < cnt; i++) free_desc(from + i); - raw_spin_lock_irqsave(&sparse_irq_lock, flags); + mutex_lock(&sparse_irq_lock); bitmap_clear(allocated_irqs, from, cnt); - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); + mutex_unlock(&sparse_irq_lock); } /** @@ -317,13 +314,12 @@ void irq_free_descs(unsigned int from, unsigned int cnt) int __ref irq_alloc_descs(int irq, unsigned int from, unsigned int cnt, int node) { - unsigned long flags; int start, ret; if (!cnt) return -EINVAL; - raw_spin_lock_irqsave(&sparse_irq_lock, flags); + mutex_lock(&sparse_irq_lock); start = bitmap_find_next_zero_area(allocated_irqs, nr_irqs, from, cnt, 0); ret = -EEXIST; @@ -335,11 +331,11 @@ irq_alloc_descs(int irq, unsigned int from, unsigned int cnt, int node) goto err; bitmap_set(allocated_irqs, start, cnt); - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); + mutex_unlock(&sparse_irq_lock); return alloc_descs(start, cnt, node); err: - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); + mutex_unlock(&sparse_irq_lock); return ret; } @@ -352,20 +348,19 @@ err: */ int irq_reserve_irqs(unsigned int from, unsigned int cnt) { - unsigned long flags; unsigned int start; int ret = 0; if (!cnt || (from + cnt) > nr_irqs) return -EINVAL; - raw_spin_lock_irqsave(&sparse_irq_lock, flags); + mutex_lock(&sparse_irq_lock); start = bitmap_find_next_zero_area(allocated_irqs, nr_irqs, from, cnt, 0); if (start == from) bitmap_set(allocated_irqs, start, cnt); else ret = -EEXIST; - raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); + mutex_unlock(&sparse_irq_lock); return ret; } -- cgit v0.10.2 From baa0d233afe765daa6dc01ff233aea8c5944f534 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 5 Oct 2010 15:14:35 +0200 Subject: genirq: Switch sparse_irq allocator to GFP_KERNEL The allocator functions are now called outside of preempt disabled regions. Switch to GFP_KERNEL. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index 0e302f9..f6f660c 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -95,7 +95,7 @@ static DECLARE_BITMAP(allocated_irqs, NR_IRQS); #ifdef CONFIG_SPARSE_IRQ -static RADIX_TREE(irq_desc_tree, GFP_ATOMIC); +static RADIX_TREE(irq_desc_tree, GFP_KERNEL); static void irq_insert_desc(unsigned int irq, struct irq_desc *desc) { @@ -126,9 +126,8 @@ static inline void free_masks(struct irq_desc *desc) { } static struct irq_desc *alloc_desc(int irq, int node) { - /* Temporary hack until we can switch to GFP_KERNEL */ - gfp_t gfp = gfp_allowed_mask == GFP_BOOT_MASK ? GFP_NOWAIT : GFP_ATOMIC; struct irq_desc *desc; + gfp_t gfp = GFP_KERNEL; desc = kzalloc_node(sizeof(*desc), gfp, node); if (!desc) -- cgit v0.10.2 From 2ee39065988d632b403f8470942b0b5edee3632b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 6 Oct 2010 16:28:51 +0200 Subject: x86: Switch sparse_irq allocations to GFP_KERNEL No callers from atomic context (except boot) anymore. Signed-off-by: Thomas Gleixner Reviewed-by: Ingo Molnar diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 057b0e1..20e47e0 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -133,7 +133,7 @@ struct irq_pin_list { static struct irq_pin_list *alloc_irq_pin_list(int node) { - return kzalloc_node(sizeof(struct irq_pin_list), GFP_ATOMIC, node); + return kzalloc_node(sizeof(struct irq_pin_list), GFP_KERNEL, node); } /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */ @@ -162,8 +162,8 @@ int __init arch_early_irq_init(void) for (i = 0; i < count; i++) { set_irq_chip_data(i, &cfg[i]); - zalloc_cpumask_var_node(&cfg[i].domain, GFP_NOWAIT, node); - zalloc_cpumask_var_node(&cfg[i].old_domain, GFP_NOWAIT, node); + zalloc_cpumask_var_node(&cfg[i].domain, GFP_KERNEL, node); + zalloc_cpumask_var_node(&cfg[i].old_domain, GFP_KERNEL, node); /* * For legacy IRQ's, start with assigning irq0 to irq15 to * IRQ0_VECTOR to IRQ15_VECTOR on cpu 0. @@ -187,12 +187,12 @@ static struct irq_cfg *alloc_irq_cfg(unsigned int irq, int node) { struct irq_cfg *cfg; - cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); + cfg = kzalloc_node(sizeof(*cfg), GFP_KERNEL, node); if (!cfg) return NULL; - if (!zalloc_cpumask_var_node(&cfg->domain, GFP_ATOMIC, node)) + if (!zalloc_cpumask_var_node(&cfg->domain, GFP_KERNEL, node)) goto out_cfg; - if (!zalloc_cpumask_var_node(&cfg->old_domain, GFP_ATOMIC, node)) + if (!zalloc_cpumask_var_node(&cfg->old_domain, GFP_KERNEL, node)) goto out_domain; return cfg; out_domain: @@ -595,14 +595,14 @@ struct IO_APIC_route_entry **alloc_ioapic_entries(void) struct IO_APIC_route_entry **ioapic_entries; ioapic_entries = kzalloc(sizeof(*ioapic_entries) * nr_ioapics, - GFP_ATOMIC); + GFP_KERNEL); if (!ioapic_entries) return 0; for (apic = 0; apic < nr_ioapics; apic++) { ioapic_entries[apic] = kzalloc(sizeof(struct IO_APIC_route_entry) * - nr_ioapic_registers[apic], GFP_ATOMIC); + nr_ioapic_registers[apic], GFP_KERNEL); if (!ioapic_entries[apic]) goto nomem; } -- cgit v0.10.2 From 10d0ff83078d5ef99aaf94a2aaceb740f5ced923 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:31:58 +0000 Subject: input: Misc/hp_sdc_rtc: semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: Dmitry Torokhov LKML-Reference: <20100907125054.888438853@linutronix.de> diff --git a/drivers/input/misc/hp_sdc_rtc.c b/drivers/input/misc/hp_sdc_rtc.c index c190664..7e2c12a 100644 --- a/drivers/input/misc/hp_sdc_rtc.c +++ b/drivers/input/misc/hp_sdc_rtc.c @@ -104,7 +104,7 @@ static int hp_sdc_rtc_do_read_bbrtc (struct rtc_time *rtctm) t.endidx = 91; t.seq = tseq; t.act.semaphore = &tsem; - init_MUTEX_LOCKED(&tsem); + sema_init(&tsem, 0); if (hp_sdc_enqueue_transaction(&t)) return -1; @@ -698,7 +698,7 @@ static int __init hp_sdc_rtc_init(void) return -ENODEV; #endif - init_MUTEX(&i8042tregs); + sema_init(&i8042tregs, 1); if ((ret = hp_sdc_request_timer_irq(&hp_sdc_rtc_isr))) return ret; -- cgit v0.10.2 From 45e8492f594dcb17e8cbb26edb2dd3018a68dada Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:01 +0000 Subject: input: Serio/hil_mlc: semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: Dmitry Torokhov LKML-Reference: <20100907125054.985090435@linutronix.de> diff --git a/drivers/input/serio/hil_mlc.c b/drivers/input/serio/hil_mlc.c index c92f4ed..e5624d8 100644 --- a/drivers/input/serio/hil_mlc.c +++ b/drivers/input/serio/hil_mlc.c @@ -915,15 +915,15 @@ int hil_mlc_register(hil_mlc *mlc) mlc->ostarted = 0; rwlock_init(&mlc->lock); - init_MUTEX(&mlc->osem); + sema_init(&mlc->osem, 1); - init_MUTEX(&mlc->isem); + sema_init(&mlc->isem, 1); mlc->icount = -1; mlc->imatch = 0; mlc->opercnt = 0; - init_MUTEX_LOCKED(&(mlc->csem)); + sema_init(&(mlc->csem), 0); hil_mlc_clear_di_scratch(mlc); hil_mlc_clear_di_map(mlc, 0); -- cgit v0.10.2 From 4a4a506f81eebe63fb1b8e9e99f4bcd1fd9aac66 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:06 +0000 Subject: input: Serio/hp_sdc: semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: Dmitry Torokhov LKML-Reference: <20100907125055.079737758@linutronix.de> diff --git a/drivers/input/serio/hp_sdc.c b/drivers/input/serio/hp_sdc.c index bcc2d30..8c0b51c 100644 --- a/drivers/input/serio/hp_sdc.c +++ b/drivers/input/serio/hp_sdc.c @@ -905,7 +905,7 @@ static int __init hp_sdc_init(void) ts_sync[1] = 0x0f; ts_sync[2] = ts_sync[3] = ts_sync[4] = ts_sync[5] = 0; t_sync.act.semaphore = &s_sync; - init_MUTEX_LOCKED(&s_sync); + sema_init(&s_sync, 0); hp_sdc_enqueue_transaction(&t_sync); down(&s_sync); /* Wait for t_sync to complete */ @@ -1039,7 +1039,7 @@ static int __init hp_sdc_register(void) return hp_sdc.dev_err; } - init_MUTEX_LOCKED(&tq_init_sem); + sema_init(&tq_init_sem, 0); tq_init.actidx = 0; tq_init.idx = 1; -- cgit v0.10.2 From 50948ee8124c2125d833e8889002aa12a948b4e9 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:10 +0000 Subject: net: 3c527: semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: David Miller LKML-Reference: <20100907125055.175750769@linutronix.de> diff --git a/drivers/net/3c527.c b/drivers/net/3c527.c index 70705d1..eca55c5 100644 --- a/drivers/net/3c527.c +++ b/drivers/net/3c527.c @@ -522,7 +522,7 @@ static int __init mc32_probe1(struct net_device *dev, int slot) lp->tx_len = lp->exec_box->data[9]; /* Transmit list count */ lp->rx_len = lp->exec_box->data[11]; /* Receive list count */ - init_MUTEX_LOCKED(&lp->cmd_mutex); + sema_init(&lp->cmd_mutex, 0); init_completion(&lp->execution_cmd); init_completion(&lp->xceiver_cmd); -- cgit v0.10.2 From 89d9f10d0b21268d3dba33db984ab03092861700 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:14 +0000 Subject: hamradio: 6pack: semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: David Miller LKML-Reference: <20100907125055.269142443@linutronix.de> diff --git a/drivers/net/hamradio/6pack.c b/drivers/net/hamradio/6pack.c index 4b52c76..3e5d0b6 100644 --- a/drivers/net/hamradio/6pack.c +++ b/drivers/net/hamradio/6pack.c @@ -608,7 +608,7 @@ static int sixpack_open(struct tty_struct *tty) spin_lock_init(&sp->lock); atomic_set(&sp->refcnt, 1); - init_MUTEX_LOCKED(&sp->dead_sem); + sema_init(&sp->dead_sem, 0); /* !!! length of the buffers. MTU is IP MTU, not PACLEN! */ -- cgit v0.10.2 From 495e64bddeb9349395ed346177057fff7d52857e Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:18 +0000 Subject: hamradio: Mkiss: semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: David Miller LKML-Reference: <20100907125055.368389976@linutronix.de> diff --git a/drivers/net/hamradio/mkiss.c b/drivers/net/hamradio/mkiss.c index 66e88bd..4c62839 100644 --- a/drivers/net/hamradio/mkiss.c +++ b/drivers/net/hamradio/mkiss.c @@ -747,7 +747,7 @@ static int mkiss_open(struct tty_struct *tty) spin_lock_init(&ax->buflock); atomic_set(&ax->refcnt, 1); - init_MUTEX_LOCKED(&ax->dead_sem); + sema_init(&ax->dead_sem, 0); ax->tty = tty; tty->disc_data = ax; -- cgit v0.10.2 From 0bce198b7688911e072bd65032ab9278b577ee93 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:22 +0000 Subject: net: Ppp_async: semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: David Miller LKML-Reference: <20100907125055.460765783@linutronix.de> diff --git a/drivers/net/ppp_async.c b/drivers/net/ppp_async.c index af50a53..78d70a6 100644 --- a/drivers/net/ppp_async.c +++ b/drivers/net/ppp_async.c @@ -184,7 +184,7 @@ ppp_asynctty_open(struct tty_struct *tty) tasklet_init(&ap->tsk, ppp_async_process, (unsigned long) ap); atomic_set(&ap->refcnt, 1); - init_MUTEX_LOCKED(&ap->dead_sem); + sema_init(&ap->dead_sem, 0); ap->chan.private = ap; ap->chan.ops = &async_ops; -- cgit v0.10.2 From d1985508320ad40a68375bdbafa8a5d6c56d2114 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:25 +0000 Subject: net: Wan/cosa.c: Convert "mutex" to semaphore Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: "David S. Miller" LKML-Reference: <20100907125055.557578360@linutronix.de> diff --git a/drivers/net/wan/cosa.c b/drivers/net/wan/cosa.c index 04c6cd4..10bafd5 100644 --- a/drivers/net/wan/cosa.c +++ b/drivers/net/wan/cosa.c @@ -575,7 +575,7 @@ static int cosa_probe(int base, int irq, int dma) /* Initialize the chardev data structures */ mutex_init(&chan->rlock); - init_MUTEX(&chan->wsem); + sema_init(&chan->wsem, 1); /* Register the network interface */ if (!(chan->netdev = alloc_hdlcdev(chan))) { -- cgit v0.10.2 From db7f859ec53c0bfaf3520afe2feafd18a27e48fb Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:30 +0000 Subject: irda: Semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Acked-by: "David S. Miller" LKML-Reference: <20100907125055.651362456@linutronix.de> diff --git a/drivers/net/irda/sir_dev.c b/drivers/net/irda/sir_dev.c index 1b051da..51d7444 100644 --- a/drivers/net/irda/sir_dev.c +++ b/drivers/net/irda/sir_dev.c @@ -909,7 +909,7 @@ struct sir_dev * sirdev_get_instance(const struct sir_driver *drv, const char *n dev->tx_skb = NULL; spin_lock_init(&dev->tx_lock); - init_MUTEX(&dev->fsm.sem); + sema_init(&dev->fsm.sem, 1); dev->drv = drv; dev->netdev = ndev; -- cgit v0.10.2 From a529f1505b6facfd0078dcb18a444976c1985962 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:32:34 +0000 Subject: parport: Semaphore cleanup Get rid of init_MUTEX[_LOCKED]() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Alan Cox Cc: Christoph Hellwig LKML-Reference: <20100907125055.743613774@linutronix.de> diff --git a/drivers/parport/share.c b/drivers/parport/share.c index dffa5d4..a2d9d1e 100644 --- a/drivers/parport/share.c +++ b/drivers/parport/share.c @@ -306,7 +306,7 @@ struct parport *parport_register_port(unsigned long base, int irq, int dma, spin_lock_init(&tmp->pardevice_lock); tmp->ieee1284.mode = IEEE1284_MODE_COMPAT; tmp->ieee1284.phase = IEEE1284_PH_FWD_IDLE; - init_MUTEX_LOCKED (&tmp->ieee1284.irq); /* actually a semaphore at 0 */ + sema_init(&tmp->ieee1284.irq, 0); tmp->spintime = parport_default_spintime; atomic_set (&tmp->ref_count, 1); INIT_LIST_HEAD(&tmp->full_list); -- cgit v0.10.2 From 8192b1f6b1a46b33213b993471356495a93ffc70 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:33:40 +0000 Subject: drivers/macintosh/adb: Do not claim that the semaphore is a mutex User DEFINE_SEMAPHORE() instead of DECLARE_MUTEX() Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Cc: Benjamin Herrenschmidt LKML-Reference: <20100907125057.086367802@linutronix.de> diff --git a/drivers/macintosh/adb.c b/drivers/macintosh/adb.c index 1c4ee6e..bf64e49 100644 --- a/drivers/macintosh/adb.c +++ b/drivers/macintosh/adb.c @@ -83,7 +83,7 @@ static struct adb_driver *adb_controller; BLOCKING_NOTIFIER_HEAD(adb_client_list); static int adb_got_sleep; static int adb_inited; -static DECLARE_MUTEX(adb_probe_mutex); +static DEFINE_SEMAPHORE(adb_probe_mutex); static int sleepy_trackpad; static int autopoll_devs; int __adb_probe_sync; -- cgit v0.10.2 From 5b8c4f23c54edda640a4850bc9b81dee70acb525 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:33:43 +0000 Subject: printk: Make console_sem a semaphore not a pseudo mutex It needs to be investigated whether it can be replaced by a real mutex, but that needs more thought. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig LKML-Reference: <20100907125057.179587334@linutronix.de> diff --git a/kernel/printk.c b/kernel/printk.c index 8fe465a..2531017 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -85,7 +85,7 @@ EXPORT_SYMBOL(oops_in_progress); * provides serialisation for access to the entire console * driver system. */ -static DECLARE_MUTEX(console_sem); +static DEFINE_SEMAPHORE(console_sem); struct console *console_drivers; EXPORT_SYMBOL_GPL(console_drivers); @@ -556,7 +556,7 @@ static void zap_locks(void) /* If a crash is occurring, make sure we can't deadlock */ spin_lock_init(&logbuf_lock); /* And make sure that we print immediately */ - init_MUTEX(&console_sem); + sema_init(&console_sem, 1); } #if defined(CONFIG_PRINTK_TIME) -- cgit v0.10.2 From 5356d94872502e77317f82e8fcae2a0b162af07c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:33:52 +0000 Subject: arm: Bcmring: semaphore cleanup Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Cc: Leo Chen LKML-Reference: <20100907125057.371771121@linutronix.de> diff --git a/arch/arm/mach-bcmring/dma.c b/arch/arm/mach-bcmring/dma.c index 29c0a91..77eb35c 100644 --- a/arch/arm/mach-bcmring/dma.c +++ b/arch/arm/mach-bcmring/dma.c @@ -691,7 +691,7 @@ int dma_init(void) memset(&gDMA, 0, sizeof(gDMA)); - init_MUTEX_LOCKED(&gDMA.lock); + sema_init(&gDMA.lock, 0); init_waitqueue_head(&gDMA.freeChannelQ); /* Initialize the Hardware */ @@ -1574,7 +1574,7 @@ int dma_init_mem_map(DMA_MemMap_t *memMap) { memset(memMap, 0, sizeof(*memMap)); - init_MUTEX(&memMap->lock); + sema_init(&memMap->lock, 1); return 0; } -- cgit v0.10.2 From 4a9410355406c31695eab9daeba694d7d9714e9b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:33:08 +0000 Subject: hfs: Convert tree_lock to mutex tree_lock is used as mutex so make it a mutex. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Acked-by: Christoph Hellwig Cc: Al Viro LKML-Reference: <20100907125056.416332114@linutronix.de> diff --git a/fs/hfs/bfind.c b/fs/hfs/bfind.c index 4129cdb..571abe9 100644 --- a/fs/hfs/bfind.c +++ b/fs/hfs/bfind.c @@ -23,7 +23,7 @@ int hfs_find_init(struct hfs_btree *tree, struct hfs_find_data *fd) fd->search_key = ptr; fd->key = ptr + tree->max_key_len + 2; dprint(DBG_BNODE_REFS, "find_init: %d (%p)\n", tree->cnid, __builtin_return_address(0)); - down(&tree->tree_lock); + mutex_lock(&tree->tree_lock); return 0; } @@ -32,7 +32,7 @@ void hfs_find_exit(struct hfs_find_data *fd) hfs_bnode_put(fd->bnode); kfree(fd->search_key); dprint(DBG_BNODE_REFS, "find_exit: %d (%p)\n", fd->tree->cnid, __builtin_return_address(0)); - up(&fd->tree->tree_lock); + mutex_unlock(&fd->tree->tree_lock); fd->tree = NULL; } diff --git a/fs/hfs/btree.c b/fs/hfs/btree.c index 38a0a99..3ebc437 100644 --- a/fs/hfs/btree.c +++ b/fs/hfs/btree.c @@ -27,7 +27,7 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id, btree_keycmp ke if (!tree) return NULL; - init_MUTEX(&tree->tree_lock); + mutex_init(&tree->tree_lock); spin_lock_init(&tree->hash_lock); /* Set the correct compare function */ tree->sb = sb; diff --git a/fs/hfs/btree.h b/fs/hfs/btree.h index cc51905..2a1d712 100644 --- a/fs/hfs/btree.h +++ b/fs/hfs/btree.h @@ -33,7 +33,7 @@ struct hfs_btree { unsigned int depth; //unsigned int map1_size, map_size; - struct semaphore tree_lock; + struct mutex tree_lock; unsigned int pages_per_bnode; spinlock_t hash_lock; -- cgit v0.10.2 From 756b0322e50aebc4b9afb4488a2d3f6c802b4e64 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 7 Sep 2010 14:33:11 +0000 Subject: affs: Use sema_init instead of init_MUTEX Get rid of init_MUTE() and use sema_init() instead. Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: Christoph Hellwig Cc: Al Viro LKML-Reference: <20100907125056.511395595@linutronix.de> diff --git a/fs/affs/super.c b/fs/affs/super.c index 33c4e7e..9581ea9 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -109,8 +109,8 @@ static void init_once(void *foo) { struct affs_inode_info *ei = (struct affs_inode_info *) foo; - init_MUTEX(&ei->i_link_lock); - init_MUTEX(&ei->i_ext_lock); + sema_init(&ei->i_link_lock, 1); + sema_init(&ei->i_ext_lock, 1); inode_init_once(&ei->vfs_inode); } -- cgit v0.10.2 From d01343244abdedd18303d0323b518ed9cdcb1988 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 12 Oct 2010 12:06:43 -0400 Subject: ring-buffer: Fix typo of time extends per page Time stamps for the ring buffer are created by the difference between two events. Each page of the ring buffer holds a full 64 bit timestamp. Each event has a 27 bit delta stamp from the last event. The unit of time is nanoseconds, so 27 bits can hold ~134 milliseconds. If two events happen more than 134 milliseconds apart, a time extend is inserted to add more bits for the delta. The time extend has 59 bits, which is good for ~18 years. Currently the time extend is committed separately from the event. If an event is discarded before it is committed, due to filtering, the time extend still exists. If all events are being filtered, then after ~134 milliseconds a new time extend will be added to the buffer. This can only happen till the end of the page. Since each page holds a full timestamp, there is no reason to add a time extend to the beginning of a page. Time extends can only fill a page that has actual data at the beginning, so there is no fear that time extends will fill more than a page without any data. When reading an event, a loop is made to skip over time extends since they are only used to maintain the time stamp and are never given to the caller. As a paranoid check to prevent the loop running forever, with the knowledge that time extends may only fill a page, a check is made that tests the iteration of the loop, and if the iteration is more than the number of time extends that can fit in a page a warning is printed and the ring buffer is disabled (all of ftrace is also disabled with it). There is another event type that is called a TIMESTAMP which can hold 64 bits of data in the theoretical case that two events happen 18 years apart. This code has not been implemented, but the name of this event exists, as well as the structure for it. The size of a TIMESTAMP is 16 bytes, where as a time extend is only 8 bytes. The macro used to calculate how many time extends can fit on a page used the TIMESTAMP size instead of the time extend size cutting the amount in half. The following test case can easily trigger the warning since we only need to have half the page filled with time extends to trigger the warning: # cd /sys/kernel/debug/tracing/ # echo function > current_tracer # echo 'common_pid < 0' > events/ftrace/function/filter # echo > trace # echo 1 > trace_marker # sleep 120 # cat trace Enabling the function tracer and then setting the filter to only trace functions where the process id is negative (no events), then clearing the trace buffer to ensure that we have nothing in the buffer, then write to trace_marker to add an event to the beginning of a page, sleep for 2 minutes (only 35 seconds is probably needed, but this guarantees the bug), and then finally reading the trace which will trigger the bug. This patch fixes the typo and prevents the false positive of that warning. Reported-by: Hans J. Koch Tested-by: Hans J. Koch Cc: Thomas Gleixner Cc: Stable Kernel Signed-off-by: Steven Rostedt diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 492197e..bca9637 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -405,7 +405,7 @@ static inline int test_time_stamp(u64 delta) #define BUF_MAX_DATA_SIZE (BUF_PAGE_SIZE - (sizeof(u32) * 2)) /* Max number of timestamps that can fit on a page */ -#define RB_TIMESTAMPS_PER_PAGE (BUF_PAGE_SIZE / RB_LEN_TIME_STAMP) +#define RB_TIMESTAMPS_PER_PAGE (BUF_PAGE_SIZE / RB_LEN_TIME_EXTEND) int ring_buffer_print_page_header(struct trace_seq *s) { -- cgit v0.10.2 From 93055c31045a2d5599ec613a0c6cdcefc481a460 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Mon, 11 Oct 2010 21:41:35 +0200 Subject: ps3disk: passing wrong variable to bvec_kunmap_irq() This should pass "buf" to bvec_kunmap_irq() instead of "bv". The api is like kmap_atomic() instead of kmap(). Signed-off-by: Dan Carpenter Acked-by: Geoff Levand Signed-off-by: Jens Axboe diff --git a/drivers/block/ps3disk.c b/drivers/block/ps3disk.c index e9da874..03688c2 100644 --- a/drivers/block/ps3disk.c +++ b/drivers/block/ps3disk.c @@ -113,7 +113,7 @@ static void ps3disk_scatter_gather(struct ps3_storage_device *dev, memcpy(buf, dev->bounce_buf+offset, size); offset += size; flush_kernel_dcache_page(bvec->bv_page); - bvec_kunmap_irq(bvec, &flags); + bvec_kunmap_irq(buf, &flags); i++; } } -- cgit v0.10.2 From c0a19ebc018222ffd1dd93af5b53d9efd779c19b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 12 Oct 2010 21:58:27 +0200 Subject: genirq: Fix CONFIG_GENIRQ_NO_DEPRECATED=y build This option can be set to verify the full conversion to the new chip functions. Fix the fallout of the patch rework, so the core code compiles and works with it. Signed-off-by: Thomas Gleixner diff --git a/kernel/irq/dummychip.c b/kernel/irq/dummychip.c index 918dea9..20dc547 100644 --- a/kernel/irq/dummychip.c +++ b/kernel/irq/dummychip.c @@ -31,7 +31,7 @@ static unsigned int noop_ret(struct irq_data *data) return 0; } -#ifndef CONFIG_GENERIC_HARDIRQS_NO_CRUFT +#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED static void compat_noop(unsigned int irq) { } #define END_INIT .end = compat_noop #else diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index f6f660c..9d917ff 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -118,7 +118,7 @@ static void free_masks(struct irq_desc *desc) #ifdef CONFIG_GENERIC_PENDING_IRQ free_cpumask_var(desc->pending_mask); #endif - free_cpumask_var(desc->affinity); + free_cpumask_var(desc->irq_data.affinity); } #else static inline void free_masks(struct irq_desc *desc) { } -- cgit v0.10.2 From 10d48b3934940c178635413b5128c33bc17abe1d Mon Sep 17 00:00:00 2001 From: Mika Westerberg Date: Tue, 12 Oct 2010 10:37:59 +0100 Subject: ARM: 6440/1: ep93xx: DMA: fix channel_disable When channel_disable() is called, it disables per channel interrupts and waits until channels state becomes STATE_STALL, and then disables the channel. Now, if the DMA transfer is disabled while the channel is in STATE_NEXT we will not wait anything and disable the channel immediately. This seems to cause weird data corruption for example in audio transfers. Fix is to wait while we are in STATE_NEXT or STATE_ON and only then disable the channel. Signed-off-by: Mika Westerberg Acked-by: Ryan Mallon Signed-off-by: Russell King diff --git a/arch/arm/mach-ep93xx/dma-m2p.c b/arch/arm/mach-ep93xx/dma-m2p.c index 8904ca4..a696d35 100644 --- a/arch/arm/mach-ep93xx/dma-m2p.c +++ b/arch/arm/mach-ep93xx/dma-m2p.c @@ -276,7 +276,7 @@ static void channel_disable(struct m2p_channel *ch) v &= ~(M2P_CONTROL_STALL_IRQ_EN | M2P_CONTROL_NFB_IRQ_EN); m2p_set_control(ch, v); - while (m2p_channel_state(ch) == STATE_ON) + while (m2p_channel_state(ch) >= STATE_ON) cpu_relax(); m2p_set_control(ch, 0x0); -- cgit v0.10.2 From 06c10884486a63a1e4ff657aaa51e848e64b9dc3 Mon Sep 17 00:00:00 2001 From: Russell King Date: Wed, 13 Oct 2010 00:15:25 +0100 Subject: ARM: relax ioremap prohibition (309caa9) for -final and -stable ... but produce a big warning about the problem as encouragement for people to fix their drivers. Signed-off-by: Russell King diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c index ab50627..17e7b0b 100644 --- a/arch/arm/mm/ioremap.c +++ b/arch/arm/mm/ioremap.c @@ -204,8 +204,12 @@ void __iomem * __arm_ioremap_pfn_caller(unsigned long pfn, /* * Don't allow RAM to be mapped - this causes problems with ARMv6+ */ - if (WARN_ON(pfn_valid(pfn))) - return NULL; + if (pfn_valid(pfn)) { + printk(KERN_WARNING "BUG: Your driver calls ioremap() on system memory. This leads\n" + KERN_WARNING "to architecturally unpredictable behaviour on ARMv6+, and ioremap()\n" + KERN_WARNING "will fail in the next kernel release. Please fix your driver.\n"); + WARN_ON(1); + } type = get_mem_type(mtype); if (!type) -- cgit v0.10.2 From bf1ebf007911d70a89de9a4a1b36d403e8eb064b Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Sun, 10 Oct 2010 10:40:32 +0100 Subject: x86, olpc: Add XO-1 poweroff support Add a pm_power_off handler for the OLPC XO-1 laptop. The driver can be built modular and follows the behaviour of the APM driver, setting pm_power_off to NULL on unload. However, the ability to unload the module will probably be removed (with a simple __module_get(THIS_MODULE)) if/when XO-1 suspend/resume support is added to this file at a later date. Signed-off-by: Daniel Drake LKML-Reference: <20101010094032.9AE669D401B@zog.reactivated.net> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c255255..0bcb876 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -2066,6 +2066,12 @@ config OLPC Add support for detecting the unique features of the OLPC XO hardware. +config OLPC_XO1 + tristate "OLPC XO-1 support" + depends on OLPC + ---help--- + Add support for non-essential features of the OLPC XO-1 laptop. + config OLPC_OPENFIRMWARE bool "Support for OLPC's Open Firmware" depends on !X86_64 && !X86_PAE diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 0925676..e977779 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -104,6 +104,7 @@ obj-$(CONFIG_SCx200) += scx200.o scx200-y += scx200_32.o obj-$(CONFIG_OLPC) += olpc.o +obj-$(CONFIG_OLPC_XO1) += olpc-xo1.o obj-$(CONFIG_OLPC_OPENFIRMWARE) += olpc_ofw.o obj-$(CONFIG_X86_MRST) += mrst.o diff --git a/arch/x86/kernel/olpc-xo1.c b/arch/x86/kernel/olpc-xo1.c new file mode 100644 index 0000000..f5442c0 --- /dev/null +++ b/arch/x86/kernel/olpc-xo1.c @@ -0,0 +1,140 @@ +/* + * Support for features of the OLPC XO-1 laptop + * + * Copyright (C) 2010 One Laptop per Child + * Copyright (C) 2006 Red Hat, Inc. + * Copyright (C) 2006 Advanced Micro Devices, Inc. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + */ + +#include +#include +#include +#include +#include + +#include +#include + +#define DRV_NAME "olpc-xo1" + +#define PMS_BAR 4 +#define ACPI_BAR 5 + +/* PMC registers (PMS block) */ +#define PM_SCLK 0x10 +#define PM_IN_SLPCTL 0x20 +#define PM_WKXD 0x34 +#define PM_WKD 0x30 +#define PM_SSC 0x54 + +/* PM registers (ACPI block) */ +#define PM1_CNT 0x08 +#define PM_GPE0_STS 0x18 + +static unsigned long acpi_base; +static unsigned long pms_base; + +static void xo1_power_off(void) +{ + printk(KERN_INFO "OLPC XO-1 power off sequence...\n"); + + /* Enable all of these controls with 0 delay */ + outl(0x40000000, pms_base + PM_SCLK); + outl(0x40000000, pms_base + PM_IN_SLPCTL); + outl(0x40000000, pms_base + PM_WKXD); + outl(0x40000000, pms_base + PM_WKD); + + /* Clear status bits (possibly unnecessary) */ + outl(0x0002ffff, pms_base + PM_SSC); + outl(0xffffffff, acpi_base + PM_GPE0_STS); + + /* Write SLP_EN bit to start the machinery */ + outl(0x00002000, acpi_base + PM1_CNT); +} + +/* Read the base addresses from the PCI BAR info */ +static int __devinit setup_bases(struct pci_dev *pdev) +{ + int r; + + r = pci_enable_device_io(pdev); + if (r) { + dev_err(&pdev->dev, "can't enable device IO\n"); + return r; + } + + r = pci_request_region(pdev, ACPI_BAR, DRV_NAME); + if (r) { + dev_err(&pdev->dev, "can't alloc PCI BAR #%d\n", ACPI_BAR); + return r; + } + + r = pci_request_region(pdev, PMS_BAR, DRV_NAME); + if (r) { + dev_err(&pdev->dev, "can't alloc PCI BAR #%d\n", PMS_BAR); + pci_release_region(pdev, ACPI_BAR); + return r; + } + + acpi_base = pci_resource_start(pdev, ACPI_BAR); + pms_base = pci_resource_start(pdev, PMS_BAR); + + return 0; +} + +static int __devinit olpc_xo1_probe(struct platform_device *pdev) +{ + struct pci_dev *pcidev; + int r; + + pcidev = pci_get_device(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_CS5536_ISA, + NULL); + if (!pdev) + return -ENODEV; + + r = setup_bases(pcidev); + if (r) + return r; + + pm_power_off = xo1_power_off; + + printk(KERN_INFO "OLPC XO-1 support registered\n"); + return 0; +} + +static int __devexit olpc_xo1_remove(struct platform_device *pdev) +{ + pm_power_off = NULL; + return 0; +} + +static struct platform_driver olpc_xo1_driver = { + .driver = { + .name = DRV_NAME, + .owner = THIS_MODULE, + }, + .probe = olpc_xo1_probe, + .remove = __devexit_p(olpc_xo1_remove), +}; + +static int __init olpc_xo1_init(void) +{ + return platform_driver_register(&olpc_xo1_driver); +} + +static void __exit olpc_xo1_exit(void) +{ + platform_driver_unregister(&olpc_xo1_driver); +} + +MODULE_AUTHOR("Daniel Drake "); +MODULE_LICENSE("GPL"); +MODULE_ALIAS("platform:olpc-xo1"); + +module_init(olpc_xo1_init); +module_exit(olpc_xo1_exit); -- cgit v0.10.2 From 272036edb77e5967c175ac48b8c88e9b3358285f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Fran=C3=A7ois=20Jaouen?= Date: Wed, 13 Oct 2010 10:47:32 +0200 Subject: HID: Add Cando touch screen 15.6-inch product id This add the product id of the touch screen found on ACER Aspire 5738PZ. Works with hid-cando driver. Signed-off-by: Francois Jaouen Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-cando.c b/drivers/hid/hid-cando.c index 4267a6f..5925bdc 100644 --- a/drivers/hid/hid-cando.c +++ b/drivers/hid/hid-cando.c @@ -237,6 +237,8 @@ static const struct hid_device_id cando_devices[] = { USB_DEVICE_ID_CANDO_MULTI_TOUCH) }, { HID_USB_DEVICE(USB_VENDOR_ID_CANDO, USB_DEVICE_ID_CANDO_MULTI_TOUCH_11_6) }, + { HID_USB_DEVICE(USB_VENDOR_ID_CANDO, + USB_DEVICE_ID_CANDO_MULTI_TOUCH_15_6) }, { } }; MODULE_DEVICE_TABLE(hid, cando_devices); diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 3f72924..a0dea3d 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1292,6 +1292,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_BTC, USB_DEVICE_ID_BTC_EMPREX_REMOTE_2) }, { HID_USB_DEVICE(USB_VENDOR_ID_CANDO, USB_DEVICE_ID_CANDO_MULTI_TOUCH) }, { HID_USB_DEVICE(USB_VENDOR_ID_CANDO, USB_DEVICE_ID_CANDO_MULTI_TOUCH_11_6) }, + { HID_USB_DEVICE(USB_VENDOR_ID_CANDO, USB_DEVICE_ID_CANDO_MULTI_TOUCH_15_6) }, { HID_USB_DEVICE(USB_VENDOR_ID_CHERRY, USB_DEVICE_ID_CHERRY_CYMOTION) }, { HID_USB_DEVICE(USB_VENDOR_ID_CHERRY, USB_DEVICE_ID_CHERRY_CYMOTION_SOLAR) }, { HID_USB_DEVICE(USB_VENDOR_ID_CHICONY, USB_DEVICE_ID_CHICONY_TACTICAL_PAD) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index a577121..c5ae5f1 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -134,6 +134,7 @@ #define USB_VENDOR_ID_CANDO 0x2087 #define USB_DEVICE_ID_CANDO_MULTI_TOUCH 0x0a01 #define USB_DEVICE_ID_CANDO_MULTI_TOUCH_11_6 0x0b03 +#define USB_DEVICE_ID_CANDO_MULTI_TOUCH_15_6 0x0f01 #define USB_VENDOR_ID_CH 0x068e #define USB_DEVICE_ID_CH_PRO_PEDALS 0x00f2 -- cgit v0.10.2 From 5bcd757f93cc713cf61bbeefceda7539d9afca55 Mon Sep 17 00:00:00 2001 From: Matthew Garrett Date: Mon, 4 Oct 2010 14:59:31 -0400 Subject: x86/amd-iommu: Reenable AMD IOMMU if it's mysteriously vanished over suspend AMD's reference BIOS code had a bug that could result in the firmware failing to reenable the iommu on resume. It transpires that this causes certain less than desirable behaviour when it comes to PCI accesses, to whit them ending up somewhere near Bristol when the more desirable outcome was Edinburgh. Sadness ensues, perhaps along with filesystem corruption. Let's make sure that it gets turned back on, and that we restore its configuration so decisions it makes bear some resemblance to those made by reasonable people rather than crack-addled lemurs who spent all your DMA on Thunderbird. Signed-off-by: Matthew Garrett Signed-off-by: Joerg Roedel diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index 0861618..bdd20c8 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -416,13 +416,22 @@ struct amd_iommu { struct dma_ops_domain *default_dom; /* - * This array is required to work around a potential BIOS bug. - * The BIOS may miss to restore parts of the PCI configuration - * space when the system resumes from S3. The result is that the - * IOMMU does not execute commands anymore which leads to system - * failure. + * We can't rely on the BIOS to restore all values on reinit, so we + * need to stash them */ - u32 cache_cfg[4]; + + /* The iommu BAR */ + u32 stored_addr_lo; + u32 stored_addr_hi; + + /* + * Each iommu has 6 l1s, each of which is documented as having 0x12 + * registers + */ + u32 stored_l1[6][0x12]; + + /* The l2 indirect registers */ + u32 stored_l2[0x83]; }; /* diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 5a170cb..44710d8 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -194,6 +194,39 @@ static inline unsigned long tbl_size(int entry_size) return 1UL << shift; } +/* Access to l1 and l2 indexed register spaces */ + +static u32 iommu_read_l1(struct amd_iommu *iommu, u16 l1, u8 address) +{ + u32 val; + + pci_write_config_dword(iommu->dev, 0xf8, (address | l1 << 16)); + pci_read_config_dword(iommu->dev, 0xfc, &val); + return val; +} + +static void iommu_write_l1(struct amd_iommu *iommu, u16 l1, u8 address, u32 val) +{ + pci_write_config_dword(iommu->dev, 0xf8, (address | l1 << 16 | 1 << 31)); + pci_write_config_dword(iommu->dev, 0xfc, val); + pci_write_config_dword(iommu->dev, 0xf8, (address | l1 << 16)); +} + +static u32 iommu_read_l2(struct amd_iommu *iommu, u8 address) +{ + u32 val; + + pci_write_config_dword(iommu->dev, 0xf0, address); + pci_read_config_dword(iommu->dev, 0xf4, &val); + return val; +} + +static void iommu_write_l2(struct amd_iommu *iommu, u8 address, u32 val) +{ + pci_write_config_dword(iommu->dev, 0xf0, (address | 1 << 8)); + pci_write_config_dword(iommu->dev, 0xf4, val); +} + /**************************************************************************** * * AMD IOMMU MMIO register space handling functions @@ -619,6 +652,7 @@ static void __init init_iommu_from_pci(struct amd_iommu *iommu) { int cap_ptr = iommu->cap_ptr; u32 range, misc; + int i, j; pci_read_config_dword(iommu->dev, cap_ptr + MMIO_CAP_HDR_OFFSET, &iommu->cap); @@ -633,12 +667,29 @@ static void __init init_iommu_from_pci(struct amd_iommu *iommu) MMIO_GET_LD(range)); iommu->evt_msi_num = MMIO_MSI_NUM(misc); - if (is_rd890_iommu(iommu->dev)) { - pci_read_config_dword(iommu->dev, 0xf0, &iommu->cache_cfg[0]); - pci_read_config_dword(iommu->dev, 0xf4, &iommu->cache_cfg[1]); - pci_read_config_dword(iommu->dev, 0xf8, &iommu->cache_cfg[2]); - pci_read_config_dword(iommu->dev, 0xfc, &iommu->cache_cfg[3]); - } + if (!is_rd890_iommu(iommu->dev)) + return; + + /* + * Some rd890 systems may not be fully reconfigured by the BIOS, so + * it's necessary for us to store this information so it can be + * reprogrammed on resume + */ + + pci_read_config_dword(iommu->dev, iommu->cap_ptr + 4, + &iommu->stored_addr_lo); + pci_read_config_dword(iommu->dev, iommu->cap_ptr + 8, + &iommu->stored_addr_hi); + + /* Low bit locks writes to configuration space */ + iommu->stored_addr_lo &= ~1; + + for (i = 0; i < 6; i++) + for (j = 0; j < 0x12; j++) + iommu->stored_l1[i][j] = iommu_read_l1(iommu, i, j); + + for (i = 0; i < 0x83; i++) + iommu->stored_l2[i] = iommu_read_l2(iommu, i); } /* @@ -1127,14 +1178,53 @@ static void iommu_init_flags(struct amd_iommu *iommu) iommu_feature_enable(iommu, CONTROL_COHERENT_EN); } -static void iommu_apply_quirks(struct amd_iommu *iommu) +static void iommu_apply_resume_quirks(struct amd_iommu *iommu) { - if (is_rd890_iommu(iommu->dev)) { - pci_write_config_dword(iommu->dev, 0xf0, iommu->cache_cfg[0]); - pci_write_config_dword(iommu->dev, 0xf4, iommu->cache_cfg[1]); - pci_write_config_dword(iommu->dev, 0xf8, iommu->cache_cfg[2]); - pci_write_config_dword(iommu->dev, 0xfc, iommu->cache_cfg[3]); - } + int i, j; + u32 ioc_feature_control; + struct pci_dev *pdev = NULL; + + /* RD890 BIOSes may not have completely reconfigured the iommu */ + if (!is_rd890_iommu(iommu->dev)) + return; + + /* + * First, we need to ensure that the iommu is enabled. This is + * controlled by a register in the northbridge + */ + pdev = pci_get_bus_and_slot(iommu->dev->bus->number, PCI_DEVFN(0, 0)); + + if (!pdev) + return; + + /* Select Northbridge indirect register 0x75 and enable writing */ + pci_write_config_dword(pdev, 0x60, 0x75 | (1 << 7)); + pci_read_config_dword(pdev, 0x64, &ioc_feature_control); + + /* Enable the iommu */ + if (!(ioc_feature_control & 0x1)) + pci_write_config_dword(pdev, 0x64, ioc_feature_control | 1); + + pci_dev_put(pdev); + + /* Restore the iommu BAR */ + pci_write_config_dword(iommu->dev, iommu->cap_ptr + 4, + iommu->stored_addr_lo); + pci_write_config_dword(iommu->dev, iommu->cap_ptr + 8, + iommu->stored_addr_hi); + + /* Restore the l1 indirect regs for each of the 6 l1s */ + for (i = 0; i < 6; i++) + for (j = 0; j < 0x12; j++) + iommu_write_l1(iommu, i, j, iommu->stored_l1[i][j]); + + /* Restore the l2 indirect regs */ + for (i = 0; i < 0x83; i++) + iommu_write_l2(iommu, i, iommu->stored_l2[i]); + + /* Lock PCI setup registers */ + pci_write_config_dword(iommu->dev, iommu->cap_ptr + 4, + iommu->stored_addr_lo | 1); } /* @@ -1147,7 +1237,6 @@ static void enable_iommus(void) for_each_iommu(iommu) { iommu_disable(iommu); - iommu_apply_quirks(iommu); iommu_init_flags(iommu); iommu_set_device_table(iommu); iommu_enable_command_buffer(iommu); @@ -1173,6 +1262,11 @@ static void disable_iommus(void) static int amd_iommu_resume(struct sys_device *dev) { + struct amd_iommu *iommu; + + for_each_iommu(iommu) + iommu_apply_resume_quirks(iommu); + /* re-load the hardware */ enable_iommus(); -- cgit v0.10.2 From 5d0d71569e671239ae0d905ced9b65cd843f99ee Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 13 Oct 2010 11:13:21 +0200 Subject: x86/amd-iommu: Update copyright headers This patch updates the copyright headers in all source files of the AMD IOMMU driver. Signed-off-by: Joerg Roedel diff --git a/arch/x86/include/asm/amd_iommu.h b/arch/x86/include/asm/amd_iommu.h index 5af2982..f16a2ca 100644 --- a/arch/x86/include/asm/amd_iommu.h +++ b/arch/x86/include/asm/amd_iommu.h @@ -1,5 +1,5 @@ /* - * Copyright (C) 2007-2009 Advanced Micro Devices, Inc. + * Copyright (C) 2007-2010 Advanced Micro Devices, Inc. * Author: Joerg Roedel * Leo Duran * diff --git a/arch/x86/include/asm/amd_iommu_proto.h b/arch/x86/include/asm/amd_iommu_proto.h index cb03037..916bc81 100644 --- a/arch/x86/include/asm/amd_iommu_proto.h +++ b/arch/x86/include/asm/amd_iommu_proto.h @@ -1,5 +1,5 @@ /* - * Copyright (C) 2009 Advanced Micro Devices, Inc. + * Copyright (C) 2009-2010 Advanced Micro Devices, Inc. * Author: Joerg Roedel * * This program is free software; you can redistribute it and/or modify it diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index bdd20c8..e3509fc 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -1,5 +1,5 @@ /* - * Copyright (C) 2007-2009 Advanced Micro Devices, Inc. + * Copyright (C) 2007-2010 Advanced Micro Devices, Inc. * Author: Joerg Roedel * Leo Duran * diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 679b645..d2fdb08 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2007-2009 Advanced Micro Devices, Inc. + * Copyright (C) 2007-2010 Advanced Micro Devices, Inc. * Author: Joerg Roedel * Leo Duran * diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 44710d8..3cb482e 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2007-2009 Advanced Micro Devices, Inc. + * Copyright (C) 2007-2010 Advanced Micro Devices, Inc. * Author: Joerg Roedel * Leo Duran * -- cgit v0.10.2 From 6fcc040f02d281c7e9563127358a77ce2bbfe284 Mon Sep 17 00:00:00 2001 From: Greg Ungerer Date: Mon, 11 Oct 2010 21:03:05 +0000 Subject: net: allow FEC driver to use fixed PHY support At least one board using the FEC driver does not have a conventional PHY attached to it, it is directly connected to a somewhat simple ethernet switch (the board is the SnapGear/LITE, and the attached 4-port ethernet switch is a RealTek RTL8305). This switch does not present the usual register interface of a PHY, it presents nothing. So a PHY scan will find nothing - it finds ID's of 0 for each PHY on the attached MII bus. After the FEC driver was changed to use phylib for supporting PHYs it no longer works on this particular board/switch setup. Add code support to use a fixed phy if no PHY is found on the MII bus. This is based on the way the cpmac.c driver solved this same problem. Signed-off-by: Greg Ungerer Signed-off-by: David S. Miller diff --git a/drivers/net/fec.c b/drivers/net/fec.c index e83f67d..cce32d4 100644 --- a/drivers/net/fec.c +++ b/drivers/net/fec.c @@ -678,24 +678,37 @@ static int fec_enet_mii_probe(struct net_device *dev) { struct fec_enet_private *fep = netdev_priv(dev); struct phy_device *phy_dev = NULL; - int ret; + char mdio_bus_id[MII_BUS_ID_SIZE]; + char phy_name[MII_BUS_ID_SIZE + 3]; + int phy_id; fep->phy_dev = NULL; - /* find the first phy */ - phy_dev = phy_find_first(fep->mii_bus); - if (!phy_dev) { - printk(KERN_ERR "%s: no PHY found\n", dev->name); - return -ENODEV; + /* check for attached phy */ + for (phy_id = 0; (phy_id < PHY_MAX_ADDR); phy_id++) { + if ((fep->mii_bus->phy_mask & (1 << phy_id))) + continue; + if (fep->mii_bus->phy_map[phy_id] == NULL) + continue; + if (fep->mii_bus->phy_map[phy_id]->phy_id == 0) + continue; + strncpy(mdio_bus_id, fep->mii_bus->id, MII_BUS_ID_SIZE); + break; } - /* attach the mac to the phy */ - ret = phy_connect_direct(dev, phy_dev, - &fec_enet_adjust_link, 0, - PHY_INTERFACE_MODE_MII); - if (ret) { - printk(KERN_ERR "%s: Could not attach to PHY\n", dev->name); - return ret; + if (phy_id >= PHY_MAX_ADDR) { + printk(KERN_INFO "%s: no PHY, assuming direct connection " + "to switch\n", dev->name); + strncpy(mdio_bus_id, "0", MII_BUS_ID_SIZE); + phy_id = 0; + } + + snprintf(phy_name, MII_BUS_ID_SIZE, PHY_ID_FMT, mdio_bus_id, phy_id); + phy_dev = phy_connect(dev, phy_name, &fec_enet_adjust_link, 0, + PHY_INTERFACE_MODE_MII); + if (IS_ERR(phy_dev)) { + printk(KERN_ERR "%s: could not attach to PHY\n", dev->name); + return PTR_ERR(phy_dev); } /* mask with MAC supported features */ @@ -738,7 +751,7 @@ static int fec_enet_mii_init(struct platform_device *pdev) fep->mii_bus->read = fec_enet_mdio_read; fep->mii_bus->write = fec_enet_mdio_write; fep->mii_bus->reset = fec_enet_mdio_reset; - snprintf(fep->mii_bus->id, MII_BUS_ID_SIZE, "%x", pdev->id); + snprintf(fep->mii_bus->id, MII_BUS_ID_SIZE, "%x", pdev->id + 1); fep->mii_bus->priv = fep; fep->mii_bus->parent = &pdev->dev; -- cgit v0.10.2 From 447b1d43ded08c11f4f3d344b4e07e6dcddbef95 Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Wed, 13 Oct 2010 19:10:42 +0100 Subject: x86, olpc: Register XO-1 platform devices The upcoming XO-1 rfkill driver (for drivers/platform/x86) will register itself with the name "xo1-rfkill", and the already-merged XO-1 poweroff code uses name "olpc-xo1" Add the necessary mechanics so that these devices are properly initialized on XO-1 laptops. Signed-off-by: Daniel Drake LKML-Reference: <20101013181042.90C8F9D401B@zog.reactivated.net> Cc: Matthew Garrett Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/olpc.c b/arch/x86/kernel/olpc.c index 7472613..edaf3fe 100644 --- a/arch/x86/kernel/olpc.c +++ b/arch/x86/kernel/olpc.c @@ -17,6 +17,7 @@ #include #include #include +#include #include #include @@ -223,8 +224,25 @@ static bool __init platform_detect(void) return true; } +static int __init add_xo1_platform_devices(void) +{ + struct platform_device *pdev; + + pdev = platform_device_register_simple("xo1-rfkill", -1, NULL, 0); + if (IS_ERR(pdev)) + return PTR_ERR(pdev); + + pdev = platform_device_register_simple("olpc-xo1", -1, NULL, 0); + if (IS_ERR(pdev)) + return PTR_ERR(pdev); + + return 0; +} + static int __init olpc_init(void) { + int r = 0; + if (!olpc_ofw_present() || !platform_detect()) return 0; @@ -251,6 +269,12 @@ static int __init olpc_init(void) olpc_platform_info.boardrev >> 4, olpc_platform_info.ecver); + if (olpc_platform_info.boardrev < olpc_board_pre(0xd0)) { /* XO-1 */ + r = add_xo1_platform_devices(); + if (r) + return r; + } + return 0; } -- cgit v0.10.2 From b1e86db1de2e8bc2be9fb94fae3451c2a776e8c1 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 13 Oct 2010 14:46:17 -0400 Subject: nfsd: fix BUG at fs/nfsd/nfsfh.h:199 on unlink As of commit 43a9aa64a2f4330a9cb59aaf5c5636566bce067c "NFSD: Fill in WCC data for REMOVE, RMDIR, MKNOD, and MKDIR", we sometimes call fh_unlock on a filehandle that isn't fully initialized. We should fix up the callers, but as a quick fix it is also sufficient just to remove this assertion. Reported-by: Marius Tolzmann Cc: Chuck Lever Signed-off-by: J. Bruce Fields diff --git a/fs/nfsd/nfsfh.h b/fs/nfsd/nfsfh.h index cdfb8c6..c16f8d8 100644 --- a/fs/nfsd/nfsfh.h +++ b/fs/nfsd/nfsfh.h @@ -196,8 +196,6 @@ fh_lock(struct svc_fh *fhp) static inline void fh_unlock(struct svc_fh *fhp) { - BUG_ON(!fhp->fh_dentry); - if (fhp->fh_locked) { fill_post_wcc(fhp); mutex_unlock(&fhp->fh_dentry->d_inode->i_mutex); -- cgit v0.10.2 From 71085ce8285dc5f3011694f6ba7259201135c6d6 Mon Sep 17 00:00:00 2001 From: Breno Leitao Date: Thu, 7 Oct 2010 13:17:33 +0000 Subject: ehea: Fix a checksum issue on the receive path Currently we set all skbs with CHECKSUM_UNNECESSARY, even those whose protocol we don't know. This patch just add the CHECKSUM_COMPLETE tag for non TCP/UDP packets. Reported-by: Eric Dumazet Signed-off-by: Breno Leitao Signed-off-by: Jay Vosburgh Signed-off-by: David S. Miller diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c index a333b42..6372610 100644 --- a/drivers/net/ehea/ehea_main.c +++ b/drivers/net/ehea/ehea_main.c @@ -533,8 +533,15 @@ static inline void ehea_fill_skb(struct net_device *dev, int length = cqe->num_bytes_transfered - 4; /*remove CRC */ skb_put(skb, length); - skb->ip_summed = CHECKSUM_UNNECESSARY; skb->protocol = eth_type_trans(skb, dev); + + /* The packet was not an IPV4 packet so a complemented checksum was + calculated. The value is found in the Internet Checksum field. */ + if (cqe->status & EHEA_CQE_BLIND_CKSUM) { + skb->ip_summed = CHECKSUM_COMPLETE; + skb->csum = csum_unfold(~cqe->inet_checksum_value); + } else + skb->ip_summed = CHECKSUM_UNNECESSARY; } static inline struct sk_buff *get_skb_by_index(struct sk_buff **skb_array, diff --git a/drivers/net/ehea/ehea_qmr.h b/drivers/net/ehea/ehea_qmr.h index f608a6c..3810473 100644 --- a/drivers/net/ehea/ehea_qmr.h +++ b/drivers/net/ehea/ehea_qmr.h @@ -150,6 +150,7 @@ struct ehea_rwqe { #define EHEA_CQE_TYPE_RQ 0x60 #define EHEA_CQE_STAT_ERR_MASK 0x700F #define EHEA_CQE_STAT_FAT_ERR_MASK 0xF +#define EHEA_CQE_BLIND_CKSUM 0x8000 #define EHEA_CQE_STAT_ERR_TCP 0x4000 #define EHEA_CQE_STAT_ERR_IP 0x2000 #define EHEA_CQE_STAT_ERR_CRC 0x1000 -- cgit v0.10.2 From c50a898fd4e736623ee175920db056194e0bb483 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Wed, 13 Oct 2010 15:43:10 -0700 Subject: ioat2: fix performance regression Commit 0793448 "DMAENGINE: generic channel status v2" changed the interface for how dma channel progress is retrieved. It inadvertently exported an internal helper function ioat_tx_status() instead of ioat_dma_tx_status(). The latter polls the hardware to get the latest completion state, while the helper just evaluates the current state without touching hardware. The effect is that we end up waiting for completion timeouts or descriptor allocation errors before the completion state is updated. iperf (before fix): [SUM] 0.0-41.3 sec 364 MBytes 73.9 Mbits/sec iperf (after fix): [SUM] 0.0- 4.5 sec 499 MBytes 940 Mbits/sec This is a regression starting with 2.6.35. Cc: Cc: Dave Jiang Cc: Jesse Brandeburg Cc: Linus Walleij Cc: Maciej Sosnowski Reported-by: Richard Scobie Signed-off-by: Dan Williams diff --git a/drivers/dma/ioat/dma_v2.c b/drivers/dma/ioat/dma_v2.c index 216f9d3..effd140 100644 --- a/drivers/dma/ioat/dma_v2.c +++ b/drivers/dma/ioat/dma_v2.c @@ -879,7 +879,7 @@ int __devinit ioat2_dma_probe(struct ioatdma_device *device, int dca) dma->device_issue_pending = ioat2_issue_pending; dma->device_alloc_chan_resources = ioat2_alloc_chan_resources; dma->device_free_chan_resources = ioat2_free_chan_resources; - dma->device_tx_status = ioat_tx_status; + dma->device_tx_status = ioat_dma_tx_status; err = ioat_probe(device); if (err) -- cgit v0.10.2 From d7acb92fea932ad2e7846480aeacddc2c03c8485 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Wed, 13 Oct 2010 16:00:29 -0700 Subject: x86-64, asm: If the assembler supports fxsave64, use it Kbuild allows for us to probe for the existence of specific constructs in the assembler, use them to find out if we can use fxsave64 and permit the compiler to generate better code. Signed-off-by: H. Peter Anvin diff --git a/arch/x86/Makefile b/arch/x86/Makefile index 8aa1b59..40668a9 100644 --- a/arch/x86/Makefile +++ b/arch/x86/Makefile @@ -96,8 +96,12 @@ cfi := $(call as-instr,.cfi_startproc\n.cfi_rel_offset $(sp-y)$(comma)0\n.cfi_en # is .cfi_signal_frame supported too? cfi-sigframe := $(call as-instr,.cfi_startproc\n.cfi_signal_frame\n.cfi_endproc,-DCONFIG_AS_CFI_SIGNAL_FRAME=1) cfi-sections := $(call as-instr,.cfi_sections .debug_frame,-DCONFIG_AS_CFI_SECTIONS=1) -KBUILD_AFLAGS += $(cfi) $(cfi-sigframe) $(cfi-sections) -KBUILD_CFLAGS += $(cfi) $(cfi-sigframe) $(cfi-sections) + +# does binutils support specific instructions? +asinstr := $(call as-instr,fxsaveq (%rax),-DCONFIG_AS_FXSAVEQ=1) + +KBUILD_AFLAGS += $(cfi) $(cfi-sigframe) $(cfi-sections) $(asinstr) +KBUILD_CFLAGS += $(cfi) $(cfi-sigframe) $(cfi-sections) $(asinstr) LDFLAGS := -m elf_$(UTS_MACHINE) diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index a73a8d5..70f105b 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -175,7 +175,7 @@ static inline void fpu_fxsave(struct fpu *fpu) uses any extended registers for addressing, a second REX prefix will be generated (to the assembler, rex64 followed by semicolon is a separate instruction), and hence the 64-bitness is lost. */ -#if 0 +#ifdef CONFIG_AS_FXSAVEQ /* Using "fxsaveq %0" would be the ideal choice, but is only supported starting with gas 2.16. */ __asm__ __volatile__("fxsaveq %0" -- cgit v0.10.2 From fb62db2ba943b1683f1d7181bb2988fce4c60870 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Wed, 13 Oct 2010 11:02:34 -0700 Subject: futex: Fix kernel-doc notation & typos Convert futex_requeue() function parameters to use @name kernel-doc notation and add @fshared & @cmpval to prevent kernel-doc warnings. Add @list to struct futex_q. Fix a few typos. Signed-off-by: Randy Dunlap Acked-by: Rusty Russell LKML-Reference: <20101013110234.89b06043.randy.dunlap@oracle.com> Signed-off-by: Ingo Molnar diff --git a/kernel/futex.c b/kernel/futex.c index 92a31d4..9b9fda7 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -91,6 +91,7 @@ struct futex_pi_state { /** * struct futex_q - The hashed futex queue entry, one per waiting task + * @list: priority-sorted list of tasks waiting on this futex * @task: the task waiting on the futex * @lock_ptr: the hash bucket lock * @key: the key the futex is hashed on @@ -104,7 +105,7 @@ struct futex_pi_state { * * A futex_q has a woken state, just like tasks have TASK_RUNNING. * It is considered woken when plist_node_empty(&q->list) || q->lock_ptr == 0. - * The order of wakup is always to make the first condition true, then + * The order of wakeup is always to make the first condition true, then * the second. * * PI futexes are typically woken before they are removed from the hash list via @@ -295,7 +296,7 @@ void put_futex_key(int fshared, union futex_key *key) * Slow path to fixup the fault we just took in the atomic write * access to @uaddr. * - * We have no generic implementation of a non destructive write to the + * We have no generic implementation of a non-destructive write to the * user address. We know that we faulted in the atomic pagefault * disabled section so we can as well avoid the #PF overhead by * calling get_user_pages() right away. @@ -515,7 +516,7 @@ lookup_pi_state(u32 uval, struct futex_hash_bucket *hb, */ pi_state = this->pi_state; /* - * Userspace might have messed up non PI and PI futexes + * Userspace might have messed up non-PI and PI futexes */ if (unlikely(!pi_state)) return -EINVAL; @@ -736,8 +737,8 @@ static void wake_futex(struct futex_q *q) /* * We set q->lock_ptr = NULL _before_ we wake up the task. If - * a non futex wake up happens on another CPU then the task - * might exit and p would dereference a non existing task + * a non-futex wake up happens on another CPU then the task + * might exit and p would dereference a non-existing task * struct. Prevent this by holding a reference on p across the * wake up. */ @@ -1131,11 +1132,13 @@ static int futex_proxy_trylock_atomic(u32 __user *pifutex, /** * futex_requeue() - Requeue waiters from uaddr1 to uaddr2 - * uaddr1: source futex user address - * uaddr2: target futex user address - * nr_wake: number of waiters to wake (must be 1 for requeue_pi) - * nr_requeue: number of waiters to requeue (0-INT_MAX) - * requeue_pi: if we are attempting to requeue from a non-pi futex to a + * @uaddr1: source futex user address + * @fshared: 0 for a PROCESS_PRIVATE futex, 1 for PROCESS_SHARED + * @uaddr2: target futex user address + * @nr_wake: number of waiters to wake (must be 1 for requeue_pi) + * @nr_requeue: number of waiters to requeue (0-INT_MAX) + * @cmpval: @uaddr1 expected value (or %NULL) + * @requeue_pi: if we are attempting to requeue from a non-pi futex to a * pi futex (pi to pi requeue is not supported) * * Requeue waiters on uaddr1 to uaddr2. In the requeue_pi case, try to acquire @@ -2651,7 +2654,7 @@ static int __init futex_init(void) * of the complex code paths. Also we want to prevent * registration of robust lists in that case. NULL is * guaranteed to fault and we get -EFAULT on functional - * implementation, the non functional ones will return + * implementation, the non-functional ones will return * -ENOSYS. */ curval = cmpxchg_futex_value_locked(NULL, 0, 0); -- cgit v0.10.2 From 864616ee6785d9fac7a2cd80c01a2da89579f2e4 Mon Sep 17 00:00:00 2001 From: Takuya Yoshikawa Date: Thu, 14 Oct 2010 16:09:13 +0900 Subject: sched: Comment updates: fix default latency and granularity numbers Targeted preemption latency and minimal preemption granularity for CPU-bound tasks have been changed. This patch updates the comments about these values. Signed-off-by: Takuya Yoshikawa Cc: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20101014160913.eb24fef4.yoshikawa.takuya@oss.ntt.co.jp> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 623e9ac..bf87192e 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -25,7 +25,7 @@ /* * Targeted preemption latency for CPU-bound tasks: - * (default: 5ms * (1 + ilog(ncpus)), units: nanoseconds) + * (default: 6ms * (1 + ilog(ncpus)), units: nanoseconds) * * NOTE: this latency value is not the same as the concept of * 'timeslice length' - timeslices in CFS are of variable length @@ -52,7 +52,7 @@ enum sched_tunable_scaling sysctl_sched_tunable_scaling /* * Minimal preemption granularity for CPU-bound tasks: - * (default: 2 msec * (1 + ilog(ncpus)), units: nanoseconds) + * (default: 0.75 msec * (1 + ilog(ncpus)), units: nanoseconds) */ unsigned int sysctl_sched_min_granularity = 750000ULL; unsigned int normalized_sysctl_sched_min_granularity = 750000ULL; -- cgit v0.10.2 From f13d4f979c518119bba5439dd2364d76d31dcd3f Mon Sep 17 00:00:00 2001 From: Salman Qazi Date: Tue, 12 Oct 2010 07:25:19 -0700 Subject: hrtimer: Preserve timer state in remove_hrtimer() The race is described as follows: CPU X CPU Y remove_hrtimer // state & QUEUED == 0 timer->state = CALLBACK unlock timer base timer->f(n) //very long hrtimer_start lock timer base remove_hrtimer // no effect hrtimer_enqueue timer->state = CALLBACK | QUEUED unlock timer base hrtimer_start lock timer base remove_hrtimer mode = INACTIVE // CALLBACK bit lost! switch_hrtimer_base CALLBACK bit not set: timer->base changes to a different CPU. lock this CPU's timer base The bug was introduced with commit ca109491f (hrtimer: removing all ur callback modes) in 2.6.29 [ tglx: Feed new state via local variable and add a comment. ] Signed-off-by: Salman Qazi Cc: akpm@linux-foundation.org Cc: Peter Zijlstra LKML-Reference: <20101012142351.8485.21823.stgit@dungbeetle.mtv.corp.google.com> Signed-off-by: Thomas Gleixner Cc: stable@kernel.org diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index 1decafb..72206cf 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -931,6 +931,7 @@ static inline int remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base) { if (hrtimer_is_queued(timer)) { + unsigned long state; int reprogram; /* @@ -944,8 +945,13 @@ remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base) debug_deactivate(timer); timer_stats_hrtimer_clear_start_info(timer); reprogram = base->cpu_base == &__get_cpu_var(hrtimer_bases); - __remove_hrtimer(timer, base, HRTIMER_STATE_INACTIVE, - reprogram); + /* + * We must preserve the CALLBACK state flag here, + * otherwise we could move the timer base in + * switch_hrtimer_base. + */ + state = timer->state & HRTIMER_STATE_CALLBACK; + __remove_hrtimer(timer, base, state, reprogram); return 1; } return 0; @@ -1231,6 +1237,9 @@ static void __run_hrtimer(struct hrtimer *timer, ktime_t *now) BUG_ON(timer->state != HRTIMER_STATE_CALLBACK); enqueue_hrtimer(timer, base); } + + WARN_ON_ONCE(!(timer->state & HRTIMER_STATE_CALLBACK)); + timer->state &= ~HRTIMER_STATE_CALLBACK; } -- cgit v0.10.2 From ee52716245877b821f5ddbb3ace85b73084fb450 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Thu, 14 Oct 2010 09:53:37 -0400 Subject: hfsplus: fix oops on mount with corrupted btree extent records A particular fsfuzzer run caused an hfs file system to crash on mount. This is due to a corrupted MDB extent record causing a miscalculation of HFSPLUS_I(inode)->first_blocks for the extent tree. If the extent records are zereod out, then it won't trigger the first_blocks special case and instead falls through to the extent code, which we're in the middle of initializing. This patch catches the 0 size extent records, reports the corruption, and fails the mount. [hch: ported of commit 47f365eb575735c6b2edf5d08e0d16d26a9c23bd from hfs] Reported-by: Ramon de Carvalho Valle Signed-off-by: Jeff Mahoney Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/btree.c b/fs/hfsplus/btree.c index 8306479..d4bd864 100644 --- a/fs/hfsplus/btree.c +++ b/fs/hfsplus/btree.c @@ -39,10 +39,16 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) goto free_tree; tree->inode = inode; + if (!HFSPLUS_I(tree->inode)->first_blocks) { + printk(KERN_ERR + "hfs: invalid btree extent records (0 size).\n"); + goto free_inode; + } + mapping = tree->inode->i_mapping; page = read_mapping_page(mapping, 0, NULL); if (IS_ERR(page)) - goto free_tree; + goto free_inode; /* Load the header */ head = (struct hfs_btree_header_rec *)(kmap(page) + sizeof(struct hfs_bnode_desc)); @@ -89,8 +95,9 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) fail_page: tree->inode->i_mapping->a_ops = &hfsplus_aops; page_cache_release(page); - free_tree: + free_inode: iput(tree->inode); + free_tree: kfree(tree); return NULL; } -- cgit v0.10.2 From b6b41424f0ec28e9a167fa29b003327860b4b71b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 14 Oct 2010 09:53:42 -0400 Subject: hfsplus: hfs_bnode_find() can fail, resulting in hfs_bnode_split() breakage oops and fs corruption; the latter can happen even on valid fs in case of oom. [hch: port of commit 3d10a15d6919488204bdb264050d156ced20d9aa from hfs] Signed-off-by: Al Viro Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/brec.c b/fs/hfsplus/brec.c index c88e5d7..fa90364 100644 --- a/fs/hfsplus/brec.c +++ b/fs/hfsplus/brec.c @@ -216,7 +216,7 @@ skip: static struct hfs_bnode *hfs_bnode_split(struct hfs_find_data *fd) { struct hfs_btree *tree; - struct hfs_bnode *node, *new_node; + struct hfs_bnode *node, *new_node, *next_node; struct hfs_bnode_desc node_desc; int num_recs, new_rec_off, new_off, old_rec_off; int data_start, data_end, size; @@ -235,6 +235,17 @@ static struct hfs_bnode *hfs_bnode_split(struct hfs_find_data *fd) new_node->type = node->type; new_node->height = node->height; + if (node->next) + next_node = hfs_bnode_find(tree, node->next); + else + next_node = NULL; + + if (IS_ERR(next_node)) { + hfs_bnode_put(node); + hfs_bnode_put(new_node); + return next_node; + } + size = tree->node_size / 2 - node->num_recs * 2 - 14; old_rec_off = tree->node_size - 4; num_recs = 1; @@ -248,6 +259,8 @@ static struct hfs_bnode *hfs_bnode_split(struct hfs_find_data *fd) /* panic? */ hfs_bnode_put(node); hfs_bnode_put(new_node); + if (next_node) + hfs_bnode_put(next_node); return ERR_PTR(-ENOSPC); } @@ -302,8 +315,7 @@ static struct hfs_bnode *hfs_bnode_split(struct hfs_find_data *fd) hfs_bnode_write(node, &node_desc, 0, sizeof(node_desc)); /* update next bnode header */ - if (new_node->next) { - struct hfs_bnode *next_node = hfs_bnode_find(tree, new_node->next); + if (next_node) { next_node->prev = new_node->this; hfs_bnode_read(next_node, &node_desc, 0, sizeof(node_desc)); node_desc.prev = cpu_to_be32(next_node->prev); -- cgit v0.10.2 From 9250f925972d03ccc0c0a4dd4e9b794d2ef6d52b Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 14 Oct 2010 09:53:48 -0400 Subject: hfsplus: handle more on-disk corruptions without oopsing hfs seems prone to bad things when it encounters on disk corruption. Many values are read from disk, and used as lengths to memcpy, as an example. This patch fixes up several of these problematic cases. o sanity check the on-disk maximum key lengths on mount (these are set to a defined value at mkfs time and shouldn't differ) o check on-disk node keylens against the maximum key length for each tree o fix hfs_btree_open so that going out via free_tree: doesn't wind up in hfs_releasepage, which wants to follow the very pointer we were trying to set up: HFS_SB(sb)->cat_tree = hfs_btree_open() . failure gets to hfs_releasepage and tries to follow HFS_SB(sb)->cat_tree Tested with the fsfuzzer; it survives more than it used to. [hch: ported of commit cf0594625083111ae522496dc1c256f7476939c2 from hfs] [hch: added the fixes from 5581d018ed3493d226e7a4d645d9c8a5af6c36b] Signed-off-by: Eric Sandeen Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/bfind.c b/fs/hfsplus/bfind.c index 68c7983..d182438 100644 --- a/fs/hfsplus/bfind.c +++ b/fs/hfsplus/bfind.c @@ -52,6 +52,10 @@ int __hfs_brec_find(struct hfs_bnode *bnode, struct hfs_find_data *fd) rec = (e + b) / 2; len = hfs_brec_lenoff(bnode, rec, &off); keylen = hfs_brec_keylen(bnode, rec); + if (keylen == 0) { + res = -EINVAL; + goto fail; + } hfs_bnode_read(bnode, fd->key, off, keylen); cmpval = bnode->tree->keycmp(fd->key, fd->search_key); if (!cmpval) { @@ -67,6 +71,10 @@ int __hfs_brec_find(struct hfs_bnode *bnode, struct hfs_find_data *fd) if (rec != e && e >= 0) { len = hfs_brec_lenoff(bnode, e, &off); keylen = hfs_brec_keylen(bnode, e); + if (keylen == 0) { + res = -EINVAL; + goto fail; + } hfs_bnode_read(bnode, fd->key, off, keylen); } done: @@ -75,6 +83,7 @@ done: fd->keylength = keylen; fd->entryoffset = off + keylen; fd->entrylength = len - keylen; +fail: return res; } @@ -198,6 +207,10 @@ int hfs_brec_goto(struct hfs_find_data *fd, int cnt) len = hfs_brec_lenoff(bnode, fd->record, &off); keylen = hfs_brec_keylen(bnode, fd->record); + if (keylen == 0) { + res = -EINVAL; + goto out; + } fd->keyoffset = off; fd->keylength = keylen; fd->entryoffset = off + keylen; diff --git a/fs/hfsplus/brec.c b/fs/hfsplus/brec.c index fa90364..790cd6a 100644 --- a/fs/hfsplus/brec.c +++ b/fs/hfsplus/brec.c @@ -42,10 +42,21 @@ u16 hfs_brec_keylen(struct hfs_bnode *node, u16 rec) recoff = hfs_bnode_read_u16(node, node->tree->node_size - (rec + 1) * 2); if (!recoff) return 0; - if (node->tree->attributes & HFS_TREE_BIGKEYS) + if (node->tree->attributes & HFS_TREE_BIGKEYS) { retval = hfs_bnode_read_u16(node, recoff) + 2; - else + if (retval > node->tree->max_key_len + 2) { + printk(KERN_ERR "hfs: keylen %d too large\n", + retval); + retval = 0; + } + } else { retval = (hfs_bnode_read_u8(node, recoff) | 1) + 1; + if (retval > node->tree->max_key_len + 1) { + printk(KERN_ERR "hfs: keylen %d too large\n", + retval); + retval = 0; + } + } } return retval; } diff --git a/fs/hfsplus/btree.c b/fs/hfsplus/btree.c index d4bd864..82caff2 100644 --- a/fs/hfsplus/btree.c +++ b/fs/hfsplus/btree.c @@ -63,10 +63,23 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) tree->max_key_len = be16_to_cpu(head->max_key_len); tree->depth = be16_to_cpu(head->depth); - /* Set the correct compare function */ - if (id == HFSPLUS_EXT_CNID) { + /* Verify the tree and set the correct compare function */ + switch (id) { + case HFSPLUS_EXT_CNID: + if (tree->max_key_len != HFSPLUS_EXT_KEYLEN - sizeof(u16)) { + printk(KERN_ERR "hfs: invalid extent max_key_len %d\n", + tree->max_key_len); + goto fail_page; + } tree->keycmp = hfsplus_ext_cmp_key; - } else if (id == HFSPLUS_CAT_CNID) { + break; + case HFSPLUS_CAT_CNID: + if (tree->max_key_len != HFSPLUS_CAT_KEYLEN - sizeof(u16)) { + printk(KERN_ERR "hfs: invalid catalog max_key_len %d\n", + tree->max_key_len); + goto fail_page; + } + if (test_bit(HFSPLUS_SB_HFSX, &HFSPLUS_SB(sb)->flags) && (head->key_type == HFSPLUS_KEY_BINARY)) tree->keycmp = hfsplus_cat_bin_cmp_key; @@ -74,7 +87,8 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) tree->keycmp = hfsplus_cat_case_cmp_key; set_bit(HFSPLUS_SB_CASEFOLD, &HFSPLUS_SB(sb)->flags); } - } else { + break; + default: printk(KERN_ERR "hfs: unknown B*Tree requested\n"); goto fail_page; } @@ -84,6 +98,7 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) goto fail_page; if (!tree->node_count) goto fail_page; + tree->node_size_shift = ffs(size) - 1; tree->pages_per_bnode = (tree->node_size + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; @@ -93,9 +108,9 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) return tree; fail_page: - tree->inode->i_mapping->a_ops = &hfsplus_aops; page_cache_release(page); free_inode: + tree->inode->i_mapping->a_ops = &hfsplus_aops; iput(tree->inode); free_tree: kfree(tree); diff --git a/fs/hfsplus/hfsplus_raw.h b/fs/hfsplus/hfsplus_raw.h index fe99fe8..6892899 100644 --- a/fs/hfsplus/hfsplus_raw.h +++ b/fs/hfsplus/hfsplus_raw.h @@ -200,6 +200,7 @@ struct hfsplus_cat_key { struct hfsplus_unistr name; } __packed; +#define HFSPLUS_CAT_KEYLEN (sizeof(struct hfsplus_cat_key)) /* Structs from hfs.h */ struct hfsp_point { @@ -323,7 +324,7 @@ struct hfsplus_ext_key { __be32 start_block; } __packed; -#define HFSPLUS_EXT_KEYLEN 12 +#define HFSPLUS_EXT_KEYLEN sizeof(struct hfsplus_ext_key) /* HFS+ generic BTree key */ typedef union { -- cgit v0.10.2 From 13571a6977f821fab7d9c3cc5f75da52b7732e40 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 14 Oct 2010 09:54:23 -0400 Subject: hfsplus: validate btree flags Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/brec.c b/fs/hfsplus/brec.c index 790cd6a..2f39d05 100644 --- a/fs/hfsplus/brec.c +++ b/fs/hfsplus/brec.c @@ -42,20 +42,12 @@ u16 hfs_brec_keylen(struct hfs_bnode *node, u16 rec) recoff = hfs_bnode_read_u16(node, node->tree->node_size - (rec + 1) * 2); if (!recoff) return 0; - if (node->tree->attributes & HFS_TREE_BIGKEYS) { - retval = hfs_bnode_read_u16(node, recoff) + 2; - if (retval > node->tree->max_key_len + 2) { - printk(KERN_ERR "hfs: keylen %d too large\n", - retval); - retval = 0; - } - } else { - retval = (hfs_bnode_read_u8(node, recoff) | 1) + 1; - if (retval > node->tree->max_key_len + 1) { - printk(KERN_ERR "hfs: keylen %d too large\n", - retval); - retval = 0; - } + + retval = hfs_bnode_read_u16(node, recoff) + 2; + if (retval > node->tree->max_key_len + 2) { + printk(KERN_ERR "hfs: keylen %d too large\n", + retval); + retval = 0; } } return retval; diff --git a/fs/hfsplus/btree.c b/fs/hfsplus/btree.c index 82caff2..22e4d4e 100644 --- a/fs/hfsplus/btree.c +++ b/fs/hfsplus/btree.c @@ -71,6 +71,11 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) tree->max_key_len); goto fail_page; } + if (tree->attributes & HFS_TREE_VARIDXKEYS) { + printk(KERN_ERR "hfs: invalid extent btree flag\n"); + goto fail_page; + } + tree->keycmp = hfsplus_ext_cmp_key; break; case HFSPLUS_CAT_CNID: @@ -79,6 +84,10 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) tree->max_key_len); goto fail_page; } + if (!(tree->attributes & HFS_TREE_VARIDXKEYS)) { + printk(KERN_ERR "hfs: invalid catalog btree flag\n"); + goto fail_page; + } if (test_bit(HFSPLUS_SB_HFSX, &HFSPLUS_SB(sb)->flags) && (head->key_type == HFSPLUS_KEY_BINARY)) @@ -93,6 +102,11 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id) goto fail_page; } + if (!(tree->attributes & HFS_TREE_BIGKEYS)) { + printk(KERN_ERR "hfs: invalid btree flag\n"); + goto fail_page; + } + size = tree->node_size; if (!is_power_of_2(size)) goto fail_page; -- cgit v0.10.2 From f6089ff87d309a8ddb7b0d4dd92a570f1b0f689b Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 14 Oct 2010 09:54:28 -0400 Subject: hfsplus: fix link corruption HFS implements hardlink by using indirect catalog entries that refer to a hidden directly. The link target is cached in the dev field in the HFS+ specific inode, which is also used for the device number for device files, and inside for passing the nlink value of the indirect node from hfsplus_cat_write_inode to a helper function. Now if we happen to write out the indirect node while hfsplus_link is creating the catalog entry we'll get a link pointing to the linkid of the current nlink value. This can easily be reproduced by a large enough loop of local git-clone operations. Stop abusing the dev field in the HFS+ inode for short term storage by refactoring the way the permission structure in the catalog entry is set up, and rename the dev field to linkid to avoid any confusion. While we're at it also prevent creating hard links to special files, as the HFS+ dev and linkid share the same space in the on-disk structure. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index 48979c4..9d1594b 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -134,7 +134,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i file->user_info.fdCreator = cpu_to_be32(HFSP_HFSPLUS_CREATOR); file->user_info.fdFlags = cpu_to_be16(0x100); file->create_date = HFSPLUS_I(sbi->hidden_dir)->create_date; - file->permissions.dev = cpu_to_be32(HFSPLUS_I(inode)->dev); + file->permissions.dev = cpu_to_be32(HFSPLUS_I(inode)->linkid); } return sizeof(*file); } diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index 33aab21..c05c877 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -102,7 +102,7 @@ again: if (IS_ERR(inode)) return ERR_CAST(inode); if (S_ISREG(inode->i_mode)) - HFSPLUS_I(inode)->dev = linkid; + HFSPLUS_I(inode)->linkid = linkid; out: d_add(dentry, inode); return NULL; @@ -252,6 +252,8 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, if (HFSPLUS_IS_RSRC(inode)) return -EPERM; + if (!S_ISREG(inode->i_mode)) + return -EPERM; mutex_lock(&sbi->vh_mutex); if (inode->i_ino == (u32)(unsigned long)src_dentry->d_fsdata) { @@ -268,7 +270,7 @@ static int hfsplus_link(struct dentry *src_dentry, struct inode *dst_dir, if (res != -EEXIST) goto out; } - HFSPLUS_I(inode)->dev = id; + HFSPLUS_I(inode)->linkid = id; cnid = sbi->next_cnid++; src_dentry->d_fsdata = (void *)(unsigned long)cnid; res = hfsplus_create_cat(cnid, src_dir, &src_dentry->d_name, inode); diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 5cda963..d92f590 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -178,7 +178,11 @@ struct hfsplus_inode_info { */ struct inode *rsrc_inode; __be32 create_date; - u32 dev; + + /* + * Protected by sbi->vh_mutex. + */ + u32 linkid; /* * Protected by i_mutex. @@ -427,6 +431,4 @@ static inline struct hfsplus_inode_info *HFSPLUS_I(struct inode *inode) #define hfsp_ut2mt(t) __hfsp_ut2mt((t).tv_sec) #define hfsp_now2mt() __hfsp_ut2mt(get_seconds()) -#define kdev_t_to_nr(x) (x) - #endif diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index a05b3af..746e0ee 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -267,7 +267,13 @@ static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) perms->mode = cpu_to_be16(inode->i_mode); perms->owner = cpu_to_be32(inode->i_uid); perms->group = cpu_to_be32(inode->i_gid); - perms->dev = cpu_to_be32(HFSPLUS_I(inode)->dev); + + if (S_ISREG(inode->i_mode)) + perms->dev = cpu_to_be32(inode->i_nlink); + else if (S_ISBLK(inode->i_mode) || S_ISCHR(inode->i_mode)) + perms->dev = cpu_to_be32(inode->i_rdev); + else + perms->dev = 0; } static int hfsplus_file_open(struct inode *inode, struct file *file) @@ -491,7 +497,7 @@ int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) type = hfs_bnode_read_u16(fd->bnode, fd->entryoffset); - HFSPLUS_I(inode)->dev = 0; + HFSPLUS_I(inode)->linkid = 0; if (type == HFSPLUS_FOLDER) { struct hfsplus_cat_folder *folder = &entry.folder; @@ -595,10 +601,6 @@ int hfsplus_cat_write_inode(struct inode *inode) hfs_bnode_read(fd.bnode, &entry, fd.entryoffset, sizeof(struct hfsplus_cat_file)); hfsplus_inode_write_fork(inode, &file->data_fork); - if (S_ISREG(inode->i_mode)) - HFSPLUS_I(inode)->dev = inode->i_nlink; - if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) - HFSPLUS_I(inode)->dev = kdev_t_to_nr(inode->i_rdev); hfsplus_set_perms(inode, &file->permissions); if ((file->permissions.rootflags | file->permissions.userflags) & HFSPLUS_FLG_IMMUTABLE) file->flags |= cpu_to_be16(HFSPLUS_FILE_LOCKED); -- cgit v0.10.2 From 722c55d13e7296cc62ed8a38f926a915ff32e4ea Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 14 Oct 2010 09:54:33 -0400 Subject: hfsplus: remove superflous rootflags field in hfsplus_inode_info The rootflags field in hfsplus_inode_info only caches the immutable and append-only flags in the VFS inode, so we can easily get rid of it. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index 9d1594b..bbcf5a2 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -77,7 +77,6 @@ static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) perms->rootflags |= HFSPLUS_FLG_APPEND; else perms->rootflags &= ~HFSPLUS_FLG_APPEND; - HFSPLUS_I(inode)->rootflags = perms->rootflags; HFSPLUS_I(inode)->userflags = perms->userflags; perms->mode = cpu_to_be16(inode->i_mode); perms->owner = cpu_to_be32(inode->i_uid); diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index d92f590..0d77844 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -188,7 +188,7 @@ struct hfsplus_inode_info { * Protected by i_mutex. */ sector_t fs_blocks; - u8 rootflags, userflags; /* BSD system and user file flags */ + u8 userflags; /* BSD user file flags */ struct list_head open_dir_list; loff_t phys_size; diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index 746e0ee..df57ea1 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -241,7 +241,6 @@ static void hfsplus_get_perms(struct inode *inode, struct hfsplus_perm *perms, i mode = S_IFREG | ((S_IRUGO|S_IWUGO) & ~(sbi->umask)); inode->i_mode = mode; - HFSPLUS_I(inode)->rootflags = perms->rootflags; HFSPLUS_I(inode)->userflags = perms->userflags; if (perms->rootflags & HFSPLUS_FLG_IMMUTABLE) inode->i_flags |= S_IMMUTABLE; diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index c9ac443..80eb5b3 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -26,9 +26,9 @@ static int hfsplus_ioctl_getflags(struct file *file, int __user *user_flags) struct hfsplus_inode_info *hip = HFSPLUS_I(inode); unsigned int flags = 0; - if (hip->rootflags & HFSPLUS_FLG_IMMUTABLE) + if (inode->i_flags & S_IMMUTABLE) flags |= FS_IMMUTABLE_FL; - if (hip->rootflags & HFSPLUS_FLG_APPEND) + if (inode->i_flags |= S_APPEND) flags |= FS_APPEND_FL; if (hip->userflags & HFSPLUS_FLG_NODUMP) flags |= FS_NODUMP_FL; @@ -59,8 +59,8 @@ static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) mutex_lock(&inode->i_mutex); - if (flags & (FS_IMMUTABLE_FL|FS_APPEND_FL) || - hip->rootflags & (HFSPLUS_FLG_IMMUTABLE|HFSPLUS_FLG_APPEND)) { + if ((flags & (FS_IMMUTABLE_FL|FS_APPEND_FL)) || + inode->i_flags & (S_IMMUTABLE|S_APPEND)) { if (!capable(CAP_LINUX_IMMUTABLE)) { err = -EPERM; goto out_unlock_inode; @@ -72,20 +72,17 @@ static int hfsplus_ioctl_setflags(struct file *file, int __user *user_flags) err = -EOPNOTSUPP; goto out_unlock_inode; } - if (flags & FS_IMMUTABLE_FL) { + + if (flags & FS_IMMUTABLE_FL) inode->i_flags |= S_IMMUTABLE; - hip->rootflags |= HFSPLUS_FLG_IMMUTABLE; - } else { + else inode->i_flags &= ~S_IMMUTABLE; - hip->rootflags &= ~HFSPLUS_FLG_IMMUTABLE; - } - if (flags & FS_APPEND_FL) { + + if (flags & FS_APPEND_FL) inode->i_flags |= S_APPEND; - hip->rootflags |= HFSPLUS_FLG_APPEND; - } else { + else inode->i_flags &= ~S_APPEND; - hip->rootflags &= ~HFSPLUS_FLG_APPEND; - } + if (flags & FS_NODUMP_FL) hip->userflags |= HFSPLUS_FLG_NODUMP; else -- cgit v0.10.2 From 90e616905a423126805186cb5754e10a704b30c8 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 14 Oct 2010 09:54:39 -0400 Subject: hfsplus: create correct initial catalog entries for device files Make sure the initial insertation of the catalog entry already contains the device number by calling init_special_inode early and setting writing out the dev field of the on-disk permission structure. The latter is facilitated by sharing the almost identical hfsplus_set_perms helpers between initial catalog entry creating and ->write_inode. Unless we crashed just after mknod this bug was harmless as the inode is marked dirty at the end of hfsplus_mknod, and hfsplus_write_inode will update the catalog entry to contain the correct value. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index bbcf5a2..8af45fc 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -67,7 +67,7 @@ static void hfsplus_cat_build_key_uni(hfsplus_btree_key *key, u32 parent, key->key_len = cpu_to_be16(6 + ustrlen); } -static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) +void hfsplus_cat_set_perms(struct inode *inode, struct hfsplus_perm *perms) { if (inode->i_flags & S_IMMUTABLE) perms->rootflags |= HFSPLUS_FLG_IMMUTABLE; @@ -77,10 +77,18 @@ static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) perms->rootflags |= HFSPLUS_FLG_APPEND; else perms->rootflags &= ~HFSPLUS_FLG_APPEND; - HFSPLUS_I(inode)->userflags = perms->userflags; + + perms->userflags = HFSPLUS_I(inode)->userflags; perms->mode = cpu_to_be16(inode->i_mode); perms->owner = cpu_to_be32(inode->i_uid); perms->group = cpu_to_be32(inode->i_gid); + + if (S_ISREG(inode->i_mode)) + perms->dev = cpu_to_be32(inode->i_nlink); + else if (S_ISBLK(inode->i_mode) || S_ISCHR(inode->i_mode)) + perms->dev = cpu_to_be32(inode->i_rdev); + else + perms->dev = 0; } static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct inode *inode) @@ -99,7 +107,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i folder->content_mod_date = folder->attribute_mod_date = folder->access_date = hfsp_now2mt(); - hfsplus_set_perms(inode, &folder->permissions); + hfsplus_cat_set_perms(inode, &folder->permissions); if (inode == sbi->hidden_dir) /* invisible and namelocked */ folder->user_info.frFlags = cpu_to_be16(0x5000); @@ -118,7 +126,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, u32 cnid, struct i file->attribute_mod_date = file->access_date = hfsp_now2mt(); if (cnid == inode->i_ino) { - hfsplus_set_perms(inode, &file->permissions); + hfsplus_cat_set_perms(inode, &file->permissions); if (S_ISLNK(inode->i_mode)) { file->user_info.fdType = cpu_to_be32(HFSP_SYMLINK_TYPE); file->user_info.fdCreator = cpu_to_be32(HFSP_SYMLINK_CREATOR); diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index c05c877..d236d85 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -418,6 +418,9 @@ static int hfsplus_mknod(struct inode *dir, struct dentry *dentry, if (!inode) goto out; + if (S_ISBLK(mode) || S_ISCHR(mode) || S_ISFIFO(mode) || S_ISSOCK(mode)) + init_special_inode(inode, mode, rdev); + res = hfsplus_create_cat(inode->i_ino, dir, &dentry->d_name, inode); if (res) { inode->i_nlink = 0; @@ -426,9 +429,6 @@ static int hfsplus_mknod(struct inode *dir, struct dentry *dentry, goto out; } - if (S_ISBLK(mode) || S_ISCHR(mode) || S_ISFIFO(mode) || S_ISSOCK(mode)) - init_special_inode(inode, mode, rdev); - hfsplus_instantiate(dentry, inode, inode->i_ino); mark_inode_dirty(inode); out: diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 0d77844..d77dfd2 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -326,6 +326,7 @@ int hfsplus_create_cat(u32, struct inode *, struct qstr *, struct inode *); int hfsplus_delete_cat(u32, struct inode *, struct qstr *); int hfsplus_rename_cat(u32, struct inode *, struct qstr *, struct inode *, struct qstr *); +void hfsplus_cat_set_perms(struct inode *inode, struct hfsplus_perm *perms); /* dir.c */ extern const struct inode_operations hfsplus_dir_inode_operations; diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index df57ea1..7844928 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -252,29 +252,6 @@ static void hfsplus_get_perms(struct inode *inode, struct hfsplus_perm *perms, i inode->i_flags &= ~S_APPEND; } -static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) -{ - if (inode->i_flags & S_IMMUTABLE) - perms->rootflags |= HFSPLUS_FLG_IMMUTABLE; - else - perms->rootflags &= ~HFSPLUS_FLG_IMMUTABLE; - if (inode->i_flags & S_APPEND) - perms->rootflags |= HFSPLUS_FLG_APPEND; - else - perms->rootflags &= ~HFSPLUS_FLG_APPEND; - perms->userflags = HFSPLUS_I(inode)->userflags; - perms->mode = cpu_to_be16(inode->i_mode); - perms->owner = cpu_to_be32(inode->i_uid); - perms->group = cpu_to_be32(inode->i_gid); - - if (S_ISREG(inode->i_mode)) - perms->dev = cpu_to_be32(inode->i_nlink); - else if (S_ISBLK(inode->i_mode) || S_ISCHR(inode->i_mode)) - perms->dev = cpu_to_be32(inode->i_rdev); - else - perms->dev = 0; -} - static int hfsplus_file_open(struct inode *inode, struct file *file) { if (HFSPLUS_IS_RSRC(inode)) @@ -578,7 +555,7 @@ int hfsplus_cat_write_inode(struct inode *inode) hfs_bnode_read(fd.bnode, &entry, fd.entryoffset, sizeof(struct hfsplus_cat_folder)); /* simple node checks? */ - hfsplus_set_perms(inode, &folder->permissions); + hfsplus_cat_set_perms(inode, &folder->permissions); folder->access_date = hfsp_ut2mt(inode->i_atime); folder->content_mod_date = hfsp_ut2mt(inode->i_mtime); folder->attribute_mod_date = hfsp_ut2mt(inode->i_ctime); @@ -600,7 +577,7 @@ int hfsplus_cat_write_inode(struct inode *inode) hfs_bnode_read(fd.bnode, &entry, fd.entryoffset, sizeof(struct hfsplus_cat_file)); hfsplus_inode_write_fork(inode, &file->data_fork); - hfsplus_set_perms(inode, &file->permissions); + hfsplus_cat_set_perms(inode, &file->permissions); if ((file->permissions.rootflags | file->permissions.userflags) & HFSPLUS_FLG_IMMUTABLE) file->flags |= cpu_to_be16(HFSPLUS_FILE_LOCKED); else -- cgit v0.10.2 From 32e39e19ccb0f25c9c1b7ff246e17e795366bbbd Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 14 Oct 2010 09:54:43 -0400 Subject: hfsplus: remove the unused hfsplus_kmap/hfsplus_kunmap helpers Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index d77dfd2..cb3653e 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -398,14 +398,6 @@ static inline struct hfsplus_inode_info *HFSPLUS_I(struct inode *inode) return list_entry(inode, struct hfsplus_inode_info, vfs_inode); } -#if 1 -#define hfsplus_kmap(p) ({ struct page *__p = (p); kmap(__p); }) -#define hfsplus_kunmap(p) ({ struct page *__p = (p); kunmap(__p); __p; }) -#else -#define hfsplus_kmap(p) kmap(p) -#define hfsplus_kunmap(p) kunmap(p) -#endif - #define sb_bread512(sb, sec, data) ({ \ struct buffer_head *__bh; \ sector_t __block; \ -- cgit v0.10.2 From 3e0f686e6bf347eea4a41d7d415f003c2f42b1a0 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Wed, 13 Oct 2010 18:07:16 -0700 Subject: ibft: fix kconfig dependencies Fix kconfig dependency warning to satisfy dependencies: warning: (BE2ISCSI && SCSI_LOWLEVEL && PCI && SCSI && NET || ISCSI_IBFT && ISCSI_IBFT_FIND && SCSI) selects ISCSI_BOOT_SYSFS which has unmet direct dependencies (SCSI_LOWLEVEL && SCSI) Signed-off-by: Randy Dunlap Signed-off-by: Peter Jones Signed-off-by: Konrad Rzeszutek Wilk diff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig index 280c9b5..88a3ae6 100644 --- a/drivers/firmware/Kconfig +++ b/drivers/firmware/Kconfig @@ -125,7 +125,7 @@ config ISCSI_IBFT_FIND config ISCSI_IBFT tristate "iSCSI Boot Firmware Table Attributes module" select ISCSI_BOOT_SYSFS - depends on ISCSI_IBFT_FIND && SCSI + depends on ISCSI_IBFT_FIND && SCSI && SCSI_LOWLEVEL default n help This option enables support for detection and exposing of iSCSI -- cgit v0.10.2 From 0eead9ab41da33644ae2c97c57ad03da636a0422 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Thu, 14 Oct 2010 10:57:40 -0700 Subject: Don't dump task struct in a.out core-dumps akiphie points out that a.out core-dumps have that odd task struct dumping that was never used and was never really a good idea (it goes back into the mists of history, probably the original core-dumping code). Just remove it. Also do the access_ok() check on dump_write(). It probably doesn't matter (since normal filesystems all seem to do it anyway), but he points out that it's normally done by the VFS layer, so ... [ I suspect that we should possibly do "vfs_write()" instead of calling ->write directly. That also does the whole fsnotify and write statistics thing, which may or may not be a good idea. ] And just to be anal, do this all for the x86-64 32-bit a.out emulation code too, even though it's not enabled (and won't currently even compile) Reported-by: akiphie Signed-off-by: Linus Torvalds diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c index 0350311..2d93bdb 100644 --- a/arch/x86/ia32/ia32_aout.c +++ b/arch/x86/ia32/ia32_aout.c @@ -34,7 +34,7 @@ #include #undef WARN_OLD -#undef CORE_DUMP /* probably broken */ +#undef CORE_DUMP /* definitely broken */ static int load_aout_binary(struct linux_binprm *, struct pt_regs *regs); static int load_aout_library(struct file *); @@ -131,21 +131,15 @@ static void set_brk(unsigned long start, unsigned long end) * macros to write out all the necessary info. */ -static int dump_write(struct file *file, const void *addr, int nr) -{ - return file->f_op->write(file, addr, nr, &file->f_pos) == nr; -} +#include #define DUMP_WRITE(addr, nr) \ if (!dump_write(file, (void *)(addr), (nr))) \ goto end_coredump; -#define DUMP_SEEK(offset) \ - if (file->f_op->llseek) { \ - if (file->f_op->llseek(file, (offset), 0) != (offset)) \ - goto end_coredump; \ - } else \ - file->f_pos = (offset) +#define DUMP_SEEK(offset) \ + if (!dump_seek(file, offset)) \ + goto end_coredump; #define START_DATA() (u.u_tsize << PAGE_SHIFT) #define START_STACK(u) (u.start_stack) @@ -217,12 +211,6 @@ static int aout_core_dump(long signr, struct pt_regs *regs, struct file *file, dump_size = dump.u_ssize << PAGE_SHIFT; DUMP_WRITE(dump_start, dump_size); } - /* - * Finally dump the task struct. Not be used by gdb, but - * could be useful - */ - set_fs(KERNEL_DS); - DUMP_WRITE(current, sizeof(*current)); end_coredump: set_fs(fs); return has_dumped; diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c index f96eff0..a6395bd 100644 --- a/fs/binfmt_aout.c +++ b/fs/binfmt_aout.c @@ -134,10 +134,6 @@ static int aout_core_dump(struct coredump_params *cprm) if (!dump_write(file, dump_start, dump_size)) goto end_coredump; } -/* Finally dump the task struct. Not be used by gdb, but could be useful */ - set_fs(KERNEL_DS); - if (!dump_write(file, current, sizeof(*current))) - goto end_coredump; end_coredump: set_fs(fs); return has_dumped; diff --git a/include/linux/coredump.h b/include/linux/coredump.h index 8ba66a9..59579cf 100644 --- a/include/linux/coredump.h +++ b/include/linux/coredump.h @@ -11,7 +11,7 @@ */ static inline int dump_write(struct file *file, const void *addr, int nr) { - return file->f_op->write(file, addr, nr, &file->f_pos) == nr; + return access_ok(VERIFY_READ, addr, nr) && file->f_op->write(file, addr, nr, &file->f_pos) == nr; } static inline int dump_seek(struct file *file, loff_t off) -- cgit v0.10.2 From 3aa0ce825ade0cf5506e32ccf51d01fc8d22a9cf Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Thu, 14 Oct 2010 14:32:06 -0700 Subject: Un-inline the core-dump helper functions Tony Luck reports that the addition of the access_ok() check in commit 0eead9ab41da ("Don't dump task struct in a.out core-dumps") broke the ia64 compile due to missing the necessary header file includes. Rather than add yet another include () to make everything happy, just uninline the silly core dump helper functions and move the bodies to fs/exec.c where they make a lot more sense. dump_seek() in particular was too big to be an inline function anyway, and none of them are in any way performance-critical. And we really don't need to mess up our include file headers more than they already are. Reported-and-tested-by: Tony Luck Signed-off-by: Linus Torvalds diff --git a/fs/exec.c b/fs/exec.c index 828dd24..03278c9 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -2014,3 +2014,41 @@ fail_creds: fail: return; } + +/* + * Core dumping helper functions. These are the only things you should + * do on a core-file: use only these functions to write out all the + * necessary info. + */ +int dump_write(struct file *file, const void *addr, int nr) +{ + return access_ok(VERIFY_READ, addr, nr) && file->f_op->write(file, addr, nr, &file->f_pos) == nr; +} + +int dump_seek(struct file *file, loff_t off) +{ + int ret = 1; + + if (file->f_op->llseek && file->f_op->llseek != no_llseek) { + if (file->f_op->llseek(file, off, SEEK_CUR) < 0) + return 0; + } else { + char *buf = (char *)get_zeroed_page(GFP_KERNEL); + + if (!buf) + return 0; + while (off > 0) { + unsigned long n = off; + + if (n > PAGE_SIZE) + n = PAGE_SIZE; + if (!dump_write(file, buf, n)) { + ret = 0; + break; + } + off -= n; + } + free_page((unsigned long)buf); + } + return ret; +} diff --git a/include/linux/coredump.h b/include/linux/coredump.h index 59579cf..ba4b85a 100644 --- a/include/linux/coredump.h +++ b/include/linux/coredump.h @@ -9,37 +9,7 @@ * These are the only things you should do on a core-file: use only these * functions to write out all the necessary info. */ -static inline int dump_write(struct file *file, const void *addr, int nr) -{ - return access_ok(VERIFY_READ, addr, nr) && file->f_op->write(file, addr, nr, &file->f_pos) == nr; -} - -static inline int dump_seek(struct file *file, loff_t off) -{ - int ret = 1; - - if (file->f_op->llseek && file->f_op->llseek != no_llseek) { - if (file->f_op->llseek(file, off, SEEK_CUR) < 0) - return 0; - } else { - char *buf = (char *)get_zeroed_page(GFP_KERNEL); - - if (!buf) - return 0; - while (off > 0) { - unsigned long n = off; - - if (n > PAGE_SIZE) - n = PAGE_SIZE; - if (!dump_write(file, buf, n)) { - ret = 0; - break; - } - off -= n; - } - free_page((unsigned long)buf); - } - return ret; -} +extern int dump_write(struct file *file, const void *addr, int nr); +extern int dump_seek(struct file *file, loff_t off); #endif /* _LINUX_COREDUMP_H */ -- cgit v0.10.2 From cd07202cc8262e1669edff0d97715f3dd9260917 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Thu, 14 Oct 2010 16:26:43 -0700 Subject: Linux 2.6.36-rc8 diff --git a/Makefile b/Makefile index 77b5c6e..7583116 100644 --- a/Makefile +++ b/Makefile @@ -1,8 +1,8 @@ VERSION = 2 PATCHLEVEL = 6 SUBLEVEL = 36 -EXTRAVERSION = -rc7 -NAME = Sheep on Meth +EXTRAVERSION = -rc8 +NAME = Flesh-Eating Bats with Fangs # *DOCUMENTATION* # To see a list of typical targets execute "make help" -- cgit v0.10.2 From 8fd01d6cfbf75465d84a4e533ed70c5f57b3ff51 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Thu, 14 Oct 2010 19:15:28 -0700 Subject: Export dump_{write,seek} to binary loader modules If you build aout support as a module, you'll want these exported. Reported-by: Tetsuo Handa Signed-off-by: Linus Torvalds diff --git a/fs/exec.c b/fs/exec.c index 03278c9..6d2b6f9 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -2024,6 +2024,7 @@ int dump_write(struct file *file, const void *addr, int nr) { return access_ok(VERIFY_READ, addr, nr) && file->f_op->write(file, addr, nr, &file->f_pos) == nr; } +EXPORT_SYMBOL(dump_write); int dump_seek(struct file *file, loff_t off) { @@ -2052,3 +2053,4 @@ int dump_seek(struct file *file, loff_t off) } return ret; } +EXPORT_SYMBOL(dump_seek); -- cgit v0.10.2 From 9e9006e9090dc1f88d5127cb69f416013e7ecd60 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Thu, 14 Oct 2010 10:13:13 -0700 Subject: x86, olpc: XO-1 uses/depends on PCI olpc-xo1 uses pci_*() interfaces so it should depend on PCI. Otherwise we get build failure like: arch/x86/kernel/olpc-xo1.c:65: error: implicit declaration of function 'pci_enable_device_io' arch/x86/kernel/olpc-xo1.c:71: error: implicit declaration of function 'pci_request_region' arch/x86/kernel/olpc-xo1.c:80: error: implicit declaration of function 'pci_release_region' Signed-off-by: Randy Dunlap Acked-by: Daniel Drake Cc: Stephen Rothwell LKML-Reference: <20101014101313.adf7eb2a.randy.dunlap@oracle.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 0bcb876..c010b8d 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -2068,7 +2068,7 @@ config OLPC config OLPC_XO1 tristate "OLPC XO-1 support" - depends on OLPC + depends on OLPC && PCI ---help--- Add support for non-essential features of the OLPC XO-1 laptop. -- cgit v0.10.2 From 47526903feb52f4c26a6350370bdf74e337fcdb1 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 15 Oct 2010 12:56:21 +0200 Subject: ubd: fix incorrect sector handling during request restart Commit f81f2f7c (ubd: drop unnecessary rq->sector manipulation) dropped request->sector manipulation in preparation for global request handling cleanup; unfortunately, it incorrectly assumed that the updated sector wasn't being used. ubd tries to issue as many requests as possible to io_thread. When issuing fails due to memory pressure or other reasons, the device is put on the restart list and issuing stops. On IO completion, devices on the restart list are scanned and IO issuing is restarted. ubd issues IOs sg-by-sg and issuing can be stopped in the middle of a request, so each device on the restart queue needs to remember where to restart in its current request. ubd needs to keep track of the issue position itself because, * blk_rq_pos(req) is now updated by the block layer to keep track of _completion_ position. * Multiple io_req's for the current request may be in flight, so it's difficult to tell where blk_rq_pos(req) currently is. Add ubd->rq_pos to keep track of the issue position and use it to correctly restart io_req issue. Signed-off-by: Tejun Heo Reported-by: Richard Weinberger Tested-by: Richard Weinberger Tested-by: Chris Frey Cc: stable@kernel.org Signed-off-by: Jens Axboe diff --git a/arch/um/drivers/ubd_kern.c b/arch/um/drivers/ubd_kern.c index 1bcd208..9734994 100644 --- a/arch/um/drivers/ubd_kern.c +++ b/arch/um/drivers/ubd_kern.c @@ -163,6 +163,7 @@ struct ubd { struct scatterlist sg[MAX_SG]; struct request *request; int start_sg, end_sg; + sector_t rq_pos; }; #define DEFAULT_COW { \ @@ -187,6 +188,7 @@ struct ubd { .request = NULL, \ .start_sg = 0, \ .end_sg = 0, \ + .rq_pos = 0, \ } /* Protected by ubd_lock */ @@ -1228,7 +1230,6 @@ static void do_ubd_request(struct request_queue *q) { struct io_thread_req *io_req; struct request *req; - sector_t sector; int n; while(1){ @@ -1239,12 +1240,12 @@ static void do_ubd_request(struct request_queue *q) return; dev->request = req; + dev->rq_pos = blk_rq_pos(req); dev->start_sg = 0; dev->end_sg = blk_rq_map_sg(q, req, dev->sg); } req = dev->request; - sector = blk_rq_pos(req); while(dev->start_sg < dev->end_sg){ struct scatterlist *sg = &dev->sg[dev->start_sg]; @@ -1256,10 +1257,9 @@ static void do_ubd_request(struct request_queue *q) return; } prepare_request(req, io_req, - (unsigned long long)sector << 9, + (unsigned long long)dev->rq_pos << 9, sg->offset, sg->length, sg_page(sg)); - sector += sg->length >> 9; n = os_write_file(thread_fd, &io_req, sizeof(struct io_thread_req *)); if(n != sizeof(struct io_thread_req *)){ @@ -1272,6 +1272,7 @@ static void do_ubd_request(struct request_queue *q) return; } + dev->rq_pos += sg->length >> 9; dev->start_sg++; } dev->end_sg = 0; -- cgit v0.10.2 From 46bf36ecec79bbe5373ef68f0ed36cbf62e03482 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 15 Oct 2010 05:45:00 -0700 Subject: hfsplus: fix getxattr return value We need to support -EOPNOTSUPP for attributes that are not supported to match other filesystems and allow userspace to detect if Posix ACLs are supported or not. setxattr already gets this right. Signed-off-by: Christoph Hellwig diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index 80eb5b3..5b4667e 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -192,7 +192,7 @@ ssize_t hfsplus_getxattr(struct dentry *dentry, const char *name, } else res = size ? -ERANGE : 4; } else - res = -ENODATA; + res = -EOPNOTSUPP; out: if (size) hfs_find_exit(&fd); -- cgit v0.10.2 From 478971600e47cb83ff2d3c63c5c24f2b04b0d6a1 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Fri, 17 Sep 2010 00:46:42 +0900 Subject: [SCSI] bsg: fix incorrect device_status value bsg incorrectly returns sg's masked_status value for device_status. [jejb: fix up expression logic] Reported-by: Douglas Gilbert Signed-off-by: FUJITA Tomonori Cc: Stable Tree Signed-off-by: James Bottomley diff --git a/block/bsg.c b/block/bsg.c index 82d5882..0c00870 100644 --- a/block/bsg.c +++ b/block/bsg.c @@ -426,7 +426,7 @@ static int blk_complete_sgv4_hdr_rq(struct request *rq, struct sg_io_v4 *hdr, /* * fill in all the output members */ - hdr->device_status = status_byte(rq->errors); + hdr->device_status = rq->errors & 0xff; hdr->transport_status = host_byte(rq->errors); hdr->driver_status = driver_byte(rq->errors); hdr->info = 0; -- cgit v0.10.2 From 444758636439810f110f86a3042c2dfd3626a9e2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 15 Oct 2010 11:49:47 -0400 Subject: ftrace: Do not process kernel/trace/ftrace.o with C recordmcount program The file kernel/trace/ftrace.c references the mcount() call to convert the mcount() callers to nops. But because it references mcount(), the mcount() address is placed in the relocation table. The C version of recordmcount reads the relocation table of all object files, and it will add all references to mcount to the __mcount_loc table that is used to find the places that call mcount() and change the call to a nop. When recordmcount finds the mcount reference in kernel/trace/ftrace.o, it saves that location even though the code is not a call, but references mcount as data. On boot up, when all calls are converted to nops, the code has a safety check to determine what op code it is actually replacing before it replaces it. If that op code at the address does not match, then a warning is printed and the function tracer is disabled. The reference to mcount in ftrace.c, causes this warning to trigger, since the reference is not a call to mcount(). The ftrace.c file is not compiled with the -pg flag, so no calls to mcount() should be expected. This patch simply makes recordmcount.c skip the kernel/trace/ftrace.c file. This was the same solution used by the perl version of recordmcount. Reported-by: Ingo Molnar Cc: John Reiser Signed-off-by: Steven Rostedt diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c index 7f7f718..26e1271 100644 --- a/scripts/recordmcount.c +++ b/scripts/recordmcount.c @@ -313,12 +313,30 @@ do_file(char const *const fname) int main(int argc, char const *argv[]) { + const char ftrace[] = "kernel/trace/ftrace.o"; + int ftrace_size = sizeof(ftrace) - 1; int n_error = 0; /* gcc-4.3.0 false positive complaint */ - if (argc <= 1) + + if (argc <= 1) { fprintf(stderr, "usage: recordmcount file.o...\n"); - else /* Process each file in turn, allowing deep failure. */ + return 0; + } + + /* Process each file in turn, allowing deep failure. */ for (--argc, ++argv; 0 < argc; --argc, ++argv) { int const sjval = setjmp(jmpenv); + int len; + + /* + * The file kernel/trace/ftrace.o references the mcount + * function but does not call it. Since ftrace.o should + * not be traced anyway, we just skip it. + */ + len = strlen(argv[0]); + if (len >= ftrace_size && + strcmp(argv[0] + (len - ftrace_size), ftrace) == 0) + continue; + switch (sjval) { default: { fprintf(stderr, "internal error: %s\n", argv[0]); -- cgit v0.10.2 From 85caa993d7f218b7c2abcfcb28e212396b6a3313 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 15 Oct 2010 12:09:25 -0400 Subject: ftrace: Use objtree for C version of recordmcount The C version of recordmcount is compiled to a binary, which will end up located in the objtree. If the kernel is built with O=path, the srctree will not include the binary recordmcount caller. Cc: Michal Marek Cc: linux-kbuild@vger.kernel.org Signed-off-by: Steven Rostedt diff --git a/scripts/Makefile.build b/scripts/Makefile.build index 4d03a7ef..4db60b2 100644 --- a/scripts/Makefile.build +++ b/scripts/Makefile.build @@ -210,7 +210,7 @@ endif ifdef CONFIG_FTRACE_MCOUNT_RECORD ifdef BUILD_C_RECORDMCOUNT -cmd_record_mcount = $(srctree)/scripts/recordmcount "$(@)"; +cmd_record_mcount = $(objtree)/scripts/recordmcount "$(@)"; else cmd_record_mcount = set -e ; perl $(srctree)/scripts/recordmcount.pl "$(ARCH)" \ "$(if $(CONFIG_CPU_BIG_ENDIAN),big,little)" \ -- cgit v0.10.2 From 1c8cf9c997a4a6b36e907c7ede5f048aeaab1644 Mon Sep 17 00:00:00 2001 From: Ohad Ben-Cohen Date: Wed, 13 Oct 2010 09:31:56 +0200 Subject: mmc: sdio: fix SDIO suspend/resume regression Fix SDIO suspend/resume regression introduced by 4c2ef25fe0b "mmc: fix all hangs related to mmc/sd card insert/removal during suspend/resume": PM: Syncing filesystems ... done. Freezing user space processes ... (elapsed 0.01 seconds) done. Freezing remaining freezable tasks ... (elapsed 0.01 seconds) done. Suspending console(s) (use no_console_suspend to debug) pm_op(): platform_pm_suspend+0x0/0x5c returns -38 PM: Device pxa2xx-mci.0 failed to suspend: error -38 PM: Some devices failed to suspend 4c2ef25fe0b moved the card removal/insertion mechanism out of MMC's suspend/resume path and into pm notifiers (mmc_pm_notify), and that broke SDIO's expectation that mmc_suspend_host() will remove the card, and squash the error, in case -ENOSYS is returned from the bus suspend handler (mmc_sdio_suspend() in this case). mmc_sdio_suspend() is using this whenever at least one of the card's SDIO function drivers does not have suspend/resume handlers - in that case it is agreed to force removal of the entire card. This patch fixes this regression by trivially bringing back that part of mmc_suspend_host(), which was removed by 4c2ef25fe0b. Reported-and-tested-by: Sven Neumann Signed-off-by: Ohad Ben-Cohen Cc: Maxim Levitsky Cc: Acked-by: Nicolas Pitre Signed-off-by: Chris Ball diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c index 5db49b1..09eee6d 100644 --- a/drivers/mmc/core/core.c +++ b/drivers/mmc/core/core.c @@ -1631,6 +1631,19 @@ int mmc_suspend_host(struct mmc_host *host) if (host->bus_ops && !host->bus_dead) { if (host->bus_ops->suspend) err = host->bus_ops->suspend(host); + if (err == -ENOSYS || !host->bus_ops->resume) { + /* + * We simply "remove" the card in this case. + * It will be redetected on resume. + */ + if (host->bus_ops->remove) + host->bus_ops->remove(host); + mmc_claim_host(host); + mmc_detach_bus(host); + mmc_release_host(host); + host->pm_flags = 0; + err = 0; + } } mmc_bus_put(host); -- cgit v0.10.2 From 940b3c7b193ec54141976a8642c00582f423690c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 15 Oct 2010 19:36:26 +0200 Subject: x86: sfi: Make local functions static Signed-off-by: Thomas Gleixner Cc: Len Brown diff --git a/arch/x86/kernel/sfi.c b/arch/x86/kernel/sfi.c index cb22acf..dd4c281 100644 --- a/arch/x86/kernel/sfi.c +++ b/arch/x86/kernel/sfi.c @@ -34,7 +34,7 @@ #ifdef CONFIG_X86_LOCAL_APIC static unsigned long sfi_lapic_addr __initdata = APIC_DEFAULT_PHYS_BASE; -void __init mp_sfi_register_lapic_address(unsigned long address) +static void __init mp_sfi_register_lapic_address(unsigned long address) { mp_lapic_addr = address; @@ -46,7 +46,7 @@ void __init mp_sfi_register_lapic_address(unsigned long address) } /* All CPUs enumerated by SFI must be present and enabled */ -void __cpuinit mp_sfi_register_lapic(u8 id) +static void __cpuinit mp_sfi_register_lapic(u8 id) { if (MAX_APICS - id <= 0) { pr_warning("Processor #%d invalid (max %d)\n", -- cgit v0.10.2 From 799c10559d60f159ab2232203f222f18fa3c4a5f Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 15 Oct 2010 11:09:28 -0700 Subject: De-pessimize rds_page_copy_user Don't try to "optimize" rds_page_copy_user() by using kmap_atomic() and the unsafe atomic user mode accessor functions. It's actually slower than the straightforward code on any reasonable modern CPU. Back when the code was written (although probably not by the time it was actually merged, though), 32-bit x86 may have been the dominant architecture. And there kmap_atomic() can be a lot faster than kmap() (unless you have very good locality, in which case the virtual address caching by kmap() can overcome all the downsides). But these days, x86-64 may not be more populous, but it's getting there (and if you care about performance, it's definitely already there - you'd have upgraded your CPU's already in the last few years). And on x86-64, the non-kmap_atomic() version is faster, simply because the code is simpler and doesn't have the "re-try page fault" case. People with old hardware are not likely to care about RDS anyway, and the optimization for the 32-bit case is simply buggy, since it doesn't verify the user addresses properly. Reported-by: Dan Rosenberg Acked-by: Andrew Morton Cc: stable@kernel.org Signed-off-by: Linus Torvalds diff --git a/net/rds/page.c b/net/rds/page.c index 595a952..1dfbfea 100644 --- a/net/rds/page.c +++ b/net/rds/page.c @@ -57,30 +57,17 @@ int rds_page_copy_user(struct page *page, unsigned long offset, unsigned long ret; void *addr; - if (to_user) + addr = kmap(page); + if (to_user) { rds_stats_add(s_copy_to_user, bytes); - else + ret = copy_to_user(ptr, addr + offset, bytes); + } else { rds_stats_add(s_copy_from_user, bytes); - - addr = kmap_atomic(page, KM_USER0); - if (to_user) - ret = __copy_to_user_inatomic(ptr, addr + offset, bytes); - else - ret = __copy_from_user_inatomic(addr + offset, ptr, bytes); - kunmap_atomic(addr, KM_USER0); - - if (ret) { - addr = kmap(page); - if (to_user) - ret = copy_to_user(ptr, addr + offset, bytes); - else - ret = copy_from_user(addr + offset, ptr, bytes); - kunmap(page); - if (ret) - return -EFAULT; + ret = copy_from_user(addr + offset, ptr, bytes); } + kunmap(page); - return 0; + return ret ? -EFAULT : 0; } EXPORT_SYMBOL_GPL(rds_page_copy_user); -- cgit v0.10.2 From 3e645d6b485446c54c6745c5e2cf5c528fe4deec Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 15 Oct 2010 11:12:38 -0700 Subject: v4l1: fix 32-bit compat microcode loading translation The compat code for the VIDIOCSMICROCODE ioctl is totally buggered. It's only used by the VIDEO_STRADIS driver, and that one is scheduled to staging and eventually removed unless somebody steps up to maintain it (at which point it should use request_firmware() rather than some magic ioctl). So we'll get rid of it eventually. But in the meantime, the compatibility ioctl code is broken, and this tries to get it to at least limp along (even if Mauro suggested just deleting it entirely, which may be the right thing to do - I don't think the compatibility translation code has ever worked unless you were very lucky). Reported-by: Kees Cook Cc: Mauro Carvalho Chehab Cc: stable@kernel.org Signed-off-by: Linus Torvalds diff --git a/drivers/media/video/v4l2-compat-ioctl32.c b/drivers/media/video/v4l2-compat-ioctl32.c index 073f013..86294ed3 100644 --- a/drivers/media/video/v4l2-compat-ioctl32.c +++ b/drivers/media/video/v4l2-compat-ioctl32.c @@ -193,17 +193,24 @@ static int put_video_window32(struct video_window *kp, struct video_window32 __u struct video_code32 { char loadwhat[16]; /* name or tag of file being passed */ compat_int_t datasize; - unsigned char *data; + compat_uptr_t data; }; -static int get_microcode32(struct video_code *kp, struct video_code32 __user *up) +static struct video_code __user *get_microcode32(struct video_code32 *kp) { - if (!access_ok(VERIFY_READ, up, sizeof(struct video_code32)) || - copy_from_user(kp->loadwhat, up->loadwhat, sizeof(up->loadwhat)) || - get_user(kp->datasize, &up->datasize) || - copy_from_user(kp->data, up->data, up->datasize)) - return -EFAULT; - return 0; + struct video_code __user *up; + + up = compat_alloc_user_space(sizeof(*up)); + + /* + * NOTE! We don't actually care if these fail. If the + * user address is invalid, the native ioctl will do + * the error handling for us + */ + (void) copy_to_user(up->loadwhat, kp->loadwhat, sizeof(up->loadwhat)); + (void) put_user(kp->datasize, &up->datasize); + (void) put_user(compat_ptr(kp->data), &up->data); + return up; } #define VIDIOCGTUNER32 _IOWR('v', 4, struct video_tuner32) @@ -739,7 +746,7 @@ static long do_video_ioctl(struct file *file, unsigned int cmd, unsigned long ar struct video_tuner vt; struct video_buffer vb; struct video_window vw; - struct video_code vc; + struct video_code32 vc; struct video_audio va; #endif struct v4l2_format v2f; @@ -818,8 +825,11 @@ static long do_video_ioctl(struct file *file, unsigned int cmd, unsigned long ar break; case VIDIOCSMICROCODE: - err = get_microcode32(&karg.vc, up); - compatible_arg = 0; + /* Copy the 32-bit "video_code32" to kernel space */ + if (copy_from_user(&karg.vc, up, sizeof(karg.vc))) + return -EFAULT; + /* Convert the 32-bit version to a 64-bit version in user space */ + up = get_microcode32(&karg.vc); break; case VIDIOCSFREQ: -- cgit v0.10.2 From 40ffa93791985ab300fd488072e9f37ccf72e88c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 15 Oct 2010 21:08:14 +0200 Subject: x86: Remove stale pmtimer_64.c This file is unused since the apic unification in 2.6.29, but nobody noticed. Signed-off-by: Thomas Gleixner diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 0925676..dd9a2e4 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -120,7 +120,6 @@ obj-$(CONFIG_SWIOTLB) += pci-swiotlb.o # 64 bit specific files ifeq ($(CONFIG_X86_64),y) obj-$(CONFIG_X86_UV) += tlb_uv.o bios_uv.o uv_irq.o uv_sysfs.o uv_time.o - obj-$(CONFIG_X86_PM_TIMER) += pmtimer_64.o obj-$(CONFIG_AUDIT) += audit_64.o obj-$(CONFIG_GART_IOMMU) += pci-gart_64.o aperture_64.o diff --git a/arch/x86/kernel/pmtimer_64.c b/arch/x86/kernel/pmtimer_64.c deleted file mode 100644 index b112406..0000000 --- a/arch/x86/kernel/pmtimer_64.c +++ /dev/null @@ -1,69 +0,0 @@ -/* Ported over from i386 by AK, original copyright was: - * - * (C) Dominik Brodowski 2003 - * - * Driver to use the Power Management Timer (PMTMR) available in some - * southbridges as primary timing source for the Linux kernel. - * - * Based on parts of linux/drivers/acpi/hardware/hwtimer.c, timer_pit.c, - * timer_hpet.c, and on Arjan van de Ven's implementation for 2.4. - * - * This file is licensed under the GPL v2. - * - * Dropped all the hardware bug workarounds for now. Hopefully they - * are not needed on 64bit chipsets. - */ - -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include - -static inline u32 cyc2us(u32 cycles) -{ - /* The Power Management Timer ticks at 3.579545 ticks per microsecond. - * 1 / PM_TIMER_FREQUENCY == 0.27936511 =~ 286/1024 [error: 0.024%] - * - * Even with HZ = 100, delta is at maximum 35796 ticks, so it can - * easily be multiplied with 286 (=0x11E) without having to fear - * u32 overflows. - */ - cycles *= 286; - return (cycles >> 10); -} - -static unsigned pmtimer_wait_tick(void) -{ - u32 a, b; - for (a = b = inl(pmtmr_ioport) & ACPI_PM_MASK; - a == b; - b = inl(pmtmr_ioport) & ACPI_PM_MASK) - cpu_relax(); - return b; -} - -/* note: wait time is rounded up to one tick */ -void pmtimer_wait(unsigned us) -{ - u32 a, b; - a = pmtimer_wait_tick(); - do { - b = inl(pmtmr_ioport); - cpu_relax(); - } while (cyc2us(b - a) < us); -} - -static int __init nopmtimer_setup(char *s) -{ - pmtmr_ioport = 0; - return 1; -} - -__setup("nopmtimer", nopmtimer_setup); diff --git a/include/linux/acpi_pmtmr.h b/include/linux/acpi_pmtmr.h index 7e3d285..1d0ef1a 100644 --- a/include/linux/acpi_pmtmr.h +++ b/include/linux/acpi_pmtmr.h @@ -25,8 +25,6 @@ static inline u32 acpi_pm_read_early(void) return acpi_pm_read_verified() & ACPI_PM_MASK; } -extern void pmtimer_wait(unsigned); - #else static inline u32 acpi_pm_read_early(void) -- cgit v0.10.2 From a9febbb4bd1302b6f01aa1203b0a804e4e5c9e25 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 15 Oct 2010 14:34:12 -0700 Subject: sysctl: min/max bounds are optional sysctl check complains with a WARN() when proc_doulongvec_minmax() or proc_doulongvec_ms_jiffies_minmax() are used by a vector of longs (with more than one element), with no min or max value specified. This is unexpected, given we had a bug on this min/max handling :) Reported-by: Jiri Slaby Signed-off-by: Eric Dumazet Cc: "Eric W. Biederman" Cc: David Miller Acked-by: WANG Cong Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/kernel/sysctl_check.c b/kernel/sysctl_check.c index 04cdcf7..10b90d8 100644 --- a/kernel/sysctl_check.c +++ b/kernel/sysctl_check.c @@ -143,15 +143,6 @@ int sysctl_check_table(struct nsproxy *namespaces, struct ctl_table *table) if (!table->maxlen) set_fail(&fail, table, "No maxlen"); } - if ((table->proc_handler == proc_doulongvec_minmax) || - (table->proc_handler == proc_doulongvec_ms_jiffies_minmax)) { - if (table->maxlen > sizeof (unsigned long)) { - if (!table->extra1) - set_fail(&fail, table, "No min"); - if (!table->extra2) - set_fail(&fail, table, "No max"); - } - } #ifdef CONFIG_PROC_SYSCTL if (table->procname && !table->proc_handler) set_fail(&fail, table, "No proc_handler"); -- cgit v0.10.2 From e3c6cf61815b0af0c697aeed4c6f11762f913002 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Fri, 15 Oct 2010 14:34:13 -0700 Subject: uml: fix build MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix a build error introduced by d6d1b650ae6acce73d55dd024 ("param: simple locking for sysfs-writable charp parameters"). CC arch/um/kernel/trap.o arch/um/drivers/hostaudio_kern.c: In function 'hostaudio_open': arch/um/drivers/hostaudio_kern.c:204: error: '__param_dsp' undeclared (first use in this function) arch/um/drivers/hostaudio_kern.c:204: error: (Each undeclared identifier is reported only once arch/um/drivers/hostaudio_kern.c:204: error: for each function it appears in.) arch/um/drivers/hostaudio_kern.c: In function 'hostmixer_open_mixdev': arch/um/drivers/hostaudio_kern.c:265: error: '__param_mixer' undeclared (first use in this function) arch/um/drivers/hostaudio_kern.c:272: error: '__param_dsp' undeclared (first use in this function) Reported-by: Toralf Förster Tested-by: Toralf Förster Cc: Rusty Russell Cc: Takashi Iwai Cc: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/arch/um/drivers/hostaudio_kern.c b/arch/um/drivers/hostaudio_kern.c index 0c46e39..63c740a 100644 --- a/arch/um/drivers/hostaudio_kern.c +++ b/arch/um/drivers/hostaudio_kern.c @@ -40,6 +40,11 @@ static char *mixer = HOSTAUDIO_DEV_MIXER; " This is used to specify the host mixer device to the hostaudio driver.\n"\ " The default is \"" HOSTAUDIO_DEV_MIXER "\".\n\n" +module_param(dsp, charp, 0644); +MODULE_PARM_DESC(dsp, DSP_HELP); +module_param(mixer, charp, 0644); +MODULE_PARM_DESC(mixer, MIXER_HELP); + #ifndef MODULE static int set_dsp(char *name, int *add) { @@ -56,15 +61,6 @@ static int set_mixer(char *name, int *add) } __uml_setup("mixer=", set_mixer, "mixer=\n" MIXER_HELP); - -#else /*MODULE*/ - -module_param(dsp, charp, 0644); -MODULE_PARM_DESC(dsp, DSP_HELP); - -module_param(mixer, charp, 0644); -MODULE_PARM_DESC(mixer, MIXER_HELP); - #endif /* /dev/dsp file operations */ -- cgit v0.10.2 From 79b5dc0c64d88cda3da23b2e22a5cec0964372ac Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Fri, 15 Oct 2010 14:34:14 -0700 Subject: types.h: define __aligned_u64 and expose to userspace We currently have a kernel internal type called aligned_u64 which aligns __u64's on 8 bytes boundaries even on systems which would normally align them on 4 byte boundaries. This patch creates a new type __aligned_u64 which does the same thing but which is exposed to userspace rather than being kernel internal. [akpm: merge early as both the net and audit trees want this] [akpm@linux-foundation.org: enhance the comment describing the reasons for using aligned_u64. Via Andreas and Andi.] Based-on-patch-by: Andreas Gruenbacher Signed-off-by: Eric Paris Cc: Jan Engelhardt Cc: David Miller Cc: Andi Kleen Cc: Arnd Bergmann Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/include/linux/types.h b/include/linux/types.h index 01a082f..357dbc1 100644 --- a/include/linux/types.h +++ b/include/linux/types.h @@ -121,7 +121,15 @@ typedef __u64 u_int64_t; typedef __s64 int64_t; #endif -/* this is a special 64bit data type that is 8-byte aligned */ +/* + * aligned_u64 should be used in defining kernel<->userspace ABIs to avoid + * common 32/64-bit compat problems. + * 64-bit values align to 4-byte boundaries on x86_32 (and possibly other + * architectures) and to 8-byte boundaries on 64-bit architetures. The new + * aligned_64 type enforces 8-byte alignment so that structs containing + * aligned_64 values have the same alignment on 32-bit and 64-bit architectures. + * No conversions are necessary between 32-bit user-space and a 64-bit kernel. + */ #define aligned_u64 __u64 __attribute__((aligned(8))) #define aligned_be64 __be64 __attribute__((aligned(8))) #define aligned_le64 __le64 __attribute__((aligned(8))) @@ -178,6 +186,11 @@ typedef __u64 __bitwise __be64; typedef __u16 __bitwise __sum16; typedef __u32 __bitwise __wsum; +/* this is a special 64bit data type that is 8-byte aligned */ +#define __aligned_u64 __u64 __attribute__((aligned(8))) +#define __aligned_be64 __be64 __attribute__((aligned(8))) +#define __aligned_le64 __le64 __attribute__((aligned(8))) + #ifdef __KERNEL__ typedef unsigned __bitwise__ gfp_t; typedef unsigned __bitwise__ fmode_t; -- cgit v0.10.2 From 99d6734f3ca66c17b81df85724953a9b29eff7cf Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Fri, 15 Oct 2010 21:16:45 -0400 Subject: m32r: restore _BLOCKABLE Commit a7f8388e accidentally removed it... Al explains: "Sorry, reordering breakage. In the signals tree here I have static inline void sig_set_blocked(struct sigset_t *set) ... and it's used all over the place (including quite a few places where we currently have sigprocmask(SIG_SETMASK, set, NULL), which is what it's equivalent to). With that done, m32r doesn't use _BLOCKABLE anywhere, so it got removed. And that chunk got picked when I'd been reordering the queue to pull the arch-specific fixes in front. Sorry." Signed-off-by: Kyle McMartin Cc: Al Viro Signed-off-by: Linus Torvalds diff --git a/arch/m32r/kernel/signal.c b/arch/m32r/kernel/signal.c index 7bbe386..cc36fe1 100644 --- a/arch/m32r/kernel/signal.c +++ b/arch/m32r/kernel/signal.c @@ -28,6 +28,8 @@ #define DEBUG_SIG 0 +#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) + asmlinkage int sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, unsigned long r2, unsigned long r3, unsigned long r4, -- cgit v0.10.2 From 388d148fe89cbdd99ec6bcab6cf34480e74c50ee Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Fri, 15 Oct 2010 21:17:09 -0400 Subject: m32r: get_user takes an lvalue, not a pointer Signed-off-by: Kyle McMartin Acked-by: Al "my fuckup" Viro Signed-off-by: Linus Torvalds diff --git a/arch/m32r/kernel/signal.c b/arch/m32r/kernel/signal.c index cc36fe1..a08697f 100644 --- a/arch/m32r/kernel/signal.c +++ b/arch/m32r/kernel/signal.c @@ -256,7 +256,7 @@ give_sigsegv: static int prev_insn(struct pt_regs *regs) { u16 inst; - if (get_user(&inst, (u16 __user *)(regs->bpc - 2))) + if (get_user(inst, (u16 __user *)(regs->bpc - 2))) return -EFAULT; if ((inst & 0xfff0) == 0x10f0) /* trap ? */ regs->bpc -= 2; -- cgit v0.10.2 From 4f515cc93248face6ba45dd296dc3e7381848ab7 Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Fri, 15 Oct 2010 21:17:13 -0400 Subject: m32r: add kernel/.gitignore and ignore vmlinux.lds Signed-off-by: Kyle McMartin Signed-off-by: Linus Torvalds diff --git a/arch/m32r/kernel/.gitignore b/arch/m32r/kernel/.gitignore new file mode 100644 index 0000000..c5f676c --- /dev/null +++ b/arch/m32r/kernel/.gitignore @@ -0,0 +1 @@ +vmlinux.lds -- cgit v0.10.2 From 2d019713b7acbc01d769a1e512ca1f9a04c30bd0 Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Fri, 15 Oct 2010 21:17:16 -0400 Subject: m32r: test __LITTLE_ENDIAN__ instead of __LITTLE_ENDIAN Fixes build for me... these are what's tested in byteorder.h... Signed-off-by: Kyle McMartin Signed-off-by: Linus Torvalds diff --git a/arch/m32r/include/asm/elf.h b/arch/m32r/include/asm/elf.h index 2f85412..b8da7d0 100644 --- a/arch/m32r/include/asm/elf.h +++ b/arch/m32r/include/asm/elf.h @@ -82,9 +82,9 @@ typedef elf_fpreg_t elf_fpregset_t; * These are used to set parameters in the core dumps. */ #define ELF_CLASS ELFCLASS32 -#if defined(__LITTLE_ENDIAN) +#if defined(__LITTLE_ENDIAN__) #define ELF_DATA ELFDATA2LSB -#elif defined(__BIG_ENDIAN) +#elif defined(__BIG_ENDIAN__) #define ELF_DATA ELFDATA2MSB #else #error no endian defined -- cgit v0.10.2 From 05d0ca85c91013d5f73c43fb566bde09a9718019 Mon Sep 17 00:00:00 2001 From: Anand Gadiyar Date: Thu, 14 Oct 2010 22:59:57 +0530 Subject: genirq, ARM: Fix boot on ARM platforms Commit b683de2b3 in linux-next as of 20101014 (genirq: Query arch for number of early descriptors) seems to have broken bootup on several ARM boards - my beagleboard gives the following dump with earlyprintk: NR_IRQS:402 Unable to handle kernel NULL pointer dereference at virtual address 00000028 pgd = c0004000 [00000028] *pgd=00000000 Internal error: Oops: 5 [#1] last sysfs file: Modules linked in: CPU: 0 Not tainted (2.6.36-rc7-next-20101014-linux-next-20101012+ #40) PC is at init_IRQ+0x14/0x48 LR is at start_kernel+0x150/0x2c0 [...] We seem to be using desc->status without assigning desc to anything. Fix this by adding back the code that was originally there. Signed-off-by: Anand Gadiyar Tested-by: Linus Walleij Cc: Russell King Cc: linux-arm-kernel@lists.infradead.org LKML-Reference: <1287077397-21781-1-git-send-email-gadiyar@ti.com> Signed-off-by: Ingo Molnar diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c index 5456d11..0e65d7e 100644 --- a/arch/arm/kernel/irq.c +++ b/arch/arm/kernel/irq.c @@ -157,8 +157,10 @@ void __init init_IRQ(void) struct irq_desc *desc; int irq; - for (irq = 0; irq < nr_irqs; irq++) + for (irq = 0; irq < nr_irqs; irq++) { + desc = irq_to_desc_alloc_node(irq, 0); desc->status |= IRQ_NOREQUEST | IRQ_NOPROBE; + } init_arch_irq(); } -- cgit v0.10.2 From 032fa36091a925f555158f4e80f8a09d951cdb10 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 16 Oct 2010 21:17:08 +0200 Subject: arm: Use ARCH_IRQ_INIT_FLAGS The core code now initializes the requested number of interrupts and sets the flags in irq_desc.status which are requested by the architecture via ARCH_IRQ_INIT_FLAGS. Add ARCH_IRQ_INIT_FLAGS and remove the loop which sets those flags after the irq descriptors are allocated. [ This patch should have been in the original irq rework and got dropped accidentaly ] Signed-off-by: Thomas Gleixner Cc: Russell King Cc: Linus Walleij Cc: Anand Gadiyar diff --git a/arch/arm/include/asm/hw_irq.h b/arch/arm/include/asm/hw_irq.h index 90831f6..5586b7c 100644 --- a/arch/arm/include/asm/hw_irq.h +++ b/arch/arm/include/asm/hw_irq.h @@ -24,4 +24,6 @@ void set_irq_flags(unsigned int irq, unsigned int flags); #define IRQF_PROBE (1 << 1) #define IRQF_NOAUTOEN (1 << 2) +#define ARCH_IRQ_INIT_FLAGS (IRQ_NOREQUEST | IRQ_NOPROBE) + #endif diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c index 0e65d7e..36ad3be 100644 --- a/arch/arm/kernel/irq.c +++ b/arch/arm/kernel/irq.c @@ -154,14 +154,6 @@ void set_irq_flags(unsigned int irq, unsigned int iflags) void __init init_IRQ(void) { - struct irq_desc *desc; - int irq; - - for (irq = 0; irq < nr_irqs; irq++) { - desc = irq_to_desc_alloc_node(irq, 0); - desc->status |= IRQ_NOREQUEST | IRQ_NOPROBE; - } - init_arch_irq(); } -- cgit v0.10.2 From aa73aec6c385e2c797ac25cc7ccf0318031de7c8 Mon Sep 17 00:00:00 2001 From: Clemens Ladisch Date: Fri, 15 Oct 2010 12:06:18 +0200 Subject: ALSA: rawmidi: fix oops (use after free) when unloading a driver module When a driver module is unloaded and the last still open file is a raw MIDI device, the card and its devices will be actually freed in the snd_card_file_remove() call when that file is closed. Afterwards, rmidi and rmidi->card point into freed memory, so the module pointer is likely to be garbage. (This was introduced by commit 9a1b64caac82aa02cb74587ffc798e6f42c6170a.) Signed-off-by: Clemens Ladisch Reported-by: Krzysztof Foltman Cc: 2.6.30-2.6.35 Signed-off-by: Takashi Iwai diff --git a/sound/core/rawmidi.c b/sound/core/rawmidi.c index a7868ad..cbbed0d 100644 --- a/sound/core/rawmidi.c +++ b/sound/core/rawmidi.c @@ -535,13 +535,15 @@ static int snd_rawmidi_release(struct inode *inode, struct file *file) { struct snd_rawmidi_file *rfile; struct snd_rawmidi *rmidi; + struct module *module; rfile = file->private_data; rmidi = rfile->rmidi; rawmidi_release_priv(rfile); kfree(rfile); + module = rmidi->card->module; snd_card_file_remove(rmidi->card, file); - module_put(rmidi->card->module); + module_put(module); return 0; } -- cgit v0.10.2 From aa0170fff3c26bf2b42159af2dd9cf86444c292a Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Sun, 17 Oct 2010 14:09:12 +0200 Subject: firewire: ohci: fix TI TSB82AA2 regression since 2.6.35 Revert commit 54672386ccf36ffa21d1de8e75624af83f9b0eeb "firewire: ohci: fix up configuration of TI chips". It caused massive slow-down and data corruption with a TSB82AA2 based StarTech EC1394B2 ExpressCard and FireWire 800 harddisks. https://bugs.launchpad.net/ubuntu/+source/linux/+bug/657081 http://thread.gmane.org/gmane.linux.kernel.firewire.user/4013 The fact that some card EEPROMs do not program these enhancements may be related to TSB81BA3 phy chip errata, if not to bugs of TSB82AA2 itself. We could re-add these configuration steps, but only conditional on a whitelist of cards on which these enhancements bring a proven positive effect. Reported-and-tested-by: Eric Shattow Cc: Clemens Ladisch Cc: 2.6.35 Signed-off-by: Stefan Richter diff --git a/drivers/firewire/ohci.c b/drivers/firewire/ohci.c index 1b05896..9dcb17d 100644 --- a/drivers/firewire/ohci.c +++ b/drivers/firewire/ohci.c @@ -2840,7 +2840,7 @@ static int __devinit pci_probe(struct pci_dev *dev, const struct pci_device_id *ent) { struct fw_ohci *ohci; - u32 bus_options, max_receive, link_speed, version, link_enh; + u32 bus_options, max_receive, link_speed, version; u64 guid; int i, err, n_ir, n_it; size_t size; @@ -2894,23 +2894,6 @@ static int __devinit pci_probe(struct pci_dev *dev, if (param_quirks) ohci->quirks = param_quirks; - /* TI OHCI-Lynx and compatible: set recommended configuration bits. */ - if (dev->vendor == PCI_VENDOR_ID_TI) { - pci_read_config_dword(dev, PCI_CFG_TI_LinkEnh, &link_enh); - - /* adjust latency of ATx FIFO: use 1.7 KB threshold */ - link_enh &= ~TI_LinkEnh_atx_thresh_mask; - link_enh |= TI_LinkEnh_atx_thresh_1_7K; - - /* use priority arbitration for asynchronous responses */ - link_enh |= TI_LinkEnh_enab_unfair; - - /* required for aPhyEnhanceEnable to work */ - link_enh |= TI_LinkEnh_enab_accel; - - pci_write_config_dword(dev, PCI_CFG_TI_LinkEnh, link_enh); - } - ar_context_init(&ohci->ar_request_ctx, ohci, OHCI1394_AsReqRcvContextControlSet); diff --git a/drivers/firewire/ohci.h b/drivers/firewire/ohci.h index 0e6c5a4..ef5e733 100644 --- a/drivers/firewire/ohci.h +++ b/drivers/firewire/ohci.h @@ -155,12 +155,4 @@ #define OHCI1394_phy_tcode 0xe -/* TI extensions */ - -#define PCI_CFG_TI_LinkEnh 0xf4 -#define TI_LinkEnh_enab_accel 0x00000002 -#define TI_LinkEnh_enab_unfair 0x00000080 -#define TI_LinkEnh_atx_thresh_mask 0x00003000 -#define TI_LinkEnh_atx_thresh_1_7K 0x00001000 - #endif /* _FIREWIRE_OHCI_H */ -- cgit v0.10.2 From 6a2a11dbea5db417d200d38dda53c30a2e5603e0 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 14 Oct 2010 17:14:57 -0400 Subject: drm/radeon/kms: avivo cursor workaround applies to evergreen as well Fixes cursor corruption in certain cases. Signed-off-by: Alex Deucher Cc: stable@kernel.org Signed-off-by: Dave Airlie diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c index 5731fc9..3eef567 100644 --- a/drivers/gpu/drm/radeon/radeon_cursor.c +++ b/drivers/gpu/drm/radeon/radeon_cursor.c @@ -203,6 +203,7 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc, struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct radeon_device *rdev = crtc->dev->dev_private; int xorigin = 0, yorigin = 0; + int w = radeon_crtc->cursor_width; if (x < 0) xorigin = -x + 1; @@ -213,22 +214,7 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc, if (yorigin >= CURSOR_HEIGHT) yorigin = CURSOR_HEIGHT - 1; - radeon_lock_cursor(crtc, true); - if (ASIC_IS_DCE4(rdev)) { - /* cursors are offset into the total surface */ - x += crtc->x; - y += crtc->y; - DRM_DEBUG("x %d y %d c->x %d c->y %d\n", x, y, crtc->x, crtc->y); - - /* XXX: check if evergreen has the same issues as avivo chips */ - WREG32(EVERGREEN_CUR_POSITION + radeon_crtc->crtc_offset, - ((xorigin ? 0 : x) << 16) | - (yorigin ? 0 : y)); - WREG32(EVERGREEN_CUR_HOT_SPOT + radeon_crtc->crtc_offset, (xorigin << 16) | yorigin); - WREG32(EVERGREEN_CUR_SIZE + radeon_crtc->crtc_offset, - ((radeon_crtc->cursor_width - 1) << 16) | (radeon_crtc->cursor_height - 1)); - } else if (ASIC_IS_AVIVO(rdev)) { - int w = radeon_crtc->cursor_width; + if (ASIC_IS_AVIVO(rdev)) { int i = 0; struct drm_crtc *crtc_p; @@ -260,7 +246,17 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc, if (w <= 0) w = 1; } + } + radeon_lock_cursor(crtc, true); + if (ASIC_IS_DCE4(rdev)) { + WREG32(EVERGREEN_CUR_POSITION + radeon_crtc->crtc_offset, + ((xorigin ? 0 : x) << 16) | + (yorigin ? 0 : y)); + WREG32(EVERGREEN_CUR_HOT_SPOT + radeon_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(EVERGREEN_CUR_SIZE + radeon_crtc->crtc_offset, + ((w - 1) << 16) | (radeon_crtc->cursor_height - 1)); + } else if (ASIC_IS_AVIVO(rdev)) { WREG32(AVIVO_D1CUR_POSITION + radeon_crtc->crtc_offset, ((xorigin ? 0 : x) << 16) | (yorigin ? 0 : y)); -- cgit v0.10.2 From c5b4afec8e669bbe265daea24037fccfc42c4fe3 Mon Sep 17 00:00:00 2001 From: Jon Povey Date: Tue, 12 Oct 2010 13:47:05 +0900 Subject: i2c-davinci: Fix TX setup for more SoCs This patch is an improvement to 4bba0fd8d1c6d405df666e2573e1a1f917098be0 which got to mainline a little early. Sudhakar Rajashekhara explains that at least OMAP-L138 requires MDR mode settings before DXR for correct behaviour, so load MDR first with STT cleared and later load again with STT set. Tested on DM355 connected to Techwell TW2836 and Wolfson WM8985 Signed-off-by: Jon Povey Acked-by: Troy Kisky Tested-by: Sudhakar Rajashekhara Acked-by: Kevin Hilman Signed-off-by: Ben Dooks diff --git a/drivers/i2c/busses/i2c-davinci.c b/drivers/i2c/busses/i2c-davinci.c index b8feac5..5795c83 100644 --- a/drivers/i2c/busses/i2c-davinci.c +++ b/drivers/i2c/busses/i2c-davinci.c @@ -331,21 +331,16 @@ i2c_davinci_xfer_msg(struct i2c_adapter *adap, struct i2c_msg *msg, int stop) INIT_COMPLETION(dev->cmd_complete); dev->cmd_err = 0; - /* Take I2C out of reset, configure it as master and set the - * start bit */ - flag = DAVINCI_I2C_MDR_IRS | DAVINCI_I2C_MDR_MST | DAVINCI_I2C_MDR_STT; + /* Take I2C out of reset and configure it as master */ + flag = DAVINCI_I2C_MDR_IRS | DAVINCI_I2C_MDR_MST; /* if the slave address is ten bit address, enable XA bit */ if (msg->flags & I2C_M_TEN) flag |= DAVINCI_I2C_MDR_XA; if (!(msg->flags & I2C_M_RD)) flag |= DAVINCI_I2C_MDR_TRX; - if (stop) - flag |= DAVINCI_I2C_MDR_STP; - if (msg->len == 0) { + if (msg->len == 0) flag |= DAVINCI_I2C_MDR_RM; - flag &= ~DAVINCI_I2C_MDR_STP; - } /* Enable receive or transmit interrupts */ w = davinci_i2c_read_reg(dev, DAVINCI_I2C_IMR_REG); @@ -358,17 +353,28 @@ i2c_davinci_xfer_msg(struct i2c_adapter *adap, struct i2c_msg *msg, int stop) dev->terminate = 0; /* + * Write mode register first as needed for correct behaviour + * on OMAP-L138, but don't set STT yet to avoid a race with XRDY + * occuring before we have loaded DXR + */ + davinci_i2c_write_reg(dev, DAVINCI_I2C_MDR_REG, flag); + + /* * First byte should be set here, not after interrupt, * because transmit-data-ready interrupt can come before * NACK-interrupt during sending of previous message and * ICDXR may have wrong data + * It also saves us one interrupt, slightly faster */ if ((!(msg->flags & I2C_M_RD)) && dev->buf_len) { davinci_i2c_write_reg(dev, DAVINCI_I2C_DXR_REG, *dev->buf++); dev->buf_len--; } - /* write the data into mode register; start transmitting */ + /* Set STT to begin transmit now DXR is loaded */ + flag |= DAVINCI_I2C_MDR_STT; + if (stop && msg->len != 0) + flag |= DAVINCI_I2C_MDR_STP; davinci_i2c_write_reg(dev, DAVINCI_I2C_MDR_REG, flag); r = wait_for_completion_interruptible_timeout(&dev->cmd_complete, -- cgit v0.10.2 From e39428d53d080ad2615b772d7f99d2a70c2aaab2 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Mon, 21 Jun 2010 09:27:05 +0200 Subject: i2c-imx: do not allow interruptions when waiting for I2C to complete The i2c_imx_trx_complete() function is using wait_event_interruptible_timeout() to wait for the I2C controller to signal that it has completed an I2C bus operation. If the process that causes the I2C operation receives a signal, the wait will be interrupted, returning an error. It is better to let the I2C operation finished before handling the signal (i.e. returning into userspace). It is safe to use wait_event_timeout() instead, because the timeout will allow the process to exit if the I2C bus hangs. It's also better to allow the I2C operation to finish, because unacknowledged I2C operations can cause the I2C bus to hang. Signed-off-by: Marc Kleine-Budde Reviewed-by: Wolfram Sang Signed-off-by: Ben Dooks diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c index d1ff940..4c2a62b 100644 --- a/drivers/i2c/busses/i2c-imx.c +++ b/drivers/i2c/busses/i2c-imx.c @@ -159,15 +159,9 @@ static int i2c_imx_bus_busy(struct imx_i2c_struct *i2c_imx, int for_busy) static int i2c_imx_trx_complete(struct imx_i2c_struct *i2c_imx) { - int result; - - result = wait_event_interruptible_timeout(i2c_imx->queue, - i2c_imx->i2csr & I2SR_IIF, HZ / 10); + wait_event_timeout(i2c_imx->queue, i2c_imx->i2csr & I2SR_IIF, HZ / 10); - if (unlikely(result < 0)) { - dev_dbg(&i2c_imx->adapter.dev, "<%s> result < 0\n", __func__); - return result; - } else if (unlikely(!(i2c_imx->i2csr & I2SR_IIF))) { + if (unlikely(!(i2c_imx->i2csr & I2SR_IIF))) { dev_dbg(&i2c_imx->adapter.dev, "<%s> Timeout\n", __func__); return -ETIMEDOUT; } @@ -295,7 +289,7 @@ static irqreturn_t i2c_imx_isr(int irq, void *dev_id) i2c_imx->i2csr = temp; temp &= ~I2SR_IIF; writeb(temp, i2c_imx->base + IMX_I2C_I2SR); - wake_up_interruptible(&i2c_imx->queue); + wake_up(&i2c_imx->queue); return IRQ_HANDLED; } -- cgit v0.10.2 From 33027af637da3f69bd17488cc3e68493c9052a7d Mon Sep 17 00:00:00 2001 From: Andrea Gelmini Date: Sat, 16 Oct 2010 15:19:22 +0200 Subject: GFS2: fixed typo Signed-off-by: Andrea Gelmini Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/glock.h b/fs/gfs2/glock.h index 2bda191..db1c26d 100644 --- a/fs/gfs2/glock.h +++ b/fs/gfs2/glock.h @@ -215,7 +215,7 @@ void gfs2_glock_dq_uninit_m(unsigned int num_gh, struct gfs2_holder *ghs); void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...); /** - * gfs2_glock_nq_init - intialize a holder and enqueue it on a glock + * gfs2_glock_nq_init - initialize a holder and enqueue it on a glock * @gl: the glock * @state: the state we're requesting * @flags: the modifier flags -- cgit v0.10.2 From 907f27840985fe6a0c62e43cd4702c6e04b4bcc7 Mon Sep 17 00:00:00 2001 From: matt mooney Date: Mon, 27 Sep 2010 19:04:53 -0700 Subject: tracing/trivial: Remove cast from void* Unnecessary cast from void* in assignment. Signed-off-by: matt mooney Signed-off-by: Steven Rostedt diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 65fb077..ebd80d5 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1638,8 +1638,8 @@ ftrace_failures_open(struct inode *inode, struct file *file) ret = ftrace_avail_open(inode, file); if (!ret) { - m = (struct seq_file *)file->private_data; - iter = (struct ftrace_iterator *)m->private; + m = file->private_data; + iter = m->private; iter->flags = FTRACE_ITER_FAILURES; } diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 9ec59f5..001bcd2 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2196,7 +2196,7 @@ int tracing_open_generic(struct inode *inode, struct file *filp) static int tracing_release(struct inode *inode, struct file *file) { - struct seq_file *m = (struct seq_file *)file->private_data; + struct seq_file *m = file->private_data; struct trace_iterator *iter; int cpu; -- cgit v0.10.2 From a9d61173dc1cb63e660ae89e874e51ba4fd2f991 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Fri, 24 Sep 2010 17:41:02 +0200 Subject: tracing: Add proper check for irq_depth routines The check_irq_entry and check_irq_return could be called from graph event context. In such case there's no graph private data allocated. Adding checks to handle this case. Signed-off-by: Jiri Olsa LKML-Reference: <20100924154102.GB1818@jolsa.brq.redhat.com> [ Fixed some grammar in the comments ] Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index ef49e93..4c58ccc 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -888,12 +888,20 @@ check_irq_entry(struct trace_iterator *iter, u32 flags, unsigned long addr, int depth) { int cpu = iter->cpu; + int *depth_irq; struct fgraph_data *data = iter->private; - int *depth_irq = &(per_cpu_ptr(data->cpu_data, cpu)->depth_irq); - if (flags & TRACE_GRAPH_PRINT_IRQS) + /* + * If we are either displaying irqs, or we got called as + * a graph event and private data does not exist, + * then we bypass the irq check. + */ + if ((flags & TRACE_GRAPH_PRINT_IRQS) || + (!data)) return 0; + depth_irq = &(per_cpu_ptr(data->cpu_data, cpu)->depth_irq); + /* * We are inside the irq code */ @@ -926,12 +934,20 @@ static int check_irq_return(struct trace_iterator *iter, u32 flags, int depth) { int cpu = iter->cpu; + int *depth_irq; struct fgraph_data *data = iter->private; - int *depth_irq = &(per_cpu_ptr(data->cpu_data, cpu)->depth_irq); - if (flags & TRACE_GRAPH_PRINT_IRQS) + /* + * If we are either displaying irqs, or we got called as + * a graph event and private data does not exist, + * then we bypass the irq check. + */ + if ((flags & TRACE_GRAPH_PRINT_IRQS) || + (!data)) return 0; + depth_irq = &(per_cpu_ptr(data->cpu_data, cpu)->depth_irq); + /* * We are not inside the irq code. */ -- cgit v0.10.2 From 0a772620a2e21fb55a02f70fe38d4b5c3a5fbbbf Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Thu, 23 Sep 2010 14:00:52 +0200 Subject: tracing: Make graph related irqs/preemptsoff functions global Move trace_graph_function() and print_graph_headers_flags() functions to the trace_function_graph.c to be globaly available. Signed-off-by: Jiri Olsa LKML-Reference: <1285243253-7372-3-git-send-email-jolsa@redhat.com> Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index d39b3c5..9021f8c 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -343,6 +343,10 @@ void trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc); +void trace_graph_function(struct trace_array *tr, + unsigned long ip, + unsigned long parent_ip, + unsigned long flags, int pc); void trace_default_header(struct seq_file *m); void print_trace_header(struct seq_file *m, struct trace_iterator *iter); int trace_empty(struct trace_iterator *iter); diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 4c58ccc..6f8fe28 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -262,6 +262,35 @@ int trace_graph_thresh_entry(struct ftrace_graph_ent *trace) return trace_graph_entry(trace); } +static void +__trace_graph_function(struct trace_array *tr, + unsigned long ip, unsigned long flags, int pc) +{ + u64 time = trace_clock_local(); + struct ftrace_graph_ent ent = { + .func = ip, + .depth = 0, + }; + struct ftrace_graph_ret ret = { + .func = ip, + .depth = 0, + .calltime = time, + .rettime = time, + }; + + __trace_graph_entry(tr, &ent, flags, pc); + __trace_graph_return(tr, &ret, flags, pc); +} + +void +trace_graph_function(struct trace_array *tr, + unsigned long ip, unsigned long parent_ip, + unsigned long flags, int pc) +{ + __trace_graph_function(tr, parent_ip, flags, pc); + __trace_graph_function(tr, ip, flags, pc); +} + void __trace_graph_return(struct trace_array *tr, struct ftrace_graph_ret *trace, unsigned long flags, @@ -1179,7 +1208,7 @@ print_graph_comment(struct trace_seq *s, struct trace_entry *ent, enum print_line_t -print_graph_function_flags(struct trace_iterator *iter, u32 flags) +__print_graph_function_flags(struct trace_iterator *iter, u32 flags) { struct ftrace_graph_ent_entry *field; struct fgraph_data *data = iter->private; @@ -1242,7 +1271,18 @@ print_graph_function_flags(struct trace_iterator *iter, u32 flags) static enum print_line_t print_graph_function(struct trace_iterator *iter) { - return print_graph_function_flags(iter, tracer_flags.val); + return __print_graph_function_flags(iter, tracer_flags.val); +} + +enum print_line_t print_graph_function_flags(struct trace_iterator *iter, + u32 flags) +{ + if (trace_flags & TRACE_ITER_LATENCY_FMT) + flags |= TRACE_GRAPH_PRINT_DURATION; + else + flags |= TRACE_GRAPH_PRINT_ABS_TIME; + + return __print_graph_function_flags(iter, flags); } static enum print_line_t @@ -1274,7 +1314,7 @@ static void print_lat_header(struct seq_file *s, u32 flags) seq_printf(s, "#%.*s|||| / \n", size, spaces); } -void print_graph_headers_flags(struct seq_file *s, u32 flags) +static void __print_graph_headers_flags(struct seq_file *s, u32 flags) { int lat = trace_flags & TRACE_ITER_LATENCY_FMT; @@ -1315,6 +1355,23 @@ void print_graph_headers(struct seq_file *s) print_graph_headers_flags(s, tracer_flags.val); } +void print_graph_headers_flags(struct seq_file *s, u32 flags) +{ + struct trace_iterator *iter = s->private; + + if (trace_flags & TRACE_ITER_LATENCY_FMT) { + /* print nothing if the buffers are empty */ + if (trace_empty(iter)) + return; + + print_trace_header(s, iter); + flags |= TRACE_GRAPH_PRINT_DURATION; + } else + flags |= TRACE_GRAPH_PRINT_ABS_TIME; + + __print_graph_headers_flags(s, flags); +} + void graph_trace_open(struct trace_iterator *iter) { /* pid and depth on the last trace processed */ diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 73a6b06..4047e98 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -229,75 +229,33 @@ static void irqsoff_trace_close(struct trace_iterator *iter) static enum print_line_t irqsoff_print_line(struct trace_iterator *iter) { - u32 flags = GRAPH_TRACER_FLAGS; - - if (trace_flags & TRACE_ITER_LATENCY_FMT) - flags |= TRACE_GRAPH_PRINT_DURATION; - else - flags |= TRACE_GRAPH_PRINT_ABS_TIME; - /* * In graph mode call the graph tracer output function, * otherwise go with the TRACE_FN event handler */ if (is_graph()) - return print_graph_function_flags(iter, flags); + return print_graph_function_flags(iter, GRAPH_TRACER_FLAGS); return TRACE_TYPE_UNHANDLED; } static void irqsoff_print_header(struct seq_file *s) { - if (is_graph()) { - struct trace_iterator *iter = s->private; - u32 flags = GRAPH_TRACER_FLAGS; - - if (trace_flags & TRACE_ITER_LATENCY_FMT) { - /* print nothing if the buffers are empty */ - if (trace_empty(iter)) - return; - - print_trace_header(s, iter); - flags |= TRACE_GRAPH_PRINT_DURATION; - } else - flags |= TRACE_GRAPH_PRINT_ABS_TIME; - - print_graph_headers_flags(s, flags); - } else + if (is_graph()) + print_graph_headers_flags(s, GRAPH_TRACER_FLAGS); + else trace_default_header(s); } static void -trace_graph_function(struct trace_array *tr, - unsigned long ip, unsigned long flags, int pc) -{ - u64 time = trace_clock_local(); - struct ftrace_graph_ent ent = { - .func = ip, - .depth = 0, - }; - struct ftrace_graph_ret ret = { - .func = ip, - .depth = 0, - .calltime = time, - .rettime = time, - }; - - __trace_graph_entry(tr, &ent, flags, pc); - __trace_graph_return(tr, &ret, flags, pc); -} - -static void __trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc) { - if (!is_graph()) + if (is_graph()) + trace_graph_function(tr, ip, parent_ip, flags, pc); + else trace_function(tr, ip, parent_ip, flags, pc); - else { - trace_graph_function(tr, parent_ip, flags, pc); - trace_graph_function(tr, ip, flags, pc); - } } #else -- cgit v0.10.2 From 7495a5beaa22f190f4888aa8cbe4827c16575d0a Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Thu, 23 Sep 2010 14:00:53 +0200 Subject: tracing: Graph support for wakeup tracer Add function graph support for wakeup latency tracer. The graph output is enabled by setting the 'display-graph' trace option. Signed-off-by: Jiri Olsa LKML-Reference: <1285243253-7372-4-git-send-email-jolsa@redhat.com> Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 4086eae6..033510d 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -31,13 +31,33 @@ static int wakeup_rt; static arch_spinlock_t wakeup_lock = (arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED; +static void wakeup_reset(struct trace_array *tr); static void __wakeup_reset(struct trace_array *tr); +static int wakeup_graph_entry(struct ftrace_graph_ent *trace); +static void wakeup_graph_return(struct ftrace_graph_ret *trace); static int save_lat_flag; +#define TRACE_DISPLAY_GRAPH 1 + +static struct tracer_opt trace_opts[] = { +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + /* display latency trace as call graph */ + { TRACER_OPT(display-graph, TRACE_DISPLAY_GRAPH) }, +#endif + { } /* Empty entry */ +}; + +static struct tracer_flags tracer_flags = { + .val = 0, + .opts = trace_opts, +}; + +#define is_graph() (tracer_flags.val & TRACE_DISPLAY_GRAPH) + #ifdef CONFIG_FUNCTION_TRACER /* - * irqsoff uses its own tracer function to keep the overhead down: + * wakeup uses its own tracer function to keep the overhead down: */ static void wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) @@ -80,8 +100,191 @@ static struct ftrace_ops trace_ops __read_mostly = { .func = wakeup_tracer_call, }; + +static int start_func_tracer(int graph) +{ + int ret; + + if (!graph) + ret = register_ftrace_function(&trace_ops); + else + ret = register_ftrace_graph(&wakeup_graph_return, + &wakeup_graph_entry); + + if (!ret && tracing_is_enabled()) + tracer_enabled = 1; + else + tracer_enabled = 0; + + return ret; +} + +static void stop_func_tracer(int graph) +{ + tracer_enabled = 0; + + if (!graph) + unregister_ftrace_function(&trace_ops); + else + unregister_ftrace_graph(); +} + #endif /* CONFIG_FUNCTION_TRACER */ +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +static int wakeup_set_flag(u32 old_flags, u32 bit, int set) +{ + + if (!(bit & TRACE_DISPLAY_GRAPH)) + return -EINVAL; + + if (!(is_graph() ^ set)) + return 0; + + stop_func_tracer(!set); + + wakeup_reset(wakeup_trace); + tracing_max_latency = 0; + + return start_func_tracer(set); +} + +static int wakeup_graph_entry(struct ftrace_graph_ent *trace) +{ + struct trace_array *tr = wakeup_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu, pc, ret = 0; + + if (likely(!wakeup_task)) + return 0; + + pc = preempt_count(); + preempt_disable_notrace(); + + cpu = raw_smp_processor_id(); + if (cpu != wakeup_current_cpu) + goto out_enable; + + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + if (unlikely(disabled != 1)) + goto out; + + local_save_flags(flags); + ret = __trace_graph_entry(tr, trace, flags, pc); + +out: + atomic_dec(&data->disabled); + +out_enable: + preempt_enable_notrace(); + return ret; +} + +static void wakeup_graph_return(struct ftrace_graph_ret *trace) +{ + struct trace_array *tr = wakeup_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu, pc; + + if (likely(!wakeup_task)) + return; + + pc = preempt_count(); + preempt_disable_notrace(); + + cpu = raw_smp_processor_id(); + if (cpu != wakeup_current_cpu) + goto out_enable; + + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + if (unlikely(disabled != 1)) + goto out; + + local_save_flags(flags); + __trace_graph_return(tr, trace, flags, pc); + +out: + atomic_dec(&data->disabled); + +out_enable: + preempt_enable_notrace(); + return; +} + +static void wakeup_trace_open(struct trace_iterator *iter) +{ + if (is_graph()) + graph_trace_open(iter); +} + +static void wakeup_trace_close(struct trace_iterator *iter) +{ + if (iter->private) + graph_trace_close(iter); +} + +#define GRAPH_TRACER_FLAGS (TRACE_GRAPH_PRINT_PROC) + +static enum print_line_t wakeup_print_line(struct trace_iterator *iter) +{ + /* + * In graph mode call the graph tracer output function, + * otherwise go with the TRACE_FN event handler + */ + if (is_graph()) + return print_graph_function_flags(iter, GRAPH_TRACER_FLAGS); + + return TRACE_TYPE_UNHANDLED; +} + +static void wakeup_print_header(struct seq_file *s) +{ + if (is_graph()) + print_graph_headers_flags(s, GRAPH_TRACER_FLAGS); + else + trace_default_header(s); +} + +static void +__trace_function(struct trace_array *tr, + unsigned long ip, unsigned long parent_ip, + unsigned long flags, int pc) +{ + if (is_graph()) + trace_graph_function(tr, ip, parent_ip, flags, pc); + else + trace_function(tr, ip, parent_ip, flags, pc); +} +#else +#define __trace_function trace_function + +static int wakeup_set_flag(u32 old_flags, u32 bit, int set) +{ + return -EINVAL; +} + +static int wakeup_graph_entry(struct ftrace_graph_ent *trace) +{ + return -1; +} + +static enum print_line_t wakeup_print_line(struct trace_iterator *iter) +{ + return TRACE_TYPE_UNHANDLED; +} + +static void wakeup_graph_return(struct ftrace_graph_ret *trace) { } +static void wakeup_print_header(struct seq_file *s) { } +static void wakeup_trace_open(struct trace_iterator *iter) { } +static void wakeup_trace_close(struct trace_iterator *iter) { } +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + /* * Should this new latency be reported/recorded? */ @@ -152,7 +355,7 @@ probe_wakeup_sched_switch(void *ignore, /* The task we are waiting for is waking up */ data = wakeup_trace->data[wakeup_cpu]; - trace_function(wakeup_trace, CALLER_ADDR0, CALLER_ADDR1, flags, pc); + __trace_function(wakeup_trace, CALLER_ADDR0, CALLER_ADDR1, flags, pc); tracing_sched_switch_trace(wakeup_trace, prev, next, flags, pc); T0 = data->preempt_timestamp; @@ -252,7 +455,7 @@ probe_wakeup(void *ignore, struct task_struct *p, int success) * is not called by an assembly function (where as schedule is) * it should be safe to use it here. */ - trace_function(wakeup_trace, CALLER_ADDR1, CALLER_ADDR2, flags, pc); + __trace_function(wakeup_trace, CALLER_ADDR1, CALLER_ADDR2, flags, pc); out_locked: arch_spin_unlock(&wakeup_lock); @@ -303,12 +506,8 @@ static void start_wakeup_tracer(struct trace_array *tr) */ smp_wmb(); - register_ftrace_function(&trace_ops); - - if (tracing_is_enabled()) - tracer_enabled = 1; - else - tracer_enabled = 0; + if (start_func_tracer(is_graph())) + printk(KERN_ERR "failed to start wakeup tracer\n"); return; fail_deprobe_wake_new: @@ -320,7 +519,7 @@ fail_deprobe: static void stop_wakeup_tracer(struct trace_array *tr) { tracer_enabled = 0; - unregister_ftrace_function(&trace_ops); + stop_func_tracer(is_graph()); unregister_trace_sched_switch(probe_wakeup_sched_switch, NULL); unregister_trace_sched_wakeup_new(probe_wakeup, NULL); unregister_trace_sched_wakeup(probe_wakeup, NULL); @@ -379,9 +578,15 @@ static struct tracer wakeup_tracer __read_mostly = .start = wakeup_tracer_start, .stop = wakeup_tracer_stop, .print_max = 1, + .print_header = wakeup_print_header, + .print_line = wakeup_print_line, + .flags = &tracer_flags, + .set_flag = wakeup_set_flag, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_wakeup, #endif + .open = wakeup_trace_open, + .close = wakeup_trace_close, .use_max_tr = 1, }; @@ -394,9 +599,15 @@ static struct tracer wakeup_rt_tracer __read_mostly = .stop = wakeup_tracer_stop, .wait_pipe = poll_wait_pipe, .print_max = 1, + .print_header = wakeup_print_header, + .print_line = wakeup_print_line, + .flags = &tracer_flags, + .set_flag = wakeup_set_flag, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_wakeup, #endif + .open = wakeup_trace_open, + .close = wakeup_trace_close, .use_max_tr = 1, }; -- cgit v0.10.2 From 542181d3769d001c59cd17573dd4381e87d215f2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 5 Oct 2010 16:38:49 -0400 Subject: tracing: Use one prologue for the wakeup tracer function tracers The wakeup tracer has three types of function tracers. Normal function tracer, function graph entry, and function graph return. Each of these use a complex dance to prevent recursion and whether to trace the data or not (depending on the wake_task variable). This patch moves the duplicate code into a single routine, to prevent future mistakes with modifying duplicate complex code. Cc: Jiri Olsa Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 033510d..31689d2 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -56,43 +56,73 @@ static struct tracer_flags tracer_flags = { #define is_graph() (tracer_flags.val & TRACE_DISPLAY_GRAPH) #ifdef CONFIG_FUNCTION_TRACER + /* - * wakeup uses its own tracer function to keep the overhead down: + * Prologue for the wakeup function tracers. + * + * Returns 1 if it is OK to continue, and preemption + * is disabled and data->disabled is incremented. + * 0 if the trace is to be ignored, and preemption + * is not disabled and data->disabled is + * kept the same. + * + * Note, this function is also used outside this ifdef but + * inside the #ifdef of the function graph tracer below. + * This is OK, since the function graph tracer is + * dependent on the function tracer. */ -static void -wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) +static int +func_prolog_preempt_disable(struct trace_array *tr, + struct trace_array_cpu **data, + int *pc) { - struct trace_array *tr = wakeup_trace; - struct trace_array_cpu *data; - unsigned long flags; long disabled; int cpu; - int pc; if (likely(!wakeup_task)) - return; + return 0; - pc = preempt_count(); + *pc = preempt_count(); preempt_disable_notrace(); cpu = raw_smp_processor_id(); if (cpu != wakeup_current_cpu) goto out_enable; - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); + *data = tr->data[cpu]; + disabled = atomic_inc_return(&(*data)->disabled); if (unlikely(disabled != 1)) goto out; - local_irq_save(flags); + return 1; - trace_function(tr, ip, parent_ip, flags, pc); +out: + atomic_dec(&(*data)->disabled); + +out_enable: + preempt_enable_notrace(); + return 0; +} + +/* + * wakeup uses its own tracer function to keep the overhead down: + */ +static void +wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = wakeup_trace; + struct trace_array_cpu *data; + unsigned long flags; + int pc; + + if (!func_prolog_preempt_disable(tr, &data, &pc)) + return; + local_irq_save(flags); + trace_function(tr, ip, parent_ip, flags, pc); local_irq_restore(flags); - out: atomic_dec(&data->disabled); - out_enable: preempt_enable_notrace(); } @@ -154,32 +184,16 @@ static int wakeup_graph_entry(struct ftrace_graph_ent *trace) struct trace_array *tr = wakeup_trace; struct trace_array_cpu *data; unsigned long flags; - long disabled; - int cpu, pc, ret = 0; + int pc, ret = 0; - if (likely(!wakeup_task)) + if (!func_prolog_preempt_disable(tr, &data, &pc)) return 0; - pc = preempt_count(); - preempt_disable_notrace(); - - cpu = raw_smp_processor_id(); - if (cpu != wakeup_current_cpu) - goto out_enable; - - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - if (unlikely(disabled != 1)) - goto out; - local_save_flags(flags); ret = __trace_graph_entry(tr, trace, flags, pc); - -out: atomic_dec(&data->disabled); - -out_enable: preempt_enable_notrace(); + return ret; } @@ -188,31 +202,15 @@ static void wakeup_graph_return(struct ftrace_graph_ret *trace) struct trace_array *tr = wakeup_trace; struct trace_array_cpu *data; unsigned long flags; - long disabled; - int cpu, pc; + int pc; - if (likely(!wakeup_task)) + if (!func_prolog_preempt_disable(tr, &data, &pc)) return; - pc = preempt_count(); - preempt_disable_notrace(); - - cpu = raw_smp_processor_id(); - if (cpu != wakeup_current_cpu) - goto out_enable; - - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - if (unlikely(disabled != 1)) - goto out; - local_save_flags(flags); __trace_graph_return(tr, trace, flags, pc); - -out: atomic_dec(&data->disabled); -out_enable: preempt_enable_notrace(); return; } -- cgit v0.10.2 From 5e6d2b9cfa3a6e7fe62fc0135bc1bd778f5db564 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 5 Oct 2010 19:41:43 -0400 Subject: tracing: Use one prologue for the preempt irqs off tracer function tracers The preempt and irqsoff tracers have three types of function tracers. Normal function tracer, function graph entry, and function graph return. Each of these use a complex dance to prevent recursion and whether to trace the data or not (depending if interrupts are enabled or not). This patch moves the duplicate code into a single routine, to prevent future mistakes with modifying duplicate complex code. Cc: Jiri Olsa Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 4047e98..5cf8c60 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -87,14 +87,22 @@ static __cacheline_aligned_in_smp unsigned long max_sequence; #ifdef CONFIG_FUNCTION_TRACER /* - * irqsoff uses its own tracer function to keep the overhead down: + * Prologue for the preempt and irqs off function tracers. + * + * Returns 1 if it is OK to continue, and data->disabled is + * incremented. + * 0 if the trace is to be ignored, and data->disabled + * is kept the same. + * + * Note, this function is also used outside this ifdef but + * inside the #ifdef of the function graph tracer below. + * This is OK, since the function graph tracer is + * dependent on the function tracer. */ -static void -irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) +static int func_prolog_dec(struct trace_array *tr, + struct trace_array_cpu **data, + unsigned long *flags) { - struct trace_array *tr = irqsoff_trace; - struct trace_array_cpu *data; - unsigned long flags; long disabled; int cpu; @@ -106,18 +114,38 @@ irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) */ cpu = raw_smp_processor_id(); if (likely(!per_cpu(tracing_cpu, cpu))) - return; + return 0; - local_save_flags(flags); + local_save_flags(*flags); /* slight chance to get a false positive on tracing_cpu */ - if (!irqs_disabled_flags(flags)) - return; + if (!irqs_disabled_flags(*flags)) + return 0; - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); + *data = tr->data[cpu]; + disabled = atomic_inc_return(&(*data)->disabled); if (likely(disabled == 1)) - trace_function(tr, ip, parent_ip, flags, preempt_count()); + return 1; + + atomic_dec(&(*data)->disabled); + + return 0; +} + +/* + * irqsoff uses its own tracer function to keep the overhead down: + */ +static void +irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = irqsoff_trace; + struct trace_array_cpu *data; + unsigned long flags; + + if (!func_prolog_dec(tr, &data, &flags)) + return; + + trace_function(tr, ip, parent_ip, flags, preempt_count()); atomic_dec(&data->disabled); } @@ -155,30 +183,16 @@ static int irqsoff_graph_entry(struct ftrace_graph_ent *trace) struct trace_array *tr = irqsoff_trace; struct trace_array_cpu *data; unsigned long flags; - long disabled; int ret; - int cpu; int pc; - cpu = raw_smp_processor_id(); - if (likely(!per_cpu(tracing_cpu, cpu))) - return 0; - - local_save_flags(flags); - /* slight chance to get a false positive on tracing_cpu */ - if (!irqs_disabled_flags(flags)) + if (!func_prolog_dec(tr, &data, &flags)) return 0; - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - - if (likely(disabled == 1)) { - pc = preempt_count(); - ret = __trace_graph_entry(tr, trace, flags, pc); - } else - ret = 0; - + pc = preempt_count(); + ret = __trace_graph_entry(tr, trace, flags, pc); atomic_dec(&data->disabled); + return ret; } @@ -187,27 +201,13 @@ static void irqsoff_graph_return(struct ftrace_graph_ret *trace) struct trace_array *tr = irqsoff_trace; struct trace_array_cpu *data; unsigned long flags; - long disabled; - int cpu; int pc; - cpu = raw_smp_processor_id(); - if (likely(!per_cpu(tracing_cpu, cpu))) - return; - - local_save_flags(flags); - /* slight chance to get a false positive on tracing_cpu */ - if (!irqs_disabled_flags(flags)) + if (!func_prolog_dec(tr, &data, &flags)) return; - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - - if (likely(disabled == 1)) { - pc = preempt_count(); - __trace_graph_return(tr, trace, flags, pc); - } - + pc = preempt_count(); + __trace_graph_return(tr, trace, flags, pc); atomic_dec(&data->disabled); } -- cgit v0.10.2 From 78c89ba121221d9224a5747803d7fffe51cd6e44 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 5 Oct 2010 23:22:19 -0400 Subject: tracing: Remove parent recording in latency tracer graph options Even though the parent is recorded with the normal function tracing of the latency tracers (irqsoff and wakeup), the function graph recording is bogus. This is due to the function graph messing with the return stack. The latency tracers pass in as the parent CALLER_ADDR0, which works fine for plain function tracing. But this causes bogus output with the graph tracer: 3) -0 | d.s3. 0.000 us | return_to_handler(); 3) -0 | d.s3. 0.000 us | _raw_spin_unlock_irqrestore(); 3) -0 | d.s3. 0.000 us | return_to_handler(); 3) -0 | d.s3. 0.000 us | trace_hardirqs_on(); The "return_to_handle()" call is the trampoline of the function graph tracer, and is meaningless in this context. Cc: Jiri Olsa Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 6f8fe28..76b0598 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -287,7 +287,6 @@ trace_graph_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc) { - __trace_graph_function(tr, parent_ip, flags, pc); __trace_graph_function(tr, ip, flags, pc); } -- cgit v0.10.2 From 0a74a1df3c13655b66b1908d76f2159654216326 Mon Sep 17 00:00:00 2001 From: Daniel Mack Date: Mon, 18 Oct 2010 08:43:30 -0700 Subject: Input: evdev - fix Ooops in EVIOCGABS/EVIOCSABS This fixes a regression introduced by the dynamic allocation of absinfo for input devices. We need to bail out early for input devices which don't have absolute axis. [ 929.664303] Pid: 2989, comm: input Not tainted 2.6.36-rc8+ #14 MS-7260/MS-7260 [ 929.664318] EIP: 0060:[] EFLAGS: 00010246 CPU: 0 [ 929.664331] EIP is at evdev_ioctl+0x4f8/0x59f [ 929.664341] EAX: 00000040 EBX: 00000000 ECX: 00000006 EDX: f45a1efc [ 929.664355] ESI: 00000000 EDI: f45a1efc EBP: f45a1f24 ESP: f45a1eb8 [ 929.664369] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 [ 929.664402] f470da74 f6a30e78 f462c240 00000018 bfe4a260 00000000 f45b06fc 00000000 [ 929.664429] <0> 000000c4 b769d000 c3544620 f470da74 f45b06fc f45b06fc f45a1f38 c107dd1f [ 929.664458] <0> f4710b74 000000c4 00000000 00000000 00000000 0000029d 00000a74 f4710b74 [ 929.664500] [] ? handle_mm_fault+0x2be/0x59a [ 929.664513] [] ? evdev_ioctl+0x0/0x59f [ 929.664524] [] ? do_vfs_ioctl+0x494/0x4d9 [ 929.664538] [] ? up_read+0x16/0x29 [ 929.664550] [] ? do_page_fault+0x2ff/0x32d [ 929.664564] [] ? do_sys_open+0xc5/0xcf [ 929.664575] [] ? sys_ioctl+0x41/0x61 [ 929.664587] [] ? sysenter_do_call+0x12/0x36 [ 929.684570] ---[ end trace 11b83e923bd8f2bb ]--- Signed-off-by: Daniel Mack Signed-off-by: Dmitry Torokhov diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c index c908c5f..afb64cf 100644 --- a/drivers/input/evdev.c +++ b/drivers/input/evdev.c @@ -669,6 +669,9 @@ static long evdev_do_ioctl(struct file *file, unsigned int cmd, if ((_IOC_NR(cmd) & ~ABS_MAX) == _IOC_NR(EVIOCGABS(0))) { + if (!dev->absinfo) + return -EINVAL; + t = _IOC_NR(cmd) & ABS_MAX; abs = dev->absinfo[t]; @@ -684,6 +687,9 @@ static long evdev_do_ioctl(struct file *file, unsigned int cmd, if ((_IOC_NR(cmd) & ~ABS_MAX) == _IOC_NR(EVIOCSABS(0))) { + if (!dev->absinfo) + return -EINVAL; + t = _IOC_NR(cmd) & ABS_MAX; if (copy_from_user(&abs, p, min_t(size_t, -- cgit v0.10.2 From f9ce6eb5b6fa8cbcf0a0fb7c5f4203f94730fc52 Mon Sep 17 00:00:00 2001 From: Daniel Mack Date: Mon, 18 Oct 2010 08:43:50 -0700 Subject: Input: evdev - fix EVIOCSABS regression 448cd16 ("Input: evdev - rearrange ioctl handling") broke EVIOCSABS by checking for the wrong direction bit. Signed-off-by: Daniel Mack Reported-by: Sven Neumann Tested-by: Sven Neumann Signed-off-by: Dmitry Torokhov diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c index afb64cf..9ddafc3 100644 --- a/drivers/input/evdev.c +++ b/drivers/input/evdev.c @@ -683,7 +683,7 @@ static long evdev_do_ioctl(struct file *file, unsigned int cmd, } } - if (_IOC_DIR(cmd) == _IOC_READ) { + if (_IOC_DIR(cmd) == _IOC_WRITE) { if ((_IOC_NR(cmd) & ~ABS_MAX) == _IOC_NR(EVIOCSABS(0))) { -- cgit v0.10.2 From 062ab57b2f8439ab506645cf6f29432e7c61497c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 28 Sep 2010 18:50:17 +0100 Subject: MIPS: Don't block signals if we'd failed to setup a sigframe Signed-off-by: Al Viro Cc: linux-kernel@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/1696/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/kernel/signal.c b/arch/mips/kernel/signal.c index 2099d5a..b3273ae 100644 --- a/arch/mips/kernel/signal.c +++ b/arch/mips/kernel/signal.c @@ -575,6 +575,9 @@ static int handle_signal(unsigned long sig, siginfo_t *info, ret = abi->setup_frame(vdso + abi->signal_return_offset, ka, regs, sig, oldset); + if (ret) + return ret; + spin_lock_irq(¤t->sighand->siglock); sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask); if (!(ka->sa.sa_flags & SA_NODEFER)) -- cgit v0.10.2 From e5b377a8fb7b05f27647698ac739a0504cb2bf80 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 28 Sep 2010 18:50:27 +0100 Subject: MIPS: secure_computing, syscall audit: syscall number should in r2, not r0. As it is, audit_syscall_entry() and secure_computing() get the bogus value (0, in fact) Signed-off-by: Al Viro Cc: linux-kernel@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/1697/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/kernel/ptrace.c b/arch/mips/kernel/ptrace.c index c51b95f..c877733 100644 --- a/arch/mips/kernel/ptrace.c +++ b/arch/mips/kernel/ptrace.c @@ -536,7 +536,7 @@ asmlinkage void do_syscall_trace(struct pt_regs *regs, int entryexit) { /* do the secure computing check first */ if (!entryexit) - secure_computing(regs->regs[0]); + secure_computing(regs->regs[2]); if (unlikely(current->audit_context) && entryexit) audit_syscall_exit(AUDITSC_RESULT(regs->regs[2]), @@ -565,7 +565,7 @@ asmlinkage void do_syscall_trace(struct pt_regs *regs, int entryexit) out: if (unlikely(current->audit_context) && !entryexit) - audit_syscall_entry(audit_arch(), regs->regs[0], + audit_syscall_entry(audit_arch(), regs->regs[2], regs->regs[4], regs->regs[5], regs->regs[6], regs->regs[7]); } -- cgit v0.10.2 From 8f5a00eb422ed86e77bb8f67e08b9fe6d30f679a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 28 Sep 2010 18:50:37 +0100 Subject: MIPS: Sanitize restart logics Put the original syscall number into ->regs[0] when we leave syscall with error. Use it in restart logics. Everything else will have it 0 since we pass through SAVE_SOME on all the ways in. Note that in places like bad_stack and inllegal_syscall we leave it 0 - it's not restartable. Signed-off-by: Al Viro Cc: linux-kernel@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/1698/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/kernel/branch.c b/arch/mips/kernel/branch.c index 0176ed0..32103cc 100644 --- a/arch/mips/kernel/branch.c +++ b/arch/mips/kernel/branch.c @@ -40,7 +40,6 @@ int __compute_return_epc(struct pt_regs *regs) return -EFAULT; } - regs->regs[0] = 0; switch (insn.i_format.opcode) { /* * jr and jalr are in r_format format. diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall32-o32.S index 17202bb..73a061e 100644 --- a/arch/mips/kernel/scall32-o32.S +++ b/arch/mips/kernel/scall32-o32.S @@ -63,9 +63,9 @@ stack_done: sw t0, PT_R7(sp) # set error flag beqz t0, 1f + lw t1, PT_R2(sp) # syscall number negu v0 # error - sw v0, PT_R0(sp) # set flag for syscall - # restarting + sw t1, PT_R0(sp) # save it for syscall restarting 1: sw v0, PT_R2(sp) # result o32_syscall_exit: @@ -104,9 +104,9 @@ syscall_trace_entry: sw t0, PT_R7(sp) # set error flag beqz t0, 1f + lw t1, PT_R2(sp) # syscall number negu v0 # error - sw v0, PT_R0(sp) # set flag for syscall - # restarting + sw t1, PT_R0(sp) # save it for syscall restarting 1: sw v0, PT_R2(sp) # result j syscall_exit @@ -170,7 +170,6 @@ stackargs: */ bad_stack: negu v0 # error - sw v0, PT_R0(sp) sw v0, PT_R2(sp) li t0, 1 # set error flag sw t0, PT_R7(sp) diff --git a/arch/mips/kernel/scall64-64.S b/arch/mips/kernel/scall64-64.S index a8a6c59..eb0bb73 100644 --- a/arch/mips/kernel/scall64-64.S +++ b/arch/mips/kernel/scall64-64.S @@ -66,9 +66,9 @@ NESTED(handle_sys64, PT_SIZE, sp) sd t0, PT_R7(sp) # set error flag beqz t0, 1f + ld t1, PT_R2(sp) # syscall number dnegu v0 # error - sd v0, PT_R0(sp) # set flag for syscall - # restarting + sd t1, PT_R0(sp) # save it for syscall restarting 1: sd v0, PT_R2(sp) # result n64_syscall_exit: @@ -109,8 +109,9 @@ syscall_trace_entry: sd t0, PT_R7(sp) # set error flag beqz t0, 1f + ld t1, PT_R2(sp) # syscall number dnegu v0 # error - sd v0, PT_R0(sp) # set flag for syscall restarting + sd t1, PT_R0(sp) # save it for syscall restarting 1: sd v0, PT_R2(sp) # result j syscall_exit diff --git a/arch/mips/kernel/scall64-n32.S b/arch/mips/kernel/scall64-n32.S index a3d6613..4da3faf 100644 --- a/arch/mips/kernel/scall64-n32.S +++ b/arch/mips/kernel/scall64-n32.S @@ -65,8 +65,9 @@ NESTED(handle_sysn32, PT_SIZE, sp) sd t0, PT_R7(sp) # set error flag beqz t0, 1f + ld t1, PT_R2(sp) # syscall number dnegu v0 # error - sd v0, PT_R0(sp) # set flag for syscall restarting + sd t1, PT_R0(sp) # save it for syscall restarting 1: sd v0, PT_R2(sp) # result local_irq_disable # make sure need_resched and @@ -106,8 +107,9 @@ n32_syscall_trace_entry: sd t0, PT_R7(sp) # set error flag beqz t0, 1f + ld t1, PT_R2(sp) # syscall number dnegu v0 # error - sd v0, PT_R0(sp) # set flag for syscall restarting + sd t1, PT_R0(sp) # save it for syscall restarting 1: sd v0, PT_R2(sp) # result j syscall_exit diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall64-o32.S index 813689e..7ce0a36 100644 --- a/arch/mips/kernel/scall64-o32.S +++ b/arch/mips/kernel/scall64-o32.S @@ -93,8 +93,9 @@ NESTED(handle_sys, PT_SIZE, sp) sd t0, PT_R7(sp) # set error flag beqz t0, 1f + ld t1, PT_R2(sp) # syscall number dnegu v0 # error - sd v0, PT_R0(sp) # flag for syscall restarting + sd t1, PT_R0(sp) # save it for syscall restarting 1: sd v0, PT_R2(sp) # result o32_syscall_exit: @@ -142,8 +143,9 @@ trace_a_syscall: sd t0, PT_R7(sp) # set error flag beqz t0, 1f + ld t1, PT_R2(sp) # syscall number dnegu v0 # error - sd v0, PT_R0(sp) # set flag for syscall restarting + sd t1, PT_R0(sp) # save it for syscall restarting 1: sd v0, PT_R2(sp) # result j syscall_exit @@ -155,7 +157,6 @@ trace_a_syscall: */ bad_stack: dnegu v0 # error - sd v0, PT_R0(sp) sd v0, PT_R2(sp) li t0, 1 # set error flag sd t0, PT_R7(sp) diff --git a/arch/mips/kernel/signal.c b/arch/mips/kernel/signal.c index b3273ae..604f077 100644 --- a/arch/mips/kernel/signal.c +++ b/arch/mips/kernel/signal.c @@ -550,23 +550,26 @@ static int handle_signal(unsigned long sig, siginfo_t *info, struct mips_abi *abi = current->thread.abi; void *vdso = current->mm->context.vdso; - switch(regs->regs[0]) { - case ERESTART_RESTARTBLOCK: - case ERESTARTNOHAND: - regs->regs[2] = EINTR; - break; - case ERESTARTSYS: - if (!(ka->sa.sa_flags & SA_RESTART)) { + if (regs->regs[0]) { + switch(regs->regs[2]) { + case ERESTART_RESTARTBLOCK: + case ERESTARTNOHAND: regs->regs[2] = EINTR; break; + case ERESTARTSYS: + if (!(ka->sa.sa_flags & SA_RESTART)) { + regs->regs[2] = EINTR; + break; + } + /* fallthrough */ + case ERESTARTNOINTR: + regs->regs[7] = regs->regs[26]; + regs->regs[2] = regs->regs[0]; + regs->cp0_epc -= 4; } - /* fallthrough */ - case ERESTARTNOINTR: /* Userland will reload $v0. */ - regs->regs[7] = regs->regs[26]; - regs->cp0_epc -= 8; - } - regs->regs[0] = 0; /* Don't deal with this again. */ + regs->regs[0] = 0; /* Don't deal with this again. */ + } if (sig_uses_siginfo(ka)) ret = abi->setup_rt_frame(vdso + abi->rt_signal_return_offset, @@ -625,17 +628,13 @@ static void do_signal(struct pt_regs *regs) return; } - /* - * Who's code doesn't conform to the restartable syscall convention - * dies here!!! The li instruction, a single machine instruction, - * must directly be followed by the syscall instruction. - */ if (regs->regs[0]) { if (regs->regs[2] == ERESTARTNOHAND || regs->regs[2] == ERESTARTSYS || regs->regs[2] == ERESTARTNOINTR) { + regs->regs[2] = regs->regs[0]; regs->regs[7] = regs->regs[26]; - regs->cp0_epc -= 8; + regs->cp0_epc -= 4; } if (regs->regs[2] == ERESTART_RESTARTBLOCK) { regs->regs[2] = current->thread.abi->restart; diff --git a/arch/mips/kernel/unaligned.c b/arch/mips/kernel/unaligned.c index 69b039c..33d5a5c 100644 --- a/arch/mips/kernel/unaligned.c +++ b/arch/mips/kernel/unaligned.c @@ -109,8 +109,6 @@ static void emulate_load_store_insn(struct pt_regs *regs, unsigned long value; unsigned int res; - regs->regs[0] = 0; - /* * This load never faults. */ -- cgit v0.10.2 From 5b89c0044e048842245193b200b61e5f998abd9a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 28 Sep 2010 18:50:47 +0100 Subject: MIPS: Fix error values in case of bad_stack We want EFAULT, not - Signed-off-by: Al Viro Cc: linux-kernel@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/1699/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall32-o32.S index 73a061e..84b4179 100644 --- a/arch/mips/kernel/scall32-o32.S +++ b/arch/mips/kernel/scall32-o32.S @@ -169,7 +169,7 @@ stackargs: * We probably should handle this case a bit more drastic. */ bad_stack: - negu v0 # error + li v0, EFAULT sw v0, PT_R2(sp) li t0, 1 # set error flag sw t0, PT_R7(sp) diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall64-o32.S index 7ce0a36..d2bc285 100644 --- a/arch/mips/kernel/scall64-o32.S +++ b/arch/mips/kernel/scall64-o32.S @@ -156,7 +156,7 @@ trace_a_syscall: * The stackpointer for a call with more than 4 arguments is bad. */ bad_stack: - dnegu v0 # error + li v0, EFAULT sd v0, PT_R2(sp) li t0, 1 # set error flag sd t0, PT_R7(sp) -- cgit v0.10.2 From 4bfb8c5c87bd167640ef5789b22e058e339df804 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 28 Sep 2010 18:50:57 +0100 Subject: MIPS: do_sigaltstack() expects userland pointers o32 compat does the right thing, native and n32 compat do not... Signed-off-by: Al Viro Cc: linux-kernel@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-mips@linux-mips.org Patchwork: http://patchwork.linux-mips.org/patch/1700/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/kernel/signal.c b/arch/mips/kernel/signal.c index 604f077..5922342 100644 --- a/arch/mips/kernel/signal.c +++ b/arch/mips/kernel/signal.c @@ -390,7 +390,6 @@ asmlinkage void sys_rt_sigreturn(nabi_no_regargs struct pt_regs regs) { struct rt_sigframe __user *frame; sigset_t set; - stack_t st; int sig; frame = (struct rt_sigframe __user *) regs.regs[29]; @@ -411,11 +410,9 @@ asmlinkage void sys_rt_sigreturn(nabi_no_regargs struct pt_regs regs) else if (sig) force_sig(sig, current); - if (__copy_from_user(&st, &frame->rs_uc.uc_stack, sizeof(st))) - goto badframe; /* It is more difficult to avoid calling this function than to call it and ignore errors. */ - do_sigaltstack((stack_t __user *)&st, NULL, regs.regs[29]); + do_sigaltstack(&frame->rs_uc.uc_stack, NULL, regs.regs[29]); /* * Don't let your children do this ... diff --git a/arch/mips/kernel/signal_n32.c b/arch/mips/kernel/signal_n32.c index 2c5df81..ee24d81 100644 --- a/arch/mips/kernel/signal_n32.c +++ b/arch/mips/kernel/signal_n32.c @@ -109,6 +109,7 @@ asmlinkage int sysn32_rt_sigsuspend(nabi_no_regargs struct pt_regs regs) asmlinkage void sysn32_rt_sigreturn(nabi_no_regargs struct pt_regs regs) { struct rt_sigframe_n32 __user *frame; + mm_segment_t old_fs; sigset_t set; stack_t st; s32 sp; @@ -143,7 +144,11 @@ asmlinkage void sysn32_rt_sigreturn(nabi_no_regargs struct pt_regs regs) /* It is more difficult to avoid calling this function than to call it and ignore errors. */ + old_fs = get_fs(); + set_fs(KERNEL_DS); do_sigaltstack((stack_t __user *)&st, NULL, regs.regs[29]); + set_fs(old_fs); + /* * Don't let your children do this ... -- cgit v0.10.2 From ec3352925b74df8043131b535794f39e2516f614 Mon Sep 17 00:00:00 2001 From: Wu Zhangjin Date: Wed, 22 Sep 2010 14:03:57 +0800 Subject: MIPS: Remove all generated vmlinuz* files on "make clean" [Ralf: I changed the patch to explicitly list all files to be deleted out of paranoia.] Signed-off-by: Wu Zhangjin Patchwork: http://patchwork.linux-mips.org/patch/1590/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/boot/compressed/Makefile b/arch/mips/boot/compressed/Makefile index ed9bb70..4df4c77 100644 --- a/arch/mips/boot/compressed/Makefile +++ b/arch/mips/boot/compressed/Makefile @@ -105,4 +105,4 @@ OBJCOPYFLAGS_vmlinuz.srec := $(OBJCOPYFLAGS) -S -O srec vmlinuz.srec: vmlinuz $(call cmd,objcopy) -clean-files := $(objtree)/vmlinuz.* +clean-files := $(objtree)/vmlinuz $(objtree)/vmlinuz.{32,ecoff,bin,srec} -- cgit v0.10.2 From 063c13a5adbd0ab3110cbefcf9ba5aeb103a84ac Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Oct 2010 01:41:39 +0900 Subject: MIPS: 32-bit: Fix build failure in asm/fcntl.h CC security/integrity/ima/ima_fs.o In file included from linux/include/linux/fcntl.h:4:0, from linux/security/integrity/ima/ima_fs.c:18: linux/arch/mips/include/asm/fcntl.h:63:2: error: expected specifier-qualifier-list before 'off_t' make[3]: *** [security/integrity/ima/ima_fs.o] Error 1 make[2]: *** [security/integrity/ima/ima_fs.o] Error 2 make[1]: *** [sub-make] Error 2 make: *** [all] Error 2 Signed-off-by: Namhyung Kim Cc: linux-kernel@vger.kernel.org Patchwork: https://patchwork.linux-mips.org/patch/1715/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/include/asm/fcntl.h b/arch/mips/include/asm/fcntl.h index e482fe9..75edded 100644 --- a/arch/mips/include/asm/fcntl.h +++ b/arch/mips/include/asm/fcntl.h @@ -56,6 +56,7 @@ */ #ifdef CONFIG_32BIT +#include struct flock { short l_type; -- cgit v0.10.2 From a35bee8a2748abe9a40865f477aba0a9b85965fd Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Mon, 18 Oct 2010 12:55:21 +0900 Subject: MIPS: Enable ISA_DMA_API config to fix build failure Add ISA_DMA_API config item and select it when GENERIC_ISA_DMA enabled. This fixes build failure on allmodconfig like following: CC sound/isa/es18xx.o sound/isa/es18xx.c: In function 'snd_es18xx_playback1_prepare': sound/isa/es18xx.c:501:9: error: implicit declaration of function 'snd_dma_program' sound/isa/es18xx.c: In function 'snd_es18xx_playback_pointer': sound/isa/es18xx.c:818:3: error: implicit declaration of function 'snd_dma_pointer' make[3]: *** [sound/isa/es18xx.o] Error 1 make[2]: *** [sound/isa/es18xx.o] Error 2 make[1]: *** [sub-make] Error 2 make: *** [all] Error 2 Signed-off-by: Namhyung Kim Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org Patchwork: https://patchwork.linux-mips.org/patch/1717/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig index 3ad59dd..2d3cf38 100644 --- a/arch/mips/Kconfig +++ b/arch/mips/Kconfig @@ -880,11 +880,15 @@ config NO_IOPORT config GENERIC_ISA_DMA bool select ZONE_DMA if GENERIC_ISA_DMA_SUPPORT_BROKEN=n + select ISA_DMA_API config GENERIC_ISA_DMA_SUPPORT_BROKEN bool select GENERIC_ISA_DMA +config ISA_DMA_API + bool + config GENERIC_GPIO bool -- cgit v0.10.2 From 4924627423d5e286136ad2520f5be536345ae590 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Sun, 17 Oct 2010 21:46:10 +0200 Subject: sched: Unindent labels Labels should be on column 0. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 2111491..7f52283 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4891,7 +4891,7 @@ long sched_setaffinity(pid_t pid, const struct cpumask *in_mask) cpuset_cpus_allowed(p, cpus_allowed); cpumask_and(new_mask, in_mask, cpus_allowed); - again: +again: retval = set_cpus_allowed_ptr(p, new_mask); if (!retval) { @@ -8141,9 +8141,9 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent) return 1; - err_free_rq: +err_free_rq: kfree(cfs_rq); - err: +err: return 0; } @@ -8231,9 +8231,9 @@ int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent) return 1; - err_free_rq: +err_free_rq: kfree(rt_rq); - err: +err: return 0; } @@ -8591,7 +8591,7 @@ static int tg_set_bandwidth(struct task_group *tg, raw_spin_unlock(&rt_rq->rt_runtime_lock); } raw_spin_unlock_irq(&tg->rt_bandwidth.rt_runtime_lock); - unlock: +unlock: read_unlock(&tasklist_lock); mutex_unlock(&rt_constraints_mutex); diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index baef30f..ab77aa0 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1140,7 +1140,7 @@ static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu) for_each_leaf_rt_rq(rt_rq, rq) { array = &rt_rq->active; idx = sched_find_first_bit(array->bitmap); - next_idx: +next_idx: if (idx >= MAX_RT_PRIO) continue; if (next && next->prio < idx) @@ -1316,7 +1316,7 @@ static int push_rt_task(struct rq *rq) if (!next_task) return 0; - retry: +retry: if (unlikely(next_task == rq->curr)) { WARN_ON(1); return 0; @@ -1464,7 +1464,7 @@ static int pull_rt_task(struct rq *this_rq) * but possible) */ } - skip: +skip: double_unlock_balance(this_rq, src_rq); } -- cgit v0.10.2 From 34f971f6f7988be4d014eec3e3526bee6d007ffa Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 22 Sep 2010 13:53:15 +0200 Subject: sched: Create special class for stop/migrate work In order to separate the stop/migrate work thread from the SCHED_FIFO implementation, create a special class for it that is of higher priority than SCHED_FIFO itself. This currently solves a problem where cpu-hotplug consumes so much cpu-time that the SCHED_FIFO class gets throttled, but has the bandwidth replenishment timer pending on the now dead cpu. It is also required for when we add the planned deadline scheduling class above SCHED_FIFO, as the stop/migrate thread still needs to transcent those tasks. Tested-by: Heiko Carstens Signed-off-by: Peter Zijlstra LKML-Reference: <1285165776.2275.1022.camel@laptop> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 7f52283..5f64fed 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -486,7 +486,7 @@ struct rq { */ unsigned long nr_uninterruptible; - struct task_struct *curr, *idle; + struct task_struct *curr, *idle, *stop; unsigned long next_balance; struct mm_struct *prev_mm; @@ -1837,7 +1837,7 @@ static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu) static const struct sched_class rt_sched_class; -#define sched_class_highest (&rt_sched_class) +#define sched_class_highest (&stop_sched_class) #define for_each_class(class) \ for (class = sched_class_highest; class; class = class->next) @@ -1917,10 +1917,41 @@ static void deactivate_task(struct rq *rq, struct task_struct *p, int flags) #include "sched_idletask.c" #include "sched_fair.c" #include "sched_rt.c" +#include "sched_stoptask.c" #ifdef CONFIG_SCHED_DEBUG # include "sched_debug.c" #endif +void sched_set_stop_task(int cpu, struct task_struct *stop) +{ + struct sched_param param = { .sched_priority = MAX_RT_PRIO - 1 }; + struct task_struct *old_stop = cpu_rq(cpu)->stop; + + if (stop) { + /* + * Make it appear like a SCHED_FIFO task, its something + * userspace knows about and won't get confused about. + * + * Also, it will make PI more or less work without too + * much confusion -- but then, stop work should not + * rely on PI working anyway. + */ + sched_setscheduler_nocheck(stop, SCHED_FIFO, ¶m); + + stop->sched_class = &stop_sched_class; + } + + cpu_rq(cpu)->stop = stop; + + if (old_stop) { + /* + * Reset it back to a normal scheduling class so that + * it can die in pieces. + */ + old_stop->sched_class = &rt_sched_class; + } +} + /* * __normal_prio - return the priority that is based on the static prio */ @@ -3720,17 +3751,13 @@ pick_next_task(struct rq *rq) return p; } - class = sched_class_highest; - for ( ; ; ) { + for_each_class(class) { p = class->pick_next_task(rq); if (p) return p; - /* - * Will never be NULL as the idle class always - * returns a non-NULL p: - */ - class = class->next; } + + BUG(); /* the idle class will always have a runnable task */ } /* @@ -4659,6 +4686,15 @@ recheck: */ rq = __task_rq_lock(p); + /* + * Changing the policy of the stop threads its a very bad idea + */ + if (p == rq->stop) { + __task_rq_unlock(rq); + raw_spin_unlock_irqrestore(&p->pi_lock, flags); + return -EINVAL; + } + #ifdef CONFIG_RT_GROUP_SCHED if (user) { /* diff --git a/kernel/sched_stoptask.c b/kernel/sched_stoptask.c new file mode 100644 index 0000000..45bddc0 --- /dev/null +++ b/kernel/sched_stoptask.c @@ -0,0 +1,108 @@ +/* + * stop-task scheduling class. + * + * The stop task is the highest priority task in the system, it preempts + * everything and will be preempted by nothing. + * + * See kernel/stop_machine.c + */ + +#ifdef CONFIG_SMP +static int +select_task_rq_stop(struct rq *rq, struct task_struct *p, + int sd_flag, int flags) +{ + return task_cpu(p); /* stop tasks as never migrate */ +} +#endif /* CONFIG_SMP */ + +static void +check_preempt_curr_stop(struct rq *rq, struct task_struct *p, int flags) +{ + resched_task(rq->curr); /* we preempt everything */ +} + +static struct task_struct *pick_next_task_stop(struct rq *rq) +{ + struct task_struct *stop = rq->stop; + + if (stop && stop->state == TASK_RUNNING) + return stop; + + return NULL; +} + +static void +enqueue_task_stop(struct rq *rq, struct task_struct *p, int flags) +{ +} + +static void +dequeue_task_stop(struct rq *rq, struct task_struct *p, int flags) +{ +} + +static void yield_task_stop(struct rq *rq) +{ + BUG(); /* the stop task should never yield, its pointless. */ +} + +static void put_prev_task_stop(struct rq *rq, struct task_struct *prev) +{ +} + +static void task_tick_stop(struct rq *rq, struct task_struct *curr, int queued) +{ +} + +static void set_curr_task_stop(struct rq *rq) +{ +} + +static void switched_to_stop(struct rq *rq, struct task_struct *p, + int running) +{ + BUG(); /* its impossible to change to this class */ +} + +static void prio_changed_stop(struct rq *rq, struct task_struct *p, + int oldprio, int running) +{ + BUG(); /* how!?, what priority? */ +} + +static unsigned int +get_rr_interval_stop(struct rq *rq, struct task_struct *task) +{ + return 0; +} + +/* + * Simple, special scheduling class for the per-CPU stop tasks: + */ +static const struct sched_class stop_sched_class = { + .next = &rt_sched_class, + + .enqueue_task = enqueue_task_stop, + .dequeue_task = dequeue_task_stop, + .yield_task = yield_task_stop, + + .check_preempt_curr = check_preempt_curr_stop, + + .pick_next_task = pick_next_task_stop, + .put_prev_task = put_prev_task_stop, + +#ifdef CONFIG_SMP + .select_task_rq = select_task_rq_stop, +#endif + + .set_curr_task = set_curr_task_stop, + .task_tick = task_tick_stop, + + .get_rr_interval = get_rr_interval_stop, + + .prio_changed = prio_changed_stop, + .switched_to = switched_to_stop, + + /* no .task_new for stop tasks */ +}; diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c index 4372ccb..090c288 100644 --- a/kernel/stop_machine.c +++ b/kernel/stop_machine.c @@ -287,11 +287,12 @@ repeat: goto repeat; } +extern void sched_set_stop_task(int cpu, struct task_struct *stop); + /* manage stopper for a cpu, mostly lifted from sched migration thread mgmt */ static int __cpuinit cpu_stop_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { - struct sched_param param = { .sched_priority = MAX_RT_PRIO - 1 }; unsigned int cpu = (unsigned long)hcpu; struct cpu_stopper *stopper = &per_cpu(cpu_stopper, cpu); struct task_struct *p; @@ -304,13 +305,13 @@ static int __cpuinit cpu_stop_cpu_callback(struct notifier_block *nfb, cpu); if (IS_ERR(p)) return NOTIFY_BAD; - sched_setscheduler_nocheck(p, SCHED_FIFO, ¶m); get_task_struct(p); + kthread_bind(p, cpu); + sched_set_stop_task(cpu, p); stopper->thread = p; break; case CPU_ONLINE: - kthread_bind(stopper->thread, cpu); /* strictly unnecessary, as first user will wake it */ wake_up_process(stopper->thread); /* mark enabled */ @@ -325,6 +326,7 @@ static int __cpuinit cpu_stop_cpu_callback(struct notifier_block *nfb, { struct cpu_stop_work *work; + sched_set_stop_task(cpu, NULL); /* kill the stopper */ kthread_stop(stopper->thread); /* drain remaining works */ -- cgit v0.10.2 From 17bdcf949d03306b308c5fb694849cd35f119807 Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Mon, 11 Oct 2010 16:36:51 +0200 Subject: sched: Drop all load weight manipulation for RT tasks Load weights are for the CFS, they do not belong in the RT task. This makes all RT scheduling classes leave the CFS weights alone. This fixes a real bug as well: I noticed the following phonomena: a process elevated to SCHED_RR forks with SCHED_RESET_ON_FORK set, and the child is indeed SCHED_OTHER, and the niceval is indeed reset to 0. However the weight inserted by set_load_weight() remains at 0, giving the task insignificat priority. With this fix, the weight is reset to what the task had before being elevated to SCHED_RR/SCHED_FIFO. Cc: Lennart Poettering Cc: stable@kernel.org Signed-off-by: Linus Walleij Signed-off-by: Peter Zijlstra LKML-Reference: <1286807811-10568-1-git-send-email-linus.walleij@stericsson.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 5f64fed..728081a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1855,12 +1855,6 @@ static void dec_nr_running(struct rq *rq) static void set_load_weight(struct task_struct *p) { - if (task_has_rt_policy(p)) { - p->se.load.weight = 0; - p->se.load.inv_weight = WMULT_CONST; - return; - } - /* * SCHED_IDLE tasks get minimal weight: */ -- cgit v0.10.2 From 620162505e5d46bc4494b1761743e4b0b3bf8e16 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Tue, 5 Oct 2010 18:01:51 +0900 Subject: lockdep: Add improved subclass caching Current lockdep_map only caches one class with subclass == 0, and looks up hash table of classes when subclass != 0. It seems that this has no problem because the case of subclass != 0 is rare. But locks of struct rq are acquired with subclass == 1 when task migration is executed. Task migration is high frequent event, so I modified lockdep to cache subclasses. I measured the score of perf bench sched messaging. This patch has slightly but certain (order of milli seconds or 10 milli seconds) effect when lots of tasks are running. I'll show the result in the tail of this description. NR_LOCKDEP_CACHING_CLASSES specifies how many classes can be cached in the instances of lockdep_map. I discussed with Peter Zijlstra in LinuxCon Japan about this approach and he taught me that caching every subclasses(8) is cleary waste of memory. So number of cached classes should be configurable. === Score comparison of benchmarks === # "min" means best score, and "max" means worst score for i in `seq 1 10`; do ./perf bench -f simple sched messaging; done before: min: 0.565000, max: 0.583000, avg: 0.572500 after: min: 0.559000, max: 0.568000, avg: 0.563300 # with more processes for i in `seq 1 10`; do ./perf bench -f simple sched messaging -g 40; done before: min: 2.274000, max: 2.298000, avg: 2.286300 after: min: 2.242000, max: 2.270000, avg: 2.259700 Signed-off-by: Hitoshi Mitake Cc: Frederic Weisbecker Signed-off-by: Peter Zijlstra LKML-Reference: <1286269311-28336-2-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h index 06aed83..2186a64 100644 --- a/include/linux/lockdep.h +++ b/include/linux/lockdep.h @@ -32,6 +32,17 @@ extern int lock_stat; #define MAX_LOCKDEP_SUBCLASSES 8UL /* + * NR_LOCKDEP_CACHING_CLASSES ... Number of classes + * cached in the instance of lockdep_map + * + * Currently main class (subclass == 0) and signle depth subclass + * are cached in lockdep_map. This optimization is mainly targeting + * on rq->lock. double_rq_lock() acquires this highly competitive with + * single depth. + */ +#define NR_LOCKDEP_CACHING_CLASSES 2 + +/* * Lock-classes are keyed via unique addresses, by embedding the * lockclass-key into the kernel (or module) .data section. (For * static locks we use the lock address itself as the key.) @@ -138,7 +149,7 @@ void clear_lock_stats(struct lock_class *class); */ struct lockdep_map { struct lock_class_key *key; - struct lock_class *class_cache; + struct lock_class *class_cache[NR_LOCKDEP_CACHING_CLASSES]; const char *name; #ifdef CONFIG_LOCK_STAT int cpu; diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 84baa71..bc4d328 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -774,7 +774,9 @@ out_unlock_set: raw_local_irq_restore(flags); if (!subclass || force) - lock->class_cache = class; + lock->class_cache[0] = class; + else if (subclass < NR_LOCKDEP_CACHING_CLASSES) + lock->class_cache[subclass] = class; if (DEBUG_LOCKS_WARN_ON(class->subclass != subclass)) return NULL; @@ -2679,7 +2681,11 @@ static int mark_lock(struct task_struct *curr, struct held_lock *this, void lockdep_init_map(struct lockdep_map *lock, const char *name, struct lock_class_key *key, int subclass) { - lock->class_cache = NULL; + int i; + + for (i = 0; i < NR_LOCKDEP_CACHING_CLASSES; i++) + lock->class_cache[i] = NULL; + #ifdef CONFIG_LOCK_STAT lock->cpu = raw_smp_processor_id(); #endif @@ -2750,10 +2756,10 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, if (lock->key == &__lockdep_no_validate__) check = 1; - if (!subclass) - class = lock->class_cache; + if (subclass < NR_LOCKDEP_CACHING_CLASSES) + class = lock->class_cache[subclass]; /* - * Not cached yet or subclass? + * Not cached? */ if (unlikely(!class)) { class = register_lock_class(lock, subclass, 0); @@ -2918,7 +2924,7 @@ static int match_held_lock(struct held_lock *hlock, struct lockdep_map *lock) return 1; if (hlock->references) { - struct lock_class *class = lock->class_cache; + struct lock_class *class = lock->class_cache[0]; if (!class) class = look_up_lock_class(lock, 0); @@ -3559,7 +3565,12 @@ void lockdep_reset_lock(struct lockdep_map *lock) if (list_empty(head)) continue; list_for_each_entry_safe(class, next, head, hash_entry) { - if (unlikely(class == lock->class_cache)) { + int match = 0; + + for (j = 0; j < NR_LOCKDEP_CACHING_CLASSES; j++) + match |= class == lock->class_cache[j]; + + if (unlikely(match)) { if (debug_locks_off_graph_unlock()) WARN_ON(1); goto out_restore; -- cgit v0.10.2 From 4ba053c04aece1f4734056f21b751eee47ea3fb1 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Wed, 13 Oct 2010 17:30:26 +0900 Subject: lockdep: Check the depth of subclass Current look_up_lock_class() doesn't check the parameter "subclass". This rarely rises problems because the main caller of this function, register_lock_class(), checks it. But register_lock_class() is not the only function which calls look_up_lock_class(). lock_set_class() and its callees also call it. And lock_set_class() doesn't check this parameter. This will rise problems when the the value of subclass is larger than MAX_LOCKDEP_SUBCLASSES. Because the address (used as the key of class) caliculated with too large subclass has a probability to point another key in different lock_class_key. Of course this problem depends on the memory layout and occurs with really low probability. Signed-off-by: Hitoshi Mitake Cc: Dmitry Torokhov Cc: Vojtech Pavlik Cc: Frederic Weisbecker Signed-off-by: Peter Zijlstra LKML-Reference: <1286958626-986-1-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar diff --git a/kernel/lockdep.c b/kernel/lockdep.c index bc4d328..42ba65d 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -639,6 +639,16 @@ look_up_lock_class(struct lockdep_map *lock, unsigned int subclass) } #endif + if (unlikely(subclass >= MAX_LOCKDEP_SUBCLASSES)) { + debug_locks_off(); + printk(KERN_ERR + "BUG: looking up invalid subclass: %u\n", subclass); + printk(KERN_ERR + "turning off the locking correctness validator.\n"); + dump_stack(); + return NULL; + } + /* * Static locks do not have their class-keys yet - for them the key * is the lock object itself: @@ -2745,14 +2755,6 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, if (DEBUG_LOCKS_WARN_ON(!irqs_disabled())) return 0; - if (unlikely(subclass >= MAX_LOCKDEP_SUBCLASSES)) { - debug_locks_off(); - printk("BUG: MAX_LOCKDEP_SUBCLASSES too low!\n"); - printk("turning off the locking correctness validator.\n"); - dump_stack(); - return 0; - } - if (lock->key == &__lockdep_no_validate__) check = 1; -- cgit v0.10.2 From c530ccd9a1864a44a7ff35826681229ce9f2357a Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Fri, 15 Oct 2010 15:26:01 +0200 Subject: perf_events: Fix bogus context time tracking You can only call update_context_time() when the context is active, i.e., the thread it is attached to is still running. However, perf_event_read() can be called even when the context is inactive, e.g., user read() the counters. The call to update_context_time() must be conditioned on the status of the context, otherwise, bogus time_enabled, time_running may be returned. Here is an example on AMD64. The task program is an example from libpfm4. The -p prints deltas every 1s. $ task -p -e cpu_clk_unhalted sleep 5 2,266,610 cpu_clk_unhalted (0.00% scaling, ena=2,158,982, run=2,158,982) 0 cpu_clk_unhalted (0.00% scaling, ena=2,158,982, run=2,158,982) 0 cpu_clk_unhalted (0.00% scaling, ena=2,158,982, run=2,158,982) 0 cpu_clk_unhalted (0.00% scaling, ena=2,158,982, run=2,158,982) 0 cpu_clk_unhalted (0.00% scaling, ena=2,158,982, run=2,158,982) 5,242,358,071 cpu_clk_unhalted (99.95% scaling, ena=5,000,359,984, run=2,319,270) Whereas if you don't read deltas, e.g., no call to perf_event_read() until the process terminates: $ task -e cpu_clk_unhalted sleep 5 2,497,783 cpu_clk_unhalted (0.00% scaling, ena=2,376,899, run=2,376,899) Notice that time_enable, time_running are bogus in the first example causing bogus scaling. This patch fixes the problem, by conditionally calling update_context_time() in perf_event_read(). Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Cc: stable@kernel.org LKML-Reference: <4cb856dc.51edd80a.5ae0.38fb@mx.google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 1ec3916..e7eeba1 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1780,7 +1780,13 @@ static u64 perf_event_read(struct perf_event *event) unsigned long flags; raw_spin_lock_irqsave(&ctx->lock, flags); - update_context_time(ctx); + /* + * may read while context is not active + * (e.g., thread is blocked), in that case + * we cannot update context time + */ + if (ctx->is_active) + update_context_time(ctx); update_event_times(event); raw_spin_unlock_irqrestore(&ctx->lock, flags); } -- cgit v0.10.2 From ba0cef3d149ce4db293c572bf36ed352b11ce7b9 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Fri, 15 Oct 2010 15:15:01 +0200 Subject: perf_events: Fix bogus AMD64 generic TLB events PERF_COUNT_HW_CACHE_DTLB:READ:MISS had a bogus umask value of 0 which counts nothing. Needed to be 0x7 (to count all possibilities). PERF_COUNT_HW_CACHE_ITLB:READ:MISS had a bogus umask value of 0 which counts nothing. Needed to be 0x3 (to count all possibilities). Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Cc: Robert Richter Cc: # as far back as it applies LKML-Reference: <4cb85478.41e9d80a.44e2.3f00@mx.google.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c index c2897b7..46d5844 100644 --- a/arch/x86/kernel/cpu/perf_event_amd.c +++ b/arch/x86/kernel/cpu/perf_event_amd.c @@ -52,7 +52,7 @@ static __initconst const u64 amd_hw_cache_event_ids [ C(DTLB) ] = { [ C(OP_READ) ] = { [ C(RESULT_ACCESS) ] = 0x0040, /* Data Cache Accesses */ - [ C(RESULT_MISS) ] = 0x0046, /* L1 DTLB and L2 DLTB Miss */ + [ C(RESULT_MISS) ] = 0x0746, /* L1_DTLB_AND_L2_DLTB_MISS.ALL */ }, [ C(OP_WRITE) ] = { [ C(RESULT_ACCESS) ] = 0, @@ -66,7 +66,7 @@ static __initconst const u64 amd_hw_cache_event_ids [ C(ITLB) ] = { [ C(OP_READ) ] = { [ C(RESULT_ACCESS) ] = 0x0080, /* Instruction fecthes */ - [ C(RESULT_MISS) ] = 0x0085, /* Instr. fetch ITLB misses */ + [ C(RESULT_MISS) ] = 0x0385, /* L1_ITLB_AND_L2_ITLB_MISS.ALL */ }, [ C(OP_WRITE) ] = { [ C(RESULT_ACCESS) ] = -1, -- cgit v0.10.2 From 8e5fc1a7320baf6076391607515dceb61319b36a Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Fri, 15 Oct 2010 16:54:01 +0200 Subject: perf_events: Fix transaction recovery in group_sched_in() The group_sched_in() function uses a transactional approach to schedule a group of events. In a group, either all events can be scheduled or none are. To schedule each event in, the function calls event_sched_in(). In case of error, event_sched_out() is called on each event in the group. The problem is that event_sched_out() does not completely cancel the effects of event_sched_in(). Furthermore event_sched_out() changes the state of the event as if it had run which is not true is this particular case. Those inconsistencies impact time tracking fields and may lead to events in a group not all reporting the same time_enabled and time_running values. This is demonstrated with the example below: $ task -eunhalted_core_cycles,baclears,baclears -e unhalted_core_cycles,baclears,baclears sleep 5 1946101 unhalted_core_cycles (32.85% scaling, ena=829181, run=556827) 11423 baclears (32.85% scaling, ena=829181, run=556827) 7671 baclears (0.00% scaling, ena=556827, run=556827) 2250443 unhalted_core_cycles (57.83% scaling, ena=962822, run=405995) 11705 baclears (57.83% scaling, ena=962822, run=405995) 11705 baclears (57.83% scaling, ena=962822, run=405995) Notice that in the first group, the last baclears event does not report the same timings as its siblings. This issue comes from the fact that tstamp_stopped is updated by event_sched_out() as if the event had actually run. To solve the issue, we must ensure that, in case of error, there is no change in the event state whatsoever. That means timings must remain as they were when entering group_sched_in(). To do this we defer updating tstamp_running until we know the transaction succeeded. Therefore, we have split event_sched_in() in two parts separating the update to tstamp_running. Similarly, in case of error, we do not want to update tstamp_stopped. Therefore, we have split event_sched_out() in two parts separating the update to tstamp_stopped. With this patch, we now get the following output: $ task -eunhalted_core_cycles,baclears,baclears -e unhalted_core_cycles,baclears,baclears sleep 5 2492050 unhalted_core_cycles (71.75% scaling, ena=1093330, run=308841) 11243 baclears (71.75% scaling, ena=1093330, run=308841) 11243 baclears (71.75% scaling, ena=1093330, run=308841) 1852746 unhalted_core_cycles (0.00% scaling, ena=784489, run=784489) 9253 baclears (0.00% scaling, ena=784489, run=784489) 9253 baclears (0.00% scaling, ena=784489, run=784489) Note that the uneven timing between groups is a side effect of the process spending most of its time sleeping, i.e., not enough event rotations (but that's a separate issue). Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra LKML-Reference: <4cb86b4c.41e9d80a.44e9.3e19@mx.google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index e7eeba1..634f86a4 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -412,8 +412,8 @@ event_filter_match(struct perf_event *event) return event->cpu == -1 || event->cpu == smp_processor_id(); } -static void -event_sched_out(struct perf_event *event, +static int +__event_sched_out(struct perf_event *event, struct perf_cpu_context *cpuctx, struct perf_event_context *ctx) { @@ -432,14 +432,13 @@ event_sched_out(struct perf_event *event, } if (event->state != PERF_EVENT_STATE_ACTIVE) - return; + return 0; event->state = PERF_EVENT_STATE_INACTIVE; if (event->pending_disable) { event->pending_disable = 0; event->state = PERF_EVENT_STATE_OFF; } - event->tstamp_stopped = ctx->time; event->pmu->del(event, 0); event->oncpu = -1; @@ -448,6 +447,19 @@ event_sched_out(struct perf_event *event, ctx->nr_active--; if (event->attr.exclusive || !cpuctx->active_oncpu) cpuctx->exclusive = 0; + return 1; +} + +static void +event_sched_out(struct perf_event *event, + struct perf_cpu_context *cpuctx, + struct perf_event_context *ctx) +{ + int ret; + + ret = __event_sched_out(event, cpuctx, ctx); + if (ret) + event->tstamp_stopped = ctx->time; } static void @@ -647,7 +659,7 @@ retry: } static int -event_sched_in(struct perf_event *event, +__event_sched_in(struct perf_event *event, struct perf_cpu_context *cpuctx, struct perf_event_context *ctx) { @@ -667,8 +679,6 @@ event_sched_in(struct perf_event *event, return -EAGAIN; } - event->tstamp_running += ctx->time - event->tstamp_stopped; - if (!is_software_event(event)) cpuctx->active_oncpu++; ctx->nr_active++; @@ -679,6 +689,35 @@ event_sched_in(struct perf_event *event, return 0; } +static inline int +event_sched_in(struct perf_event *event, + struct perf_cpu_context *cpuctx, + struct perf_event_context *ctx) +{ + int ret = __event_sched_in(event, cpuctx, ctx); + if (ret) + return ret; + event->tstamp_running += ctx->time - event->tstamp_stopped; + return 0; +} + +static void +group_commit_event_sched_in(struct perf_event *group_event, + struct perf_cpu_context *cpuctx, + struct perf_event_context *ctx) +{ + struct perf_event *event; + u64 now = ctx->time; + + group_event->tstamp_running += now - group_event->tstamp_stopped; + /* + * Schedule in siblings as one group (if any): + */ + list_for_each_entry(event, &group_event->sibling_list, group_entry) { + event->tstamp_running += now - event->tstamp_stopped; + } +} + static int group_sched_in(struct perf_event *group_event, struct perf_cpu_context *cpuctx, @@ -692,7 +731,13 @@ group_sched_in(struct perf_event *group_event, pmu->start_txn(pmu); - if (event_sched_in(group_event, cpuctx, ctx)) { + /* + * use __event_sched_in() to delay updating tstamp_running + * until the transaction is committed. In case of failure + * we will keep an unmodified tstamp_running which is a + * requirement to get correct timing information + */ + if (__event_sched_in(group_event, cpuctx, ctx)) { pmu->cancel_txn(pmu); return -EAGAIN; } @@ -701,26 +746,31 @@ group_sched_in(struct perf_event *group_event, * Schedule in siblings as one group (if any): */ list_for_each_entry(event, &group_event->sibling_list, group_entry) { - if (event_sched_in(event, cpuctx, ctx)) { + if (__event_sched_in(event, cpuctx, ctx)) { partial_group = event; goto group_error; } } - if (!pmu->commit_txn(pmu)) + if (!pmu->commit_txn(pmu)) { + /* commit tstamp_running */ + group_commit_event_sched_in(group_event, cpuctx, ctx); return 0; - + } group_error: /* * Groups can be scheduled in as one unit only, so undo any * partial group before returning: + * + * use __event_sched_out() to avoid updating tstamp_stopped + * because the event never actually ran */ list_for_each_entry(event, &group_event->sibling_list, group_entry) { if (event == partial_group) break; - event_sched_out(event, cpuctx, ctx); + __event_sched_out(event, cpuctx, ctx); } - event_sched_out(group_event, cpuctx, ctx); + __event_sched_out(group_event, cpuctx, ctx); pmu->cancel_txn(pmu); -- cgit v0.10.2 From e360adbe29241a0194e10e20595360dd7b98a2b3 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 14:01:34 +0800 Subject: irq_work: Add generic hardirq context callbacks Provide a mechanism that allows running code in IRQ context. It is most useful for NMI code that needs to interact with the rest of the system -- like wakeup a task to drain buffers. Perf currently has such a mechanism, so extract that and provide it as a generic feature, independent of perf so that others may also benefit. The IRQ context callback is generated through self-IPIs where possible, or on architectures like powerpc the decrementer (the built-in timer facility) is set to generate an interrupt immediately. Architectures that don't have anything like this get to do with a callback from the timer tick. These architectures can call irq_work_run() at the tail of any IRQ handlers that might enqueue such work (like the perf IRQ handler) to avoid undue latencies in processing the work. Signed-off-by: Peter Zijlstra Acked-by: Kyle McMartin Acked-by: Martin Schwidefsky [ various fixes ] Signed-off-by: Huang Ying LKML-Reference: <1287036094.7768.291.camel@yhuang-dev> Signed-off-by: Ingo Molnar diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig index b9647bb..d04ccd7 100644 --- a/arch/alpha/Kconfig +++ b/arch/alpha/Kconfig @@ -9,6 +9,7 @@ config ALPHA select HAVE_IDE select HAVE_OPROFILE select HAVE_SYSCALL_WRAPPERS + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select HAVE_DMA_ATTRS help diff --git a/arch/alpha/include/asm/perf_event.h b/arch/alpha/include/asm/perf_event.h index 4157cd3..fe792ca 100644 --- a/arch/alpha/include/asm/perf_event.h +++ b/arch/alpha/include/asm/perf_event.h @@ -1,11 +1,6 @@ #ifndef __ASM_ALPHA_PERF_EVENT_H #define __ASM_ALPHA_PERF_EVENT_H -/* Alpha only supports software events through this interface. */ -extern void set_perf_event_pending(void); - -#define PERF_EVENT_INDEX_OFFSET 0 - #ifdef CONFIG_PERF_EVENTS extern void init_hw_perf_events(void); #else diff --git a/arch/alpha/kernel/time.c b/arch/alpha/kernel/time.c index 396af17..0f1d849 100644 --- a/arch/alpha/kernel/time.c +++ b/arch/alpha/kernel/time.c @@ -41,7 +41,7 @@ #include #include #include -#include +#include #include #include @@ -83,25 +83,25 @@ static struct { unsigned long est_cycle_freq; -#ifdef CONFIG_PERF_EVENTS +#ifdef CONFIG_IRQ_WORK -DEFINE_PER_CPU(u8, perf_event_pending); +DEFINE_PER_CPU(u8, irq_work_pending); -#define set_perf_event_pending_flag() __get_cpu_var(perf_event_pending) = 1 -#define test_perf_event_pending() __get_cpu_var(perf_event_pending) -#define clear_perf_event_pending() __get_cpu_var(perf_event_pending) = 0 +#define set_irq_work_pending_flag() __get_cpu_var(irq_work_pending) = 1 +#define test_irq_work_pending() __get_cpu_var(irq_work_pending) +#define clear_irq_work_pending() __get_cpu_var(irq_work_pending) = 0 -void set_perf_event_pending(void) +void set_irq_work_pending(void) { - set_perf_event_pending_flag(); + set_irq_work_pending_flag(); } -#else /* CONFIG_PERF_EVENTS */ +#else /* CONFIG_IRQ_WORK */ -#define test_perf_event_pending() 0 -#define clear_perf_event_pending() +#define test_irq_work_pending() 0 +#define clear_irq_work_pending() -#endif /* CONFIG_PERF_EVENTS */ +#endif /* CONFIG_IRQ_WORK */ static inline __u32 rpcc(void) @@ -191,9 +191,9 @@ irqreturn_t timer_interrupt(int irq, void *dev) write_sequnlock(&xtime_lock); - if (test_perf_event_pending()) { - clear_perf_event_pending(); - perf_event_do_pending(); + if (test_irq_work_pending()) { + clear_irq_work_pending(); + irq_work_run(); } #ifndef CONFIG_SMP diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 88c97bc..7c0dfcc 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -23,6 +23,7 @@ config ARM select HAVE_KERNEL_GZIP select HAVE_KERNEL_LZO select HAVE_KERNEL_LZMA + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC select HAVE_REGS_AND_STACK_ACCESS_API diff --git a/arch/arm/include/asm/perf_event.h b/arch/arm/include/asm/perf_event.h index b5799a3..c4aa4e8 100644 --- a/arch/arm/include/asm/perf_event.h +++ b/arch/arm/include/asm/perf_event.h @@ -12,18 +12,6 @@ #ifndef __ARM_PERF_EVENT_H__ #define __ARM_PERF_EVENT_H__ -/* - * NOP: on *most* (read: all supported) ARM platforms, the performance - * counter interrupts are regular interrupts and not an NMI. This - * means that when we receive the interrupt we can call - * perf_event_do_pending() that handles all of the work with - * interrupts disabled. - */ -static inline void -set_perf_event_pending(void) -{ -} - /* ARM performance counters start from 1 (in the cp15 accesses) so use the * same indexes here for consistency. */ #define PERF_EVENT_INDEX_OFFSET 1 diff --git a/arch/arm/kernel/perf_event.c b/arch/arm/kernel/perf_event.c index 6cc6521..49643b1 100644 --- a/arch/arm/kernel/perf_event.c +++ b/arch/arm/kernel/perf_event.c @@ -1092,7 +1092,7 @@ armv6pmu_handle_irq(int irq_num, * platforms that can have the PMU interrupts raised as an NMI, this * will not work. */ - perf_event_do_pending(); + irq_work_run(); return IRQ_HANDLED; } @@ -2068,7 +2068,7 @@ static irqreturn_t armv7pmu_handle_irq(int irq_num, void *dev) * platforms that can have the PMU interrupts raised as an NMI, this * will not work. */ - perf_event_do_pending(); + irq_work_run(); return IRQ_HANDLED; } @@ -2436,7 +2436,7 @@ xscale1pmu_handle_irq(int irq_num, void *dev) armpmu->disable(hwc, idx); } - perf_event_do_pending(); + irq_work_run(); /* * Re-enable the PMU. @@ -2763,7 +2763,7 @@ xscale2pmu_handle_irq(int irq_num, void *dev) armpmu->disable(hwc, idx); } - perf_event_do_pending(); + irq_work_run(); /* * Re-enable the PMU. diff --git a/arch/frv/Kconfig b/arch/frv/Kconfig index 16399bd..0f2417d 100644 --- a/arch/frv/Kconfig +++ b/arch/frv/Kconfig @@ -7,6 +7,7 @@ config FRV default y select HAVE_IDE select HAVE_ARCH_TRACEHOOK + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS config ZONE_DMA diff --git a/arch/frv/lib/Makefile b/arch/frv/lib/Makefile index f470975..4ff2fb1 100644 --- a/arch/frv/lib/Makefile +++ b/arch/frv/lib/Makefile @@ -5,4 +5,4 @@ lib-y := \ __ashldi3.o __lshrdi3.o __muldi3.o __ashrdi3.o __negdi2.o __ucmpdi2.o \ checksum.o memcpy.o memset.o atomic-ops.o atomic64-ops.o \ - outsl_ns.o outsl_sw.o insl_ns.o insl_sw.o cache.o perf_event.o + outsl_ns.o outsl_sw.o insl_ns.o insl_sw.o cache.o diff --git a/arch/frv/lib/perf_event.c b/arch/frv/lib/perf_event.c deleted file mode 100644 index 9ac5acf..0000000 --- a/arch/frv/lib/perf_event.c +++ /dev/null @@ -1,19 +0,0 @@ -/* Performance event handling - * - * Copyright (C) 2009 Red Hat, Inc. All Rights Reserved. - * Written by David Howells (dhowells@redhat.com) - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public Licence - * as published by the Free Software Foundation; either version - * 2 of the Licence, or (at your option) any later version. - */ - -#include - -/* - * mark the performance event as pending - */ -void set_perf_event_pending(void) -{ -} diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig index 907417d..79a04a9 100644 --- a/arch/parisc/Kconfig +++ b/arch/parisc/Kconfig @@ -16,6 +16,7 @@ config PARISC select RTC_DRV_GENERIC select INIT_ALL_POSSIBLE select BUG + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select GENERIC_ATOMIC64 if !64BIT help diff --git a/arch/parisc/include/asm/perf_event.h b/arch/parisc/include/asm/perf_event.h index cc146427..1e0fd8b 100644 --- a/arch/parisc/include/asm/perf_event.h +++ b/arch/parisc/include/asm/perf_event.h @@ -1,7 +1,6 @@ #ifndef __ASM_PARISC_PERF_EVENT_H #define __ASM_PARISC_PERF_EVENT_H -/* parisc only supports software events through this interface. */ -static inline void set_perf_event_pending(void) { } +/* Empty, just to avoid compiling error */ #endif /* __ASM_PARISC_PERF_EVENT_H */ diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 631e5a0..4b1e521 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -138,6 +138,7 @@ config PPC select HAVE_OPROFILE select HAVE_SYSCALL_WRAPPERS if PPC64 select GENERIC_ATOMIC64 if PPC32 + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select HAVE_REGS_AND_STACK_ACCESS_API select HAVE_HW_BREAKPOINT if PERF_EVENTS && PPC_BOOK3S_64 diff --git a/arch/powerpc/include/asm/paca.h b/arch/powerpc/include/asm/paca.h index 1ff6662..9b287fd 100644 --- a/arch/powerpc/include/asm/paca.h +++ b/arch/powerpc/include/asm/paca.h @@ -129,7 +129,7 @@ struct paca_struct { u8 soft_enabled; /* irq soft-enable flag */ u8 hard_enabled; /* set if irqs are enabled in MSR */ u8 io_sync; /* writel() needs spin_unlock sync */ - u8 perf_event_pending; /* PM interrupt while soft-disabled */ + u8 irq_work_pending; /* IRQ_WORK interrupt while soft-disable */ /* Stuff for accurate time accounting */ u64 user_time; /* accumulated usermode TB ticks */ diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 8533b3b..54888eb 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -53,7 +53,7 @@ #include #include #include -#include +#include #include #include @@ -493,60 +493,60 @@ void __init iSeries_time_init_early(void) } #endif /* CONFIG_PPC_ISERIES */ -#ifdef CONFIG_PERF_EVENTS +#ifdef CONFIG_IRQ_WORK /* * 64-bit uses a byte in the PACA, 32-bit uses a per-cpu variable... */ #ifdef CONFIG_PPC64 -static inline unsigned long test_perf_event_pending(void) +static inline unsigned long test_irq_work_pending(void) { unsigned long x; asm volatile("lbz %0,%1(13)" : "=r" (x) - : "i" (offsetof(struct paca_struct, perf_event_pending))); + : "i" (offsetof(struct paca_struct, irq_work_pending))); return x; } -static inline void set_perf_event_pending_flag(void) +static inline void set_irq_work_pending_flag(void) { asm volatile("stb %0,%1(13)" : : "r" (1), - "i" (offsetof(struct paca_struct, perf_event_pending))); + "i" (offsetof(struct paca_struct, irq_work_pending))); } -static inline void clear_perf_event_pending(void) +static inline void clear_irq_work_pending(void) { asm volatile("stb %0,%1(13)" : : "r" (0), - "i" (offsetof(struct paca_struct, perf_event_pending))); + "i" (offsetof(struct paca_struct, irq_work_pending))); } #else /* 32-bit */ -DEFINE_PER_CPU(u8, perf_event_pending); +DEFINE_PER_CPU(u8, irq_work_pending); -#define set_perf_event_pending_flag() __get_cpu_var(perf_event_pending) = 1 -#define test_perf_event_pending() __get_cpu_var(perf_event_pending) -#define clear_perf_event_pending() __get_cpu_var(perf_event_pending) = 0 +#define set_irq_work_pending_flag() __get_cpu_var(irq_work_pending) = 1 +#define test_irq_work_pending() __get_cpu_var(irq_work_pending) +#define clear_irq_work_pending() __get_cpu_var(irq_work_pending) = 0 #endif /* 32 vs 64 bit */ -void set_perf_event_pending(void) +void set_irq_work_pending(void) { preempt_disable(); - set_perf_event_pending_flag(); + set_irq_work_pending_flag(); set_dec(1); preempt_enable(); } -#else /* CONFIG_PERF_EVENTS */ +#else /* CONFIG_IRQ_WORK */ -#define test_perf_event_pending() 0 -#define clear_perf_event_pending() +#define test_irq_work_pending() 0 +#define clear_irq_work_pending() -#endif /* CONFIG_PERF_EVENTS */ +#endif /* CONFIG_IRQ_WORK */ /* * For iSeries shared processors, we have to let the hypervisor @@ -587,9 +587,9 @@ void timer_interrupt(struct pt_regs * regs) calculate_steal_time(); - if (test_perf_event_pending()) { - clear_perf_event_pending(); - perf_event_do_pending(); + if (test_irq_work_pending()) { + clear_irq_work_pending(); + irq_work_run(); } #ifdef CONFIG_PPC_ISERIES diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index f0777a4..958f0da 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -95,6 +95,7 @@ config S390 select HAVE_KVM if 64BIT select HAVE_ARCH_TRACEHOOK select INIT_ALL_POSSIBLE + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select HAVE_KERNEL_GZIP select HAVE_KERNEL_BZIP2 diff --git a/arch/s390/include/asm/perf_event.h b/arch/s390/include/asm/perf_event.h index 3840cbe..a75f168 100644 --- a/arch/s390/include/asm/perf_event.h +++ b/arch/s390/include/asm/perf_event.h @@ -4,7 +4,6 @@ * Copyright 2009 Martin Schwidefsky, IBM Corporation. */ -static inline void set_perf_event_pending(void) {} -static inline void clear_perf_event_pending(void) {} +/* Empty, just to avoid compiling error */ #define PERF_EVENT_INDEX_OFFSET 0 diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 35b6c3f..35b6879 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -16,6 +16,7 @@ config SUPERH select HAVE_ARCH_TRACEHOOK select HAVE_DMA_API_DEBUG select HAVE_DMA_ATTRS + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC select HAVE_KERNEL_GZIP diff --git a/arch/sh/include/asm/perf_event.h b/arch/sh/include/asm/perf_event.h index 3d0c9f3..14308be 100644 --- a/arch/sh/include/asm/perf_event.h +++ b/arch/sh/include/asm/perf_event.h @@ -26,11 +26,4 @@ extern int register_sh_pmu(struct sh_pmu *); extern int reserve_pmc_hardware(void); extern void release_pmc_hardware(void); -static inline void set_perf_event_pending(void) -{ - /* Nothing to see here, move along. */ -} - -#define PERF_EVENT_INDEX_OFFSET 0 - #endif /* __ASM_SH_PERF_EVENT_H */ diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig index 9212cd4..3e9d314 100644 --- a/arch/sparc/Kconfig +++ b/arch/sparc/Kconfig @@ -26,6 +26,7 @@ config SPARC select ARCH_WANT_OPTIONAL_GPIOLIB select RTC_CLASS select RTC_DRV_M48T59 + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC select HAVE_DMA_ATTRS @@ -54,6 +55,7 @@ config SPARC64 select RTC_DRV_BQ4802 select RTC_DRV_SUN4V select RTC_DRV_STARFIRE + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC diff --git a/arch/sparc/include/asm/perf_event.h b/arch/sparc/include/asm/perf_event.h index 727af70..6e8bfa1 100644 --- a/arch/sparc/include/asm/perf_event.h +++ b/arch/sparc/include/asm/perf_event.h @@ -1,10 +1,6 @@ #ifndef __ASM_SPARC_PERF_EVENT_H #define __ASM_SPARC_PERF_EVENT_H -extern void set_perf_event_pending(void); - -#define PERF_EVENT_INDEX_OFFSET 0 - #ifdef CONFIG_PERF_EVENTS #include diff --git a/arch/sparc/kernel/pcr.c b/arch/sparc/kernel/pcr.c index c4a6a50..b87873c 100644 --- a/arch/sparc/kernel/pcr.c +++ b/arch/sparc/kernel/pcr.c @@ -7,7 +7,7 @@ #include #include -#include +#include #include #include @@ -43,14 +43,14 @@ void __irq_entry deferred_pcr_work_irq(int irq, struct pt_regs *regs) old_regs = set_irq_regs(regs); irq_enter(); -#ifdef CONFIG_PERF_EVENTS - perf_event_do_pending(); +#ifdef CONFIG_IRQ_WORK + irq_work_run(); #endif irq_exit(); set_irq_regs(old_regs); } -void set_perf_event_pending(void) +void arch_irq_work_raise(void) { set_softint(1 << PIL_DEFERRED_PCR_WORK); } diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 9815221..fd227d6 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -25,6 +25,7 @@ config X86 select HAVE_IDE select HAVE_OPROFILE select HAVE_PERF_EVENTS if (!M386 && !M486) + select HAVE_IRQ_WORK select HAVE_IOREMAP_PROT select HAVE_KPROBES select ARCH_WANT_OPTIONAL_GPIOLIB diff --git a/arch/x86/include/asm/entry_arch.h b/arch/x86/include/asm/entry_arch.h index 8e8ec66..b8e96a1 100644 --- a/arch/x86/include/asm/entry_arch.h +++ b/arch/x86/include/asm/entry_arch.h @@ -49,8 +49,8 @@ BUILD_INTERRUPT(apic_timer_interrupt,LOCAL_TIMER_VECTOR) BUILD_INTERRUPT(error_interrupt,ERROR_APIC_VECTOR) BUILD_INTERRUPT(spurious_interrupt,SPURIOUS_APIC_VECTOR) -#ifdef CONFIG_PERF_EVENTS -BUILD_INTERRUPT(perf_pending_interrupt, LOCAL_PENDING_VECTOR) +#ifdef CONFIG_IRQ_WORK +BUILD_INTERRUPT(irq_work_interrupt, IRQ_WORK_VECTOR) #endif #ifdef CONFIG_X86_THERMAL_VECTOR diff --git a/arch/x86/include/asm/hardirq.h b/arch/x86/include/asm/hardirq.h index aeab29a..55e4de6 100644 --- a/arch/x86/include/asm/hardirq.h +++ b/arch/x86/include/asm/hardirq.h @@ -14,7 +14,7 @@ typedef struct { #endif unsigned int x86_platform_ipis; /* arch dependent */ unsigned int apic_perf_irqs; - unsigned int apic_pending_irqs; + unsigned int apic_irq_work_irqs; #ifdef CONFIG_SMP unsigned int irq_resched_count; unsigned int irq_call_count; diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index 46c0fe0..3a54a1c 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -29,7 +29,7 @@ extern void apic_timer_interrupt(void); extern void x86_platform_ipi(void); extern void error_interrupt(void); -extern void perf_pending_interrupt(void); +extern void irq_work_interrupt(void); extern void spurious_interrupt(void); extern void thermal_interrupt(void); diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h index e2ca300..6af0894 100644 --- a/arch/x86/include/asm/irq_vectors.h +++ b/arch/x86/include/asm/irq_vectors.h @@ -114,9 +114,9 @@ #define X86_PLATFORM_IPI_VECTOR 0xed /* - * Performance monitoring pending work vector: + * IRQ work vector: */ -#define LOCAL_PENDING_VECTOR 0xec +#define IRQ_WORK_VECTOR 0xec #define UV_BAU_MESSAGE 0xea diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 9d3f485..7490bf8 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -35,6 +35,7 @@ obj-y := process_$(BITS).o signal.o entry_$(BITS).o obj-y += traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o obj-y += time.o ioport.o ldt.o dumpstack.o obj-y += setup.o x86_init.o i8259.o irqinit.o jump_label.o +obj-$(CONFIG_IRQ_WORK) += irq_work.o obj-$(CONFIG_X86_VISWS) += visws_quirks.o obj-$(CONFIG_X86_32) += probe_roms_32.o obj-$(CONFIG_X86_32) += sys_i386_32.o i386_ksyms_32.o diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index e2513f2..fe73c18 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1196,25 +1196,6 @@ static int x86_pmu_handle_irq(struct pt_regs *regs) return handled; } -void smp_perf_pending_interrupt(struct pt_regs *regs) -{ - irq_enter(); - ack_APIC_irq(); - inc_irq_stat(apic_pending_irqs); - perf_event_do_pending(); - irq_exit(); -} - -void set_perf_event_pending(void) -{ -#ifdef CONFIG_X86_LOCAL_APIC - if (!x86_pmu.apic || !x86_pmu_initialized()) - return; - - apic->send_IPI_self(LOCAL_PENDING_VECTOR); -#endif -} - void perf_events_lapic_init(void) { if (!x86_pmu.apic || !x86_pmu_initialized()) diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 17be5ec..c375c79 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1023,9 +1023,9 @@ apicinterrupt ERROR_APIC_VECTOR \ apicinterrupt SPURIOUS_APIC_VECTOR \ spurious_interrupt smp_spurious_interrupt -#ifdef CONFIG_PERF_EVENTS -apicinterrupt LOCAL_PENDING_VECTOR \ - perf_pending_interrupt smp_perf_pending_interrupt +#ifdef CONFIG_IRQ_WORK +apicinterrupt IRQ_WORK_VECTOR \ + irq_work_interrupt smp_irq_work_interrupt #endif /* diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index 91fd0c7..44edb03 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -67,10 +67,10 @@ static int show_other_interrupts(struct seq_file *p, int prec) for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->apic_perf_irqs); seq_printf(p, " Performance monitoring interrupts\n"); - seq_printf(p, "%*s: ", prec, "PND"); + seq_printf(p, "%*s: ", prec, "IWI"); for_each_online_cpu(j) - seq_printf(p, "%10u ", irq_stats(j)->apic_pending_irqs); - seq_printf(p, " Performance pending work\n"); + seq_printf(p, "%10u ", irq_stats(j)->apic_irq_work_irqs); + seq_printf(p, " IRQ work interrupts\n"); #endif if (x86_platform_ipi_callback) { seq_printf(p, "%*s: ", prec, "PLT"); @@ -185,7 +185,7 @@ u64 arch_irq_stat_cpu(unsigned int cpu) sum += irq_stats(cpu)->apic_timer_irqs; sum += irq_stats(cpu)->irq_spurious_count; sum += irq_stats(cpu)->apic_perf_irqs; - sum += irq_stats(cpu)->apic_pending_irqs; + sum += irq_stats(cpu)->apic_irq_work_irqs; #endif if (x86_platform_ipi_callback) sum += irq_stats(cpu)->x86_platform_ipis; diff --git a/arch/x86/kernel/irq_work.c b/arch/x86/kernel/irq_work.c new file mode 100644 index 0000000..ca8f703 --- /dev/null +++ b/arch/x86/kernel/irq_work.c @@ -0,0 +1,30 @@ +/* + * x86 specific code for irq_work + * + * Copyright (C) 2010 Red Hat, Inc., Peter Zijlstra + */ + +#include +#include +#include +#include + +void smp_irq_work_interrupt(struct pt_regs *regs) +{ + irq_enter(); + ack_APIC_irq(); + inc_irq_stat(apic_irq_work_irqs); + irq_work_run(); + irq_exit(); +} + +void arch_irq_work_raise(void) +{ +#ifdef CONFIG_X86_LOCAL_APIC + if (!cpu_has_apic) + return; + + apic->send_IPI_self(IRQ_WORK_VECTOR); + apic_wait_icr_idle(); +#endif +} diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c index 990ae7c..713969b 100644 --- a/arch/x86/kernel/irqinit.c +++ b/arch/x86/kernel/irqinit.c @@ -224,9 +224,9 @@ static void __init apic_intr_init(void) alloc_intr_gate(SPURIOUS_APIC_VECTOR, spurious_interrupt); alloc_intr_gate(ERROR_APIC_VECTOR, error_interrupt); - /* Performance monitoring interrupts: */ -# ifdef CONFIG_PERF_EVENTS - alloc_intr_gate(LOCAL_PENDING_VECTOR, perf_pending_interrupt); + /* IRQ work interrupts: */ +# ifdef CONFIG_IRQ_WORK + alloc_intr_gate(IRQ_WORK_VECTOR, irq_work_interrupt); # endif #endif diff --git a/include/linux/irq_work.h b/include/linux/irq_work.h new file mode 100644 index 0000000..4fa09d4 --- /dev/null +++ b/include/linux/irq_work.h @@ -0,0 +1,20 @@ +#ifndef _LINUX_IRQ_WORK_H +#define _LINUX_IRQ_WORK_H + +struct irq_work { + struct irq_work *next; + void (*func)(struct irq_work *); +}; + +static inline +void init_irq_work(struct irq_work *entry, void (*func)(struct irq_work *)) +{ + entry->next = NULL; + entry->func = func; +} + +bool irq_work_queue(struct irq_work *entry); +void irq_work_run(void); +void irq_work_sync(struct irq_work *entry); + +#endif /* _LINUX_IRQ_WORK_H */ diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index a9227e9..2ebfc9a 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -486,6 +486,7 @@ struct perf_guest_info_callbacks { #include #include #include +#include #include #include @@ -672,11 +673,6 @@ struct perf_buffer { void *data_pages[0]; }; -struct perf_pending_entry { - struct perf_pending_entry *next; - void (*func)(struct perf_pending_entry *); -}; - struct perf_sample_data; typedef void (*perf_overflow_handler_t)(struct perf_event *, int, @@ -784,7 +780,7 @@ struct perf_event { int pending_wakeup; int pending_kill; int pending_disable; - struct perf_pending_entry pending; + struct irq_work pending; atomic_t event_limit; @@ -898,8 +894,6 @@ extern int perf_event_init_task(struct task_struct *child); extern void perf_event_exit_task(struct task_struct *child); extern void perf_event_free_task(struct task_struct *task); extern void perf_event_delayed_put(struct task_struct *task); -extern void set_perf_event_pending(void); -extern void perf_event_do_pending(void); extern void perf_event_print_debug(void); extern void perf_pmu_disable(struct pmu *pmu); extern void perf_pmu_enable(struct pmu *pmu); @@ -1078,7 +1072,6 @@ static inline int perf_event_init_task(struct task_struct *child) { return 0; } static inline void perf_event_exit_task(struct task_struct *child) { } static inline void perf_event_free_task(struct task_struct *task) { } static inline void perf_event_delayed_put(struct task_struct *task) { } -static inline void perf_event_do_pending(void) { } static inline void perf_event_print_debug(void) { } static inline int perf_event_task_disable(void) { return -EINVAL; } static inline int perf_event_task_enable(void) { return -EINVAL; } diff --git a/init/Kconfig b/init/Kconfig index 2de5b1c..1ef0b43 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -21,6 +21,13 @@ config CONSTRUCTORS depends on !UML default y +config HAVE_IRQ_WORK + bool + +config IRQ_WORK + bool + depends on HAVE_IRQ_WORK + menu "General setup" config EXPERIMENTAL @@ -987,6 +994,7 @@ config PERF_EVENTS default y if (PROFILING || PERF_COUNTERS) depends on HAVE_PERF_EVENTS select ANON_INODES + select IRQ_WORK help Enable kernel support for various performance events provided by software and hardware. diff --git a/kernel/Makefile b/kernel/Makefile index d52b473..4d9bf5f 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -23,6 +23,7 @@ CFLAGS_REMOVE_rtmutex-debug.o = -pg CFLAGS_REMOVE_cgroup-debug.o = -pg CFLAGS_REMOVE_sched_clock.o = -pg CFLAGS_REMOVE_perf_event.o = -pg +CFLAGS_REMOVE_irq_work.o = -pg endif obj-$(CONFIG_FREEZER) += freezer.o @@ -100,6 +101,7 @@ obj-$(CONFIG_TRACING) += trace/ obj-$(CONFIG_X86_DS) += trace/ obj-$(CONFIG_RING_BUFFER) += trace/ obj-$(CONFIG_SMP) += sched_cpupri.o +obj-$(CONFIG_IRQ_WORK) += irq_work.o obj-$(CONFIG_PERF_EVENTS) += perf_event.o obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o obj-$(CONFIG_USER_RETURN_NOTIFIER) += user-return-notifier.o diff --git a/kernel/irq_work.c b/kernel/irq_work.c new file mode 100644 index 0000000..f16763f --- /dev/null +++ b/kernel/irq_work.c @@ -0,0 +1,164 @@ +/* + * Copyright (C) 2010 Red Hat, Inc., Peter Zijlstra + * + * Provides a framework for enqueueing and running callbacks from hardirq + * context. The enqueueing is NMI-safe. + */ + +#include +#include +#include +#include + +/* + * An entry can be in one of four states: + * + * free NULL, 0 -> {claimed} : free to be used + * claimed NULL, 3 -> {pending} : claimed to be enqueued + * pending next, 3 -> {busy} : queued, pending callback + * busy NULL, 2 -> {free, claimed} : callback in progress, can be claimed + * + * We use the lower two bits of the next pointer to keep PENDING and BUSY + * flags. + */ + +#define IRQ_WORK_PENDING 1UL +#define IRQ_WORK_BUSY 2UL +#define IRQ_WORK_FLAGS 3UL + +static inline bool irq_work_is_set(struct irq_work *entry, int flags) +{ + return (unsigned long)entry->next & flags; +} + +static inline struct irq_work *irq_work_next(struct irq_work *entry) +{ + unsigned long next = (unsigned long)entry->next; + next &= ~IRQ_WORK_FLAGS; + return (struct irq_work *)next; +} + +static inline struct irq_work *next_flags(struct irq_work *entry, int flags) +{ + unsigned long next = (unsigned long)entry; + next |= flags; + return (struct irq_work *)next; +} + +static DEFINE_PER_CPU(struct irq_work *, irq_work_list); + +/* + * Claim the entry so that no one else will poke at it. + */ +static bool irq_work_claim(struct irq_work *entry) +{ + struct irq_work *next, *nflags; + + do { + next = entry->next; + if ((unsigned long)next & IRQ_WORK_PENDING) + return false; + nflags = next_flags(next, IRQ_WORK_FLAGS); + } while (cmpxchg(&entry->next, next, nflags) != next); + + return true; +} + + +void __weak arch_irq_work_raise(void) +{ + /* + * Lame architectures will get the timer tick callback + */ +} + +/* + * Queue the entry and raise the IPI if needed. + */ +static void __irq_work_queue(struct irq_work *entry) +{ + struct irq_work **head, *next; + + head = &get_cpu_var(irq_work_list); + + do { + next = *head; + /* Can assign non-atomic because we keep the flags set. */ + entry->next = next_flags(next, IRQ_WORK_FLAGS); + } while (cmpxchg(head, next, entry) != next); + + /* The list was empty, raise self-interrupt to start processing. */ + if (!irq_work_next(entry)) + arch_irq_work_raise(); + + put_cpu_var(irq_work_list); +} + +/* + * Enqueue the irq_work @entry, returns true on success, failure when the + * @entry was already enqueued by someone else. + * + * Can be re-enqueued while the callback is still in progress. + */ +bool irq_work_queue(struct irq_work *entry) +{ + if (!irq_work_claim(entry)) { + /* + * Already enqueued, can't do! + */ + return false; + } + + __irq_work_queue(entry); + return true; +} +EXPORT_SYMBOL_GPL(irq_work_queue); + +/* + * Run the irq_work entries on this cpu. Requires to be ran from hardirq + * context with local IRQs disabled. + */ +void irq_work_run(void) +{ + struct irq_work *list, **head; + + head = &__get_cpu_var(irq_work_list); + if (*head == NULL) + return; + + BUG_ON(!in_irq()); + BUG_ON(!irqs_disabled()); + + list = xchg(head, NULL); + while (list != NULL) { + struct irq_work *entry = list; + + list = irq_work_next(list); + + /* + * Clear the PENDING bit, after this point the @entry + * can be re-used. + */ + entry->next = next_flags(NULL, IRQ_WORK_BUSY); + entry->func(entry); + /* + * Clear the BUSY bit and return to the free state if + * no-one else claimed it meanwhile. + */ + cmpxchg(&entry->next, next_flags(NULL, IRQ_WORK_BUSY), NULL); + } +} +EXPORT_SYMBOL_GPL(irq_work_run); + +/* + * Synchronize against the irq_work @entry, ensures the entry is not + * currently in use. + */ +void irq_work_sync(struct irq_work *entry) +{ + WARN_ON_ONCE(irqs_disabled()); + + while (irq_work_is_set(entry, IRQ_WORK_BUSY)) + cpu_relax(); +} +EXPORT_SYMBOL_GPL(irq_work_sync); diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 634f86a4..99b9700 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -2206,12 +2206,11 @@ static void free_event_rcu(struct rcu_head *head) kfree(event); } -static void perf_pending_sync(struct perf_event *event); static void perf_buffer_put(struct perf_buffer *buffer); static void free_event(struct perf_event *event) { - perf_pending_sync(event); + irq_work_sync(&event->pending); if (!event->parent) { atomic_dec(&nr_events); @@ -3162,16 +3161,7 @@ void perf_event_wakeup(struct perf_event *event) } } -/* - * Pending wakeups - * - * Handle the case where we need to wakeup up from NMI (or rq->lock) context. - * - * The NMI bit means we cannot possibly take locks. Therefore, maintain a - * single linked list and use cmpxchg() to add entries lockless. - */ - -static void perf_pending_event(struct perf_pending_entry *entry) +static void perf_pending_event(struct irq_work *entry) { struct perf_event *event = container_of(entry, struct perf_event, pending); @@ -3187,89 +3177,6 @@ static void perf_pending_event(struct perf_pending_entry *entry) } } -#define PENDING_TAIL ((struct perf_pending_entry *)-1UL) - -static DEFINE_PER_CPU(struct perf_pending_entry *, perf_pending_head) = { - PENDING_TAIL, -}; - -static void perf_pending_queue(struct perf_pending_entry *entry, - void (*func)(struct perf_pending_entry *)) -{ - struct perf_pending_entry **head; - - if (cmpxchg(&entry->next, NULL, PENDING_TAIL) != NULL) - return; - - entry->func = func; - - head = &get_cpu_var(perf_pending_head); - - do { - entry->next = *head; - } while (cmpxchg(head, entry->next, entry) != entry->next); - - set_perf_event_pending(); - - put_cpu_var(perf_pending_head); -} - -static int __perf_pending_run(void) -{ - struct perf_pending_entry *list; - int nr = 0; - - list = xchg(&__get_cpu_var(perf_pending_head), PENDING_TAIL); - while (list != PENDING_TAIL) { - void (*func)(struct perf_pending_entry *); - struct perf_pending_entry *entry = list; - - list = list->next; - - func = entry->func; - entry->next = NULL; - /* - * Ensure we observe the unqueue before we issue the wakeup, - * so that we won't be waiting forever. - * -- see perf_not_pending(). - */ - smp_wmb(); - - func(entry); - nr++; - } - - return nr; -} - -static inline int perf_not_pending(struct perf_event *event) -{ - /* - * If we flush on whatever cpu we run, there is a chance we don't - * need to wait. - */ - get_cpu(); - __perf_pending_run(); - put_cpu(); - - /* - * Ensure we see the proper queue state before going to sleep - * so that we do not miss the wakeup. -- see perf_pending_handle() - */ - smp_rmb(); - return event->pending.next == NULL; -} - -static void perf_pending_sync(struct perf_event *event) -{ - wait_event(event->waitq, perf_not_pending(event)); -} - -void perf_event_do_pending(void) -{ - __perf_pending_run(); -} - /* * We assume there is only KVM supporting the callbacks. * Later on, we might change it to a list if there is @@ -3319,8 +3226,7 @@ static void perf_output_wakeup(struct perf_output_handle *handle) if (handle->nmi) { handle->event->pending_wakeup = 1; - perf_pending_queue(&handle->event->pending, - perf_pending_event); + irq_work_queue(&handle->event->pending); } else perf_event_wakeup(handle->event); } @@ -4356,8 +4262,7 @@ static int __perf_event_overflow(struct perf_event *event, int nmi, event->pending_kill = POLL_HUP; if (nmi) { event->pending_disable = 1; - perf_pending_queue(&event->pending, - perf_pending_event); + irq_work_queue(&event->pending); } else perf_event_disable(event); } @@ -5374,6 +5279,7 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, INIT_LIST_HEAD(&event->event_entry); INIT_LIST_HEAD(&event->sibling_list); init_waitqueue_head(&event->waitq); + init_irq_work(&event->pending, perf_pending_event); mutex_init(&event->mmap_mutex); diff --git a/kernel/timer.c b/kernel/timer.c index 97bf05b..68a9ae7 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -37,7 +37,7 @@ #include #include #include -#include +#include #include #include @@ -1279,7 +1279,10 @@ void update_process_times(int user_tick) run_local_timers(); rcu_check_callbacks(cpu, user_tick); printk_tick(); - perf_event_do_pending(); +#ifdef CONFIG_IRQ_WORK + if (in_irq()) + irq_work_run(); +#endif scheduler_tick(); run_posix_cpu_timers(p); } -- cgit v0.10.2 From 74c3337c2fc6389d3a57a622a936036b6db6b2e8 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 15 Oct 2010 11:40:29 +0200 Subject: perf: Fix group moving Matt found we trigger the WARN_ON_ONCE() in perf_group_attach() when we take the move_group path in perf_event_open(). Since we cannot de-construct the group (we rely on it to move the events), we have to simply ignore the double attach. The group state is context invariant and doesn't need changing. Reported-by: Matt Fleming Signed-off-by: Peter Zijlstra LKML-Reference: <1287135757.29097.1368.camel@twins> Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 99b9700..346dc0e 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -315,7 +315,12 @@ static void perf_group_attach(struct perf_event *event) { struct perf_event *group_leader = event->group_leader; - WARN_ON_ONCE(event->attach_state & PERF_ATTACH_GROUP); + /* + * We can have double attach due to group movement in perf_event_open. + */ + if (event->attach_state & PERF_ATTACH_GROUP) + return; + event->attach_state |= PERF_ATTACH_GROUP; if (group_leader == event) -- cgit v0.10.2 From e7d0bc047548d76feee6b23f7d3d9da927189a50 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 16:54:51 +0200 Subject: perf: Fix task refcount bugs Currently it looks like find_lively_task_by_vpid() takes a task ref and relies on find_get_context() to drop it. The problem is that perf_event_create_kernel_counter() shouldn't be dropping task refs. Signed-off-by: Peter Zijlstra Acked-by: Frederic Weisbecker Acked-by: Matt Helsley LKML-Reference: <20101014203625.278436085@chello.nl> Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 346dc0e..f928878 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -2190,11 +2190,9 @@ retry: } } - put_task_struct(task); return ctx; errout: - put_task_struct(task); return ERR_PTR(err); } @@ -5602,7 +5600,7 @@ SYSCALL_DEFINE5(perf_event_open, ctx = find_get_context(pmu, task, cpu); if (IS_ERR(ctx)) { err = PTR_ERR(ctx); - goto err_group_fd; + goto err_task; } /* @@ -5698,6 +5696,9 @@ SYSCALL_DEFINE5(perf_event_open, err_context: put_ctx(ctx); +err_task: + if (task) + put_task_struct(task); err_group_fd: fput_light(group_file, fput_needed); free_event(event); -- cgit v0.10.2 From c6be5a5cb62592d9d661899a2aa78236eb00ffa5 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 16:59:46 +0200 Subject: perf: Find task before event alloc So that we can pass the task pointer to the event allocation, so that we can use task associated data during event initialization. Signed-off-by: Peter Zijlstra LKML-Reference: <20101014203625.340789919@chello.nl> Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index f928878..b21d06a 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -5551,10 +5551,18 @@ SYSCALL_DEFINE5(perf_event_open, group_leader = NULL; } + if (pid != -1) { + task = find_lively_task_by_vpid(pid); + if (IS_ERR(task)) { + err = PTR_ERR(task); + goto err_group_fd; + } + } + event = perf_event_alloc(&attr, cpu, group_leader, NULL, NULL); if (IS_ERR(event)) { err = PTR_ERR(event); - goto err_fd; + goto err_task; } /* @@ -5586,21 +5594,13 @@ SYSCALL_DEFINE5(perf_event_open, } } - if (pid != -1) { - task = find_lively_task_by_vpid(pid); - if (IS_ERR(task)) { - err = PTR_ERR(task); - goto err_group_fd; - } - } - /* * Get the target context (task or percpu): */ ctx = find_get_context(pmu, task, cpu); if (IS_ERR(ctx)) { err = PTR_ERR(ctx); - goto err_task; + goto err_alloc; } /* @@ -5696,12 +5696,13 @@ SYSCALL_DEFINE5(perf_event_open, err_context: put_ctx(ctx); +err_alloc: + free_event(event); err_task: if (task) put_task_struct(task); err_group_fd: fput_light(group_file, fput_needed); - free_event(event); err_fd: put_unused_fd(event_fd); return err; -- cgit v0.10.2 From d580ff8699e8811a9af37e9de4dea375401bdeec Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 17:43:23 +0200 Subject: perf, hw_breakpoint: Fix crash in hw_breakpoint creation hw_breakpoint creation needs to account stuff per-task to ensure there is always sufficient hardware resources to back these things due to ptrace. With the perf per pmu context changes the event initialization no longer has access to the event context, for the simple reason that we need to first find the pmu (result of initialization) before we can find the context. This makes hw_breakpoints unhappy, because it can no longer do per task accounting, cure this by frobbing a task pointer in the event::hw bits for now... Signed-off-by: Peter Zijlstra Cc: Frederic Weisbecker LKML-Reference: <20101014203625.391543667@chello.nl> Signed-off-by: Ingo Molnar diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 2ebfc9a..97965fa 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -536,6 +536,12 @@ struct hw_perf_event { struct { /* breakpoint */ struct arch_hw_breakpoint info; struct list_head bp_list; + /* + * Crufty hack to avoid the chicken and egg + * problem hw_breakpoint has with context + * creation and event initalization. + */ + struct task_struct *bp_target; }; #endif }; @@ -693,6 +699,7 @@ struct swevent_hlist { #define PERF_ATTACH_CONTEXT 0x01 #define PERF_ATTACH_GROUP 0x02 +#define PERF_ATTACH_TASK 0x04 /** * struct perf_event - performance event kernel representation: diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index 3b714e8..2c9120f 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -113,12 +113,12 @@ static unsigned int max_task_bp_pinned(int cpu, enum bp_type_idx type) */ static int task_bp_pinned(struct perf_event *bp, enum bp_type_idx type) { - struct perf_event_context *ctx = bp->ctx; + struct task_struct *tsk = bp->hw.bp_target; struct perf_event *iter; int count = 0; list_for_each_entry(iter, &bp_task_head, hw.bp_list) { - if (iter->ctx == ctx && find_slot_idx(iter) == type) + if (iter->hw.bp_target == tsk && find_slot_idx(iter) == type) count += hw_breakpoint_weight(iter); } @@ -134,7 +134,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, enum bp_type_idx type) { int cpu = bp->cpu; - struct task_struct *tsk = bp->ctx->task; + struct task_struct *tsk = bp->hw.bp_target; if (cpu >= 0) { slots->pinned = per_cpu(nr_cpu_bp_pinned[type], cpu); @@ -213,7 +213,7 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, int weight) { int cpu = bp->cpu; - struct task_struct *tsk = bp->ctx->task; + struct task_struct *tsk = bp->hw.bp_target; /* Pinned counter cpu profiling */ if (!tsk) { diff --git a/kernel/perf_event.c b/kernel/perf_event.c index b21d06a..856e20b 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -5255,9 +5255,10 @@ unlock: */ static struct perf_event * perf_event_alloc(struct perf_event_attr *attr, int cpu, - struct perf_event *group_leader, - struct perf_event *parent_event, - perf_overflow_handler_t overflow_handler) + struct task_struct *task, + struct perf_event *group_leader, + struct perf_event *parent_event, + perf_overflow_handler_t overflow_handler) { struct pmu *pmu; struct perf_event *event; @@ -5299,6 +5300,17 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, event->state = PERF_EVENT_STATE_INACTIVE; + if (task) { + event->attach_state = PERF_ATTACH_TASK; +#ifdef CONFIG_HAVE_HW_BREAKPOINT + /* + * hw_breakpoint is a bit difficult here.. + */ + if (attr->type == PERF_TYPE_BREAKPOINT) + event->hw.bp_target = task; +#endif + } + if (!overflow_handler && parent_event) overflow_handler = parent_event->overflow_handler; @@ -5559,7 +5571,7 @@ SYSCALL_DEFINE5(perf_event_open, } } - event = perf_event_alloc(&attr, cpu, group_leader, NULL, NULL); + event = perf_event_alloc(&attr, cpu, task, group_leader, NULL, NULL); if (IS_ERR(event)) { err = PTR_ERR(event); goto err_task; @@ -5728,7 +5740,7 @@ perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, * Get the target context (task or percpu): */ - event = perf_event_alloc(attr, cpu, NULL, NULL, overflow_handler); + event = perf_event_alloc(attr, cpu, task, NULL, NULL, overflow_handler); if (IS_ERR(event)) { err = PTR_ERR(event); goto err; @@ -5996,6 +6008,7 @@ inherit_event(struct perf_event *parent_event, child_event = perf_event_alloc(&parent_event->attr, parent_event->cpu, + child, group_leader, parent_event, NULL); if (IS_ERR(child_event)) -- cgit v0.10.2 From 3b6e901f839f42afb40f614418df82c08b01320a Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 21:10:38 +0200 Subject: jump_label: Use more consistent naming Now that there's still only a few users around, rename things to make them more consistent. Signed-off-by: Peter Zijlstra LKML-Reference: <20101014203625.448565169@chello.nl> Signed-off-by: Ingo Molnar diff --git a/include/linux/jump_label.h b/include/linux/jump_label.h index b72cd9f..81be496 100644 --- a/include/linux/jump_label.h +++ b/include/linux/jump_label.h @@ -25,10 +25,10 @@ extern void jump_label_update(unsigned long key, enum jump_label_type type); extern void jump_label_apply_nops(struct module *mod); extern int jump_label_text_reserved(void *start, void *end); -#define enable_jump_label(key) \ +#define jump_label_enable(key) \ jump_label_update((unsigned long)key, JUMP_LABEL_ENABLE); -#define disable_jump_label(key) \ +#define jump_label_disable(key) \ jump_label_update((unsigned long)key, JUMP_LABEL_DISABLE); #else @@ -39,12 +39,12 @@ do { \ goto label; \ } while (0) -#define enable_jump_label(cond_var) \ +#define jump_label_enable(cond_var) \ do { \ *(cond_var) = 1; \ } while (0) -#define disable_jump_label(cond_var) \ +#define jump_label_disable(cond_var) \ do { \ *(cond_var) = 0; \ } while (0) diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c index d6073a5..e95ee7f 100644 --- a/kernel/tracepoint.c +++ b/kernel/tracepoint.c @@ -265,10 +265,10 @@ static void set_tracepoint(struct tracepoint_entry **entry, */ rcu_assign_pointer(elem->funcs, (*entry)->funcs); if (!elem->state && active) { - enable_jump_label(&elem->state); + jump_label_enable(&elem->state); elem->state = active; } else if (elem->state && !active) { - disable_jump_label(&elem->state); + jump_label_disable(&elem->state); elem->state = active; } } @@ -285,7 +285,7 @@ static void disable_tracepoint(struct tracepoint *elem) elem->unregfunc(); if (elem->state) { - disable_jump_label(&elem->state); + jump_label_disable(&elem->state); elem->state = 0; } rcu_assign_pointer(elem->funcs, NULL); diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c index e925c7b..7bd6df7 100644 --- a/lib/dynamic_debug.c +++ b/lib/dynamic_debug.c @@ -142,9 +142,9 @@ static void ddebug_change(const struct ddebug_query *query, dt->num_enabled++; dp->flags = newflags; if (newflags) { - enable_jump_label(&dp->enabled); + jump_label_enable(&dp->enabled); } else { - disable_jump_label(&dp->enabled); + jump_label_disable(&dp->enabled); } if (verbose) printk(KERN_INFO -- cgit v0.10.2 From 8b92538d84e50062560ba33adbaed7887b6e4a42 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 21:39:02 +0200 Subject: jump_label: Add atomic_t interface Add an interface to allow usage of jump_labels with atomic counters. Signed-off-by: Peter Zijlstra Acked-by: Frederic Weisbecker LKML-Reference: <20101014203625.501657727@chello.nl> Signed-off-by: Ingo Molnar diff --git a/include/linux/jump_label_ref.h b/include/linux/jump_label_ref.h new file mode 100644 index 0000000..e5d012a --- /dev/null +++ b/include/linux/jump_label_ref.h @@ -0,0 +1,44 @@ +#ifndef _LINUX_JUMP_LABEL_REF_H +#define _LINUX_JUMP_LABEL_REF_H + +#include +#include + +#ifdef HAVE_JUMP_LABEL + +static inline void jump_label_inc(atomic_t *key) +{ + if (atomic_add_return(1, key) == 1) + jump_label_enable(key); +} + +static inline void jump_label_dec(atomic_t *key) +{ + if (atomic_dec_and_test(key)) + jump_label_disable(key); +} + +#else /* !HAVE_JUMP_LABEL */ + +static inline void jump_label_inc(atomic_t *key) +{ + atomic_inc(key); +} + +static inline void jump_label_dec(atomic_t *key) +{ + atomic_dec(key); +} + +#undef JUMP_LABEL +#define JUMP_LABEL(key, label) \ +do { \ + if (unlikely(__builtin_choose_expr( \ + __builtin_types_compatible_p(typeof(key), atomic_t *), \ + atomic_read((atomic_t *)(key)), *(key)))) \ + goto label; \ +} while (0) + +#endif /* HAVE_JUMP_LABEL */ + +#endif /* _LINUX_JUMP_LABEL_REF_H */ -- cgit v0.10.2 From 82cd6def9806dcb6a325fb6abbc1d61388a15f6a Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 17:57:23 +0200 Subject: perf: Use jump_labels to optimize the scheduler hooks Trades a call + conditional + ret for an unconditional jmp. Acked-by: Frederic Weisbecker Signed-off-by: Peter Zijlstra LKML-Reference: <20101014203625.501657727@chello.nl> Signed-off-by: Ingo Molnar diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 97965fa..7f0e7f5 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -487,6 +487,7 @@ struct perf_guest_info_callbacks { #include #include #include +#include #include #include @@ -895,8 +896,30 @@ extern void perf_pmu_unregister(struct pmu *pmu); extern int perf_num_counters(void); extern const char *perf_pmu_name(void); -extern void perf_event_task_sched_in(struct task_struct *task); -extern void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next); +extern void __perf_event_task_sched_in(struct task_struct *task); +extern void __perf_event_task_sched_out(struct task_struct *task, struct task_struct *next); + +extern atomic_t perf_task_events; + +static inline void perf_event_task_sched_in(struct task_struct *task) +{ + JUMP_LABEL(&perf_task_events, have_events); + return; + +have_events: + __perf_event_task_sched_in(task); +} + +static inline +void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next) +{ + JUMP_LABEL(&perf_task_events, have_events); + return; + +have_events: + __perf_event_task_sched_out(task, next); +} + extern int perf_event_init_task(struct task_struct *child); extern void perf_event_exit_task(struct task_struct *child); extern void perf_event_free_task(struct task_struct *task); diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 856e20b..f7febb0 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -34,7 +34,7 @@ #include -static atomic_t nr_events __read_mostly; +atomic_t perf_task_events __read_mostly; static atomic_t nr_mmap_events __read_mostly; static atomic_t nr_comm_events __read_mostly; static atomic_t nr_task_events __read_mostly; @@ -1311,8 +1311,8 @@ void perf_event_context_sched_out(struct task_struct *task, int ctxn, * accessing the event control register. If a NMI hits, then it will * not restart the event. */ -void perf_event_task_sched_out(struct task_struct *task, - struct task_struct *next) +void __perf_event_task_sched_out(struct task_struct *task, + struct task_struct *next) { int ctxn; @@ -1340,14 +1340,6 @@ static void task_ctx_sched_out(struct perf_event_context *ctx, /* * Called with IRQs disabled */ -static void __perf_event_task_sched_out(struct perf_event_context *ctx) -{ - task_ctx_sched_out(ctx, EVENT_ALL); -} - -/* - * Called with IRQs disabled - */ static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx, enum event_type_t event_type) { @@ -1494,7 +1486,7 @@ void perf_event_context_sched_in(struct perf_event_context *ctx) * accessing the event control register. If a NMI hits, then it will * keep the event running. */ -void perf_event_task_sched_in(struct task_struct *task) +void __perf_event_task_sched_in(struct task_struct *task) { struct perf_event_context *ctx; int ctxn; @@ -2216,7 +2208,8 @@ static void free_event(struct perf_event *event) irq_work_sync(&event->pending); if (!event->parent) { - atomic_dec(&nr_events); + if (event->attach_state & PERF_ATTACH_TASK) + jump_label_dec(&perf_task_events); if (event->attr.mmap || event->attr.mmap_data) atomic_dec(&nr_mmap_events); if (event->attr.comm) @@ -5354,7 +5347,8 @@ done: event->pmu = pmu; if (!event->parent) { - atomic_inc(&nr_events); + if (event->attach_state & PERF_ATTACH_TASK) + jump_label_inc(&perf_task_events); if (event->attr.mmap || event->attr.mmap_data) atomic_inc(&nr_mmap_events); if (event->attr.comm) @@ -5849,7 +5843,7 @@ static void perf_event_exit_task_context(struct task_struct *child, int ctxn) * our context. */ child_ctx = child->perf_event_ctxp[ctxn]; - __perf_event_task_sched_out(child_ctx); + task_ctx_sched_out(child_ctx, EVENT_ALL); /* * Take the context lock here so that if find_get_context is -- cgit v0.10.2 From 7e54a5a0b655734326dc78c2b5efc1eb35497bb6 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 14 Oct 2010 22:32:45 +0200 Subject: perf: Optimize sw events Acked-by: Frederic Weisbecker Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7f0e7f5..3b80cbf 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -1012,18 +1012,20 @@ static inline void perf_fetch_caller_regs(struct pt_regs *regs) perf_arch_fetch_caller_regs(regs, CALLER_ADDR0); } -static inline void +static __always_inline void perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr) { - if (atomic_read(&perf_swevent_enabled[event_id])) { - struct pt_regs hot_regs; - - if (!regs) { - perf_fetch_caller_regs(&hot_regs); - regs = &hot_regs; - } - __perf_sw_event(event_id, nr, nmi, regs, addr); + struct pt_regs hot_regs; + + JUMP_LABEL(&perf_swevent_enabled[event_id], have_event); + return; + +have_event: + if (!regs) { + perf_fetch_caller_regs(&hot_regs); + regs = &hot_regs; } + __perf_sw_event(event_id, nr, nmi, regs, addr); } extern void perf_event_mmap(struct vm_area_struct *vma); diff --git a/kernel/perf_event.c b/kernel/perf_event.c index f7febb0..05ecf6f 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -4669,7 +4669,7 @@ static void sw_perf_event_destroy(struct perf_event *event) WARN_ON(event->parent); - atomic_dec(&perf_swevent_enabled[event_id]); + jump_label_dec(&perf_swevent_enabled[event_id]); swevent_hlist_put(event); } @@ -4699,7 +4699,7 @@ static int perf_swevent_init(struct perf_event *event) if (err) return err; - atomic_inc(&perf_swevent_enabled[event_id]); + jump_label_inc(&perf_swevent_enabled[event_id]); event->destroy = sw_perf_event_destroy; } -- cgit v0.10.2 From ebf31f502492527e2b6b5e5cf85a4ebc7fc8a52e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Sun, 17 Oct 2010 12:15:00 +0200 Subject: jump_label: Add COND_STMT(), reducer wrappery The use of the JUMP_LABEL() construct ends up creating endless silly wrappers, create a higher level construct to reduce this clutter. Signed-off-by: Peter Zijlstra Cc: Jason Baron Cc: Steven Rostedt Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: Paul Mackerras LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/include/linux/jump_label.h b/include/linux/jump_label.h index 81be496..b67cb18 100644 --- a/include/linux/jump_label.h +++ b/include/linux/jump_label.h @@ -61,4 +61,14 @@ static inline int jump_label_text_reserved(void *start, void *end) #endif +#define COND_STMT(key, stmt) \ +do { \ + __label__ jl_enabled; \ + JUMP_LABEL(key, jl_enabled); \ + if (0) { \ +jl_enabled: \ + stmt; \ + } \ +} while (0) + #endif diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 3b80cbf..057bf22 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -903,21 +903,13 @@ extern atomic_t perf_task_events; static inline void perf_event_task_sched_in(struct task_struct *task) { - JUMP_LABEL(&perf_task_events, have_events); - return; - -have_events: - __perf_event_task_sched_in(task); + COND_STMT(&perf_task_events, __perf_event_task_sched_in(task)); } static inline void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next) { - JUMP_LABEL(&perf_task_events, have_events); - return; - -have_events: - __perf_event_task_sched_out(task, next); + COND_STMT(&perf_task_events, __perf_event_task_sched_out(task, next)); } extern int perf_event_init_task(struct task_struct *child); -- cgit v0.10.2 From ef8002f6848236de5adc613063ebeabddea8a6fb Mon Sep 17 00:00:00 2001 From: Nikhil Rao Date: Wed, 13 Oct 2010 12:09:35 -0700 Subject: sched: Do not consider SCHED_IDLE tasks to be cache hot This patch adds a check in task_hot to return if the task has SCHED_IDLE policy. SCHED_IDLE tasks have very low weight, and when run with regular workloads, are typically scheduled many milliseconds apart. There is no need to consider these tasks hot for load balancing. Signed-off-by: Nikhil Rao Signed-off-by: Peter Zijlstra LKML-Reference: <1287173550-30365-2-git-send-email-ncrao@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 728081a..771b518 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2025,6 +2025,9 @@ task_hot(struct task_struct *p, u64 now, struct sched_domain *sd) if (p->sched_class != &fair_sched_class) return 0; + if (unlikely(p->policy == SCHED_IDLE)) + return 0; + /* * Buddy candidates are cache hot: */ -- cgit v0.10.2 From 2582f0eba54066b5e98ff2b27ef0cfa833b59f54 Mon Sep 17 00:00:00 2001 From: Nikhil Rao Date: Wed, 13 Oct 2010 12:09:36 -0700 Subject: sched: Set group_imb only a task can be pulled from the busiest cpu When cycling through sched groups to determine the busiest group, set group_imb only if the busiest cpu has more than 1 runnable task. This patch fixes the case where two cpus in a group have one runnable task each, but there is a large weight differential between these two tasks. The load balancer is unable to migrate any task from this group, and hence do not consider this group to be imbalanced. Signed-off-by: Nikhil Rao Signed-off-by: Peter Zijlstra LKML-Reference: <1286996978-7007-3-git-send-email-ncrao@google.com> [ small code readability edits ] Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index bf87192e..3656480 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2378,7 +2378,7 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, int local_group, const struct cpumask *cpus, int *balance, struct sg_lb_stats *sgs) { - unsigned long load, max_cpu_load, min_cpu_load; + unsigned long load, max_cpu_load, min_cpu_load, max_nr_running; int i; unsigned int balance_cpu = -1, first_idle_cpu = 0; unsigned long avg_load_per_task = 0; @@ -2389,6 +2389,7 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, /* Tally up the load of all CPUs in the group */ max_cpu_load = 0; min_cpu_load = ~0UL; + max_nr_running = 0; for_each_cpu_and(i, sched_group_cpus(group), cpus) { struct rq *rq = cpu_rq(i); @@ -2406,8 +2407,10 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, load = target_load(i, load_idx); } else { load = source_load(i, load_idx); - if (load > max_cpu_load) + if (load > max_cpu_load) { max_cpu_load = load; + max_nr_running = rq->nr_running; + } if (min_cpu_load > load) min_cpu_load = load; } @@ -2447,11 +2450,10 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, if (sgs->sum_nr_running) avg_load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; - if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) + if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task && max_nr_running > 1) sgs->group_imb = 1; - sgs->group_capacity = - DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); + sgs->group_capacity = DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); if (!sgs->group_capacity) sgs->group_capacity = fix_small_capacity(sd, group); } -- cgit v0.10.2 From fab476228ba37907ad75216d0fd9732ada9c119e Mon Sep 17 00:00:00 2001 From: Nikhil Rao Date: Fri, 15 Oct 2010 13:12:29 -0700 Subject: sched: Force balancing on newidle balance if local group has capacity This patch forces a load balance on a newly idle cpu when the local group has extra capacity and the busiest group does not have any. It improves system utilization when balancing tasks with a large weight differential. Under certain situations, such as a niced down task (i.e. nice = -15) in the presence of nr_cpus NICE0 tasks, the niced task lands on a sched group and kicks away other tasks because of its large weight. This leads to sub-optimal utilization of the machine. Even though the sched group has capacity, it does not pull tasks because sds.this_load >> sds.max_load, and f_b_g() returns NULL. With this patch, if the local group has extra capacity, we shortcut the checks in f_b_g() and try to pull a task over. A sched group has extra capacity if the group capacity is greater than the number of running tasks in that group. Thanks to Mike Galbraith for discussions leading to this patch and for the insight to reuse SD_NEWIDLE_BALANCE. Signed-off-by: Nikhil Rao Signed-off-by: Peter Zijlstra LKML-Reference: <1287173550-30365-4-git-send-email-ncrao@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 3656480..032b548 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1764,6 +1764,10 @@ static void pull_task(struct rq *src_rq, struct task_struct *p, set_task_cpu(p, this_cpu); activate_task(this_rq, p, 0); check_preempt_curr(this_rq, p, 0); + + /* re-arm NEWIDLE balancing when moving tasks */ + src_rq->avg_idle = this_rq->avg_idle = 2*sysctl_sched_migration_cost; + this_rq->idle_stamp = 0; } /* @@ -2030,12 +2034,14 @@ struct sd_lb_stats { unsigned long this_load; unsigned long this_load_per_task; unsigned long this_nr_running; + unsigned long this_has_capacity; /* Statistics of the busiest group */ unsigned long max_load; unsigned long busiest_load_per_task; unsigned long busiest_nr_running; unsigned long busiest_group_capacity; + unsigned long busiest_has_capacity; int group_imb; /* Is there imbalance in this sd */ #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) @@ -2058,6 +2064,7 @@ struct sg_lb_stats { unsigned long sum_weighted_load; /* Weighted load of group's tasks */ unsigned long group_capacity; int group_imb; /* Is there an imbalance in the group ? */ + int group_has_capacity; /* Is there extra capacity in the group? */ }; /** @@ -2456,6 +2463,9 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, sgs->group_capacity = DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); if (!sgs->group_capacity) sgs->group_capacity = fix_small_capacity(sd, group); + + if (sgs->group_capacity > sgs->sum_nr_running) + sgs->group_has_capacity = 1; } /** @@ -2554,12 +2564,14 @@ static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, sds->this = sg; sds->this_nr_running = sgs.sum_nr_running; sds->this_load_per_task = sgs.sum_weighted_load; + sds->this_has_capacity = sgs.group_has_capacity; } else if (update_sd_pick_busiest(sd, sds, sg, &sgs, this_cpu)) { sds->max_load = sgs.avg_load; sds->busiest = sg; sds->busiest_nr_running = sgs.sum_nr_running; sds->busiest_group_capacity = sgs.group_capacity; sds->busiest_load_per_task = sgs.sum_weighted_load; + sds->busiest_has_capacity = sgs.group_has_capacity; sds->group_imb = sgs.group_imb; } @@ -2756,6 +2768,7 @@ static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, return fix_small_imbalance(sds, this_cpu, imbalance); } + /******* find_busiest_group() helpers end here *********************/ /** @@ -2807,6 +2820,11 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, * 4) This group is more busy than the avg busieness at this * sched_domain. * 5) The imbalance is within the specified limit. + * + * Note: when doing newidle balance, if the local group has excess + * capacity (i.e. nr_running < group_capacity) and the busiest group + * does not have any capacity, we force a load balance to pull tasks + * to the local group. In this case, we skip past checks 3, 4 and 5. */ if (!(*balance)) goto ret; @@ -2818,6 +2836,11 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, if (!sds.busiest || sds.busiest_nr_running == 0) goto out_balanced; + /* SD_BALANCE_NEWIDLE trumps SMP nice when underutilized */ + if (idle == CPU_NEWLY_IDLE && sds.this_has_capacity && + !sds.busiest_has_capacity) + goto force_balance; + if (sds.this_load >= sds.max_load) goto out_balanced; @@ -2829,6 +2852,7 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) goto out_balanced; +force_balance: /* Looks like there is an imbalance. Compute it */ calculate_imbalance(&sds, this_cpu, imbalance); return sds.busiest; @@ -3162,10 +3186,8 @@ static void idle_balance(int this_cpu, struct rq *this_rq) interval = msecs_to_jiffies(sd->balance_interval); if (time_after(next_balance, sd->last_balance + interval)) next_balance = sd->last_balance + interval; - if (pulled_task) { - this_rq->idle_stamp = 0; + if (pulled_task) break; - } } raw_spin_lock(&this_rq->lock); -- cgit v0.10.2 From 75dd321d79d495a0ee579e6249ebc38ddbb2667f Mon Sep 17 00:00:00 2001 From: Nikhil Rao Date: Fri, 15 Oct 2010 13:12:30 -0700 Subject: sched: Drop group_capacity to 1 only if local group has extra capacity When SD_PREFER_SIBLING is set on a sched domain, drop group_capacity to 1 only if the local group has extra capacity. The extra check prevents the case where you always pull from the heaviest group when it is already under-utilized (possible with a large weight task outweighs the tasks on the system). For example, consider a 16-cpu quad-core quad-socket machine with MC and NUMA scheduling domains. Let's say we spawn 15 nice0 tasks and one nice-15 task, and each task is running on one core. In this case, we observe the following events when balancing at the NUMA domain: - find_busiest_group() will always pick the sched group containing the niced task to be the busiest group. - find_busiest_queue() will then always pick one of the cpus running the nice0 task (never picks the cpu with the nice -15 task since weighted_cpuload > imbalance). - The load balancer fails to migrate the task since it is the running task and increments sd->nr_balance_failed. - It repeats the above steps a few more times until sd->nr_balance_failed > 5, at which point it kicks off the active load balancer, wakes up the migration thread and kicks the nice 0 task off the cpu. The load balancer doesn't stop until we kick out all nice 0 tasks from the sched group, leaving you with 3 idle cpus and one cpu running the nice -15 task. When balancing at the NUMA domain, we drop sgs.group_capacity to 1 if the child domain (in this case MC) has SD_PREFER_SIBLING set. Subsequent load checks are not relevant because the niced task has a very large weight. In this patch, we add an extra condition to the "if(prefer_sibling)" check in update_sd_lb_stats(). We drop the capacity of a group only if the local group has extra capacity, ie. nr_running < group_capacity. This patch preserves the original intent of the prefer_siblings check (to spread tasks across the system in low utilization scenarios) and fixes the case above. It helps in the following ways: - In low utilization cases (where nr_tasks << nr_cpus), we still drop group_capacity down to 1 if we prefer siblings. - On very busy systems (where nr_tasks >> nr_cpus), sgs.nr_running will most likely be > sgs.group_capacity. - When balancing large weight tasks, if the local group does not have extra capacity, we do not pick the group with the niced task as the busiest group. This prevents failed balances, active migration and the under-utilization described above. Signed-off-by: Nikhil Rao Signed-off-by: Peter Zijlstra LKML-Reference: <1287173550-30365-5-git-send-email-ncrao@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 032b548..f1c615f 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2554,9 +2554,14 @@ static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, /* * In case the child domain prefers tasks go to siblings * first, lower the sg capacity to one so that we'll try - * and move all the excess tasks away. + * and move all the excess tasks away. We lower the capacity + * of a group only if the local group has the capacity to fit + * these excess tasks, i.e. nr_running < group_capacity. The + * extra check prevents the case where you always pull from the + * heaviest group when it is already under-utilized (possible + * with a large weight task outweighs the tasks on the system). */ - if (prefer_sibling) + if (prefer_sibling && !local_group && sds->this_has_capacity) sgs.group_capacity = min(sgs.group_capacity, 1UL); if (local_group) { -- cgit v0.10.2 From 75e1056f5c57050415b64cb761a3acc35d91f013 Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:16 -0700 Subject: sched: Fix softirq time accounting Peter Zijlstra found a bug in the way softirq time is accounted in VIRT_CPU_ACCOUNTING on this thread: http://lkml.indiana.edu/hypermail//linux/kernel/1009.2/01366.html The problem is, softirq processing uses local_bh_disable internally. There is no way, later in the flow, to differentiate between whether softirq is being processed or is it just that bh has been disabled. So, a hardirq when bh is disabled results in time being wrongly accounted as softirq. Looking at the code a bit more, the problem exists in !VIRT_CPU_ACCOUNTING as well. As account_system_time() in normal tick based accouting also uses softirq_count, which will be set even when not in softirq with bh disabled. Peter also suggested solution of using 2*SOFTIRQ_OFFSET as irq count for local_bh_{disable,enable} and using just SOFTIRQ_OFFSET while softirq processing. The patch below does that and adds API in_serving_softirq() which returns whether we are currently processing softirq or not. Also changes one of the usages of softirq_count in net/sched/cls_cgroup.c to in_serving_softirq. Looks like many usages of in_softirq really want in_serving_softirq. Those changes can be made individually on a case by case basis. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-2-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index d5b3876..e37a77c 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -64,6 +64,8 @@ #define HARDIRQ_OFFSET (1UL << HARDIRQ_SHIFT) #define NMI_OFFSET (1UL << NMI_SHIFT) +#define SOFTIRQ_DISABLE_OFFSET (2 * SOFTIRQ_OFFSET) + #ifndef PREEMPT_ACTIVE #define PREEMPT_ACTIVE_BITS 1 #define PREEMPT_ACTIVE_SHIFT (NMI_SHIFT + NMI_BITS) @@ -82,10 +84,13 @@ /* * Are we doing bottom half or hardware interrupt processing? * Are we in a softirq context? Interrupt context? + * in_softirq - Are we currently processing softirq or have bh disabled? + * in_serving_softirq - Are we currently processing softirq? */ #define in_irq() (hardirq_count()) #define in_softirq() (softirq_count()) #define in_interrupt() (irq_count()) +#define in_serving_softirq() (softirq_count() & SOFTIRQ_OFFSET) /* * Are we in NMI context? diff --git a/include/linux/sched.h b/include/linux/sched.h index cdf5669..8744e50 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2366,9 +2366,9 @@ extern int __cond_resched_lock(spinlock_t *lock); extern int __cond_resched_softirq(void); -#define cond_resched_softirq() ({ \ - __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \ - __cond_resched_softirq(); \ +#define cond_resched_softirq() ({ \ + __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \ + __cond_resched_softirq(); \ }) /* diff --git a/kernel/sched.c b/kernel/sched.c index 771b518..089be8a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3422,7 +3422,7 @@ void account_system_time(struct task_struct *p, int hardirq_offset, tmp = cputime_to_cputime64(cputime); if (hardirq_count() - hardirq_offset) cpustat->irq = cputime64_add(cpustat->irq, tmp); - else if (softirq_count()) + else if (in_serving_softirq()) cpustat->softirq = cputime64_add(cpustat->softirq, tmp); else cpustat->system = cputime64_add(cpustat->system, tmp); diff --git a/kernel/softirq.c b/kernel/softirq.c index 07b4f1b..988dfbe 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -77,11 +77,21 @@ void wakeup_softirqd(void) } /* + * preempt_count and SOFTIRQ_OFFSET usage: + * - preempt_count is changed by SOFTIRQ_OFFSET on entering or leaving + * softirq processing. + * - preempt_count is changed by SOFTIRQ_DISABLE_OFFSET (= 2 * SOFTIRQ_OFFSET) + * on local_bh_disable or local_bh_enable. + * This lets us distinguish between whether we are currently processing + * softirq and whether we just have bh disabled. + */ + +/* * This one is for softirq.c-internal use, * where hardirqs are disabled legitimately: */ #ifdef CONFIG_TRACE_IRQFLAGS -static void __local_bh_disable(unsigned long ip) +static void __local_bh_disable(unsigned long ip, unsigned int cnt) { unsigned long flags; @@ -95,32 +105,43 @@ static void __local_bh_disable(unsigned long ip) * We must manually increment preempt_count here and manually * call the trace_preempt_off later. */ - preempt_count() += SOFTIRQ_OFFSET; + preempt_count() += cnt; /* * Were softirqs turned off above: */ - if (softirq_count() == SOFTIRQ_OFFSET) + if (softirq_count() == cnt) trace_softirqs_off(ip); raw_local_irq_restore(flags); - if (preempt_count() == SOFTIRQ_OFFSET) + if (preempt_count() == cnt) trace_preempt_off(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1)); } #else /* !CONFIG_TRACE_IRQFLAGS */ -static inline void __local_bh_disable(unsigned long ip) +static inline void __local_bh_disable(unsigned long ip, unsigned int cnt) { - add_preempt_count(SOFTIRQ_OFFSET); + add_preempt_count(cnt); barrier(); } #endif /* CONFIG_TRACE_IRQFLAGS */ void local_bh_disable(void) { - __local_bh_disable((unsigned long)__builtin_return_address(0)); + __local_bh_disable((unsigned long)__builtin_return_address(0), + SOFTIRQ_DISABLE_OFFSET); } EXPORT_SYMBOL(local_bh_disable); +static void __local_bh_enable(unsigned int cnt) +{ + WARN_ON_ONCE(in_irq()); + WARN_ON_ONCE(!irqs_disabled()); + + if (softirq_count() == cnt) + trace_softirqs_on((unsigned long)__builtin_return_address(0)); + sub_preempt_count(cnt); +} + /* * Special-case - softirqs can safely be enabled in * cond_resched_softirq(), or by __do_softirq(), @@ -128,12 +149,7 @@ EXPORT_SYMBOL(local_bh_disable); */ void _local_bh_enable(void) { - WARN_ON_ONCE(in_irq()); - WARN_ON_ONCE(!irqs_disabled()); - - if (softirq_count() == SOFTIRQ_OFFSET) - trace_softirqs_on((unsigned long)__builtin_return_address(0)); - sub_preempt_count(SOFTIRQ_OFFSET); + __local_bh_enable(SOFTIRQ_DISABLE_OFFSET); } EXPORT_SYMBOL(_local_bh_enable); @@ -147,13 +163,13 @@ static inline void _local_bh_enable_ip(unsigned long ip) /* * Are softirqs going to be turned on now: */ - if (softirq_count() == SOFTIRQ_OFFSET) + if (softirq_count() == SOFTIRQ_DISABLE_OFFSET) trace_softirqs_on(ip); /* * Keep preemption disabled until we are done with * softirq processing: */ - sub_preempt_count(SOFTIRQ_OFFSET - 1); + sub_preempt_count(SOFTIRQ_DISABLE_OFFSET - 1); if (unlikely(!in_interrupt() && local_softirq_pending())) do_softirq(); @@ -198,7 +214,8 @@ asmlinkage void __do_softirq(void) pending = local_softirq_pending(); account_system_vtime(current); - __local_bh_disable((unsigned long)__builtin_return_address(0)); + __local_bh_disable((unsigned long)__builtin_return_address(0), + SOFTIRQ_OFFSET); lockdep_softirq_enter(); cpu = smp_processor_id(); @@ -245,7 +262,7 @@ restart: lockdep_softirq_exit(); account_system_vtime(current); - _local_bh_enable(); + __local_bh_enable(SOFTIRQ_OFFSET); } #ifndef __ARCH_HAS_DO_SOFTIRQ diff --git a/net/sched/cls_cgroup.c b/net/sched/cls_cgroup.c index 78ef2c5..37dff78 100644 --- a/net/sched/cls_cgroup.c +++ b/net/sched/cls_cgroup.c @@ -123,7 +123,7 @@ static int cls_cgroup_classify(struct sk_buff *skb, struct tcf_proto *tp, * calls by looking at the number of nested bh disable calls because * softirqs always disables bh. */ - if (softirq_count() != SOFTIRQ_OFFSET) { + if (in_serving_softirq()) { /* If there is an sk_classid we'll use that. */ if (!skb->sk) return -1; -- cgit v0.10.2 From e1e10a265d28273ab8c70be19d43dcbdeead6c5a Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:17 -0700 Subject: sched: Consolidate account_system_vtime extern declaration Just a minor cleanup patch that makes things easier to the following patches. No functionality change in this patch. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-3-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/arch/ia64/include/asm/system.h b/arch/ia64/include/asm/system.h index 9f342a5..dd028f2 100644 --- a/arch/ia64/include/asm/system.h +++ b/arch/ia64/include/asm/system.h @@ -272,10 +272,6 @@ void cpu_idle_wait(void); void default_idle(void); -#ifdef CONFIG_VIRT_CPU_ACCOUNTING -extern void account_system_vtime(struct task_struct *); -#endif - #endif /* __KERNEL__ */ #endif /* __ASSEMBLY__ */ diff --git a/arch/powerpc/include/asm/system.h b/arch/powerpc/include/asm/system.h index 6c294ac..9c3d160 100644 --- a/arch/powerpc/include/asm/system.h +++ b/arch/powerpc/include/asm/system.h @@ -542,10 +542,6 @@ extern void reloc_got2(unsigned long); #define PTRRELOC(x) ((typeof(x)) add_reloc_offset((unsigned long)(x))) -#ifdef CONFIG_VIRT_CPU_ACCOUNTING -extern void account_system_vtime(struct task_struct *); -#endif - extern struct dentry *powerpc_debugfs_root; #endif /* __KERNEL__ */ diff --git a/arch/s390/include/asm/system.h b/arch/s390/include/asm/system.h index cef6621..38ddd8a 100644 --- a/arch/s390/include/asm/system.h +++ b/arch/s390/include/asm/system.h @@ -97,7 +97,6 @@ static inline void restore_access_regs(unsigned int *acrs) extern void account_vtime(struct task_struct *, struct task_struct *); extern void account_tick_vtime(struct task_struct *); -extern void account_system_vtime(struct task_struct *); #ifdef CONFIG_PFAULT extern void pfault_irq_init(void); diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index e37a77c..41367c5 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -141,6 +141,8 @@ struct task_struct; static inline void account_system_vtime(struct task_struct *tsk) { } +#else +extern void account_system_vtime(struct task_struct *tsk); #endif #if defined(CONFIG_NO_HZ) -- cgit v0.10.2 From 6cdd5199daf0cb7b0fcc8dca941af08492612887 Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:18 -0700 Subject: sched: Add a PF flag for ksoftirqd identification To account softirq time cleanly in scheduler, we need to identify whether softirq is invoked in ksoftirqd context or softirq at hardirq tail context. Add PF_KSOFTIRQD for that purpose. As all PF flag bits are currently taken, create space by moving one of the infrequently used bits (PF_THREAD_BOUND) down in task_struct to be along with some other state fields. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-4-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/sched.h b/include/linux/sched.h index 8744e50..aca0ce6 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1682,6 +1682,7 @@ extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t * /* * Per process flags */ +#define PF_KSOFTIRQD 0x00000001 /* I am ksoftirqd */ #define PF_STARTING 0x00000002 /* being created */ #define PF_EXITING 0x00000004 /* getting shut down */ #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */ diff --git a/kernel/softirq.c b/kernel/softirq.c index 988dfbe..267f7b7 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -713,6 +713,7 @@ static int run_ksoftirqd(void * __bind_cpu) { set_current_state(TASK_INTERRUPTIBLE); + current->flags |= PF_KSOFTIRQD; while (!kthread_should_stop()) { preempt_disable(); if (!local_softirq_pending()) { -- cgit v0.10.2 From b52bfee445d315549d41eacf2fa7c156e7d153d5 Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:19 -0700 Subject: sched: Add IRQ_TIME_ACCOUNTING, finer accounting of irq time s390/powerpc/ia64 have support for CONFIG_VIRT_CPU_ACCOUNTING which does the fine granularity accounting of user, system, hardirq, softirq times. Adding that option on archs like x86 will be challenging however, given the state of TSC reliability on various platforms and also the overhead it will add in syscall entry exit. Instead, add a lighter variant that only does finer accounting of hardirq and softirq times, providing precise irq times (instead of timer tick based samples). This accounting is added with a new config option CONFIG_IRQ_TIME_ACCOUNTING so that there won't be any overhead for users not interested in paying the perf penalty. This accounting is based on sched_clock, with the code being generic. So, other archs may find it useful as well. This patch just adds the core logic and does not enable this logic yet. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-5-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 41367c5..ff43e92 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -137,7 +137,7 @@ extern void synchronize_irq(unsigned int irq); struct task_struct; -#ifndef CONFIG_VIRT_CPU_ACCOUNTING +#if !defined(CONFIG_VIRT_CPU_ACCOUNTING) && !defined(CONFIG_IRQ_TIME_ACCOUNTING) static inline void account_system_vtime(struct task_struct *tsk) { } diff --git a/include/linux/sched.h b/include/linux/sched.h index aca0ce6..2cca9a9 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1826,6 +1826,19 @@ extern void sched_clock_idle_sleep_event(void); extern void sched_clock_idle_wakeup_event(u64 delta_ns); #endif +#ifdef CONFIG_IRQ_TIME_ACCOUNTING +/* + * An i/f to runtime opt-in for irq time accounting based off of sched_clock. + * The reason for this explicit opt-in is not to have perf penalty with + * slow sched_clocks. + */ +extern void enable_sched_clock_irqtime(void); +extern void disable_sched_clock_irqtime(void); +#else +static inline void enable_sched_clock_irqtime(void) {} +static inline void disable_sched_clock_irqtime(void) {} +#endif + extern unsigned long long task_sched_runtime(struct task_struct *task); extern unsigned long long thread_group_sched_runtime(struct task_struct *task); diff --git a/kernel/sched.c b/kernel/sched.c index 089be8a..9b302e3 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1908,6 +1908,55 @@ static void deactivate_task(struct rq *rq, struct task_struct *p, int flags) dec_nr_running(rq); } +#ifdef CONFIG_IRQ_TIME_ACCOUNTING + +static DEFINE_PER_CPU(u64, cpu_hardirq_time); +static DEFINE_PER_CPU(u64, cpu_softirq_time); + +static DEFINE_PER_CPU(u64, irq_start_time); +static int sched_clock_irqtime; + +void enable_sched_clock_irqtime(void) +{ + sched_clock_irqtime = 1; +} + +void disable_sched_clock_irqtime(void) +{ + sched_clock_irqtime = 0; +} + +void account_system_vtime(struct task_struct *curr) +{ + unsigned long flags; + int cpu; + u64 now, delta; + + if (!sched_clock_irqtime) + return; + + local_irq_save(flags); + + now = sched_clock(); + cpu = smp_processor_id(); + delta = now - per_cpu(irq_start_time, cpu); + per_cpu(irq_start_time, cpu) = now; + /* + * We do not account for softirq time from ksoftirqd here. + * We want to continue accounting softirq time to ksoftirqd thread + * in that case, so as not to confuse scheduler with a special task + * that do not consume any time, but still wants to run. + */ + if (hardirq_count()) + per_cpu(cpu_hardirq_time, cpu) += delta; + else if (in_serving_softirq() && !(curr->flags & PF_KSOFTIRQD)) + per_cpu(cpu_softirq_time, cpu) += delta; + + local_irq_restore(flags); +} + +#endif + #include "sched_idletask.c" #include "sched_fair.c" #include "sched_rt.c" -- cgit v0.10.2 From e82b8e4ea4f3dffe6e7939f90e78da675fcc450e Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:20 -0700 Subject: x86: Add IRQ_TIME_ACCOUNTING This patch adds IRQ_TIME_ACCOUNTING option on x86 and runtime enables it when TSC is enabled. This change just enables fine grained irq time accounting, isn't used yet. Following patches use it for different purposes. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-6-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 8dd7248..ed05a4a 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -2435,6 +2435,10 @@ and is between 256 and 4096 characters. It is defined in the file disables clocksource verification at runtime. Used to enable high-resolution timer mode on older hardware, and in virtualized environment. + [x86] noirqtime: Do not use TSC to do irq accounting. + Used to run time disable IRQ_TIME_ACCOUNTING on any + platforms where RDTSC is slow and this accounting + can add overhead. turbografx.map[2|3]= [HW,JOY] TurboGraFX parallel port interface diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..f4c70c2 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -795,6 +795,17 @@ config SCHED_MC making when dealing with multi-core CPU chips at a cost of slightly increased overhead in some places. If unsure say N here. +config IRQ_TIME_ACCOUNTING + bool "Fine granularity task level IRQ time accounting" + default n + ---help--- + Select this option to enable fine granularity task irq time + accounting. This is done by reading a timestamp on each + transitions between softirq and hardirq state, so there can be a + small performance impact. + + If in doubt, say N here. + source "kernel/Kconfig.preempt" config X86_UP_APIC diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index 26a863a..a1c2cd7 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -104,10 +104,14 @@ int __init notsc_setup(char *str) __setup("notsc", notsc_setup); +static int no_sched_irq_time; + static int __init tsc_setup(char *str) { if (!strcmp(str, "reliable")) tsc_clocksource_reliable = 1; + if (!strncmp(str, "noirqtime", 9)) + no_sched_irq_time = 1; return 1; } @@ -801,6 +805,7 @@ void mark_tsc_unstable(char *reason) if (!tsc_unstable) { tsc_unstable = 1; sched_clock_stable = 0; + disable_sched_clock_irqtime(); printk(KERN_INFO "Marking TSC unstable due to %s\n", reason); /* Change only the rating, when not registered */ if (clocksource_tsc.mult) @@ -987,6 +992,9 @@ void __init tsc_init(void) /* now allow native_sched_clock() to use rdtsc */ tsc_disabled = 0; + if (!no_sched_irq_time) + enable_sched_clock_irqtime(); + lpj = ((u64)tsc_khz * 1000); do_div(lpj, HZ); lpj_fine = lpj; -- cgit v0.10.2 From 305e6835e05513406fa12820e40e4a8ecb63743c Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:21 -0700 Subject: sched: Do not account irq time to current task Scheduler accounts both softirq and interrupt processing times to the currently running task. This means, if the interrupt processing was for some other task in the system, then the current task ends up being penalized as it gets shorter runtime than otherwise. Change sched task accounting to acoount only actual task time from currently running task. Now update_curr(), modifies the delta_exec to depend on rq->clock_task. Note that this change only handles CONFIG_IRQ_TIME_ACCOUNTING case. We can extend this to CONFIG_VIRT_CPU_ACCOUNTING with minimal effort. But, thats for later. This change will impact scheduling behavior in interrupt heavy conditions. Tested on a 4-way system with eth0 handled by CPU 2 and a network heavy task (nc) running on CPU 3 (and no RSS/RFS). With that I have CPU 2 spending 75%+ of its time in irq processing. CPU 3 spending around 35% time running nc task. Now, if I run another CPU intensive task on CPU 2, without this change /proc//schedstat shows 100% of time accounted to this task. With this change, it rightly shows less than 25% accounted to this task as remaining time is actually spent on irq processing. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-7-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 9b302e3..9e01b71 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -491,6 +491,7 @@ struct rq { struct mm_struct *prev_mm; u64 clock; + u64 clock_task; atomic_t nr_iowait; @@ -641,10 +642,19 @@ static inline struct task_group *task_group(struct task_struct *p) #endif /* CONFIG_CGROUP_SCHED */ +static u64 irq_time_cpu(int cpu); + inline void update_rq_clock(struct rq *rq) { - if (!rq->skip_clock_update) - rq->clock = sched_clock_cpu(cpu_of(rq)); + if (!rq->skip_clock_update) { + int cpu = cpu_of(rq); + u64 irq_time; + + rq->clock = sched_clock_cpu(cpu); + irq_time = irq_time_cpu(cpu); + if (rq->clock - irq_time > rq->clock_task) + rq->clock_task = rq->clock - irq_time; + } } /* @@ -1910,6 +1920,18 @@ static void deactivate_task(struct rq *rq, struct task_struct *p, int flags) #ifdef CONFIG_IRQ_TIME_ACCOUNTING +/* + * There are no locks covering percpu hardirq/softirq time. + * They are only modified in account_system_vtime, on corresponding CPU + * with interrupts disabled. So, writes are safe. + * They are read and saved off onto struct rq in update_rq_clock(). + * This may result in other CPU reading this CPU's irq time and can + * race with irq/account_system_vtime on this CPU. We would either get old + * or new value (or semi updated value on 32 bit) with a side effect of + * accounting a slice of irq time to wrong task when irq is in progress + * while we read rq->clock. That is a worthy compromise in place of having + * locks on each irq in account_system_time. + */ static DEFINE_PER_CPU(u64, cpu_hardirq_time); static DEFINE_PER_CPU(u64, cpu_softirq_time); @@ -1926,6 +1948,14 @@ void disable_sched_clock_irqtime(void) sched_clock_irqtime = 0; } +static u64 irq_time_cpu(int cpu) +{ + if (!sched_clock_irqtime) + return 0; + + return per_cpu(cpu_softirq_time, cpu) + per_cpu(cpu_hardirq_time, cpu); +} + void account_system_vtime(struct task_struct *curr) { unsigned long flags; @@ -1955,6 +1985,13 @@ void account_system_vtime(struct task_struct *curr) local_irq_restore(flags); } +#else + +static u64 irq_time_cpu(int cpu) +{ + return 0; +} + #endif #include "sched_idletask.c" @@ -3322,7 +3359,7 @@ static u64 do_task_delta_exec(struct task_struct *p, struct rq *rq) if (task_current(rq, p)) { update_rq_clock(rq); - ns = rq->clock - p->se.exec_start; + ns = rq->clock_task - p->se.exec_start; if ((s64)ns < 0) ns = 0; } diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index f1c615f..c358d40 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -519,7 +519,7 @@ __update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr, static void update_curr(struct cfs_rq *cfs_rq) { struct sched_entity *curr = cfs_rq->curr; - u64 now = rq_of(cfs_rq)->clock; + u64 now = rq_of(cfs_rq)->clock_task; unsigned long delta_exec; if (unlikely(!curr)) @@ -602,7 +602,7 @@ update_stats_curr_start(struct cfs_rq *cfs_rq, struct sched_entity *se) /* * We are starting a new run period: */ - se->exec_start = rq_of(cfs_rq)->clock; + se->exec_start = rq_of(cfs_rq)->clock_task; } /************************************************** @@ -1802,7 +1802,7 @@ int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, * 2) too many balance attempts have failed. */ - tsk_cache_hot = task_hot(p, rq->clock, sd); + tsk_cache_hot = task_hot(p, rq->clock_task, sd); if (!tsk_cache_hot || sd->nr_balance_failed > sd->cache_nice_tries) { #ifdef CONFIG_SCHEDSTATS diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index ab77aa0..bea7d79 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -609,7 +609,7 @@ static void update_curr_rt(struct rq *rq) if (!task_has_rt_policy(curr)) return; - delta_exec = rq->clock - curr->se.exec_start; + delta_exec = rq->clock_task - curr->se.exec_start; if (unlikely((s64)delta_exec < 0)) delta_exec = 0; @@ -618,7 +618,7 @@ static void update_curr_rt(struct rq *rq) curr->se.sum_exec_runtime += delta_exec; account_group_exec_runtime(curr, delta_exec); - curr->se.exec_start = rq->clock; + curr->se.exec_start = rq->clock_task; cpuacct_charge(curr, delta_exec); sched_rt_avg_update(rq, delta_exec); @@ -1075,7 +1075,7 @@ static struct task_struct *_pick_next_task_rt(struct rq *rq) } while (rt_rq); p = rt_task_of(rt_se); - p->se.exec_start = rq->clock; + p->se.exec_start = rq->clock_task; return p; } @@ -1713,7 +1713,7 @@ static void set_curr_task_rt(struct rq *rq) { struct task_struct *p = rq->curr; - p->se.exec_start = rq->clock; + p->se.exec_start = rq->clock_task; /* The running task is never eligible for pushing */ dequeue_pushable_task(rq, p); -- cgit v0.10.2 From aa483808516ca5cacfa0e5849691f64fec25828e Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:22 -0700 Subject: sched: Remove irq time from available CPU power The idea was suggested by Peter Zijlstra here: http://marc.info/?l=linux-kernel&m=127476934517534&w=2 irq time is technically not available to the tasks running on the CPU. This patch removes irq time from CPU power piggybacking on sched_rt_avg_update(). Tested this by keeping CPU X busy with a network intensive task having 75% oa a single CPU irq processing (hard+soft) on a 4-way system. And start seven cycle soakers on the system. Without this change, there will be two tasks on each CPU. With this change, there is a single task on irq busy CPU X and remaining 7 tasks are spread around among other 3 CPUs. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-8-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 9e01b71..bff9ef5 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -519,6 +519,10 @@ struct rq { u64 avg_idle; #endif +#ifdef CONFIG_IRQ_TIME_ACCOUNTING + u64 prev_irq_time; +#endif + /* calc_load related fields */ unsigned long calc_load_update; long calc_load_active; @@ -643,6 +647,7 @@ static inline struct task_group *task_group(struct task_struct *p) #endif /* CONFIG_CGROUP_SCHED */ static u64 irq_time_cpu(int cpu); +static void sched_irq_time_avg_update(struct rq *rq, u64 irq_time); inline void update_rq_clock(struct rq *rq) { @@ -654,6 +659,8 @@ inline void update_rq_clock(struct rq *rq) irq_time = irq_time_cpu(cpu); if (rq->clock - irq_time > rq->clock_task) rq->clock_task = rq->clock - irq_time; + + sched_irq_time_avg_update(rq, irq_time); } } @@ -1985,6 +1992,15 @@ void account_system_vtime(struct task_struct *curr) local_irq_restore(flags); } +static void sched_irq_time_avg_update(struct rq *rq, u64 curr_irq_time) +{ + if (sched_clock_irqtime && sched_feat(NONIRQ_POWER)) { + u64 delta_irq = curr_irq_time - rq->prev_irq_time; + rq->prev_irq_time = curr_irq_time; + sched_rt_avg_update(rq, delta_irq); + } +} + #else static u64 irq_time_cpu(int cpu) @@ -1992,6 +2008,8 @@ static u64 irq_time_cpu(int cpu) return 0; } +static void sched_irq_time_avg_update(struct rq *rq, u64 curr_irq_time) { } + #endif #include "sched_idletask.c" diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index c358d40..74cccfa 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2275,7 +2275,13 @@ unsigned long scale_rt_power(int cpu) u64 total, available; total = sched_avg_period() + (rq->clock - rq->age_stamp); - available = total - rq->rt_avg; + + if (unlikely(total < rq->rt_avg)) { + /* Ensures that power won't end up being negative */ + available = 0; + } else { + available = total - rq->rt_avg; + } if (unlikely((s64)total < SCHED_LOAD_SCALE)) total = SCHED_LOAD_SCALE; diff --git a/kernel/sched_features.h b/kernel/sched_features.h index 83c66e8..185f920 100644 --- a/kernel/sched_features.h +++ b/kernel/sched_features.h @@ -61,3 +61,8 @@ SCHED_FEAT(ASYM_EFF_LOAD, 1) * release the lock. Decreases scheduling overhead. */ SCHED_FEAT(OWNER_SPIN, 1) + +/* + * Decrement CPU power based on irq activity + */ +SCHED_FEAT(NONIRQ_POWER, 1) -- cgit v0.10.2 From d267f87fb8179c6dba03d08b91952e81bc3723c7 Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Mon, 4 Oct 2010 17:03:23 -0700 Subject: sched: Call tick_check_idle before __irq_enter When CPU is idle and on first interrupt, irq_enter calls tick_check_idle() to notify interruption from idle. But, there is a problem if this call is done after __irq_enter, as all routines in __irq_enter may find stale time due to yet to be done tick_check_idle. Specifically, trace calls in __irq_enter when they use global clock and also account_system_vtime change in this patch as it wants to use sched_clock_cpu() to do proper irq timing. But, tick_check_idle was moved after __irq_enter intentionally to prevent problem of unneeded ksoftirqd wakeups by the commit ee5f80a: irq: call __irq_enter() before calling the tick_idle_check Impact: avoid spurious ksoftirqd wakeups Moving tick_check_idle() before __irq_enter and wrapping it with local_bh_enable/disable would solve both the problems. Fixed-by: Yong Zhang Signed-off-by: Venkatesh Pallipadi Signed-off-by: Peter Zijlstra LKML-Reference: <1286237003-12406-9-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index bff9ef5..567f5cb 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1974,8 +1974,8 @@ void account_system_vtime(struct task_struct *curr) local_irq_save(flags); - now = sched_clock(); cpu = smp_processor_id(); + now = sched_clock_cpu(cpu); delta = now - per_cpu(irq_start_time, cpu); per_cpu(irq_start_time, cpu) = now; /* diff --git a/kernel/softirq.c b/kernel/softirq.c index 267f7b7..79ee8f1 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -296,10 +296,16 @@ void irq_enter(void) rcu_irq_enter(); if (idle_cpu(cpu) && !in_interrupt()) { - __irq_enter(); + /* + * Prevent raise_softirq from needlessly waking up ksoftirqd + * here, as softirq will be serviced on return from interrupt. + */ + local_bh_disable(); tick_check_idle(cpu); - } else - __irq_enter(); + _local_bh_enable(); + } + + __irq_enter(); } #ifdef __ARCH_IRQ_EXIT_IRQS_DISABLED -- cgit v0.10.2 From b7dadc38797584f6203386da1947ed5edf516646 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 18 Oct 2010 20:00:37 +0200 Subject: sched: Export account_system_vtime() KVM uses it for example: ERROR: "account_system_vtime" [arch/x86/kvm/kvm.ko] undefined! Cc: Venkatesh Pallipadi Cc: Peter Zijlstra LKML-Reference: <1286237003-12406-3-git-send-email-venki@google.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 567f5cb..5998222 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1991,6 +1991,7 @@ void account_system_vtime(struct task_struct *curr) local_irq_restore(flags); } +EXPORT_SYMBOL_GPL(account_system_vtime); static void sched_irq_time_avg_update(struct rq *rq, u64 curr_irq_time) { -- cgit v0.10.2 From 63f1474c69c061f923068a25f136dca0c49cd478 Mon Sep 17 00:00:00 2001 From: Sascha Hauer Date: Mon, 18 Oct 2010 10:16:26 +0200 Subject: mxc_nand: do not depend on disabling the irq in the interrupt handler This patch reverts the driver to enabling/disabling the NFC interrupt mask rather than enabling/disabling the system interrupt. This cleans up the driver so that it doesn't rely on interrupts being disabled within the interrupt handler. For i.MX21 we keep the current behaviour, that is calling enable_irq/disable_irq_nosync to enable/disable interrupts. This patch is based on earlier work by John Ogness. Signed-off-by: Sascha Hauer Acked-by: John Ogness Tested-by: John Ogness Signed-off-by: David Woodhouse Signed-off-by: Linus Torvalds diff --git a/drivers/mtd/nand/mxc_nand.c b/drivers/mtd/nand/mxc_nand.c index b2828e8..214b03a 100644 --- a/drivers/mtd/nand/mxc_nand.c +++ b/drivers/mtd/nand/mxc_nand.c @@ -30,6 +30,8 @@ #include #include #include +#include +#include #include #include @@ -151,7 +153,7 @@ struct mxc_nand_host { int irq; int eccsize; - wait_queue_head_t irq_waitq; + struct completion op_completion; uint8_t *data_buf; unsigned int buf_start; @@ -164,6 +166,7 @@ struct mxc_nand_host { void (*send_read_id)(struct mxc_nand_host *); uint16_t (*get_dev_status)(struct mxc_nand_host *); int (*check_int)(struct mxc_nand_host *); + void (*irq_control)(struct mxc_nand_host *, int); }; /* OOB placement block for use with hardware ecc generation */ @@ -216,9 +219,12 @@ static irqreturn_t mxc_nfc_irq(int irq, void *dev_id) { struct mxc_nand_host *host = dev_id; - disable_irq_nosync(irq); + if (!host->check_int(host)) + return IRQ_NONE; - wake_up(&host->irq_waitq); + host->irq_control(host, 0); + + complete(&host->op_completion); return IRQ_HANDLED; } @@ -245,11 +251,54 @@ static int check_int_v1_v2(struct mxc_nand_host *host) if (!(tmp & NFC_V1_V2_CONFIG2_INT)) return 0; - writew(tmp & ~NFC_V1_V2_CONFIG2_INT, NFC_V1_V2_CONFIG2); + if (!cpu_is_mx21()) + writew(tmp & ~NFC_V1_V2_CONFIG2_INT, NFC_V1_V2_CONFIG2); return 1; } +/* + * It has been observed that the i.MX21 cannot read the CONFIG2:INT bit + * if interrupts are masked (CONFIG1:INT_MSK is set). To handle this, the + * driver can enable/disable the irq line rather than simply masking the + * interrupts. + */ +static void irq_control_mx21(struct mxc_nand_host *host, int activate) +{ + if (activate) + enable_irq(host->irq); + else + disable_irq_nosync(host->irq); +} + +static void irq_control_v1_v2(struct mxc_nand_host *host, int activate) +{ + uint16_t tmp; + + tmp = readw(NFC_V1_V2_CONFIG1); + + if (activate) + tmp &= ~NFC_V1_V2_CONFIG1_INT_MSK; + else + tmp |= NFC_V1_V2_CONFIG1_INT_MSK; + + writew(tmp, NFC_V1_V2_CONFIG1); +} + +static void irq_control_v3(struct mxc_nand_host *host, int activate) +{ + uint32_t tmp; + + tmp = readl(NFC_V3_CONFIG2); + + if (activate) + tmp &= ~NFC_V3_CONFIG2_INT_MSK; + else + tmp |= NFC_V3_CONFIG2_INT_MSK; + + writel(tmp, NFC_V3_CONFIG2); +} + /* This function polls the NANDFC to wait for the basic operation to * complete by checking the INT bit of config2 register. */ @@ -259,10 +308,9 @@ static void wait_op_done(struct mxc_nand_host *host, int useirq) if (useirq) { if (!host->check_int(host)) { - - enable_irq(host->irq); - - wait_event(host->irq_waitq, host->check_int(host)); + INIT_COMPLETION(host->op_completion); + host->irq_control(host, 1); + wait_for_completion(&host->op_completion); } } else { while (max_retries-- > 0) { @@ -799,6 +847,7 @@ static void preset_v3(struct mtd_info *mtd) NFC_V3_CONFIG2_2CMD_PHASES | NFC_V3_CONFIG2_SPAS(mtd->oobsize >> 1) | NFC_V3_CONFIG2_ST_CMD(0x70) | + NFC_V3_CONFIG2_INT_MSK | NFC_V3_CONFIG2_NUM_ADDR_PHASE0; if (chip->ecc.mode == NAND_ECC_HW) @@ -1024,6 +1073,10 @@ static int __init mxcnd_probe(struct platform_device *pdev) host->send_read_id = send_read_id_v1_v2; host->get_dev_status = get_dev_status_v1_v2; host->check_int = check_int_v1_v2; + if (cpu_is_mx21()) + host->irq_control = irq_control_mx21; + else + host->irq_control = irq_control_v1_v2; } if (nfc_is_v21()) { @@ -1062,6 +1115,7 @@ static int __init mxcnd_probe(struct platform_device *pdev) host->send_read_id = send_read_id_v3; host->check_int = check_int_v3; host->get_dev_status = get_dev_status_v3; + host->irq_control = irq_control_v3; oob_smallpage = &nandv2_hw_eccoob_smallpage; oob_largepage = &nandv2_hw_eccoob_largepage; } else @@ -1093,14 +1147,34 @@ static int __init mxcnd_probe(struct platform_device *pdev) this->options |= NAND_USE_FLASH_BBT; } - init_waitqueue_head(&host->irq_waitq); + init_completion(&host->op_completion); host->irq = platform_get_irq(pdev, 0); + /* + * mask the interrupt. For i.MX21 explicitely call + * irq_control_v1_v2 to use the mask bit. We can't call + * disable_irq_nosync() for an interrupt we do not own yet. + */ + if (cpu_is_mx21()) + irq_control_v1_v2(host, 0); + else + host->irq_control(host, 0); + err = request_irq(host->irq, mxc_nfc_irq, IRQF_DISABLED, DRIVER_NAME, host); if (err) goto eirq; + host->irq_control(host, 0); + + /* + * Now that the interrupt is disabled make sure the interrupt + * mask bit is cleared on i.MX21. Otherwise we can't read + * the interrupt status bit on this machine. + */ + if (cpu_is_mx21()) + irq_control_v1_v2(host, 1); + /* first scan to find the device and get the page size */ if (nand_scan_ident(mtd, 1, NULL)) { err = -ENXIO; -- cgit v0.10.2 From d7b4d6de57d414a6384376880f2caf7125a45494 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 18 Oct 2010 14:42:00 -0400 Subject: ftrace: Remove recursion between recordmcount and scripts/mod/empty When DYNAMIC_FTRACE is enabled and we use the C version of recordmcount, all objects are run through the recordmcount program to create a separate section that stores all the callers of mcount. The build process has a special file: scripts/mod/empty.o. This is built from empty.c which is literally an empty file (except for a single comment). This file is used to find information about the target elf format, like endianness and word size. The problem comes up when we need to build recordmcount. The build process requires that empty.o is built first. The build rules for empty.o will try to execute recordmcount on the empty.o file. We get an error that recordmcount does not exist. To avoid this recursion, the build file will skip running recordmcount if the file that it is building is script/mod/empty.o. [ extra comment Suggested-by: Sam Ravnborg ] Reported-by: Ingo Molnar Tested-by: Ingo Molnar Cc: Michal Marek Cc: linux-kbuild@vger.kernel.org Signed-off-by: Steven Rostedt diff --git a/scripts/Makefile.build b/scripts/Makefile.build index 4db60b2..843bd4f 100644 --- a/scripts/Makefile.build +++ b/scripts/Makefile.build @@ -210,7 +210,13 @@ endif ifdef CONFIG_FTRACE_MCOUNT_RECORD ifdef BUILD_C_RECORDMCOUNT -cmd_record_mcount = $(objtree)/scripts/recordmcount "$(@)"; +# Due to recursion, we must skip empty.o. +# The empty.o file is created in the make process in order to determine +# the target endianness and word size. It is made before all other C +# files, including recordmcount. +cmd_record_mcount = if [ $(@) != "scripts/mod/empty.o" ]; then \ + $(objtree)/scripts/recordmcount "$(@)"; \ + fi; else cmd_record_mcount = set -e ; perl $(srctree)/scripts/recordmcount.pl "$(ARCH)" \ "$(if $(CONFIG_CPU_BIG_ENDIAN),big,little)" \ -- cgit v0.10.2 From 57fa7214330be2e292ddb1402834ff0b221ef29a Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Tue, 19 Oct 2010 16:55:35 +1100 Subject: perf, powerpc: Fix power_pmu_event_init to not use event->ctx Commit c3f00c70 ("perf: Separate find_get_context() from event initialization") changed the generic perf_event code to call perf_event_alloc, which calls the arch-specific event_init code, before looking up the context for the new event. Unfortunately, power_pmu_event_init uses event->ctx->task to see whether the new event is a per-task event or a system-wide event, and thus crashes since event->ctx is NULL at the point where power_pmu_event_init gets called. (The reason it needs to know whether it is a per-task event is because there are some hardware events on Power systems which only count when the processor is not idle, and there are some fixed-function counters which count such events. For example, the "run cycles" event counts cycles when the processor is not idle. If the user asks to count cycles, we can use "run cycles" if this is a per-task event, since the processor is running when the task is running, by definition. We can't use "run cycles" if the user asks for "cycles" on a system-wide counter.) Fortunately the information we need is in the event->attach_state field, so we just use that instead. Signed-off-by: Paul Mackerras Cc: Benjamin Herrenschmidt Cc: Peter Zijlstra Cc: Frederic Weisbecker LKML-Reference: <20101019055535.GA10398@drongo> Signed-off-by: Ingo Molnar Reported-by: Alexey Kardashevskiy Signed-off-by: Ingo Molnar diff --git a/arch/powerpc/kernel/perf_event.c b/arch/powerpc/kernel/perf_event.c index 9cb4924..3129c85 100644 --- a/arch/powerpc/kernel/perf_event.c +++ b/arch/powerpc/kernel/perf_event.c @@ -1092,7 +1092,7 @@ static int power_pmu_event_init(struct perf_event *event) * XXX we should check if the task is an idle task. */ flags = 0; - if (event->ctx->task) + if (event->attach_state & PERF_ATTACH_TASK) flags |= PPMU_ONLY_COUNT_RUN; /* -- cgit v0.10.2 From 9717967c4b704ce344c954afb5bb160aa9c01c34 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 18 Oct 2010 13:47:48 -0700 Subject: x86: ioapic: Call free_irte only if interrupt remapping enabled On a system that support intr-rempping when booting with "intremap=off" [ 177.895501] BUG: unable to handle kernel NULL pointer dereference at 00000000000000f8 [ 177.913316] IP: [] free_irte+0x47/0xc0 ... [ 178.173326] Call Trace: [ 178.173574] [] destroy_irq+0x3a/0x75 [ 178.192934] [] arch_teardown_msi_irq+0xe/0x10 [ 178.193418] [] arch_teardown_msi_irqs+0x56/0x7f [ 178.213021] [] free_msi_irqs+0x8d/0xeb Call free_irte only when interrupt remapping is enabled. Signed-off-by: Yinghai Lu LKML-Reference: <4CBCB274.7010108@kernel.org> Signed-off-by: Thomas Gleixner diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 20e47e0..8ae808d 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3109,7 +3109,8 @@ void destroy_irq(unsigned int irq) irq_set_status_flags(irq, IRQ_NOREQUEST|IRQ_NOPROBE); - free_irte(irq); + if (intr_remapping_enabled) + free_irte(irq); raw_spin_lock_irqsave(&vector_lock, flags); __clear_irq_vector(irq, cfg); raw_spin_unlock_irqrestore(&vector_lock, flags); -- cgit v0.10.2 From 7ada876a8703f23befbb20a7465a702ee39b1704 Mon Sep 17 00:00:00 2001 From: Darren Hart Date: Sun, 17 Oct 2010 08:35:04 -0700 Subject: futex: Fix errors in nested key ref-counting MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit futex_wait() is leaking key references due to futex_wait_setup() acquiring an additional reference via the queue_lock() routine. The nested key ref-counting has been masking bugs and complicating code analysis. queue_lock() is only called with a previously ref-counted key, so remove the additional ref-counting from the queue_(un)lock() functions. Also futex_wait_requeue_pi() drops one key reference too many in unqueue_me_pi(). Remove the key reference handling from unqueue_me_pi(). This was paired with a queue_lock() in futex_lock_pi(), so the count remains unchanged. Document remaining nested key ref-counting sites. Signed-off-by: Darren Hart Reported-and-tested-by: Matthieu Fertré Reported-by: Louis Rilling Cc: Peter Zijlstra Cc: Eric Dumazet Cc: John Kacur Cc: Rusty Russell LKML-Reference: <4CBB17A8.70401@linux.intel.com> Signed-off-by: Thomas Gleixner Cc: stable@kernel.org diff --git a/kernel/futex.c b/kernel/futex.c index 6a3a5fa..e328f57 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -1363,7 +1363,6 @@ static inline struct futex_hash_bucket *queue_lock(struct futex_q *q) { struct futex_hash_bucket *hb; - get_futex_key_refs(&q->key); hb = hash_futex(&q->key); q->lock_ptr = &hb->lock; @@ -1375,7 +1374,6 @@ static inline void queue_unlock(struct futex_q *q, struct futex_hash_bucket *hb) { spin_unlock(&hb->lock); - drop_futex_key_refs(&q->key); } /** @@ -1480,8 +1478,6 @@ static void unqueue_me_pi(struct futex_q *q) q->pi_state = NULL; spin_unlock(q->lock_ptr); - - drop_futex_key_refs(&q->key); } /* @@ -1812,7 +1808,10 @@ static int futex_wait(u32 __user *uaddr, int fshared, } retry: - /* Prepare to wait on uaddr. */ + /* + * Prepare to wait on uaddr. On success, holds hb lock and increments + * q.key refs. + */ ret = futex_wait_setup(uaddr, val, fshared, &q, &hb); if (ret) goto out; @@ -1822,24 +1821,23 @@ retry: /* If we were woken (and unqueued), we succeeded, whatever. */ ret = 0; + /* unqueue_me() drops q.key ref */ if (!unqueue_me(&q)) - goto out_put_key; + goto out; ret = -ETIMEDOUT; if (to && !to->task) - goto out_put_key; + goto out; /* * We expect signal_pending(current), but we might be the * victim of a spurious wakeup as well. */ - if (!signal_pending(current)) { - put_futex_key(fshared, &q.key); + if (!signal_pending(current)) goto retry; - } ret = -ERESTARTSYS; if (!abs_time) - goto out_put_key; + goto out; restart = ¤t_thread_info()->restart_block; restart->fn = futex_wait_restart; @@ -1856,8 +1854,6 @@ retry: ret = -ERESTART_RESTARTBLOCK; -out_put_key: - put_futex_key(fshared, &q.key); out: if (to) { hrtimer_cancel(&to->timer); @@ -2236,7 +2232,10 @@ static int futex_wait_requeue_pi(u32 __user *uaddr, int fshared, q.rt_waiter = &rt_waiter; q.requeue_pi_key = &key2; - /* Prepare to wait on uaddr. */ + /* + * Prepare to wait on uaddr. On success, increments q.key (key1) ref + * count. + */ ret = futex_wait_setup(uaddr, val, fshared, &q, &hb); if (ret) goto out_key2; @@ -2254,7 +2253,9 @@ static int futex_wait_requeue_pi(u32 __user *uaddr, int fshared, * In order for us to be here, we know our q.key == key2, and since * we took the hb->lock above, we also know that futex_requeue() has * completed and we no longer have to concern ourselves with a wakeup - * race with the atomic proxy lock acquition by the requeue code. + * race with the atomic proxy lock acquisition by the requeue code. The + * futex_requeue dropped our key1 reference and incremented our key2 + * reference count. */ /* Check if the requeue code acquired the second futex for us. */ -- cgit v0.10.2 From 3c4ea5b4985ecf831e22034d5d5db6e6391f3911 Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Mon, 20 Sep 2010 13:15:07 -0700 Subject: [IA64] Move local_softirq_pending() definition Ugly #include dependencies. We need to have local_softirq_pending() defined before it gets used in . But provides the definition *after* this #include chain: Signed-off-by: Tony Luck [ cherry-picked from the ia64 tree to fix "2bf2160: irq: Add tracepoint to softirq_raise" ] Signed-off-by: Ingo Molnar diff --git a/arch/ia64/include/asm/hardirq.h b/arch/ia64/include/asm/hardirq.h index d514cd9..8fb7d33 100644 --- a/arch/ia64/include/asm/hardirq.h +++ b/arch/ia64/include/asm/hardirq.h @@ -6,12 +6,6 @@ * David Mosberger-Tang */ - -#include -#include - -#include - /* * No irq_cpustat_t for IA-64. The data is held in the per-CPU data structure. */ @@ -20,6 +14,11 @@ #define local_softirq_pending() (local_cpu_data->softirq_pending) +#include +#include + +#include + extern void __iomem *ipi_base_addr; void ack_bad_irq(unsigned int irq); -- cgit v0.10.2 From 3f7edb1656e5beba2b5e617d31e4064e1eed0bc0 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 12 Oct 2010 11:03:44 +0200 Subject: [S390] hardirq: remove pointless header file includes Remove a couple of pointless header file includes. Fixes a compile bug caused by header file include dependencies with "irq: Add tracepoint to softirq_raise" within linux-next. Reported-by: Sachin Sant Signed-off-by: Heiko Carstens Signed-off-by: Martin Schwidefsky [ cherry-picked from the s390 tree to fix "2bf2160: irq: Add tracepoint to softirq_raise" ] Signed-off-by: Ingo Molnar diff --git a/arch/s390/include/asm/hardirq.h b/arch/s390/include/asm/hardirq.h index 498bc38..881d945 100644 --- a/arch/s390/include/asm/hardirq.h +++ b/arch/s390/include/asm/hardirq.h @@ -12,10 +12,6 @@ #ifndef __ASM_HARDIRQ_H #define __ASM_HARDIRQ_H -#include -#include -#include -#include #include #define local_softirq_pending() (S390_lowcore.softirq_pending) -- cgit v0.10.2 From 7e40798f406fe73f9bac496a390daabd8768a8f7 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 19 Oct 2010 10:56:19 -0400 Subject: tracing: Fix compile issue for trace_sched_wakeup.c The function start_func_tracer() was incorrectly added in the #ifdef CONFIG_FUNCTION_TRACER condition, but is still used even when function tracing is not enabled. The calls to register_ftrace_function() and register_ftrace_graph() become nops (and their arguments are even ignored), thus there is no reason to hide start_func_tracer() when function tracing is not enabled. Reported-by: Ingo Molnar Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 31689d2..7319559 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -130,6 +130,7 @@ static struct ftrace_ops trace_ops __read_mostly = { .func = wakeup_tracer_call, }; +#endif /* CONFIG_FUNCTION_TRACER */ static int start_func_tracer(int graph) { @@ -159,8 +160,6 @@ static void stop_func_tracer(int graph) unregister_ftrace_graph(); } -#endif /* CONFIG_FUNCTION_TRACER */ - #ifdef CONFIG_FUNCTION_GRAPH_TRACER static int wakeup_set_flag(u32 old_flags, u32 bit, int set) { -- cgit v0.10.2 From 9581d442b9058d3699b4be568b6e5eae38a41493 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Tue, 19 Oct 2010 16:46:55 +0200 Subject: KVM: Fix fs/gs reload oops with invalid ldt kvm reloads the host's fs and gs blindly, however the underlying segment descriptors may be invalid due to the user modifying the ldt after loading them. Fix by using the safe accessors (loadsegment() and load_gs_index()) instead of home grown unsafe versions. This is CVE-2010-3698. KVM-Stable-Tag. Signed-off-by: Avi Kivity Signed-off-by: Marcelo Tosatti diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 502e53f..c52e2eb 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -652,20 +652,6 @@ static inline struct kvm_mmu_page *page_header(hpa_t shadow_page) return (struct kvm_mmu_page *)page_private(page); } -static inline u16 kvm_read_fs(void) -{ - u16 seg; - asm("mov %%fs, %0" : "=g"(seg)); - return seg; -} - -static inline u16 kvm_read_gs(void) -{ - u16 seg; - asm("mov %%gs, %0" : "=g"(seg)); - return seg; -} - static inline u16 kvm_read_ldt(void) { u16 ldt; @@ -673,16 +659,6 @@ static inline u16 kvm_read_ldt(void) return ldt; } -static inline void kvm_load_fs(u16 sel) -{ - asm("mov %0, %%fs" : : "rm"(sel)); -} - -static inline void kvm_load_gs(u16 sel) -{ - asm("mov %0, %%gs" : : "rm"(sel)); -} - static inline void kvm_load_ldt(u16 sel) { asm("lldt %0" : : "rm"(sel)); diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 81ed28c..8a3f9f6 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -3163,8 +3163,8 @@ static void svm_vcpu_run(struct kvm_vcpu *vcpu) sync_lapic_to_cr8(vcpu); save_host_msrs(vcpu); - fs_selector = kvm_read_fs(); - gs_selector = kvm_read_gs(); + savesegment(fs, fs_selector); + savesegment(gs, gs_selector); ldt_selector = kvm_read_ldt(); svm->vmcb->save.cr2 = vcpu->arch.cr2; /* required for live migration with NPT */ @@ -3251,10 +3251,15 @@ static void svm_vcpu_run(struct kvm_vcpu *vcpu) vcpu->arch.regs[VCPU_REGS_RSP] = svm->vmcb->save.rsp; vcpu->arch.regs[VCPU_REGS_RIP] = svm->vmcb->save.rip; - kvm_load_fs(fs_selector); - kvm_load_gs(gs_selector); - kvm_load_ldt(ldt_selector); load_host_msrs(vcpu); + loadsegment(fs, fs_selector); +#ifdef CONFIG_X86_64 + load_gs_index(gs_selector); + wrmsrl(MSR_KERNEL_GS_BASE, current->thread.gs); +#else + loadsegment(gs, gs_selector); +#endif + kvm_load_ldt(ldt_selector); reload_tss(vcpu); diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 49b25ee..7bddfab 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -803,7 +803,7 @@ static void vmx_save_host_state(struct kvm_vcpu *vcpu) */ vmx->host_state.ldt_sel = kvm_read_ldt(); vmx->host_state.gs_ldt_reload_needed = vmx->host_state.ldt_sel; - vmx->host_state.fs_sel = kvm_read_fs(); + savesegment(fs, vmx->host_state.fs_sel); if (!(vmx->host_state.fs_sel & 7)) { vmcs_write16(HOST_FS_SELECTOR, vmx->host_state.fs_sel); vmx->host_state.fs_reload_needed = 0; @@ -811,7 +811,7 @@ static void vmx_save_host_state(struct kvm_vcpu *vcpu) vmcs_write16(HOST_FS_SELECTOR, 0); vmx->host_state.fs_reload_needed = 1; } - vmx->host_state.gs_sel = kvm_read_gs(); + savesegment(gs, vmx->host_state.gs_sel); if (!(vmx->host_state.gs_sel & 7)) vmcs_write16(HOST_GS_SELECTOR, vmx->host_state.gs_sel); else { @@ -841,27 +841,21 @@ static void vmx_save_host_state(struct kvm_vcpu *vcpu) static void __vmx_load_host_state(struct vcpu_vmx *vmx) { - unsigned long flags; - if (!vmx->host_state.loaded) return; ++vmx->vcpu.stat.host_state_reload; vmx->host_state.loaded = 0; if (vmx->host_state.fs_reload_needed) - kvm_load_fs(vmx->host_state.fs_sel); + loadsegment(fs, vmx->host_state.fs_sel); if (vmx->host_state.gs_ldt_reload_needed) { kvm_load_ldt(vmx->host_state.ldt_sel); - /* - * If we have to reload gs, we must take care to - * preserve our gs base. - */ - local_irq_save(flags); - kvm_load_gs(vmx->host_state.gs_sel); #ifdef CONFIG_X86_64 - wrmsrl(MSR_GS_BASE, vmcs_readl(HOST_GS_BASE)); + load_gs_index(vmx->host_state.gs_sel); + wrmsrl(MSR_KERNEL_GS_BASE, current->thread.gs); +#else + loadsegment(gs, vmx->host_state.gs_sel); #endif - local_irq_restore(flags); } reload_tss(); #ifdef CONFIG_X86_64 @@ -2589,8 +2583,8 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx) vmcs_write16(HOST_CS_SELECTOR, __KERNEL_CS); /* 22.2.4 */ vmcs_write16(HOST_DS_SELECTOR, __KERNEL_DS); /* 22.2.4 */ vmcs_write16(HOST_ES_SELECTOR, __KERNEL_DS); /* 22.2.4 */ - vmcs_write16(HOST_FS_SELECTOR, kvm_read_fs()); /* 22.2.4 */ - vmcs_write16(HOST_GS_SELECTOR, kvm_read_gs()); /* 22.2.4 */ + vmcs_write16(HOST_FS_SELECTOR, 0); /* 22.2.4 */ + vmcs_write16(HOST_GS_SELECTOR, 0); /* 22.2.4 */ vmcs_write16(HOST_SS_SELECTOR, __KERNEL_DS); /* 22.2.4 */ #ifdef CONFIG_X86_64 rdmsrl(MSR_FS_BASE, a); -- cgit v0.10.2 From ad4b2b627c8ac479e842bf586108c7a7fd63264b Mon Sep 17 00:00:00 2001 From: David Daney Date: Mon, 18 Oct 2010 17:51:26 -0700 Subject: MIPS: Repair Kbuild make clean breakage. When running make clean, Kbuild doesn't process the .config file, so nothing generates a platform-y variable. We can get it to descend into the platform directories by setting $(obj-). The dec Platform file was unconditionally setting platform-, obliterating its previous contents and preventing some directories from being cleaned. This is change to an append operation '+=' to allow cavium-octeon to be cleaned. Signed-off-by: David Daney Cc: Sam Ravnborg Patchwork: https://patchwork.linux-mips.org/patch/1718/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/Kbuild b/arch/mips/Kbuild index e322d65..7dd65cf 100644 --- a/arch/mips/Kbuild +++ b/arch/mips/Kbuild @@ -7,6 +7,10 @@ subdir-ccflags-y := -Werror include arch/mips/Kbuild.platforms obj-y := $(platform-y) +# make clean traverses $(obj-) without having included .config, so +# everything ends up here +obj- := $(platform-) + # mips object files # The object files are linked as core-y files would be linked diff --git a/arch/mips/dec/Platform b/arch/mips/dec/Platform index 3adbcbd..cf55a6f 100644 --- a/arch/mips/dec/Platform +++ b/arch/mips/dec/Platform @@ -1,7 +1,7 @@ # # DECstation family # -platform-$(CONFIG_MACH_DECSTATION) = dec/ +platform-$(CONFIG_MACH_DECSTATION) += dec/ cflags-$(CONFIG_MACH_DECSTATION) += \ -I$(srctree)/arch/mips/include/asm/mach-dec libs-$(CONFIG_MACH_DECSTATION) += arch/mips/dec/prom/ -- cgit v0.10.2 From 08be7b2bc7393d5cf4beb00aecb5eb9e5cde5f85 Mon Sep 17 00:00:00 2001 From: David Daney Date: Mon, 18 Oct 2010 18:05:49 -0700 Subject: MIPS: jz4740: Fix Kbuild Platform file. The platform specific files should be included via the platform-y variable. Signed-off-by: David Daney Cc: Lars-Peter Clausen Patchwork: https://patchwork.linux-mips.org/patch/1719/ Signed-off-by: Ralf Baechle diff --git a/arch/mips/jz4740/Platform b/arch/mips/jz4740/Platform index 6a97230..ba91be9 100644 --- a/arch/mips/jz4740/Platform +++ b/arch/mips/jz4740/Platform @@ -1,3 +1,3 @@ -core-$(CONFIG_MACH_JZ4740) += arch/mips/jz4740/ +platform-$(CONFIG_MACH_JZ4740) += jz4740/ cflags-$(CONFIG_MACH_JZ4740) += -I$(srctree)/arch/mips/include/asm/mach-jz4740 load-$(CONFIG_MACH_JZ4740) += 0xffffffff80010000 -- cgit v0.10.2 From 6cc0cc4a35337b48e09c82c5f3fb83534884fca8 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Fri, 15 Oct 2010 15:00:09 +0200 Subject: SERIAL: ioc3_serial: Return -ENOMEM on memory allocation failure In this code, 0 is returned on memory allocation failure, even though other failures return -ENOMEM or other similar values. A simplified version of the semantic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // @@ expression ret; expression x,e1,e2,e3; @@ ret = 0 ... when != ret = e1 *x = \(kmalloc\|kcalloc\|kzalloc\)(...) ... when != ret = e2 if (x == NULL) { ... when != ret = e3 return ret; } // Signed-off-by: Julia Lawall To: Pat Gefre Cc: kernel-janitors@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org Patchwork: https://patchwork.linux-mips.org/patch/1704/ Signed-off-by: Ralf Baechle diff --git a/drivers/serial/ioc3_serial.c b/drivers/serial/ioc3_serial.c index 93de907..800c546 100644 --- a/drivers/serial/ioc3_serial.c +++ b/drivers/serial/ioc3_serial.c @@ -2044,6 +2044,7 @@ ioc3uart_probe(struct ioc3_submodule *is, struct ioc3_driver_data *idd) if (!port) { printk(KERN_WARNING "IOC3 serial memory not available for port\n"); + ret = -ENOMEM; goto out4; } spin_lock_init(&port->ip_lock); -- cgit v0.10.2 From d39e072166493cbae3840ecfcf2b34bea0d2bebe Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Tue, 19 Oct 2010 18:32:41 +0100 Subject: MAINTAINERS: Change list for ioc_serial to linux-serial. IOC3 is also being used on SGI MIPS systems but this particular driver is only being used on IA64 systems so linux-mips made no sense as a list. Pat also thinks linux-serial@vger.kernel.org is the better list. Signed-off-by: Ralf Baechle diff --git a/MAINTAINERS b/MAINTAINERS index f46d8e6..1117403 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3140,7 +3140,7 @@ F: drivers/net/ioc3-eth.c IOC3 SERIAL DRIVER M: Pat Gefre -L: linux-mips@linux-mips.org +L: linux-serial@vger.kernel.org S: Maintained F: drivers/serial/ioc3_serial.c -- cgit v0.10.2 From e2cc502c3f71d8de812893527dc435e61b7fe604 Mon Sep 17 00:00:00 2001 From: Michel Thebeau Date: Tue, 28 Sep 2010 14:15:37 -0400 Subject: MIPS: O32 compat/N32: Fix to use compat syscall wrappers for AIO syscalls. [Ralf: Michel's original patch only fixed N32; I replicated the same fix for O32.] Signed-off-by: Michel Thebeau Cc: paul.gortmaker@windriver.com Cc: bruce.ashfield@windriver.com Signed-off-by: Ralf Baechle diff --git a/arch/mips/kernel/scall64-n32.S b/arch/mips/kernel/scall64-n32.S index a3d6613..79c27d1 100644 --- a/arch/mips/kernel/scall64-n32.S +++ b/arch/mips/kernel/scall64-n32.S @@ -320,10 +320,10 @@ EXPORT(sysn32_call_table) PTR sys_cacheflush PTR sys_cachectl PTR sys_sysmips - PTR sys_io_setup /* 6200 */ + PTR compat_sys_io_setup /* 6200 */ PTR sys_io_destroy - PTR sys_io_getevents - PTR sys_io_submit + PTR compat_sys_io_getevents + PTR compat_sys_io_submit PTR sys_io_cancel PTR sys_exit_group /* 6205 */ PTR sys_lookup_dcookie diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall64-o32.S index 813689e..c2aa30a 100644 --- a/arch/mips/kernel/scall64-o32.S +++ b/arch/mips/kernel/scall64-o32.S @@ -444,10 +444,10 @@ sys_call_table: PTR compat_sys_futex PTR compat_sys_sched_setaffinity PTR compat_sys_sched_getaffinity /* 4240 */ - PTR sys_io_setup + PTR compat_sys_io_setup PTR sys_io_destroy - PTR sys_io_getevents - PTR sys_io_submit + PTR compat_sys_io_getevents + PTR compat_sys_io_submit PTR sys_io_cancel /* 4245 */ PTR sys_exit_group PTR sys32_lookup_dcookie -- cgit v0.10.2 From 44235dcde416104b8e1db7606c283f4c0149c760 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Thu, 14 Oct 2010 17:04:59 -0700 Subject: x86, mm: Fix bogus whitespace in sync_global_pgds() Whitespace cleanup only. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: H. Peter Anvin diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 74f0f35..1ad7c0f 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -103,28 +103,28 @@ __setup("noexec32=", nonx32_setup); */ void sync_global_pgds(unsigned long start, unsigned long end) { - unsigned long address; - - for (address = start; address <= end; address += PGDIR_SIZE) { - const pgd_t *pgd_ref = pgd_offset_k(address); - unsigned long flags; - struct page *page; - - if (pgd_none(*pgd_ref)) - continue; - - spin_lock_irqsave(&pgd_lock, flags); - list_for_each_entry(page, &pgd_list, lru) { - pgd_t *pgd; - pgd = (pgd_t *)page_address(page) + pgd_index(address); - if (pgd_none(*pgd)) - set_pgd(pgd, *pgd_ref); - else - BUG_ON(pgd_page_vaddr(*pgd) - != pgd_page_vaddr(*pgd_ref)); - } - spin_unlock_irqrestore(&pgd_lock, flags); - } + unsigned long address; + + for (address = start; address <= end; address += PGDIR_SIZE) { + const pgd_t *pgd_ref = pgd_offset_k(address); + unsigned long flags; + struct page *page; + + if (pgd_none(*pgd_ref)) + continue; + + spin_lock_irqsave(&pgd_lock, flags); + list_for_each_entry(page, &pgd_list, lru) { + pgd_t *pgd; + pgd = (pgd_t *)page_address(page) + pgd_index(address); + if (pgd_none(*pgd)) + set_pgd(pgd, *pgd_ref); + else + BUG_ON(pgd_page_vaddr(*pgd) + != pgd_page_vaddr(*pgd_ref)); + } + spin_unlock_irqrestore(&pgd_lock, flags); + } } /* -- cgit v0.10.2 From 617d34d9e5d8326ec8f188c616aa06ac59d083fe Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 21 Sep 2010 12:01:51 -0700 Subject: x86, mm: Hold mm->page_table_lock while doing vmalloc_sync Take mm->page_table_lock while syncing the vmalloc region. This prevents a race with the Xen pagetable pin/unpin code, which expects that the page_table_lock is already held. If this race occurs, then Xen can see an inconsistent page type (a page can either be read/write or a pagetable page, and pin/unpin converts it between them), which will cause either the pin or the set_p[gm]d to fail; either will crash the kernel. vmalloc_sync_all() should be called rarely, so this extra use of page_table_lock should not interfere with its normal users. The mm pointer is stashed in the pgd page's index field, as that won't be otherwise used for pgds. Reported-by: Ian Campbell Originally-by: Jan Beulich LKML-Reference: <4CB88A4C.1080305@goop.org> Signed-off-by: Jeremy Fitzhardinge Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 2d0a33b..ada823a 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -28,6 +28,8 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)]; extern spinlock_t pgd_lock; extern struct list_head pgd_list; +extern struct mm_struct *pgd_page_get_mm(struct page *page); + #ifdef CONFIG_PARAVIRT #include #else /* !CONFIG_PARAVIRT */ diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index caec229..6c27c39 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -229,7 +229,16 @@ void vmalloc_sync_all(void) spin_lock_irqsave(&pgd_lock, flags); list_for_each_entry(page, &pgd_list, lru) { - if (!vmalloc_sync_one(page_address(page), address)) + spinlock_t *pgt_lock; + int ret; + + pgt_lock = &pgd_page_get_mm(page)->page_table_lock; + + spin_lock(pgt_lock); + ret = vmalloc_sync_one(page_address(page), address); + spin_unlock(pgt_lock); + + if (!ret) break; } spin_unlock_irqrestore(&pgd_lock, flags); diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 1ad7c0f..4d323fb 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -116,12 +116,19 @@ void sync_global_pgds(unsigned long start, unsigned long end) spin_lock_irqsave(&pgd_lock, flags); list_for_each_entry(page, &pgd_list, lru) { pgd_t *pgd; + spinlock_t *pgt_lock; + pgd = (pgd_t *)page_address(page) + pgd_index(address); + pgt_lock = &pgd_page_get_mm(page)->page_table_lock; + spin_lock(pgt_lock); + if (pgd_none(*pgd)) set_pgd(pgd, *pgd_ref); else BUG_ON(pgd_page_vaddr(*pgd) != pgd_page_vaddr(*pgd_ref)); + + spin_unlock(pgt_lock); } spin_unlock_irqrestore(&pgd_lock, flags); } diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c index 5c4ee42..c70e57d 100644 --- a/arch/x86/mm/pgtable.c +++ b/arch/x86/mm/pgtable.c @@ -87,7 +87,19 @@ static inline void pgd_list_del(pgd_t *pgd) #define UNSHARED_PTRS_PER_PGD \ (SHARED_KERNEL_PMD ? KERNEL_PGD_BOUNDARY : PTRS_PER_PGD) -static void pgd_ctor(pgd_t *pgd) + +static void pgd_set_mm(pgd_t *pgd, struct mm_struct *mm) +{ + BUILD_BUG_ON(sizeof(virt_to_page(pgd)->index) < sizeof(mm)); + virt_to_page(pgd)->index = (pgoff_t)mm; +} + +struct mm_struct *pgd_page_get_mm(struct page *page) +{ + return (struct mm_struct *)page->index; +} + +static void pgd_ctor(struct mm_struct *mm, pgd_t *pgd) { /* If the pgd points to a shared pagetable level (either the ptes in non-PAE, or shared PMD in PAE), then just copy the @@ -105,8 +117,10 @@ static void pgd_ctor(pgd_t *pgd) } /* list required to sync kernel mapping updates */ - if (!SHARED_KERNEL_PMD) + if (!SHARED_KERNEL_PMD) { + pgd_set_mm(pgd, mm); pgd_list_add(pgd); + } } static void pgd_dtor(pgd_t *pgd) @@ -272,7 +286,7 @@ pgd_t *pgd_alloc(struct mm_struct *mm) */ spin_lock_irqsave(&pgd_lock, flags); - pgd_ctor(pgd); + pgd_ctor(mm, pgd); pgd_prepopulate_pmd(mm, pgd, pmds); spin_unlock_irqrestore(&pgd_lock, flags); -- cgit v0.10.2 From a68c439b1966c91f0ef474e2bf275d6792312726 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 6 Oct 2010 12:27:53 +0200 Subject: apic, x86: Check if EILVT APIC registers are available (AMD only) This patch implements checks for the availability of LVT entries (APIC500-530) and reserves it if used. The check becomes necessary since we want to let the BIOS provide the LVT offsets. The offsets should be determined by the subsystems using it like those for MCE threshold or IBS. On K8 only offset 0 (APIC500) and MCE interrupts are supported. Beginning with family 10h at least 4 offsets are available. Since offsets must be consistent for all cores, we keep track of the LVT offsets in software and reserve the offset for the same vector also to be used on other cores. An offset is freed by setting the entry to APIC_EILVT_MASKED. If the BIOS is right, there should be no conflicts. Otherwise a "[Firmware Bug]: ..." error message is generated. However, if software does not properly determines the offsets, it is not necessarily a BIOS bug. Signed-off-by: Robert Richter LKML-Reference: <1286360874-1471-2-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/apicdef.h b/arch/x86/include/asm/apicdef.h index 7fe3b30..a859ca4 100644 --- a/arch/x86/include/asm/apicdef.h +++ b/arch/x86/include/asm/apicdef.h @@ -131,6 +131,7 @@ #define APIC_EILVTn(n) (0x500 + 0x10 * n) #define APIC_EILVT_NR_AMD_K8 1 /* # of extended interrupts */ #define APIC_EILVT_NR_AMD_10H 4 +#define APIC_EILVT_NR_MAX APIC_EILVT_NR_AMD_10H #define APIC_EILVT_LVTOFF(x) (((x) >> 4) & 0xF) #define APIC_EILVT_MSG_FIX 0x0 #define APIC_EILVT_MSG_SMI 0x2 diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 8cf86fb..2bfeafd 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -52,6 +52,7 @@ #include #include #include +#include unsigned int num_processors; @@ -370,24 +371,88 @@ static void __setup_APIC_LVTT(unsigned int clocks, int oneshot, int irqen) } /* - * Setup extended LVT, AMD specific (K8, family 10h) + * Setup extended LVT, AMD specific * - * Vector mappings are hard coded. On K8 only offset 0 (APIC500) and - * MCE interrupts are supported. Thus MCE offset must be set to 0. + * Software should use the LVT offsets the BIOS provides. The offsets + * are determined by the subsystems using it like those for MCE + * threshold or IBS. On K8 only offset 0 (APIC500) and MCE interrupts + * are supported. Beginning with family 10h at least 4 offsets are + * available. * - * If mask=1, the LVT entry does not generate interrupts while mask=0 - * enables the vector. See also the BKDGs. + * Since the offsets must be consistent for all cores, we keep track + * of the LVT offsets in software and reserve the offset for the same + * vector also to be used on other cores. An offset is freed by + * setting the entry to APIC_EILVT_MASKED. + * + * If the BIOS is right, there should be no conflicts. Otherwise a + * "[Firmware Bug]: ..." error message is generated. However, if + * software does not properly determines the offsets, it is not + * necessarily a BIOS bug. */ #define APIC_EILVT_LVTOFF_MCE 0 #define APIC_EILVT_LVTOFF_IBS 1 -static void setup_APIC_eilvt(u8 lvt_off, u8 vector, u8 msg_type, u8 mask) +static atomic_t eilvt_offsets[APIC_EILVT_NR_MAX]; + +static inline int eilvt_entry_is_changeable(unsigned int old, unsigned int new) +{ + return (old & APIC_EILVT_MASKED) + || (new == APIC_EILVT_MASKED) + || ((new & ~APIC_EILVT_MASKED) == old); +} + +static unsigned int reserve_eilvt_offset(int offset, unsigned int new) +{ + unsigned int rsvd; /* 0: uninitialized */ + + if (offset >= APIC_EILVT_NR_MAX) + return ~0; + + rsvd = atomic_read(&eilvt_offsets[offset]) & ~APIC_EILVT_MASKED; + do { + if (rsvd && + !eilvt_entry_is_changeable(rsvd, new)) + /* may not change if vectors are different */ + return rsvd; + rsvd = atomic_cmpxchg(&eilvt_offsets[offset], rsvd, new); + } while (rsvd != new); + + return new; +} + +/* + * If mask=1, the LVT entry does not generate interrupts while mask=0 + * enables the vector. See also the BKDGs. + */ + +static int setup_APIC_eilvt(u8 offset, u8 vector, u8 msg_type, u8 mask) { - unsigned long reg = (lvt_off << 4) + APIC_EILVTn(0); - unsigned int v = (mask << 16) | (msg_type << 8) | vector; + unsigned long reg = APIC_EILVTn(offset); + unsigned int new, old, reserved; + + new = (mask << 16) | (msg_type << 8) | vector; + old = apic_read(reg); + reserved = reserve_eilvt_offset(offset, new); - apic_write(reg, v); + if (reserved != new) { + pr_err(FW_BUG "cpu %d, try to setup vector 0x%x, but " + "vector 0x%x was already reserved by another core, " + "APIC%lX=0x%x\n", + smp_processor_id(), new, reserved, reg, old); + return -EINVAL; + } + + if (!eilvt_entry_is_changeable(old, new)) { + pr_err(FW_BUG "cpu %d, try to setup vector 0x%x but " + "register already in use, APIC%lX=0x%x\n", + smp_processor_id(), new, reg, old); + return -EBUSY; + } + + apic_write(reg, new); + + return 0; } u8 setup_APIC_eilvt_mce(u8 vector, u8 msg_type, u8 mask) -- cgit v0.10.2 From 27afdf2008da0b8878a73e32e4eb12381b84e224 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 6 Oct 2010 12:27:54 +0200 Subject: apic, x86: Use BIOS settings for IBS and MCE threshold interrupt LVT offsets We want the BIOS to setup the EILVT APIC registers. The offsets were hardcoded and BIOS settings were overwritten by the OS. Now, the subsystems for MCE threshold and IBS determine the LVT offset from the registers the BIOS has setup. If the BIOS setup is buggy on a family 10h system, a workaround enables IBS. If the OS determines an invalid register setup, a "[Firmware Bug]: " error message is reported. We need this change also for upcomming cpu families. Signed-off-by: Robert Richter LKML-Reference: <1286360874-1471-3-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index 1fa03e0..286de34 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -252,9 +252,7 @@ static inline int apic_is_clustered_box(void) } #endif -extern u8 setup_APIC_eilvt_mce(u8 vector, u8 msg_type, u8 mask); -extern u8 setup_APIC_eilvt_ibs(u8 vector, u8 msg_type, u8 mask); - +extern int setup_APIC_eilvt(u8 lvt_off, u8 vector, u8 msg_type, u8 mask); #else /* !CONFIG_X86_LOCAL_APIC */ static inline void lapic_shutdown(void) { } diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 2bfeafd..850657d 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -390,9 +390,6 @@ static void __setup_APIC_LVTT(unsigned int clocks, int oneshot, int irqen) * necessarily a BIOS bug. */ -#define APIC_EILVT_LVTOFF_MCE 0 -#define APIC_EILVT_LVTOFF_IBS 1 - static atomic_t eilvt_offsets[APIC_EILVT_NR_MAX]; static inline int eilvt_entry_is_changeable(unsigned int old, unsigned int new) @@ -426,7 +423,7 @@ static unsigned int reserve_eilvt_offset(int offset, unsigned int new) * enables the vector. See also the BKDGs. */ -static int setup_APIC_eilvt(u8 offset, u8 vector, u8 msg_type, u8 mask) +int setup_APIC_eilvt(u8 offset, u8 vector, u8 msg_type, u8 mask) { unsigned long reg = APIC_EILVTn(offset); unsigned int new, old, reserved; @@ -454,19 +451,7 @@ static int setup_APIC_eilvt(u8 offset, u8 vector, u8 msg_type, u8 mask) return 0; } - -u8 setup_APIC_eilvt_mce(u8 vector, u8 msg_type, u8 mask) -{ - setup_APIC_eilvt(APIC_EILVT_LVTOFF_MCE, vector, msg_type, mask); - return APIC_EILVT_LVTOFF_MCE; -} - -u8 setup_APIC_eilvt_ibs(u8 vector, u8 msg_type, u8 mask) -{ - setup_APIC_eilvt(APIC_EILVT_LVTOFF_IBS, vector, msg_type, mask); - return APIC_EILVT_LVTOFF_IBS; -} -EXPORT_SYMBOL_GPL(setup_APIC_eilvt_ibs); +EXPORT_SYMBOL_GPL(setup_APIC_eilvt); /* * Program the next event, relative to now diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c index 39aaee5..80c4823 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_amd.c +++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c @@ -131,7 +131,8 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c) u32 low = 0, high = 0, address = 0; unsigned int bank, block; struct thresh_restart tr; - u8 lvt_off; + int lvt_off = -1; + u8 offset; for (bank = 0; bank < NR_BANKS; ++bank) { for (block = 0; block < NR_BLOCKS; ++block) { @@ -162,8 +163,28 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c) if (shared_bank[bank] && c->cpu_core_id) break; #endif - lvt_off = setup_APIC_eilvt_mce(THRESHOLD_APIC_VECTOR, - APIC_EILVT_MSG_FIX, 0); + offset = (high & MASK_LVTOFF_HI) >> 20; + if (lvt_off < 0) { + if (setup_APIC_eilvt(offset, + THRESHOLD_APIC_VECTOR, + APIC_EILVT_MSG_FIX, 0)) { + pr_err(FW_BUG "cpu %d, failed to " + "setup threshold interrupt " + "for bank %d, block %d " + "(MSR%08X=0x%x%08x)", + smp_processor_id(), bank, block, + address, high, low); + continue; + } + lvt_off = offset; + } else if (lvt_off != offset) { + pr_err(FW_BUG "cpu %d, invalid threshold " + "interrupt offset %d for bank %d," + "block %d (MSR%08X=0x%x%08x)", + smp_processor_id(), lvt_off, bank, + block, address, high, low); + continue; + } high &= ~MASK_LVTOFF_HI; high |= lvt_off << 20; diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index b67a6b5..42fb46f 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -64,15 +64,22 @@ static u64 ibs_op_ctl; * IBS cpuid feature detection */ -#define IBS_CPUID_FEATURES 0x8000001b +#define IBS_CPUID_FEATURES 0x8000001b /* * Same bit mask as for IBS cpuid feature flags (Fn8000_001B_EAX), but * bit 0 is used to indicate the existence of IBS. */ -#define IBS_CAPS_AVAIL (1LL<<0) -#define IBS_CAPS_RDWROPCNT (1LL<<3) -#define IBS_CAPS_OPCNT (1LL<<4) +#define IBS_CAPS_AVAIL (1U<<0) +#define IBS_CAPS_RDWROPCNT (1U<<3) +#define IBS_CAPS_OPCNT (1U<<4) + +/* + * IBS APIC setup + */ +#define IBSCTL 0x1cc +#define IBSCTL_LVT_OFFSET_VALID (1ULL<<8) +#define IBSCTL_LVT_OFFSET_MASK 0x0F /* * IBS randomization macros @@ -266,6 +273,74 @@ static void op_amd_stop_ibs(void) wrmsrl(MSR_AMD64_IBSOPCTL, 0); } +static inline int eilvt_is_available(int offset) +{ + /* check if we may assign a vector */ + return !setup_APIC_eilvt(offset, 0, APIC_EILVT_MSG_NMI, 1); +} + +static inline int ibs_eilvt_valid(void) +{ + u64 val; + int offset; + + rdmsrl(MSR_AMD64_IBSCTL, val); + if (!(val & IBSCTL_LVT_OFFSET_VALID)) { + pr_err(FW_BUG "cpu %d, invalid IBS " + "interrupt offset %d (MSR%08X=0x%016llx)", + smp_processor_id(), offset, + MSR_AMD64_IBSCTL, val); + return 0; + } + + offset = val & IBSCTL_LVT_OFFSET_MASK; + + if (eilvt_is_available(offset)) + return !0; + + pr_err(FW_BUG "cpu %d, IBS interrupt offset %d " + "not available (MSR%08X=0x%016llx)", + smp_processor_id(), offset, + MSR_AMD64_IBSCTL, val); + + return 0; +} + +static inline int get_ibs_offset(void) +{ + u64 val; + + rdmsrl(MSR_AMD64_IBSCTL, val); + if (!(val & IBSCTL_LVT_OFFSET_VALID)) + return -EINVAL; + + return val & IBSCTL_LVT_OFFSET_MASK; +} + +static void setup_APIC_ibs(void) +{ + int offset; + + offset = get_ibs_offset(); + if (offset < 0) + goto failed; + + if (!setup_APIC_eilvt(offset, 0, APIC_EILVT_MSG_NMI, 0)) + return; +failed: + pr_warn("oprofile: IBS APIC setup failed on cpu #%d\n", + smp_processor_id()); +} + +static void clear_APIC_ibs(void) +{ + int offset; + + offset = get_ibs_offset(); + if (offset >= 0) + setup_APIC_eilvt(offset, 0, APIC_EILVT_MSG_FIX, 1); +} + #ifdef CONFIG_OPROFILE_EVENT_MULTIPLEX static void op_mux_switch_ctrl(struct op_x86_model_spec const *model, @@ -376,13 +451,13 @@ static void op_amd_setup_ctrs(struct op_x86_model_spec const *model, } if (ibs_caps) - setup_APIC_eilvt_ibs(0, APIC_EILVT_MSG_NMI, 0); + setup_APIC_ibs(); } static void op_amd_cpu_shutdown(void) { if (ibs_caps) - setup_APIC_eilvt_ibs(0, APIC_EILVT_MSG_FIX, 1); + clear_APIC_ibs(); } static int op_amd_check_ctrs(struct pt_regs * const regs, @@ -445,16 +520,11 @@ static void op_amd_stop(struct op_msrs const * const msrs) op_amd_stop_ibs(); } -static int __init_ibs_nmi(void) +static int setup_ibs_ctl(int ibs_eilvt_off) { -#define IBSCTL_LVTOFFSETVAL (1 << 8) -#define IBSCTL 0x1cc struct pci_dev *cpu_cfg; int nodes; u32 value = 0; - u8 ibs_eilvt_off; - - ibs_eilvt_off = setup_APIC_eilvt_ibs(0, APIC_EILVT_MSG_FIX, 1); nodes = 0; cpu_cfg = NULL; @@ -466,21 +536,60 @@ static int __init_ibs_nmi(void) break; ++nodes; pci_write_config_dword(cpu_cfg, IBSCTL, ibs_eilvt_off - | IBSCTL_LVTOFFSETVAL); + | IBSCTL_LVT_OFFSET_VALID); pci_read_config_dword(cpu_cfg, IBSCTL, &value); - if (value != (ibs_eilvt_off | IBSCTL_LVTOFFSETVAL)) { + if (value != (ibs_eilvt_off | IBSCTL_LVT_OFFSET_VALID)) { pci_dev_put(cpu_cfg); printk(KERN_DEBUG "Failed to setup IBS LVT offset, " - "IBSCTL = 0x%08x", value); - return 1; + "IBSCTL = 0x%08x\n", value); + return -EINVAL; } } while (1); if (!nodes) { - printk(KERN_DEBUG "No CPU node configured for IBS"); - return 1; + printk(KERN_DEBUG "No CPU node configured for IBS\n"); + return -ENODEV; + } + + return 0; +} + +static int force_ibs_eilvt_setup(void) +{ + int i; + int ret; + + /* find the next free available EILVT entry */ + for (i = 1; i < 4; i++) { + if (!eilvt_is_available(i)) + continue; + ret = setup_ibs_ctl(i); + if (ret) + return ret; + return 0; } + printk(KERN_DEBUG "No EILVT entry available\n"); + + return -EBUSY; +} + +static int __init_ibs_nmi(void) +{ + int ret; + + if (ibs_eilvt_valid()) + return 0; + + ret = force_ibs_eilvt_setup(); + if (ret) + return ret; + + if (!ibs_eilvt_valid()) + return -EFAULT; + + pr_err(FW_BUG "workaround enabled for IBS LVT offset\n"); + return 0; } -- cgit v0.10.2 From f01f7c56a1425b9749a99af821e1de334fb64d7e Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Tue, 19 Oct 2010 22:17:37 +0000 Subject: x86, mm: Fix incorrect data type in vmalloc_sync_all() arch/x86/mm/fault.c: In function 'vmalloc_sync_all': arch/x86/mm/fault.c:238: warning: assignment makes integer from pointer without a cast introduced by 617d34d9e5d8326ec8f188c616aa06ac59d083fe. Signed-off-by: Borislav Petkov LKML-Reference: <20101020103642.GA3135@kryptos.osrc.amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 6c27c39..0cdb8d4 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -230,7 +230,7 @@ void vmalloc_sync_all(void) spin_lock_irqsave(&pgd_lock, flags); list_for_each_entry(page, &pgd_list, lru) { spinlock_t *pgt_lock; - int ret; + pmd_t *ret; pgt_lock = &pgd_page_get_mm(page)->page_table_lock; -- cgit v0.10.2 From 531295e63be8c2b8b909d7400739e8b8df60d61f Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Wed, 20 Oct 2010 13:45:43 +1030 Subject: virtio: console: Don't block entire guest if host doesn't read data If the host is slow in reading data or doesn't read data at all, blocking write calls not only blocked the program that called write() but the entire guest itself. To overcome this, let's not block till the host signals it has given back the virtio ring element we passed it. Instead, send the buffer to the host and return to userspace. This operation then becomes similar to how non-blocking writes work, so let's use the existing code for this path as well. This code change also ensures blocking write calls do get blocked if there's not enough room in the virtio ring as well as they don't return -EAGAIN to userspace. Signed-off-by: Amit Shah Acked-by: Hans de Goede CC: stable@kernel.org Signed-off-by: Rusty Russell Signed-off-by: Linus Torvalds diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index c810481..0f69c5e 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -459,9 +459,12 @@ static ssize_t send_buf(struct port *port, void *in_buf, size_t in_count, /* * Wait till the host acknowledges it pushed out the data we - * sent. This is done for ports in blocking mode or for data - * from the hvc_console; the tty operations are performed with - * spinlocks held so we can't sleep here. + * sent. This is done for data from the hvc_console; the tty + * operations are performed with spinlocks held so we can't + * sleep here. An alternative would be to copy the data to a + * buffer and relax the spinning requirement. The downside is + * we need to kmalloc a GFP_ATOMIC buffer each time the + * console driver writes something out. */ while (!virtqueue_get_buf(out_vq, &len)) cpu_relax(); @@ -626,6 +629,14 @@ static ssize_t port_fops_write(struct file *filp, const char __user *ubuf, goto free_buf; } + /* + * We now ask send_buf() to not spin for generic ports -- we + * can re-use the same code path that non-blocking file + * descriptors take for blocking file descriptors since the + * wait is already done and we're certain the write will go + * through to the host. + */ + nonblock = true; ret = send_buf(port, buf, count, nonblock); if (nonblock && ret > 0) -- cgit v0.10.2 From f6f94e2ab1b33f0082ac22d71f66385a60d8157f Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 20 Oct 2010 13:30:22 -0700 Subject: Linux 2.6.36 diff --git a/Makefile b/Makefile index 7583116..860c26a 100644 --- a/Makefile +++ b/Makefile @@ -1,7 +1,7 @@ VERSION = 2 PATCHLEVEL = 6 SUBLEVEL = 36 -EXTRAVERSION = -rc8 +EXTRAVERSION = NAME = Flesh-Eating Bats with Fangs # *DOCUMENTATION* -- cgit v0.10.2 From c957ef2c59e952803766ddc22e89981ab534606f Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Wed, 20 Oct 2010 11:07:02 +0800 Subject: percpu: Introduce a read-mostly percpu API Add a new readmostly percpu section and API. This can be used to avoid dirtying data lines which are generally not written to, which is especially important for data which may be accessed by processors other than the one for which the percpu area belongs to. [ hpa: moved it *after* the page-aligned section, for obvious reasons. ] Signed-off-by: Shaohua Li LKML-Reference: <1287544022.4571.7.camel@sli10-conroe.sh.intel.com> Cc: Eric Dumazet Signed-off-by: H. Peter Anvin diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 8a92a17..d7e7b21 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -677,7 +677,9 @@ - LOAD_OFFSET) { \ VMLINUX_SYMBOL(__per_cpu_start) = .; \ *(.data..percpu..first) \ + . = ALIGN(PAGE_SIZE); \ *(.data..percpu..page_aligned) \ + *(.data..percpu..readmostly) \ *(.data..percpu) \ *(.data..percpu..shared_aligned) \ VMLINUX_SYMBOL(__per_cpu_end) = .; \ @@ -703,6 +705,8 @@ VMLINUX_SYMBOL(__per_cpu_load) = .; \ VMLINUX_SYMBOL(__per_cpu_start) = .; \ *(.data..percpu..first) \ + . = ALIGN(PAGE_SIZE); \ + *(.data..percpu..readmostly) \ *(.data..percpu..page_aligned) \ *(.data..percpu) \ *(.data..percpu..shared_aligned) \ diff --git a/include/linux/percpu-defs.h b/include/linux/percpu-defs.h index ce2dc65..27ef6b1 100644 --- a/include/linux/percpu-defs.h +++ b/include/linux/percpu-defs.h @@ -139,6 +139,15 @@ __aligned(PAGE_SIZE) /* + * Declaration/definition used for per-CPU variables that must be read mostly. + */ +#define DECLARE_PER_CPU_READ_MOSTLY(type, name) \ + DECLARE_PER_CPU_SECTION(type, name, "..readmostly") + +#define DEFINE_PER_CPU_READ_MOSTLY(type, name) \ + DEFINE_PER_CPU_SECTION(type, name, "..readmostly") + +/* * Intermodule exports for per-CPU variables. sparse forgets about * address space across EXPORT_SYMBOL(), change EXPORT_SYMBOL() to * noop if __CHECKER__. -- cgit v0.10.2 From 932967202182743c01a2eee4bdfa2c42697bc586 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Wed, 20 Oct 2010 11:07:03 +0800 Subject: x86: Spread tlb flush vector between nodes Currently flush tlb vector allocation is based on below equation: sender = smp_processor_id() % 8 This isn't optimal, CPUs from different node can have the same vector, this causes a lot of lock contention. Instead, we can assign the same vectors to CPUs from the same node, while different node has different vectors. This has below advantages: a. if there is lock contention, the lock contention is between CPUs from one node. This should be much cheaper than the contention between nodes. b. completely avoid lock contention between nodes. This especially benefits kswapd, which is the biggest user of tlb flush, since kswapd sets its affinity to specific node. In my test, this could reduce > 20% CPU overhead in extreme case.The test machine has 4 nodes and each node has 16 CPUs. I then bind each node's kswapd to the first CPU of the node. I run a workload with 4 sequential mmap file read thread. The files are empty sparse file. This workload will trigger a lot of page reclaim and tlbflush. The kswapd bind is to easy trigger the extreme tlb flush lock contention because otherwise kswapd keeps migrating between CPUs of a node and I can't get stable result. Sure in real workload, we can't always see so big tlb flush lock contention, but it's possible. [ hpa: folded in fix from Eric Dumazet to use this_cpu_read() ] Signed-off-by: Shaohua Li LKML-Reference: <1287544023.4571.8.camel@sli10-conroe.sh.intel.com> Cc: Eric Dumazet Signed-off-by: H. Peter Anvin diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c index c03f14a..4935848 100644 --- a/arch/x86/mm/tlb.c +++ b/arch/x86/mm/tlb.c @@ -5,6 +5,7 @@ #include #include #include +#include #include #include @@ -52,6 +53,8 @@ union smp_flush_state { want false sharing in the per cpu data segment. */ static union smp_flush_state flush_state[NUM_INVALIDATE_TLB_VECTORS]; +static DEFINE_PER_CPU_READ_MOSTLY(int, tlb_vector_offset); + /* * We cannot call mmdrop() because we are in interrupt context, * instead update mm->cpu_vm_mask. @@ -173,7 +176,7 @@ static void flush_tlb_others_ipi(const struct cpumask *cpumask, union smp_flush_state *f; /* Caller has disabled preemption */ - sender = smp_processor_id() % NUM_INVALIDATE_TLB_VECTORS; + sender = this_cpu_read(tlb_vector_offset); f = &flush_state[sender]; /* @@ -218,6 +221,47 @@ void native_flush_tlb_others(const struct cpumask *cpumask, flush_tlb_others_ipi(cpumask, mm, va); } +static void __cpuinit calculate_tlb_offset(void) +{ + int cpu, node, nr_node_vecs; + /* + * we are changing tlb_vector_offset for each CPU in runtime, but this + * will not cause inconsistency, as the write is atomic under X86. we + * might see more lock contentions in a short time, but after all CPU's + * tlb_vector_offset are changed, everything should go normal + * + * Note: if NUM_INVALIDATE_TLB_VECTORS % nr_online_nodes !=0, we might + * waste some vectors. + **/ + if (nr_online_nodes > NUM_INVALIDATE_TLB_VECTORS) + nr_node_vecs = 1; + else + nr_node_vecs = NUM_INVALIDATE_TLB_VECTORS/nr_online_nodes; + + for_each_online_node(node) { + int node_offset = (node % NUM_INVALIDATE_TLB_VECTORS) * + nr_node_vecs; + int cpu_offset = 0; + for_each_cpu(cpu, cpumask_of_node(node)) { + per_cpu(tlb_vector_offset, cpu) = node_offset + + cpu_offset; + cpu_offset++; + cpu_offset = cpu_offset % nr_node_vecs; + } + } +} + +static int tlb_cpuhp_notify(struct notifier_block *n, + unsigned long action, void *hcpu) +{ + switch (action & 0xf) { + case CPU_ONLINE: + case CPU_DEAD: + calculate_tlb_offset(); + } + return NOTIFY_OK; +} + static int __cpuinit init_smp_flush(void) { int i; @@ -225,6 +269,8 @@ static int __cpuinit init_smp_flush(void) for (i = 0; i < ARRAY_SIZE(flush_state); i++) raw_spin_lock_init(&flush_state[i].tlbstate_lock); + calculate_tlb_offset(); + hotcpu_notifier(tlb_cpuhp_notify, 0); return 0; } core_initcall(init_smp_flush); -- cgit v0.10.2 From 7669a2c95e502a77f93f27e5449fc93a00d588b6 Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Mon, 17 May 2010 12:31:35 -0700 Subject: ceph: lookup pool in osdmap by name Implement a pool lookup by name. This will be used by rbd. Signed-off-by: Yehuda Sadeh Signed-off-by: Sage Weil diff --git a/fs/ceph/osdmap.c b/fs/ceph/osdmap.c index e31f118..3ccd117 100644 --- a/fs/ceph/osdmap.c +++ b/fs/ceph/osdmap.c @@ -417,6 +417,19 @@ static struct ceph_pg_pool_info *__lookup_pg_pool(struct rb_root *root, int id) return NULL; } +int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name) +{ + struct rb_node *rbp; + + for (rbp = rb_first(&map->pg_pools); rbp; rbp = rb_next(rbp)) { + struct ceph_pg_pool_info *pi = + rb_entry(rbp, struct ceph_pg_pool_info, node); + if (pi->name && strcmp(pi->name, name) == 0) + return pi->id; + } + return -ENOENT; +} + static void __remove_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *pi) { rb_erase(&pi->node, root); diff --git a/fs/ceph/osdmap.h b/fs/ceph/osdmap.h index 970b547..a592b21 100644 --- a/fs/ceph/osdmap.h +++ b/fs/ceph/osdmap.h @@ -125,4 +125,6 @@ extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid, extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg pgid); +extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name); + #endif -- cgit v0.10.2 From 3499e8a5d4dbb083324efd942e2c4fb7eb65f27c Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Tue, 6 Apr 2010 14:51:47 -0700 Subject: ceph: refactor osdc requests creation functions The osd requests creation are being decoupled from the vino parameter, allowing clients using the osd to use other arbitrary object names that are not necessarily vino based. Also, calc_raw_layout now takes a snap id. Signed-off-by: Yehuda Sadeh Signed-off-by: Sage Weil diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c index 3b5571b..2647daf 100644 --- a/fs/ceph/osd_client.c +++ b/fs/ceph/osd_client.c @@ -22,6 +22,35 @@ static int __kick_requests(struct ceph_osd_client *osdc, static void kick_requests(struct ceph_osd_client *osdc, struct ceph_osd *osd); +void ceph_calc_raw_layout(struct ceph_osd_client *osdc, + struct ceph_file_layout *layout, + u64 snapid, + u64 off, u64 len, u64 *bno, + struct ceph_osd_request *req) +{ + struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; + struct ceph_osd_op *op = (void *)(reqhead + 1); + u64 orig_len = len; + u64 objoff, objlen; /* extent in object */ + + reqhead->snapid = cpu_to_le64(snapid); + + /* object extent? */ + ceph_calc_file_object_mapping(layout, off, &len, bno, + &objoff, &objlen); + if (len < orig_len) + dout(" skipping last %llu, final file extent %llu~%llu\n", + orig_len - len, off, len); + + op->extent.offset = cpu_to_le64(objoff); + op->extent.length = cpu_to_le64(objlen); + req->r_num_pages = calc_pages_for(off, len); + + dout("calc_layout bno=%llx %llu~%llu (%d pages)\n", + *bno, objoff, objlen, req->r_num_pages); + +} + /* * Implement client access to distributed object storage cluster. * @@ -48,34 +77,17 @@ static void kick_requests(struct ceph_osd_client *osdc, struct ceph_osd *osd); * fill osd op in request message. */ static void calc_layout(struct ceph_osd_client *osdc, - struct ceph_vino vino, struct ceph_file_layout *layout, + struct ceph_vino vino, + struct ceph_file_layout *layout, u64 off, u64 *plen, struct ceph_osd_request *req) { - struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; - struct ceph_osd_op *op = (void *)(reqhead + 1); - u64 orig_len = *plen; - u64 objoff, objlen; /* extent in object */ u64 bno; - reqhead->snapid = cpu_to_le64(vino.snap); - - /* object extent? */ - ceph_calc_file_object_mapping(layout, off, plen, &bno, - &objoff, &objlen); - if (*plen < orig_len) - dout(" skipping last %llu, final file extent %llu~%llu\n", - orig_len - *plen, off, *plen); + ceph_calc_raw_layout(osdc, layout, vino.snap, off, *plen, &bno, req); sprintf(req->r_oid, "%llx.%08llx", vino.ino, bno); req->r_oid_len = strlen(req->r_oid); - - op->extent.offset = cpu_to_le64(objoff); - op->extent.length = cpu_to_le64(objlen); - req->r_num_pages = calc_pages_for(off, *plen); - - dout("calc_layout %s (%d) %llu~%llu (%d pages)\n", - req->r_oid, req->r_oid_len, objoff, objlen, req->r_num_pages); } /* @@ -108,43 +120,34 @@ void ceph_osdc_release_request(struct kref *kref) kfree(req); } -/* - * build new request AND message, calculate layout, and adjust file - * extent as needed. - * - * if the file was recently truncated, we include information about its - * old and new size so that the object can be updated appropriately. (we - * avoid synchronously deleting truncated objects because it's slow.) - * - * if @do_sync, include a 'startsync' command so that the osd will flush - * data quickly. - */ -struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, - struct ceph_file_layout *layout, - struct ceph_vino vino, - u64 off, u64 *plen, - int opcode, int flags, +struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, + int flags, struct ceph_snap_context *snapc, int do_sync, - u32 truncate_seq, - u64 truncate_size, - struct timespec *mtime, - bool use_mempool, int num_reply) + bool use_mempool, + gfp_t gfp_flags, + struct page **pages) { struct ceph_osd_request *req; struct ceph_msg *msg; - struct ceph_osd_request_head *head; - struct ceph_osd_op *op; - void *p; int num_op = 1 + do_sync; - size_t msg_size = sizeof(*head) + num_op*sizeof(*op); - int i; + size_t msg_size = sizeof(struct ceph_osd_request_head) + + num_op*sizeof(struct ceph_osd_op); + + if (use_mempool) { + req = mempool_alloc(osdc->req_mempool, gfp_flags); + memset(req, 0, sizeof(*req)); + } else { + req = kzalloc(sizeof(*req), gfp_flags); + } + if (!req) + return NULL; if (use_mempool) { - req = mempool_alloc(osdc->req_mempool, GFP_NOFS); + req = mempool_alloc(osdc->req_mempool, gfp_flags); memset(req, 0, sizeof(*req)); } else { - req = kzalloc(sizeof(*req), GFP_NOFS); + req = kzalloc(sizeof(*req), gfp_flags); } if (req == NULL) return NULL; @@ -164,7 +167,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0); else msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, - OSD_OPREPLY_FRONT_LEN, GFP_NOFS); + OSD_OPREPLY_FRONT_LEN, gfp_flags); if (!msg) { ceph_osdc_put_request(req); return NULL; @@ -178,18 +181,48 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, if (use_mempool) msg = ceph_msgpool_get(&osdc->msgpool_op, 0); else - msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, GFP_NOFS); + msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags); if (!msg) { ceph_osdc_put_request(req); return NULL; } msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP); memset(msg->front.iov_base, 0, msg->front.iov_len); + + req->r_request = msg; + req->r_pages = pages; + + return req; +} + +/* + * build new request AND message + * + */ +void ceph_osdc_build_request(struct ceph_osd_request *req, + u64 off, u64 *plen, + int opcode, + struct ceph_snap_context *snapc, + int do_sync, + u32 truncate_seq, + u64 truncate_size, + struct timespec *mtime, + const char *oid, + int oid_len) +{ + struct ceph_msg *msg = req->r_request; + struct ceph_osd_request_head *head; + struct ceph_osd_op *op; + void *p; + int num_op = 1 + do_sync; + size_t msg_size = sizeof(*head) + num_op*sizeof(*op); + int i; + int flags = req->r_flags; + head = msg->front.iov_base; op = (void *)(head + 1); p = (void *)(op + num_op); - req->r_request = msg; req->r_snapc = ceph_get_snap_context(snapc); head->client_inc = cpu_to_le32(1); /* always, for now. */ @@ -199,10 +232,6 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, head->num_ops = cpu_to_le16(num_op); op->op = cpu_to_le16(opcode); - /* calculate max write size */ - calc_layout(osdc, vino, layout, off, plen, req); - req->r_file_layout = *layout; /* keep a copy */ - if (flags & CEPH_OSD_FLAG_WRITE) { req->r_request->hdr.data_off = cpu_to_le16(off); req->r_request->hdr.data_len = cpu_to_le32(*plen); @@ -212,9 +241,9 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, op->extent.truncate_seq = cpu_to_le32(truncate_seq); /* fill in oid */ - head->object_len = cpu_to_le32(req->r_oid_len); - memcpy(p, req->r_oid, req->r_oid_len); - p += req->r_oid_len; + head->object_len = cpu_to_le32(oid_len); + memcpy(p, oid, oid_len); + p += oid_len; if (do_sync) { op++; @@ -233,6 +262,50 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, msg_size = p - msg->front.iov_base; msg->front.iov_len = msg_size; msg->hdr.front_len = cpu_to_le32(msg_size); + return; +} + +/* + * build new request AND message, calculate layout, and adjust file + * extent as needed. + * + * if the file was recently truncated, we include information about its + * old and new size so that the object can be updated appropriately. (we + * avoid synchronously deleting truncated objects because it's slow.) + * + * if @do_sync, include a 'startsync' command so that the osd will flush + * data quickly. + */ +struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, + struct ceph_file_layout *layout, + struct ceph_vino vino, + u64 off, u64 *plen, + int opcode, int flags, + struct ceph_snap_context *snapc, + int do_sync, + u32 truncate_seq, + u64 truncate_size, + struct timespec *mtime, + bool use_mempool, int num_reply) +{ + struct ceph_osd_request *req = + ceph_osdc_alloc_request(osdc, flags, + snapc, do_sync, + use_mempool, + GFP_NOFS, NULL); + if (IS_ERR(req)) + return req; + + /* calculate max write size */ + calc_layout(osdc, vino, layout, off, plen, req); + req->r_file_layout = *layout; /* keep a copy */ + + ceph_osdc_build_request(req, off, plen, opcode, + snapc, do_sync, + truncate_seq, truncate_size, + mtime, + req->r_oid, req->r_oid_len); + return req; } diff --git a/fs/ceph/osd_client.h b/fs/ceph/osd_client.h index ce77698..b687c2e 100644 --- a/fs/ceph/osd_client.h +++ b/fs/ceph/osd_client.h @@ -119,6 +119,31 @@ extern void ceph_osdc_handle_reply(struct ceph_osd_client *osdc, extern void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg); +extern void ceph_calc_raw_layout(struct ceph_osd_client *osdc, + struct ceph_file_layout *layout, + u64 snapid, + u64 off, u64 len, u64 *bno, + struct ceph_osd_request *req); + +extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, + int flags, + struct ceph_snap_context *snapc, + int do_sync, + bool use_mempool, + gfp_t gfp_flags, + struct page **pages); + +extern void ceph_osdc_build_request(struct ceph_osd_request *req, + u64 off, u64 *plen, + int opcode, + struct ceph_snap_context *snapc, + int do_sync, + u32 truncate_seq, + u64 truncate_size, + struct timespec *mtime, + const char *oid, + int oid_len); + extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *, struct ceph_file_layout *layout, struct ceph_vino vino, -- cgit v0.10.2 From 68b4476b0bc13fef18266b4140309a30e86739d2 Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Tue, 6 Apr 2010 15:01:27 -0700 Subject: ceph: messenger and osdc changes for rbd Allow the messenger to send/receive data in a bio. This is added so that we wouldn't need to copy the data into pages or some other buffer when doing IO for an rbd block device. We can now have trailing variable sized data for osd ops. Also osd ops encoding is more modular. Signed-off-by: Yehuda Sadeh Signed-off-by: Sage Weil diff --git a/fs/ceph/messenger.c b/fs/ceph/messenger.c index 2502d76..17a09b3 100644 --- a/fs/ceph/messenger.c +++ b/fs/ceph/messenger.c @@ -9,6 +9,8 @@ #include #include #include +#include +#include #include #include "super.h" @@ -529,8 +531,11 @@ static void prepare_write_message(struct ceph_connection *con) if (le32_to_cpu(m->hdr.data_len) > 0) { /* initialize page iterator */ con->out_msg_pos.page = 0; - con->out_msg_pos.page_pos = - le16_to_cpu(m->hdr.data_off) & ~PAGE_MASK; + if (m->pages) + con->out_msg_pos.page_pos = + le16_to_cpu(m->hdr.data_off) & ~PAGE_MASK; + else + con->out_msg_pos.page_pos = 0; con->out_msg_pos.data_pos = 0; con->out_msg_pos.did_page_crc = 0; con->out_more = 1; /* data + footer will follow */ @@ -712,6 +717,31 @@ out: return ret; /* done! */ } +#ifdef CONFIG_BLOCK +static void init_bio_iter(struct bio *bio, struct bio **iter, int *seg) +{ + if (!bio) { + *iter = NULL; + *seg = 0; + return; + } + *iter = bio; + *seg = bio->bi_idx; +} + +static void iter_bio_next(struct bio **bio_iter, int *seg) +{ + if (*bio_iter == NULL) + return; + + BUG_ON(*seg >= (*bio_iter)->bi_vcnt); + + (*seg)++; + if (*seg == (*bio_iter)->bi_vcnt) + init_bio_iter((*bio_iter)->bi_next, bio_iter, seg); +} +#endif + /* * Write as much message data payload as we can. If we finish, queue * up the footer. @@ -726,21 +756,46 @@ static int write_partial_msg_pages(struct ceph_connection *con) size_t len; int crc = con->msgr->nocrc; int ret; + int total_max_write; + int in_trail = 0; + size_t trail_len = (msg->trail ? msg->trail->length : 0); dout("write_partial_msg_pages %p msg %p page %d/%d offset %d\n", con, con->out_msg, con->out_msg_pos.page, con->out_msg->nr_pages, con->out_msg_pos.page_pos); - while (con->out_msg_pos.page < con->out_msg->nr_pages) { +#ifdef CONFIG_BLOCK + if (msg->bio && !msg->bio_iter) + init_bio_iter(msg->bio, &msg->bio_iter, &msg->bio_seg); +#endif + + while (data_len > con->out_msg_pos.data_pos) { struct page *page = NULL; void *kaddr = NULL; + int max_write = PAGE_SIZE; + int page_shift = 0; + + total_max_write = data_len - trail_len - + con->out_msg_pos.data_pos; /* * if we are calculating the data crc (the default), we need * to map the page. if our pages[] has been revoked, use the * zero page. */ - if (msg->pages) { + + /* have we reached the trail part of the data? */ + if (con->out_msg_pos.data_pos >= data_len - trail_len) { + in_trail = 1; + + total_max_write = data_len - con->out_msg_pos.data_pos; + + page = list_first_entry(&msg->trail->head, + struct page, lru); + if (crc) + kaddr = kmap(page); + max_write = PAGE_SIZE; + } else if (msg->pages) { page = msg->pages[con->out_msg_pos.page]; if (crc) kaddr = kmap(page); @@ -749,13 +804,25 @@ static int write_partial_msg_pages(struct ceph_connection *con) struct page, lru); if (crc) kaddr = kmap(page); +#ifdef CONFIG_BLOCK + } else if (msg->bio) { + struct bio_vec *bv; + + bv = bio_iovec_idx(msg->bio_iter, msg->bio_seg); + page = bv->bv_page; + page_shift = bv->bv_offset; + if (crc) + kaddr = kmap(page) + page_shift; + max_write = bv->bv_len; +#endif } else { page = con->msgr->zero_page; if (crc) kaddr = page_address(con->msgr->zero_page); } - len = min((int)(PAGE_SIZE - con->out_msg_pos.page_pos), - (int)(data_len - con->out_msg_pos.data_pos)); + len = min_t(int, max_write - con->out_msg_pos.page_pos, + total_max_write); + if (crc && !con->out_msg_pos.did_page_crc) { void *base = kaddr + con->out_msg_pos.page_pos; u32 tmpcrc = le32_to_cpu(con->out_msg->footer.data_crc); @@ -765,13 +832,14 @@ static int write_partial_msg_pages(struct ceph_connection *con) cpu_to_le32(crc32c(tmpcrc, base, len)); con->out_msg_pos.did_page_crc = 1; } - ret = kernel_sendpage(con->sock, page, - con->out_msg_pos.page_pos, len, + con->out_msg_pos.page_pos + page_shift, + len, MSG_DONTWAIT | MSG_NOSIGNAL | MSG_MORE); - if (crc && (msg->pages || msg->pagelist)) + if (crc && + (msg->pages || msg->pagelist || msg->bio || in_trail)) kunmap(page); if (ret <= 0) @@ -783,9 +851,16 @@ static int write_partial_msg_pages(struct ceph_connection *con) con->out_msg_pos.page_pos = 0; con->out_msg_pos.page++; con->out_msg_pos.did_page_crc = 0; - if (msg->pagelist) + if (in_trail) + list_move_tail(&page->lru, + &msg->trail->head); + else if (msg->pagelist) list_move_tail(&page->lru, &msg->pagelist->head); +#ifdef CONFIG_BLOCK + else if (msg->bio) + iter_bio_next(&msg->bio_iter, &msg->bio_seg); +#endif } } @@ -1305,8 +1380,7 @@ static int read_partial_message_section(struct ceph_connection *con, struct kvec *section, unsigned int sec_len, u32 *crc) { - int left; - int ret; + int ret, left; BUG_ON(!section); @@ -1329,13 +1403,83 @@ static int read_partial_message_section(struct ceph_connection *con, static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con, struct ceph_msg_header *hdr, int *skip); + + +static int read_partial_message_pages(struct ceph_connection *con, + struct page **pages, + unsigned data_len, int datacrc) +{ + void *p; + int ret; + int left; + + left = min((int)(data_len - con->in_msg_pos.data_pos), + (int)(PAGE_SIZE - con->in_msg_pos.page_pos)); + /* (page) data */ + BUG_ON(pages == NULL); + p = kmap(pages[con->in_msg_pos.page]); + ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos, + left); + if (ret > 0 && datacrc) + con->in_data_crc = + crc32c(con->in_data_crc, + p + con->in_msg_pos.page_pos, ret); + kunmap(pages[con->in_msg_pos.page]); + if (ret <= 0) + return ret; + con->in_msg_pos.data_pos += ret; + con->in_msg_pos.page_pos += ret; + if (con->in_msg_pos.page_pos == PAGE_SIZE) { + con->in_msg_pos.page_pos = 0; + con->in_msg_pos.page++; + } + + return ret; +} + +#ifdef CONFIG_BLOCK +static int read_partial_message_bio(struct ceph_connection *con, + struct bio **bio_iter, int *bio_seg, + unsigned data_len, int datacrc) +{ + struct bio_vec *bv = bio_iovec_idx(*bio_iter, *bio_seg); + void *p; + int ret, left; + + if (IS_ERR(bv)) + return PTR_ERR(bv); + + left = min((int)(data_len - con->in_msg_pos.data_pos), + (int)(bv->bv_len - con->in_msg_pos.page_pos)); + + p = kmap(bv->bv_page) + bv->bv_offset; + + ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos, + left); + if (ret > 0 && datacrc) + con->in_data_crc = + crc32c(con->in_data_crc, + p + con->in_msg_pos.page_pos, ret); + kunmap(bv->bv_page); + if (ret <= 0) + return ret; + con->in_msg_pos.data_pos += ret; + con->in_msg_pos.page_pos += ret; + if (con->in_msg_pos.page_pos == bv->bv_len) { + con->in_msg_pos.page_pos = 0; + iter_bio_next(bio_iter, bio_seg); + } + + return ret; +} +#endif + /* * read (part of) a message. */ static int read_partial_message(struct ceph_connection *con) { struct ceph_msg *m = con->in_msg; - void *p; int ret; int to, left; unsigned front_len, middle_len, data_len, data_off; @@ -1422,7 +1566,10 @@ static int read_partial_message(struct ceph_connection *con) m->middle->vec.iov_len = 0; con->in_msg_pos.page = 0; - con->in_msg_pos.page_pos = data_off & ~PAGE_MASK; + if (m->pages) + con->in_msg_pos.page_pos = data_off & ~PAGE_MASK; + else + con->in_msg_pos.page_pos = 0; con->in_msg_pos.data_pos = 0; } @@ -1440,27 +1587,29 @@ static int read_partial_message(struct ceph_connection *con) if (ret <= 0) return ret; } +#ifdef CONFIG_BLOCK + if (m->bio && !m->bio_iter) + init_bio_iter(m->bio, &m->bio_iter, &m->bio_seg); +#endif /* (page) data */ while (con->in_msg_pos.data_pos < data_len) { - left = min((int)(data_len - con->in_msg_pos.data_pos), - (int)(PAGE_SIZE - con->in_msg_pos.page_pos)); - BUG_ON(m->pages == NULL); - p = kmap(m->pages[con->in_msg_pos.page]); - ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos, - left); - if (ret > 0 && datacrc) - con->in_data_crc = - crc32c(con->in_data_crc, - p + con->in_msg_pos.page_pos, ret); - kunmap(m->pages[con->in_msg_pos.page]); - if (ret <= 0) - return ret; - con->in_msg_pos.data_pos += ret; - con->in_msg_pos.page_pos += ret; - if (con->in_msg_pos.page_pos == PAGE_SIZE) { - con->in_msg_pos.page_pos = 0; - con->in_msg_pos.page++; + if (m->pages) { + ret = read_partial_message_pages(con, m->pages, + data_len, datacrc); + if (ret <= 0) + return ret; +#ifdef CONFIG_BLOCK + } else if (m->bio) { + + ret = read_partial_message_bio(con, + &m->bio_iter, &m->bio_seg, + data_len, datacrc); + if (ret <= 0) + return ret; +#endif + } else { + BUG_ON(1); } } @@ -2136,6 +2285,10 @@ struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags) m->nr_pages = 0; m->pages = NULL; m->pagelist = NULL; + m->bio = NULL; + m->bio_iter = NULL; + m->bio_seg = 0; + m->trail = NULL; dout("ceph_msg_new %p front %d\n", m, front_len); return m; @@ -2250,6 +2403,8 @@ void ceph_msg_last_put(struct kref *kref) m->pagelist = NULL; } + m->trail = NULL; + if (m->pool) ceph_msgpool_put(m->pool, m); else diff --git a/fs/ceph/messenger.h b/fs/ceph/messenger.h index 76fbc95..5a79450 100644 --- a/fs/ceph/messenger.h +++ b/fs/ceph/messenger.h @@ -82,6 +82,10 @@ struct ceph_msg { struct ceph_pagelist *pagelist; /* instead of pages */ struct list_head list_head; struct kref kref; + struct bio *bio; /* instead of pages/pagelist */ + struct bio *bio_iter; /* bio iterator */ + int bio_seg; /* current bio segment */ + struct ceph_pagelist *trail; /* the trailing part of the data */ bool front_is_vmalloc; bool more_to_follow; bool needs_out_seq; diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c index 2647daf..c5d818e 100644 --- a/fs/ceph/osd_client.c +++ b/fs/ceph/osd_client.c @@ -6,12 +6,16 @@ #include #include #include +#ifdef CONFIG_BLOCK +#include +#endif #include "super.h" #include "osd_client.h" #include "messenger.h" #include "decode.h" #include "auth.h" +#include "pagelist.h" #define OSD_OP_FRONT_LEN 4096 #define OSD_OPREPLY_FRONT_LEN 512 @@ -22,29 +26,50 @@ static int __kick_requests(struct ceph_osd_client *osdc, static void kick_requests(struct ceph_osd_client *osdc, struct ceph_osd *osd); +static int op_needs_trail(int op) +{ + switch (op) { + case CEPH_OSD_OP_GETXATTR: + case CEPH_OSD_OP_SETXATTR: + case CEPH_OSD_OP_CMPXATTR: + case CEPH_OSD_OP_CALL: + return 1; + default: + return 0; + } +} + +static int op_has_extent(int op) +{ + return (op == CEPH_OSD_OP_READ || + op == CEPH_OSD_OP_WRITE); +} + void ceph_calc_raw_layout(struct ceph_osd_client *osdc, struct ceph_file_layout *layout, u64 snapid, - u64 off, u64 len, u64 *bno, - struct ceph_osd_request *req) + u64 off, u64 *plen, u64 *bno, + struct ceph_osd_request *req, + struct ceph_osd_req_op *op) { struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; - struct ceph_osd_op *op = (void *)(reqhead + 1); - u64 orig_len = len; + u64 orig_len = *plen; u64 objoff, objlen; /* extent in object */ reqhead->snapid = cpu_to_le64(snapid); /* object extent? */ - ceph_calc_file_object_mapping(layout, off, &len, bno, + ceph_calc_file_object_mapping(layout, off, plen, bno, &objoff, &objlen); - if (len < orig_len) + if (*plen < orig_len) dout(" skipping last %llu, final file extent %llu~%llu\n", - orig_len - len, off, len); + orig_len - *plen, off, *plen); - op->extent.offset = cpu_to_le64(objoff); - op->extent.length = cpu_to_le64(objlen); - req->r_num_pages = calc_pages_for(off, len); + if (op_has_extent(op->op)) { + op->extent.offset = objoff; + op->extent.length = objlen; + } + req->r_num_pages = calc_pages_for(off, *plen); dout("calc_layout bno=%llx %llu~%llu (%d pages)\n", *bno, objoff, objlen, req->r_num_pages); @@ -80,11 +105,13 @@ static void calc_layout(struct ceph_osd_client *osdc, struct ceph_vino vino, struct ceph_file_layout *layout, u64 off, u64 *plen, - struct ceph_osd_request *req) + struct ceph_osd_request *req, + struct ceph_osd_req_op *op) { u64 bno; - ceph_calc_raw_layout(osdc, layout, vino.snap, off, *plen, &bno, req); + ceph_calc_raw_layout(osdc, layout, vino.snap, off, + plen, &bno, req, op); sprintf(req->r_oid, "%llx.%08llx", vino.ino, bno); req->r_oid_len = strlen(req->r_oid); @@ -113,35 +140,64 @@ void ceph_osdc_release_request(struct kref *kref) if (req->r_own_pages) ceph_release_page_vector(req->r_pages, req->r_num_pages); +#ifdef CONFIG_BLOCK + if (req->r_bio) + bio_put(req->r_bio); +#endif ceph_put_snap_context(req->r_snapc); + if (req->r_trail) { + ceph_pagelist_release(req->r_trail); + kfree(req->r_trail); + } if (req->r_mempool) mempool_free(req, req->r_osdc->req_mempool); else kfree(req); } +static int op_needs_trail(int op) +{ + switch (op) { + case CEPH_OSD_OP_GETXATTR: + case CEPH_OSD_OP_SETXATTR: + case CEPH_OSD_OP_CMPXATTR: + return 1; + default: + return 0; + } +} + +static int get_num_ops(struct ceph_osd_req_op *ops, int *needs_trail) +{ + int i = 0; + + if (needs_trail) + *needs_trail = 0; + while (ops[i].op) { + if (needs_trail && op_needs_trail(ops[i].op)) + *needs_trail = 1; + i++; + } + + return i; +} + struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, int flags, struct ceph_snap_context *snapc, - int do_sync, + struct ceph_osd_req_op *ops, bool use_mempool, gfp_t gfp_flags, - struct page **pages) + struct page **pages, + struct bio *bio) { struct ceph_osd_request *req; struct ceph_msg *msg; - int num_op = 1 + do_sync; - size_t msg_size = sizeof(struct ceph_osd_request_head) + - num_op*sizeof(struct ceph_osd_op); + int needs_trail; + int num_op = get_num_ops(ops, &needs_trail); + size_t msg_size = sizeof(struct ceph_osd_request_head); - if (use_mempool) { - req = mempool_alloc(osdc->req_mempool, gfp_flags); - memset(req, 0, sizeof(*req)); - } else { - req = kzalloc(sizeof(*req), gfp_flags); - } - if (!req) - return NULL; + msg_size += num_op*sizeof(struct ceph_osd_op); if (use_mempool) { req = mempool_alloc(osdc->req_mempool, gfp_flags); @@ -154,6 +210,7 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, req->r_osdc = osdc; req->r_mempool = use_mempool; + kref_init(&req->r_kref); init_completion(&req->r_completion); init_completion(&req->r_safe_completion); @@ -174,6 +231,15 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, } req->r_reply = msg; + /* allocate space for the trailing data */ + if (needs_trail) { + req->r_trail = kmalloc(sizeof(struct ceph_pagelist), gfp_flags); + if (!req->r_trail) { + ceph_osdc_put_request(req); + return NULL; + } + ceph_pagelist_init(req->r_trail); + } /* create request message; allow space for oid */ msg_size += 40; if (snapc) @@ -186,38 +252,87 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, ceph_osdc_put_request(req); return NULL; } + msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP); memset(msg->front.iov_base, 0, msg->front.iov_len); req->r_request = msg; req->r_pages = pages; +#ifdef CONFIG_BLOCK + if (bio) { + req->r_bio = bio; + bio_get(req->r_bio); + } +#endif return req; } +static void osd_req_encode_op(struct ceph_osd_request *req, + struct ceph_osd_op *dst, + struct ceph_osd_req_op *src) +{ + dst->op = cpu_to_le16(src->op); + + switch (dst->op) { + case CEPH_OSD_OP_READ: + case CEPH_OSD_OP_WRITE: + dst->extent.offset = + cpu_to_le64(src->extent.offset); + dst->extent.length = + cpu_to_le64(src->extent.length); + dst->extent.truncate_size = + cpu_to_le64(src->extent.truncate_size); + dst->extent.truncate_seq = + cpu_to_le32(src->extent.truncate_seq); + break; + + case CEPH_OSD_OP_GETXATTR: + case CEPH_OSD_OP_SETXATTR: + case CEPH_OSD_OP_CMPXATTR: + BUG_ON(!req->r_trail); + + dst->xattr.name_len = cpu_to_le32(src->xattr.name_len); + dst->xattr.value_len = cpu_to_le32(src->xattr.value_len); + dst->xattr.cmp_op = src->xattr.cmp_op; + dst->xattr.cmp_mode = src->xattr.cmp_mode; + ceph_pagelist_append(req->r_trail, src->xattr.name, + src->xattr.name_len); + ceph_pagelist_append(req->r_trail, src->xattr.val, + src->xattr.value_len); + break; + case CEPH_OSD_OP_STARTSYNC: + break; + default: + pr_err("unrecognized osd opcode %d\n", dst->op); + WARN_ON(1); + break; + } + dst->payload_len = cpu_to_le32(src->payload_len); +} + /* * build new request AND message * */ void ceph_osdc_build_request(struct ceph_osd_request *req, - u64 off, u64 *plen, - int opcode, - struct ceph_snap_context *snapc, - int do_sync, - u32 truncate_seq, - u64 truncate_size, - struct timespec *mtime, - const char *oid, - int oid_len) + u64 off, u64 *plen, + struct ceph_osd_req_op *src_ops, + struct ceph_snap_context *snapc, + struct timespec *mtime, + const char *oid, + int oid_len) { struct ceph_msg *msg = req->r_request; struct ceph_osd_request_head *head; + struct ceph_osd_req_op *src_op; struct ceph_osd_op *op; void *p; - int num_op = 1 + do_sync; + int num_op = get_num_ops(src_ops, NULL); size_t msg_size = sizeof(*head) + num_op*sizeof(*op); - int i; int flags = req->r_flags; + u64 data_len = 0; + int i; head = msg->front.iov_base; op = (void *)(head + 1); @@ -230,25 +345,23 @@ void ceph_osdc_build_request(struct ceph_osd_request *req, if (flags & CEPH_OSD_FLAG_WRITE) ceph_encode_timespec(&head->mtime, mtime); head->num_ops = cpu_to_le16(num_op); - op->op = cpu_to_le16(opcode); - if (flags & CEPH_OSD_FLAG_WRITE) { - req->r_request->hdr.data_off = cpu_to_le16(off); - req->r_request->hdr.data_len = cpu_to_le32(*plen); - op->payload_len = cpu_to_le32(*plen); - } - op->extent.truncate_size = cpu_to_le64(truncate_size); - op->extent.truncate_seq = cpu_to_le32(truncate_seq); /* fill in oid */ head->object_len = cpu_to_le32(oid_len); memcpy(p, oid, oid_len); p += oid_len; - if (do_sync) { + src_op = src_ops; + while (src_op->op) { + osd_req_encode_op(req, op, src_op); + src_op++; op++; - op->op = cpu_to_le16(CEPH_OSD_OP_STARTSYNC); } + + if (req->r_trail) + data_len += req->r_trail->length; + if (snapc) { head->snap_seq = cpu_to_le64(snapc->seq); head->num_snaps = cpu_to_le32(snapc->num_snaps); @@ -258,6 +371,14 @@ void ceph_osdc_build_request(struct ceph_osd_request *req, } } + if (flags & CEPH_OSD_FLAG_WRITE) { + req->r_request->hdr.data_off = cpu_to_le16(off); + req->r_request->hdr.data_len = cpu_to_le32(*plen + data_len); + } else if (data_len) { + req->r_request->hdr.data_off = 0; + req->r_request->hdr.data_len = cpu_to_le32(data_len); + } + BUG_ON(p > msg->front.iov_base + msg->front.iov_len); msg_size = p - msg->front.iov_base; msg->front.iov_len = msg_size; @@ -288,21 +409,34 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, struct timespec *mtime, bool use_mempool, int num_reply) { - struct ceph_osd_request *req = - ceph_osdc_alloc_request(osdc, flags, - snapc, do_sync, + struct ceph_osd_req_op ops[3]; + struct ceph_osd_request *req; + + ops[0].op = opcode; + ops[0].extent.truncate_seq = truncate_seq; + ops[0].extent.truncate_size = truncate_size; + ops[0].payload_len = 0; + + if (do_sync) { + ops[1].op = CEPH_OSD_OP_STARTSYNC; + ops[1].payload_len = 0; + ops[2].op = 0; + } else + ops[1].op = 0; + + req = ceph_osdc_alloc_request(osdc, flags, + snapc, ops, use_mempool, - GFP_NOFS, NULL); + GFP_NOFS, NULL, NULL); if (IS_ERR(req)) return req; /* calculate max write size */ - calc_layout(osdc, vino, layout, off, plen, req); + calc_layout(osdc, vino, layout, off, plen, req, ops); req->r_file_layout = *layout; /* keep a copy */ - ceph_osdc_build_request(req, off, plen, opcode, - snapc, do_sync, - truncate_seq, truncate_size, + ceph_osdc_build_request(req, off, plen, ops, + snapc, mtime, req->r_oid, req->r_oid_len); @@ -1177,6 +1311,10 @@ int ceph_osdc_start_request(struct ceph_osd_client *osdc, req->r_request->pages = req->r_pages; req->r_request->nr_pages = req->r_num_pages; +#ifdef CONFIG_BLOCK + req->r_request->bio = req->r_bio; +#endif + req->r_request->trail = req->r_trail; register_request(osdc, req); @@ -1493,6 +1631,9 @@ static struct ceph_msg *get_reply(struct ceph_connection *con, } m->pages = req->r_pages; m->nr_pages = req->r_num_pages; +#ifdef CONFIG_BLOCK + m->bio = req->r_bio; +#endif } *skip = 0; req->r_con_filling_msg = ceph_con_get(con); diff --git a/fs/ceph/osd_client.h b/fs/ceph/osd_client.h index b687c2e..d583d1b 100644 --- a/fs/ceph/osd_client.h +++ b/fs/ceph/osd_client.h @@ -15,6 +15,7 @@ struct ceph_snap_context; struct ceph_osd_request; struct ceph_osd_client; struct ceph_authorizer; +struct ceph_pagelist; /* * completion callback for async writepages @@ -80,6 +81,11 @@ struct ceph_osd_request { struct page **r_pages; /* pages for data payload */ int r_pages_from_pool; int r_own_pages; /* if true, i own page list */ +#ifdef CONFIG_BLOCK + struct bio *r_bio; /* instead of pages */ +#endif + + struct ceph_pagelist *r_trail; /* trailing part of the data */ }; struct ceph_osd_client { @@ -110,6 +116,36 @@ struct ceph_osd_client { struct ceph_msgpool msgpool_op_reply; }; +struct ceph_osd_req_op { + u16 op; /* CEPH_OSD_OP_* */ + u32 flags; /* CEPH_OSD_FLAG_* */ + union { + struct { + u64 offset, length; + u64 truncate_size; + u32 truncate_seq; + } extent; + struct { + const char *name; + u32 name_len; + const char *val; + u32 value_len; + __u8 cmp_op; /* CEPH_OSD_CMPXATTR_OP_* */ + __u8 cmp_mode; /* CEPH_OSD_CMPXATTR_MODE_* */ + } xattr; + struct { + __u8 class_len; + __u8 method_len; + __u8 argc; + u32 indata_len; + } cls; + struct { + u64 cookie, count; + } pgls; + }; + u32 payload_len; +}; + extern int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client); extern void ceph_osdc_stop(struct ceph_osd_client *osdc); @@ -122,27 +158,26 @@ extern void ceph_osdc_handle_map(struct ceph_osd_client *osdc, extern void ceph_calc_raw_layout(struct ceph_osd_client *osdc, struct ceph_file_layout *layout, u64 snapid, - u64 off, u64 len, u64 *bno, - struct ceph_osd_request *req); + u64 off, u64 *plen, u64 *bno, + struct ceph_osd_request *req, + struct ceph_osd_req_op *op); extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, int flags, struct ceph_snap_context *snapc, - int do_sync, + struct ceph_osd_req_op *ops, bool use_mempool, gfp_t gfp_flags, - struct page **pages); + struct page **pages, + struct bio *bio); extern void ceph_osdc_build_request(struct ceph_osd_request *req, - u64 off, u64 *plen, - int opcode, - struct ceph_snap_context *snapc, - int do_sync, - u32 truncate_seq, - u64 truncate_size, - struct timespec *mtime, - const char *oid, - int oid_len); + u64 off, u64 *plen, + struct ceph_osd_req_op *src_ops, + struct ceph_snap_context *snapc, + struct timespec *mtime, + const char *oid, + int oid_len); extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *, struct ceph_file_layout *layout, diff --git a/fs/ceph/pagelist.c b/fs/ceph/pagelist.c index 46a368b..326e1c0 100644 --- a/fs/ceph/pagelist.c +++ b/fs/ceph/pagelist.c @@ -39,7 +39,7 @@ static int ceph_pagelist_addpage(struct ceph_pagelist *pl) return 0; } -int ceph_pagelist_append(struct ceph_pagelist *pl, void *buf, size_t len) +int ceph_pagelist_append(struct ceph_pagelist *pl, const void *buf, size_t len) { while (pl->room < len) { size_t bit = pl->room; diff --git a/fs/ceph/pagelist.h b/fs/ceph/pagelist.h index e8a4187..cc9327a 100644 --- a/fs/ceph/pagelist.h +++ b/fs/ceph/pagelist.h @@ -19,7 +19,7 @@ static inline void ceph_pagelist_init(struct ceph_pagelist *pl) } extern int ceph_pagelist_release(struct ceph_pagelist *pl); -extern int ceph_pagelist_append(struct ceph_pagelist *pl, void *d, size_t l); +extern int ceph_pagelist_append(struct ceph_pagelist *pl, const void *d, size_t l); static inline int ceph_pagelist_encode_64(struct ceph_pagelist *pl, u64 v) { -- cgit v0.10.2 From ae1533b62b3369e6ae32338f4a77d64d0e88f676 Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Tue, 18 May 2010 16:38:08 -0700 Subject: ceph-rbd: osdc support for osd call and rollback operations This will be used for rbd snapshots administration. Signed-off-by: Yehuda Sadeh diff --git a/fs/ceph/decode.h b/fs/ceph/decode.h index 3d25415..c5b6939 100644 --- a/fs/ceph/decode.h +++ b/fs/ceph/decode.h @@ -191,6 +191,11 @@ static inline void ceph_encode_string(void **p, void *end, ceph_encode_need(p, end, n, bad); \ ceph_encode_copy(p, pv, n); \ } while (0) +#define ceph_encode_string_safe(p, end, s, n, bad) \ + do { \ + ceph_encode_need(p, end, n, bad); \ + ceph_encode_string(p, end, s, n); \ + } while (0) #endif diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c index c5d818e..0edb43f 100644 --- a/fs/ceph/osd_client.c +++ b/fs/ceph/osd_client.c @@ -161,6 +161,7 @@ static int op_needs_trail(int op) case CEPH_OSD_OP_GETXATTR: case CEPH_OSD_OP_SETXATTR: case CEPH_OSD_OP_CMPXATTR: + case CEPH_OSD_OP_CALL: return 1; default: return 0; @@ -301,6 +302,23 @@ static void osd_req_encode_op(struct ceph_osd_request *req, ceph_pagelist_append(req->r_trail, src->xattr.val, src->xattr.value_len); break; + case CEPH_OSD_OP_CALL: + BUG_ON(!req->r_trail); + + dst->cls.class_len = src->cls.class_len; + dst->cls.method_len = src->cls.method_len; + dst->cls.indata_len = cpu_to_le32(src->cls.indata_len); + + ceph_pagelist_append(req->r_trail, src->cls.class_name, + src->cls.class_len); + ceph_pagelist_append(req->r_trail, src->cls.method_name, + src->cls.method_len); + ceph_pagelist_append(req->r_trail, src->cls.indata, + src->cls.indata_len); + break; + case CEPH_OSD_OP_ROLLBACK: + dst->snap.snapid = cpu_to_le64(src->snap.snapid); + break; case CEPH_OSD_OP_STARTSYNC: break; default: diff --git a/fs/ceph/osd_client.h b/fs/ceph/osd_client.h index d583d1b..65c9c56 100644 --- a/fs/ceph/osd_client.h +++ b/fs/ceph/osd_client.h @@ -134,14 +134,20 @@ struct ceph_osd_req_op { __u8 cmp_mode; /* CEPH_OSD_CMPXATTR_MODE_* */ } xattr; struct { + const char *class_name; __u8 class_len; + const char *method_name; __u8 method_len; __u8 argc; + const char *indata; u32 indata_len; } cls; struct { u64 cookie, count; } pgls; + struct { + u64 snapid; + } snap; }; u32 payload_len; }; -- cgit v0.10.2 From 3d14c5d2b6e15c21d8e5467dc62d33127c23a644 Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Tue, 6 Apr 2010 15:14:15 -0700 Subject: ceph: factor out libceph from Ceph file system This factors out protocol and low-level storage parts of ceph into a separate libceph module living in net/ceph and include/linux/ceph. This is mostly a matter of moving files around. However, a few key pieces of the interface change as well: - ceph_client becomes ceph_fs_client and ceph_client, where the latter captures the mon and osd clients, and the fs_client gets the mds client and file system specific pieces. - Mount option parsing and debugfs setup is correspondingly broken into two pieces. - The mon client gets a generic handler callback for otherwise unknown messages (mds map, in this case). - The basic supported/required feature bits can be expanded (and are by ceph_fs_client). No functional change, aside from some subtle error handling cases that got cleaned up in the refactoring process. Signed-off-by: Sage Weil diff --git a/MAINTAINERS b/MAINTAINERS index 7679bf3..48d0565 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1527,6 +1527,8 @@ T: git git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client.git S: Supported F: Documentation/filesystems/ceph.txt F: fs/ceph +F: net/ceph +F: include/linux/ceph CERTIFIED WIRELESS USB (WUSB) SUBSYSTEM: M: David Vrabel diff --git a/fs/ceph/Kconfig b/fs/ceph/Kconfig index 0fcd264..9eb134e 100644 --- a/fs/ceph/Kconfig +++ b/fs/ceph/Kconfig @@ -1,9 +1,11 @@ config CEPH_FS tristate "Ceph distributed file system (EXPERIMENTAL)" depends on INET && EXPERIMENTAL + select CEPH_LIB select LIBCRC32C select CRYPTO_AES select CRYPTO + default n help Choose Y or M here to include support for mounting the experimental Ceph distributed file system. Ceph is an extremely @@ -14,15 +16,3 @@ config CEPH_FS If unsure, say N. -config CEPH_FS_PRETTYDEBUG - bool "Include file:line in ceph debug output" - depends on CEPH_FS - default n - help - If you say Y here, debug output will include a filename and - line to aid debugging. This icnreases kernel size and slows - execution slightly when debug call sites are enabled (e.g., - via CONFIG_DYNAMIC_DEBUG). - - If unsure, say N. - diff --git a/fs/ceph/Makefile b/fs/ceph/Makefile index 278e117..9e6c4f2 100644 --- a/fs/ceph/Makefile +++ b/fs/ceph/Makefile @@ -8,15 +8,8 @@ obj-$(CONFIG_CEPH_FS) += ceph.o ceph-objs := super.o inode.o dir.o file.o locks.o addr.o ioctl.o \ export.o caps.o snap.o xattr.o \ - messenger.o msgpool.o buffer.o pagelist.o \ - mds_client.o mdsmap.o \ - mon_client.o \ - osd_client.o osdmap.o crush/crush.o crush/mapper.o crush/hash.o \ - debugfs.o \ - auth.o auth_none.o \ - crypto.o armor.o \ - auth_x.o \ - ceph_fs.o ceph_strings.o ceph_hash.o ceph_frag.o + mds_client.o mdsmap.o strings.o ceph_frag.o \ + debugfs.o else #Otherwise we were called directly from the command diff --git a/fs/ceph/README b/fs/ceph/README deleted file mode 100644 index 18352fa..0000000 --- a/fs/ceph/README +++ /dev/null @@ -1,20 +0,0 @@ -# -# The following files are shared by (and manually synchronized -# between) the Ceph userland and kernel client. -# -# userland kernel -src/include/ceph_fs.h fs/ceph/ceph_fs.h -src/include/ceph_fs.cc fs/ceph/ceph_fs.c -src/include/msgr.h fs/ceph/msgr.h -src/include/rados.h fs/ceph/rados.h -src/include/ceph_strings.cc fs/ceph/ceph_strings.c -src/include/ceph_frag.h fs/ceph/ceph_frag.h -src/include/ceph_frag.cc fs/ceph/ceph_frag.c -src/include/ceph_hash.h fs/ceph/ceph_hash.h -src/include/ceph_hash.cc fs/ceph/ceph_hash.c -src/crush/crush.c fs/ceph/crush/crush.c -src/crush/crush.h fs/ceph/crush/crush.h -src/crush/mapper.c fs/ceph/crush/mapper.c -src/crush/mapper.h fs/ceph/crush/mapper.h -src/crush/hash.h fs/ceph/crush/hash.h -src/crush/hash.c fs/ceph/crush/hash.c diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index efbc604..51bcc5c 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1,4 +1,4 @@ -#include "ceph_debug.h" +#include #include #include @@ -10,7 +10,8 @@ #include #include "super.h" -#include "osd_client.h" +#include "mds_client.h" +#include /* * Ceph address space ops. @@ -193,7 +194,8 @@ static int readpage_nounlock(struct file *filp, struct page *page) { struct inode *inode = filp->f_dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_osd_client *osdc = &ceph_inode_to_client(inode)->osdc; + struct ceph_osd_client *osdc = + &ceph_inode_to_client(inode)->client->osdc; int err = 0; u64 len = PAGE_CACHE_SIZE; @@ -265,7 +267,8 @@ static int ceph_readpages(struct file *file, struct address_space *mapping, { struct inode *inode = file->f_dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_osd_client *osdc = &ceph_inode_to_client(inode)->osdc; + struct ceph_osd_client *osdc = + &ceph_inode_to_client(inode)->client->osdc; int rc = 0; struct page **pages; loff_t offset; @@ -365,7 +368,7 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) { struct inode *inode; struct ceph_inode_info *ci; - struct ceph_client *client; + struct ceph_fs_client *fsc; struct ceph_osd_client *osdc; loff_t page_off = page->index << PAGE_CACHE_SHIFT; int len = PAGE_CACHE_SIZE; @@ -383,8 +386,8 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) } inode = page->mapping->host; ci = ceph_inode(inode); - client = ceph_inode_to_client(inode); - osdc = &client->osdc; + fsc = ceph_inode_to_client(inode); + osdc = &fsc->client->osdc; /* verify this is a writeable snap context */ snapc = (void *)page->private; @@ -414,10 +417,10 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) dout("writepage %p page %p index %lu on %llu~%u snapc %p\n", inode, page, page->index, page_off, len, snapc); - writeback_stat = atomic_long_inc_return(&client->writeback_count); + writeback_stat = atomic_long_inc_return(&fsc->writeback_count); if (writeback_stat > - CONGESTION_ON_THRESH(client->mount_args->congestion_kb)) - set_bdi_congested(&client->backing_dev_info, BLK_RW_ASYNC); + CONGESTION_ON_THRESH(fsc->mount_options->congestion_kb)) + set_bdi_congested(&fsc->backing_dev_info, BLK_RW_ASYNC); set_page_writeback(page); err = ceph_osdc_writepages(osdc, ceph_vino(inode), @@ -496,7 +499,7 @@ static void writepages_finish(struct ceph_osd_request *req, struct address_space *mapping = inode->i_mapping; __s32 rc = -EIO; u64 bytes = 0; - struct ceph_client *client = ceph_inode_to_client(inode); + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); long writeback_stat; unsigned issued = ceph_caps_issued(ci); @@ -529,10 +532,10 @@ static void writepages_finish(struct ceph_osd_request *req, WARN_ON(!PageUptodate(page)); writeback_stat = - atomic_long_dec_return(&client->writeback_count); + atomic_long_dec_return(&fsc->writeback_count); if (writeback_stat < - CONGESTION_OFF_THRESH(client->mount_args->congestion_kb)) - clear_bdi_congested(&client->backing_dev_info, + CONGESTION_OFF_THRESH(fsc->mount_options->congestion_kb)) + clear_bdi_congested(&fsc->backing_dev_info, BLK_RW_ASYNC); ceph_put_snap_context((void *)page->private); @@ -569,13 +572,13 @@ static void writepages_finish(struct ceph_osd_request *req, * mempool. we avoid the mempool if we can because req->r_num_pages * may be less than the maximum write size. */ -static void alloc_page_vec(struct ceph_client *client, +static void alloc_page_vec(struct ceph_fs_client *fsc, struct ceph_osd_request *req) { req->r_pages = kmalloc(sizeof(struct page *) * req->r_num_pages, GFP_NOFS); if (!req->r_pages) { - req->r_pages = mempool_alloc(client->wb_pagevec_pool, GFP_NOFS); + req->r_pages = mempool_alloc(fsc->wb_pagevec_pool, GFP_NOFS); req->r_pages_from_pool = 1; WARN_ON(!req->r_pages); } @@ -590,7 +593,7 @@ static int ceph_writepages_start(struct address_space *mapping, struct inode *inode = mapping->host; struct backing_dev_info *bdi = mapping->backing_dev_info; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_client *client; + struct ceph_fs_client *fsc; pgoff_t index, start, end; int range_whole = 0; int should_loop = 1; @@ -617,13 +620,13 @@ static int ceph_writepages_start(struct address_space *mapping, wbc->sync_mode == WB_SYNC_NONE ? "NONE" : (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD")); - client = ceph_inode_to_client(inode); - if (client->mount_state == CEPH_MOUNT_SHUTDOWN) { + fsc = ceph_inode_to_client(inode); + if (fsc->mount_state == CEPH_MOUNT_SHUTDOWN) { pr_warning("writepage_start %p on forced umount\n", inode); return -EIO; /* we're in a forced umount, don't write! */ } - if (client->mount_args->wsize && client->mount_args->wsize < wsize) - wsize = client->mount_args->wsize; + if (fsc->mount_options->wsize && fsc->mount_options->wsize < wsize) + wsize = fsc->mount_options->wsize; if (wsize < PAGE_CACHE_SIZE) wsize = PAGE_CACHE_SIZE; max_pages_ever = wsize >> PAGE_CACHE_SHIFT; @@ -769,7 +772,7 @@ get_more_pages: offset = (unsigned long long)page->index << PAGE_CACHE_SHIFT; len = wsize; - req = ceph_osdc_new_request(&client->osdc, + req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout, ceph_vino(inode), offset, &len, @@ -782,7 +785,7 @@ get_more_pages: &inode->i_mtime, true, 1); max_pages = req->r_num_pages; - alloc_page_vec(client, req); + alloc_page_vec(fsc, req); req->r_callback = writepages_finish; req->r_inode = inode; } @@ -794,10 +797,10 @@ get_more_pages: inode, page, page->index); writeback_stat = - atomic_long_inc_return(&client->writeback_count); + atomic_long_inc_return(&fsc->writeback_count); if (writeback_stat > CONGESTION_ON_THRESH( - client->mount_args->congestion_kb)) { - set_bdi_congested(&client->backing_dev_info, + fsc->mount_options->congestion_kb)) { + set_bdi_congested(&fsc->backing_dev_info, BLK_RW_ASYNC); } @@ -846,7 +849,7 @@ get_more_pages: op->payload_len = cpu_to_le32(len); req->r_request->hdr.data_len = cpu_to_le32(len); - ceph_osdc_start_request(&client->osdc, req, true); + ceph_osdc_start_request(&fsc->client->osdc, req, true); req = NULL; /* continue? */ @@ -915,7 +918,7 @@ static int ceph_update_writeable_page(struct file *file, { struct inode *inode = file->f_dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_mds_client *mdsc = &ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; loff_t page_off = pos & PAGE_CACHE_MASK; int pos_in_page = pos & ~PAGE_CACHE_MASK; int end_in_page = pos_in_page + len; @@ -1053,8 +1056,8 @@ static int ceph_write_end(struct file *file, struct address_space *mapping, struct page *page, void *fsdata) { struct inode *inode = file->f_dentry->d_inode; - struct ceph_client *client = ceph_inode_to_client(inode); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + struct ceph_mds_client *mdsc = fsc->mdsc; unsigned from = pos & (PAGE_CACHE_SIZE - 1); int check_cap = 0; @@ -1123,7 +1126,7 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) { struct inode *inode = vma->vm_file->f_dentry->d_inode; struct page *page = vmf->page; - struct ceph_mds_client *mdsc = &ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; loff_t off = page->index << PAGE_CACHE_SHIFT; loff_t size, len; int ret; diff --git a/fs/ceph/armor.c b/fs/ceph/armor.c deleted file mode 100644 index eb2a666..0000000 --- a/fs/ceph/armor.c +++ /dev/null @@ -1,103 +0,0 @@ - -#include - -int ceph_armor(char *dst, const char *src, const char *end); -int ceph_unarmor(char *dst, const char *src, const char *end); - -/* - * base64 encode/decode. - */ - -static const char *pem_key = - "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; - -static int encode_bits(int c) -{ - return pem_key[c]; -} - -static int decode_bits(char c) -{ - if (c >= 'A' && c <= 'Z') - return c - 'A'; - if (c >= 'a' && c <= 'z') - return c - 'a' + 26; - if (c >= '0' && c <= '9') - return c - '0' + 52; - if (c == '+') - return 62; - if (c == '/') - return 63; - if (c == '=') - return 0; /* just non-negative, please */ - return -EINVAL; -} - -int ceph_armor(char *dst, const char *src, const char *end) -{ - int olen = 0; - int line = 0; - - while (src < end) { - unsigned char a, b, c; - - a = *src++; - *dst++ = encode_bits(a >> 2); - if (src < end) { - b = *src++; - *dst++ = encode_bits(((a & 3) << 4) | (b >> 4)); - if (src < end) { - c = *src++; - *dst++ = encode_bits(((b & 15) << 2) | - (c >> 6)); - *dst++ = encode_bits(c & 63); - } else { - *dst++ = encode_bits((b & 15) << 2); - *dst++ = '='; - } - } else { - *dst++ = encode_bits(((a & 3) << 4)); - *dst++ = '='; - *dst++ = '='; - } - olen += 4; - line += 4; - if (line == 64) { - line = 0; - *(dst++) = '\n'; - olen++; - } - } - return olen; -} - -int ceph_unarmor(char *dst, const char *src, const char *end) -{ - int olen = 0; - - while (src < end) { - int a, b, c, d; - - if (src < end && src[0] == '\n') - src++; - if (src + 4 > end) - return -EINVAL; - a = decode_bits(src[0]); - b = decode_bits(src[1]); - c = decode_bits(src[2]); - d = decode_bits(src[3]); - if (a < 0 || b < 0 || c < 0 || d < 0) - return -EINVAL; - - *dst++ = (a << 2) | (b >> 4); - if (src[2] == '=') - return olen + 1; - *dst++ = ((b & 15) << 4) | (c >> 2); - if (src[3] == '=') - return olen + 2; - *dst++ = ((c & 3) << 6) | d; - olen += 3; - src += 4; - } - return olen; -} diff --git a/fs/ceph/auth.c b/fs/ceph/auth.c deleted file mode 100644 index 6d2e306..0000000 --- a/fs/ceph/auth.c +++ /dev/null @@ -1,259 +0,0 @@ -#include "ceph_debug.h" - -#include -#include -#include - -#include "types.h" -#include "auth_none.h" -#include "auth_x.h" -#include "decode.h" -#include "super.h" - -#include "messenger.h" - -/* - * get protocol handler - */ -static u32 supported_protocols[] = { - CEPH_AUTH_NONE, - CEPH_AUTH_CEPHX -}; - -static int ceph_auth_init_protocol(struct ceph_auth_client *ac, int protocol) -{ - switch (protocol) { - case CEPH_AUTH_NONE: - return ceph_auth_none_init(ac); - case CEPH_AUTH_CEPHX: - return ceph_x_init(ac); - default: - return -ENOENT; - } -} - -/* - * setup, teardown. - */ -struct ceph_auth_client *ceph_auth_init(const char *name, const char *secret) -{ - struct ceph_auth_client *ac; - int ret; - - dout("auth_init name '%s' secret '%s'\n", name, secret); - - ret = -ENOMEM; - ac = kzalloc(sizeof(*ac), GFP_NOFS); - if (!ac) - goto out; - - ac->negotiating = true; - if (name) - ac->name = name; - else - ac->name = CEPH_AUTH_NAME_DEFAULT; - dout("auth_init name %s secret %s\n", ac->name, secret); - ac->secret = secret; - return ac; - -out: - return ERR_PTR(ret); -} - -void ceph_auth_destroy(struct ceph_auth_client *ac) -{ - dout("auth_destroy %p\n", ac); - if (ac->ops) - ac->ops->destroy(ac); - kfree(ac); -} - -/* - * Reset occurs when reconnecting to the monitor. - */ -void ceph_auth_reset(struct ceph_auth_client *ac) -{ - dout("auth_reset %p\n", ac); - if (ac->ops && !ac->negotiating) - ac->ops->reset(ac); - ac->negotiating = true; -} - -int ceph_entity_name_encode(const char *name, void **p, void *end) -{ - int len = strlen(name); - - if (*p + 2*sizeof(u32) + len > end) - return -ERANGE; - ceph_encode_32(p, CEPH_ENTITY_TYPE_CLIENT); - ceph_encode_32(p, len); - ceph_encode_copy(p, name, len); - return 0; -} - -/* - * Initiate protocol negotiation with monitor. Include entity name - * and list supported protocols. - */ -int ceph_auth_build_hello(struct ceph_auth_client *ac, void *buf, size_t len) -{ - struct ceph_mon_request_header *monhdr = buf; - void *p = monhdr + 1, *end = buf + len, *lenp; - int i, num; - int ret; - - dout("auth_build_hello\n"); - monhdr->have_version = 0; - monhdr->session_mon = cpu_to_le16(-1); - monhdr->session_mon_tid = 0; - - ceph_encode_32(&p, 0); /* no protocol, yet */ - - lenp = p; - p += sizeof(u32); - - ceph_decode_need(&p, end, 1 + sizeof(u32), bad); - ceph_encode_8(&p, 1); - num = ARRAY_SIZE(supported_protocols); - ceph_encode_32(&p, num); - ceph_decode_need(&p, end, num * sizeof(u32), bad); - for (i = 0; i < num; i++) - ceph_encode_32(&p, supported_protocols[i]); - - ret = ceph_entity_name_encode(ac->name, &p, end); - if (ret < 0) - return ret; - ceph_decode_need(&p, end, sizeof(u64), bad); - ceph_encode_64(&p, ac->global_id); - - ceph_encode_32(&lenp, p - lenp - sizeof(u32)); - return p - buf; - -bad: - return -ERANGE; -} - -static int ceph_build_auth_request(struct ceph_auth_client *ac, - void *msg_buf, size_t msg_len) -{ - struct ceph_mon_request_header *monhdr = msg_buf; - void *p = monhdr + 1; - void *end = msg_buf + msg_len; - int ret; - - monhdr->have_version = 0; - monhdr->session_mon = cpu_to_le16(-1); - monhdr->session_mon_tid = 0; - - ceph_encode_32(&p, ac->protocol); - - ret = ac->ops->build_request(ac, p + sizeof(u32), end); - if (ret < 0) { - pr_err("error %d building auth method %s request\n", ret, - ac->ops->name); - return ret; - } - dout(" built request %d bytes\n", ret); - ceph_encode_32(&p, ret); - return p + ret - msg_buf; -} - -/* - * Handle auth message from monitor. - */ -int ceph_handle_auth_reply(struct ceph_auth_client *ac, - void *buf, size_t len, - void *reply_buf, size_t reply_len) -{ - void *p = buf; - void *end = buf + len; - int protocol; - s32 result; - u64 global_id; - void *payload, *payload_end; - int payload_len; - char *result_msg; - int result_msg_len; - int ret = -EINVAL; - - dout("handle_auth_reply %p %p\n", p, end); - ceph_decode_need(&p, end, sizeof(u32) * 3 + sizeof(u64), bad); - protocol = ceph_decode_32(&p); - result = ceph_decode_32(&p); - global_id = ceph_decode_64(&p); - payload_len = ceph_decode_32(&p); - payload = p; - p += payload_len; - ceph_decode_need(&p, end, sizeof(u32), bad); - result_msg_len = ceph_decode_32(&p); - result_msg = p; - p += result_msg_len; - if (p != end) - goto bad; - - dout(" result %d '%.*s' gid %llu len %d\n", result, result_msg_len, - result_msg, global_id, payload_len); - - payload_end = payload + payload_len; - - if (global_id && ac->global_id != global_id) { - dout(" set global_id %lld -> %lld\n", ac->global_id, global_id); - ac->global_id = global_id; - } - - if (ac->negotiating) { - /* server does not support our protocols? */ - if (!protocol && result < 0) { - ret = result; - goto out; - } - /* set up (new) protocol handler? */ - if (ac->protocol && ac->protocol != protocol) { - ac->ops->destroy(ac); - ac->protocol = 0; - ac->ops = NULL; - } - if (ac->protocol != protocol) { - ret = ceph_auth_init_protocol(ac, protocol); - if (ret) { - pr_err("error %d on auth protocol %d init\n", - ret, protocol); - goto out; - } - } - - ac->negotiating = false; - } - - ret = ac->ops->handle_reply(ac, result, payload, payload_end); - if (ret == -EAGAIN) { - return ceph_build_auth_request(ac, reply_buf, reply_len); - } else if (ret) { - pr_err("auth method '%s' error %d\n", ac->ops->name, ret); - return ret; - } - return 0; - -bad: - pr_err("failed to decode auth msg\n"); -out: - return ret; -} - -int ceph_build_auth(struct ceph_auth_client *ac, - void *msg_buf, size_t msg_len) -{ - if (!ac->protocol) - return ceph_auth_build_hello(ac, msg_buf, msg_len); - BUG_ON(!ac->ops); - if (ac->ops->should_authenticate(ac)) - return ceph_build_auth_request(ac, msg_buf, msg_len); - return 0; -} - -int ceph_auth_is_authenticated(struct ceph_auth_client *ac) -{ - if (!ac->ops) - return 0; - return ac->ops->is_authenticated(ac); -} diff --git a/fs/ceph/auth.h b/fs/ceph/auth.h deleted file mode 100644 index d38a2fb..0000000 --- a/fs/ceph/auth.h +++ /dev/null @@ -1,92 +0,0 @@ -#ifndef _FS_CEPH_AUTH_H -#define _FS_CEPH_AUTH_H - -#include "types.h" -#include "buffer.h" - -/* - * Abstract interface for communicating with the authenticate module. - * There is some handshake that takes place between us and the monitor - * to acquire the necessary keys. These are used to generate an - * 'authorizer' that we use when connecting to a service (mds, osd). - */ - -struct ceph_auth_client; -struct ceph_authorizer; - -struct ceph_auth_client_ops { - const char *name; - - /* - * true if we are authenticated and can connect to - * services. - */ - int (*is_authenticated)(struct ceph_auth_client *ac); - - /* - * true if we should (re)authenticate, e.g., when our tickets - * are getting old and crusty. - */ - int (*should_authenticate)(struct ceph_auth_client *ac); - - /* - * build requests and process replies during monitor - * handshake. if handle_reply returns -EAGAIN, we build - * another request. - */ - int (*build_request)(struct ceph_auth_client *ac, void *buf, void *end); - int (*handle_reply)(struct ceph_auth_client *ac, int result, - void *buf, void *end); - - /* - * Create authorizer for connecting to a service, and verify - * the response to authenticate the service. - */ - int (*create_authorizer)(struct ceph_auth_client *ac, int peer_type, - struct ceph_authorizer **a, - void **buf, size_t *len, - void **reply_buf, size_t *reply_len); - int (*verify_authorizer_reply)(struct ceph_auth_client *ac, - struct ceph_authorizer *a, size_t len); - void (*destroy_authorizer)(struct ceph_auth_client *ac, - struct ceph_authorizer *a); - void (*invalidate_authorizer)(struct ceph_auth_client *ac, - int peer_type); - - /* reset when we (re)connect to a monitor */ - void (*reset)(struct ceph_auth_client *ac); - - void (*destroy)(struct ceph_auth_client *ac); -}; - -struct ceph_auth_client { - u32 protocol; /* CEPH_AUTH_* */ - void *private; /* for use by protocol implementation */ - const struct ceph_auth_client_ops *ops; /* null iff protocol==0 */ - - bool negotiating; /* true if negotiating protocol */ - const char *name; /* entity name */ - u64 global_id; /* our unique id in system */ - const char *secret; /* our secret key */ - unsigned want_keys; /* which services we want */ -}; - -extern struct ceph_auth_client *ceph_auth_init(const char *name, - const char *secret); -extern void ceph_auth_destroy(struct ceph_auth_client *ac); - -extern void ceph_auth_reset(struct ceph_auth_client *ac); - -extern int ceph_auth_build_hello(struct ceph_auth_client *ac, - void *buf, size_t len); -extern int ceph_handle_auth_reply(struct ceph_auth_client *ac, - void *buf, size_t len, - void *reply_buf, size_t reply_len); -extern int ceph_entity_name_encode(const char *name, void **p, void *end); - -extern int ceph_build_auth(struct ceph_auth_client *ac, - void *msg_buf, size_t msg_len); - -extern int ceph_auth_is_authenticated(struct ceph_auth_client *ac); - -#endif diff --git a/fs/ceph/auth_none.c b/fs/ceph/auth_none.c deleted file mode 100644 index ad1dc21..0000000 --- a/fs/ceph/auth_none.c +++ /dev/null @@ -1,131 +0,0 @@ - -#include "ceph_debug.h" - -#include -#include -#include -#include - -#include "auth_none.h" -#include "auth.h" -#include "decode.h" - -static void reset(struct ceph_auth_client *ac) -{ - struct ceph_auth_none_info *xi = ac->private; - - xi->starting = true; - xi->built_authorizer = false; -} - -static void destroy(struct ceph_auth_client *ac) -{ - kfree(ac->private); - ac->private = NULL; -} - -static int is_authenticated(struct ceph_auth_client *ac) -{ - struct ceph_auth_none_info *xi = ac->private; - - return !xi->starting; -} - -static int should_authenticate(struct ceph_auth_client *ac) -{ - struct ceph_auth_none_info *xi = ac->private; - - return xi->starting; -} - -/* - * the generic auth code decode the global_id, and we carry no actual - * authenticate state, so nothing happens here. - */ -static int handle_reply(struct ceph_auth_client *ac, int result, - void *buf, void *end) -{ - struct ceph_auth_none_info *xi = ac->private; - - xi->starting = false; - return result; -} - -/* - * build an 'authorizer' with our entity_name and global_id. we can - * reuse a single static copy since it is identical for all services - * we connect to. - */ -static int ceph_auth_none_create_authorizer( - struct ceph_auth_client *ac, int peer_type, - struct ceph_authorizer **a, - void **buf, size_t *len, - void **reply_buf, size_t *reply_len) -{ - struct ceph_auth_none_info *ai = ac->private; - struct ceph_none_authorizer *au = &ai->au; - void *p, *end; - int ret; - - if (!ai->built_authorizer) { - p = au->buf; - end = p + sizeof(au->buf); - ceph_encode_8(&p, 1); - ret = ceph_entity_name_encode(ac->name, &p, end - 8); - if (ret < 0) - goto bad; - ceph_decode_need(&p, end, sizeof(u64), bad2); - ceph_encode_64(&p, ac->global_id); - au->buf_len = p - (void *)au->buf; - ai->built_authorizer = true; - dout("built authorizer len %d\n", au->buf_len); - } - - *a = (struct ceph_authorizer *)au; - *buf = au->buf; - *len = au->buf_len; - *reply_buf = au->reply_buf; - *reply_len = sizeof(au->reply_buf); - return 0; - -bad2: - ret = -ERANGE; -bad: - return ret; -} - -static void ceph_auth_none_destroy_authorizer(struct ceph_auth_client *ac, - struct ceph_authorizer *a) -{ - /* nothing to do */ -} - -static const struct ceph_auth_client_ops ceph_auth_none_ops = { - .name = "none", - .reset = reset, - .destroy = destroy, - .is_authenticated = is_authenticated, - .should_authenticate = should_authenticate, - .handle_reply = handle_reply, - .create_authorizer = ceph_auth_none_create_authorizer, - .destroy_authorizer = ceph_auth_none_destroy_authorizer, -}; - -int ceph_auth_none_init(struct ceph_auth_client *ac) -{ - struct ceph_auth_none_info *xi; - - dout("ceph_auth_none_init %p\n", ac); - xi = kzalloc(sizeof(*xi), GFP_NOFS); - if (!xi) - return -ENOMEM; - - xi->starting = true; - xi->built_authorizer = false; - - ac->protocol = CEPH_AUTH_NONE; - ac->private = xi; - ac->ops = &ceph_auth_none_ops; - return 0; -} - diff --git a/fs/ceph/auth_none.h b/fs/ceph/auth_none.h deleted file mode 100644 index 8164df1..0000000 --- a/fs/ceph/auth_none.h +++ /dev/null @@ -1,30 +0,0 @@ -#ifndef _FS_CEPH_AUTH_NONE_H -#define _FS_CEPH_AUTH_NONE_H - -#include - -#include "auth.h" - -/* - * null security mode. - * - * we use a single static authorizer that simply encodes our entity name - * and global id. - */ - -struct ceph_none_authorizer { - char buf[128]; - int buf_len; - char reply_buf[0]; -}; - -struct ceph_auth_none_info { - bool starting; - bool built_authorizer; - struct ceph_none_authorizer au; /* we only need one; it's static */ -}; - -extern int ceph_auth_none_init(struct ceph_auth_client *ac); - -#endif - diff --git a/fs/ceph/auth_x.c b/fs/ceph/auth_x.c deleted file mode 100644 index a2d002c..0000000 --- a/fs/ceph/auth_x.c +++ /dev/null @@ -1,687 +0,0 @@ - -#include "ceph_debug.h" - -#include -#include -#include -#include - -#include "auth_x.h" -#include "auth_x_protocol.h" -#include "crypto.h" -#include "auth.h" -#include "decode.h" - -#define TEMP_TICKET_BUF_LEN 256 - -static void ceph_x_validate_tickets(struct ceph_auth_client *ac, int *pneed); - -static int ceph_x_is_authenticated(struct ceph_auth_client *ac) -{ - struct ceph_x_info *xi = ac->private; - int need; - - ceph_x_validate_tickets(ac, &need); - dout("ceph_x_is_authenticated want=%d need=%d have=%d\n", - ac->want_keys, need, xi->have_keys); - return (ac->want_keys & xi->have_keys) == ac->want_keys; -} - -static int ceph_x_should_authenticate(struct ceph_auth_client *ac) -{ - struct ceph_x_info *xi = ac->private; - int need; - - ceph_x_validate_tickets(ac, &need); - dout("ceph_x_should_authenticate want=%d need=%d have=%d\n", - ac->want_keys, need, xi->have_keys); - return need != 0; -} - -static int ceph_x_encrypt_buflen(int ilen) -{ - return sizeof(struct ceph_x_encrypt_header) + ilen + 16 + - sizeof(u32); -} - -static int ceph_x_encrypt(struct ceph_crypto_key *secret, - void *ibuf, int ilen, void *obuf, size_t olen) -{ - struct ceph_x_encrypt_header head = { - .struct_v = 1, - .magic = cpu_to_le64(CEPHX_ENC_MAGIC) - }; - size_t len = olen - sizeof(u32); - int ret; - - ret = ceph_encrypt2(secret, obuf + sizeof(u32), &len, - &head, sizeof(head), ibuf, ilen); - if (ret) - return ret; - ceph_encode_32(&obuf, len); - return len + sizeof(u32); -} - -static int ceph_x_decrypt(struct ceph_crypto_key *secret, - void **p, void *end, void *obuf, size_t olen) -{ - struct ceph_x_encrypt_header head; - size_t head_len = sizeof(head); - int len, ret; - - len = ceph_decode_32(p); - if (*p + len > end) - return -EINVAL; - - dout("ceph_x_decrypt len %d\n", len); - ret = ceph_decrypt2(secret, &head, &head_len, obuf, &olen, - *p, len); - if (ret) - return ret; - if (head.struct_v != 1 || le64_to_cpu(head.magic) != CEPHX_ENC_MAGIC) - return -EPERM; - *p += len; - return olen; -} - -/* - * get existing (or insert new) ticket handler - */ -static struct ceph_x_ticket_handler * -get_ticket_handler(struct ceph_auth_client *ac, int service) -{ - struct ceph_x_ticket_handler *th; - struct ceph_x_info *xi = ac->private; - struct rb_node *parent = NULL, **p = &xi->ticket_handlers.rb_node; - - while (*p) { - parent = *p; - th = rb_entry(parent, struct ceph_x_ticket_handler, node); - if (service < th->service) - p = &(*p)->rb_left; - else if (service > th->service) - p = &(*p)->rb_right; - else - return th; - } - - /* add it */ - th = kzalloc(sizeof(*th), GFP_NOFS); - if (!th) - return ERR_PTR(-ENOMEM); - th->service = service; - rb_link_node(&th->node, parent, p); - rb_insert_color(&th->node, &xi->ticket_handlers); - return th; -} - -static void remove_ticket_handler(struct ceph_auth_client *ac, - struct ceph_x_ticket_handler *th) -{ - struct ceph_x_info *xi = ac->private; - - dout("remove_ticket_handler %p %d\n", th, th->service); - rb_erase(&th->node, &xi->ticket_handlers); - ceph_crypto_key_destroy(&th->session_key); - if (th->ticket_blob) - ceph_buffer_put(th->ticket_blob); - kfree(th); -} - -static int ceph_x_proc_ticket_reply(struct ceph_auth_client *ac, - struct ceph_crypto_key *secret, - void *buf, void *end) -{ - struct ceph_x_info *xi = ac->private; - int num; - void *p = buf; - int ret; - char *dbuf; - char *ticket_buf; - u8 reply_struct_v; - - dbuf = kmalloc(TEMP_TICKET_BUF_LEN, GFP_NOFS); - if (!dbuf) - return -ENOMEM; - - ret = -ENOMEM; - ticket_buf = kmalloc(TEMP_TICKET_BUF_LEN, GFP_NOFS); - if (!ticket_buf) - goto out_dbuf; - - ceph_decode_need(&p, end, 1 + sizeof(u32), bad); - reply_struct_v = ceph_decode_8(&p); - if (reply_struct_v != 1) - goto bad; - num = ceph_decode_32(&p); - dout("%d tickets\n", num); - while (num--) { - int type; - u8 tkt_struct_v, blob_struct_v; - struct ceph_x_ticket_handler *th; - void *dp, *dend; - int dlen; - char is_enc; - struct timespec validity; - struct ceph_crypto_key old_key; - void *tp, *tpend; - struct ceph_timespec new_validity; - struct ceph_crypto_key new_session_key; - struct ceph_buffer *new_ticket_blob; - unsigned long new_expires, new_renew_after; - u64 new_secret_id; - - ceph_decode_need(&p, end, sizeof(u32) + 1, bad); - - type = ceph_decode_32(&p); - dout(" ticket type %d %s\n", type, ceph_entity_type_name(type)); - - tkt_struct_v = ceph_decode_8(&p); - if (tkt_struct_v != 1) - goto bad; - - th = get_ticket_handler(ac, type); - if (IS_ERR(th)) { - ret = PTR_ERR(th); - goto out; - } - - /* blob for me */ - dlen = ceph_x_decrypt(secret, &p, end, dbuf, - TEMP_TICKET_BUF_LEN); - if (dlen <= 0) { - ret = dlen; - goto out; - } - dout(" decrypted %d bytes\n", dlen); - dend = dbuf + dlen; - dp = dbuf; - - tkt_struct_v = ceph_decode_8(&dp); - if (tkt_struct_v != 1) - goto bad; - - memcpy(&old_key, &th->session_key, sizeof(old_key)); - ret = ceph_crypto_key_decode(&new_session_key, &dp, dend); - if (ret) - goto out; - - ceph_decode_copy(&dp, &new_validity, sizeof(new_validity)); - ceph_decode_timespec(&validity, &new_validity); - new_expires = get_seconds() + validity.tv_sec; - new_renew_after = new_expires - (validity.tv_sec / 4); - dout(" expires=%lu renew_after=%lu\n", new_expires, - new_renew_after); - - /* ticket blob for service */ - ceph_decode_8_safe(&p, end, is_enc, bad); - tp = ticket_buf; - if (is_enc) { - /* encrypted */ - dout(" encrypted ticket\n"); - dlen = ceph_x_decrypt(&old_key, &p, end, ticket_buf, - TEMP_TICKET_BUF_LEN); - if (dlen < 0) { - ret = dlen; - goto out; - } - dlen = ceph_decode_32(&tp); - } else { - /* unencrypted */ - ceph_decode_32_safe(&p, end, dlen, bad); - ceph_decode_need(&p, end, dlen, bad); - ceph_decode_copy(&p, ticket_buf, dlen); - } - tpend = tp + dlen; - dout(" ticket blob is %d bytes\n", dlen); - ceph_decode_need(&tp, tpend, 1 + sizeof(u64), bad); - blob_struct_v = ceph_decode_8(&tp); - new_secret_id = ceph_decode_64(&tp); - ret = ceph_decode_buffer(&new_ticket_blob, &tp, tpend); - if (ret) - goto out; - - /* all is well, update our ticket */ - ceph_crypto_key_destroy(&th->session_key); - if (th->ticket_blob) - ceph_buffer_put(th->ticket_blob); - th->session_key = new_session_key; - th->ticket_blob = new_ticket_blob; - th->validity = new_validity; - th->secret_id = new_secret_id; - th->expires = new_expires; - th->renew_after = new_renew_after; - dout(" got ticket service %d (%s) secret_id %lld len %d\n", - type, ceph_entity_type_name(type), th->secret_id, - (int)th->ticket_blob->vec.iov_len); - xi->have_keys |= th->service; - } - - ret = 0; -out: - kfree(ticket_buf); -out_dbuf: - kfree(dbuf); - return ret; - -bad: - ret = -EINVAL; - goto out; -} - -static int ceph_x_build_authorizer(struct ceph_auth_client *ac, - struct ceph_x_ticket_handler *th, - struct ceph_x_authorizer *au) -{ - int maxlen; - struct ceph_x_authorize_a *msg_a; - struct ceph_x_authorize_b msg_b; - void *p, *end; - int ret; - int ticket_blob_len = - (th->ticket_blob ? th->ticket_blob->vec.iov_len : 0); - - dout("build_authorizer for %s %p\n", - ceph_entity_type_name(th->service), au); - - maxlen = sizeof(*msg_a) + sizeof(msg_b) + - ceph_x_encrypt_buflen(ticket_blob_len); - dout(" need len %d\n", maxlen); - if (au->buf && au->buf->alloc_len < maxlen) { - ceph_buffer_put(au->buf); - au->buf = NULL; - } - if (!au->buf) { - au->buf = ceph_buffer_new(maxlen, GFP_NOFS); - if (!au->buf) - return -ENOMEM; - } - au->service = th->service; - - msg_a = au->buf->vec.iov_base; - msg_a->struct_v = 1; - msg_a->global_id = cpu_to_le64(ac->global_id); - msg_a->service_id = cpu_to_le32(th->service); - msg_a->ticket_blob.struct_v = 1; - msg_a->ticket_blob.secret_id = cpu_to_le64(th->secret_id); - msg_a->ticket_blob.blob_len = cpu_to_le32(ticket_blob_len); - if (ticket_blob_len) { - memcpy(msg_a->ticket_blob.blob, th->ticket_blob->vec.iov_base, - th->ticket_blob->vec.iov_len); - } - dout(" th %p secret_id %lld %lld\n", th, th->secret_id, - le64_to_cpu(msg_a->ticket_blob.secret_id)); - - p = msg_a + 1; - p += ticket_blob_len; - end = au->buf->vec.iov_base + au->buf->vec.iov_len; - - get_random_bytes(&au->nonce, sizeof(au->nonce)); - msg_b.struct_v = 1; - msg_b.nonce = cpu_to_le64(au->nonce); - ret = ceph_x_encrypt(&th->session_key, &msg_b, sizeof(msg_b), - p, end - p); - if (ret < 0) - goto out_buf; - p += ret; - au->buf->vec.iov_len = p - au->buf->vec.iov_base; - dout(" built authorizer nonce %llx len %d\n", au->nonce, - (int)au->buf->vec.iov_len); - BUG_ON(au->buf->vec.iov_len > maxlen); - return 0; - -out_buf: - ceph_buffer_put(au->buf); - au->buf = NULL; - return ret; -} - -static int ceph_x_encode_ticket(struct ceph_x_ticket_handler *th, - void **p, void *end) -{ - ceph_decode_need(p, end, 1 + sizeof(u64), bad); - ceph_encode_8(p, 1); - ceph_encode_64(p, th->secret_id); - if (th->ticket_blob) { - const char *buf = th->ticket_blob->vec.iov_base; - u32 len = th->ticket_blob->vec.iov_len; - - ceph_encode_32_safe(p, end, len, bad); - ceph_encode_copy_safe(p, end, buf, len, bad); - } else { - ceph_encode_32_safe(p, end, 0, bad); - } - - return 0; -bad: - return -ERANGE; -} - -static void ceph_x_validate_tickets(struct ceph_auth_client *ac, int *pneed) -{ - int want = ac->want_keys; - struct ceph_x_info *xi = ac->private; - int service; - - *pneed = ac->want_keys & ~(xi->have_keys); - - for (service = 1; service <= want; service <<= 1) { - struct ceph_x_ticket_handler *th; - - if (!(ac->want_keys & service)) - continue; - - if (*pneed & service) - continue; - - th = get_ticket_handler(ac, service); - - if (IS_ERR(th)) { - *pneed |= service; - continue; - } - - if (get_seconds() >= th->renew_after) - *pneed |= service; - if (get_seconds() >= th->expires) - xi->have_keys &= ~service; - } -} - - -static int ceph_x_build_request(struct ceph_auth_client *ac, - void *buf, void *end) -{ - struct ceph_x_info *xi = ac->private; - int need; - struct ceph_x_request_header *head = buf; - int ret; - struct ceph_x_ticket_handler *th = - get_ticket_handler(ac, CEPH_ENTITY_TYPE_AUTH); - - if (IS_ERR(th)) - return PTR_ERR(th); - - ceph_x_validate_tickets(ac, &need); - - dout("build_request want %x have %x need %x\n", - ac->want_keys, xi->have_keys, need); - - if (need & CEPH_ENTITY_TYPE_AUTH) { - struct ceph_x_authenticate *auth = (void *)(head + 1); - void *p = auth + 1; - struct ceph_x_challenge_blob tmp; - char tmp_enc[40]; - u64 *u; - - if (p > end) - return -ERANGE; - - dout(" get_auth_session_key\n"); - head->op = cpu_to_le16(CEPHX_GET_AUTH_SESSION_KEY); - - /* encrypt and hash */ - get_random_bytes(&auth->client_challenge, sizeof(u64)); - tmp.client_challenge = auth->client_challenge; - tmp.server_challenge = cpu_to_le64(xi->server_challenge); - ret = ceph_x_encrypt(&xi->secret, &tmp, sizeof(tmp), - tmp_enc, sizeof(tmp_enc)); - if (ret < 0) - return ret; - - auth->struct_v = 1; - auth->key = 0; - for (u = (u64 *)tmp_enc; u + 1 <= (u64 *)(tmp_enc + ret); u++) - auth->key ^= *(__le64 *)u; - dout(" server_challenge %llx client_challenge %llx key %llx\n", - xi->server_challenge, le64_to_cpu(auth->client_challenge), - le64_to_cpu(auth->key)); - - /* now encode the old ticket if exists */ - ret = ceph_x_encode_ticket(th, &p, end); - if (ret < 0) - return ret; - - return p - buf; - } - - if (need) { - void *p = head + 1; - struct ceph_x_service_ticket_request *req; - - if (p > end) - return -ERANGE; - head->op = cpu_to_le16(CEPHX_GET_PRINCIPAL_SESSION_KEY); - - ret = ceph_x_build_authorizer(ac, th, &xi->auth_authorizer); - if (ret) - return ret; - ceph_encode_copy(&p, xi->auth_authorizer.buf->vec.iov_base, - xi->auth_authorizer.buf->vec.iov_len); - - req = p; - req->keys = cpu_to_le32(need); - p += sizeof(*req); - return p - buf; - } - - return 0; -} - -static int ceph_x_handle_reply(struct ceph_auth_client *ac, int result, - void *buf, void *end) -{ - struct ceph_x_info *xi = ac->private; - struct ceph_x_reply_header *head = buf; - struct ceph_x_ticket_handler *th; - int len = end - buf; - int op; - int ret; - - if (result) - return result; /* XXX hmm? */ - - if (xi->starting) { - /* it's a hello */ - struct ceph_x_server_challenge *sc = buf; - - if (len != sizeof(*sc)) - return -EINVAL; - xi->server_challenge = le64_to_cpu(sc->server_challenge); - dout("handle_reply got server challenge %llx\n", - xi->server_challenge); - xi->starting = false; - xi->have_keys &= ~CEPH_ENTITY_TYPE_AUTH; - return -EAGAIN; - } - - op = le16_to_cpu(head->op); - result = le32_to_cpu(head->result); - dout("handle_reply op %d result %d\n", op, result); - switch (op) { - case CEPHX_GET_AUTH_SESSION_KEY: - /* verify auth key */ - ret = ceph_x_proc_ticket_reply(ac, &xi->secret, - buf + sizeof(*head), end); - break; - - case CEPHX_GET_PRINCIPAL_SESSION_KEY: - th = get_ticket_handler(ac, CEPH_ENTITY_TYPE_AUTH); - if (IS_ERR(th)) - return PTR_ERR(th); - ret = ceph_x_proc_ticket_reply(ac, &th->session_key, - buf + sizeof(*head), end); - break; - - default: - return -EINVAL; - } - if (ret) - return ret; - if (ac->want_keys == xi->have_keys) - return 0; - return -EAGAIN; -} - -static int ceph_x_create_authorizer( - struct ceph_auth_client *ac, int peer_type, - struct ceph_authorizer **a, - void **buf, size_t *len, - void **reply_buf, size_t *reply_len) -{ - struct ceph_x_authorizer *au; - struct ceph_x_ticket_handler *th; - int ret; - - th = get_ticket_handler(ac, peer_type); - if (IS_ERR(th)) - return PTR_ERR(th); - - au = kzalloc(sizeof(*au), GFP_NOFS); - if (!au) - return -ENOMEM; - - ret = ceph_x_build_authorizer(ac, th, au); - if (ret) { - kfree(au); - return ret; - } - - *a = (struct ceph_authorizer *)au; - *buf = au->buf->vec.iov_base; - *len = au->buf->vec.iov_len; - *reply_buf = au->reply_buf; - *reply_len = sizeof(au->reply_buf); - return 0; -} - -static int ceph_x_verify_authorizer_reply(struct ceph_auth_client *ac, - struct ceph_authorizer *a, size_t len) -{ - struct ceph_x_authorizer *au = (void *)a; - struct ceph_x_ticket_handler *th; - int ret = 0; - struct ceph_x_authorize_reply reply; - void *p = au->reply_buf; - void *end = p + sizeof(au->reply_buf); - - th = get_ticket_handler(ac, au->service); - if (IS_ERR(th)) - return PTR_ERR(th); - ret = ceph_x_decrypt(&th->session_key, &p, end, &reply, sizeof(reply)); - if (ret < 0) - return ret; - if (ret != sizeof(reply)) - return -EPERM; - - if (au->nonce + 1 != le64_to_cpu(reply.nonce_plus_one)) - ret = -EPERM; - else - ret = 0; - dout("verify_authorizer_reply nonce %llx got %llx ret %d\n", - au->nonce, le64_to_cpu(reply.nonce_plus_one), ret); - return ret; -} - -static void ceph_x_destroy_authorizer(struct ceph_auth_client *ac, - struct ceph_authorizer *a) -{ - struct ceph_x_authorizer *au = (void *)a; - - ceph_buffer_put(au->buf); - kfree(au); -} - - -static void ceph_x_reset(struct ceph_auth_client *ac) -{ - struct ceph_x_info *xi = ac->private; - - dout("reset\n"); - xi->starting = true; - xi->server_challenge = 0; -} - -static void ceph_x_destroy(struct ceph_auth_client *ac) -{ - struct ceph_x_info *xi = ac->private; - struct rb_node *p; - - dout("ceph_x_destroy %p\n", ac); - ceph_crypto_key_destroy(&xi->secret); - - while ((p = rb_first(&xi->ticket_handlers)) != NULL) { - struct ceph_x_ticket_handler *th = - rb_entry(p, struct ceph_x_ticket_handler, node); - remove_ticket_handler(ac, th); - } - - if (xi->auth_authorizer.buf) - ceph_buffer_put(xi->auth_authorizer.buf); - - kfree(ac->private); - ac->private = NULL; -} - -static void ceph_x_invalidate_authorizer(struct ceph_auth_client *ac, - int peer_type) -{ - struct ceph_x_ticket_handler *th; - - th = get_ticket_handler(ac, peer_type); - if (!IS_ERR(th)) - remove_ticket_handler(ac, th); -} - - -static const struct ceph_auth_client_ops ceph_x_ops = { - .name = "x", - .is_authenticated = ceph_x_is_authenticated, - .should_authenticate = ceph_x_should_authenticate, - .build_request = ceph_x_build_request, - .handle_reply = ceph_x_handle_reply, - .create_authorizer = ceph_x_create_authorizer, - .verify_authorizer_reply = ceph_x_verify_authorizer_reply, - .destroy_authorizer = ceph_x_destroy_authorizer, - .invalidate_authorizer = ceph_x_invalidate_authorizer, - .reset = ceph_x_reset, - .destroy = ceph_x_destroy, -}; - - -int ceph_x_init(struct ceph_auth_client *ac) -{ - struct ceph_x_info *xi; - int ret; - - dout("ceph_x_init %p\n", ac); - ret = -ENOMEM; - xi = kzalloc(sizeof(*xi), GFP_NOFS); - if (!xi) - goto out; - - ret = -EINVAL; - if (!ac->secret) { - pr_err("no secret set (for auth_x protocol)\n"); - goto out_nomem; - } - - ret = ceph_crypto_key_unarmor(&xi->secret, ac->secret); - if (ret) - goto out_nomem; - - xi->starting = true; - xi->ticket_handlers = RB_ROOT; - - ac->protocol = CEPH_AUTH_CEPHX; - ac->private = xi; - ac->ops = &ceph_x_ops; - return 0; - -out_nomem: - kfree(xi); -out: - return ret; -} - - diff --git a/fs/ceph/auth_x.h b/fs/ceph/auth_x.h deleted file mode 100644 index ff6f818..0000000 --- a/fs/ceph/auth_x.h +++ /dev/null @@ -1,49 +0,0 @@ -#ifndef _FS_CEPH_AUTH_X_H -#define _FS_CEPH_AUTH_X_H - -#include - -#include "crypto.h" -#include "auth.h" -#include "auth_x_protocol.h" - -/* - * Handle ticket for a single service. - */ -struct ceph_x_ticket_handler { - struct rb_node node; - unsigned service; - - struct ceph_crypto_key session_key; - struct ceph_timespec validity; - - u64 secret_id; - struct ceph_buffer *ticket_blob; - - unsigned long renew_after, expires; -}; - - -struct ceph_x_authorizer { - struct ceph_buffer *buf; - unsigned service; - u64 nonce; - char reply_buf[128]; /* big enough for encrypted blob */ -}; - -struct ceph_x_info { - struct ceph_crypto_key secret; - - bool starting; - u64 server_challenge; - - unsigned have_keys; - struct rb_root ticket_handlers; - - struct ceph_x_authorizer auth_authorizer; -}; - -extern int ceph_x_init(struct ceph_auth_client *ac); - -#endif - diff --git a/fs/ceph/auth_x_protocol.h b/fs/ceph/auth_x_protocol.h deleted file mode 100644 index 671d305..0000000 --- a/fs/ceph/auth_x_protocol.h +++ /dev/null @@ -1,90 +0,0 @@ -#ifndef __FS_CEPH_AUTH_X_PROTOCOL -#define __FS_CEPH_AUTH_X_PROTOCOL - -#define CEPHX_GET_AUTH_SESSION_KEY 0x0100 -#define CEPHX_GET_PRINCIPAL_SESSION_KEY 0x0200 -#define CEPHX_GET_ROTATING_KEY 0x0400 - -/* common bits */ -struct ceph_x_ticket_blob { - __u8 struct_v; - __le64 secret_id; - __le32 blob_len; - char blob[]; -} __attribute__ ((packed)); - - -/* common request/reply headers */ -struct ceph_x_request_header { - __le16 op; -} __attribute__ ((packed)); - -struct ceph_x_reply_header { - __le16 op; - __le32 result; -} __attribute__ ((packed)); - - -/* authenticate handshake */ - -/* initial hello (no reply header) */ -struct ceph_x_server_challenge { - __u8 struct_v; - __le64 server_challenge; -} __attribute__ ((packed)); - -struct ceph_x_authenticate { - __u8 struct_v; - __le64 client_challenge; - __le64 key; - /* ticket blob */ -} __attribute__ ((packed)); - -struct ceph_x_service_ticket_request { - __u8 struct_v; - __le32 keys; -} __attribute__ ((packed)); - -struct ceph_x_challenge_blob { - __le64 server_challenge; - __le64 client_challenge; -} __attribute__ ((packed)); - - - -/* authorize handshake */ - -/* - * The authorizer consists of two pieces: - * a - service id, ticket blob - * b - encrypted with session key - */ -struct ceph_x_authorize_a { - __u8 struct_v; - __le64 global_id; - __le32 service_id; - struct ceph_x_ticket_blob ticket_blob; -} __attribute__ ((packed)); - -struct ceph_x_authorize_b { - __u8 struct_v; - __le64 nonce; -} __attribute__ ((packed)); - -struct ceph_x_authorize_reply { - __u8 struct_v; - __le64 nonce_plus_one; -} __attribute__ ((packed)); - - -/* - * encyption bundle - */ -#define CEPHX_ENC_MAGIC 0xff009cad8826aa55ull - -struct ceph_x_encrypt_header { - __u8 struct_v; - __le64 magic; -} __attribute__ ((packed)); - -#endif diff --git a/fs/ceph/buffer.c b/fs/ceph/buffer.c deleted file mode 100644 index cd39f17..0000000 --- a/fs/ceph/buffer.c +++ /dev/null @@ -1,65 +0,0 @@ - -#include "ceph_debug.h" - -#include - -#include "buffer.h" -#include "decode.h" - -struct ceph_buffer *ceph_buffer_new(size_t len, gfp_t gfp) -{ - struct ceph_buffer *b; - - b = kmalloc(sizeof(*b), gfp); - if (!b) - return NULL; - - b->vec.iov_base = kmalloc(len, gfp | __GFP_NOWARN); - if (b->vec.iov_base) { - b->is_vmalloc = false; - } else { - b->vec.iov_base = __vmalloc(len, gfp, PAGE_KERNEL); - if (!b->vec.iov_base) { - kfree(b); - return NULL; - } - b->is_vmalloc = true; - } - - kref_init(&b->kref); - b->alloc_len = len; - b->vec.iov_len = len; - dout("buffer_new %p\n", b); - return b; -} - -void ceph_buffer_release(struct kref *kref) -{ - struct ceph_buffer *b = container_of(kref, struct ceph_buffer, kref); - - dout("buffer_release %p\n", b); - if (b->vec.iov_base) { - if (b->is_vmalloc) - vfree(b->vec.iov_base); - else - kfree(b->vec.iov_base); - } - kfree(b); -} - -int ceph_decode_buffer(struct ceph_buffer **b, void **p, void *end) -{ - size_t len; - - ceph_decode_need(p, end, sizeof(u32), bad); - len = ceph_decode_32(p); - dout("decode_buffer len %d\n", (int)len); - ceph_decode_need(p, end, len, bad); - *b = ceph_buffer_new(len, GFP_NOFS); - if (!*b) - return -ENOMEM; - ceph_decode_copy(p, (*b)->vec.iov_base, len); - return 0; -bad: - return -EINVAL; -} diff --git a/fs/ceph/buffer.h b/fs/ceph/buffer.h deleted file mode 100644 index 58d1901..0000000 --- a/fs/ceph/buffer.h +++ /dev/null @@ -1,39 +0,0 @@ -#ifndef __FS_CEPH_BUFFER_H -#define __FS_CEPH_BUFFER_H - -#include -#include -#include -#include -#include - -/* - * a simple reference counted buffer. - * - * use kmalloc for small sizes (<= one page), vmalloc for larger - * sizes. - */ -struct ceph_buffer { - struct kref kref; - struct kvec vec; - size_t alloc_len; - bool is_vmalloc; -}; - -extern struct ceph_buffer *ceph_buffer_new(size_t len, gfp_t gfp); -extern void ceph_buffer_release(struct kref *kref); - -static inline struct ceph_buffer *ceph_buffer_get(struct ceph_buffer *b) -{ - kref_get(&b->kref); - return b; -} - -static inline void ceph_buffer_put(struct ceph_buffer *b) -{ - kref_put(&b->kref, ceph_buffer_release); -} - -extern int ceph_decode_buffer(struct ceph_buffer **b, void **p, void *end); - -#endif diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 5e9da99..3cff67c 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1,4 +1,4 @@ -#include "ceph_debug.h" +#include #include #include @@ -9,8 +9,9 @@ #include #include "super.h" -#include "decode.h" -#include "messenger.h" +#include "mds_client.h" +#include +#include /* * Capability management @@ -287,11 +288,11 @@ void ceph_put_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap) spin_unlock(&mdsc->caps_list_lock); } -void ceph_reservation_status(struct ceph_client *client, +void ceph_reservation_status(struct ceph_fs_client *fsc, int *total, int *avail, int *used, int *reserved, int *min) { - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_mds_client *mdsc = fsc->mdsc; if (total) *total = mdsc->caps_total_count; @@ -399,7 +400,7 @@ static void __insert_cap_node(struct ceph_inode_info *ci, static void __cap_set_timeouts(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) { - struct ceph_mount_args *ma = mdsc->client->mount_args; + struct ceph_mount_options *ma = mdsc->fsc->mount_options; ci->i_hold_caps_min = round_jiffies(jiffies + ma->caps_wanted_delay_min * HZ); @@ -515,7 +516,7 @@ int ceph_add_cap(struct inode *inode, unsigned seq, unsigned mseq, u64 realmino, int flags, struct ceph_cap_reservation *caps_reservation) { - struct ceph_mds_client *mdsc = &ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_cap *new_cap = NULL; struct ceph_cap *cap; @@ -873,7 +874,7 @@ void __ceph_remove_cap(struct ceph_cap *cap) struct ceph_mds_session *session = cap->session; struct ceph_inode_info *ci = cap->ci; struct ceph_mds_client *mdsc = - &ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; + ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; int removed = 0; dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode); @@ -1210,7 +1211,7 @@ void __ceph_flush_snaps(struct ceph_inode_info *ci, int mds; struct ceph_cap_snap *capsnap; u32 mseq; - struct ceph_mds_client *mdsc = &ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; struct ceph_mds_session *session = NULL; /* if session != NULL, we hold session->s_mutex */ u64 next_follows = 0; /* keep track of how far we've gotten through the @@ -1336,7 +1337,7 @@ static void ceph_flush_snaps(struct ceph_inode_info *ci) void __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask) { struct ceph_mds_client *mdsc = - &ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; + ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; struct inode *inode = &ci->vfs_inode; int was = ci->i_dirty_caps; int dirty = 0; @@ -1378,7 +1379,7 @@ void __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask) static int __mark_caps_flushing(struct inode *inode, struct ceph_mds_session *session) { - struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_inode_info *ci = ceph_inode(inode); int flushing; @@ -1462,8 +1463,8 @@ static int try_nonblocking_invalidate(struct inode *inode) void ceph_check_caps(struct ceph_inode_info *ci, int flags, struct ceph_mds_session *session) { - struct ceph_client *client = ceph_inode_to_client(&ci->vfs_inode); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode); + struct ceph_mds_client *mdsc = fsc->mdsc; struct inode *inode = &ci->vfs_inode; struct ceph_cap *cap; int file_wanted, used; @@ -1706,7 +1707,7 @@ ack: static int try_flush_caps(struct inode *inode, struct ceph_mds_session *session, unsigned *flush_tid) { - struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_inode_info *ci = ceph_inode(inode); int unlock_session = session ? 0 : 1; int flushing = 0; @@ -1872,7 +1873,7 @@ int ceph_write_inode(struct inode *inode, struct writeback_control *wbc) caps_are_flushed(inode, flush_tid)); } else { struct ceph_mds_client *mdsc = - &ceph_sb_to_client(inode->i_sb)->mdsc; + ceph_sb_to_client(inode->i_sb)->mdsc; spin_lock(&inode->i_lock); if (__ceph_caps_dirty(ci)) @@ -2465,7 +2466,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, __releases(inode->i_lock) { struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; unsigned seq = le32_to_cpu(m->seq); int dirty = le32_to_cpu(m->dirty); int cleaned = 0; @@ -2713,7 +2714,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, struct ceph_msg *msg) { struct ceph_mds_client *mdsc = session->s_mdsc; - struct super_block *sb = mdsc->client->sb; + struct super_block *sb = mdsc->fsc->sb; struct inode *inode; struct ceph_cap *cap; struct ceph_mds_caps *h; diff --git a/fs/ceph/ceph_debug.h b/fs/ceph/ceph_debug.h deleted file mode 100644 index 1818c23..0000000 --- a/fs/ceph/ceph_debug.h +++ /dev/null @@ -1,37 +0,0 @@ -#ifndef _FS_CEPH_DEBUG_H -#define _FS_CEPH_DEBUG_H - -#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt - -#ifdef CONFIG_CEPH_FS_PRETTYDEBUG - -/* - * wrap pr_debug to include a filename:lineno prefix on each line. - * this incurs some overhead (kernel size and execution time) due to - * the extra function call at each call site. - */ - -# if defined(DEBUG) || defined(CONFIG_DYNAMIC_DEBUG) -extern const char *ceph_file_part(const char *s, int len); -# define dout(fmt, ...) \ - pr_debug(" %12.12s:%-4d : " fmt, \ - ceph_file_part(__FILE__, sizeof(__FILE__)), \ - __LINE__, ##__VA_ARGS__) -# else -/* faux printk call just to see any compiler warnings. */ -# define dout(fmt, ...) do { \ - if (0) \ - printk(KERN_DEBUG fmt, ##__VA_ARGS__); \ - } while (0) -# endif - -#else - -/* - * or, just wrap pr_debug - */ -# define dout(fmt, ...) pr_debug(" " fmt, ##__VA_ARGS__) - -#endif - -#endif diff --git a/fs/ceph/ceph_frag.c b/fs/ceph/ceph_frag.c index ab6cf35..bdce8b1 100644 --- a/fs/ceph/ceph_frag.c +++ b/fs/ceph/ceph_frag.c @@ -1,7 +1,8 @@ /* * Ceph 'frag' type */ -#include "types.h" +#include +#include int ceph_frag_compare(__u32 a, __u32 b) { diff --git a/fs/ceph/ceph_frag.h b/fs/ceph/ceph_frag.h deleted file mode 100644 index 5babb8e..0000000 --- a/fs/ceph/ceph_frag.h +++ /dev/null @@ -1,109 +0,0 @@ -#ifndef FS_CEPH_FRAG_H -#define FS_CEPH_FRAG_H - -/* - * "Frags" are a way to describe a subset of a 32-bit number space, - * using a mask and a value to match against that mask. Any given frag - * (subset of the number space) can be partitioned into 2^n sub-frags. - * - * Frags are encoded into a 32-bit word: - * 8 upper bits = "bits" - * 24 lower bits = "value" - * (We could go to 5+27 bits, but who cares.) - * - * We use the _most_ significant bits of the 24 bit value. This makes - * values logically sort. - * - * Unfortunately, because the "bits" field is still in the high bits, we - * can't sort encoded frags numerically. However, it does allow you - * to feed encoded frags as values into frag_contains_value. - */ -static inline __u32 ceph_frag_make(__u32 b, __u32 v) -{ - return (b << 24) | - (v & (0xffffffu << (24-b)) & 0xffffffu); -} -static inline __u32 ceph_frag_bits(__u32 f) -{ - return f >> 24; -} -static inline __u32 ceph_frag_value(__u32 f) -{ - return f & 0xffffffu; -} -static inline __u32 ceph_frag_mask(__u32 f) -{ - return (0xffffffu << (24-ceph_frag_bits(f))) & 0xffffffu; -} -static inline __u32 ceph_frag_mask_shift(__u32 f) -{ - return 24 - ceph_frag_bits(f); -} - -static inline int ceph_frag_contains_value(__u32 f, __u32 v) -{ - return (v & ceph_frag_mask(f)) == ceph_frag_value(f); -} -static inline int ceph_frag_contains_frag(__u32 f, __u32 sub) -{ - /* is sub as specific as us, and contained by us? */ - return ceph_frag_bits(sub) >= ceph_frag_bits(f) && - (ceph_frag_value(sub) & ceph_frag_mask(f)) == ceph_frag_value(f); -} - -static inline __u32 ceph_frag_parent(__u32 f) -{ - return ceph_frag_make(ceph_frag_bits(f) - 1, - ceph_frag_value(f) & (ceph_frag_mask(f) << 1)); -} -static inline int ceph_frag_is_left_child(__u32 f) -{ - return ceph_frag_bits(f) > 0 && - (ceph_frag_value(f) & (0x1000000 >> ceph_frag_bits(f))) == 0; -} -static inline int ceph_frag_is_right_child(__u32 f) -{ - return ceph_frag_bits(f) > 0 && - (ceph_frag_value(f) & (0x1000000 >> ceph_frag_bits(f))) == 1; -} -static inline __u32 ceph_frag_sibling(__u32 f) -{ - return ceph_frag_make(ceph_frag_bits(f), - ceph_frag_value(f) ^ (0x1000000 >> ceph_frag_bits(f))); -} -static inline __u32 ceph_frag_left_child(__u32 f) -{ - return ceph_frag_make(ceph_frag_bits(f)+1, ceph_frag_value(f)); -} -static inline __u32 ceph_frag_right_child(__u32 f) -{ - return ceph_frag_make(ceph_frag_bits(f)+1, - ceph_frag_value(f) | (0x1000000 >> (1+ceph_frag_bits(f)))); -} -static inline __u32 ceph_frag_make_child(__u32 f, int by, int i) -{ - int newbits = ceph_frag_bits(f) + by; - return ceph_frag_make(newbits, - ceph_frag_value(f) | (i << (24 - newbits))); -} -static inline int ceph_frag_is_leftmost(__u32 f) -{ - return ceph_frag_value(f) == 0; -} -static inline int ceph_frag_is_rightmost(__u32 f) -{ - return ceph_frag_value(f) == ceph_frag_mask(f); -} -static inline __u32 ceph_frag_next(__u32 f) -{ - return ceph_frag_make(ceph_frag_bits(f), - ceph_frag_value(f) + (0x1000000 >> ceph_frag_bits(f))); -} - -/* - * comparator to sort frags logically, as when traversing the - * number space in ascending order... - */ -int ceph_frag_compare(__u32 a, __u32 b); - -#endif diff --git a/fs/ceph/ceph_fs.c b/fs/ceph/ceph_fs.c deleted file mode 100644 index 3ac6cc7..0000000 --- a/fs/ceph/ceph_fs.c +++ /dev/null @@ -1,72 +0,0 @@ -/* - * Some non-inline ceph helpers - */ -#include "types.h" - -/* - * return true if @layout appears to be valid - */ -int ceph_file_layout_is_valid(const struct ceph_file_layout *layout) -{ - __u32 su = le32_to_cpu(layout->fl_stripe_unit); - __u32 sc = le32_to_cpu(layout->fl_stripe_count); - __u32 os = le32_to_cpu(layout->fl_object_size); - - /* stripe unit, object size must be non-zero, 64k increment */ - if (!su || (su & (CEPH_MIN_STRIPE_UNIT-1))) - return 0; - if (!os || (os & (CEPH_MIN_STRIPE_UNIT-1))) - return 0; - /* object size must be a multiple of stripe unit */ - if (os < su || os % su) - return 0; - /* stripe count must be non-zero */ - if (!sc) - return 0; - return 1; -} - - -int ceph_flags_to_mode(int flags) -{ - int mode; - -#ifdef O_DIRECTORY /* fixme */ - if ((flags & O_DIRECTORY) == O_DIRECTORY) - return CEPH_FILE_MODE_PIN; -#endif - if ((flags & O_APPEND) == O_APPEND) - flags |= O_WRONLY; - - if ((flags & O_ACCMODE) == O_RDWR) - mode = CEPH_FILE_MODE_RDWR; - else if ((flags & O_ACCMODE) == O_WRONLY) - mode = CEPH_FILE_MODE_WR; - else - mode = CEPH_FILE_MODE_RD; - -#ifdef O_LAZY - if (flags & O_LAZY) - mode |= CEPH_FILE_MODE_LAZY; -#endif - - return mode; -} - -int ceph_caps_for_mode(int mode) -{ - int caps = CEPH_CAP_PIN; - - if (mode & CEPH_FILE_MODE_RD) - caps |= CEPH_CAP_FILE_SHARED | - CEPH_CAP_FILE_RD | CEPH_CAP_FILE_CACHE; - if (mode & CEPH_FILE_MODE_WR) - caps |= CEPH_CAP_FILE_EXCL | - CEPH_CAP_FILE_WR | CEPH_CAP_FILE_BUFFER | - CEPH_CAP_AUTH_SHARED | CEPH_CAP_AUTH_EXCL | - CEPH_CAP_XATTR_SHARED | CEPH_CAP_XATTR_EXCL; - if (mode & CEPH_FILE_MODE_LAZY) - caps |= CEPH_CAP_FILE_LAZYIO; - - return caps; -} diff --git a/fs/ceph/ceph_fs.h b/fs/ceph/ceph_fs.h deleted file mode 100644 index d5619ac..0000000 --- a/fs/ceph/ceph_fs.h +++ /dev/null @@ -1,728 +0,0 @@ -/* - * ceph_fs.h - Ceph constants and data types to share between kernel and - * user space. - * - * Most types in this file are defined as little-endian, and are - * primarily intended to describe data structures that pass over the - * wire or that are stored on disk. - * - * LGPL2 - */ - -#ifndef CEPH_FS_H -#define CEPH_FS_H - -#include "msgr.h" -#include "rados.h" - -/* - * subprotocol versions. when specific messages types or high-level - * protocols change, bump the affected components. we keep rev - * internal cluster protocols separately from the public, - * client-facing protocol. - */ -#define CEPH_OSD_PROTOCOL 8 /* cluster internal */ -#define CEPH_MDS_PROTOCOL 12 /* cluster internal */ -#define CEPH_MON_PROTOCOL 5 /* cluster internal */ -#define CEPH_OSDC_PROTOCOL 24 /* server/client */ -#define CEPH_MDSC_PROTOCOL 32 /* server/client */ -#define CEPH_MONC_PROTOCOL 15 /* server/client */ - - -#define CEPH_INO_ROOT 1 -#define CEPH_INO_CEPH 2 /* hidden .ceph dir */ - -/* arbitrary limit on max # of monitors (cluster of 3 is typical) */ -#define CEPH_MAX_MON 31 - - -/* - * feature bits - */ -#define CEPH_FEATURE_UID (1<<0) -#define CEPH_FEATURE_NOSRCADDR (1<<1) -#define CEPH_FEATURE_MONCLOCKCHECK (1<<2) -#define CEPH_FEATURE_FLOCK (1<<3) - - -/* - * ceph_file_layout - describe data layout for a file/inode - */ -struct ceph_file_layout { - /* file -> object mapping */ - __le32 fl_stripe_unit; /* stripe unit, in bytes. must be multiple - of page size. */ - __le32 fl_stripe_count; /* over this many objects */ - __le32 fl_object_size; /* until objects are this big, then move to - new objects */ - __le32 fl_cas_hash; /* 0 = none; 1 = sha256 */ - - /* pg -> disk layout */ - __le32 fl_object_stripe_unit; /* for per-object parity, if any */ - - /* object -> pg layout */ - __le32 fl_pg_preferred; /* preferred primary for pg (-1 for none) */ - __le32 fl_pg_pool; /* namespace, crush ruleset, rep level */ -} __attribute__ ((packed)); - -#define CEPH_MIN_STRIPE_UNIT 65536 - -int ceph_file_layout_is_valid(const struct ceph_file_layout *layout); - - -/* crypto algorithms */ -#define CEPH_CRYPTO_NONE 0x0 -#define CEPH_CRYPTO_AES 0x1 - -#define CEPH_AES_IV "cephsageyudagreg" - -/* security/authentication protocols */ -#define CEPH_AUTH_UNKNOWN 0x0 -#define CEPH_AUTH_NONE 0x1 -#define CEPH_AUTH_CEPHX 0x2 - -#define CEPH_AUTH_UID_DEFAULT ((__u64) -1) - - -/********************************************* - * message layer - */ - -/* - * message types - */ - -/* misc */ -#define CEPH_MSG_SHUTDOWN 1 -#define CEPH_MSG_PING 2 - -/* client <-> monitor */ -#define CEPH_MSG_MON_MAP 4 -#define CEPH_MSG_MON_GET_MAP 5 -#define CEPH_MSG_STATFS 13 -#define CEPH_MSG_STATFS_REPLY 14 -#define CEPH_MSG_MON_SUBSCRIBE 15 -#define CEPH_MSG_MON_SUBSCRIBE_ACK 16 -#define CEPH_MSG_AUTH 17 -#define CEPH_MSG_AUTH_REPLY 18 - -/* client <-> mds */ -#define CEPH_MSG_MDS_MAP 21 - -#define CEPH_MSG_CLIENT_SESSION 22 -#define CEPH_MSG_CLIENT_RECONNECT 23 - -#define CEPH_MSG_CLIENT_REQUEST 24 -#define CEPH_MSG_CLIENT_REQUEST_FORWARD 25 -#define CEPH_MSG_CLIENT_REPLY 26 -#define CEPH_MSG_CLIENT_CAPS 0x310 -#define CEPH_MSG_CLIENT_LEASE 0x311 -#define CEPH_MSG_CLIENT_SNAP 0x312 -#define CEPH_MSG_CLIENT_CAPRELEASE 0x313 - -/* pool ops */ -#define CEPH_MSG_POOLOP_REPLY 48 -#define CEPH_MSG_POOLOP 49 - - -/* osd */ -#define CEPH_MSG_OSD_MAP 41 -#define CEPH_MSG_OSD_OP 42 -#define CEPH_MSG_OSD_OPREPLY 43 - -/* pool operations */ -enum { - POOL_OP_CREATE = 0x01, - POOL_OP_DELETE = 0x02, - POOL_OP_AUID_CHANGE = 0x03, - POOL_OP_CREATE_SNAP = 0x11, - POOL_OP_DELETE_SNAP = 0x12, - POOL_OP_CREATE_UNMANAGED_SNAP = 0x21, - POOL_OP_DELETE_UNMANAGED_SNAP = 0x22, -}; - -struct ceph_mon_request_header { - __le64 have_version; - __le16 session_mon; - __le64 session_mon_tid; -} __attribute__ ((packed)); - -struct ceph_mon_statfs { - struct ceph_mon_request_header monhdr; - struct ceph_fsid fsid; -} __attribute__ ((packed)); - -struct ceph_statfs { - __le64 kb, kb_used, kb_avail; - __le64 num_objects; -} __attribute__ ((packed)); - -struct ceph_mon_statfs_reply { - struct ceph_fsid fsid; - __le64 version; - struct ceph_statfs st; -} __attribute__ ((packed)); - -const char *ceph_pool_op_name(int op); - -struct ceph_mon_poolop { - struct ceph_mon_request_header monhdr; - struct ceph_fsid fsid; - __le32 pool; - __le32 op; - __le64 auid; - __le64 snapid; - __le32 name_len; -} __attribute__ ((packed)); - -struct ceph_mon_poolop_reply { - struct ceph_mon_request_header monhdr; - struct ceph_fsid fsid; - __le32 reply_code; - __le32 epoch; - char has_data; - char data[0]; -} __attribute__ ((packed)); - -struct ceph_mon_unmanaged_snap { - __le64 snapid; -} __attribute__ ((packed)); - -struct ceph_osd_getmap { - struct ceph_mon_request_header monhdr; - struct ceph_fsid fsid; - __le32 start; -} __attribute__ ((packed)); - -struct ceph_mds_getmap { - struct ceph_mon_request_header monhdr; - struct ceph_fsid fsid; -} __attribute__ ((packed)); - -struct ceph_client_mount { - struct ceph_mon_request_header monhdr; -} __attribute__ ((packed)); - -struct ceph_mon_subscribe_item { - __le64 have_version; __le64 have; - __u8 onetime; -} __attribute__ ((packed)); - -struct ceph_mon_subscribe_ack { - __le32 duration; /* seconds */ - struct ceph_fsid fsid; -} __attribute__ ((packed)); - -/* - * mds states - * > 0 -> in - * <= 0 -> out - */ -#define CEPH_MDS_STATE_DNE 0 /* down, does not exist. */ -#define CEPH_MDS_STATE_STOPPED -1 /* down, once existed, but no subtrees. - empty log. */ -#define CEPH_MDS_STATE_BOOT -4 /* up, boot announcement. */ -#define CEPH_MDS_STATE_STANDBY -5 /* up, idle. waiting for assignment. */ -#define CEPH_MDS_STATE_CREATING -6 /* up, creating MDS instance. */ -#define CEPH_MDS_STATE_STARTING -7 /* up, starting previously stopped mds */ -#define CEPH_MDS_STATE_STANDBY_REPLAY -8 /* up, tailing active node's journal */ - -#define CEPH_MDS_STATE_REPLAY 8 /* up, replaying journal. */ -#define CEPH_MDS_STATE_RESOLVE 9 /* up, disambiguating distributed - operations (import, rename, etc.) */ -#define CEPH_MDS_STATE_RECONNECT 10 /* up, reconnect to clients */ -#define CEPH_MDS_STATE_REJOIN 11 /* up, rejoining distributed cache */ -#define CEPH_MDS_STATE_CLIENTREPLAY 12 /* up, replaying client operations */ -#define CEPH_MDS_STATE_ACTIVE 13 /* up, active */ -#define CEPH_MDS_STATE_STOPPING 14 /* up, but exporting metadata */ - -extern const char *ceph_mds_state_name(int s); - - -/* - * metadata lock types. - * - these are bitmasks.. we can compose them - * - they also define the lock ordering by the MDS - * - a few of these are internal to the mds - */ -#define CEPH_LOCK_DVERSION 1 -#define CEPH_LOCK_DN 2 -#define CEPH_LOCK_ISNAP 16 -#define CEPH_LOCK_IVERSION 32 /* mds internal */ -#define CEPH_LOCK_IFILE 64 -#define CEPH_LOCK_IAUTH 128 -#define CEPH_LOCK_ILINK 256 -#define CEPH_LOCK_IDFT 512 /* dir frag tree */ -#define CEPH_LOCK_INEST 1024 /* mds internal */ -#define CEPH_LOCK_IXATTR 2048 -#define CEPH_LOCK_IFLOCK 4096 /* advisory file locks */ -#define CEPH_LOCK_INO 8192 /* immutable inode bits; not a lock */ - -/* client_session ops */ -enum { - CEPH_SESSION_REQUEST_OPEN, - CEPH_SESSION_OPEN, - CEPH_SESSION_REQUEST_CLOSE, - CEPH_SESSION_CLOSE, - CEPH_SESSION_REQUEST_RENEWCAPS, - CEPH_SESSION_RENEWCAPS, - CEPH_SESSION_STALE, - CEPH_SESSION_RECALL_STATE, -}; - -extern const char *ceph_session_op_name(int op); - -struct ceph_mds_session_head { - __le32 op; - __le64 seq; - struct ceph_timespec stamp; - __le32 max_caps, max_leases; -} __attribute__ ((packed)); - -/* client_request */ -/* - * metadata ops. - * & 0x001000 -> write op - * & 0x010000 -> follow symlink (e.g. stat(), not lstat()). - & & 0x100000 -> use weird ino/path trace - */ -#define CEPH_MDS_OP_WRITE 0x001000 -enum { - CEPH_MDS_OP_LOOKUP = 0x00100, - CEPH_MDS_OP_GETATTR = 0x00101, - CEPH_MDS_OP_LOOKUPHASH = 0x00102, - CEPH_MDS_OP_LOOKUPPARENT = 0x00103, - - CEPH_MDS_OP_SETXATTR = 0x01105, - CEPH_MDS_OP_RMXATTR = 0x01106, - CEPH_MDS_OP_SETLAYOUT = 0x01107, - CEPH_MDS_OP_SETATTR = 0x01108, - CEPH_MDS_OP_SETFILELOCK= 0x01109, - CEPH_MDS_OP_GETFILELOCK= 0x00110, - - CEPH_MDS_OP_MKNOD = 0x01201, - CEPH_MDS_OP_LINK = 0x01202, - CEPH_MDS_OP_UNLINK = 0x01203, - CEPH_MDS_OP_RENAME = 0x01204, - CEPH_MDS_OP_MKDIR = 0x01220, - CEPH_MDS_OP_RMDIR = 0x01221, - CEPH_MDS_OP_SYMLINK = 0x01222, - - CEPH_MDS_OP_CREATE = 0x01301, - CEPH_MDS_OP_OPEN = 0x00302, - CEPH_MDS_OP_READDIR = 0x00305, - - CEPH_MDS_OP_LOOKUPSNAP = 0x00400, - CEPH_MDS_OP_MKSNAP = 0x01400, - CEPH_MDS_OP_RMSNAP = 0x01401, - CEPH_MDS_OP_LSSNAP = 0x00402, -}; - -extern const char *ceph_mds_op_name(int op); - - -#define CEPH_SETATTR_MODE 1 -#define CEPH_SETATTR_UID 2 -#define CEPH_SETATTR_GID 4 -#define CEPH_SETATTR_MTIME 8 -#define CEPH_SETATTR_ATIME 16 -#define CEPH_SETATTR_SIZE 32 -#define CEPH_SETATTR_CTIME 64 - -union ceph_mds_request_args { - struct { - __le32 mask; /* CEPH_CAP_* */ - } __attribute__ ((packed)) getattr; - struct { - __le32 mode; - __le32 uid; - __le32 gid; - struct ceph_timespec mtime; - struct ceph_timespec atime; - __le64 size, old_size; /* old_size needed by truncate */ - __le32 mask; /* CEPH_SETATTR_* */ - } __attribute__ ((packed)) setattr; - struct { - __le32 frag; /* which dir fragment */ - __le32 max_entries; /* how many dentries to grab */ - __le32 max_bytes; - } __attribute__ ((packed)) readdir; - struct { - __le32 mode; - __le32 rdev; - } __attribute__ ((packed)) mknod; - struct { - __le32 mode; - } __attribute__ ((packed)) mkdir; - struct { - __le32 flags; - __le32 mode; - __le32 stripe_unit; /* layout for newly created file */ - __le32 stripe_count; /* ... */ - __le32 object_size; - __le32 file_replication; - __le32 preferred; - } __attribute__ ((packed)) open; - struct { - __le32 flags; - } __attribute__ ((packed)) setxattr; - struct { - struct ceph_file_layout layout; - } __attribute__ ((packed)) setlayout; - struct { - __u8 rule; /* currently fcntl or flock */ - __u8 type; /* shared, exclusive, remove*/ - __le64 pid; /* process id requesting the lock */ - __le64 pid_namespace; - __le64 start; /* initial location to lock */ - __le64 length; /* num bytes to lock from start */ - __u8 wait; /* will caller wait for lock to become available? */ - } __attribute__ ((packed)) filelock_change; -} __attribute__ ((packed)); - -#define CEPH_MDS_FLAG_REPLAY 1 /* this is a replayed op */ -#define CEPH_MDS_FLAG_WANT_DENTRY 2 /* want dentry in reply */ - -struct ceph_mds_request_head { - __le64 oldest_client_tid; - __le32 mdsmap_epoch; /* on client */ - __le32 flags; /* CEPH_MDS_FLAG_* */ - __u8 num_retry, num_fwd; /* count retry, fwd attempts */ - __le16 num_releases; /* # include cap/lease release records */ - __le32 op; /* mds op code */ - __le32 caller_uid, caller_gid; - __le64 ino; /* use this ino for openc, mkdir, mknod, - etc. (if replaying) */ - union ceph_mds_request_args args; -} __attribute__ ((packed)); - -/* cap/lease release record */ -struct ceph_mds_request_release { - __le64 ino, cap_id; /* ino and unique cap id */ - __le32 caps, wanted; /* new issued, wanted */ - __le32 seq, issue_seq, mseq; - __le32 dname_seq; /* if releasing a dentry lease, a */ - __le32 dname_len; /* string follows. */ -} __attribute__ ((packed)); - -/* client reply */ -struct ceph_mds_reply_head { - __le32 op; - __le32 result; - __le32 mdsmap_epoch; - __u8 safe; /* true if committed to disk */ - __u8 is_dentry, is_target; /* true if dentry, target inode records - are included with reply */ -} __attribute__ ((packed)); - -/* one for each node split */ -struct ceph_frag_tree_split { - __le32 frag; /* this frag splits... */ - __le32 by; /* ...by this many bits */ -} __attribute__ ((packed)); - -struct ceph_frag_tree_head { - __le32 nsplits; /* num ceph_frag_tree_split records */ - struct ceph_frag_tree_split splits[]; -} __attribute__ ((packed)); - -/* capability issue, for bundling with mds reply */ -struct ceph_mds_reply_cap { - __le32 caps, wanted; /* caps issued, wanted */ - __le64 cap_id; - __le32 seq, mseq; - __le64 realm; /* snap realm */ - __u8 flags; /* CEPH_CAP_FLAG_* */ -} __attribute__ ((packed)); - -#define CEPH_CAP_FLAG_AUTH 1 /* cap is issued by auth mds */ - -/* inode record, for bundling with mds reply */ -struct ceph_mds_reply_inode { - __le64 ino; - __le64 snapid; - __le32 rdev; - __le64 version; /* inode version */ - __le64 xattr_version; /* version for xattr blob */ - struct ceph_mds_reply_cap cap; /* caps issued for this inode */ - struct ceph_file_layout layout; - struct ceph_timespec ctime, mtime, atime; - __le32 time_warp_seq; - __le64 size, max_size, truncate_size; - __le32 truncate_seq; - __le32 mode, uid, gid; - __le32 nlink; - __le64 files, subdirs, rbytes, rfiles, rsubdirs; /* dir stats */ - struct ceph_timespec rctime; - struct ceph_frag_tree_head fragtree; /* (must be at end of struct) */ -} __attribute__ ((packed)); -/* followed by frag array, then symlink string, then xattr blob */ - -/* reply_lease follows dname, and reply_inode */ -struct ceph_mds_reply_lease { - __le16 mask; /* lease type(s) */ - __le32 duration_ms; /* lease duration */ - __le32 seq; -} __attribute__ ((packed)); - -struct ceph_mds_reply_dirfrag { - __le32 frag; /* fragment */ - __le32 auth; /* auth mds, if this is a delegation point */ - __le32 ndist; /* number of mds' this is replicated on */ - __le32 dist[]; -} __attribute__ ((packed)); - -#define CEPH_LOCK_FCNTL 1 -#define CEPH_LOCK_FLOCK 2 - -#define CEPH_LOCK_SHARED 1 -#define CEPH_LOCK_EXCL 2 -#define CEPH_LOCK_UNLOCK 4 - -struct ceph_filelock { - __le64 start;/* file offset to start lock at */ - __le64 length; /* num bytes to lock; 0 for all following start */ - __le64 client; /* which client holds the lock */ - __le64 pid; /* process id holding the lock on the client */ - __le64 pid_namespace; - __u8 type; /* shared lock, exclusive lock, or unlock */ -} __attribute__ ((packed)); - - -/* file access modes */ -#define CEPH_FILE_MODE_PIN 0 -#define CEPH_FILE_MODE_RD 1 -#define CEPH_FILE_MODE_WR 2 -#define CEPH_FILE_MODE_RDWR 3 /* RD | WR */ -#define CEPH_FILE_MODE_LAZY 4 /* lazy io */ -#define CEPH_FILE_MODE_NUM 8 /* bc these are bit fields.. mostly */ - -int ceph_flags_to_mode(int flags); - - -/* capability bits */ -#define CEPH_CAP_PIN 1 /* no specific capabilities beyond the pin */ - -/* generic cap bits */ -#define CEPH_CAP_GSHARED 1 /* client can reads */ -#define CEPH_CAP_GEXCL 2 /* client can read and update */ -#define CEPH_CAP_GCACHE 4 /* (file) client can cache reads */ -#define CEPH_CAP_GRD 8 /* (file) client can read */ -#define CEPH_CAP_GWR 16 /* (file) client can write */ -#define CEPH_CAP_GBUFFER 32 /* (file) client can buffer writes */ -#define CEPH_CAP_GWREXTEND 64 /* (file) client can extend EOF */ -#define CEPH_CAP_GLAZYIO 128 /* (file) client can perform lazy io */ - -/* per-lock shift */ -#define CEPH_CAP_SAUTH 2 -#define CEPH_CAP_SLINK 4 -#define CEPH_CAP_SXATTR 6 -#define CEPH_CAP_SFILE 8 -#define CEPH_CAP_SFLOCK 20 - -#define CEPH_CAP_BITS 22 - -/* composed values */ -#define CEPH_CAP_AUTH_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SAUTH) -#define CEPH_CAP_AUTH_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SAUTH) -#define CEPH_CAP_LINK_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SLINK) -#define CEPH_CAP_LINK_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SLINK) -#define CEPH_CAP_XATTR_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SXATTR) -#define CEPH_CAP_XATTR_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SXATTR) -#define CEPH_CAP_FILE(x) (x << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_CACHE (CEPH_CAP_GCACHE << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_RD (CEPH_CAP_GRD << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_WR (CEPH_CAP_GWR << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_BUFFER (CEPH_CAP_GBUFFER << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_WREXTEND (CEPH_CAP_GWREXTEND << CEPH_CAP_SFILE) -#define CEPH_CAP_FILE_LAZYIO (CEPH_CAP_GLAZYIO << CEPH_CAP_SFILE) -#define CEPH_CAP_FLOCK_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SFLOCK) -#define CEPH_CAP_FLOCK_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SFLOCK) - - -/* cap masks (for getattr) */ -#define CEPH_STAT_CAP_INODE CEPH_CAP_PIN -#define CEPH_STAT_CAP_TYPE CEPH_CAP_PIN /* mode >> 12 */ -#define CEPH_STAT_CAP_SYMLINK CEPH_CAP_PIN -#define CEPH_STAT_CAP_UID CEPH_CAP_AUTH_SHARED -#define CEPH_STAT_CAP_GID CEPH_CAP_AUTH_SHARED -#define CEPH_STAT_CAP_MODE CEPH_CAP_AUTH_SHARED -#define CEPH_STAT_CAP_NLINK CEPH_CAP_LINK_SHARED -#define CEPH_STAT_CAP_LAYOUT CEPH_CAP_FILE_SHARED -#define CEPH_STAT_CAP_MTIME CEPH_CAP_FILE_SHARED -#define CEPH_STAT_CAP_SIZE CEPH_CAP_FILE_SHARED -#define CEPH_STAT_CAP_ATIME CEPH_CAP_FILE_SHARED /* fixme */ -#define CEPH_STAT_CAP_XATTR CEPH_CAP_XATTR_SHARED -#define CEPH_STAT_CAP_INODE_ALL (CEPH_CAP_PIN | \ - CEPH_CAP_AUTH_SHARED | \ - CEPH_CAP_LINK_SHARED | \ - CEPH_CAP_FILE_SHARED | \ - CEPH_CAP_XATTR_SHARED) - -#define CEPH_CAP_ANY_SHARED (CEPH_CAP_AUTH_SHARED | \ - CEPH_CAP_LINK_SHARED | \ - CEPH_CAP_XATTR_SHARED | \ - CEPH_CAP_FILE_SHARED) -#define CEPH_CAP_ANY_RD (CEPH_CAP_ANY_SHARED | CEPH_CAP_FILE_RD | \ - CEPH_CAP_FILE_CACHE) - -#define CEPH_CAP_ANY_EXCL (CEPH_CAP_AUTH_EXCL | \ - CEPH_CAP_LINK_EXCL | \ - CEPH_CAP_XATTR_EXCL | \ - CEPH_CAP_FILE_EXCL) -#define CEPH_CAP_ANY_FILE_WR (CEPH_CAP_FILE_WR | CEPH_CAP_FILE_BUFFER | \ - CEPH_CAP_FILE_EXCL) -#define CEPH_CAP_ANY_WR (CEPH_CAP_ANY_EXCL | CEPH_CAP_ANY_FILE_WR) -#define CEPH_CAP_ANY (CEPH_CAP_ANY_RD | CEPH_CAP_ANY_EXCL | \ - CEPH_CAP_ANY_FILE_WR | CEPH_CAP_FILE_LAZYIO | \ - CEPH_CAP_PIN) - -#define CEPH_CAP_LOCKS (CEPH_LOCK_IFILE | CEPH_LOCK_IAUTH | CEPH_LOCK_ILINK | \ - CEPH_LOCK_IXATTR) - -int ceph_caps_for_mode(int mode); - -enum { - CEPH_CAP_OP_GRANT, /* mds->client grant */ - CEPH_CAP_OP_REVOKE, /* mds->client revoke */ - CEPH_CAP_OP_TRUNC, /* mds->client trunc notify */ - CEPH_CAP_OP_EXPORT, /* mds has exported the cap */ - CEPH_CAP_OP_IMPORT, /* mds has imported the cap */ - CEPH_CAP_OP_UPDATE, /* client->mds update */ - CEPH_CAP_OP_DROP, /* client->mds drop cap bits */ - CEPH_CAP_OP_FLUSH, /* client->mds cap writeback */ - CEPH_CAP_OP_FLUSH_ACK, /* mds->client flushed */ - CEPH_CAP_OP_FLUSHSNAP, /* client->mds flush snapped metadata */ - CEPH_CAP_OP_FLUSHSNAP_ACK, /* mds->client flushed snapped metadata */ - CEPH_CAP_OP_RELEASE, /* client->mds release (clean) cap */ - CEPH_CAP_OP_RENEW, /* client->mds renewal request */ -}; - -extern const char *ceph_cap_op_name(int op); - -/* - * caps message, used for capability callbacks, acks, requests, etc. - */ -struct ceph_mds_caps { - __le32 op; /* CEPH_CAP_OP_* */ - __le64 ino, realm; - __le64 cap_id; - __le32 seq, issue_seq; - __le32 caps, wanted, dirty; /* latest issued/wanted/dirty */ - __le32 migrate_seq; - __le64 snap_follows; - __le32 snap_trace_len; - - /* authlock */ - __le32 uid, gid, mode; - - /* linklock */ - __le32 nlink; - - /* xattrlock */ - __le32 xattr_len; - __le64 xattr_version; - - /* filelock */ - __le64 size, max_size, truncate_size; - __le32 truncate_seq; - struct ceph_timespec mtime, atime, ctime; - struct ceph_file_layout layout; - __le32 time_warp_seq; -} __attribute__ ((packed)); - -/* cap release msg head */ -struct ceph_mds_cap_release { - __le32 num; /* number of cap_items that follow */ -} __attribute__ ((packed)); - -struct ceph_mds_cap_item { - __le64 ino; - __le64 cap_id; - __le32 migrate_seq, seq; -} __attribute__ ((packed)); - -#define CEPH_MDS_LEASE_REVOKE 1 /* mds -> client */ -#define CEPH_MDS_LEASE_RELEASE 2 /* client -> mds */ -#define CEPH_MDS_LEASE_RENEW 3 /* client <-> mds */ -#define CEPH_MDS_LEASE_REVOKE_ACK 4 /* client -> mds */ - -extern const char *ceph_lease_op_name(int o); - -/* lease msg header */ -struct ceph_mds_lease { - __u8 action; /* CEPH_MDS_LEASE_* */ - __le16 mask; /* which lease */ - __le64 ino; - __le64 first, last; /* snap range */ - __le32 seq; - __le32 duration_ms; /* duration of renewal */ -} __attribute__ ((packed)); -/* followed by a __le32+string for dname */ - -/* client reconnect */ -struct ceph_mds_cap_reconnect { - __le64 cap_id; - __le32 wanted; - __le32 issued; - __le64 snaprealm; - __le64 pathbase; /* base ino for our path to this ino */ - __le32 flock_len; /* size of flock state blob, if any */ -} __attribute__ ((packed)); -/* followed by flock blob */ - -struct ceph_mds_cap_reconnect_v1 { - __le64 cap_id; - __le32 wanted; - __le32 issued; - __le64 size; - struct ceph_timespec mtime, atime; - __le64 snaprealm; - __le64 pathbase; /* base ino for our path to this ino */ -} __attribute__ ((packed)); - -struct ceph_mds_snaprealm_reconnect { - __le64 ino; /* snap realm base */ - __le64 seq; /* snap seq for this snap realm */ - __le64 parent; /* parent realm */ -} __attribute__ ((packed)); - -/* - * snaps - */ -enum { - CEPH_SNAP_OP_UPDATE, /* CREATE or DESTROY */ - CEPH_SNAP_OP_CREATE, - CEPH_SNAP_OP_DESTROY, - CEPH_SNAP_OP_SPLIT, -}; - -extern const char *ceph_snap_op_name(int o); - -/* snap msg header */ -struct ceph_mds_snap_head { - __le32 op; /* CEPH_SNAP_OP_* */ - __le64 split; /* ino to split off, if any */ - __le32 num_split_inos; /* # inos belonging to new child realm */ - __le32 num_split_realms; /* # child realms udner new child realm */ - __le32 trace_len; /* size of snap trace blob */ -} __attribute__ ((packed)); -/* followed by split ino list, then split realms, then the trace blob */ - -/* - * encode info about a snaprealm, as viewed by a client - */ -struct ceph_mds_snap_realm { - __le64 ino; /* ino */ - __le64 created; /* snap: when created */ - __le64 parent; /* ino: parent realm */ - __le64 parent_since; /* snap: same parent since */ - __le64 seq; /* snap: version */ - __le32 num_snaps; - __le32 num_prior_parent_snaps; -} __attribute__ ((packed)); -/* followed by my snap list, then prior parent snap list */ - -#endif diff --git a/fs/ceph/ceph_hash.c b/fs/ceph/ceph_hash.c deleted file mode 100644 index bd57001..0000000 --- a/fs/ceph/ceph_hash.c +++ /dev/null @@ -1,118 +0,0 @@ - -#include "types.h" - -/* - * Robert Jenkin's hash function. - * http://burtleburtle.net/bob/hash/evahash.html - * This is in the public domain. - */ -#define mix(a, b, c) \ - do { \ - a = a - b; a = a - c; a = a ^ (c >> 13); \ - b = b - c; b = b - a; b = b ^ (a << 8); \ - c = c - a; c = c - b; c = c ^ (b >> 13); \ - a = a - b; a = a - c; a = a ^ (c >> 12); \ - b = b - c; b = b - a; b = b ^ (a << 16); \ - c = c - a; c = c - b; c = c ^ (b >> 5); \ - a = a - b; a = a - c; a = a ^ (c >> 3); \ - b = b - c; b = b - a; b = b ^ (a << 10); \ - c = c - a; c = c - b; c = c ^ (b >> 15); \ - } while (0) - -unsigned ceph_str_hash_rjenkins(const char *str, unsigned length) -{ - const unsigned char *k = (const unsigned char *)str; - __u32 a, b, c; /* the internal state */ - __u32 len; /* how many key bytes still need mixing */ - - /* Set up the internal state */ - len = length; - a = 0x9e3779b9; /* the golden ratio; an arbitrary value */ - b = a; - c = 0; /* variable initialization of internal state */ - - /* handle most of the key */ - while (len >= 12) { - a = a + (k[0] + ((__u32)k[1] << 8) + ((__u32)k[2] << 16) + - ((__u32)k[3] << 24)); - b = b + (k[4] + ((__u32)k[5] << 8) + ((__u32)k[6] << 16) + - ((__u32)k[7] << 24)); - c = c + (k[8] + ((__u32)k[9] << 8) + ((__u32)k[10] << 16) + - ((__u32)k[11] << 24)); - mix(a, b, c); - k = k + 12; - len = len - 12; - } - - /* handle the last 11 bytes */ - c = c + length; - switch (len) { /* all the case statements fall through */ - case 11: - c = c + ((__u32)k[10] << 24); - case 10: - c = c + ((__u32)k[9] << 16); - case 9: - c = c + ((__u32)k[8] << 8); - /* the first byte of c is reserved for the length */ - case 8: - b = b + ((__u32)k[7] << 24); - case 7: - b = b + ((__u32)k[6] << 16); - case 6: - b = b + ((__u32)k[5] << 8); - case 5: - b = b + k[4]; - case 4: - a = a + ((__u32)k[3] << 24); - case 3: - a = a + ((__u32)k[2] << 16); - case 2: - a = a + ((__u32)k[1] << 8); - case 1: - a = a + k[0]; - /* case 0: nothing left to add */ - } - mix(a, b, c); - - return c; -} - -/* - * linux dcache hash - */ -unsigned ceph_str_hash_linux(const char *str, unsigned length) -{ - unsigned long hash = 0; - unsigned char c; - - while (length--) { - c = *str++; - hash = (hash + (c << 4) + (c >> 4)) * 11; - } - return hash; -} - - -unsigned ceph_str_hash(int type, const char *s, unsigned len) -{ - switch (type) { - case CEPH_STR_HASH_LINUX: - return ceph_str_hash_linux(s, len); - case CEPH_STR_HASH_RJENKINS: - return ceph_str_hash_rjenkins(s, len); - default: - return -1; - } -} - -const char *ceph_str_hash_name(int type) -{ - switch (type) { - case CEPH_STR_HASH_LINUX: - return "linux"; - case CEPH_STR_HASH_RJENKINS: - return "rjenkins"; - default: - return "unknown"; - } -} diff --git a/fs/ceph/ceph_hash.h b/fs/ceph/ceph_hash.h deleted file mode 100644 index d099c3f..0000000 --- a/fs/ceph/ceph_hash.h +++ /dev/null @@ -1,13 +0,0 @@ -#ifndef FS_CEPH_HASH_H -#define FS_CEPH_HASH_H - -#define CEPH_STR_HASH_LINUX 0x1 /* linux dcache hash */ -#define CEPH_STR_HASH_RJENKINS 0x2 /* robert jenkins' */ - -extern unsigned ceph_str_hash_linux(const char *s, unsigned len); -extern unsigned ceph_str_hash_rjenkins(const char *s, unsigned len); - -extern unsigned ceph_str_hash(int type, const char *s, unsigned len); -extern const char *ceph_str_hash_name(int type); - -#endif diff --git a/fs/ceph/ceph_strings.c b/fs/ceph/ceph_strings.c deleted file mode 100644 index c6179d3..0000000 --- a/fs/ceph/ceph_strings.c +++ /dev/null @@ -1,193 +0,0 @@ -/* - * Ceph string constants - */ -#include "types.h" - -const char *ceph_entity_type_name(int type) -{ - switch (type) { - case CEPH_ENTITY_TYPE_MDS: return "mds"; - case CEPH_ENTITY_TYPE_OSD: return "osd"; - case CEPH_ENTITY_TYPE_MON: return "mon"; - case CEPH_ENTITY_TYPE_CLIENT: return "client"; - case CEPH_ENTITY_TYPE_AUTH: return "auth"; - default: return "unknown"; - } -} - -const char *ceph_osd_op_name(int op) -{ - switch (op) { - case CEPH_OSD_OP_READ: return "read"; - case CEPH_OSD_OP_STAT: return "stat"; - - case CEPH_OSD_OP_MASKTRUNC: return "masktrunc"; - - case CEPH_OSD_OP_WRITE: return "write"; - case CEPH_OSD_OP_DELETE: return "delete"; - case CEPH_OSD_OP_TRUNCATE: return "truncate"; - case CEPH_OSD_OP_ZERO: return "zero"; - case CEPH_OSD_OP_WRITEFULL: return "writefull"; - case CEPH_OSD_OP_ROLLBACK: return "rollback"; - - case CEPH_OSD_OP_APPEND: return "append"; - case CEPH_OSD_OP_STARTSYNC: return "startsync"; - case CEPH_OSD_OP_SETTRUNC: return "settrunc"; - case CEPH_OSD_OP_TRIMTRUNC: return "trimtrunc"; - - case CEPH_OSD_OP_TMAPUP: return "tmapup"; - case CEPH_OSD_OP_TMAPGET: return "tmapget"; - case CEPH_OSD_OP_TMAPPUT: return "tmapput"; - - case CEPH_OSD_OP_GETXATTR: return "getxattr"; - case CEPH_OSD_OP_GETXATTRS: return "getxattrs"; - case CEPH_OSD_OP_SETXATTR: return "setxattr"; - case CEPH_OSD_OP_SETXATTRS: return "setxattrs"; - case CEPH_OSD_OP_RESETXATTRS: return "resetxattrs"; - case CEPH_OSD_OP_RMXATTR: return "rmxattr"; - case CEPH_OSD_OP_CMPXATTR: return "cmpxattr"; - - case CEPH_OSD_OP_PULL: return "pull"; - case CEPH_OSD_OP_PUSH: return "push"; - case CEPH_OSD_OP_BALANCEREADS: return "balance-reads"; - case CEPH_OSD_OP_UNBALANCEREADS: return "unbalance-reads"; - case CEPH_OSD_OP_SCRUB: return "scrub"; - - case CEPH_OSD_OP_WRLOCK: return "wrlock"; - case CEPH_OSD_OP_WRUNLOCK: return "wrunlock"; - case CEPH_OSD_OP_RDLOCK: return "rdlock"; - case CEPH_OSD_OP_RDUNLOCK: return "rdunlock"; - case CEPH_OSD_OP_UPLOCK: return "uplock"; - case CEPH_OSD_OP_DNLOCK: return "dnlock"; - - case CEPH_OSD_OP_CALL: return "call"; - - case CEPH_OSD_OP_PGLS: return "pgls"; - } - return "???"; -} - -const char *ceph_mds_state_name(int s) -{ - switch (s) { - /* down and out */ - case CEPH_MDS_STATE_DNE: return "down:dne"; - case CEPH_MDS_STATE_STOPPED: return "down:stopped"; - /* up and out */ - case CEPH_MDS_STATE_BOOT: return "up:boot"; - case CEPH_MDS_STATE_STANDBY: return "up:standby"; - case CEPH_MDS_STATE_STANDBY_REPLAY: return "up:standby-replay"; - case CEPH_MDS_STATE_CREATING: return "up:creating"; - case CEPH_MDS_STATE_STARTING: return "up:starting"; - /* up and in */ - case CEPH_MDS_STATE_REPLAY: return "up:replay"; - case CEPH_MDS_STATE_RESOLVE: return "up:resolve"; - case CEPH_MDS_STATE_RECONNECT: return "up:reconnect"; - case CEPH_MDS_STATE_REJOIN: return "up:rejoin"; - case CEPH_MDS_STATE_CLIENTREPLAY: return "up:clientreplay"; - case CEPH_MDS_STATE_ACTIVE: return "up:active"; - case CEPH_MDS_STATE_STOPPING: return "up:stopping"; - } - return "???"; -} - -const char *ceph_session_op_name(int op) -{ - switch (op) { - case CEPH_SESSION_REQUEST_OPEN: return "request_open"; - case CEPH_SESSION_OPEN: return "open"; - case CEPH_SESSION_REQUEST_CLOSE: return "request_close"; - case CEPH_SESSION_CLOSE: return "close"; - case CEPH_SESSION_REQUEST_RENEWCAPS: return "request_renewcaps"; - case CEPH_SESSION_RENEWCAPS: return "renewcaps"; - case CEPH_SESSION_STALE: return "stale"; - case CEPH_SESSION_RECALL_STATE: return "recall_state"; - } - return "???"; -} - -const char *ceph_mds_op_name(int op) -{ - switch (op) { - case CEPH_MDS_OP_LOOKUP: return "lookup"; - case CEPH_MDS_OP_LOOKUPHASH: return "lookuphash"; - case CEPH_MDS_OP_LOOKUPPARENT: return "lookupparent"; - case CEPH_MDS_OP_GETATTR: return "getattr"; - case CEPH_MDS_OP_SETXATTR: return "setxattr"; - case CEPH_MDS_OP_SETATTR: return "setattr"; - case CEPH_MDS_OP_RMXATTR: return "rmxattr"; - case CEPH_MDS_OP_READDIR: return "readdir"; - case CEPH_MDS_OP_MKNOD: return "mknod"; - case CEPH_MDS_OP_LINK: return "link"; - case CEPH_MDS_OP_UNLINK: return "unlink"; - case CEPH_MDS_OP_RENAME: return "rename"; - case CEPH_MDS_OP_MKDIR: return "mkdir"; - case CEPH_MDS_OP_RMDIR: return "rmdir"; - case CEPH_MDS_OP_SYMLINK: return "symlink"; - case CEPH_MDS_OP_CREATE: return "create"; - case CEPH_MDS_OP_OPEN: return "open"; - case CEPH_MDS_OP_LOOKUPSNAP: return "lookupsnap"; - case CEPH_MDS_OP_LSSNAP: return "lssnap"; - case CEPH_MDS_OP_MKSNAP: return "mksnap"; - case CEPH_MDS_OP_RMSNAP: return "rmsnap"; - case CEPH_MDS_OP_SETFILELOCK: return "setfilelock"; - case CEPH_MDS_OP_GETFILELOCK: return "getfilelock"; - } - return "???"; -} - -const char *ceph_cap_op_name(int op) -{ - switch (op) { - case CEPH_CAP_OP_GRANT: return "grant"; - case CEPH_CAP_OP_REVOKE: return "revoke"; - case CEPH_CAP_OP_TRUNC: return "trunc"; - case CEPH_CAP_OP_EXPORT: return "export"; - case CEPH_CAP_OP_IMPORT: return "import"; - case CEPH_CAP_OP_UPDATE: return "update"; - case CEPH_CAP_OP_DROP: return "drop"; - case CEPH_CAP_OP_FLUSH: return "flush"; - case CEPH_CAP_OP_FLUSH_ACK: return "flush_ack"; - case CEPH_CAP_OP_FLUSHSNAP: return "flushsnap"; - case CEPH_CAP_OP_FLUSHSNAP_ACK: return "flushsnap_ack"; - case CEPH_CAP_OP_RELEASE: return "release"; - case CEPH_CAP_OP_RENEW: return "renew"; - } - return "???"; -} - -const char *ceph_lease_op_name(int o) -{ - switch (o) { - case CEPH_MDS_LEASE_REVOKE: return "revoke"; - case CEPH_MDS_LEASE_RELEASE: return "release"; - case CEPH_MDS_LEASE_RENEW: return "renew"; - case CEPH_MDS_LEASE_REVOKE_ACK: return "revoke_ack"; - } - return "???"; -} - -const char *ceph_snap_op_name(int o) -{ - switch (o) { - case CEPH_SNAP_OP_UPDATE: return "update"; - case CEPH_SNAP_OP_CREATE: return "create"; - case CEPH_SNAP_OP_DESTROY: return "destroy"; - case CEPH_SNAP_OP_SPLIT: return "split"; - } - return "???"; -} - -const char *ceph_pool_op_name(int op) -{ - switch (op) { - case POOL_OP_CREATE: return "create"; - case POOL_OP_DELETE: return "delete"; - case POOL_OP_AUID_CHANGE: return "auid change"; - case POOL_OP_CREATE_SNAP: return "create snap"; - case POOL_OP_DELETE_SNAP: return "delete snap"; - case POOL_OP_CREATE_UNMANAGED_SNAP: return "create unmanaged snap"; - case POOL_OP_DELETE_UNMANAGED_SNAP: return "delete unmanaged snap"; - } - return "???"; -} diff --git a/fs/ceph/crush/crush.c b/fs/ceph/crush/crush.c deleted file mode 100644 index fabd302..0000000 --- a/fs/ceph/crush/crush.c +++ /dev/null @@ -1,151 +0,0 @@ - -#ifdef __KERNEL__ -# include -#else -# include -# include -# define kfree(x) do { if (x) free(x); } while (0) -# define BUG_ON(x) assert(!(x)) -#endif - -#include "crush.h" - -const char *crush_bucket_alg_name(int alg) -{ - switch (alg) { - case CRUSH_BUCKET_UNIFORM: return "uniform"; - case CRUSH_BUCKET_LIST: return "list"; - case CRUSH_BUCKET_TREE: return "tree"; - case CRUSH_BUCKET_STRAW: return "straw"; - default: return "unknown"; - } -} - -/** - * crush_get_bucket_item_weight - Get weight of an item in given bucket - * @b: bucket pointer - * @p: item index in bucket - */ -int crush_get_bucket_item_weight(struct crush_bucket *b, int p) -{ - if (p >= b->size) - return 0; - - switch (b->alg) { - case CRUSH_BUCKET_UNIFORM: - return ((struct crush_bucket_uniform *)b)->item_weight; - case CRUSH_BUCKET_LIST: - return ((struct crush_bucket_list *)b)->item_weights[p]; - case CRUSH_BUCKET_TREE: - if (p & 1) - return ((struct crush_bucket_tree *)b)->node_weights[p]; - return 0; - case CRUSH_BUCKET_STRAW: - return ((struct crush_bucket_straw *)b)->item_weights[p]; - } - return 0; -} - -/** - * crush_calc_parents - Calculate parent vectors for the given crush map. - * @map: crush_map pointer - */ -void crush_calc_parents(struct crush_map *map) -{ - int i, b, c; - - for (b = 0; b < map->max_buckets; b++) { - if (map->buckets[b] == NULL) - continue; - for (i = 0; i < map->buckets[b]->size; i++) { - c = map->buckets[b]->items[i]; - BUG_ON(c >= map->max_devices || - c < -map->max_buckets); - if (c >= 0) - map->device_parents[c] = map->buckets[b]->id; - else - map->bucket_parents[-1-c] = map->buckets[b]->id; - } - } -} - -void crush_destroy_bucket_uniform(struct crush_bucket_uniform *b) -{ - kfree(b->h.perm); - kfree(b->h.items); - kfree(b); -} - -void crush_destroy_bucket_list(struct crush_bucket_list *b) -{ - kfree(b->item_weights); - kfree(b->sum_weights); - kfree(b->h.perm); - kfree(b->h.items); - kfree(b); -} - -void crush_destroy_bucket_tree(struct crush_bucket_tree *b) -{ - kfree(b->node_weights); - kfree(b); -} - -void crush_destroy_bucket_straw(struct crush_bucket_straw *b) -{ - kfree(b->straws); - kfree(b->item_weights); - kfree(b->h.perm); - kfree(b->h.items); - kfree(b); -} - -void crush_destroy_bucket(struct crush_bucket *b) -{ - switch (b->alg) { - case CRUSH_BUCKET_UNIFORM: - crush_destroy_bucket_uniform((struct crush_bucket_uniform *)b); - break; - case CRUSH_BUCKET_LIST: - crush_destroy_bucket_list((struct crush_bucket_list *)b); - break; - case CRUSH_BUCKET_TREE: - crush_destroy_bucket_tree((struct crush_bucket_tree *)b); - break; - case CRUSH_BUCKET_STRAW: - crush_destroy_bucket_straw((struct crush_bucket_straw *)b); - break; - } -} - -/** - * crush_destroy - Destroy a crush_map - * @map: crush_map pointer - */ -void crush_destroy(struct crush_map *map) -{ - int b; - - /* buckets */ - if (map->buckets) { - for (b = 0; b < map->max_buckets; b++) { - if (map->buckets[b] == NULL) - continue; - crush_destroy_bucket(map->buckets[b]); - } - kfree(map->buckets); - } - - /* rules */ - if (map->rules) { - for (b = 0; b < map->max_rules; b++) - kfree(map->rules[b]); - kfree(map->rules); - } - - kfree(map->bucket_parents); - kfree(map->device_parents); - kfree(map); -} - - diff --git a/fs/ceph/crush/crush.h b/fs/ceph/crush/crush.h deleted file mode 100644 index 97e435b..0000000 --- a/fs/ceph/crush/crush.h +++ /dev/null @@ -1,180 +0,0 @@ -#ifndef CEPH_CRUSH_CRUSH_H -#define CEPH_CRUSH_CRUSH_H - -#include - -/* - * CRUSH is a pseudo-random data distribution algorithm that - * efficiently distributes input values (typically, data objects) - * across a heterogeneous, structured storage cluster. - * - * The algorithm was originally described in detail in this paper - * (although the algorithm has evolved somewhat since then): - * - * http://www.ssrc.ucsc.edu/Papers/weil-sc06.pdf - * - * LGPL2 - */ - - -#define CRUSH_MAGIC 0x00010000ul /* for detecting algorithm revisions */ - - -#define CRUSH_MAX_DEPTH 10 /* max crush hierarchy depth */ -#define CRUSH_MAX_SET 10 /* max size of a mapping result */ - - -/* - * CRUSH uses user-defined "rules" to describe how inputs should be - * mapped to devices. A rule consists of sequence of steps to perform - * to generate the set of output devices. - */ -struct crush_rule_step { - __u32 op; - __s32 arg1; - __s32 arg2; -}; - -/* step op codes */ -enum { - CRUSH_RULE_NOOP = 0, - CRUSH_RULE_TAKE = 1, /* arg1 = value to start with */ - CRUSH_RULE_CHOOSE_FIRSTN = 2, /* arg1 = num items to pick */ - /* arg2 = type */ - CRUSH_RULE_CHOOSE_INDEP = 3, /* same */ - CRUSH_RULE_EMIT = 4, /* no args */ - CRUSH_RULE_CHOOSE_LEAF_FIRSTN = 6, - CRUSH_RULE_CHOOSE_LEAF_INDEP = 7, -}; - -/* - * for specifying choose num (arg1) relative to the max parameter - * passed to do_rule - */ -#define CRUSH_CHOOSE_N 0 -#define CRUSH_CHOOSE_N_MINUS(x) (-(x)) - -/* - * The rule mask is used to describe what the rule is intended for. - * Given a ruleset and size of output set, we search through the - * rule list for a matching rule_mask. - */ -struct crush_rule_mask { - __u8 ruleset; - __u8 type; - __u8 min_size; - __u8 max_size; -}; - -struct crush_rule { - __u32 len; - struct crush_rule_mask mask; - struct crush_rule_step steps[0]; -}; - -#define crush_rule_size(len) (sizeof(struct crush_rule) + \ - (len)*sizeof(struct crush_rule_step)) - - - -/* - * A bucket is a named container of other items (either devices or - * other buckets). Items within a bucket are chosen using one of a - * few different algorithms. The table summarizes how the speed of - * each option measures up against mapping stability when items are - * added or removed. - * - * Bucket Alg Speed Additions Removals - * ------------------------------------------------ - * uniform O(1) poor poor - * list O(n) optimal poor - * tree O(log n) good good - * straw O(n) optimal optimal - */ -enum { - CRUSH_BUCKET_UNIFORM = 1, - CRUSH_BUCKET_LIST = 2, - CRUSH_BUCKET_TREE = 3, - CRUSH_BUCKET_STRAW = 4 -}; -extern const char *crush_bucket_alg_name(int alg); - -struct crush_bucket { - __s32 id; /* this'll be negative */ - __u16 type; /* non-zero; type=0 is reserved for devices */ - __u8 alg; /* one of CRUSH_BUCKET_* */ - __u8 hash; /* which hash function to use, CRUSH_HASH_* */ - __u32 weight; /* 16-bit fixed point */ - __u32 size; /* num items */ - __s32 *items; - - /* - * cached random permutation: used for uniform bucket and for - * the linear search fallback for the other bucket types. - */ - __u32 perm_x; /* @x for which *perm is defined */ - __u32 perm_n; /* num elements of *perm that are permuted/defined */ - __u32 *perm; -}; - -struct crush_bucket_uniform { - struct crush_bucket h; - __u32 item_weight; /* 16-bit fixed point; all items equally weighted */ -}; - -struct crush_bucket_list { - struct crush_bucket h; - __u32 *item_weights; /* 16-bit fixed point */ - __u32 *sum_weights; /* 16-bit fixed point. element i is sum - of weights 0..i, inclusive */ -}; - -struct crush_bucket_tree { - struct crush_bucket h; /* note: h.size is _tree_ size, not number of - actual items */ - __u8 num_nodes; - __u32 *node_weights; -}; - -struct crush_bucket_straw { - struct crush_bucket h; - __u32 *item_weights; /* 16-bit fixed point */ - __u32 *straws; /* 16-bit fixed point */ -}; - - - -/* - * CRUSH map includes all buckets, rules, etc. - */ -struct crush_map { - struct crush_bucket **buckets; - struct crush_rule **rules; - - /* - * Parent pointers to identify the parent bucket a device or - * bucket in the hierarchy. If an item appears more than - * once, this is the _last_ time it appeared (where buckets - * are processed in bucket id order, from -1 on down to - * -max_buckets. - */ - __u32 *bucket_parents; - __u32 *device_parents; - - __s32 max_buckets; - __u32 max_rules; - __s32 max_devices; -}; - - -/* crush.c */ -extern int crush_get_bucket_item_weight(struct crush_bucket *b, int pos); -extern void crush_calc_parents(struct crush_map *map); -extern void crush_destroy_bucket_uniform(struct crush_bucket_uniform *b); -extern void crush_destroy_bucket_list(struct crush_bucket_list *b); -extern void crush_destroy_bucket_tree(struct crush_bucket_tree *b); -extern void crush_destroy_bucket_straw(struct crush_bucket_straw *b); -extern void crush_destroy_bucket(struct crush_bucket *b); -extern void crush_destroy(struct crush_map *map); - -#endif diff --git a/fs/ceph/crush/hash.c b/fs/ceph/crush/hash.c deleted file mode 100644 index 5873aed..0000000 --- a/fs/ceph/crush/hash.c +++ /dev/null @@ -1,149 +0,0 @@ - -#include -#include "hash.h" - -/* - * Robert Jenkins' function for mixing 32-bit values - * http://burtleburtle.net/bob/hash/evahash.html - * a, b = random bits, c = input and output - */ -#define crush_hashmix(a, b, c) do { \ - a = a-b; a = a-c; a = a^(c>>13); \ - b = b-c; b = b-a; b = b^(a<<8); \ - c = c-a; c = c-b; c = c^(b>>13); \ - a = a-b; a = a-c; a = a^(c>>12); \ - b = b-c; b = b-a; b = b^(a<<16); \ - c = c-a; c = c-b; c = c^(b>>5); \ - a = a-b; a = a-c; a = a^(c>>3); \ - b = b-c; b = b-a; b = b^(a<<10); \ - c = c-a; c = c-b; c = c^(b>>15); \ - } while (0) - -#define crush_hash_seed 1315423911 - -static __u32 crush_hash32_rjenkins1(__u32 a) -{ - __u32 hash = crush_hash_seed ^ a; - __u32 b = a; - __u32 x = 231232; - __u32 y = 1232; - crush_hashmix(b, x, hash); - crush_hashmix(y, a, hash); - return hash; -} - -static __u32 crush_hash32_rjenkins1_2(__u32 a, __u32 b) -{ - __u32 hash = crush_hash_seed ^ a ^ b; - __u32 x = 231232; - __u32 y = 1232; - crush_hashmix(a, b, hash); - crush_hashmix(x, a, hash); - crush_hashmix(b, y, hash); - return hash; -} - -static __u32 crush_hash32_rjenkins1_3(__u32 a, __u32 b, __u32 c) -{ - __u32 hash = crush_hash_seed ^ a ^ b ^ c; - __u32 x = 231232; - __u32 y = 1232; - crush_hashmix(a, b, hash); - crush_hashmix(c, x, hash); - crush_hashmix(y, a, hash); - crush_hashmix(b, x, hash); - crush_hashmix(y, c, hash); - return hash; -} - -static __u32 crush_hash32_rjenkins1_4(__u32 a, __u32 b, __u32 c, __u32 d) -{ - __u32 hash = crush_hash_seed ^ a ^ b ^ c ^ d; - __u32 x = 231232; - __u32 y = 1232; - crush_hashmix(a, b, hash); - crush_hashmix(c, d, hash); - crush_hashmix(a, x, hash); - crush_hashmix(y, b, hash); - crush_hashmix(c, x, hash); - crush_hashmix(y, d, hash); - return hash; -} - -static __u32 crush_hash32_rjenkins1_5(__u32 a, __u32 b, __u32 c, __u32 d, - __u32 e) -{ - __u32 hash = crush_hash_seed ^ a ^ b ^ c ^ d ^ e; - __u32 x = 231232; - __u32 y = 1232; - crush_hashmix(a, b, hash); - crush_hashmix(c, d, hash); - crush_hashmix(e, x, hash); - crush_hashmix(y, a, hash); - crush_hashmix(b, x, hash); - crush_hashmix(y, c, hash); - crush_hashmix(d, x, hash); - crush_hashmix(y, e, hash); - return hash; -} - - -__u32 crush_hash32(int type, __u32 a) -{ - switch (type) { - case CRUSH_HASH_RJENKINS1: - return crush_hash32_rjenkins1(a); - default: - return 0; - } -} - -__u32 crush_hash32_2(int type, __u32 a, __u32 b) -{ - switch (type) { - case CRUSH_HASH_RJENKINS1: - return crush_hash32_rjenkins1_2(a, b); - default: - return 0; - } -} - -__u32 crush_hash32_3(int type, __u32 a, __u32 b, __u32 c) -{ - switch (type) { - case CRUSH_HASH_RJENKINS1: - return crush_hash32_rjenkins1_3(a, b, c); - default: - return 0; - } -} - -__u32 crush_hash32_4(int type, __u32 a, __u32 b, __u32 c, __u32 d) -{ - switch (type) { - case CRUSH_HASH_RJENKINS1: - return crush_hash32_rjenkins1_4(a, b, c, d); - default: - return 0; - } -} - -__u32 crush_hash32_5(int type, __u32 a, __u32 b, __u32 c, __u32 d, __u32 e) -{ - switch (type) { - case CRUSH_HASH_RJENKINS1: - return crush_hash32_rjenkins1_5(a, b, c, d, e); - default: - return 0; - } -} - -const char *crush_hash_name(int type) -{ - switch (type) { - case CRUSH_HASH_RJENKINS1: - return "rjenkins1"; - default: - return "unknown"; - } -} diff --git a/fs/ceph/crush/hash.h b/fs/ceph/crush/hash.h deleted file mode 100644 index 91e8842..0000000 --- a/fs/ceph/crush/hash.h +++ /dev/null @@ -1,17 +0,0 @@ -#ifndef CEPH_CRUSH_HASH_H -#define CEPH_CRUSH_HASH_H - -#define CRUSH_HASH_RJENKINS1 0 - -#define CRUSH_HASH_DEFAULT CRUSH_HASH_RJENKINS1 - -extern const char *crush_hash_name(int type); - -extern __u32 crush_hash32(int type, __u32 a); -extern __u32 crush_hash32_2(int type, __u32 a, __u32 b); -extern __u32 crush_hash32_3(int type, __u32 a, __u32 b, __u32 c); -extern __u32 crush_hash32_4(int type, __u32 a, __u32 b, __u32 c, __u32 d); -extern __u32 crush_hash32_5(int type, __u32 a, __u32 b, __u32 c, __u32 d, - __u32 e); - -#endif diff --git a/fs/ceph/crush/mapper.c b/fs/ceph/crush/mapper.c deleted file mode 100644 index a4eec13..0000000 --- a/fs/ceph/crush/mapper.c +++ /dev/null @@ -1,609 +0,0 @@ - -#ifdef __KERNEL__ -# include -# include -# include -# include -# ifndef dprintk -# define dprintk(args...) -# endif -#else -# include -# include -# include -# include -# define BUG_ON(x) assert(!(x)) -# define dprintk(args...) /* printf(args) */ -# define kmalloc(x, f) malloc(x) -# define kfree(x) free(x) -#endif - -#include "crush.h" -#include "hash.h" - -/* - * Implement the core CRUSH mapping algorithm. - */ - -/** - * crush_find_rule - find a crush_rule id for a given ruleset, type, and size. - * @map: the crush_map - * @ruleset: the storage ruleset id (user defined) - * @type: storage ruleset type (user defined) - * @size: output set size - */ -int crush_find_rule(struct crush_map *map, int ruleset, int type, int size) -{ - int i; - - for (i = 0; i < map->max_rules; i++) { - if (map->rules[i] && - map->rules[i]->mask.ruleset == ruleset && - map->rules[i]->mask.type == type && - map->rules[i]->mask.min_size <= size && - map->rules[i]->mask.max_size >= size) - return i; - } - return -1; -} - - -/* - * bucket choose methods - * - * For each bucket algorithm, we have a "choose" method that, given a - * crush input @x and replica position (usually, position in output set) @r, - * will produce an item in the bucket. - */ - -/* - * Choose based on a random permutation of the bucket. - * - * We used to use some prime number arithmetic to do this, but it - * wasn't very random, and had some other bad behaviors. Instead, we - * calculate an actual random permutation of the bucket members. - * Since this is expensive, we optimize for the r=0 case, which - * captures the vast majority of calls. - */ -static int bucket_perm_choose(struct crush_bucket *bucket, - int x, int r) -{ - unsigned pr = r % bucket->size; - unsigned i, s; - - /* start a new permutation if @x has changed */ - if (bucket->perm_x != x || bucket->perm_n == 0) { - dprintk("bucket %d new x=%d\n", bucket->id, x); - bucket->perm_x = x; - - /* optimize common r=0 case */ - if (pr == 0) { - s = crush_hash32_3(bucket->hash, x, bucket->id, 0) % - bucket->size; - bucket->perm[0] = s; - bucket->perm_n = 0xffff; /* magic value, see below */ - goto out; - } - - for (i = 0; i < bucket->size; i++) - bucket->perm[i] = i; - bucket->perm_n = 0; - } else if (bucket->perm_n == 0xffff) { - /* clean up after the r=0 case above */ - for (i = 1; i < bucket->size; i++) - bucket->perm[i] = i; - bucket->perm[bucket->perm[0]] = 0; - bucket->perm_n = 1; - } - - /* calculate permutation up to pr */ - for (i = 0; i < bucket->perm_n; i++) - dprintk(" perm_choose have %d: %d\n", i, bucket->perm[i]); - while (bucket->perm_n <= pr) { - unsigned p = bucket->perm_n; - /* no point in swapping the final entry */ - if (p < bucket->size - 1) { - i = crush_hash32_3(bucket->hash, x, bucket->id, p) % - (bucket->size - p); - if (i) { - unsigned t = bucket->perm[p + i]; - bucket->perm[p + i] = bucket->perm[p]; - bucket->perm[p] = t; - } - dprintk(" perm_choose swap %d with %d\n", p, p+i); - } - bucket->perm_n++; - } - for (i = 0; i < bucket->size; i++) - dprintk(" perm_choose %d: %d\n", i, bucket->perm[i]); - - s = bucket->perm[pr]; -out: - dprintk(" perm_choose %d sz=%d x=%d r=%d (%d) s=%d\n", bucket->id, - bucket->size, x, r, pr, s); - return bucket->items[s]; -} - -/* uniform */ -static int bucket_uniform_choose(struct crush_bucket_uniform *bucket, - int x, int r) -{ - return bucket_perm_choose(&bucket->h, x, r); -} - -/* list */ -static int bucket_list_choose(struct crush_bucket_list *bucket, - int x, int r) -{ - int i; - - for (i = bucket->h.size-1; i >= 0; i--) { - __u64 w = crush_hash32_4(bucket->h.hash,x, bucket->h.items[i], - r, bucket->h.id); - w &= 0xffff; - dprintk("list_choose i=%d x=%d r=%d item %d weight %x " - "sw %x rand %llx", - i, x, r, bucket->h.items[i], bucket->item_weights[i], - bucket->sum_weights[i], w); - w *= bucket->sum_weights[i]; - w = w >> 16; - /*dprintk(" scaled %llx\n", w);*/ - if (w < bucket->item_weights[i]) - return bucket->h.items[i]; - } - - BUG_ON(1); - return 0; -} - - -/* (binary) tree */ -static int height(int n) -{ - int h = 0; - while ((n & 1) == 0) { - h++; - n = n >> 1; - } - return h; -} - -static int left(int x) -{ - int h = height(x); - return x - (1 << (h-1)); -} - -static int right(int x) -{ - int h = height(x); - return x + (1 << (h-1)); -} - -static int terminal(int x) -{ - return x & 1; -} - -static int bucket_tree_choose(struct crush_bucket_tree *bucket, - int x, int r) -{ - int n, l; - __u32 w; - __u64 t; - - /* start at root */ - n = bucket->num_nodes >> 1; - - while (!terminal(n)) { - /* pick point in [0, w) */ - w = bucket->node_weights[n]; - t = (__u64)crush_hash32_4(bucket->h.hash, x, n, r, - bucket->h.id) * (__u64)w; - t = t >> 32; - - /* descend to the left or right? */ - l = left(n); - if (t < bucket->node_weights[l]) - n = l; - else - n = right(n); - } - - return bucket->h.items[n >> 1]; -} - - -/* straw */ - -static int bucket_straw_choose(struct crush_bucket_straw *bucket, - int x, int r) -{ - int i; - int high = 0; - __u64 high_draw = 0; - __u64 draw; - - for (i = 0; i < bucket->h.size; i++) { - draw = crush_hash32_3(bucket->h.hash, x, bucket->h.items[i], r); - draw &= 0xffff; - draw *= bucket->straws[i]; - if (i == 0 || draw > high_draw) { - high = i; - high_draw = draw; - } - } - return bucket->h.items[high]; -} - -static int crush_bucket_choose(struct crush_bucket *in, int x, int r) -{ - dprintk(" crush_bucket_choose %d x=%d r=%d\n", in->id, x, r); - switch (in->alg) { - case CRUSH_BUCKET_UNIFORM: - return bucket_uniform_choose((struct crush_bucket_uniform *)in, - x, r); - case CRUSH_BUCKET_LIST: - return bucket_list_choose((struct crush_bucket_list *)in, - x, r); - case CRUSH_BUCKET_TREE: - return bucket_tree_choose((struct crush_bucket_tree *)in, - x, r); - case CRUSH_BUCKET_STRAW: - return bucket_straw_choose((struct crush_bucket_straw *)in, - x, r); - default: - BUG_ON(1); - return in->items[0]; - } -} - -/* - * true if device is marked "out" (failed, fully offloaded) - * of the cluster - */ -static int is_out(struct crush_map *map, __u32 *weight, int item, int x) -{ - if (weight[item] >= 0x10000) - return 0; - if (weight[item] == 0) - return 1; - if ((crush_hash32_2(CRUSH_HASH_RJENKINS1, x, item) & 0xffff) - < weight[item]) - return 0; - return 1; -} - -/** - * crush_choose - choose numrep distinct items of given type - * @map: the crush_map - * @bucket: the bucket we are choose an item from - * @x: crush input value - * @numrep: the number of items to choose - * @type: the type of item to choose - * @out: pointer to output vector - * @outpos: our position in that vector - * @firstn: true if choosing "first n" items, false if choosing "indep" - * @recurse_to_leaf: true if we want one device under each item of given type - * @out2: second output vector for leaf items (if @recurse_to_leaf) - */ -static int crush_choose(struct crush_map *map, - struct crush_bucket *bucket, - __u32 *weight, - int x, int numrep, int type, - int *out, int outpos, - int firstn, int recurse_to_leaf, - int *out2) -{ - int rep; - int ftotal, flocal; - int retry_descent, retry_bucket, skip_rep; - struct crush_bucket *in = bucket; - int r; - int i; - int item = 0; - int itemtype; - int collide, reject; - const int orig_tries = 5; /* attempts before we fall back to search */ - - dprintk("CHOOSE%s bucket %d x %d outpos %d numrep %d\n", recurse_to_leaf ? "_LEAF" : "", - bucket->id, x, outpos, numrep); - - for (rep = outpos; rep < numrep; rep++) { - /* keep trying until we get a non-out, non-colliding item */ - ftotal = 0; - skip_rep = 0; - do { - retry_descent = 0; - in = bucket; /* initial bucket */ - - /* choose through intervening buckets */ - flocal = 0; - do { - collide = 0; - retry_bucket = 0; - r = rep; - if (in->alg == CRUSH_BUCKET_UNIFORM) { - /* be careful */ - if (firstn || numrep >= in->size) - /* r' = r + f_total */ - r += ftotal; - else if (in->size % numrep == 0) - /* r'=r+(n+1)*f_local */ - r += (numrep+1) * - (flocal+ftotal); - else - /* r' = r + n*f_local */ - r += numrep * (flocal+ftotal); - } else { - if (firstn) - /* r' = r + f_total */ - r += ftotal; - else - /* r' = r + n*f_local */ - r += numrep * (flocal+ftotal); - } - - /* bucket choose */ - if (in->size == 0) { - reject = 1; - goto reject; - } - if (flocal >= (in->size>>1) && - flocal > orig_tries) - item = bucket_perm_choose(in, x, r); - else - item = crush_bucket_choose(in, x, r); - BUG_ON(item >= map->max_devices); - - /* desired type? */ - if (item < 0) - itemtype = map->buckets[-1-item]->type; - else - itemtype = 0; - dprintk(" item %d type %d\n", item, itemtype); - - /* keep going? */ - if (itemtype != type) { - BUG_ON(item >= 0 || - (-1-item) >= map->max_buckets); - in = map->buckets[-1-item]; - retry_bucket = 1; - continue; - } - - /* collision? */ - for (i = 0; i < outpos; i++) { - if (out[i] == item) { - collide = 1; - break; - } - } - - reject = 0; - if (recurse_to_leaf) { - if (item < 0) { - if (crush_choose(map, - map->buckets[-1-item], - weight, - x, outpos+1, 0, - out2, outpos, - firstn, 0, - NULL) <= outpos) - /* didn't get leaf */ - reject = 1; - } else { - /* we already have a leaf! */ - out2[outpos] = item; - } - } - - if (!reject) { - /* out? */ - if (itemtype == 0) - reject = is_out(map, weight, - item, x); - else - reject = 0; - } - -reject: - if (reject || collide) { - ftotal++; - flocal++; - - if (collide && flocal < 3) - /* retry locally a few times */ - retry_bucket = 1; - else if (flocal < in->size + orig_tries) - /* exhaustive bucket search */ - retry_bucket = 1; - else if (ftotal < 20) - /* then retry descent */ - retry_descent = 1; - else - /* else give up */ - skip_rep = 1; - dprintk(" reject %d collide %d " - "ftotal %d flocal %d\n", - reject, collide, ftotal, - flocal); - } - } while (retry_bucket); - } while (retry_descent); - - if (skip_rep) { - dprintk("skip rep\n"); - continue; - } - - dprintk("CHOOSE got %d\n", item); - out[outpos] = item; - outpos++; - } - - dprintk("CHOOSE returns %d\n", outpos); - return outpos; -} - - -/** - * crush_do_rule - calculate a mapping with the given input and rule - * @map: the crush_map - * @ruleno: the rule id - * @x: hash input - * @result: pointer to result vector - * @result_max: maximum result size - * @force: force initial replica choice; -1 for none - */ -int crush_do_rule(struct crush_map *map, - int ruleno, int x, int *result, int result_max, - int force, __u32 *weight) -{ - int result_len; - int force_context[CRUSH_MAX_DEPTH]; - int force_pos = -1; - int a[CRUSH_MAX_SET]; - int b[CRUSH_MAX_SET]; - int c[CRUSH_MAX_SET]; - int recurse_to_leaf; - int *w; - int wsize = 0; - int *o; - int osize; - int *tmp; - struct crush_rule *rule; - int step; - int i, j; - int numrep; - int firstn; - int rc = -1; - - BUG_ON(ruleno >= map->max_rules); - - rule = map->rules[ruleno]; - result_len = 0; - w = a; - o = b; - - /* - * determine hierarchical context of force, if any. note - * that this may or may not correspond to the specific types - * referenced by the crush rule. - */ - if (force >= 0) { - if (force >= map->max_devices || - map->device_parents[force] == 0) { - /*dprintk("CRUSH: forcefed device dne\n");*/ - rc = -1; /* force fed device dne */ - goto out; - } - if (!is_out(map, weight, force, x)) { - while (1) { - force_context[++force_pos] = force; - if (force >= 0) - force = map->device_parents[force]; - else - force = map->bucket_parents[-1-force]; - if (force == 0) - break; - } - } - } - - for (step = 0; step < rule->len; step++) { - firstn = 0; - switch (rule->steps[step].op) { - case CRUSH_RULE_TAKE: - w[0] = rule->steps[step].arg1; - if (force_pos >= 0) { - BUG_ON(force_context[force_pos] != w[0]); - force_pos--; - } - wsize = 1; - break; - - case CRUSH_RULE_CHOOSE_LEAF_FIRSTN: - case CRUSH_RULE_CHOOSE_FIRSTN: - firstn = 1; - case CRUSH_RULE_CHOOSE_LEAF_INDEP: - case CRUSH_RULE_CHOOSE_INDEP: - BUG_ON(wsize == 0); - - recurse_to_leaf = - rule->steps[step].op == - CRUSH_RULE_CHOOSE_LEAF_FIRSTN || - rule->steps[step].op == - CRUSH_RULE_CHOOSE_LEAF_INDEP; - - /* reset output */ - osize = 0; - - for (i = 0; i < wsize; i++) { - /* - * see CRUSH_N, CRUSH_N_MINUS macros. - * basically, numrep <= 0 means relative to - * the provided result_max - */ - numrep = rule->steps[step].arg1; - if (numrep <= 0) { - numrep += result_max; - if (numrep <= 0) - continue; - } - j = 0; - if (osize == 0 && force_pos >= 0) { - /* skip any intermediate types */ - while (force_pos && - force_context[force_pos] < 0 && - rule->steps[step].arg2 != - map->buckets[-1 - - force_context[force_pos]]->type) - force_pos--; - o[osize] = force_context[force_pos]; - if (recurse_to_leaf) - c[osize] = force_context[0]; - j++; - force_pos--; - } - osize += crush_choose(map, - map->buckets[-1-w[i]], - weight, - x, numrep, - rule->steps[step].arg2, - o+osize, j, - firstn, - recurse_to_leaf, c+osize); - } - - if (recurse_to_leaf) - /* copy final _leaf_ values to output set */ - memcpy(o, c, osize*sizeof(*o)); - - /* swap t and w arrays */ - tmp = o; - o = w; - w = tmp; - wsize = osize; - break; - - - case CRUSH_RULE_EMIT: - for (i = 0; i < wsize && result_len < result_max; i++) { - result[result_len] = w[i]; - result_len++; - } - wsize = 0; - break; - - default: - BUG_ON(1); - } - } - rc = result_len; - -out: - return rc; -} - - diff --git a/fs/ceph/crush/mapper.h b/fs/ceph/crush/mapper.h deleted file mode 100644 index c46b99c..0000000 --- a/fs/ceph/crush/mapper.h +++ /dev/null @@ -1,20 +0,0 @@ -#ifndef CEPH_CRUSH_MAPPER_H -#define CEPH_CRUSH_MAPPER_H - -/* - * CRUSH functions for find rules and then mapping an input to an - * output set. - * - * LGPL2 - */ - -#include "crush.h" - -extern int crush_find_rule(struct crush_map *map, int pool, int type, int size); -extern int crush_do_rule(struct crush_map *map, - int ruleno, - int x, int *result, int result_max, - int forcefeed, /* -1 for none */ - __u32 *weights); - -#endif diff --git a/fs/ceph/crypto.c b/fs/ceph/crypto.c deleted file mode 100644 index a3e627f..0000000 --- a/fs/ceph/crypto.c +++ /dev/null @@ -1,412 +0,0 @@ - -#include "ceph_debug.h" - -#include -#include -#include -#include - -#include "crypto.h" -#include "decode.h" - -int ceph_crypto_key_encode(struct ceph_crypto_key *key, void **p, void *end) -{ - if (*p + sizeof(u16) + sizeof(key->created) + - sizeof(u16) + key->len > end) - return -ERANGE; - ceph_encode_16(p, key->type); - ceph_encode_copy(p, &key->created, sizeof(key->created)); - ceph_encode_16(p, key->len); - ceph_encode_copy(p, key->key, key->len); - return 0; -} - -int ceph_crypto_key_decode(struct ceph_crypto_key *key, void **p, void *end) -{ - ceph_decode_need(p, end, 2*sizeof(u16) + sizeof(key->created), bad); - key->type = ceph_decode_16(p); - ceph_decode_copy(p, &key->created, sizeof(key->created)); - key->len = ceph_decode_16(p); - ceph_decode_need(p, end, key->len, bad); - key->key = kmalloc(key->len, GFP_NOFS); - if (!key->key) - return -ENOMEM; - ceph_decode_copy(p, key->key, key->len); - return 0; - -bad: - dout("failed to decode crypto key\n"); - return -EINVAL; -} - -int ceph_crypto_key_unarmor(struct ceph_crypto_key *key, const char *inkey) -{ - int inlen = strlen(inkey); - int blen = inlen * 3 / 4; - void *buf, *p; - int ret; - - dout("crypto_key_unarmor %s\n", inkey); - buf = kmalloc(blen, GFP_NOFS); - if (!buf) - return -ENOMEM; - blen = ceph_unarmor(buf, inkey, inkey+inlen); - if (blen < 0) { - kfree(buf); - return blen; - } - - p = buf; - ret = ceph_crypto_key_decode(key, &p, p + blen); - kfree(buf); - if (ret) - return ret; - dout("crypto_key_unarmor key %p type %d len %d\n", key, - key->type, key->len); - return 0; -} - - - -#define AES_KEY_SIZE 16 - -static struct crypto_blkcipher *ceph_crypto_alloc_cipher(void) -{ - return crypto_alloc_blkcipher("cbc(aes)", 0, CRYPTO_ALG_ASYNC); -} - -static const u8 *aes_iv = (u8 *)CEPH_AES_IV; - -static int ceph_aes_encrypt(const void *key, int key_len, - void *dst, size_t *dst_len, - const void *src, size_t src_len) -{ - struct scatterlist sg_in[2], sg_out[1]; - struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); - struct blkcipher_desc desc = { .tfm = tfm, .flags = 0 }; - int ret; - void *iv; - int ivsize; - size_t zero_padding = (0x10 - (src_len & 0x0f)); - char pad[16]; - - if (IS_ERR(tfm)) - return PTR_ERR(tfm); - - memset(pad, zero_padding, zero_padding); - - *dst_len = src_len + zero_padding; - - crypto_blkcipher_setkey((void *)tfm, key, key_len); - sg_init_table(sg_in, 2); - sg_set_buf(&sg_in[0], src, src_len); - sg_set_buf(&sg_in[1], pad, zero_padding); - sg_init_table(sg_out, 1); - sg_set_buf(sg_out, dst, *dst_len); - iv = crypto_blkcipher_crt(tfm)->iv; - ivsize = crypto_blkcipher_ivsize(tfm); - - memcpy(iv, aes_iv, ivsize); - /* - print_hex_dump(KERN_ERR, "enc key: ", DUMP_PREFIX_NONE, 16, 1, - key, key_len, 1); - print_hex_dump(KERN_ERR, "enc src: ", DUMP_PREFIX_NONE, 16, 1, - src, src_len, 1); - print_hex_dump(KERN_ERR, "enc pad: ", DUMP_PREFIX_NONE, 16, 1, - pad, zero_padding, 1); - */ - ret = crypto_blkcipher_encrypt(&desc, sg_out, sg_in, - src_len + zero_padding); - crypto_free_blkcipher(tfm); - if (ret < 0) - pr_err("ceph_aes_crypt failed %d\n", ret); - /* - print_hex_dump(KERN_ERR, "enc out: ", DUMP_PREFIX_NONE, 16, 1, - dst, *dst_len, 1); - */ - return 0; -} - -static int ceph_aes_encrypt2(const void *key, int key_len, void *dst, - size_t *dst_len, - const void *src1, size_t src1_len, - const void *src2, size_t src2_len) -{ - struct scatterlist sg_in[3], sg_out[1]; - struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); - struct blkcipher_desc desc = { .tfm = tfm, .flags = 0 }; - int ret; - void *iv; - int ivsize; - size_t zero_padding = (0x10 - ((src1_len + src2_len) & 0x0f)); - char pad[16]; - - if (IS_ERR(tfm)) - return PTR_ERR(tfm); - - memset(pad, zero_padding, zero_padding); - - *dst_len = src1_len + src2_len + zero_padding; - - crypto_blkcipher_setkey((void *)tfm, key, key_len); - sg_init_table(sg_in, 3); - sg_set_buf(&sg_in[0], src1, src1_len); - sg_set_buf(&sg_in[1], src2, src2_len); - sg_set_buf(&sg_in[2], pad, zero_padding); - sg_init_table(sg_out, 1); - sg_set_buf(sg_out, dst, *dst_len); - iv = crypto_blkcipher_crt(tfm)->iv; - ivsize = crypto_blkcipher_ivsize(tfm); - - memcpy(iv, aes_iv, ivsize); - /* - print_hex_dump(KERN_ERR, "enc key: ", DUMP_PREFIX_NONE, 16, 1, - key, key_len, 1); - print_hex_dump(KERN_ERR, "enc src1: ", DUMP_PREFIX_NONE, 16, 1, - src1, src1_len, 1); - print_hex_dump(KERN_ERR, "enc src2: ", DUMP_PREFIX_NONE, 16, 1, - src2, src2_len, 1); - print_hex_dump(KERN_ERR, "enc pad: ", DUMP_PREFIX_NONE, 16, 1, - pad, zero_padding, 1); - */ - ret = crypto_blkcipher_encrypt(&desc, sg_out, sg_in, - src1_len + src2_len + zero_padding); - crypto_free_blkcipher(tfm); - if (ret < 0) - pr_err("ceph_aes_crypt2 failed %d\n", ret); - /* - print_hex_dump(KERN_ERR, "enc out: ", DUMP_PREFIX_NONE, 16, 1, - dst, *dst_len, 1); - */ - return 0; -} - -static int ceph_aes_decrypt(const void *key, int key_len, - void *dst, size_t *dst_len, - const void *src, size_t src_len) -{ - struct scatterlist sg_in[1], sg_out[2]; - struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); - struct blkcipher_desc desc = { .tfm = tfm }; - char pad[16]; - void *iv; - int ivsize; - int ret; - int last_byte; - - if (IS_ERR(tfm)) - return PTR_ERR(tfm); - - crypto_blkcipher_setkey((void *)tfm, key, key_len); - sg_init_table(sg_in, 1); - sg_init_table(sg_out, 2); - sg_set_buf(sg_in, src, src_len); - sg_set_buf(&sg_out[0], dst, *dst_len); - sg_set_buf(&sg_out[1], pad, sizeof(pad)); - - iv = crypto_blkcipher_crt(tfm)->iv; - ivsize = crypto_blkcipher_ivsize(tfm); - - memcpy(iv, aes_iv, ivsize); - - /* - print_hex_dump(KERN_ERR, "dec key: ", DUMP_PREFIX_NONE, 16, 1, - key, key_len, 1); - print_hex_dump(KERN_ERR, "dec in: ", DUMP_PREFIX_NONE, 16, 1, - src, src_len, 1); - */ - - ret = crypto_blkcipher_decrypt(&desc, sg_out, sg_in, src_len); - crypto_free_blkcipher(tfm); - if (ret < 0) { - pr_err("ceph_aes_decrypt failed %d\n", ret); - return ret; - } - - if (src_len <= *dst_len) - last_byte = ((char *)dst)[src_len - 1]; - else - last_byte = pad[src_len - *dst_len - 1]; - if (last_byte <= 16 && src_len >= last_byte) { - *dst_len = src_len - last_byte; - } else { - pr_err("ceph_aes_decrypt got bad padding %d on src len %d\n", - last_byte, (int)src_len); - return -EPERM; /* bad padding */ - } - /* - print_hex_dump(KERN_ERR, "dec out: ", DUMP_PREFIX_NONE, 16, 1, - dst, *dst_len, 1); - */ - return 0; -} - -static int ceph_aes_decrypt2(const void *key, int key_len, - void *dst1, size_t *dst1_len, - void *dst2, size_t *dst2_len, - const void *src, size_t src_len) -{ - struct scatterlist sg_in[1], sg_out[3]; - struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); - struct blkcipher_desc desc = { .tfm = tfm }; - char pad[16]; - void *iv; - int ivsize; - int ret; - int last_byte; - - if (IS_ERR(tfm)) - return PTR_ERR(tfm); - - sg_init_table(sg_in, 1); - sg_set_buf(sg_in, src, src_len); - sg_init_table(sg_out, 3); - sg_set_buf(&sg_out[0], dst1, *dst1_len); - sg_set_buf(&sg_out[1], dst2, *dst2_len); - sg_set_buf(&sg_out[2], pad, sizeof(pad)); - - crypto_blkcipher_setkey((void *)tfm, key, key_len); - iv = crypto_blkcipher_crt(tfm)->iv; - ivsize = crypto_blkcipher_ivsize(tfm); - - memcpy(iv, aes_iv, ivsize); - - /* - print_hex_dump(KERN_ERR, "dec key: ", DUMP_PREFIX_NONE, 16, 1, - key, key_len, 1); - print_hex_dump(KERN_ERR, "dec in: ", DUMP_PREFIX_NONE, 16, 1, - src, src_len, 1); - */ - - ret = crypto_blkcipher_decrypt(&desc, sg_out, sg_in, src_len); - crypto_free_blkcipher(tfm); - if (ret < 0) { - pr_err("ceph_aes_decrypt failed %d\n", ret); - return ret; - } - - if (src_len <= *dst1_len) - last_byte = ((char *)dst1)[src_len - 1]; - else if (src_len <= *dst1_len + *dst2_len) - last_byte = ((char *)dst2)[src_len - *dst1_len - 1]; - else - last_byte = pad[src_len - *dst1_len - *dst2_len - 1]; - if (last_byte <= 16 && src_len >= last_byte) { - src_len -= last_byte; - } else { - pr_err("ceph_aes_decrypt got bad padding %d on src len %d\n", - last_byte, (int)src_len); - return -EPERM; /* bad padding */ - } - - if (src_len < *dst1_len) { - *dst1_len = src_len; - *dst2_len = 0; - } else { - *dst2_len = src_len - *dst1_len; - } - /* - print_hex_dump(KERN_ERR, "dec out1: ", DUMP_PREFIX_NONE, 16, 1, - dst1, *dst1_len, 1); - print_hex_dump(KERN_ERR, "dec out2: ", DUMP_PREFIX_NONE, 16, 1, - dst2, *dst2_len, 1); - */ - - return 0; -} - - -int ceph_decrypt(struct ceph_crypto_key *secret, void *dst, size_t *dst_len, - const void *src, size_t src_len) -{ - switch (secret->type) { - case CEPH_CRYPTO_NONE: - if (*dst_len < src_len) - return -ERANGE; - memcpy(dst, src, src_len); - *dst_len = src_len; - return 0; - - case CEPH_CRYPTO_AES: - return ceph_aes_decrypt(secret->key, secret->len, dst, - dst_len, src, src_len); - - default: - return -EINVAL; - } -} - -int ceph_decrypt2(struct ceph_crypto_key *secret, - void *dst1, size_t *dst1_len, - void *dst2, size_t *dst2_len, - const void *src, size_t src_len) -{ - size_t t; - - switch (secret->type) { - case CEPH_CRYPTO_NONE: - if (*dst1_len + *dst2_len < src_len) - return -ERANGE; - t = min(*dst1_len, src_len); - memcpy(dst1, src, t); - *dst1_len = t; - src += t; - src_len -= t; - if (src_len) { - t = min(*dst2_len, src_len); - memcpy(dst2, src, t); - *dst2_len = t; - } - return 0; - - case CEPH_CRYPTO_AES: - return ceph_aes_decrypt2(secret->key, secret->len, - dst1, dst1_len, dst2, dst2_len, - src, src_len); - - default: - return -EINVAL; - } -} - -int ceph_encrypt(struct ceph_crypto_key *secret, void *dst, size_t *dst_len, - const void *src, size_t src_len) -{ - switch (secret->type) { - case CEPH_CRYPTO_NONE: - if (*dst_len < src_len) - return -ERANGE; - memcpy(dst, src, src_len); - *dst_len = src_len; - return 0; - - case CEPH_CRYPTO_AES: - return ceph_aes_encrypt(secret->key, secret->len, dst, - dst_len, src, src_len); - - default: - return -EINVAL; - } -} - -int ceph_encrypt2(struct ceph_crypto_key *secret, void *dst, size_t *dst_len, - const void *src1, size_t src1_len, - const void *src2, size_t src2_len) -{ - switch (secret->type) { - case CEPH_CRYPTO_NONE: - if (*dst_len < src1_len + src2_len) - return -ERANGE; - memcpy(dst, src1, src1_len); - memcpy(dst + src1_len, src2, src2_len); - *dst_len = src1_len + src2_len; - return 0; - - case CEPH_CRYPTO_AES: - return ceph_aes_encrypt2(secret->key, secret->len, dst, dst_len, - src1, src1_len, src2, src2_len); - - default: - return -EINVAL; - } -} diff --git a/fs/ceph/crypto.h b/fs/ceph/crypto.h deleted file mode 100644 index bdf3860..0000000 --- a/fs/ceph/crypto.h +++ /dev/null @@ -1,48 +0,0 @@ -#ifndef _FS_CEPH_CRYPTO_H -#define _FS_CEPH_CRYPTO_H - -#include "types.h" -#include "buffer.h" - -/* - * cryptographic secret - */ -struct ceph_crypto_key { - int type; - struct ceph_timespec created; - int len; - void *key; -}; - -static inline void ceph_crypto_key_destroy(struct ceph_crypto_key *key) -{ - kfree(key->key); -} - -extern int ceph_crypto_key_encode(struct ceph_crypto_key *key, - void **p, void *end); -extern int ceph_crypto_key_decode(struct ceph_crypto_key *key, - void **p, void *end); -extern int ceph_crypto_key_unarmor(struct ceph_crypto_key *key, const char *in); - -/* crypto.c */ -extern int ceph_decrypt(struct ceph_crypto_key *secret, - void *dst, size_t *dst_len, - const void *src, size_t src_len); -extern int ceph_encrypt(struct ceph_crypto_key *secret, - void *dst, size_t *dst_len, - const void *src, size_t src_len); -extern int ceph_decrypt2(struct ceph_crypto_key *secret, - void *dst1, size_t *dst1_len, - void *dst2, size_t *dst2_len, - const void *src, size_t src_len); -extern int ceph_encrypt2(struct ceph_crypto_key *secret, - void *dst, size_t *dst_len, - const void *src1, size_t src1_len, - const void *src2, size_t src2_len); - -/* armor.c */ -extern int ceph_armor(char *dst, const char *src, const char *end); -extern int ceph_unarmor(char *dst, const char *src, const char *end); - -#endif diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c index 6fd8b20..94fe1d2 100644 --- a/fs/ceph/debugfs.c +++ b/fs/ceph/debugfs.c @@ -1,4 +1,4 @@ -#include "ceph_debug.h" +#include #include #include @@ -7,143 +7,48 @@ #include #include -#include "super.h" -#include "mds_client.h" -#include "mon_client.h" -#include "auth.h" +#include +#include +#include +#include #ifdef CONFIG_DEBUG_FS -/* - * Implement /sys/kernel/debug/ceph fun - * - * /sys/kernel/debug/ceph/client* - an instance of the ceph client - * .../osdmap - current osdmap - * .../mdsmap - current mdsmap - * .../monmap - current monmap - * .../osdc - active osd requests - * .../mdsc - active mds requests - * .../monc - mon client state - * .../dentry_lru - dump contents of dentry lru - * .../caps - expose cap (reservation) stats - * .../bdi - symlink to ../../bdi/something - */ - -static struct dentry *ceph_debugfs_dir; - -static int monmap_show(struct seq_file *s, void *p) -{ - int i; - struct ceph_client *client = s->private; - - if (client->monc.monmap == NULL) - return 0; - - seq_printf(s, "epoch %d\n", client->monc.monmap->epoch); - for (i = 0; i < client->monc.monmap->num_mon; i++) { - struct ceph_entity_inst *inst = - &client->monc.monmap->mon_inst[i]; - - seq_printf(s, "\t%s%lld\t%s\n", - ENTITY_NAME(inst->name), - pr_addr(&inst->addr.in_addr)); - } - return 0; -} +#include "super.h" +#include "mds_client.h" static int mdsmap_show(struct seq_file *s, void *p) { int i; - struct ceph_client *client = s->private; + struct ceph_fs_client *fsc = s->private; - if (client->mdsc.mdsmap == NULL) + if (fsc->mdsc == NULL || fsc->mdsc->mdsmap == NULL) return 0; - seq_printf(s, "epoch %d\n", client->mdsc.mdsmap->m_epoch); - seq_printf(s, "root %d\n", client->mdsc.mdsmap->m_root); + seq_printf(s, "epoch %d\n", fsc->mdsc->mdsmap->m_epoch); + seq_printf(s, "root %d\n", fsc->mdsc->mdsmap->m_root); seq_printf(s, "session_timeout %d\n", - client->mdsc.mdsmap->m_session_timeout); + fsc->mdsc->mdsmap->m_session_timeout); seq_printf(s, "session_autoclose %d\n", - client->mdsc.mdsmap->m_session_autoclose); - for (i = 0; i < client->mdsc.mdsmap->m_max_mds; i++) { + fsc->mdsc->mdsmap->m_session_autoclose); + for (i = 0; i < fsc->mdsc->mdsmap->m_max_mds; i++) { struct ceph_entity_addr *addr = - &client->mdsc.mdsmap->m_info[i].addr; - int state = client->mdsc.mdsmap->m_info[i].state; + &fsc->mdsc->mdsmap->m_info[i].addr; + int state = fsc->mdsc->mdsmap->m_info[i].state; - seq_printf(s, "\tmds%d\t%s\t(%s)\n", i, pr_addr(&addr->in_addr), + seq_printf(s, "\tmds%d\t%s\t(%s)\n", i, + ceph_pr_addr(&addr->in_addr), ceph_mds_state_name(state)); } return 0; } -static int osdmap_show(struct seq_file *s, void *p) -{ - int i; - struct ceph_client *client = s->private; - struct rb_node *n; - - if (client->osdc.osdmap == NULL) - return 0; - seq_printf(s, "epoch %d\n", client->osdc.osdmap->epoch); - seq_printf(s, "flags%s%s\n", - (client->osdc.osdmap->flags & CEPH_OSDMAP_NEARFULL) ? - " NEARFULL" : "", - (client->osdc.osdmap->flags & CEPH_OSDMAP_FULL) ? - " FULL" : ""); - for (n = rb_first(&client->osdc.osdmap->pg_pools); n; n = rb_next(n)) { - struct ceph_pg_pool_info *pool = - rb_entry(n, struct ceph_pg_pool_info, node); - seq_printf(s, "pg_pool %d pg_num %d / %d, lpg_num %d / %d\n", - pool->id, pool->v.pg_num, pool->pg_num_mask, - pool->v.lpg_num, pool->lpg_num_mask); - } - for (i = 0; i < client->osdc.osdmap->max_osd; i++) { - struct ceph_entity_addr *addr = - &client->osdc.osdmap->osd_addr[i]; - int state = client->osdc.osdmap->osd_state[i]; - char sb[64]; - - seq_printf(s, "\tosd%d\t%s\t%3d%%\t(%s)\n", - i, pr_addr(&addr->in_addr), - ((client->osdc.osdmap->osd_weight[i]*100) >> 16), - ceph_osdmap_state_str(sb, sizeof(sb), state)); - } - return 0; -} - -static int monc_show(struct seq_file *s, void *p) -{ - struct ceph_client *client = s->private; - struct ceph_mon_generic_request *req; - struct ceph_mon_client *monc = &client->monc; - struct rb_node *rp; - - mutex_lock(&monc->mutex); - - if (monc->have_mdsmap) - seq_printf(s, "have mdsmap %u\n", (unsigned)monc->have_mdsmap); - if (monc->have_osdmap) - seq_printf(s, "have osdmap %u\n", (unsigned)monc->have_osdmap); - if (monc->want_next_osdmap) - seq_printf(s, "want next osdmap\n"); - - for (rp = rb_first(&monc->generic_request_tree); rp; rp = rb_next(rp)) { - __u16 op; - req = rb_entry(rp, struct ceph_mon_generic_request, node); - op = le16_to_cpu(req->request->hdr.type); - if (op == CEPH_MSG_STATFS) - seq_printf(s, "%lld statfs\n", req->tid); - else - seq_printf(s, "%lld unknown\n", req->tid); - } - - mutex_unlock(&monc->mutex); - return 0; -} - +/* + * mdsc debugfs + */ static int mdsc_show(struct seq_file *s, void *p) { - struct ceph_client *client = s->private; - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = s->private; + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; struct rb_node *rp; int pathlen; @@ -214,61 +119,12 @@ static int mdsc_show(struct seq_file *s, void *p) return 0; } -static int osdc_show(struct seq_file *s, void *pp) -{ - struct ceph_client *client = s->private; - struct ceph_osd_client *osdc = &client->osdc; - struct rb_node *p; - - mutex_lock(&osdc->request_mutex); - for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { - struct ceph_osd_request *req; - struct ceph_osd_request_head *head; - struct ceph_osd_op *op; - int num_ops; - int opcode, olen; - int i; - - req = rb_entry(p, struct ceph_osd_request, r_node); - - seq_printf(s, "%lld\tosd%d\t%d.%x\t", req->r_tid, - req->r_osd ? req->r_osd->o_osd : -1, - le32_to_cpu(req->r_pgid.pool), - le16_to_cpu(req->r_pgid.ps)); - - head = req->r_request->front.iov_base; - op = (void *)(head + 1); - - num_ops = le16_to_cpu(head->num_ops); - olen = le32_to_cpu(head->object_len); - seq_printf(s, "%.*s", olen, - (const char *)(head->ops + num_ops)); - - if (req->r_reassert_version.epoch) - seq_printf(s, "\t%u'%llu", - (unsigned)le32_to_cpu(req->r_reassert_version.epoch), - le64_to_cpu(req->r_reassert_version.version)); - else - seq_printf(s, "\t"); - - for (i = 0; i < num_ops; i++) { - opcode = le16_to_cpu(op->op); - seq_printf(s, "\t%s", ceph_osd_op_name(opcode)); - op++; - } - - seq_printf(s, "\n"); - } - mutex_unlock(&osdc->request_mutex); - return 0; -} - static int caps_show(struct seq_file *s, void *p) { - struct ceph_client *client = s->private; + struct ceph_fs_client *fsc = s->private; int total, avail, used, reserved, min; - ceph_reservation_status(client, &total, &avail, &used, &reserved, &min); + ceph_reservation_status(fsc, &total, &avail, &used, &reserved, &min); seq_printf(s, "total\t\t%d\n" "avail\t\t%d\n" "used\t\t%d\n" @@ -280,8 +136,8 @@ static int caps_show(struct seq_file *s, void *p) static int dentry_lru_show(struct seq_file *s, void *ptr) { - struct ceph_client *client = s->private; - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = s->private; + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_dentry_info *di; spin_lock(&mdsc->dentry_lru_lock); @@ -295,199 +151,124 @@ static int dentry_lru_show(struct seq_file *s, void *ptr) return 0; } -#define DEFINE_SHOW_FUNC(name) \ -static int name##_open(struct inode *inode, struct file *file) \ -{ \ - struct seq_file *sf; \ - int ret; \ - \ - ret = single_open(file, name, NULL); \ - sf = file->private_data; \ - sf->private = inode->i_private; \ - return ret; \ -} \ - \ -static const struct file_operations name##_fops = { \ - .open = name##_open, \ - .read = seq_read, \ - .llseek = seq_lseek, \ - .release = single_release, \ -}; - -DEFINE_SHOW_FUNC(monmap_show) -DEFINE_SHOW_FUNC(mdsmap_show) -DEFINE_SHOW_FUNC(osdmap_show) -DEFINE_SHOW_FUNC(monc_show) -DEFINE_SHOW_FUNC(mdsc_show) -DEFINE_SHOW_FUNC(osdc_show) -DEFINE_SHOW_FUNC(dentry_lru_show) -DEFINE_SHOW_FUNC(caps_show) +CEPH_DEFINE_SHOW_FUNC(mdsmap_show) +CEPH_DEFINE_SHOW_FUNC(mdsc_show) +CEPH_DEFINE_SHOW_FUNC(caps_show) +CEPH_DEFINE_SHOW_FUNC(dentry_lru_show) + +/* + * debugfs + */ static int congestion_kb_set(void *data, u64 val) { - struct ceph_client *client = (struct ceph_client *)data; - - if (client) - client->mount_args->congestion_kb = (int)val; + struct ceph_fs_client *fsc = (struct ceph_fs_client *)data; + fsc->mount_options->congestion_kb = (int)val; return 0; } static int congestion_kb_get(void *data, u64 *val) { - struct ceph_client *client = (struct ceph_client *)data; - - if (client) - *val = (u64)client->mount_args->congestion_kb; + struct ceph_fs_client *fsc = (struct ceph_fs_client *)data; + *val = (u64)fsc->mount_options->congestion_kb; return 0; } - DEFINE_SIMPLE_ATTRIBUTE(congestion_kb_fops, congestion_kb_get, congestion_kb_set, "%llu\n"); -int __init ceph_debugfs_init(void) -{ - ceph_debugfs_dir = debugfs_create_dir("ceph", NULL); - if (!ceph_debugfs_dir) - return -ENOMEM; - return 0; -} -void ceph_debugfs_cleanup(void) +void ceph_fs_debugfs_cleanup(struct ceph_fs_client *fsc) { - debugfs_remove(ceph_debugfs_dir); + dout("ceph_fs_debugfs_cleanup\n"); + debugfs_remove(fsc->debugfs_bdi); + debugfs_remove(fsc->debugfs_congestion_kb); + debugfs_remove(fsc->debugfs_mdsmap); + debugfs_remove(fsc->debugfs_caps); + debugfs_remove(fsc->debugfs_mdsc); + debugfs_remove(fsc->debugfs_dentry_lru); } -int ceph_debugfs_client_init(struct ceph_client *client) +int ceph_fs_debugfs_init(struct ceph_fs_client *fsc) { - int ret = 0; - char name[80]; - - snprintf(name, sizeof(name), "%pU.client%lld", &client->fsid, - client->monc.auth->global_id); - - client->debugfs_dir = debugfs_create_dir(name, ceph_debugfs_dir); - if (!client->debugfs_dir) - goto out; + char name[100]; + int err = -ENOMEM; - client->monc.debugfs_file = debugfs_create_file("monc", - 0600, - client->debugfs_dir, - client, - &monc_show_fops); - if (!client->monc.debugfs_file) - goto out; - - client->mdsc.debugfs_file = debugfs_create_file("mdsc", - 0600, - client->debugfs_dir, - client, - &mdsc_show_fops); - if (!client->mdsc.debugfs_file) + dout("ceph_fs_debugfs_init\n"); + fsc->debugfs_congestion_kb = + debugfs_create_file("writeback_congestion_kb", + 0600, + fsc->client->debugfs_dir, + fsc, + &congestion_kb_fops); + if (!fsc->debugfs_congestion_kb) goto out; - client->osdc.debugfs_file = debugfs_create_file("osdc", - 0600, - client->debugfs_dir, - client, - &osdc_show_fops); - if (!client->osdc.debugfs_file) - goto out; + dout("a\n"); - client->debugfs_monmap = debugfs_create_file("monmap", - 0600, - client->debugfs_dir, - client, - &monmap_show_fops); - if (!client->debugfs_monmap) + snprintf(name, sizeof(name), "../../bdi/%s", + dev_name(fsc->backing_dev_info.dev)); + fsc->debugfs_bdi = + debugfs_create_symlink("bdi", + fsc->client->debugfs_dir, + name); + if (!fsc->debugfs_bdi) goto out; - client->debugfs_mdsmap = debugfs_create_file("mdsmap", + dout("b\n"); + fsc->debugfs_mdsmap = debugfs_create_file("mdsmap", 0600, - client->debugfs_dir, - client, + fsc->client->debugfs_dir, + fsc, &mdsmap_show_fops); - if (!client->debugfs_mdsmap) + if (!fsc->debugfs_mdsmap) goto out; - client->debugfs_osdmap = debugfs_create_file("osdmap", - 0600, - client->debugfs_dir, - client, - &osdmap_show_fops); - if (!client->debugfs_osdmap) + dout("ca\n"); + fsc->debugfs_mdsc = debugfs_create_file("mdsc", + 0600, + fsc->client->debugfs_dir, + fsc, + &mdsc_show_fops); + if (!fsc->debugfs_mdsc) goto out; - client->debugfs_dentry_lru = debugfs_create_file("dentry_lru", - 0600, - client->debugfs_dir, - client, - &dentry_lru_show_fops); - if (!client->debugfs_dentry_lru) - goto out; - - client->debugfs_caps = debugfs_create_file("caps", + dout("da\n"); + fsc->debugfs_caps = debugfs_create_file("caps", 0400, - client->debugfs_dir, - client, + fsc->client->debugfs_dir, + fsc, &caps_show_fops); - if (!client->debugfs_caps) + if (!fsc->debugfs_caps) goto out; - client->debugfs_congestion_kb = - debugfs_create_file("writeback_congestion_kb", - 0600, - client->debugfs_dir, - client, - &congestion_kb_fops); - if (!client->debugfs_congestion_kb) + dout("ea\n"); + fsc->debugfs_dentry_lru = debugfs_create_file("dentry_lru", + 0600, + fsc->client->debugfs_dir, + fsc, + &dentry_lru_show_fops); + if (!fsc->debugfs_dentry_lru) goto out; - sprintf(name, "../../bdi/%s", dev_name(client->sb->s_bdi->dev)); - client->debugfs_bdi = debugfs_create_symlink("bdi", client->debugfs_dir, - name); - return 0; out: - ceph_debugfs_client_cleanup(client); - return ret; + ceph_fs_debugfs_cleanup(fsc); + return err; } -void ceph_debugfs_client_cleanup(struct ceph_client *client) -{ - debugfs_remove(client->debugfs_bdi); - debugfs_remove(client->debugfs_caps); - debugfs_remove(client->debugfs_dentry_lru); - debugfs_remove(client->debugfs_osdmap); - debugfs_remove(client->debugfs_mdsmap); - debugfs_remove(client->debugfs_monmap); - debugfs_remove(client->osdc.debugfs_file); - debugfs_remove(client->mdsc.debugfs_file); - debugfs_remove(client->monc.debugfs_file); - debugfs_remove(client->debugfs_congestion_kb); - debugfs_remove(client->debugfs_dir); -} #else /* CONFIG_DEBUG_FS */ -int __init ceph_debugfs_init(void) -{ - return 0; -} - -void ceph_debugfs_cleanup(void) -{ -} - -int ceph_debugfs_client_init(struct ceph_client *client) +int ceph_fs_debugfs_init(struct ceph_fs_client *fsc) { return 0; } -void ceph_debugfs_client_cleanup(struct ceph_client *client) +void ceph_fs_debugfs_cleanup(struct ceph_fs_client *fsc) { } diff --git a/fs/ceph/decode.h b/fs/ceph/decode.h deleted file mode 100644 index c5b6939..0000000 --- a/fs/ceph/decode.h +++ /dev/null @@ -1,201 +0,0 @@ -#ifndef __CEPH_DECODE_H -#define __CEPH_DECODE_H - -#include -#include - -#include "types.h" - -/* - * in all cases, - * void **p pointer to position pointer - * void *end pointer to end of buffer (last byte + 1) - */ - -static inline u64 ceph_decode_64(void **p) -{ - u64 v = get_unaligned_le64(*p); - *p += sizeof(u64); - return v; -} -static inline u32 ceph_decode_32(void **p) -{ - u32 v = get_unaligned_le32(*p); - *p += sizeof(u32); - return v; -} -static inline u16 ceph_decode_16(void **p) -{ - u16 v = get_unaligned_le16(*p); - *p += sizeof(u16); - return v; -} -static inline u8 ceph_decode_8(void **p) -{ - u8 v = *(u8 *)*p; - (*p)++; - return v; -} -static inline void ceph_decode_copy(void **p, void *pv, size_t n) -{ - memcpy(pv, *p, n); - *p += n; -} - -/* - * bounds check input. - */ -#define ceph_decode_need(p, end, n, bad) \ - do { \ - if (unlikely(*(p) + (n) > (end))) \ - goto bad; \ - } while (0) - -#define ceph_decode_64_safe(p, end, v, bad) \ - do { \ - ceph_decode_need(p, end, sizeof(u64), bad); \ - v = ceph_decode_64(p); \ - } while (0) -#define ceph_decode_32_safe(p, end, v, bad) \ - do { \ - ceph_decode_need(p, end, sizeof(u32), bad); \ - v = ceph_decode_32(p); \ - } while (0) -#define ceph_decode_16_safe(p, end, v, bad) \ - do { \ - ceph_decode_need(p, end, sizeof(u16), bad); \ - v = ceph_decode_16(p); \ - } while (0) -#define ceph_decode_8_safe(p, end, v, bad) \ - do { \ - ceph_decode_need(p, end, sizeof(u8), bad); \ - v = ceph_decode_8(p); \ - } while (0) - -#define ceph_decode_copy_safe(p, end, pv, n, bad) \ - do { \ - ceph_decode_need(p, end, n, bad); \ - ceph_decode_copy(p, pv, n); \ - } while (0) - -/* - * struct ceph_timespec <-> struct timespec - */ -static inline void ceph_decode_timespec(struct timespec *ts, - const struct ceph_timespec *tv) -{ - ts->tv_sec = le32_to_cpu(tv->tv_sec); - ts->tv_nsec = le32_to_cpu(tv->tv_nsec); -} -static inline void ceph_encode_timespec(struct ceph_timespec *tv, - const struct timespec *ts) -{ - tv->tv_sec = cpu_to_le32(ts->tv_sec); - tv->tv_nsec = cpu_to_le32(ts->tv_nsec); -} - -/* - * sockaddr_storage <-> ceph_sockaddr - */ -static inline void ceph_encode_addr(struct ceph_entity_addr *a) -{ - __be16 ss_family = htons(a->in_addr.ss_family); - a->in_addr.ss_family = *(__u16 *)&ss_family; -} -static inline void ceph_decode_addr(struct ceph_entity_addr *a) -{ - __be16 ss_family = *(__be16 *)&a->in_addr.ss_family; - a->in_addr.ss_family = ntohs(ss_family); - WARN_ON(a->in_addr.ss_family == 512); -} - -/* - * encoders - */ -static inline void ceph_encode_64(void **p, u64 v) -{ - put_unaligned_le64(v, (__le64 *)*p); - *p += sizeof(u64); -} -static inline void ceph_encode_32(void **p, u32 v) -{ - put_unaligned_le32(v, (__le32 *)*p); - *p += sizeof(u32); -} -static inline void ceph_encode_16(void **p, u16 v) -{ - put_unaligned_le16(v, (__le16 *)*p); - *p += sizeof(u16); -} -static inline void ceph_encode_8(void **p, u8 v) -{ - *(u8 *)*p = v; - (*p)++; -} -static inline void ceph_encode_copy(void **p, const void *s, int len) -{ - memcpy(*p, s, len); - *p += len; -} - -/* - * filepath, string encoders - */ -static inline void ceph_encode_filepath(void **p, void *end, - u64 ino, const char *path) -{ - u32 len = path ? strlen(path) : 0; - BUG_ON(*p + sizeof(ino) + sizeof(len) + len > end); - ceph_encode_8(p, 1); - ceph_encode_64(p, ino); - ceph_encode_32(p, len); - if (len) - memcpy(*p, path, len); - *p += len; -} - -static inline void ceph_encode_string(void **p, void *end, - const char *s, u32 len) -{ - BUG_ON(*p + sizeof(len) + len > end); - ceph_encode_32(p, len); - if (len) - memcpy(*p, s, len); - *p += len; -} - -#define ceph_encode_need(p, end, n, bad) \ - do { \ - if (unlikely(*(p) + (n) > (end))) \ - goto bad; \ - } while (0) - -#define ceph_encode_64_safe(p, end, v, bad) \ - do { \ - ceph_encode_need(p, end, sizeof(u64), bad); \ - ceph_encode_64(p, v); \ - } while (0) -#define ceph_encode_32_safe(p, end, v, bad) \ - do { \ - ceph_encode_need(p, end, sizeof(u32), bad); \ - ceph_encode_32(p, v); \ - } while (0) -#define ceph_encode_16_safe(p, end, v, bad) \ - do { \ - ceph_encode_need(p, end, sizeof(u16), bad); \ - ceph_encode_16(p, v); \ - } while (0) - -#define ceph_encode_copy_safe(p, end, pv, n, bad) \ - do { \ - ceph_encode_need(p, end, n, bad); \ - ceph_encode_copy(p, pv, n); \ - } while (0) -#define ceph_encode_string_safe(p, end, s, n, bad) \ - do { \ - ceph_encode_need(p, end, n, bad); \ - ceph_encode_string(p, end, s, n); \ - } while (0) - - -#endif diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index a1986eb..a8d2aac 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -1,4 +1,4 @@ -#include "ceph_debug.h" +#include #include #include @@ -7,6 +7,7 @@ #include #include "super.h" +#include "mds_client.h" /* * Directory operations: readdir, lookup, create, link, unlink, @@ -227,15 +228,15 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir) struct ceph_file_info *fi = filp->private_data; struct inode *inode = filp->f_dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_client *client = ceph_inode_to_client(inode); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + struct ceph_mds_client *mdsc = fsc->mdsc; unsigned frag = fpos_frag(filp->f_pos); int off = fpos_off(filp->f_pos); int err; u32 ftype; struct ceph_mds_reply_info_parsed *rinfo; - const int max_entries = client->mount_args->max_readdir; - const int max_bytes = client->mount_args->max_readdir_bytes; + const int max_entries = fsc->mount_options->max_readdir; + const int max_bytes = fsc->mount_options->max_readdir_bytes; dout("readdir %p filp %p frag %u off %u\n", inode, filp, frag, off); if (fi->at_end) @@ -267,7 +268,7 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir) /* can we use the dcache? */ spin_lock(&inode->i_lock); if ((filp->f_pos == 2 || fi->dentry) && - !ceph_test_opt(client, NOASYNCREADDIR) && + !ceph_test_mount_opt(fsc, NOASYNCREADDIR) && ceph_snap(inode) != CEPH_SNAPDIR && (ci->i_ceph_flags & CEPH_I_COMPLETE) && __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) { @@ -487,14 +488,14 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int origin) struct dentry *ceph_finish_lookup(struct ceph_mds_request *req, struct dentry *dentry, int err) { - struct ceph_client *client = ceph_sb_to_client(dentry->d_sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); struct inode *parent = dentry->d_parent->d_inode; /* .snap dir? */ if (err == -ENOENT && ceph_vino(parent).ino != CEPH_INO_ROOT && /* no .snap in root dir */ strcmp(dentry->d_name.name, - client->mount_args->snapdir_name) == 0) { + fsc->mount_options->snapdir_name) == 0) { struct inode *inode = ceph_get_snapdir(parent); dout("ENOENT on snapdir %p '%.*s', linking to snapdir %p\n", dentry, dentry->d_name.len, dentry->d_name.name, inode); @@ -539,8 +540,8 @@ static int is_root_ceph_dentry(struct inode *inode, struct dentry *dentry) static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd) { - struct ceph_client *client = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int op; int err; @@ -572,7 +573,7 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, spin_lock(&dir->i_lock); dout(" dir %p flags are %d\n", dir, ci->i_ceph_flags); if (strncmp(dentry->d_name.name, - client->mount_args->snapdir_name, + fsc->mount_options->snapdir_name, dentry->d_name.len) && !is_root_ceph_dentry(dir, dentry) && (ci->i_ceph_flags & CEPH_I_COMPLETE) && @@ -629,8 +630,8 @@ int ceph_handle_notrace_create(struct inode *dir, struct dentry *dentry) static int ceph_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev) { - struct ceph_client *client = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int err; @@ -685,8 +686,8 @@ static int ceph_create(struct inode *dir, struct dentry *dentry, int mode, static int ceph_symlink(struct inode *dir, struct dentry *dentry, const char *dest) { - struct ceph_client *client = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int err; @@ -716,8 +717,8 @@ static int ceph_symlink(struct inode *dir, struct dentry *dentry, static int ceph_mkdir(struct inode *dir, struct dentry *dentry, int mode) { - struct ceph_client *client = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int err = -EROFS; int op; @@ -758,8 +759,8 @@ out: static int ceph_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry) { - struct ceph_client *client = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int err; @@ -813,8 +814,8 @@ static int drop_caps_for_unlink(struct inode *inode) */ static int ceph_unlink(struct inode *dir, struct dentry *dentry) { - struct ceph_client *client = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct inode *inode = dentry->d_inode; struct ceph_mds_request *req; int err = -EROFS; @@ -854,8 +855,8 @@ out: static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry) { - struct ceph_client *client = ceph_sb_to_client(old_dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(old_dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int err; @@ -1076,7 +1077,7 @@ static ssize_t ceph_read_dir(struct file *file, char __user *buf, size_t size, struct ceph_inode_info *ci = ceph_inode(inode); int left; - if (!ceph_test_opt(ceph_sb_to_client(inode->i_sb), DIRSTAT)) + if (!ceph_test_mount_opt(ceph_sb_to_client(inode->i_sb), DIRSTAT)) return -EISDIR; if (!cf->dir_info) { @@ -1177,7 +1178,7 @@ void ceph_dentry_lru_add(struct dentry *dn) dout("dentry_lru_add %p %p '%.*s'\n", di, dn, dn->d_name.len, dn->d_name.name); if (di) { - mdsc = &ceph_sb_to_client(dn->d_sb)->mdsc; + mdsc = ceph_sb_to_client(dn->d_sb)->mdsc; spin_lock(&mdsc->dentry_lru_lock); list_add_tail(&di->lru, &mdsc->dentry_lru); mdsc->num_dentry++; @@ -1193,7 +1194,7 @@ void ceph_dentry_lru_touch(struct dentry *dn) dout("dentry_lru_touch %p %p '%.*s' (offset %lld)\n", di, dn, dn->d_name.len, dn->d_name.name, di->offset); if (di) { - mdsc = &ceph_sb_to_client(dn->d_sb)->mdsc; + mdsc = ceph_sb_to_client(dn->d_sb)->mdsc; spin_lock(&mdsc->dentry_lru_lock); list_move_tail(&di->lru, &mdsc->dentry_lru); spin_unlock(&mdsc->dentry_lru_lock); @@ -1208,7 +1209,7 @@ void ceph_dentry_lru_del(struct dentry *dn) dout("dentry_lru_del %p %p '%.*s'\n", di, dn, dn->d_name.len, dn->d_name.name); if (di) { - mdsc = &ceph_sb_to_client(dn->d_sb)->mdsc; + mdsc = ceph_sb_to_client(dn->d_sb)->mdsc; spin_lock(&mdsc->dentry_lru_lock); list_del_init(&di->lru); mdsc->num_dentry--; diff --git a/fs/ceph/export.c b/fs/ceph/export.c index e38423e..2297d94 100644 --- a/fs/ceph/export.c +++ b/fs/ceph/export.c @@ -1,10 +1,11 @@ -#include "ceph_debug.h" +#include #include #include #include #include "super.h" +#include "mds_client.h" /* * NFS export support @@ -120,7 +121,7 @@ static struct dentry *__fh_to_dentry(struct super_block *sb, static struct dentry *__cfh_to_dentry(struct super_block *sb, struct ceph_nfs_confh *cfh) { - struct ceph_mds_client *mdsc = &ceph_sb_to_client(sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(sb)->mdsc; struct inode *inode; struct dentry *dentry; struct ceph_vino vino; diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 66e4da6..e77c28c 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -1,5 +1,6 @@ -#include "ceph_debug.h" +#include +#include #include #include #include @@ -38,8 +39,8 @@ static struct ceph_mds_request * prepare_open_request(struct super_block *sb, int flags, int create_mode) { - struct ceph_client *client = ceph_sb_to_client(sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int want_auth = USE_ANY_MDS; int op = (flags & O_CREAT) ? CEPH_MDS_OP_CREATE : CEPH_MDS_OP_OPEN; @@ -117,8 +118,8 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode) int ceph_open(struct inode *inode, struct file *file) { struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_client *client = ceph_sb_to_client(inode->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; struct ceph_file_info *cf = file->private_data; struct inode *parent_inode = file->f_dentry->d_parent->d_inode; @@ -216,8 +217,8 @@ struct dentry *ceph_lookup_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd, int mode, int locked_dir) { - struct ceph_client *client = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct file *file = nd->intent.open.file; struct inode *parent_inode = get_dentry_parent_inode(file->f_dentry); struct ceph_mds_request *req; @@ -270,163 +271,6 @@ int ceph_release(struct inode *inode, struct file *file) } /* - * build a vector of user pages - */ -static struct page **get_direct_page_vector(const char __user *data, - int num_pages, - loff_t off, size_t len) -{ - struct page **pages; - int rc; - - pages = kmalloc(sizeof(*pages) * num_pages, GFP_NOFS); - if (!pages) - return ERR_PTR(-ENOMEM); - - down_read(¤t->mm->mmap_sem); - rc = get_user_pages(current, current->mm, (unsigned long)data, - num_pages, 0, 0, pages, NULL); - up_read(¤t->mm->mmap_sem); - if (rc < 0) - goto fail; - return pages; - -fail: - kfree(pages); - return ERR_PTR(rc); -} - -static void put_page_vector(struct page **pages, int num_pages) -{ - int i; - - for (i = 0; i < num_pages; i++) - put_page(pages[i]); - kfree(pages); -} - -void ceph_release_page_vector(struct page **pages, int num_pages) -{ - int i; - - for (i = 0; i < num_pages; i++) - __free_pages(pages[i], 0); - kfree(pages); -} - -/* - * allocate a vector new pages - */ -static struct page **ceph_alloc_page_vector(int num_pages, gfp_t flags) -{ - struct page **pages; - int i; - - pages = kmalloc(sizeof(*pages) * num_pages, flags); - if (!pages) - return ERR_PTR(-ENOMEM); - for (i = 0; i < num_pages; i++) { - pages[i] = __page_cache_alloc(flags); - if (pages[i] == NULL) { - ceph_release_page_vector(pages, i); - return ERR_PTR(-ENOMEM); - } - } - return pages; -} - -/* - * copy user data into a page vector - */ -static int copy_user_to_page_vector(struct page **pages, - const char __user *data, - loff_t off, size_t len) -{ - int i = 0; - int po = off & ~PAGE_CACHE_MASK; - int left = len; - int l, bad; - - while (left > 0) { - l = min_t(int, PAGE_CACHE_SIZE-po, left); - bad = copy_from_user(page_address(pages[i]) + po, data, l); - if (bad == l) - return -EFAULT; - data += l - bad; - left -= l - bad; - po += l - bad; - if (po == PAGE_CACHE_SIZE) { - po = 0; - i++; - } - } - return len; -} - -/* - * copy user data from a page vector into a user pointer - */ -static int copy_page_vector_to_user(struct page **pages, char __user *data, - loff_t off, size_t len) -{ - int i = 0; - int po = off & ~PAGE_CACHE_MASK; - int left = len; - int l, bad; - - while (left > 0) { - l = min_t(int, left, PAGE_CACHE_SIZE-po); - bad = copy_to_user(data, page_address(pages[i]) + po, l); - if (bad == l) - return -EFAULT; - data += l - bad; - left -= l - bad; - if (po) { - po += l - bad; - if (po == PAGE_CACHE_SIZE) - po = 0; - } - i++; - } - return len; -} - -/* - * Zero an extent within a page vector. Offset is relative to the - * start of the first page. - */ -static void zero_page_vector_range(int off, int len, struct page **pages) -{ - int i = off >> PAGE_CACHE_SHIFT; - - off &= ~PAGE_CACHE_MASK; - - dout("zero_page_vector_page %u~%u\n", off, len); - - /* leading partial page? */ - if (off) { - int end = min((int)PAGE_CACHE_SIZE, off + len); - dout("zeroing %d %p head from %d\n", i, pages[i], - (int)off); - zero_user_segment(pages[i], off, end); - len -= (end - off); - i++; - } - while (len >= PAGE_CACHE_SIZE) { - dout("zeroing %d %p len=%d\n", i, pages[i], len); - zero_user_segment(pages[i], 0, PAGE_CACHE_SIZE); - len -= PAGE_CACHE_SIZE; - i++; - } - /* trailing partial page? */ - if (len) { - dout("zeroing %d %p tail to %d\n", i, pages[i], (int)len); - zero_user_segment(pages[i], 0, len); - } -} - - -/* * Read a range of bytes striped over one or more objects. Iterate over * objects we stripe over. (That's not atomic, but good enough for now.) * @@ -438,7 +282,7 @@ static int striped_read(struct inode *inode, struct page **pages, int num_pages, int *checkeof) { - struct ceph_client *client = ceph_inode_to_client(inode); + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_inode_info *ci = ceph_inode(inode); u64 pos, this_len; int page_off = off & ~PAGE_CACHE_MASK; /* first byte's offset in page */ @@ -459,7 +303,7 @@ static int striped_read(struct inode *inode, more: this_len = left; - ret = ceph_osdc_readpages(&client->osdc, ceph_vino(inode), + ret = ceph_osdc_readpages(&fsc->client->osdc, ceph_vino(inode), &ci->i_layout, pos, &this_len, ci->i_truncate_seq, ci->i_truncate_size, @@ -477,8 +321,8 @@ more: if (read < pos - off) { dout(" zero gap %llu to %llu\n", off + read, pos); - zero_page_vector_range(page_off + read, - pos - off - read, pages); + ceph_zero_page_vector_range(page_off + read, + pos - off - read, pages); } pos += ret; read = pos - off; @@ -495,8 +339,8 @@ more: /* was original extent fully inside i_size? */ if (pos + left <= inode->i_size) { dout("zero tail\n"); - zero_page_vector_range(page_off + read, len - read, - pages); + ceph_zero_page_vector_range(page_off + read, len - read, + pages); read = len; goto out; } @@ -531,7 +375,7 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data, (file->f_flags & O_DIRECT) ? "O_DIRECT" : ""); if (file->f_flags & O_DIRECT) { - pages = get_direct_page_vector(data, num_pages, off, len); + pages = ceph_get_direct_page_vector(data, num_pages, off, len); /* * flush any page cache pages in this range. this @@ -552,13 +396,13 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data, ret = striped_read(inode, off, len, pages, num_pages, checkeof); if (ret >= 0 && (file->f_flags & O_DIRECT) == 0) - ret = copy_page_vector_to_user(pages, data, off, ret); + ret = ceph_copy_page_vector_to_user(pages, data, off, ret); if (ret >= 0) *poff = off + ret; done: if (file->f_flags & O_DIRECT) - put_page_vector(pages, num_pages); + ceph_put_page_vector(pages, num_pages); else ceph_release_page_vector(pages, num_pages); dout("sync_read result %d\n", ret); @@ -594,7 +438,7 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data, { struct inode *inode = file->f_dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_client *client = ceph_inode_to_client(inode); + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_osd_request *req; struct page **pages; int num_pages; @@ -642,7 +486,7 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data, */ more: len = left; - req = ceph_osdc_new_request(&client->osdc, &ci->i_layout, + req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout, ceph_vino(inode), pos, &len, CEPH_OSD_OP_WRITE, flags, ci->i_snap_realm->cached_context, @@ -655,7 +499,7 @@ more: num_pages = calc_pages_for(pos, len); if (file->f_flags & O_DIRECT) { - pages = get_direct_page_vector(data, num_pages, pos, len); + pages = ceph_get_direct_page_vector(data, num_pages, pos, len); if (IS_ERR(pages)) { ret = PTR_ERR(pages); goto out; @@ -673,7 +517,7 @@ more: ret = PTR_ERR(pages); goto out; } - ret = copy_user_to_page_vector(pages, data, pos, len); + ret = ceph_copy_user_to_page_vector(pages, data, pos, len); if (ret < 0) { ceph_release_page_vector(pages, num_pages); goto out; @@ -689,7 +533,7 @@ more: req->r_num_pages = num_pages; req->r_inode = inode; - ret = ceph_osdc_start_request(&client->osdc, req, false); + ret = ceph_osdc_start_request(&fsc->client->osdc, req, false); if (!ret) { if (req->r_safe_callback) { /* @@ -701,11 +545,11 @@ more: spin_unlock(&ci->i_unsafe_lock); ceph_get_cap_refs(ci, CEPH_CAP_FILE_WR); } - ret = ceph_osdc_wait_request(&client->osdc, req); + ret = ceph_osdc_wait_request(&fsc->client->osdc, req); } if (file->f_flags & O_DIRECT) - put_page_vector(pages, num_pages); + ceph_put_page_vector(pages, num_pages); else if (file->f_flags & O_SYNC) ceph_release_page_vector(pages, num_pages); @@ -814,7 +658,8 @@ static ssize_t ceph_aio_write(struct kiocb *iocb, const struct iovec *iov, struct ceph_file_info *fi = file->private_data; struct inode *inode = file->f_dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_osd_client *osdc = &ceph_sb_to_client(inode->i_sb)->osdc; + struct ceph_osd_client *osdc = + &ceph_sb_to_client(inode->i_sb)->client->osdc; loff_t endoff = pos + iov->iov_len; int want, got = 0; int ret, err; diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 62377ec..1d6a45b 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1,4 +1,4 @@ -#include "ceph_debug.h" +#include #include #include @@ -13,7 +13,8 @@ #include #include "super.h" -#include "decode.h" +#include "mds_client.h" +#include /* * Ceph inode operations @@ -384,7 +385,7 @@ void ceph_destroy_inode(struct inode *inode) */ if (ci->i_snap_realm) { struct ceph_mds_client *mdsc = - &ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; + ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; struct ceph_snap_realm *realm = ci->i_snap_realm; dout(" dropping residual ref to snap realm %p\n", realm); @@ -685,7 +686,7 @@ static int fill_inode(struct inode *inode, } /* it may be better to set st_size in getattr instead? */ - if (ceph_test_opt(ceph_sb_to_client(inode->i_sb), RBYTES)) + if (ceph_test_mount_opt(ceph_sb_to_client(inode->i_sb), RBYTES)) inode->i_size = ci->i_rbytes; break; default: @@ -901,7 +902,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, struct inode *in = NULL; struct ceph_mds_reply_inode *ininfo; struct ceph_vino vino; - struct ceph_client *client = ceph_sb_to_client(sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(sb); int i = 0; int err = 0; @@ -965,7 +966,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, */ if (rinfo->head->is_dentry && !req->r_aborted && (rinfo->head->is_target || strncmp(req->r_dentry->d_name.name, - client->mount_args->snapdir_name, + fsc->mount_options->snapdir_name, req->r_dentry->d_name.len))) { /* * lookup link rename : null -> possibly existing inode @@ -1533,7 +1534,7 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) struct inode *parent_inode = dentry->d_parent->d_inode; const unsigned int ia_valid = attr->ia_valid; struct ceph_mds_request *req; - struct ceph_mds_client *mdsc = &ceph_sb_to_client(dentry->d_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(dentry->d_sb)->mdsc; int issued; int release = 0, dirtied = 0; int mask = 0; @@ -1728,8 +1729,8 @@ out: */ int ceph_do_getattr(struct inode *inode, int mask) { - struct ceph_client *client = ceph_sb_to_client(inode->i_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req; int err; diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 76e307d..899578b 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -1,8 +1,10 @@ #include -#include "ioctl.h" #include "super.h" -#include "ceph_debug.h" +#include "mds_client.h" +#include + +#include "ioctl.h" /* @@ -37,7 +39,7 @@ static long ceph_ioctl_set_layout(struct file *file, void __user *arg) { struct inode *inode = file->f_dentry->d_inode; struct inode *parent_inode = file->f_dentry->d_parent->d_inode; - struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_mds_request *req; struct ceph_ioctl_layout l; int err, i; @@ -98,7 +100,8 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) struct ceph_ioctl_dataloc dl; struct inode *inode = file->f_dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_osd_client *osdc = &ceph_sb_to_client(inode->i_sb)->osdc; + struct ceph_osd_client *osdc = + &ceph_sb_to_client(inode->i_sb)->client->osdc; u64 len = 1, olen; u64 tmp; struct ceph_object_layout ol; diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c index ff4e753..087afe9 100644 --- a/fs/ceph/locks.c +++ b/fs/ceph/locks.c @@ -1,11 +1,11 @@ -#include "ceph_debug.h" +#include #include #include #include "super.h" #include "mds_client.h" -#include "pagelist.h" +#include /** * Implement fcntl and flock locking functions. @@ -16,7 +16,7 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, { struct inode *inode = file->f_dentry->d_inode; struct ceph_mds_client *mdsc = - &ceph_sb_to_client(inode->i_sb)->mdsc; + ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_mds_request *req; int err; diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index fad95f8..3356823 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1,17 +1,20 @@ -#include "ceph_debug.h" +#include #include #include #include #include +#include +#include -#include "mds_client.h" -#include "mon_client.h" #include "super.h" -#include "messenger.h" -#include "decode.h" -#include "auth.h" -#include "pagelist.h" +#include "mds_client.h" + +#include +#include +#include +#include +#include /* * A cluster of MDS (metadata server) daemons is responsible for @@ -286,8 +289,9 @@ void ceph_put_mds_session(struct ceph_mds_session *s) atomic_read(&s->s_ref), atomic_read(&s->s_ref)-1); if (atomic_dec_and_test(&s->s_ref)) { if (s->s_authorizer) - s->s_mdsc->client->monc.auth->ops->destroy_authorizer( - s->s_mdsc->client->monc.auth, s->s_authorizer); + s->s_mdsc->fsc->client->monc.auth->ops->destroy_authorizer( + s->s_mdsc->fsc->client->monc.auth, + s->s_authorizer); kfree(s); } } @@ -344,7 +348,7 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc, s->s_seq = 0; mutex_init(&s->s_mutex); - ceph_con_init(mdsc->client->msgr, &s->s_con); + ceph_con_init(mdsc->fsc->client->msgr, &s->s_con); s->s_con.private = s; s->s_con.ops = &mds_con_ops; s->s_con.peer_name.type = CEPH_ENTITY_TYPE_MDS; @@ -599,7 +603,7 @@ static int __choose_mds(struct ceph_mds_client *mdsc, } else if (req->r_dentry) { struct inode *dir = req->r_dentry->d_parent->d_inode; - if (dir->i_sb != mdsc->client->sb) { + if (dir->i_sb != mdsc->fsc->sb) { /* not this fs! */ inode = req->r_dentry->d_inode; } else if (ceph_snap(dir) != CEPH_NOSNAP) { @@ -884,7 +888,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, __ceph_remove_cap(cap); if (!__ceph_is_any_real_caps(ci)) { struct ceph_mds_client *mdsc = - &ceph_sb_to_client(inode->i_sb)->mdsc; + ceph_sb_to_client(inode->i_sb)->mdsc; spin_lock(&mdsc->cap_dirty_lock); if (!list_empty(&ci->i_dirty_item)) { @@ -1146,7 +1150,7 @@ int ceph_add_cap_releases(struct ceph_mds_client *mdsc, struct ceph_msg *msg, *partial = NULL; struct ceph_mds_cap_release *head; int err = -ENOMEM; - int extra = mdsc->client->mount_args->cap_release_safety; + int extra = mdsc->fsc->mount_options->cap_release_safety; int num; dout("add_cap_releases %p mds%d extra %d\n", session, session->s_mds, @@ -2085,7 +2089,7 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg) /* insert trace into our cache */ mutex_lock(&req->r_fill_mutex); - err = ceph_fill_trace(mdsc->client->sb, req, req->r_session); + err = ceph_fill_trace(mdsc->fsc->sb, req, req->r_session); if (err == 0) { if (result == 0 && rinfo->dir_nr) ceph_readdir_prepopulate(req, req->r_session); @@ -2613,7 +2617,7 @@ static void handle_lease(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, struct ceph_msg *msg) { - struct super_block *sb = mdsc->client->sb; + struct super_block *sb = mdsc->fsc->sb; struct inode *inode; struct ceph_inode_info *ci; struct dentry *parent, *dentry; @@ -2891,10 +2895,16 @@ static void delayed_work(struct work_struct *work) schedule_delayed(mdsc); } +int ceph_mdsc_init(struct ceph_fs_client *fsc) -int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client) { - mdsc->client = client; + struct ceph_mds_client *mdsc; + + mdsc = kzalloc(sizeof(struct ceph_mds_client), GFP_NOFS); + if (!mdsc) + return -ENOMEM; + mdsc->fsc = fsc; + fsc->mdsc = mdsc; mutex_init(&mdsc->mutex); mdsc->mdsmap = kzalloc(sizeof(*mdsc->mdsmap), GFP_NOFS); if (mdsc->mdsmap == NULL) @@ -2927,7 +2937,7 @@ int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client) INIT_LIST_HEAD(&mdsc->dentry_lru); ceph_caps_init(mdsc); - ceph_adjust_min_caps(mdsc, client->min_caps); + ceph_adjust_min_caps(mdsc, fsc->min_caps); return 0; } @@ -2939,7 +2949,7 @@ int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client) static void wait_requests(struct ceph_mds_client *mdsc) { struct ceph_mds_request *req; - struct ceph_client *client = mdsc->client; + struct ceph_fs_client *fsc = mdsc->fsc; mutex_lock(&mdsc->mutex); if (__get_oldest_req(mdsc)) { @@ -2947,7 +2957,7 @@ static void wait_requests(struct ceph_mds_client *mdsc) dout("wait_requests waiting for requests\n"); wait_for_completion_timeout(&mdsc->safe_umount_waiters, - client->mount_args->mount_timeout * HZ); + fsc->client->options->mount_timeout * HZ); /* tear down remaining requests */ mutex_lock(&mdsc->mutex); @@ -3030,7 +3040,7 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc) { u64 want_tid, want_flush; - if (mdsc->client->mount_state == CEPH_MOUNT_SHUTDOWN) + if (mdsc->fsc->mount_state == CEPH_MOUNT_SHUTDOWN) return; dout("sync\n"); @@ -3053,7 +3063,7 @@ bool done_closing_sessions(struct ceph_mds_client *mdsc) { int i, n = 0; - if (mdsc->client->mount_state == CEPH_MOUNT_SHUTDOWN) + if (mdsc->fsc->mount_state == CEPH_MOUNT_SHUTDOWN) return true; mutex_lock(&mdsc->mutex); @@ -3071,8 +3081,8 @@ void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc) { struct ceph_mds_session *session; int i; - struct ceph_client *client = mdsc->client; - unsigned long timeout = client->mount_args->mount_timeout * HZ; + struct ceph_fs_client *fsc = mdsc->fsc; + unsigned long timeout = fsc->client->options->mount_timeout * HZ; dout("close_sessions\n"); @@ -3119,7 +3129,7 @@ void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc) dout("stopped\n"); } -void ceph_mdsc_stop(struct ceph_mds_client *mdsc) +static void ceph_mdsc_stop(struct ceph_mds_client *mdsc) { dout("stop\n"); cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */ @@ -3129,6 +3139,15 @@ void ceph_mdsc_stop(struct ceph_mds_client *mdsc) ceph_caps_finalize(mdsc); } +void ceph_mdsc_destroy(struct ceph_fs_client *fsc) +{ + struct ceph_mds_client *mdsc = fsc->mdsc; + + ceph_mdsc_stop(mdsc); + fsc->mdsc = NULL; + kfree(mdsc); +} + /* * handle mds map update. @@ -3145,14 +3164,14 @@ void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, struct ceph_msg *msg) ceph_decode_need(&p, end, sizeof(fsid)+2*sizeof(u32), bad); ceph_decode_copy(&p, &fsid, sizeof(fsid)); - if (ceph_check_fsid(mdsc->client, &fsid) < 0) + if (ceph_check_fsid(mdsc->fsc->client, &fsid) < 0) return; epoch = ceph_decode_32(&p); maplen = ceph_decode_32(&p); dout("handle_map epoch %u len %d\n", epoch, (int)maplen); /* do we need it? */ - ceph_monc_got_mdsmap(&mdsc->client->monc, epoch); + ceph_monc_got_mdsmap(&mdsc->fsc->client->monc, epoch); mutex_lock(&mdsc->mutex); if (mdsc->mdsmap && epoch <= mdsc->mdsmap->m_epoch) { dout("handle_map epoch %u <= our %u\n", @@ -3176,7 +3195,7 @@ void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, struct ceph_msg *msg) } else { mdsc->mdsmap = newmap; /* first mds map */ } - mdsc->client->sb->s_maxbytes = mdsc->mdsmap->m_max_file_size; + mdsc->fsc->sb->s_maxbytes = mdsc->mdsmap->m_max_file_size; __wake_requests(mdsc, &mdsc->waiting_for_map); @@ -3277,7 +3296,7 @@ static int get_authorizer(struct ceph_connection *con, { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; - struct ceph_auth_client *ac = mdsc->client->monc.auth; + struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth; int ret = 0; if (force_new && s->s_authorizer) { @@ -3311,7 +3330,7 @@ static int verify_authorizer_reply(struct ceph_connection *con, int len) { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; - struct ceph_auth_client *ac = mdsc->client->monc.auth; + struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth; return ac->ops->verify_authorizer_reply(ac, s->s_authorizer, len); } @@ -3320,12 +3339,12 @@ static int invalidate_authorizer(struct ceph_connection *con) { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; - struct ceph_auth_client *ac = mdsc->client->monc.auth; + struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth; if (ac->ops->invalidate_authorizer) ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_MDS); - return ceph_monc_validate_auth(&mdsc->client->monc); + return ceph_monc_validate_auth(&mdsc->fsc->client->monc); } static const struct ceph_connection_operations mds_con_ops = { @@ -3338,7 +3357,4 @@ static const struct ceph_connection_operations mds_con_ops = { .peer_reset = peer_reset, }; - - - /* eof */ diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index c98267c..d66d63c 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -8,9 +8,9 @@ #include #include -#include "types.h" -#include "messenger.h" -#include "mdsmap.h" +#include +#include +#include /* * Some lock dependencies: @@ -26,7 +26,7 @@ * */ -struct ceph_client; +struct ceph_fs_client; struct ceph_cap; /* @@ -230,7 +230,7 @@ struct ceph_mds_request { * mds client state */ struct ceph_mds_client { - struct ceph_client *client; + struct ceph_fs_client *fsc; struct mutex mutex; /* all nested structures */ struct ceph_mdsmap *mdsmap; @@ -289,11 +289,6 @@ struct ceph_mds_client { int caps_avail_count; /* unused, unreserved */ int caps_min_count; /* keep at least this many (unreserved) */ - -#ifdef CONFIG_DEBUG_FS - struct dentry *debugfs_file; -#endif - spinlock_t dentry_lru_lock; struct list_head dentry_lru; int num_dentry; @@ -316,10 +311,9 @@ extern void ceph_put_mds_session(struct ceph_mds_session *s); extern int ceph_send_msg_mds(struct ceph_mds_client *mdsc, struct ceph_msg *msg, int mds); -extern int ceph_mdsc_init(struct ceph_mds_client *mdsc, - struct ceph_client *client); +extern int ceph_mdsc_init(struct ceph_fs_client *fsc); extern void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc); -extern void ceph_mdsc_stop(struct ceph_mds_client *mdsc); +extern void ceph_mdsc_destroy(struct ceph_fs_client *fsc); extern void ceph_mdsc_sync(struct ceph_mds_client *mdsc); diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c index 040be6d..73b7d44 100644 --- a/fs/ceph/mdsmap.c +++ b/fs/ceph/mdsmap.c @@ -1,4 +1,4 @@ -#include "ceph_debug.h" +#include #include #include @@ -6,9 +6,9 @@ #include #include -#include "mdsmap.h" -#include "messenger.h" -#include "decode.h" +#include +#include +#include #include "super.h" @@ -117,7 +117,8 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) } dout("mdsmap_decode %d/%d %lld mds%d.%d %s %s\n", - i+1, n, global_id, mds, inc, pr_addr(&addr.in_addr), + i+1, n, global_id, mds, inc, + ceph_pr_addr(&addr.in_addr), ceph_mds_state_name(state)); if (mds >= 0 && mds < m->m_max_mds && state > 0) { m->m_info[mds].global_id = global_id; diff --git a/fs/ceph/mdsmap.h b/fs/ceph/mdsmap.h deleted file mode 100644 index 4c5cb08..0000000 --- a/fs/ceph/mdsmap.h +++ /dev/null @@ -1,62 +0,0 @@ -#ifndef _FS_CEPH_MDSMAP_H -#define _FS_CEPH_MDSMAP_H - -#include "types.h" - -/* - * mds map - describe servers in the mds cluster. - * - * we limit fields to those the client actually xcares about - */ -struct ceph_mds_info { - u64 global_id; - struct ceph_entity_addr addr; - s32 state; - int num_export_targets; - bool laggy; - u32 *export_targets; -}; - -struct ceph_mdsmap { - u32 m_epoch, m_client_epoch, m_last_failure; - u32 m_root; - u32 m_session_timeout; /* seconds */ - u32 m_session_autoclose; /* seconds */ - u64 m_max_file_size; - u32 m_max_mds; /* size of m_addr, m_state arrays */ - struct ceph_mds_info *m_info; - - /* which object pools file data can be stored in */ - int m_num_data_pg_pools; - u32 *m_data_pg_pools; - u32 m_cas_pg_pool; -}; - -static inline struct ceph_entity_addr * -ceph_mdsmap_get_addr(struct ceph_mdsmap *m, int w) -{ - if (w >= m->m_max_mds) - return NULL; - return &m->m_info[w].addr; -} - -static inline int ceph_mdsmap_get_state(struct ceph_mdsmap *m, int w) -{ - BUG_ON(w < 0); - if (w >= m->m_max_mds) - return CEPH_MDS_STATE_DNE; - return m->m_info[w].state; -} - -static inline bool ceph_mdsmap_is_laggy(struct ceph_mdsmap *m, int w) -{ - if (w >= 0 && w < m->m_max_mds) - return m->m_info[w].laggy; - return false; -} - -extern int ceph_mdsmap_get_random_mds(struct ceph_mdsmap *m); -extern struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end); -extern void ceph_mdsmap_destroy(struct ceph_mdsmap *m); - -#endif diff --git a/fs/ceph/messenger.c b/fs/ceph/messenger.c deleted file mode 100644 index 17a09b3..0000000 --- a/fs/ceph/messenger.c +++ /dev/null @@ -1,2432 +0,0 @@ -#include "ceph_debug.h" - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "super.h" -#include "messenger.h" -#include "decode.h" -#include "pagelist.h" - -/* - * Ceph uses the messenger to exchange ceph_msg messages with other - * hosts in the system. The messenger provides ordered and reliable - * delivery. We tolerate TCP disconnects by reconnecting (with - * exponential backoff) in the case of a fault (disconnection, bad - * crc, protocol error). Acks allow sent messages to be discarded by - * the sender. - */ - -/* static tag bytes (protocol control messages) */ -static char tag_msg = CEPH_MSGR_TAG_MSG; -static char tag_ack = CEPH_MSGR_TAG_ACK; -static char tag_keepalive = CEPH_MSGR_TAG_KEEPALIVE; - -#ifdef CONFIG_LOCKDEP -static struct lock_class_key socket_class; -#endif - - -static void queue_con(struct ceph_connection *con); -static void con_work(struct work_struct *); -static void ceph_fault(struct ceph_connection *con); - -/* - * nicely render a sockaddr as a string. - */ -#define MAX_ADDR_STR 20 -#define MAX_ADDR_STR_LEN 60 -static char addr_str[MAX_ADDR_STR][MAX_ADDR_STR_LEN]; -static DEFINE_SPINLOCK(addr_str_lock); -static int last_addr_str; - -const char *pr_addr(const struct sockaddr_storage *ss) -{ - int i; - char *s; - struct sockaddr_in *in4 = (void *)ss; - struct sockaddr_in6 *in6 = (void *)ss; - - spin_lock(&addr_str_lock); - i = last_addr_str++; - if (last_addr_str == MAX_ADDR_STR) - last_addr_str = 0; - spin_unlock(&addr_str_lock); - s = addr_str[i]; - - switch (ss->ss_family) { - case AF_INET: - snprintf(s, MAX_ADDR_STR_LEN, "%pI4:%u", &in4->sin_addr, - (unsigned int)ntohs(in4->sin_port)); - break; - - case AF_INET6: - snprintf(s, MAX_ADDR_STR_LEN, "[%pI6c]:%u", &in6->sin6_addr, - (unsigned int)ntohs(in6->sin6_port)); - break; - - default: - sprintf(s, "(unknown sockaddr family %d)", (int)ss->ss_family); - } - - return s; -} - -static void encode_my_addr(struct ceph_messenger *msgr) -{ - memcpy(&msgr->my_enc_addr, &msgr->inst.addr, sizeof(msgr->my_enc_addr)); - ceph_encode_addr(&msgr->my_enc_addr); -} - -/* - * work queue for all reading and writing to/from the socket. - */ -struct workqueue_struct *ceph_msgr_wq; - -int __init ceph_msgr_init(void) -{ - ceph_msgr_wq = create_workqueue("ceph-msgr"); - if (IS_ERR(ceph_msgr_wq)) { - int ret = PTR_ERR(ceph_msgr_wq); - pr_err("msgr_init failed to create workqueue: %d\n", ret); - ceph_msgr_wq = NULL; - return ret; - } - return 0; -} - -void ceph_msgr_exit(void) -{ - destroy_workqueue(ceph_msgr_wq); -} - -void ceph_msgr_flush(void) -{ - flush_workqueue(ceph_msgr_wq); -} - - -/* - * socket callback functions - */ - -/* data available on socket, or listen socket received a connect */ -static void ceph_data_ready(struct sock *sk, int count_unused) -{ - struct ceph_connection *con = - (struct ceph_connection *)sk->sk_user_data; - if (sk->sk_state != TCP_CLOSE_WAIT) { - dout("ceph_data_ready on %p state = %lu, queueing work\n", - con, con->state); - queue_con(con); - } -} - -/* socket has buffer space for writing */ -static void ceph_write_space(struct sock *sk) -{ - struct ceph_connection *con = - (struct ceph_connection *)sk->sk_user_data; - - /* only queue to workqueue if there is data we want to write. */ - if (test_bit(WRITE_PENDING, &con->state)) { - dout("ceph_write_space %p queueing write work\n", con); - queue_con(con); - } else { - dout("ceph_write_space %p nothing to write\n", con); - } - - /* since we have our own write_space, clear the SOCK_NOSPACE flag */ - clear_bit(SOCK_NOSPACE, &sk->sk_socket->flags); -} - -/* socket's state has changed */ -static void ceph_state_change(struct sock *sk) -{ - struct ceph_connection *con = - (struct ceph_connection *)sk->sk_user_data; - - dout("ceph_state_change %p state = %lu sk_state = %u\n", - con, con->state, sk->sk_state); - - if (test_bit(CLOSED, &con->state)) - return; - - switch (sk->sk_state) { - case TCP_CLOSE: - dout("ceph_state_change TCP_CLOSE\n"); - case TCP_CLOSE_WAIT: - dout("ceph_state_change TCP_CLOSE_WAIT\n"); - if (test_and_set_bit(SOCK_CLOSED, &con->state) == 0) { - if (test_bit(CONNECTING, &con->state)) - con->error_msg = "connection failed"; - else - con->error_msg = "socket closed"; - queue_con(con); - } - break; - case TCP_ESTABLISHED: - dout("ceph_state_change TCP_ESTABLISHED\n"); - queue_con(con); - break; - } -} - -/* - * set up socket callbacks - */ -static void set_sock_callbacks(struct socket *sock, - struct ceph_connection *con) -{ - struct sock *sk = sock->sk; - sk->sk_user_data = (void *)con; - sk->sk_data_ready = ceph_data_ready; - sk->sk_write_space = ceph_write_space; - sk->sk_state_change = ceph_state_change; -} - - -/* - * socket helpers - */ - -/* - * initiate connection to a remote socket. - */ -static struct socket *ceph_tcp_connect(struct ceph_connection *con) -{ - struct sockaddr_storage *paddr = &con->peer_addr.in_addr; - struct socket *sock; - int ret; - - BUG_ON(con->sock); - ret = sock_create_kern(con->peer_addr.in_addr.ss_family, SOCK_STREAM, - IPPROTO_TCP, &sock); - if (ret) - return ERR_PTR(ret); - con->sock = sock; - sock->sk->sk_allocation = GFP_NOFS; - -#ifdef CONFIG_LOCKDEP - lockdep_set_class(&sock->sk->sk_lock, &socket_class); -#endif - - set_sock_callbacks(sock, con); - - dout("connect %s\n", pr_addr(&con->peer_addr.in_addr)); - - ret = sock->ops->connect(sock, (struct sockaddr *)paddr, sizeof(*paddr), - O_NONBLOCK); - if (ret == -EINPROGRESS) { - dout("connect %s EINPROGRESS sk_state = %u\n", - pr_addr(&con->peer_addr.in_addr), - sock->sk->sk_state); - ret = 0; - } - if (ret < 0) { - pr_err("connect %s error %d\n", - pr_addr(&con->peer_addr.in_addr), ret); - sock_release(sock); - con->sock = NULL; - con->error_msg = "connect error"; - } - - if (ret < 0) - return ERR_PTR(ret); - return sock; -} - -static int ceph_tcp_recvmsg(struct socket *sock, void *buf, size_t len) -{ - struct kvec iov = {buf, len}; - struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL }; - - return kernel_recvmsg(sock, &msg, &iov, 1, len, msg.msg_flags); -} - -/* - * write something. @more is true if caller will be sending more data - * shortly. - */ -static int ceph_tcp_sendmsg(struct socket *sock, struct kvec *iov, - size_t kvlen, size_t len, int more) -{ - struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL }; - - if (more) - msg.msg_flags |= MSG_MORE; - else - msg.msg_flags |= MSG_EOR; /* superfluous, but what the hell */ - - return kernel_sendmsg(sock, &msg, iov, kvlen, len); -} - - -/* - * Shutdown/close the socket for the given connection. - */ -static int con_close_socket(struct ceph_connection *con) -{ - int rc; - - dout("con_close_socket on %p sock %p\n", con, con->sock); - if (!con->sock) - return 0; - set_bit(SOCK_CLOSED, &con->state); - rc = con->sock->ops->shutdown(con->sock, SHUT_RDWR); - sock_release(con->sock); - con->sock = NULL; - clear_bit(SOCK_CLOSED, &con->state); - return rc; -} - -/* - * Reset a connection. Discard all incoming and outgoing messages - * and clear *_seq state. - */ -static void ceph_msg_remove(struct ceph_msg *msg) -{ - list_del_init(&msg->list_head); - ceph_msg_put(msg); -} -static void ceph_msg_remove_list(struct list_head *head) -{ - while (!list_empty(head)) { - struct ceph_msg *msg = list_first_entry(head, struct ceph_msg, - list_head); - ceph_msg_remove(msg); - } -} - -static void reset_connection(struct ceph_connection *con) -{ - /* reset connection, out_queue, msg_ and connect_seq */ - /* discard existing out_queue and msg_seq */ - ceph_msg_remove_list(&con->out_queue); - ceph_msg_remove_list(&con->out_sent); - - if (con->in_msg) { - ceph_msg_put(con->in_msg); - con->in_msg = NULL; - } - - con->connect_seq = 0; - con->out_seq = 0; - if (con->out_msg) { - ceph_msg_put(con->out_msg); - con->out_msg = NULL; - } - con->out_keepalive_pending = false; - con->in_seq = 0; - con->in_seq_acked = 0; -} - -/* - * mark a peer down. drop any open connections. - */ -void ceph_con_close(struct ceph_connection *con) -{ - dout("con_close %p peer %s\n", con, pr_addr(&con->peer_addr.in_addr)); - set_bit(CLOSED, &con->state); /* in case there's queued work */ - clear_bit(STANDBY, &con->state); /* avoid connect_seq bump */ - clear_bit(LOSSYTX, &con->state); /* so we retry next connect */ - clear_bit(KEEPALIVE_PENDING, &con->state); - clear_bit(WRITE_PENDING, &con->state); - mutex_lock(&con->mutex); - reset_connection(con); - con->peer_global_seq = 0; - cancel_delayed_work(&con->work); - mutex_unlock(&con->mutex); - queue_con(con); -} - -/* - * Reopen a closed connection, with a new peer address. - */ -void ceph_con_open(struct ceph_connection *con, struct ceph_entity_addr *addr) -{ - dout("con_open %p %s\n", con, pr_addr(&addr->in_addr)); - set_bit(OPENING, &con->state); - clear_bit(CLOSED, &con->state); - memcpy(&con->peer_addr, addr, sizeof(*addr)); - con->delay = 0; /* reset backoff memory */ - queue_con(con); -} - -/* - * return true if this connection ever successfully opened - */ -bool ceph_con_opened(struct ceph_connection *con) -{ - return con->connect_seq > 0; -} - -/* - * generic get/put - */ -struct ceph_connection *ceph_con_get(struct ceph_connection *con) -{ - dout("con_get %p nref = %d -> %d\n", con, - atomic_read(&con->nref), atomic_read(&con->nref) + 1); - if (atomic_inc_not_zero(&con->nref)) - return con; - return NULL; -} - -void ceph_con_put(struct ceph_connection *con) -{ - dout("con_put %p nref = %d -> %d\n", con, - atomic_read(&con->nref), atomic_read(&con->nref) - 1); - BUG_ON(atomic_read(&con->nref) == 0); - if (atomic_dec_and_test(&con->nref)) { - BUG_ON(con->sock); - kfree(con); - } -} - -/* - * initialize a new connection. - */ -void ceph_con_init(struct ceph_messenger *msgr, struct ceph_connection *con) -{ - dout("con_init %p\n", con); - memset(con, 0, sizeof(*con)); - atomic_set(&con->nref, 1); - con->msgr = msgr; - mutex_init(&con->mutex); - INIT_LIST_HEAD(&con->out_queue); - INIT_LIST_HEAD(&con->out_sent); - INIT_DELAYED_WORK(&con->work, con_work); -} - - -/* - * We maintain a global counter to order connection attempts. Get - * a unique seq greater than @gt. - */ -static u32 get_global_seq(struct ceph_messenger *msgr, u32 gt) -{ - u32 ret; - - spin_lock(&msgr->global_seq_lock); - if (msgr->global_seq < gt) - msgr->global_seq = gt; - ret = ++msgr->global_seq; - spin_unlock(&msgr->global_seq_lock); - return ret; -} - - -/* - * Prepare footer for currently outgoing message, and finish things - * off. Assumes out_kvec* are already valid.. we just add on to the end. - */ -static void prepare_write_message_footer(struct ceph_connection *con, int v) -{ - struct ceph_msg *m = con->out_msg; - - dout("prepare_write_message_footer %p\n", con); - con->out_kvec_is_msg = true; - con->out_kvec[v].iov_base = &m->footer; - con->out_kvec[v].iov_len = sizeof(m->footer); - con->out_kvec_bytes += sizeof(m->footer); - con->out_kvec_left++; - con->out_more = m->more_to_follow; - con->out_msg_done = true; -} - -/* - * Prepare headers for the next outgoing message. - */ -static void prepare_write_message(struct ceph_connection *con) -{ - struct ceph_msg *m; - int v = 0; - - con->out_kvec_bytes = 0; - con->out_kvec_is_msg = true; - con->out_msg_done = false; - - /* Sneak an ack in there first? If we can get it into the same - * TCP packet that's a good thing. */ - if (con->in_seq > con->in_seq_acked) { - con->in_seq_acked = con->in_seq; - con->out_kvec[v].iov_base = &tag_ack; - con->out_kvec[v++].iov_len = 1; - con->out_temp_ack = cpu_to_le64(con->in_seq_acked); - con->out_kvec[v].iov_base = &con->out_temp_ack; - con->out_kvec[v++].iov_len = sizeof(con->out_temp_ack); - con->out_kvec_bytes = 1 + sizeof(con->out_temp_ack); - } - - m = list_first_entry(&con->out_queue, - struct ceph_msg, list_head); - con->out_msg = m; - if (test_bit(LOSSYTX, &con->state)) { - list_del_init(&m->list_head); - } else { - /* put message on sent list */ - ceph_msg_get(m); - list_move_tail(&m->list_head, &con->out_sent); - } - - /* - * only assign outgoing seq # if we haven't sent this message - * yet. if it is requeued, resend with it's original seq. - */ - if (m->needs_out_seq) { - m->hdr.seq = cpu_to_le64(++con->out_seq); - m->needs_out_seq = false; - } - - dout("prepare_write_message %p seq %lld type %d len %d+%d+%d %d pgs\n", - m, con->out_seq, le16_to_cpu(m->hdr.type), - le32_to_cpu(m->hdr.front_len), le32_to_cpu(m->hdr.middle_len), - le32_to_cpu(m->hdr.data_len), - m->nr_pages); - BUG_ON(le32_to_cpu(m->hdr.front_len) != m->front.iov_len); - - /* tag + hdr + front + middle */ - con->out_kvec[v].iov_base = &tag_msg; - con->out_kvec[v++].iov_len = 1; - con->out_kvec[v].iov_base = &m->hdr; - con->out_kvec[v++].iov_len = sizeof(m->hdr); - con->out_kvec[v++] = m->front; - if (m->middle) - con->out_kvec[v++] = m->middle->vec; - con->out_kvec_left = v; - con->out_kvec_bytes += 1 + sizeof(m->hdr) + m->front.iov_len + - (m->middle ? m->middle->vec.iov_len : 0); - con->out_kvec_cur = con->out_kvec; - - /* fill in crc (except data pages), footer */ - con->out_msg->hdr.crc = - cpu_to_le32(crc32c(0, (void *)&m->hdr, - sizeof(m->hdr) - sizeof(m->hdr.crc))); - con->out_msg->footer.flags = CEPH_MSG_FOOTER_COMPLETE; - con->out_msg->footer.front_crc = - cpu_to_le32(crc32c(0, m->front.iov_base, m->front.iov_len)); - if (m->middle) - con->out_msg->footer.middle_crc = - cpu_to_le32(crc32c(0, m->middle->vec.iov_base, - m->middle->vec.iov_len)); - else - con->out_msg->footer.middle_crc = 0; - con->out_msg->footer.data_crc = 0; - dout("prepare_write_message front_crc %u data_crc %u\n", - le32_to_cpu(con->out_msg->footer.front_crc), - le32_to_cpu(con->out_msg->footer.middle_crc)); - - /* is there a data payload? */ - if (le32_to_cpu(m->hdr.data_len) > 0) { - /* initialize page iterator */ - con->out_msg_pos.page = 0; - if (m->pages) - con->out_msg_pos.page_pos = - le16_to_cpu(m->hdr.data_off) & ~PAGE_MASK; - else - con->out_msg_pos.page_pos = 0; - con->out_msg_pos.data_pos = 0; - con->out_msg_pos.did_page_crc = 0; - con->out_more = 1; /* data + footer will follow */ - } else { - /* no, queue up footer too and be done */ - prepare_write_message_footer(con, v); - } - - set_bit(WRITE_PENDING, &con->state); -} - -/* - * Prepare an ack. - */ -static void prepare_write_ack(struct ceph_connection *con) -{ - dout("prepare_write_ack %p %llu -> %llu\n", con, - con->in_seq_acked, con->in_seq); - con->in_seq_acked = con->in_seq; - - con->out_kvec[0].iov_base = &tag_ack; - con->out_kvec[0].iov_len = 1; - con->out_temp_ack = cpu_to_le64(con->in_seq_acked); - con->out_kvec[1].iov_base = &con->out_temp_ack; - con->out_kvec[1].iov_len = sizeof(con->out_temp_ack); - con->out_kvec_left = 2; - con->out_kvec_bytes = 1 + sizeof(con->out_temp_ack); - con->out_kvec_cur = con->out_kvec; - con->out_more = 1; /* more will follow.. eventually.. */ - set_bit(WRITE_PENDING, &con->state); -} - -/* - * Prepare to write keepalive byte. - */ -static void prepare_write_keepalive(struct ceph_connection *con) -{ - dout("prepare_write_keepalive %p\n", con); - con->out_kvec[0].iov_base = &tag_keepalive; - con->out_kvec[0].iov_len = 1; - con->out_kvec_left = 1; - con->out_kvec_bytes = 1; - con->out_kvec_cur = con->out_kvec; - set_bit(WRITE_PENDING, &con->state); -} - -/* - * Connection negotiation. - */ - -static void prepare_connect_authorizer(struct ceph_connection *con) -{ - void *auth_buf; - int auth_len = 0; - int auth_protocol = 0; - - mutex_unlock(&con->mutex); - if (con->ops->get_authorizer) - con->ops->get_authorizer(con, &auth_buf, &auth_len, - &auth_protocol, &con->auth_reply_buf, - &con->auth_reply_buf_len, - con->auth_retry); - mutex_lock(&con->mutex); - - con->out_connect.authorizer_protocol = cpu_to_le32(auth_protocol); - con->out_connect.authorizer_len = cpu_to_le32(auth_len); - - con->out_kvec[con->out_kvec_left].iov_base = auth_buf; - con->out_kvec[con->out_kvec_left].iov_len = auth_len; - con->out_kvec_left++; - con->out_kvec_bytes += auth_len; -} - -/* - * We connected to a peer and are saying hello. - */ -static void prepare_write_banner(struct ceph_messenger *msgr, - struct ceph_connection *con) -{ - int len = strlen(CEPH_BANNER); - - con->out_kvec[0].iov_base = CEPH_BANNER; - con->out_kvec[0].iov_len = len; - con->out_kvec[1].iov_base = &msgr->my_enc_addr; - con->out_kvec[1].iov_len = sizeof(msgr->my_enc_addr); - con->out_kvec_left = 2; - con->out_kvec_bytes = len + sizeof(msgr->my_enc_addr); - con->out_kvec_cur = con->out_kvec; - con->out_more = 0; - set_bit(WRITE_PENDING, &con->state); -} - -static void prepare_write_connect(struct ceph_messenger *msgr, - struct ceph_connection *con, - int after_banner) -{ - unsigned global_seq = get_global_seq(con->msgr, 0); - int proto; - - switch (con->peer_name.type) { - case CEPH_ENTITY_TYPE_MON: - proto = CEPH_MONC_PROTOCOL; - break; - case CEPH_ENTITY_TYPE_OSD: - proto = CEPH_OSDC_PROTOCOL; - break; - case CEPH_ENTITY_TYPE_MDS: - proto = CEPH_MDSC_PROTOCOL; - break; - default: - BUG(); - } - - dout("prepare_write_connect %p cseq=%d gseq=%d proto=%d\n", con, - con->connect_seq, global_seq, proto); - - con->out_connect.features = cpu_to_le64(CEPH_FEATURE_SUPPORTED); - con->out_connect.host_type = cpu_to_le32(CEPH_ENTITY_TYPE_CLIENT); - con->out_connect.connect_seq = cpu_to_le32(con->connect_seq); - con->out_connect.global_seq = cpu_to_le32(global_seq); - con->out_connect.protocol_version = cpu_to_le32(proto); - con->out_connect.flags = 0; - - if (!after_banner) { - con->out_kvec_left = 0; - con->out_kvec_bytes = 0; - } - con->out_kvec[con->out_kvec_left].iov_base = &con->out_connect; - con->out_kvec[con->out_kvec_left].iov_len = sizeof(con->out_connect); - con->out_kvec_left++; - con->out_kvec_bytes += sizeof(con->out_connect); - con->out_kvec_cur = con->out_kvec; - con->out_more = 0; - set_bit(WRITE_PENDING, &con->state); - - prepare_connect_authorizer(con); -} - - -/* - * write as much of pending kvecs to the socket as we can. - * 1 -> done - * 0 -> socket full, but more to do - * <0 -> error - */ -static int write_partial_kvec(struct ceph_connection *con) -{ - int ret; - - dout("write_partial_kvec %p %d left\n", con, con->out_kvec_bytes); - while (con->out_kvec_bytes > 0) { - ret = ceph_tcp_sendmsg(con->sock, con->out_kvec_cur, - con->out_kvec_left, con->out_kvec_bytes, - con->out_more); - if (ret <= 0) - goto out; - con->out_kvec_bytes -= ret; - if (con->out_kvec_bytes == 0) - break; /* done */ - while (ret > 0) { - if (ret >= con->out_kvec_cur->iov_len) { - ret -= con->out_kvec_cur->iov_len; - con->out_kvec_cur++; - con->out_kvec_left--; - } else { - con->out_kvec_cur->iov_len -= ret; - con->out_kvec_cur->iov_base += ret; - ret = 0; - break; - } - } - } - con->out_kvec_left = 0; - con->out_kvec_is_msg = false; - ret = 1; -out: - dout("write_partial_kvec %p %d left in %d kvecs ret = %d\n", con, - con->out_kvec_bytes, con->out_kvec_left, ret); - return ret; /* done! */ -} - -#ifdef CONFIG_BLOCK -static void init_bio_iter(struct bio *bio, struct bio **iter, int *seg) -{ - if (!bio) { - *iter = NULL; - *seg = 0; - return; - } - *iter = bio; - *seg = bio->bi_idx; -} - -static void iter_bio_next(struct bio **bio_iter, int *seg) -{ - if (*bio_iter == NULL) - return; - - BUG_ON(*seg >= (*bio_iter)->bi_vcnt); - - (*seg)++; - if (*seg == (*bio_iter)->bi_vcnt) - init_bio_iter((*bio_iter)->bi_next, bio_iter, seg); -} -#endif - -/* - * Write as much message data payload as we can. If we finish, queue - * up the footer. - * 1 -> done, footer is now queued in out_kvec[]. - * 0 -> socket full, but more to do - * <0 -> error - */ -static int write_partial_msg_pages(struct ceph_connection *con) -{ - struct ceph_msg *msg = con->out_msg; - unsigned data_len = le32_to_cpu(msg->hdr.data_len); - size_t len; - int crc = con->msgr->nocrc; - int ret; - int total_max_write; - int in_trail = 0; - size_t trail_len = (msg->trail ? msg->trail->length : 0); - - dout("write_partial_msg_pages %p msg %p page %d/%d offset %d\n", - con, con->out_msg, con->out_msg_pos.page, con->out_msg->nr_pages, - con->out_msg_pos.page_pos); - -#ifdef CONFIG_BLOCK - if (msg->bio && !msg->bio_iter) - init_bio_iter(msg->bio, &msg->bio_iter, &msg->bio_seg); -#endif - - while (data_len > con->out_msg_pos.data_pos) { - struct page *page = NULL; - void *kaddr = NULL; - int max_write = PAGE_SIZE; - int page_shift = 0; - - total_max_write = data_len - trail_len - - con->out_msg_pos.data_pos; - - /* - * if we are calculating the data crc (the default), we need - * to map the page. if our pages[] has been revoked, use the - * zero page. - */ - - /* have we reached the trail part of the data? */ - if (con->out_msg_pos.data_pos >= data_len - trail_len) { - in_trail = 1; - - total_max_write = data_len - con->out_msg_pos.data_pos; - - page = list_first_entry(&msg->trail->head, - struct page, lru); - if (crc) - kaddr = kmap(page); - max_write = PAGE_SIZE; - } else if (msg->pages) { - page = msg->pages[con->out_msg_pos.page]; - if (crc) - kaddr = kmap(page); - } else if (msg->pagelist) { - page = list_first_entry(&msg->pagelist->head, - struct page, lru); - if (crc) - kaddr = kmap(page); -#ifdef CONFIG_BLOCK - } else if (msg->bio) { - struct bio_vec *bv; - - bv = bio_iovec_idx(msg->bio_iter, msg->bio_seg); - page = bv->bv_page; - page_shift = bv->bv_offset; - if (crc) - kaddr = kmap(page) + page_shift; - max_write = bv->bv_len; -#endif - } else { - page = con->msgr->zero_page; - if (crc) - kaddr = page_address(con->msgr->zero_page); - } - len = min_t(int, max_write - con->out_msg_pos.page_pos, - total_max_write); - - if (crc && !con->out_msg_pos.did_page_crc) { - void *base = kaddr + con->out_msg_pos.page_pos; - u32 tmpcrc = le32_to_cpu(con->out_msg->footer.data_crc); - - BUG_ON(kaddr == NULL); - con->out_msg->footer.data_crc = - cpu_to_le32(crc32c(tmpcrc, base, len)); - con->out_msg_pos.did_page_crc = 1; - } - ret = kernel_sendpage(con->sock, page, - con->out_msg_pos.page_pos + page_shift, - len, - MSG_DONTWAIT | MSG_NOSIGNAL | - MSG_MORE); - - if (crc && - (msg->pages || msg->pagelist || msg->bio || in_trail)) - kunmap(page); - - if (ret <= 0) - goto out; - - con->out_msg_pos.data_pos += ret; - con->out_msg_pos.page_pos += ret; - if (ret == len) { - con->out_msg_pos.page_pos = 0; - con->out_msg_pos.page++; - con->out_msg_pos.did_page_crc = 0; - if (in_trail) - list_move_tail(&page->lru, - &msg->trail->head); - else if (msg->pagelist) - list_move_tail(&page->lru, - &msg->pagelist->head); -#ifdef CONFIG_BLOCK - else if (msg->bio) - iter_bio_next(&msg->bio_iter, &msg->bio_seg); -#endif - } - } - - dout("write_partial_msg_pages %p msg %p done\n", con, msg); - - /* prepare and queue up footer, too */ - if (!crc) - con->out_msg->footer.flags |= CEPH_MSG_FOOTER_NOCRC; - con->out_kvec_bytes = 0; - con->out_kvec_left = 0; - con->out_kvec_cur = con->out_kvec; - prepare_write_message_footer(con, 0); - ret = 1; -out: - return ret; -} - -/* - * write some zeros - */ -static int write_partial_skip(struct ceph_connection *con) -{ - int ret; - - while (con->out_skip > 0) { - struct kvec iov = { - .iov_base = page_address(con->msgr->zero_page), - .iov_len = min(con->out_skip, (int)PAGE_CACHE_SIZE) - }; - - ret = ceph_tcp_sendmsg(con->sock, &iov, 1, iov.iov_len, 1); - if (ret <= 0) - goto out; - con->out_skip -= ret; - } - ret = 1; -out: - return ret; -} - -/* - * Prepare to read connection handshake, or an ack. - */ -static void prepare_read_banner(struct ceph_connection *con) -{ - dout("prepare_read_banner %p\n", con); - con->in_base_pos = 0; -} - -static void prepare_read_connect(struct ceph_connection *con) -{ - dout("prepare_read_connect %p\n", con); - con->in_base_pos = 0; -} - -static void prepare_read_ack(struct ceph_connection *con) -{ - dout("prepare_read_ack %p\n", con); - con->in_base_pos = 0; -} - -static void prepare_read_tag(struct ceph_connection *con) -{ - dout("prepare_read_tag %p\n", con); - con->in_base_pos = 0; - con->in_tag = CEPH_MSGR_TAG_READY; -} - -/* - * Prepare to read a message. - */ -static int prepare_read_message(struct ceph_connection *con) -{ - dout("prepare_read_message %p\n", con); - BUG_ON(con->in_msg != NULL); - con->in_base_pos = 0; - con->in_front_crc = con->in_middle_crc = con->in_data_crc = 0; - return 0; -} - - -static int read_partial(struct ceph_connection *con, - int *to, int size, void *object) -{ - *to += size; - while (con->in_base_pos < *to) { - int left = *to - con->in_base_pos; - int have = size - left; - int ret = ceph_tcp_recvmsg(con->sock, object + have, left); - if (ret <= 0) - return ret; - con->in_base_pos += ret; - } - return 1; -} - - -/* - * Read all or part of the connect-side handshake on a new connection - */ -static int read_partial_banner(struct ceph_connection *con) -{ - int ret, to = 0; - - dout("read_partial_banner %p at %d\n", con, con->in_base_pos); - - /* peer's banner */ - ret = read_partial(con, &to, strlen(CEPH_BANNER), con->in_banner); - if (ret <= 0) - goto out; - ret = read_partial(con, &to, sizeof(con->actual_peer_addr), - &con->actual_peer_addr); - if (ret <= 0) - goto out; - ret = read_partial(con, &to, sizeof(con->peer_addr_for_me), - &con->peer_addr_for_me); - if (ret <= 0) - goto out; -out: - return ret; -} - -static int read_partial_connect(struct ceph_connection *con) -{ - int ret, to = 0; - - dout("read_partial_connect %p at %d\n", con, con->in_base_pos); - - ret = read_partial(con, &to, sizeof(con->in_reply), &con->in_reply); - if (ret <= 0) - goto out; - ret = read_partial(con, &to, le32_to_cpu(con->in_reply.authorizer_len), - con->auth_reply_buf); - if (ret <= 0) - goto out; - - dout("read_partial_connect %p tag %d, con_seq = %u, g_seq = %u\n", - con, (int)con->in_reply.tag, - le32_to_cpu(con->in_reply.connect_seq), - le32_to_cpu(con->in_reply.global_seq)); -out: - return ret; - -} - -/* - * Verify the hello banner looks okay. - */ -static int verify_hello(struct ceph_connection *con) -{ - if (memcmp(con->in_banner, CEPH_BANNER, strlen(CEPH_BANNER))) { - pr_err("connect to %s got bad banner\n", - pr_addr(&con->peer_addr.in_addr)); - con->error_msg = "protocol error, bad banner"; - return -1; - } - return 0; -} - -static bool addr_is_blank(struct sockaddr_storage *ss) -{ - switch (ss->ss_family) { - case AF_INET: - return ((struct sockaddr_in *)ss)->sin_addr.s_addr == 0; - case AF_INET6: - return - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[0] == 0 && - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[1] == 0 && - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[2] == 0 && - ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[3] == 0; - } - return false; -} - -static int addr_port(struct sockaddr_storage *ss) -{ - switch (ss->ss_family) { - case AF_INET: - return ntohs(((struct sockaddr_in *)ss)->sin_port); - case AF_INET6: - return ntohs(((struct sockaddr_in6 *)ss)->sin6_port); - } - return 0; -} - -static void addr_set_port(struct sockaddr_storage *ss, int p) -{ - switch (ss->ss_family) { - case AF_INET: - ((struct sockaddr_in *)ss)->sin_port = htons(p); - case AF_INET6: - ((struct sockaddr_in6 *)ss)->sin6_port = htons(p); - } -} - -/* - * Parse an ip[:port] list into an addr array. Use the default - * monitor port if a port isn't specified. - */ -int ceph_parse_ips(const char *c, const char *end, - struct ceph_entity_addr *addr, - int max_count, int *count) -{ - int i; - const char *p = c; - - dout("parse_ips on '%.*s'\n", (int)(end-c), c); - for (i = 0; i < max_count; i++) { - const char *ipend; - struct sockaddr_storage *ss = &addr[i].in_addr; - struct sockaddr_in *in4 = (void *)ss; - struct sockaddr_in6 *in6 = (void *)ss; - int port; - char delim = ','; - - if (*p == '[') { - delim = ']'; - p++; - } - - memset(ss, 0, sizeof(*ss)); - if (in4_pton(p, end - p, (u8 *)&in4->sin_addr.s_addr, - delim, &ipend)) - ss->ss_family = AF_INET; - else if (in6_pton(p, end - p, (u8 *)&in6->sin6_addr.s6_addr, - delim, &ipend)) - ss->ss_family = AF_INET6; - else - goto bad; - p = ipend; - - if (delim == ']') { - if (*p != ']') { - dout("missing matching ']'\n"); - goto bad; - } - p++; - } - - /* port? */ - if (p < end && *p == ':') { - port = 0; - p++; - while (p < end && *p >= '0' && *p <= '9') { - port = (port * 10) + (*p - '0'); - p++; - } - if (port > 65535 || port == 0) - goto bad; - } else { - port = CEPH_MON_PORT; - } - - addr_set_port(ss, port); - - dout("parse_ips got %s\n", pr_addr(ss)); - - if (p == end) - break; - if (*p != ',') - goto bad; - p++; - } - - if (p != end) - goto bad; - - if (count) - *count = i + 1; - return 0; - -bad: - pr_err("parse_ips bad ip '%.*s'\n", (int)(end - c), c); - return -EINVAL; -} - -static int process_banner(struct ceph_connection *con) -{ - dout("process_banner on %p\n", con); - - if (verify_hello(con) < 0) - return -1; - - ceph_decode_addr(&con->actual_peer_addr); - ceph_decode_addr(&con->peer_addr_for_me); - - /* - * Make sure the other end is who we wanted. note that the other - * end may not yet know their ip address, so if it's 0.0.0.0, give - * them the benefit of the doubt. - */ - if (memcmp(&con->peer_addr, &con->actual_peer_addr, - sizeof(con->peer_addr)) != 0 && - !(addr_is_blank(&con->actual_peer_addr.in_addr) && - con->actual_peer_addr.nonce == con->peer_addr.nonce)) { - pr_warning("wrong peer, want %s/%d, got %s/%d\n", - pr_addr(&con->peer_addr.in_addr), - (int)le32_to_cpu(con->peer_addr.nonce), - pr_addr(&con->actual_peer_addr.in_addr), - (int)le32_to_cpu(con->actual_peer_addr.nonce)); - con->error_msg = "wrong peer at address"; - return -1; - } - - /* - * did we learn our address? - */ - if (addr_is_blank(&con->msgr->inst.addr.in_addr)) { - int port = addr_port(&con->msgr->inst.addr.in_addr); - - memcpy(&con->msgr->inst.addr.in_addr, - &con->peer_addr_for_me.in_addr, - sizeof(con->peer_addr_for_me.in_addr)); - addr_set_port(&con->msgr->inst.addr.in_addr, port); - encode_my_addr(con->msgr); - dout("process_banner learned my addr is %s\n", - pr_addr(&con->msgr->inst.addr.in_addr)); - } - - set_bit(NEGOTIATING, &con->state); - prepare_read_connect(con); - return 0; -} - -static void fail_protocol(struct ceph_connection *con) -{ - reset_connection(con); - set_bit(CLOSED, &con->state); /* in case there's queued work */ - - mutex_unlock(&con->mutex); - if (con->ops->bad_proto) - con->ops->bad_proto(con); - mutex_lock(&con->mutex); -} - -static int process_connect(struct ceph_connection *con) -{ - u64 sup_feat = CEPH_FEATURE_SUPPORTED; - u64 req_feat = CEPH_FEATURE_REQUIRED; - u64 server_feat = le64_to_cpu(con->in_reply.features); - - dout("process_connect on %p tag %d\n", con, (int)con->in_tag); - - switch (con->in_reply.tag) { - case CEPH_MSGR_TAG_FEATURES: - pr_err("%s%lld %s feature set mismatch," - " my %llx < server's %llx, missing %llx\n", - ENTITY_NAME(con->peer_name), - pr_addr(&con->peer_addr.in_addr), - sup_feat, server_feat, server_feat & ~sup_feat); - con->error_msg = "missing required protocol features"; - fail_protocol(con); - return -1; - - case CEPH_MSGR_TAG_BADPROTOVER: - pr_err("%s%lld %s protocol version mismatch," - " my %d != server's %d\n", - ENTITY_NAME(con->peer_name), - pr_addr(&con->peer_addr.in_addr), - le32_to_cpu(con->out_connect.protocol_version), - le32_to_cpu(con->in_reply.protocol_version)); - con->error_msg = "protocol version mismatch"; - fail_protocol(con); - return -1; - - case CEPH_MSGR_TAG_BADAUTHORIZER: - con->auth_retry++; - dout("process_connect %p got BADAUTHORIZER attempt %d\n", con, - con->auth_retry); - if (con->auth_retry == 2) { - con->error_msg = "connect authorization failure"; - reset_connection(con); - set_bit(CLOSED, &con->state); - return -1; - } - con->auth_retry = 1; - prepare_write_connect(con->msgr, con, 0); - prepare_read_connect(con); - break; - - case CEPH_MSGR_TAG_RESETSESSION: - /* - * If we connected with a large connect_seq but the peer - * has no record of a session with us (no connection, or - * connect_seq == 0), they will send RESETSESION to indicate - * that they must have reset their session, and may have - * dropped messages. - */ - dout("process_connect got RESET peer seq %u\n", - le32_to_cpu(con->in_connect.connect_seq)); - pr_err("%s%lld %s connection reset\n", - ENTITY_NAME(con->peer_name), - pr_addr(&con->peer_addr.in_addr)); - reset_connection(con); - prepare_write_connect(con->msgr, con, 0); - prepare_read_connect(con); - - /* Tell ceph about it. */ - mutex_unlock(&con->mutex); - pr_info("reset on %s%lld\n", ENTITY_NAME(con->peer_name)); - if (con->ops->peer_reset) - con->ops->peer_reset(con); - mutex_lock(&con->mutex); - break; - - case CEPH_MSGR_TAG_RETRY_SESSION: - /* - * If we sent a smaller connect_seq than the peer has, try - * again with a larger value. - */ - dout("process_connect got RETRY my seq = %u, peer_seq = %u\n", - le32_to_cpu(con->out_connect.connect_seq), - le32_to_cpu(con->in_connect.connect_seq)); - con->connect_seq = le32_to_cpu(con->in_connect.connect_seq); - prepare_write_connect(con->msgr, con, 0); - prepare_read_connect(con); - break; - - case CEPH_MSGR_TAG_RETRY_GLOBAL: - /* - * If we sent a smaller global_seq than the peer has, try - * again with a larger value. - */ - dout("process_connect got RETRY_GLOBAL my %u peer_gseq %u\n", - con->peer_global_seq, - le32_to_cpu(con->in_connect.global_seq)); - get_global_seq(con->msgr, - le32_to_cpu(con->in_connect.global_seq)); - prepare_write_connect(con->msgr, con, 0); - prepare_read_connect(con); - break; - - case CEPH_MSGR_TAG_READY: - if (req_feat & ~server_feat) { - pr_err("%s%lld %s protocol feature mismatch," - " my required %llx > server's %llx, need %llx\n", - ENTITY_NAME(con->peer_name), - pr_addr(&con->peer_addr.in_addr), - req_feat, server_feat, req_feat & ~server_feat); - con->error_msg = "missing required protocol features"; - fail_protocol(con); - return -1; - } - clear_bit(CONNECTING, &con->state); - con->peer_global_seq = le32_to_cpu(con->in_reply.global_seq); - con->connect_seq++; - con->peer_features = server_feat; - dout("process_connect got READY gseq %d cseq %d (%d)\n", - con->peer_global_seq, - le32_to_cpu(con->in_reply.connect_seq), - con->connect_seq); - WARN_ON(con->connect_seq != - le32_to_cpu(con->in_reply.connect_seq)); - - if (con->in_reply.flags & CEPH_MSG_CONNECT_LOSSY) - set_bit(LOSSYTX, &con->state); - - prepare_read_tag(con); - break; - - case CEPH_MSGR_TAG_WAIT: - /* - * If there is a connection race (we are opening - * connections to each other), one of us may just have - * to WAIT. This shouldn't happen if we are the - * client. - */ - pr_err("process_connect peer connecting WAIT\n"); - - default: - pr_err("connect protocol error, will retry\n"); - con->error_msg = "protocol error, garbage tag during connect"; - return -1; - } - return 0; -} - - -/* - * read (part of) an ack - */ -static int read_partial_ack(struct ceph_connection *con) -{ - int to = 0; - - return read_partial(con, &to, sizeof(con->in_temp_ack), - &con->in_temp_ack); -} - - -/* - * We can finally discard anything that's been acked. - */ -static void process_ack(struct ceph_connection *con) -{ - struct ceph_msg *m; - u64 ack = le64_to_cpu(con->in_temp_ack); - u64 seq; - - while (!list_empty(&con->out_sent)) { - m = list_first_entry(&con->out_sent, struct ceph_msg, - list_head); - seq = le64_to_cpu(m->hdr.seq); - if (seq > ack) - break; - dout("got ack for seq %llu type %d at %p\n", seq, - le16_to_cpu(m->hdr.type), m); - ceph_msg_remove(m); - } - prepare_read_tag(con); -} - - - - -static int read_partial_message_section(struct ceph_connection *con, - struct kvec *section, - unsigned int sec_len, u32 *crc) -{ - int ret, left; - - BUG_ON(!section); - - while (section->iov_len < sec_len) { - BUG_ON(section->iov_base == NULL); - left = sec_len - section->iov_len; - ret = ceph_tcp_recvmsg(con->sock, (char *)section->iov_base + - section->iov_len, left); - if (ret <= 0) - return ret; - section->iov_len += ret; - if (section->iov_len == sec_len) - *crc = crc32c(0, section->iov_base, - section->iov_len); - } - - return 1; -} - -static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con, - struct ceph_msg_header *hdr, - int *skip); - - -static int read_partial_message_pages(struct ceph_connection *con, - struct page **pages, - unsigned data_len, int datacrc) -{ - void *p; - int ret; - int left; - - left = min((int)(data_len - con->in_msg_pos.data_pos), - (int)(PAGE_SIZE - con->in_msg_pos.page_pos)); - /* (page) data */ - BUG_ON(pages == NULL); - p = kmap(pages[con->in_msg_pos.page]); - ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos, - left); - if (ret > 0 && datacrc) - con->in_data_crc = - crc32c(con->in_data_crc, - p + con->in_msg_pos.page_pos, ret); - kunmap(pages[con->in_msg_pos.page]); - if (ret <= 0) - return ret; - con->in_msg_pos.data_pos += ret; - con->in_msg_pos.page_pos += ret; - if (con->in_msg_pos.page_pos == PAGE_SIZE) { - con->in_msg_pos.page_pos = 0; - con->in_msg_pos.page++; - } - - return ret; -} - -#ifdef CONFIG_BLOCK -static int read_partial_message_bio(struct ceph_connection *con, - struct bio **bio_iter, int *bio_seg, - unsigned data_len, int datacrc) -{ - struct bio_vec *bv = bio_iovec_idx(*bio_iter, *bio_seg); - void *p; - int ret, left; - - if (IS_ERR(bv)) - return PTR_ERR(bv); - - left = min((int)(data_len - con->in_msg_pos.data_pos), - (int)(bv->bv_len - con->in_msg_pos.page_pos)); - - p = kmap(bv->bv_page) + bv->bv_offset; - - ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos, - left); - if (ret > 0 && datacrc) - con->in_data_crc = - crc32c(con->in_data_crc, - p + con->in_msg_pos.page_pos, ret); - kunmap(bv->bv_page); - if (ret <= 0) - return ret; - con->in_msg_pos.data_pos += ret; - con->in_msg_pos.page_pos += ret; - if (con->in_msg_pos.page_pos == bv->bv_len) { - con->in_msg_pos.page_pos = 0; - iter_bio_next(bio_iter, bio_seg); - } - - return ret; -} -#endif - -/* - * read (part of) a message. - */ -static int read_partial_message(struct ceph_connection *con) -{ - struct ceph_msg *m = con->in_msg; - int ret; - int to, left; - unsigned front_len, middle_len, data_len, data_off; - int datacrc = con->msgr->nocrc; - int skip; - u64 seq; - - dout("read_partial_message con %p msg %p\n", con, m); - - /* header */ - while (con->in_base_pos < sizeof(con->in_hdr)) { - left = sizeof(con->in_hdr) - con->in_base_pos; - ret = ceph_tcp_recvmsg(con->sock, - (char *)&con->in_hdr + con->in_base_pos, - left); - if (ret <= 0) - return ret; - con->in_base_pos += ret; - if (con->in_base_pos == sizeof(con->in_hdr)) { - u32 crc = crc32c(0, (void *)&con->in_hdr, - sizeof(con->in_hdr) - sizeof(con->in_hdr.crc)); - if (crc != le32_to_cpu(con->in_hdr.crc)) { - pr_err("read_partial_message bad hdr " - " crc %u != expected %u\n", - crc, con->in_hdr.crc); - return -EBADMSG; - } - } - } - front_len = le32_to_cpu(con->in_hdr.front_len); - if (front_len > CEPH_MSG_MAX_FRONT_LEN) - return -EIO; - middle_len = le32_to_cpu(con->in_hdr.middle_len); - if (middle_len > CEPH_MSG_MAX_DATA_LEN) - return -EIO; - data_len = le32_to_cpu(con->in_hdr.data_len); - if (data_len > CEPH_MSG_MAX_DATA_LEN) - return -EIO; - data_off = le16_to_cpu(con->in_hdr.data_off); - - /* verify seq# */ - seq = le64_to_cpu(con->in_hdr.seq); - if ((s64)seq - (s64)con->in_seq < 1) { - pr_info("skipping %s%lld %s seq %lld, expected %lld\n", - ENTITY_NAME(con->peer_name), - pr_addr(&con->peer_addr.in_addr), - seq, con->in_seq + 1); - con->in_base_pos = -front_len - middle_len - data_len - - sizeof(m->footer); - con->in_tag = CEPH_MSGR_TAG_READY; - con->in_seq++; - return 0; - } else if ((s64)seq - (s64)con->in_seq > 1) { - pr_err("read_partial_message bad seq %lld expected %lld\n", - seq, con->in_seq + 1); - con->error_msg = "bad message sequence # for incoming message"; - return -EBADMSG; - } - - /* allocate message? */ - if (!con->in_msg) { - dout("got hdr type %d front %d data %d\n", con->in_hdr.type, - con->in_hdr.front_len, con->in_hdr.data_len); - skip = 0; - con->in_msg = ceph_alloc_msg(con, &con->in_hdr, &skip); - if (skip) { - /* skip this message */ - dout("alloc_msg said skip message\n"); - BUG_ON(con->in_msg); - con->in_base_pos = -front_len - middle_len - data_len - - sizeof(m->footer); - con->in_tag = CEPH_MSGR_TAG_READY; - con->in_seq++; - return 0; - } - if (!con->in_msg) { - con->error_msg = - "error allocating memory for incoming message"; - return -ENOMEM; - } - m = con->in_msg; - m->front.iov_len = 0; /* haven't read it yet */ - if (m->middle) - m->middle->vec.iov_len = 0; - - con->in_msg_pos.page = 0; - if (m->pages) - con->in_msg_pos.page_pos = data_off & ~PAGE_MASK; - else - con->in_msg_pos.page_pos = 0; - con->in_msg_pos.data_pos = 0; - } - - /* front */ - ret = read_partial_message_section(con, &m->front, front_len, - &con->in_front_crc); - if (ret <= 0) - return ret; - - /* middle */ - if (m->middle) { - ret = read_partial_message_section(con, &m->middle->vec, - middle_len, - &con->in_middle_crc); - if (ret <= 0) - return ret; - } -#ifdef CONFIG_BLOCK - if (m->bio && !m->bio_iter) - init_bio_iter(m->bio, &m->bio_iter, &m->bio_seg); -#endif - - /* (page) data */ - while (con->in_msg_pos.data_pos < data_len) { - if (m->pages) { - ret = read_partial_message_pages(con, m->pages, - data_len, datacrc); - if (ret <= 0) - return ret; -#ifdef CONFIG_BLOCK - } else if (m->bio) { - - ret = read_partial_message_bio(con, - &m->bio_iter, &m->bio_seg, - data_len, datacrc); - if (ret <= 0) - return ret; -#endif - } else { - BUG_ON(1); - } - } - - /* footer */ - to = sizeof(m->hdr) + sizeof(m->footer); - while (con->in_base_pos < to) { - left = to - con->in_base_pos; - ret = ceph_tcp_recvmsg(con->sock, (char *)&m->footer + - (con->in_base_pos - sizeof(m->hdr)), - left); - if (ret <= 0) - return ret; - con->in_base_pos += ret; - } - dout("read_partial_message got msg %p %d (%u) + %d (%u) + %d (%u)\n", - m, front_len, m->footer.front_crc, middle_len, - m->footer.middle_crc, data_len, m->footer.data_crc); - - /* crc ok? */ - if (con->in_front_crc != le32_to_cpu(m->footer.front_crc)) { - pr_err("read_partial_message %p front crc %u != exp. %u\n", - m, con->in_front_crc, m->footer.front_crc); - return -EBADMSG; - } - if (con->in_middle_crc != le32_to_cpu(m->footer.middle_crc)) { - pr_err("read_partial_message %p middle crc %u != exp %u\n", - m, con->in_middle_crc, m->footer.middle_crc); - return -EBADMSG; - } - if (datacrc && - (m->footer.flags & CEPH_MSG_FOOTER_NOCRC) == 0 && - con->in_data_crc != le32_to_cpu(m->footer.data_crc)) { - pr_err("read_partial_message %p data crc %u != exp. %u\n", m, - con->in_data_crc, le32_to_cpu(m->footer.data_crc)); - return -EBADMSG; - } - - return 1; /* done! */ -} - -/* - * Process message. This happens in the worker thread. The callback should - * be careful not to do anything that waits on other incoming messages or it - * may deadlock. - */ -static void process_message(struct ceph_connection *con) -{ - struct ceph_msg *msg; - - msg = con->in_msg; - con->in_msg = NULL; - - /* if first message, set peer_name */ - if (con->peer_name.type == 0) - con->peer_name = msg->hdr.src; - - con->in_seq++; - mutex_unlock(&con->mutex); - - dout("===== %p %llu from %s%lld %d=%s len %d+%d (%u %u %u) =====\n", - msg, le64_to_cpu(msg->hdr.seq), - ENTITY_NAME(msg->hdr.src), - le16_to_cpu(msg->hdr.type), - ceph_msg_type_name(le16_to_cpu(msg->hdr.type)), - le32_to_cpu(msg->hdr.front_len), - le32_to_cpu(msg->hdr.data_len), - con->in_front_crc, con->in_middle_crc, con->in_data_crc); - con->ops->dispatch(con, msg); - - mutex_lock(&con->mutex); - prepare_read_tag(con); -} - - -/* - * Write something to the socket. Called in a worker thread when the - * socket appears to be writeable and we have something ready to send. - */ -static int try_write(struct ceph_connection *con) -{ - struct ceph_messenger *msgr = con->msgr; - int ret = 1; - - dout("try_write start %p state %lu nref %d\n", con, con->state, - atomic_read(&con->nref)); - -more: - dout("try_write out_kvec_bytes %d\n", con->out_kvec_bytes); - - /* open the socket first? */ - if (con->sock == NULL) { - /* - * if we were STANDBY and are reconnecting _this_ - * connection, bump connect_seq now. Always bump - * global_seq. - */ - if (test_and_clear_bit(STANDBY, &con->state)) - con->connect_seq++; - - prepare_write_banner(msgr, con); - prepare_write_connect(msgr, con, 1); - prepare_read_banner(con); - set_bit(CONNECTING, &con->state); - clear_bit(NEGOTIATING, &con->state); - - BUG_ON(con->in_msg); - con->in_tag = CEPH_MSGR_TAG_READY; - dout("try_write initiating connect on %p new state %lu\n", - con, con->state); - con->sock = ceph_tcp_connect(con); - if (IS_ERR(con->sock)) { - con->sock = NULL; - con->error_msg = "connect error"; - ret = -1; - goto out; - } - } - -more_kvec: - /* kvec data queued? */ - if (con->out_skip) { - ret = write_partial_skip(con); - if (ret <= 0) - goto done; - if (ret < 0) { - dout("try_write write_partial_skip err %d\n", ret); - goto done; - } - } - if (con->out_kvec_left) { - ret = write_partial_kvec(con); - if (ret <= 0) - goto done; - } - - /* msg pages? */ - if (con->out_msg) { - if (con->out_msg_done) { - ceph_msg_put(con->out_msg); - con->out_msg = NULL; /* we're done with this one */ - goto do_next; - } - - ret = write_partial_msg_pages(con); - if (ret == 1) - goto more_kvec; /* we need to send the footer, too! */ - if (ret == 0) - goto done; - if (ret < 0) { - dout("try_write write_partial_msg_pages err %d\n", - ret); - goto done; - } - } - -do_next: - if (!test_bit(CONNECTING, &con->state)) { - /* is anything else pending? */ - if (!list_empty(&con->out_queue)) { - prepare_write_message(con); - goto more; - } - if (con->in_seq > con->in_seq_acked) { - prepare_write_ack(con); - goto more; - } - if (test_and_clear_bit(KEEPALIVE_PENDING, &con->state)) { - prepare_write_keepalive(con); - goto more; - } - } - - /* Nothing to do! */ - clear_bit(WRITE_PENDING, &con->state); - dout("try_write nothing else to write.\n"); -done: - ret = 0; -out: - dout("try_write done on %p\n", con); - return ret; -} - - - -/* - * Read what we can from the socket. - */ -static int try_read(struct ceph_connection *con) -{ - int ret = -1; - - if (!con->sock) - return 0; - - if (test_bit(STANDBY, &con->state)) - return 0; - - dout("try_read start on %p\n", con); - -more: - dout("try_read tag %d in_base_pos %d\n", (int)con->in_tag, - con->in_base_pos); - if (test_bit(CONNECTING, &con->state)) { - if (!test_bit(NEGOTIATING, &con->state)) { - dout("try_read connecting\n"); - ret = read_partial_banner(con); - if (ret <= 0) - goto done; - if (process_banner(con) < 0) { - ret = -1; - goto out; - } - } - ret = read_partial_connect(con); - if (ret <= 0) - goto done; - if (process_connect(con) < 0) { - ret = -1; - goto out; - } - goto more; - } - - if (con->in_base_pos < 0) { - /* - * skipping + discarding content. - * - * FIXME: there must be a better way to do this! - */ - static char buf[1024]; - int skip = min(1024, -con->in_base_pos); - dout("skipping %d / %d bytes\n", skip, -con->in_base_pos); - ret = ceph_tcp_recvmsg(con->sock, buf, skip); - if (ret <= 0) - goto done; - con->in_base_pos += ret; - if (con->in_base_pos) - goto more; - } - if (con->in_tag == CEPH_MSGR_TAG_READY) { - /* - * what's next? - */ - ret = ceph_tcp_recvmsg(con->sock, &con->in_tag, 1); - if (ret <= 0) - goto done; - dout("try_read got tag %d\n", (int)con->in_tag); - switch (con->in_tag) { - case CEPH_MSGR_TAG_MSG: - prepare_read_message(con); - break; - case CEPH_MSGR_TAG_ACK: - prepare_read_ack(con); - break; - case CEPH_MSGR_TAG_CLOSE: - set_bit(CLOSED, &con->state); /* fixme */ - goto done; - default: - goto bad_tag; - } - } - if (con->in_tag == CEPH_MSGR_TAG_MSG) { - ret = read_partial_message(con); - if (ret <= 0) { - switch (ret) { - case -EBADMSG: - con->error_msg = "bad crc"; - ret = -EIO; - goto out; - case -EIO: - con->error_msg = "io error"; - goto out; - default: - goto done; - } - } - if (con->in_tag == CEPH_MSGR_TAG_READY) - goto more; - process_message(con); - goto more; - } - if (con->in_tag == CEPH_MSGR_TAG_ACK) { - ret = read_partial_ack(con); - if (ret <= 0) - goto done; - process_ack(con); - goto more; - } - -done: - ret = 0; -out: - dout("try_read done on %p\n", con); - return ret; - -bad_tag: - pr_err("try_read bad con->in_tag = %d\n", (int)con->in_tag); - con->error_msg = "protocol error, garbage tag"; - ret = -1; - goto out; -} - - -/* - * Atomically queue work on a connection. Bump @con reference to - * avoid races with connection teardown. - * - * There is some trickery going on with QUEUED and BUSY because we - * only want a _single_ thread operating on each connection at any - * point in time, but we want to use all available CPUs. - * - * The worker thread only proceeds if it can atomically set BUSY. It - * clears QUEUED and does it's thing. When it thinks it's done, it - * clears BUSY, then rechecks QUEUED.. if it's set again, it loops - * (tries again to set BUSY). - * - * To queue work, we first set QUEUED, _then_ if BUSY isn't set, we - * try to queue work. If that fails (work is already queued, or BUSY) - * we give up (work also already being done or is queued) but leave QUEUED - * set so that the worker thread will loop if necessary. - */ -static void queue_con(struct ceph_connection *con) -{ - if (test_bit(DEAD, &con->state)) { - dout("queue_con %p ignoring: DEAD\n", - con); - return; - } - - if (!con->ops->get(con)) { - dout("queue_con %p ref count 0\n", con); - return; - } - - set_bit(QUEUED, &con->state); - if (test_bit(BUSY, &con->state)) { - dout("queue_con %p - already BUSY\n", con); - con->ops->put(con); - } else if (!queue_work(ceph_msgr_wq, &con->work.work)) { - dout("queue_con %p - already queued\n", con); - con->ops->put(con); - } else { - dout("queue_con %p\n", con); - } -} - -/* - * Do some work on a connection. Drop a connection ref when we're done. - */ -static void con_work(struct work_struct *work) -{ - struct ceph_connection *con = container_of(work, struct ceph_connection, - work.work); - int backoff = 0; - -more: - if (test_and_set_bit(BUSY, &con->state) != 0) { - dout("con_work %p BUSY already set\n", con); - goto out; - } - dout("con_work %p start, clearing QUEUED\n", con); - clear_bit(QUEUED, &con->state); - - mutex_lock(&con->mutex); - - if (test_bit(CLOSED, &con->state)) { /* e.g. if we are replaced */ - dout("con_work CLOSED\n"); - con_close_socket(con); - goto done; - } - if (test_and_clear_bit(OPENING, &con->state)) { - /* reopen w/ new peer */ - dout("con_work OPENING\n"); - con_close_socket(con); - } - - if (test_and_clear_bit(SOCK_CLOSED, &con->state) || - try_read(con) < 0 || - try_write(con) < 0) { - mutex_unlock(&con->mutex); - backoff = 1; - ceph_fault(con); /* error/fault path */ - goto done_unlocked; - } - -done: - mutex_unlock(&con->mutex); - -done_unlocked: - clear_bit(BUSY, &con->state); - dout("con->state=%lu\n", con->state); - if (test_bit(QUEUED, &con->state)) { - if (!backoff || test_bit(OPENING, &con->state)) { - dout("con_work %p QUEUED reset, looping\n", con); - goto more; - } - dout("con_work %p QUEUED reset, but just faulted\n", con); - clear_bit(QUEUED, &con->state); - } - dout("con_work %p done\n", con); - -out: - con->ops->put(con); -} - - -/* - * Generic error/fault handler. A retry mechanism is used with - * exponential backoff - */ -static void ceph_fault(struct ceph_connection *con) -{ - pr_err("%s%lld %s %s\n", ENTITY_NAME(con->peer_name), - pr_addr(&con->peer_addr.in_addr), con->error_msg); - dout("fault %p state %lu to peer %s\n", - con, con->state, pr_addr(&con->peer_addr.in_addr)); - - if (test_bit(LOSSYTX, &con->state)) { - dout("fault on LOSSYTX channel\n"); - goto out; - } - - mutex_lock(&con->mutex); - if (test_bit(CLOSED, &con->state)) - goto out_unlock; - - con_close_socket(con); - - if (con->in_msg) { - ceph_msg_put(con->in_msg); - con->in_msg = NULL; - } - - /* Requeue anything that hasn't been acked */ - list_splice_init(&con->out_sent, &con->out_queue); - - /* If there are no messages in the queue, place the connection - * in a STANDBY state (i.e., don't try to reconnect just yet). */ - if (list_empty(&con->out_queue) && !con->out_keepalive_pending) { - dout("fault setting STANDBY\n"); - set_bit(STANDBY, &con->state); - } else { - /* retry after a delay. */ - if (con->delay == 0) - con->delay = BASE_DELAY_INTERVAL; - else if (con->delay < MAX_DELAY_INTERVAL) - con->delay *= 2; - dout("fault queueing %p delay %lu\n", con, con->delay); - con->ops->get(con); - if (queue_delayed_work(ceph_msgr_wq, &con->work, - round_jiffies_relative(con->delay)) == 0) - con->ops->put(con); - } - -out_unlock: - mutex_unlock(&con->mutex); -out: - /* - * in case we faulted due to authentication, invalidate our - * current tickets so that we can get new ones. - */ - if (con->auth_retry && con->ops->invalidate_authorizer) { - dout("calling invalidate_authorizer()\n"); - con->ops->invalidate_authorizer(con); - } - - if (con->ops->fault) - con->ops->fault(con); -} - - - -/* - * create a new messenger instance - */ -struct ceph_messenger *ceph_messenger_create(struct ceph_entity_addr *myaddr) -{ - struct ceph_messenger *msgr; - - msgr = kzalloc(sizeof(*msgr), GFP_KERNEL); - if (msgr == NULL) - return ERR_PTR(-ENOMEM); - - spin_lock_init(&msgr->global_seq_lock); - - /* the zero page is needed if a request is "canceled" while the message - * is being written over the socket */ - msgr->zero_page = __page_cache_alloc(GFP_KERNEL | __GFP_ZERO); - if (!msgr->zero_page) { - kfree(msgr); - return ERR_PTR(-ENOMEM); - } - kmap(msgr->zero_page); - - if (myaddr) - msgr->inst.addr = *myaddr; - - /* select a random nonce */ - msgr->inst.addr.type = 0; - get_random_bytes(&msgr->inst.addr.nonce, sizeof(msgr->inst.addr.nonce)); - encode_my_addr(msgr); - - dout("messenger_create %p\n", msgr); - return msgr; -} - -void ceph_messenger_destroy(struct ceph_messenger *msgr) -{ - dout("destroy %p\n", msgr); - kunmap(msgr->zero_page); - __free_page(msgr->zero_page); - kfree(msgr); - dout("destroyed messenger %p\n", msgr); -} - -/* - * Queue up an outgoing message on the given connection. - */ -void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg) -{ - if (test_bit(CLOSED, &con->state)) { - dout("con_send %p closed, dropping %p\n", con, msg); - ceph_msg_put(msg); - return; - } - - /* set src+dst */ - msg->hdr.src = con->msgr->inst.name; - - BUG_ON(msg->front.iov_len != le32_to_cpu(msg->hdr.front_len)); - - msg->needs_out_seq = true; - - /* queue */ - mutex_lock(&con->mutex); - BUG_ON(!list_empty(&msg->list_head)); - list_add_tail(&msg->list_head, &con->out_queue); - dout("----- %p to %s%lld %d=%s len %d+%d+%d -----\n", msg, - ENTITY_NAME(con->peer_name), le16_to_cpu(msg->hdr.type), - ceph_msg_type_name(le16_to_cpu(msg->hdr.type)), - le32_to_cpu(msg->hdr.front_len), - le32_to_cpu(msg->hdr.middle_len), - le32_to_cpu(msg->hdr.data_len)); - mutex_unlock(&con->mutex); - - /* if there wasn't anything waiting to send before, queue - * new work */ - if (test_and_set_bit(WRITE_PENDING, &con->state) == 0) - queue_con(con); -} - -/* - * Revoke a message that was previously queued for send - */ -void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg) -{ - mutex_lock(&con->mutex); - if (!list_empty(&msg->list_head)) { - dout("con_revoke %p msg %p - was on queue\n", con, msg); - list_del_init(&msg->list_head); - ceph_msg_put(msg); - msg->hdr.seq = 0; - } - if (con->out_msg == msg) { - dout("con_revoke %p msg %p - was sending\n", con, msg); - con->out_msg = NULL; - if (con->out_kvec_is_msg) { - con->out_skip = con->out_kvec_bytes; - con->out_kvec_is_msg = false; - } - ceph_msg_put(msg); - msg->hdr.seq = 0; - } - mutex_unlock(&con->mutex); -} - -/* - * Revoke a message that we may be reading data into - */ -void ceph_con_revoke_message(struct ceph_connection *con, struct ceph_msg *msg) -{ - mutex_lock(&con->mutex); - if (con->in_msg && con->in_msg == msg) { - unsigned front_len = le32_to_cpu(con->in_hdr.front_len); - unsigned middle_len = le32_to_cpu(con->in_hdr.middle_len); - unsigned data_len = le32_to_cpu(con->in_hdr.data_len); - - /* skip rest of message */ - dout("con_revoke_pages %p msg %p revoked\n", con, msg); - con->in_base_pos = con->in_base_pos - - sizeof(struct ceph_msg_header) - - front_len - - middle_len - - data_len - - sizeof(struct ceph_msg_footer); - ceph_msg_put(con->in_msg); - con->in_msg = NULL; - con->in_tag = CEPH_MSGR_TAG_READY; - con->in_seq++; - } else { - dout("con_revoke_pages %p msg %p pages %p no-op\n", - con, con->in_msg, msg); - } - mutex_unlock(&con->mutex); -} - -/* - * Queue a keepalive byte to ensure the tcp connection is alive. - */ -void ceph_con_keepalive(struct ceph_connection *con) -{ - if (test_and_set_bit(KEEPALIVE_PENDING, &con->state) == 0 && - test_and_set_bit(WRITE_PENDING, &con->state) == 0) - queue_con(con); -} - - -/* - * construct a new message with given type, size - * the new msg has a ref count of 1. - */ -struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags) -{ - struct ceph_msg *m; - - m = kmalloc(sizeof(*m), flags); - if (m == NULL) - goto out; - kref_init(&m->kref); - INIT_LIST_HEAD(&m->list_head); - - m->hdr.tid = 0; - m->hdr.type = cpu_to_le16(type); - m->hdr.priority = cpu_to_le16(CEPH_MSG_PRIO_DEFAULT); - m->hdr.version = 0; - m->hdr.front_len = cpu_to_le32(front_len); - m->hdr.middle_len = 0; - m->hdr.data_len = 0; - m->hdr.data_off = 0; - m->hdr.reserved = 0; - m->footer.front_crc = 0; - m->footer.middle_crc = 0; - m->footer.data_crc = 0; - m->footer.flags = 0; - m->front_max = front_len; - m->front_is_vmalloc = false; - m->more_to_follow = false; - m->pool = NULL; - - /* front */ - if (front_len) { - if (front_len > PAGE_CACHE_SIZE) { - m->front.iov_base = __vmalloc(front_len, flags, - PAGE_KERNEL); - m->front_is_vmalloc = true; - } else { - m->front.iov_base = kmalloc(front_len, flags); - } - if (m->front.iov_base == NULL) { - pr_err("msg_new can't allocate %d bytes\n", - front_len); - goto out2; - } - } else { - m->front.iov_base = NULL; - } - m->front.iov_len = front_len; - - /* middle */ - m->middle = NULL; - - /* data */ - m->nr_pages = 0; - m->pages = NULL; - m->pagelist = NULL; - m->bio = NULL; - m->bio_iter = NULL; - m->bio_seg = 0; - m->trail = NULL; - - dout("ceph_msg_new %p front %d\n", m, front_len); - return m; - -out2: - ceph_msg_put(m); -out: - pr_err("msg_new can't create type %d front %d\n", type, front_len); - return NULL; -} - -/* - * Allocate "middle" portion of a message, if it is needed and wasn't - * allocated by alloc_msg. This allows us to read a small fixed-size - * per-type header in the front and then gracefully fail (i.e., - * propagate the error to the caller based on info in the front) when - * the middle is too large. - */ -static int ceph_alloc_middle(struct ceph_connection *con, struct ceph_msg *msg) -{ - int type = le16_to_cpu(msg->hdr.type); - int middle_len = le32_to_cpu(msg->hdr.middle_len); - - dout("alloc_middle %p type %d %s middle_len %d\n", msg, type, - ceph_msg_type_name(type), middle_len); - BUG_ON(!middle_len); - BUG_ON(msg->middle); - - msg->middle = ceph_buffer_new(middle_len, GFP_NOFS); - if (!msg->middle) - return -ENOMEM; - return 0; -} - -/* - * Generic message allocator, for incoming messages. - */ -static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con, - struct ceph_msg_header *hdr, - int *skip) -{ - int type = le16_to_cpu(hdr->type); - int front_len = le32_to_cpu(hdr->front_len); - int middle_len = le32_to_cpu(hdr->middle_len); - struct ceph_msg *msg = NULL; - int ret; - - if (con->ops->alloc_msg) { - mutex_unlock(&con->mutex); - msg = con->ops->alloc_msg(con, hdr, skip); - mutex_lock(&con->mutex); - if (!msg || *skip) - return NULL; - } - if (!msg) { - *skip = 0; - msg = ceph_msg_new(type, front_len, GFP_NOFS); - if (!msg) { - pr_err("unable to allocate msg type %d len %d\n", - type, front_len); - return NULL; - } - } - memcpy(&msg->hdr, &con->in_hdr, sizeof(con->in_hdr)); - - if (middle_len && !msg->middle) { - ret = ceph_alloc_middle(con, msg); - if (ret < 0) { - ceph_msg_put(msg); - return NULL; - } - } - - return msg; -} - - -/* - * Free a generically kmalloc'd message. - */ -void ceph_msg_kfree(struct ceph_msg *m) -{ - dout("msg_kfree %p\n", m); - if (m->front_is_vmalloc) - vfree(m->front.iov_base); - else - kfree(m->front.iov_base); - kfree(m); -} - -/* - * Drop a msg ref. Destroy as needed. - */ -void ceph_msg_last_put(struct kref *kref) -{ - struct ceph_msg *m = container_of(kref, struct ceph_msg, kref); - - dout("ceph_msg_put last one on %p\n", m); - WARN_ON(!list_empty(&m->list_head)); - - /* drop middle, data, if any */ - if (m->middle) { - ceph_buffer_put(m->middle); - m->middle = NULL; - } - m->nr_pages = 0; - m->pages = NULL; - - if (m->pagelist) { - ceph_pagelist_release(m->pagelist); - kfree(m->pagelist); - m->pagelist = NULL; - } - - m->trail = NULL; - - if (m->pool) - ceph_msgpool_put(m->pool, m); - else - ceph_msg_kfree(m); -} - -void ceph_msg_dump(struct ceph_msg *msg) -{ - pr_debug("msg_dump %p (front_max %d nr_pages %d)\n", msg, - msg->front_max, msg->nr_pages); - print_hex_dump(KERN_DEBUG, "header: ", - DUMP_PREFIX_OFFSET, 16, 1, - &msg->hdr, sizeof(msg->hdr), true); - print_hex_dump(KERN_DEBUG, " front: ", - DUMP_PREFIX_OFFSET, 16, 1, - msg->front.iov_base, msg->front.iov_len, true); - if (msg->middle) - print_hex_dump(KERN_DEBUG, "middle: ", - DUMP_PREFIX_OFFSET, 16, 1, - msg->middle->vec.iov_base, - msg->middle->vec.iov_len, true); - print_hex_dump(KERN_DEBUG, "footer: ", - DUMP_PREFIX_OFFSET, 16, 1, - &msg->footer, sizeof(msg->footer), true); -} diff --git a/fs/ceph/messenger.h b/fs/ceph/messenger.h deleted file mode 100644 index 5a79450..0000000 --- a/fs/ceph/messenger.h +++ /dev/null @@ -1,257 +0,0 @@ -#ifndef __FS_CEPH_MESSENGER_H -#define __FS_CEPH_MESSENGER_H - -#include -#include -#include -#include -#include -#include -#include - -#include "types.h" -#include "buffer.h" - -struct ceph_msg; -struct ceph_connection; - -extern struct workqueue_struct *ceph_msgr_wq; /* receive work queue */ - -/* - * Ceph defines these callbacks for handling connection events. - */ -struct ceph_connection_operations { - struct ceph_connection *(*get)(struct ceph_connection *); - void (*put)(struct ceph_connection *); - - /* handle an incoming message. */ - void (*dispatch) (struct ceph_connection *con, struct ceph_msg *m); - - /* authorize an outgoing connection */ - int (*get_authorizer) (struct ceph_connection *con, - void **buf, int *len, int *proto, - void **reply_buf, int *reply_len, int force_new); - int (*verify_authorizer_reply) (struct ceph_connection *con, int len); - int (*invalidate_authorizer)(struct ceph_connection *con); - - /* protocol version mismatch */ - void (*bad_proto) (struct ceph_connection *con); - - /* there was some error on the socket (disconnect, whatever) */ - void (*fault) (struct ceph_connection *con); - - /* a remote host as terminated a message exchange session, and messages - * we sent (or they tried to send us) may be lost. */ - void (*peer_reset) (struct ceph_connection *con); - - struct ceph_msg * (*alloc_msg) (struct ceph_connection *con, - struct ceph_msg_header *hdr, - int *skip); -}; - -/* use format string %s%d */ -#define ENTITY_NAME(n) ceph_entity_type_name((n).type), le64_to_cpu((n).num) - -struct ceph_messenger { - struct ceph_entity_inst inst; /* my name+address */ - struct ceph_entity_addr my_enc_addr; - struct page *zero_page; /* used in certain error cases */ - - bool nocrc; - - /* - * the global_seq counts connections i (attempt to) initiate - * in order to disambiguate certain connect race conditions. - */ - u32 global_seq; - spinlock_t global_seq_lock; -}; - -/* - * a single message. it contains a header (src, dest, message type, etc.), - * footer (crc values, mainly), a "front" message body, and possibly a - * data payload (stored in some number of pages). - */ -struct ceph_msg { - struct ceph_msg_header hdr; /* header */ - struct ceph_msg_footer footer; /* footer */ - struct kvec front; /* unaligned blobs of message */ - struct ceph_buffer *middle; - struct page **pages; /* data payload. NOT OWNER. */ - unsigned nr_pages; /* size of page array */ - struct ceph_pagelist *pagelist; /* instead of pages */ - struct list_head list_head; - struct kref kref; - struct bio *bio; /* instead of pages/pagelist */ - struct bio *bio_iter; /* bio iterator */ - int bio_seg; /* current bio segment */ - struct ceph_pagelist *trail; /* the trailing part of the data */ - bool front_is_vmalloc; - bool more_to_follow; - bool needs_out_seq; - int front_max; - - struct ceph_msgpool *pool; -}; - -struct ceph_msg_pos { - int page, page_pos; /* which page; offset in page */ - int data_pos; /* offset in data payload */ - int did_page_crc; /* true if we've calculated crc for current page */ -}; - -/* ceph connection fault delay defaults, for exponential backoff */ -#define BASE_DELAY_INTERVAL (HZ/2) -#define MAX_DELAY_INTERVAL (5 * 60 * HZ) - -/* - * ceph_connection state bit flags - * - * QUEUED and BUSY are used together to ensure that only a single - * thread is currently opening, reading or writing data to the socket. - */ -#define LOSSYTX 0 /* we can close channel or drop messages on errors */ -#define CONNECTING 1 -#define NEGOTIATING 2 -#define KEEPALIVE_PENDING 3 -#define WRITE_PENDING 4 /* we have data ready to send */ -#define QUEUED 5 /* there is work queued on this connection */ -#define BUSY 6 /* work is being done */ -#define STANDBY 8 /* no outgoing messages, socket closed. we keep - * the ceph_connection around to maintain shared - * state with the peer. */ -#define CLOSED 10 /* we've closed the connection */ -#define SOCK_CLOSED 11 /* socket state changed to closed */ -#define OPENING 13 /* open connection w/ (possibly new) peer */ -#define DEAD 14 /* dead, about to kfree */ - -/* - * A single connection with another host. - * - * We maintain a queue of outgoing messages, and some session state to - * ensure that we can preserve the lossless, ordered delivery of - * messages in the case of a TCP disconnect. - */ -struct ceph_connection { - void *private; - atomic_t nref; - - const struct ceph_connection_operations *ops; - - struct ceph_messenger *msgr; - struct socket *sock; - unsigned long state; /* connection state (see flags above) */ - const char *error_msg; /* error message, if any */ - - struct ceph_entity_addr peer_addr; /* peer address */ - struct ceph_entity_name peer_name; /* peer name */ - struct ceph_entity_addr peer_addr_for_me; - unsigned peer_features; - u32 connect_seq; /* identify the most recent connection - attempt for this connection, client */ - u32 peer_global_seq; /* peer's global seq for this connection */ - - int auth_retry; /* true if we need a newer authorizer */ - void *auth_reply_buf; /* where to put the authorizer reply */ - int auth_reply_buf_len; - - struct mutex mutex; - - /* out queue */ - struct list_head out_queue; - struct list_head out_sent; /* sending or sent but unacked */ - u64 out_seq; /* last message queued for send */ - bool out_keepalive_pending; - - u64 in_seq, in_seq_acked; /* last message received, acked */ - - /* connection negotiation temps */ - char in_banner[CEPH_BANNER_MAX_LEN]; - union { - struct { /* outgoing connection */ - struct ceph_msg_connect out_connect; - struct ceph_msg_connect_reply in_reply; - }; - struct { /* incoming */ - struct ceph_msg_connect in_connect; - struct ceph_msg_connect_reply out_reply; - }; - }; - struct ceph_entity_addr actual_peer_addr; - - /* message out temps */ - struct ceph_msg *out_msg; /* sending message (== tail of - out_sent) */ - bool out_msg_done; - struct ceph_msg_pos out_msg_pos; - - struct kvec out_kvec[8], /* sending header/footer data */ - *out_kvec_cur; - int out_kvec_left; /* kvec's left in out_kvec */ - int out_skip; /* skip this many bytes */ - int out_kvec_bytes; /* total bytes left */ - bool out_kvec_is_msg; /* kvec refers to out_msg */ - int out_more; /* there is more data after the kvecs */ - __le64 out_temp_ack; /* for writing an ack */ - - /* message in temps */ - struct ceph_msg_header in_hdr; - struct ceph_msg *in_msg; - struct ceph_msg_pos in_msg_pos; - u32 in_front_crc, in_middle_crc, in_data_crc; /* calculated crc */ - - char in_tag; /* protocol control byte */ - int in_base_pos; /* bytes read */ - __le64 in_temp_ack; /* for reading an ack */ - - struct delayed_work work; /* send|recv work */ - unsigned long delay; /* current delay interval */ -}; - - -extern const char *pr_addr(const struct sockaddr_storage *ss); -extern int ceph_parse_ips(const char *c, const char *end, - struct ceph_entity_addr *addr, - int max_count, int *count); - - -extern int ceph_msgr_init(void); -extern void ceph_msgr_exit(void); -extern void ceph_msgr_flush(void); - -extern struct ceph_messenger *ceph_messenger_create( - struct ceph_entity_addr *myaddr); -extern void ceph_messenger_destroy(struct ceph_messenger *); - -extern void ceph_con_init(struct ceph_messenger *msgr, - struct ceph_connection *con); -extern void ceph_con_open(struct ceph_connection *con, - struct ceph_entity_addr *addr); -extern bool ceph_con_opened(struct ceph_connection *con); -extern void ceph_con_close(struct ceph_connection *con); -extern void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg); -extern void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg); -extern void ceph_con_revoke_message(struct ceph_connection *con, - struct ceph_msg *msg); -extern void ceph_con_keepalive(struct ceph_connection *con); -extern struct ceph_connection *ceph_con_get(struct ceph_connection *con); -extern void ceph_con_put(struct ceph_connection *con); - -extern struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags); -extern void ceph_msg_kfree(struct ceph_msg *m); - - -static inline struct ceph_msg *ceph_msg_get(struct ceph_msg *msg) -{ - kref_get(&msg->kref); - return msg; -} -extern void ceph_msg_last_put(struct kref *kref); -static inline void ceph_msg_put(struct ceph_msg *msg) -{ - kref_put(&msg->kref, ceph_msg_last_put); -} - -extern void ceph_msg_dump(struct ceph_msg *msg); - -#endif diff --git a/fs/ceph/mon_client.c b/fs/ceph/mon_client.c deleted file mode 100644 index b2a5a3e..0000000 --- a/fs/ceph/mon_client.c +++ /dev/null @@ -1,1018 +0,0 @@ -#include "ceph_debug.h" - -#include -#include -#include -#include - -#include "mon_client.h" -#include "super.h" -#include "auth.h" -#include "decode.h" - -/* - * Interact with Ceph monitor cluster. Handle requests for new map - * versions, and periodically resend as needed. Also implement - * statfs() and umount(). - * - * A small cluster of Ceph "monitors" are responsible for managing critical - * cluster configuration and state information. An odd number (e.g., 3, 5) - * of cmon daemons use a modified version of the Paxos part-time parliament - * algorithm to manage the MDS map (mds cluster membership), OSD map, and - * list of clients who have mounted the file system. - * - * We maintain an open, active session with a monitor at all times in order to - * receive timely MDSMap updates. We periodically send a keepalive byte on the - * TCP socket to ensure we detect a failure. If the connection does break, we - * randomly hunt for a new monitor. Once the connection is reestablished, we - * resend any outstanding requests. - */ - -static const struct ceph_connection_operations mon_con_ops; - -static int __validate_auth(struct ceph_mon_client *monc); - -/* - * Decode a monmap blob (e.g., during mount). - */ -struct ceph_monmap *ceph_monmap_decode(void *p, void *end) -{ - struct ceph_monmap *m = NULL; - int i, err = -EINVAL; - struct ceph_fsid fsid; - u32 epoch, num_mon; - u16 version; - u32 len; - - ceph_decode_32_safe(&p, end, len, bad); - ceph_decode_need(&p, end, len, bad); - - dout("monmap_decode %p %p len %d\n", p, end, (int)(end-p)); - - ceph_decode_16_safe(&p, end, version, bad); - - ceph_decode_need(&p, end, sizeof(fsid) + 2*sizeof(u32), bad); - ceph_decode_copy(&p, &fsid, sizeof(fsid)); - epoch = ceph_decode_32(&p); - - num_mon = ceph_decode_32(&p); - ceph_decode_need(&p, end, num_mon*sizeof(m->mon_inst[0]), bad); - - if (num_mon >= CEPH_MAX_MON) - goto bad; - m = kmalloc(sizeof(*m) + sizeof(m->mon_inst[0])*num_mon, GFP_NOFS); - if (m == NULL) - return ERR_PTR(-ENOMEM); - m->fsid = fsid; - m->epoch = epoch; - m->num_mon = num_mon; - ceph_decode_copy(&p, m->mon_inst, num_mon*sizeof(m->mon_inst[0])); - for (i = 0; i < num_mon; i++) - ceph_decode_addr(&m->mon_inst[i].addr); - - dout("monmap_decode epoch %d, num_mon %d\n", m->epoch, - m->num_mon); - for (i = 0; i < m->num_mon; i++) - dout("monmap_decode mon%d is %s\n", i, - pr_addr(&m->mon_inst[i].addr.in_addr)); - return m; - -bad: - dout("monmap_decode failed with %d\n", err); - kfree(m); - return ERR_PTR(err); -} - -/* - * return true if *addr is included in the monmap. - */ -int ceph_monmap_contains(struct ceph_monmap *m, struct ceph_entity_addr *addr) -{ - int i; - - for (i = 0; i < m->num_mon; i++) - if (memcmp(addr, &m->mon_inst[i].addr, sizeof(*addr)) == 0) - return 1; - return 0; -} - -/* - * Send an auth request. - */ -static void __send_prepared_auth_request(struct ceph_mon_client *monc, int len) -{ - monc->pending_auth = 1; - monc->m_auth->front.iov_len = len; - monc->m_auth->hdr.front_len = cpu_to_le32(len); - ceph_con_revoke(monc->con, monc->m_auth); - ceph_msg_get(monc->m_auth); /* keep our ref */ - ceph_con_send(monc->con, monc->m_auth); -} - -/* - * Close monitor session, if any. - */ -static void __close_session(struct ceph_mon_client *monc) -{ - if (monc->con) { - dout("__close_session closing mon%d\n", monc->cur_mon); - ceph_con_revoke(monc->con, monc->m_auth); - ceph_con_close(monc->con); - monc->cur_mon = -1; - monc->pending_auth = 0; - ceph_auth_reset(monc->auth); - } -} - -/* - * Open a session with a (new) monitor. - */ -static int __open_session(struct ceph_mon_client *monc) -{ - char r; - int ret; - - if (monc->cur_mon < 0) { - get_random_bytes(&r, 1); - monc->cur_mon = r % monc->monmap->num_mon; - dout("open_session num=%d r=%d -> mon%d\n", - monc->monmap->num_mon, r, monc->cur_mon); - monc->sub_sent = 0; - monc->sub_renew_after = jiffies; /* i.e., expired */ - monc->want_next_osdmap = !!monc->want_next_osdmap; - - dout("open_session mon%d opening\n", monc->cur_mon); - monc->con->peer_name.type = CEPH_ENTITY_TYPE_MON; - monc->con->peer_name.num = cpu_to_le64(monc->cur_mon); - ceph_con_open(monc->con, - &monc->monmap->mon_inst[monc->cur_mon].addr); - - /* initiatiate authentication handshake */ - ret = ceph_auth_build_hello(monc->auth, - monc->m_auth->front.iov_base, - monc->m_auth->front_max); - __send_prepared_auth_request(monc, ret); - } else { - dout("open_session mon%d already open\n", monc->cur_mon); - } - return 0; -} - -static bool __sub_expired(struct ceph_mon_client *monc) -{ - return time_after_eq(jiffies, monc->sub_renew_after); -} - -/* - * Reschedule delayed work timer. - */ -static void __schedule_delayed(struct ceph_mon_client *monc) -{ - unsigned delay; - - if (monc->cur_mon < 0 || __sub_expired(monc)) - delay = 10 * HZ; - else - delay = 20 * HZ; - dout("__schedule_delayed after %u\n", delay); - schedule_delayed_work(&monc->delayed_work, delay); -} - -/* - * Send subscribe request for mdsmap and/or osdmap. - */ -static void __send_subscribe(struct ceph_mon_client *monc) -{ - dout("__send_subscribe sub_sent=%u exp=%u want_osd=%d\n", - (unsigned)monc->sub_sent, __sub_expired(monc), - monc->want_next_osdmap); - if ((__sub_expired(monc) && !monc->sub_sent) || - monc->want_next_osdmap == 1) { - struct ceph_msg *msg = monc->m_subscribe; - struct ceph_mon_subscribe_item *i; - void *p, *end; - - p = msg->front.iov_base; - end = p + msg->front_max; - - dout("__send_subscribe to 'mdsmap' %u+\n", - (unsigned)monc->have_mdsmap); - if (monc->want_next_osdmap) { - dout("__send_subscribe to 'osdmap' %u\n", - (unsigned)monc->have_osdmap); - ceph_encode_32(&p, 3); - ceph_encode_string(&p, end, "osdmap", 6); - i = p; - i->have = cpu_to_le64(monc->have_osdmap); - i->onetime = 1; - p += sizeof(*i); - monc->want_next_osdmap = 2; /* requested */ - } else { - ceph_encode_32(&p, 2); - } - ceph_encode_string(&p, end, "mdsmap", 6); - i = p; - i->have = cpu_to_le64(monc->have_mdsmap); - i->onetime = 0; - p += sizeof(*i); - ceph_encode_string(&p, end, "monmap", 6); - i = p; - i->have = 0; - i->onetime = 0; - p += sizeof(*i); - - msg->front.iov_len = p - msg->front.iov_base; - msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); - ceph_con_revoke(monc->con, msg); - ceph_con_send(monc->con, ceph_msg_get(msg)); - - monc->sub_sent = jiffies | 1; /* never 0 */ - } -} - -static void handle_subscribe_ack(struct ceph_mon_client *monc, - struct ceph_msg *msg) -{ - unsigned seconds; - struct ceph_mon_subscribe_ack *h = msg->front.iov_base; - - if (msg->front.iov_len < sizeof(*h)) - goto bad; - seconds = le32_to_cpu(h->duration); - - mutex_lock(&monc->mutex); - if (monc->hunting) { - pr_info("mon%d %s session established\n", - monc->cur_mon, pr_addr(&monc->con->peer_addr.in_addr)); - monc->hunting = false; - } - dout("handle_subscribe_ack after %d seconds\n", seconds); - monc->sub_renew_after = monc->sub_sent + (seconds >> 1)*HZ - 1; - monc->sub_sent = 0; - mutex_unlock(&monc->mutex); - return; -bad: - pr_err("got corrupt subscribe-ack msg\n"); - ceph_msg_dump(msg); -} - -/* - * Keep track of which maps we have - */ -int ceph_monc_got_mdsmap(struct ceph_mon_client *monc, u32 got) -{ - mutex_lock(&monc->mutex); - monc->have_mdsmap = got; - mutex_unlock(&monc->mutex); - return 0; -} - -int ceph_monc_got_osdmap(struct ceph_mon_client *monc, u32 got) -{ - mutex_lock(&monc->mutex); - monc->have_osdmap = got; - monc->want_next_osdmap = 0; - mutex_unlock(&monc->mutex); - return 0; -} - -/* - * Register interest in the next osdmap - */ -void ceph_monc_request_next_osdmap(struct ceph_mon_client *monc) -{ - dout("request_next_osdmap have %u\n", monc->have_osdmap); - mutex_lock(&monc->mutex); - if (!monc->want_next_osdmap) - monc->want_next_osdmap = 1; - if (monc->want_next_osdmap < 2) - __send_subscribe(monc); - mutex_unlock(&monc->mutex); -} - -/* - * - */ -int ceph_monc_open_session(struct ceph_mon_client *monc) -{ - if (!monc->con) { - monc->con = kmalloc(sizeof(*monc->con), GFP_KERNEL); - if (!monc->con) - return -ENOMEM; - ceph_con_init(monc->client->msgr, monc->con); - monc->con->private = monc; - monc->con->ops = &mon_con_ops; - } - - mutex_lock(&monc->mutex); - __open_session(monc); - __schedule_delayed(monc); - mutex_unlock(&monc->mutex); - return 0; -} - -/* - * The monitor responds with mount ack indicate mount success. The - * included client ticket allows the client to talk to MDSs and OSDs. - */ -static void ceph_monc_handle_map(struct ceph_mon_client *monc, - struct ceph_msg *msg) -{ - struct ceph_client *client = monc->client; - struct ceph_monmap *monmap = NULL, *old = monc->monmap; - void *p, *end; - - mutex_lock(&monc->mutex); - - dout("handle_monmap\n"); - p = msg->front.iov_base; - end = p + msg->front.iov_len; - - monmap = ceph_monmap_decode(p, end); - if (IS_ERR(monmap)) { - pr_err("problem decoding monmap, %d\n", - (int)PTR_ERR(monmap)); - goto out; - } - - if (ceph_check_fsid(monc->client, &monmap->fsid) < 0) { - kfree(monmap); - goto out; - } - - client->monc.monmap = monmap; - kfree(old); - -out: - mutex_unlock(&monc->mutex); - wake_up_all(&client->auth_wq); -} - -/* - * generic requests (e.g., statfs, poolop) - */ -static struct ceph_mon_generic_request *__lookup_generic_req( - struct ceph_mon_client *monc, u64 tid) -{ - struct ceph_mon_generic_request *req; - struct rb_node *n = monc->generic_request_tree.rb_node; - - while (n) { - req = rb_entry(n, struct ceph_mon_generic_request, node); - if (tid < req->tid) - n = n->rb_left; - else if (tid > req->tid) - n = n->rb_right; - else - return req; - } - return NULL; -} - -static void __insert_generic_request(struct ceph_mon_client *monc, - struct ceph_mon_generic_request *new) -{ - struct rb_node **p = &monc->generic_request_tree.rb_node; - struct rb_node *parent = NULL; - struct ceph_mon_generic_request *req = NULL; - - while (*p) { - parent = *p; - req = rb_entry(parent, struct ceph_mon_generic_request, node); - if (new->tid < req->tid) - p = &(*p)->rb_left; - else if (new->tid > req->tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&new->node, parent, p); - rb_insert_color(&new->node, &monc->generic_request_tree); -} - -static void release_generic_request(struct kref *kref) -{ - struct ceph_mon_generic_request *req = - container_of(kref, struct ceph_mon_generic_request, kref); - - if (req->reply) - ceph_msg_put(req->reply); - if (req->request) - ceph_msg_put(req->request); - - kfree(req); -} - -static void put_generic_request(struct ceph_mon_generic_request *req) -{ - kref_put(&req->kref, release_generic_request); -} - -static void get_generic_request(struct ceph_mon_generic_request *req) -{ - kref_get(&req->kref); -} - -static struct ceph_msg *get_generic_reply(struct ceph_connection *con, - struct ceph_msg_header *hdr, - int *skip) -{ - struct ceph_mon_client *monc = con->private; - struct ceph_mon_generic_request *req; - u64 tid = le64_to_cpu(hdr->tid); - struct ceph_msg *m; - - mutex_lock(&monc->mutex); - req = __lookup_generic_req(monc, tid); - if (!req) { - dout("get_generic_reply %lld dne\n", tid); - *skip = 1; - m = NULL; - } else { - dout("get_generic_reply %lld got %p\n", tid, req->reply); - m = ceph_msg_get(req->reply); - /* - * we don't need to track the connection reading into - * this reply because we only have one open connection - * at a time, ever. - */ - } - mutex_unlock(&monc->mutex); - return m; -} - -static int do_generic_request(struct ceph_mon_client *monc, - struct ceph_mon_generic_request *req) -{ - int err; - - /* register request */ - mutex_lock(&monc->mutex); - req->tid = ++monc->last_tid; - req->request->hdr.tid = cpu_to_le64(req->tid); - __insert_generic_request(monc, req); - monc->num_generic_requests++; - ceph_con_send(monc->con, ceph_msg_get(req->request)); - mutex_unlock(&monc->mutex); - - err = wait_for_completion_interruptible(&req->completion); - - mutex_lock(&monc->mutex); - rb_erase(&req->node, &monc->generic_request_tree); - monc->num_generic_requests--; - mutex_unlock(&monc->mutex); - - if (!err) - err = req->result; - return err; -} - -/* - * statfs - */ -static void handle_statfs_reply(struct ceph_mon_client *monc, - struct ceph_msg *msg) -{ - struct ceph_mon_generic_request *req; - struct ceph_mon_statfs_reply *reply = msg->front.iov_base; - u64 tid = le64_to_cpu(msg->hdr.tid); - - if (msg->front.iov_len != sizeof(*reply)) - goto bad; - dout("handle_statfs_reply %p tid %llu\n", msg, tid); - - mutex_lock(&monc->mutex); - req = __lookup_generic_req(monc, tid); - if (req) { - *(struct ceph_statfs *)req->buf = reply->st; - req->result = 0; - get_generic_request(req); - } - mutex_unlock(&monc->mutex); - if (req) { - complete_all(&req->completion); - put_generic_request(req); - } - return; - -bad: - pr_err("corrupt generic reply, tid %llu\n", tid); - ceph_msg_dump(msg); -} - -/* - * Do a synchronous statfs(). - */ -int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf) -{ - struct ceph_mon_generic_request *req; - struct ceph_mon_statfs *h; - int err; - - req = kzalloc(sizeof(*req), GFP_NOFS); - if (!req) - return -ENOMEM; - - kref_init(&req->kref); - req->buf = buf; - req->buf_len = sizeof(*buf); - init_completion(&req->completion); - - err = -ENOMEM; - req->request = ceph_msg_new(CEPH_MSG_STATFS, sizeof(*h), GFP_NOFS); - if (!req->request) - goto out; - req->reply = ceph_msg_new(CEPH_MSG_STATFS_REPLY, 1024, GFP_NOFS); - if (!req->reply) - goto out; - - /* fill out request */ - h = req->request->front.iov_base; - h->monhdr.have_version = 0; - h->monhdr.session_mon = cpu_to_le16(-1); - h->monhdr.session_mon_tid = 0; - h->fsid = monc->monmap->fsid; - - err = do_generic_request(monc, req); - -out: - kref_put(&req->kref, release_generic_request); - return err; -} - -/* - * pool ops - */ -static int get_poolop_reply_buf(const char *src, size_t src_len, - char *dst, size_t dst_len) -{ - u32 buf_len; - - if (src_len != sizeof(u32) + dst_len) - return -EINVAL; - - buf_len = le32_to_cpu(*(u32 *)src); - if (buf_len != dst_len) - return -EINVAL; - - memcpy(dst, src + sizeof(u32), dst_len); - return 0; -} - -static void handle_poolop_reply(struct ceph_mon_client *monc, - struct ceph_msg *msg) -{ - struct ceph_mon_generic_request *req; - struct ceph_mon_poolop_reply *reply = msg->front.iov_base; - u64 tid = le64_to_cpu(msg->hdr.tid); - - if (msg->front.iov_len < sizeof(*reply)) - goto bad; - dout("handle_poolop_reply %p tid %llu\n", msg, tid); - - mutex_lock(&monc->mutex); - req = __lookup_generic_req(monc, tid); - if (req) { - if (req->buf_len && - get_poolop_reply_buf(msg->front.iov_base + sizeof(*reply), - msg->front.iov_len - sizeof(*reply), - req->buf, req->buf_len) < 0) { - mutex_unlock(&monc->mutex); - goto bad; - } - req->result = le32_to_cpu(reply->reply_code); - get_generic_request(req); - } - mutex_unlock(&monc->mutex); - if (req) { - complete(&req->completion); - put_generic_request(req); - } - return; - -bad: - pr_err("corrupt generic reply, tid %llu\n", tid); - ceph_msg_dump(msg); -} - -/* - * Do a synchronous pool op. - */ -int ceph_monc_do_poolop(struct ceph_mon_client *monc, u32 op, - u32 pool, u64 snapid, - char *buf, int len) -{ - struct ceph_mon_generic_request *req; - struct ceph_mon_poolop *h; - int err; - - req = kzalloc(sizeof(*req), GFP_NOFS); - if (!req) - return -ENOMEM; - - kref_init(&req->kref); - req->buf = buf; - req->buf_len = len; - init_completion(&req->completion); - - err = -ENOMEM; - req->request = ceph_msg_new(CEPH_MSG_POOLOP, sizeof(*h), GFP_NOFS); - if (!req->request) - goto out; - req->reply = ceph_msg_new(CEPH_MSG_POOLOP_REPLY, 1024, GFP_NOFS); - if (!req->reply) - goto out; - - /* fill out request */ - req->request->hdr.version = cpu_to_le16(2); - h = req->request->front.iov_base; - h->monhdr.have_version = 0; - h->monhdr.session_mon = cpu_to_le16(-1); - h->monhdr.session_mon_tid = 0; - h->fsid = monc->monmap->fsid; - h->pool = cpu_to_le32(pool); - h->op = cpu_to_le32(op); - h->auid = 0; - h->snapid = cpu_to_le64(snapid); - h->name_len = 0; - - err = do_generic_request(monc, req); - -out: - kref_put(&req->kref, release_generic_request); - return err; -} - -int ceph_monc_create_snapid(struct ceph_mon_client *monc, - u32 pool, u64 *snapid) -{ - return ceph_monc_do_poolop(monc, POOL_OP_CREATE_UNMANAGED_SNAP, - pool, 0, (char *)snapid, sizeof(*snapid)); - -} - -int ceph_monc_delete_snapid(struct ceph_mon_client *monc, - u32 pool, u64 snapid) -{ - return ceph_monc_do_poolop(monc, POOL_OP_CREATE_UNMANAGED_SNAP, - pool, snapid, 0, 0); - -} - -/* - * Resend pending generic requests. - */ -static void __resend_generic_request(struct ceph_mon_client *monc) -{ - struct ceph_mon_generic_request *req; - struct rb_node *p; - - for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) { - req = rb_entry(p, struct ceph_mon_generic_request, node); - ceph_con_revoke(monc->con, req->request); - ceph_con_send(monc->con, ceph_msg_get(req->request)); - } -} - -/* - * Delayed work. If we haven't mounted yet, retry. Otherwise, - * renew/retry subscription as needed (in case it is timing out, or we - * got an ENOMEM). And keep the monitor connection alive. - */ -static void delayed_work(struct work_struct *work) -{ - struct ceph_mon_client *monc = - container_of(work, struct ceph_mon_client, delayed_work.work); - - dout("monc delayed_work\n"); - mutex_lock(&monc->mutex); - if (monc->hunting) { - __close_session(monc); - __open_session(monc); /* continue hunting */ - } else { - ceph_con_keepalive(monc->con); - - __validate_auth(monc); - - if (monc->auth->ops->is_authenticated(monc->auth)) - __send_subscribe(monc); - } - __schedule_delayed(monc); - mutex_unlock(&monc->mutex); -} - -/* - * On startup, we build a temporary monmap populated with the IPs - * provided by mount(2). - */ -static int build_initial_monmap(struct ceph_mon_client *monc) -{ - struct ceph_mount_args *args = monc->client->mount_args; - struct ceph_entity_addr *mon_addr = args->mon_addr; - int num_mon = args->num_mon; - int i; - - /* build initial monmap */ - monc->monmap = kzalloc(sizeof(*monc->monmap) + - num_mon*sizeof(monc->monmap->mon_inst[0]), - GFP_KERNEL); - if (!monc->monmap) - return -ENOMEM; - for (i = 0; i < num_mon; i++) { - monc->monmap->mon_inst[i].addr = mon_addr[i]; - monc->monmap->mon_inst[i].addr.nonce = 0; - monc->monmap->mon_inst[i].name.type = - CEPH_ENTITY_TYPE_MON; - monc->monmap->mon_inst[i].name.num = cpu_to_le64(i); - } - monc->monmap->num_mon = num_mon; - monc->have_fsid = false; - - /* release addr memory */ - kfree(args->mon_addr); - args->mon_addr = NULL; - args->num_mon = 0; - return 0; -} - -int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl) -{ - int err = 0; - - dout("init\n"); - memset(monc, 0, sizeof(*monc)); - monc->client = cl; - monc->monmap = NULL; - mutex_init(&monc->mutex); - - err = build_initial_monmap(monc); - if (err) - goto out; - - monc->con = NULL; - - /* authentication */ - monc->auth = ceph_auth_init(cl->mount_args->name, - cl->mount_args->secret); - if (IS_ERR(monc->auth)) - return PTR_ERR(monc->auth); - monc->auth->want_keys = - CEPH_ENTITY_TYPE_AUTH | CEPH_ENTITY_TYPE_MON | - CEPH_ENTITY_TYPE_OSD | CEPH_ENTITY_TYPE_MDS; - - /* msgs */ - err = -ENOMEM; - monc->m_subscribe_ack = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE_ACK, - sizeof(struct ceph_mon_subscribe_ack), - GFP_NOFS); - if (!monc->m_subscribe_ack) - goto out_monmap; - - monc->m_subscribe = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE, 96, GFP_NOFS); - if (!monc->m_subscribe) - goto out_subscribe_ack; - - monc->m_auth_reply = ceph_msg_new(CEPH_MSG_AUTH_REPLY, 4096, GFP_NOFS); - if (!monc->m_auth_reply) - goto out_subscribe; - - monc->m_auth = ceph_msg_new(CEPH_MSG_AUTH, 4096, GFP_NOFS); - monc->pending_auth = 0; - if (!monc->m_auth) - goto out_auth_reply; - - monc->cur_mon = -1; - monc->hunting = true; - monc->sub_renew_after = jiffies; - monc->sub_sent = 0; - - INIT_DELAYED_WORK(&monc->delayed_work, delayed_work); - monc->generic_request_tree = RB_ROOT; - monc->num_generic_requests = 0; - monc->last_tid = 0; - - monc->have_mdsmap = 0; - monc->have_osdmap = 0; - monc->want_next_osdmap = 1; - return 0; - -out_auth_reply: - ceph_msg_put(monc->m_auth_reply); -out_subscribe: - ceph_msg_put(monc->m_subscribe); -out_subscribe_ack: - ceph_msg_put(monc->m_subscribe_ack); -out_monmap: - kfree(monc->monmap); -out: - return err; -} - -void ceph_monc_stop(struct ceph_mon_client *monc) -{ - dout("stop\n"); - cancel_delayed_work_sync(&monc->delayed_work); - - mutex_lock(&monc->mutex); - __close_session(monc); - if (monc->con) { - monc->con->private = NULL; - monc->con->ops->put(monc->con); - monc->con = NULL; - } - mutex_unlock(&monc->mutex); - - ceph_auth_destroy(monc->auth); - - ceph_msg_put(monc->m_auth); - ceph_msg_put(monc->m_auth_reply); - ceph_msg_put(monc->m_subscribe); - ceph_msg_put(monc->m_subscribe_ack); - - kfree(monc->monmap); -} - -static void handle_auth_reply(struct ceph_mon_client *monc, - struct ceph_msg *msg) -{ - int ret; - int was_auth = 0; - - mutex_lock(&monc->mutex); - if (monc->auth->ops) - was_auth = monc->auth->ops->is_authenticated(monc->auth); - monc->pending_auth = 0; - ret = ceph_handle_auth_reply(monc->auth, msg->front.iov_base, - msg->front.iov_len, - monc->m_auth->front.iov_base, - monc->m_auth->front_max); - if (ret < 0) { - monc->client->auth_err = ret; - wake_up_all(&monc->client->auth_wq); - } else if (ret > 0) { - __send_prepared_auth_request(monc, ret); - } else if (!was_auth && monc->auth->ops->is_authenticated(monc->auth)) { - dout("authenticated, starting session\n"); - - monc->client->msgr->inst.name.type = CEPH_ENTITY_TYPE_CLIENT; - monc->client->msgr->inst.name.num = - cpu_to_le64(monc->auth->global_id); - - __send_subscribe(monc); - __resend_generic_request(monc); - } - mutex_unlock(&monc->mutex); -} - -static int __validate_auth(struct ceph_mon_client *monc) -{ - int ret; - - if (monc->pending_auth) - return 0; - - ret = ceph_build_auth(monc->auth, monc->m_auth->front.iov_base, - monc->m_auth->front_max); - if (ret <= 0) - return ret; /* either an error, or no need to authenticate */ - __send_prepared_auth_request(monc, ret); - return 0; -} - -int ceph_monc_validate_auth(struct ceph_mon_client *monc) -{ - int ret; - - mutex_lock(&monc->mutex); - ret = __validate_auth(monc); - mutex_unlock(&monc->mutex); - return ret; -} - -/* - * handle incoming message - */ -static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) -{ - struct ceph_mon_client *monc = con->private; - int type = le16_to_cpu(msg->hdr.type); - - if (!monc) - return; - - switch (type) { - case CEPH_MSG_AUTH_REPLY: - handle_auth_reply(monc, msg); - break; - - case CEPH_MSG_MON_SUBSCRIBE_ACK: - handle_subscribe_ack(monc, msg); - break; - - case CEPH_MSG_STATFS_REPLY: - handle_statfs_reply(monc, msg); - break; - - case CEPH_MSG_POOLOP_REPLY: - handle_poolop_reply(monc, msg); - break; - - case CEPH_MSG_MON_MAP: - ceph_monc_handle_map(monc, msg); - break; - - case CEPH_MSG_MDS_MAP: - ceph_mdsc_handle_map(&monc->client->mdsc, msg); - break; - - case CEPH_MSG_OSD_MAP: - ceph_osdc_handle_map(&monc->client->osdc, msg); - break; - - default: - pr_err("received unknown message type %d %s\n", type, - ceph_msg_type_name(type)); - } - ceph_msg_put(msg); -} - -/* - * Allocate memory for incoming message - */ -static struct ceph_msg *mon_alloc_msg(struct ceph_connection *con, - struct ceph_msg_header *hdr, - int *skip) -{ - struct ceph_mon_client *monc = con->private; - int type = le16_to_cpu(hdr->type); - int front_len = le32_to_cpu(hdr->front_len); - struct ceph_msg *m = NULL; - - *skip = 0; - - switch (type) { - case CEPH_MSG_MON_SUBSCRIBE_ACK: - m = ceph_msg_get(monc->m_subscribe_ack); - break; - case CEPH_MSG_POOLOP_REPLY: - case CEPH_MSG_STATFS_REPLY: - return get_generic_reply(con, hdr, skip); - case CEPH_MSG_AUTH_REPLY: - m = ceph_msg_get(monc->m_auth_reply); - break; - case CEPH_MSG_MON_MAP: - case CEPH_MSG_MDS_MAP: - case CEPH_MSG_OSD_MAP: - m = ceph_msg_new(type, front_len, GFP_NOFS); - break; - } - - if (!m) { - pr_info("alloc_msg unknown type %d\n", type); - *skip = 1; - } - return m; -} - -/* - * If the monitor connection resets, pick a new monitor and resubmit - * any pending requests. - */ -static void mon_fault(struct ceph_connection *con) -{ - struct ceph_mon_client *monc = con->private; - - if (!monc) - return; - - dout("mon_fault\n"); - mutex_lock(&monc->mutex); - if (!con->private) - goto out; - - if (monc->con && !monc->hunting) - pr_info("mon%d %s session lost, " - "hunting for new mon\n", monc->cur_mon, - pr_addr(&monc->con->peer_addr.in_addr)); - - __close_session(monc); - if (!monc->hunting) { - /* start hunting */ - monc->hunting = true; - __open_session(monc); - } else { - /* already hunting, let's wait a bit */ - __schedule_delayed(monc); - } -out: - mutex_unlock(&monc->mutex); -} - -static const struct ceph_connection_operations mon_con_ops = { - .get = ceph_con_get, - .put = ceph_con_put, - .dispatch = dispatch, - .fault = mon_fault, - .alloc_msg = mon_alloc_msg, -}; diff --git a/fs/ceph/mon_client.h b/fs/ceph/mon_client.h deleted file mode 100644 index 8e396f2..0000000 --- a/fs/ceph/mon_client.h +++ /dev/null @@ -1,121 +0,0 @@ -#ifndef _FS_CEPH_MON_CLIENT_H -#define _FS_CEPH_MON_CLIENT_H - -#include -#include -#include - -#include "messenger.h" - -struct ceph_client; -struct ceph_mount_args; -struct ceph_auth_client; - -/* - * The monitor map enumerates the set of all monitors. - */ -struct ceph_monmap { - struct ceph_fsid fsid; - u32 epoch; - u32 num_mon; - struct ceph_entity_inst mon_inst[0]; -}; - -struct ceph_mon_client; -struct ceph_mon_generic_request; - - -/* - * Generic mechanism for resending monitor requests. - */ -typedef void (*ceph_monc_request_func_t)(struct ceph_mon_client *monc, - int newmon); - -/* a pending monitor request */ -struct ceph_mon_request { - struct ceph_mon_client *monc; - struct delayed_work delayed_work; - unsigned long delay; - ceph_monc_request_func_t do_request; -}; - -/* - * ceph_mon_generic_request is being used for the statfs and poolop requests - * which are bening done a bit differently because we need to get data back - * to the caller - */ -struct ceph_mon_generic_request { - struct kref kref; - u64 tid; - struct rb_node node; - int result; - void *buf; - int buf_len; - struct completion completion; - struct ceph_msg *request; /* original request */ - struct ceph_msg *reply; /* and reply */ -}; - -struct ceph_mon_client { - struct ceph_client *client; - struct ceph_monmap *monmap; - - struct mutex mutex; - struct delayed_work delayed_work; - - struct ceph_auth_client *auth; - struct ceph_msg *m_auth, *m_auth_reply, *m_subscribe, *m_subscribe_ack; - int pending_auth; - - bool hunting; - int cur_mon; /* last monitor i contacted */ - unsigned long sub_sent, sub_renew_after; - struct ceph_connection *con; - bool have_fsid; - - /* pending generic requests */ - struct rb_root generic_request_tree; - int num_generic_requests; - u64 last_tid; - - /* mds/osd map */ - int want_next_osdmap; /* 1 = want, 2 = want+asked */ - u32 have_osdmap, have_mdsmap; - -#ifdef CONFIG_DEBUG_FS - struct dentry *debugfs_file; -#endif -}; - -extern struct ceph_monmap *ceph_monmap_decode(void *p, void *end); -extern int ceph_monmap_contains(struct ceph_monmap *m, - struct ceph_entity_addr *addr); - -extern int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl); -extern void ceph_monc_stop(struct ceph_mon_client *monc); - -/* - * The model here is to indicate that we need a new map of at least - * epoch @want, and also call in when we receive a map. We will - * periodically rerequest the map from the monitor cluster until we - * get what we want. - */ -extern int ceph_monc_got_mdsmap(struct ceph_mon_client *monc, u32 have); -extern int ceph_monc_got_osdmap(struct ceph_mon_client *monc, u32 have); - -extern void ceph_monc_request_next_osdmap(struct ceph_mon_client *monc); - -extern int ceph_monc_do_statfs(struct ceph_mon_client *monc, - struct ceph_statfs *buf); - -extern int ceph_monc_open_session(struct ceph_mon_client *monc); - -extern int ceph_monc_validate_auth(struct ceph_mon_client *monc); - -extern int ceph_monc_create_snapid(struct ceph_mon_client *monc, - u32 pool, u64 *snapid); - -extern int ceph_monc_delete_snapid(struct ceph_mon_client *monc, - u32 pool, u64 snapid); - -#endif diff --git a/fs/ceph/msgpool.c b/fs/ceph/msgpool.c deleted file mode 100644 index dd65a64..0000000 --- a/fs/ceph/msgpool.c +++ /dev/null @@ -1,64 +0,0 @@ -#include "ceph_debug.h" - -#include -#include -#include -#include - -#include "msgpool.h" - -static void *alloc_fn(gfp_t gfp_mask, void *arg) -{ - struct ceph_msgpool *pool = arg; - void *p; - - p = ceph_msg_new(0, pool->front_len, gfp_mask); - if (!p) - pr_err("msgpool %s alloc failed\n", pool->name); - return p; -} - -static void free_fn(void *element, void *arg) -{ - ceph_msg_put(element); -} - -int ceph_msgpool_init(struct ceph_msgpool *pool, - int front_len, int size, bool blocking, const char *name) -{ - pool->front_len = front_len; - pool->pool = mempool_create(size, alloc_fn, free_fn, pool); - if (!pool->pool) - return -ENOMEM; - pool->name = name; - return 0; -} - -void ceph_msgpool_destroy(struct ceph_msgpool *pool) -{ - mempool_destroy(pool->pool); -} - -struct ceph_msg *ceph_msgpool_get(struct ceph_msgpool *pool, - int front_len) -{ - if (front_len > pool->front_len) { - pr_err("msgpool_get pool %s need front %d, pool size is %d\n", - pool->name, front_len, pool->front_len); - WARN_ON(1); - - /* try to alloc a fresh message */ - return ceph_msg_new(0, front_len, GFP_NOFS); - } - - return mempool_alloc(pool->pool, GFP_NOFS); -} - -void ceph_msgpool_put(struct ceph_msgpool *pool, struct ceph_msg *msg) -{ - /* reset msg front_len; user may have changed it */ - msg->front.iov_len = pool->front_len; - msg->hdr.front_len = cpu_to_le32(pool->front_len); - - kref_init(&msg->kref); /* retake single ref */ -} diff --git a/fs/ceph/msgpool.h b/fs/ceph/msgpool.h deleted file mode 100644 index a362605f..0000000 --- a/fs/ceph/msgpool.h +++ /dev/null @@ -1,25 +0,0 @@ -#ifndef _FS_CEPH_MSGPOOL -#define _FS_CEPH_MSGPOOL - -#include -#include "messenger.h" - -/* - * we use memory pools for preallocating messages we may receive, to - * avoid unexpected OOM conditions. - */ -struct ceph_msgpool { - const char *name; - mempool_t *pool; - int front_len; /* preallocated payload size */ -}; - -extern int ceph_msgpool_init(struct ceph_msgpool *pool, - int front_len, int size, bool blocking, - const char *name); -extern void ceph_msgpool_destroy(struct ceph_msgpool *pool); -extern struct ceph_msg *ceph_msgpool_get(struct ceph_msgpool *, - int front_len); -extern void ceph_msgpool_put(struct ceph_msgpool *, struct ceph_msg *); - -#endif diff --git a/fs/ceph/msgr.h b/fs/ceph/msgr.h deleted file mode 100644 index 680d3d6..0000000 --- a/fs/ceph/msgr.h +++ /dev/null @@ -1,175 +0,0 @@ -#ifndef CEPH_MSGR_H -#define CEPH_MSGR_H - -/* - * Data types for message passing layer used by Ceph. - */ - -#define CEPH_MON_PORT 6789 /* default monitor port */ - -/* - * client-side processes will try to bind to ports in this - * range, simply for the benefit of tools like nmap or wireshark - * that would like to identify the protocol. - */ -#define CEPH_PORT_FIRST 6789 -#define CEPH_PORT_START 6800 /* non-monitors start here */ -#define CEPH_PORT_LAST 6900 - -/* - * tcp connection banner. include a protocol version. and adjust - * whenever the wire protocol changes. try to keep this string length - * constant. - */ -#define CEPH_BANNER "ceph v027" -#define CEPH_BANNER_MAX_LEN 30 - - -/* - * Rollover-safe type and comparator for 32-bit sequence numbers. - * Comparator returns -1, 0, or 1. - */ -typedef __u32 ceph_seq_t; - -static inline __s32 ceph_seq_cmp(__u32 a, __u32 b) -{ - return (__s32)a - (__s32)b; -} - - -/* - * entity_name -- logical name for a process participating in the - * network, e.g. 'mds0' or 'osd3'. - */ -struct ceph_entity_name { - __u8 type; /* CEPH_ENTITY_TYPE_* */ - __le64 num; -} __attribute__ ((packed)); - -#define CEPH_ENTITY_TYPE_MON 0x01 -#define CEPH_ENTITY_TYPE_MDS 0x02 -#define CEPH_ENTITY_TYPE_OSD 0x04 -#define CEPH_ENTITY_TYPE_CLIENT 0x08 -#define CEPH_ENTITY_TYPE_AUTH 0x20 - -#define CEPH_ENTITY_TYPE_ANY 0xFF - -extern const char *ceph_entity_type_name(int type); - -/* - * entity_addr -- network address - */ -struct ceph_entity_addr { - __le32 type; - __le32 nonce; /* unique id for process (e.g. pid) */ - struct sockaddr_storage in_addr; -} __attribute__ ((packed)); - -struct ceph_entity_inst { - struct ceph_entity_name name; - struct ceph_entity_addr addr; -} __attribute__ ((packed)); - - -/* used by message exchange protocol */ -#define CEPH_MSGR_TAG_READY 1 /* server->client: ready for messages */ -#define CEPH_MSGR_TAG_RESETSESSION 2 /* server->client: reset, try again */ -#define CEPH_MSGR_TAG_WAIT 3 /* server->client: wait for racing - incoming connection */ -#define CEPH_MSGR_TAG_RETRY_SESSION 4 /* server->client + cseq: try again - with higher cseq */ -#define CEPH_MSGR_TAG_RETRY_GLOBAL 5 /* server->client + gseq: try again - with higher gseq */ -#define CEPH_MSGR_TAG_CLOSE 6 /* closing pipe */ -#define CEPH_MSGR_TAG_MSG 7 /* message */ -#define CEPH_MSGR_TAG_ACK 8 /* message ack */ -#define CEPH_MSGR_TAG_KEEPALIVE 9 /* just a keepalive byte! */ -#define CEPH_MSGR_TAG_BADPROTOVER 10 /* bad protocol version */ -#define CEPH_MSGR_TAG_BADAUTHORIZER 11 /* bad authorizer */ -#define CEPH_MSGR_TAG_FEATURES 12 /* insufficient features */ - - -/* - * connection negotiation - */ -struct ceph_msg_connect { - __le64 features; /* supported feature bits */ - __le32 host_type; /* CEPH_ENTITY_TYPE_* */ - __le32 global_seq; /* count connections initiated by this host */ - __le32 connect_seq; /* count connections initiated in this session */ - __le32 protocol_version; - __le32 authorizer_protocol; - __le32 authorizer_len; - __u8 flags; /* CEPH_MSG_CONNECT_* */ -} __attribute__ ((packed)); - -struct ceph_msg_connect_reply { - __u8 tag; - __le64 features; /* feature bits for this session */ - __le32 global_seq; - __le32 connect_seq; - __le32 protocol_version; - __le32 authorizer_len; - __u8 flags; -} __attribute__ ((packed)); - -#define CEPH_MSG_CONNECT_LOSSY 1 /* messages i send may be safely dropped */ - - -/* - * message header - */ -struct ceph_msg_header_old { - __le64 seq; /* message seq# for this session */ - __le64 tid; /* transaction id */ - __le16 type; /* message type */ - __le16 priority; /* priority. higher value == higher priority */ - __le16 version; /* version of message encoding */ - - __le32 front_len; /* bytes in main payload */ - __le32 middle_len;/* bytes in middle payload */ - __le32 data_len; /* bytes of data payload */ - __le16 data_off; /* sender: include full offset; - receiver: mask against ~PAGE_MASK */ - - struct ceph_entity_inst src, orig_src; - __le32 reserved; - __le32 crc; /* header crc32c */ -} __attribute__ ((packed)); - -struct ceph_msg_header { - __le64 seq; /* message seq# for this session */ - __le64 tid; /* transaction id */ - __le16 type; /* message type */ - __le16 priority; /* priority. higher value == higher priority */ - __le16 version; /* version of message encoding */ - - __le32 front_len; /* bytes in main payload */ - __le32 middle_len;/* bytes in middle payload */ - __le32 data_len; /* bytes of data payload */ - __le16 data_off; /* sender: include full offset; - receiver: mask against ~PAGE_MASK */ - - struct ceph_entity_name src; - __le32 reserved; - __le32 crc; /* header crc32c */ -} __attribute__ ((packed)); - -#define CEPH_MSG_PRIO_LOW 64 -#define CEPH_MSG_PRIO_DEFAULT 127 -#define CEPH_MSG_PRIO_HIGH 196 -#define CEPH_MSG_PRIO_HIGHEST 255 - -/* - * follows data payload - */ -struct ceph_msg_footer { - __le32 front_crc, middle_crc, data_crc; - __u8 flags; -} __attribute__ ((packed)); - -#define CEPH_MSG_FOOTER_COMPLETE (1<<0) /* msg wasn't aborted */ -#define CEPH_MSG_FOOTER_NOCRC (1<<1) /* no data crc */ - - -#endif diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c deleted file mode 100644 index 0edb43f..0000000 --- a/fs/ceph/osd_client.c +++ /dev/null @@ -1,1771 +0,0 @@ -#include "ceph_debug.h" - -#include -#include -#include -#include -#include -#include -#ifdef CONFIG_BLOCK -#include -#endif - -#include "super.h" -#include "osd_client.h" -#include "messenger.h" -#include "decode.h" -#include "auth.h" -#include "pagelist.h" - -#define OSD_OP_FRONT_LEN 4096 -#define OSD_OPREPLY_FRONT_LEN 512 - -static const struct ceph_connection_operations osd_con_ops; -static int __kick_requests(struct ceph_osd_client *osdc, - struct ceph_osd *kickosd); - -static void kick_requests(struct ceph_osd_client *osdc, struct ceph_osd *osd); - -static int op_needs_trail(int op) -{ - switch (op) { - case CEPH_OSD_OP_GETXATTR: - case CEPH_OSD_OP_SETXATTR: - case CEPH_OSD_OP_CMPXATTR: - case CEPH_OSD_OP_CALL: - return 1; - default: - return 0; - } -} - -static int op_has_extent(int op) -{ - return (op == CEPH_OSD_OP_READ || - op == CEPH_OSD_OP_WRITE); -} - -void ceph_calc_raw_layout(struct ceph_osd_client *osdc, - struct ceph_file_layout *layout, - u64 snapid, - u64 off, u64 *plen, u64 *bno, - struct ceph_osd_request *req, - struct ceph_osd_req_op *op) -{ - struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; - u64 orig_len = *plen; - u64 objoff, objlen; /* extent in object */ - - reqhead->snapid = cpu_to_le64(snapid); - - /* object extent? */ - ceph_calc_file_object_mapping(layout, off, plen, bno, - &objoff, &objlen); - if (*plen < orig_len) - dout(" skipping last %llu, final file extent %llu~%llu\n", - orig_len - *plen, off, *plen); - - if (op_has_extent(op->op)) { - op->extent.offset = objoff; - op->extent.length = objlen; - } - req->r_num_pages = calc_pages_for(off, *plen); - - dout("calc_layout bno=%llx %llu~%llu (%d pages)\n", - *bno, objoff, objlen, req->r_num_pages); - -} - -/* - * Implement client access to distributed object storage cluster. - * - * All data objects are stored within a cluster/cloud of OSDs, or - * "object storage devices." (Note that Ceph OSDs have _nothing_ to - * do with the T10 OSD extensions to SCSI.) Ceph OSDs are simply - * remote daemons serving up and coordinating consistent and safe - * access to storage. - * - * Cluster membership and the mapping of data objects onto storage devices - * are described by the osd map. - * - * We keep track of pending OSD requests (read, write), resubmit - * requests to different OSDs when the cluster topology/data layout - * change, or retry the affected requests when the communications - * channel with an OSD is reset. - */ - -/* - * calculate the mapping of a file extent onto an object, and fill out the - * request accordingly. shorten extent as necessary if it crosses an - * object boundary. - * - * fill osd op in request message. - */ -static void calc_layout(struct ceph_osd_client *osdc, - struct ceph_vino vino, - struct ceph_file_layout *layout, - u64 off, u64 *plen, - struct ceph_osd_request *req, - struct ceph_osd_req_op *op) -{ - u64 bno; - - ceph_calc_raw_layout(osdc, layout, vino.snap, off, - plen, &bno, req, op); - - sprintf(req->r_oid, "%llx.%08llx", vino.ino, bno); - req->r_oid_len = strlen(req->r_oid); -} - -/* - * requests - */ -void ceph_osdc_release_request(struct kref *kref) -{ - struct ceph_osd_request *req = container_of(kref, - struct ceph_osd_request, - r_kref); - - if (req->r_request) - ceph_msg_put(req->r_request); - if (req->r_reply) - ceph_msg_put(req->r_reply); - if (req->r_con_filling_msg) { - dout("release_request revoking pages %p from con %p\n", - req->r_pages, req->r_con_filling_msg); - ceph_con_revoke_message(req->r_con_filling_msg, - req->r_reply); - ceph_con_put(req->r_con_filling_msg); - } - if (req->r_own_pages) - ceph_release_page_vector(req->r_pages, - req->r_num_pages); -#ifdef CONFIG_BLOCK - if (req->r_bio) - bio_put(req->r_bio); -#endif - ceph_put_snap_context(req->r_snapc); - if (req->r_trail) { - ceph_pagelist_release(req->r_trail); - kfree(req->r_trail); - } - if (req->r_mempool) - mempool_free(req, req->r_osdc->req_mempool); - else - kfree(req); -} - -static int op_needs_trail(int op) -{ - switch (op) { - case CEPH_OSD_OP_GETXATTR: - case CEPH_OSD_OP_SETXATTR: - case CEPH_OSD_OP_CMPXATTR: - case CEPH_OSD_OP_CALL: - return 1; - default: - return 0; - } -} - -static int get_num_ops(struct ceph_osd_req_op *ops, int *needs_trail) -{ - int i = 0; - - if (needs_trail) - *needs_trail = 0; - while (ops[i].op) { - if (needs_trail && op_needs_trail(ops[i].op)) - *needs_trail = 1; - i++; - } - - return i; -} - -struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, - int flags, - struct ceph_snap_context *snapc, - struct ceph_osd_req_op *ops, - bool use_mempool, - gfp_t gfp_flags, - struct page **pages, - struct bio *bio) -{ - struct ceph_osd_request *req; - struct ceph_msg *msg; - int needs_trail; - int num_op = get_num_ops(ops, &needs_trail); - size_t msg_size = sizeof(struct ceph_osd_request_head); - - msg_size += num_op*sizeof(struct ceph_osd_op); - - if (use_mempool) { - req = mempool_alloc(osdc->req_mempool, gfp_flags); - memset(req, 0, sizeof(*req)); - } else { - req = kzalloc(sizeof(*req), gfp_flags); - } - if (req == NULL) - return NULL; - - req->r_osdc = osdc; - req->r_mempool = use_mempool; - - kref_init(&req->r_kref); - init_completion(&req->r_completion); - init_completion(&req->r_safe_completion); - INIT_LIST_HEAD(&req->r_unsafe_item); - req->r_flags = flags; - - WARN_ON((flags & (CEPH_OSD_FLAG_READ|CEPH_OSD_FLAG_WRITE)) == 0); - - /* create reply message */ - if (use_mempool) - msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0); - else - msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, - OSD_OPREPLY_FRONT_LEN, gfp_flags); - if (!msg) { - ceph_osdc_put_request(req); - return NULL; - } - req->r_reply = msg; - - /* allocate space for the trailing data */ - if (needs_trail) { - req->r_trail = kmalloc(sizeof(struct ceph_pagelist), gfp_flags); - if (!req->r_trail) { - ceph_osdc_put_request(req); - return NULL; - } - ceph_pagelist_init(req->r_trail); - } - /* create request message; allow space for oid */ - msg_size += 40; - if (snapc) - msg_size += sizeof(u64) * snapc->num_snaps; - if (use_mempool) - msg = ceph_msgpool_get(&osdc->msgpool_op, 0); - else - msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags); - if (!msg) { - ceph_osdc_put_request(req); - return NULL; - } - - msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP); - memset(msg->front.iov_base, 0, msg->front.iov_len); - - req->r_request = msg; - req->r_pages = pages; -#ifdef CONFIG_BLOCK - if (bio) { - req->r_bio = bio; - bio_get(req->r_bio); - } -#endif - - return req; -} - -static void osd_req_encode_op(struct ceph_osd_request *req, - struct ceph_osd_op *dst, - struct ceph_osd_req_op *src) -{ - dst->op = cpu_to_le16(src->op); - - switch (dst->op) { - case CEPH_OSD_OP_READ: - case CEPH_OSD_OP_WRITE: - dst->extent.offset = - cpu_to_le64(src->extent.offset); - dst->extent.length = - cpu_to_le64(src->extent.length); - dst->extent.truncate_size = - cpu_to_le64(src->extent.truncate_size); - dst->extent.truncate_seq = - cpu_to_le32(src->extent.truncate_seq); - break; - - case CEPH_OSD_OP_GETXATTR: - case CEPH_OSD_OP_SETXATTR: - case CEPH_OSD_OP_CMPXATTR: - BUG_ON(!req->r_trail); - - dst->xattr.name_len = cpu_to_le32(src->xattr.name_len); - dst->xattr.value_len = cpu_to_le32(src->xattr.value_len); - dst->xattr.cmp_op = src->xattr.cmp_op; - dst->xattr.cmp_mode = src->xattr.cmp_mode; - ceph_pagelist_append(req->r_trail, src->xattr.name, - src->xattr.name_len); - ceph_pagelist_append(req->r_trail, src->xattr.val, - src->xattr.value_len); - break; - case CEPH_OSD_OP_CALL: - BUG_ON(!req->r_trail); - - dst->cls.class_len = src->cls.class_len; - dst->cls.method_len = src->cls.method_len; - dst->cls.indata_len = cpu_to_le32(src->cls.indata_len); - - ceph_pagelist_append(req->r_trail, src->cls.class_name, - src->cls.class_len); - ceph_pagelist_append(req->r_trail, src->cls.method_name, - src->cls.method_len); - ceph_pagelist_append(req->r_trail, src->cls.indata, - src->cls.indata_len); - break; - case CEPH_OSD_OP_ROLLBACK: - dst->snap.snapid = cpu_to_le64(src->snap.snapid); - break; - case CEPH_OSD_OP_STARTSYNC: - break; - default: - pr_err("unrecognized osd opcode %d\n", dst->op); - WARN_ON(1); - break; - } - dst->payload_len = cpu_to_le32(src->payload_len); -} - -/* - * build new request AND message - * - */ -void ceph_osdc_build_request(struct ceph_osd_request *req, - u64 off, u64 *plen, - struct ceph_osd_req_op *src_ops, - struct ceph_snap_context *snapc, - struct timespec *mtime, - const char *oid, - int oid_len) -{ - struct ceph_msg *msg = req->r_request; - struct ceph_osd_request_head *head; - struct ceph_osd_req_op *src_op; - struct ceph_osd_op *op; - void *p; - int num_op = get_num_ops(src_ops, NULL); - size_t msg_size = sizeof(*head) + num_op*sizeof(*op); - int flags = req->r_flags; - u64 data_len = 0; - int i; - - head = msg->front.iov_base; - op = (void *)(head + 1); - p = (void *)(op + num_op); - - req->r_snapc = ceph_get_snap_context(snapc); - - head->client_inc = cpu_to_le32(1); /* always, for now. */ - head->flags = cpu_to_le32(flags); - if (flags & CEPH_OSD_FLAG_WRITE) - ceph_encode_timespec(&head->mtime, mtime); - head->num_ops = cpu_to_le16(num_op); - - - /* fill in oid */ - head->object_len = cpu_to_le32(oid_len); - memcpy(p, oid, oid_len); - p += oid_len; - - src_op = src_ops; - while (src_op->op) { - osd_req_encode_op(req, op, src_op); - src_op++; - op++; - } - - if (req->r_trail) - data_len += req->r_trail->length; - - if (snapc) { - head->snap_seq = cpu_to_le64(snapc->seq); - head->num_snaps = cpu_to_le32(snapc->num_snaps); - for (i = 0; i < snapc->num_snaps; i++) { - put_unaligned_le64(snapc->snaps[i], p); - p += sizeof(u64); - } - } - - if (flags & CEPH_OSD_FLAG_WRITE) { - req->r_request->hdr.data_off = cpu_to_le16(off); - req->r_request->hdr.data_len = cpu_to_le32(*plen + data_len); - } else if (data_len) { - req->r_request->hdr.data_off = 0; - req->r_request->hdr.data_len = cpu_to_le32(data_len); - } - - BUG_ON(p > msg->front.iov_base + msg->front.iov_len); - msg_size = p - msg->front.iov_base; - msg->front.iov_len = msg_size; - msg->hdr.front_len = cpu_to_le32(msg_size); - return; -} - -/* - * build new request AND message, calculate layout, and adjust file - * extent as needed. - * - * if the file was recently truncated, we include information about its - * old and new size so that the object can be updated appropriately. (we - * avoid synchronously deleting truncated objects because it's slow.) - * - * if @do_sync, include a 'startsync' command so that the osd will flush - * data quickly. - */ -struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, - struct ceph_file_layout *layout, - struct ceph_vino vino, - u64 off, u64 *plen, - int opcode, int flags, - struct ceph_snap_context *snapc, - int do_sync, - u32 truncate_seq, - u64 truncate_size, - struct timespec *mtime, - bool use_mempool, int num_reply) -{ - struct ceph_osd_req_op ops[3]; - struct ceph_osd_request *req; - - ops[0].op = opcode; - ops[0].extent.truncate_seq = truncate_seq; - ops[0].extent.truncate_size = truncate_size; - ops[0].payload_len = 0; - - if (do_sync) { - ops[1].op = CEPH_OSD_OP_STARTSYNC; - ops[1].payload_len = 0; - ops[2].op = 0; - } else - ops[1].op = 0; - - req = ceph_osdc_alloc_request(osdc, flags, - snapc, ops, - use_mempool, - GFP_NOFS, NULL, NULL); - if (IS_ERR(req)) - return req; - - /* calculate max write size */ - calc_layout(osdc, vino, layout, off, plen, req, ops); - req->r_file_layout = *layout; /* keep a copy */ - - ceph_osdc_build_request(req, off, plen, ops, - snapc, - mtime, - req->r_oid, req->r_oid_len); - - return req; -} - -/* - * We keep osd requests in an rbtree, sorted by ->r_tid. - */ -static void __insert_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *new) -{ - struct rb_node **p = &osdc->requests.rb_node; - struct rb_node *parent = NULL; - struct ceph_osd_request *req = NULL; - - while (*p) { - parent = *p; - req = rb_entry(parent, struct ceph_osd_request, r_node); - if (new->r_tid < req->r_tid) - p = &(*p)->rb_left; - else if (new->r_tid > req->r_tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&new->r_node, parent, p); - rb_insert_color(&new->r_node, &osdc->requests); -} - -static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc, - u64 tid) -{ - struct ceph_osd_request *req; - struct rb_node *n = osdc->requests.rb_node; - - while (n) { - req = rb_entry(n, struct ceph_osd_request, r_node); - if (tid < req->r_tid) - n = n->rb_left; - else if (tid > req->r_tid) - n = n->rb_right; - else - return req; - } - return NULL; -} - -static struct ceph_osd_request * -__lookup_request_ge(struct ceph_osd_client *osdc, - u64 tid) -{ - struct ceph_osd_request *req; - struct rb_node *n = osdc->requests.rb_node; - - while (n) { - req = rb_entry(n, struct ceph_osd_request, r_node); - if (tid < req->r_tid) { - if (!n->rb_left) - return req; - n = n->rb_left; - } else if (tid > req->r_tid) { - n = n->rb_right; - } else { - return req; - } - } - return NULL; -} - - -/* - * If the osd connection drops, we need to resubmit all requests. - */ -static void osd_reset(struct ceph_connection *con) -{ - struct ceph_osd *osd = con->private; - struct ceph_osd_client *osdc; - - if (!osd) - return; - dout("osd_reset osd%d\n", osd->o_osd); - osdc = osd->o_osdc; - down_read(&osdc->map_sem); - kick_requests(osdc, osd); - up_read(&osdc->map_sem); -} - -/* - * Track open sessions with osds. - */ -static struct ceph_osd *create_osd(struct ceph_osd_client *osdc) -{ - struct ceph_osd *osd; - - osd = kzalloc(sizeof(*osd), GFP_NOFS); - if (!osd) - return NULL; - - atomic_set(&osd->o_ref, 1); - osd->o_osdc = osdc; - INIT_LIST_HEAD(&osd->o_requests); - INIT_LIST_HEAD(&osd->o_osd_lru); - osd->o_incarnation = 1; - - ceph_con_init(osdc->client->msgr, &osd->o_con); - osd->o_con.private = osd; - osd->o_con.ops = &osd_con_ops; - osd->o_con.peer_name.type = CEPH_ENTITY_TYPE_OSD; - - INIT_LIST_HEAD(&osd->o_keepalive_item); - return osd; -} - -static struct ceph_osd *get_osd(struct ceph_osd *osd) -{ - if (atomic_inc_not_zero(&osd->o_ref)) { - dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1, - atomic_read(&osd->o_ref)); - return osd; - } else { - dout("get_osd %p FAIL\n", osd); - return NULL; - } -} - -static void put_osd(struct ceph_osd *osd) -{ - dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref), - atomic_read(&osd->o_ref) - 1); - if (atomic_dec_and_test(&osd->o_ref)) { - struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth; - - if (osd->o_authorizer) - ac->ops->destroy_authorizer(ac, osd->o_authorizer); - kfree(osd); - } -} - -/* - * remove an osd from our map - */ -static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) -{ - dout("__remove_osd %p\n", osd); - BUG_ON(!list_empty(&osd->o_requests)); - rb_erase(&osd->o_node, &osdc->osds); - list_del_init(&osd->o_osd_lru); - ceph_con_close(&osd->o_con); - put_osd(osd); -} - -static void __move_osd_to_lru(struct ceph_osd_client *osdc, - struct ceph_osd *osd) -{ - dout("__move_osd_to_lru %p\n", osd); - BUG_ON(!list_empty(&osd->o_osd_lru)); - list_add_tail(&osd->o_osd_lru, &osdc->osd_lru); - osd->lru_ttl = jiffies + osdc->client->mount_args->osd_idle_ttl * HZ; -} - -static void __remove_osd_from_lru(struct ceph_osd *osd) -{ - dout("__remove_osd_from_lru %p\n", osd); - if (!list_empty(&osd->o_osd_lru)) - list_del_init(&osd->o_osd_lru); -} - -static void remove_old_osds(struct ceph_osd_client *osdc, int remove_all) -{ - struct ceph_osd *osd, *nosd; - - dout("__remove_old_osds %p\n", osdc); - mutex_lock(&osdc->request_mutex); - list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) { - if (!remove_all && time_before(jiffies, osd->lru_ttl)) - break; - __remove_osd(osdc, osd); - } - mutex_unlock(&osdc->request_mutex); -} - -/* - * reset osd connect - */ -static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) -{ - struct ceph_osd_request *req; - int ret = 0; - - dout("__reset_osd %p osd%d\n", osd, osd->o_osd); - if (list_empty(&osd->o_requests)) { - __remove_osd(osdc, osd); - } else if (memcmp(&osdc->osdmap->osd_addr[osd->o_osd], - &osd->o_con.peer_addr, - sizeof(osd->o_con.peer_addr)) == 0 && - !ceph_con_opened(&osd->o_con)) { - dout(" osd addr hasn't changed and connection never opened," - " letting msgr retry"); - /* touch each r_stamp for handle_timeout()'s benfit */ - list_for_each_entry(req, &osd->o_requests, r_osd_item) - req->r_stamp = jiffies; - ret = -EAGAIN; - } else { - ceph_con_close(&osd->o_con); - ceph_con_open(&osd->o_con, &osdc->osdmap->osd_addr[osd->o_osd]); - osd->o_incarnation++; - } - return ret; -} - -static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new) -{ - struct rb_node **p = &osdc->osds.rb_node; - struct rb_node *parent = NULL; - struct ceph_osd *osd = NULL; - - while (*p) { - parent = *p; - osd = rb_entry(parent, struct ceph_osd, o_node); - if (new->o_osd < osd->o_osd) - p = &(*p)->rb_left; - else if (new->o_osd > osd->o_osd) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&new->o_node, parent, p); - rb_insert_color(&new->o_node, &osdc->osds); -} - -static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o) -{ - struct ceph_osd *osd; - struct rb_node *n = osdc->osds.rb_node; - - while (n) { - osd = rb_entry(n, struct ceph_osd, o_node); - if (o < osd->o_osd) - n = n->rb_left; - else if (o > osd->o_osd) - n = n->rb_right; - else - return osd; - } - return NULL; -} - -static void __schedule_osd_timeout(struct ceph_osd_client *osdc) -{ - schedule_delayed_work(&osdc->timeout_work, - osdc->client->mount_args->osd_keepalive_timeout * HZ); -} - -static void __cancel_osd_timeout(struct ceph_osd_client *osdc) -{ - cancel_delayed_work(&osdc->timeout_work); -} - -/* - * Register request, assign tid. If this is the first request, set up - * the timeout event. - */ -static void register_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) -{ - mutex_lock(&osdc->request_mutex); - req->r_tid = ++osdc->last_tid; - req->r_request->hdr.tid = cpu_to_le64(req->r_tid); - INIT_LIST_HEAD(&req->r_req_lru_item); - - dout("register_request %p tid %lld\n", req, req->r_tid); - __insert_request(osdc, req); - ceph_osdc_get_request(req); - osdc->num_requests++; - - if (osdc->num_requests == 1) { - dout(" first request, scheduling timeout\n"); - __schedule_osd_timeout(osdc); - } - mutex_unlock(&osdc->request_mutex); -} - -/* - * called under osdc->request_mutex - */ -static void __unregister_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) -{ - dout("__unregister_request %p tid %lld\n", req, req->r_tid); - rb_erase(&req->r_node, &osdc->requests); - osdc->num_requests--; - - if (req->r_osd) { - /* make sure the original request isn't in flight. */ - ceph_con_revoke(&req->r_osd->o_con, req->r_request); - - list_del_init(&req->r_osd_item); - if (list_empty(&req->r_osd->o_requests)) - __move_osd_to_lru(osdc, req->r_osd); - req->r_osd = NULL; - } - - ceph_osdc_put_request(req); - - list_del_init(&req->r_req_lru_item); - if (osdc->num_requests == 0) { - dout(" no requests, canceling timeout\n"); - __cancel_osd_timeout(osdc); - } -} - -/* - * Cancel a previously queued request message - */ -static void __cancel_request(struct ceph_osd_request *req) -{ - if (req->r_sent && req->r_osd) { - ceph_con_revoke(&req->r_osd->o_con, req->r_request); - req->r_sent = 0; - } - list_del_init(&req->r_req_lru_item); -} - -/* - * Pick an osd (the first 'up' osd in the pg), allocate the osd struct - * (as needed), and set the request r_osd appropriately. If there is - * no up osd, set r_osd to NULL. - * - * Return 0 if unchanged, 1 if changed, or negative on error. - * - * Caller should hold map_sem for read and request_mutex. - */ -static int __map_osds(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) -{ - struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; - struct ceph_pg pgid; - int acting[CEPH_PG_MAX_SIZE]; - int o = -1, num = 0; - int err; - - dout("map_osds %p tid %lld\n", req, req->r_tid); - err = ceph_calc_object_layout(&reqhead->layout, req->r_oid, - &req->r_file_layout, osdc->osdmap); - if (err) - return err; - pgid = reqhead->layout.ol_pgid; - req->r_pgid = pgid; - - err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting); - if (err > 0) { - o = acting[0]; - num = err; - } - - if ((req->r_osd && req->r_osd->o_osd == o && - req->r_sent >= req->r_osd->o_incarnation && - req->r_num_pg_osds == num && - memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) || - (req->r_osd == NULL && o == -1)) - return 0; /* no change */ - - dout("map_osds tid %llu pgid %d.%x osd%d (was osd%d)\n", - req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o, - req->r_osd ? req->r_osd->o_osd : -1); - - /* record full pg acting set */ - memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num); - req->r_num_pg_osds = num; - - if (req->r_osd) { - __cancel_request(req); - list_del_init(&req->r_osd_item); - req->r_osd = NULL; - } - - req->r_osd = __lookup_osd(osdc, o); - if (!req->r_osd && o >= 0) { - err = -ENOMEM; - req->r_osd = create_osd(osdc); - if (!req->r_osd) - goto out; - - dout("map_osds osd %p is osd%d\n", req->r_osd, o); - req->r_osd->o_osd = o; - req->r_osd->o_con.peer_name.num = cpu_to_le64(o); - __insert_osd(osdc, req->r_osd); - - ceph_con_open(&req->r_osd->o_con, &osdc->osdmap->osd_addr[o]); - } - - if (req->r_osd) { - __remove_osd_from_lru(req->r_osd); - list_add(&req->r_osd_item, &req->r_osd->o_requests); - } - err = 1; /* osd or pg changed */ - -out: - return err; -} - -/* - * caller should hold map_sem (for read) and request_mutex - */ -static int __send_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) -{ - struct ceph_osd_request_head *reqhead; - int err; - - err = __map_osds(osdc, req); - if (err < 0) - return err; - if (req->r_osd == NULL) { - dout("send_request %p no up osds in pg\n", req); - ceph_monc_request_next_osdmap(&osdc->client->monc); - return 0; - } - - dout("send_request %p tid %llu to osd%d flags %d\n", - req, req->r_tid, req->r_osd->o_osd, req->r_flags); - - reqhead = req->r_request->front.iov_base; - reqhead->osdmap_epoch = cpu_to_le32(osdc->osdmap->epoch); - reqhead->flags |= cpu_to_le32(req->r_flags); /* e.g., RETRY */ - reqhead->reassert_version = req->r_reassert_version; - - req->r_stamp = jiffies; - list_move_tail(&req->r_req_lru_item, &osdc->req_lru); - - ceph_msg_get(req->r_request); /* send consumes a ref */ - ceph_con_send(&req->r_osd->o_con, req->r_request); - req->r_sent = req->r_osd->o_incarnation; - return 0; -} - -/* - * Timeout callback, called every N seconds when 1 or more osd - * requests has been active for more than N seconds. When this - * happens, we ping all OSDs with requests who have timed out to - * ensure any communications channel reset is detected. Reset the - * request timeouts another N seconds in the future as we go. - * Reschedule the timeout event another N seconds in future (unless - * there are no open requests). - */ -static void handle_timeout(struct work_struct *work) -{ - struct ceph_osd_client *osdc = - container_of(work, struct ceph_osd_client, timeout_work.work); - struct ceph_osd_request *req, *last_req = NULL; - struct ceph_osd *osd; - unsigned long timeout = osdc->client->mount_args->osd_timeout * HZ; - unsigned long keepalive = - osdc->client->mount_args->osd_keepalive_timeout * HZ; - unsigned long last_stamp = 0; - struct rb_node *p; - struct list_head slow_osds; - - dout("timeout\n"); - down_read(&osdc->map_sem); - - ceph_monc_request_next_osdmap(&osdc->client->monc); - - mutex_lock(&osdc->request_mutex); - for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { - req = rb_entry(p, struct ceph_osd_request, r_node); - - if (req->r_resend) { - int err; - - dout("osdc resending prev failed %lld\n", req->r_tid); - err = __send_request(osdc, req); - if (err) - dout("osdc failed again on %lld\n", req->r_tid); - else - req->r_resend = false; - continue; - } - } - - /* - * reset osds that appear to be _really_ unresponsive. this - * is a failsafe measure.. we really shouldn't be getting to - * this point if the system is working properly. the monitors - * should mark the osd as failed and we should find out about - * it from an updated osd map. - */ - while (timeout && !list_empty(&osdc->req_lru)) { - req = list_entry(osdc->req_lru.next, struct ceph_osd_request, - r_req_lru_item); - - if (time_before(jiffies, req->r_stamp + timeout)) - break; - - BUG_ON(req == last_req && req->r_stamp == last_stamp); - last_req = req; - last_stamp = req->r_stamp; - - osd = req->r_osd; - BUG_ON(!osd); - pr_warning(" tid %llu timed out on osd%d, will reset osd\n", - req->r_tid, osd->o_osd); - __kick_requests(osdc, osd); - } - - /* - * ping osds that are a bit slow. this ensures that if there - * is a break in the TCP connection we will notice, and reopen - * a connection with that osd (from the fault callback). - */ - INIT_LIST_HEAD(&slow_osds); - list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) { - if (time_before(jiffies, req->r_stamp + keepalive)) - break; - - osd = req->r_osd; - BUG_ON(!osd); - dout(" tid %llu is slow, will send keepalive on osd%d\n", - req->r_tid, osd->o_osd); - list_move_tail(&osd->o_keepalive_item, &slow_osds); - } - while (!list_empty(&slow_osds)) { - osd = list_entry(slow_osds.next, struct ceph_osd, - o_keepalive_item); - list_del_init(&osd->o_keepalive_item); - ceph_con_keepalive(&osd->o_con); - } - - __schedule_osd_timeout(osdc); - mutex_unlock(&osdc->request_mutex); - - up_read(&osdc->map_sem); -} - -static void handle_osds_timeout(struct work_struct *work) -{ - struct ceph_osd_client *osdc = - container_of(work, struct ceph_osd_client, - osds_timeout_work.work); - unsigned long delay = - osdc->client->mount_args->osd_idle_ttl * HZ >> 2; - - dout("osds timeout\n"); - down_read(&osdc->map_sem); - remove_old_osds(osdc, 0); - up_read(&osdc->map_sem); - - schedule_delayed_work(&osdc->osds_timeout_work, - round_jiffies_relative(delay)); -} - -/* - * handle osd op reply. either call the callback if it is specified, - * or do the completion to wake up the waiting thread. - */ -static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg, - struct ceph_connection *con) -{ - struct ceph_osd_reply_head *rhead = msg->front.iov_base; - struct ceph_osd_request *req; - u64 tid; - int numops, object_len, flags; - s32 result; - - tid = le64_to_cpu(msg->hdr.tid); - if (msg->front.iov_len < sizeof(*rhead)) - goto bad; - numops = le32_to_cpu(rhead->num_ops); - object_len = le32_to_cpu(rhead->object_len); - result = le32_to_cpu(rhead->result); - if (msg->front.iov_len != sizeof(*rhead) + object_len + - numops * sizeof(struct ceph_osd_op)) - goto bad; - dout("handle_reply %p tid %llu result %d\n", msg, tid, (int)result); - - /* lookup */ - mutex_lock(&osdc->request_mutex); - req = __lookup_request(osdc, tid); - if (req == NULL) { - dout("handle_reply tid %llu dne\n", tid); - mutex_unlock(&osdc->request_mutex); - return; - } - ceph_osdc_get_request(req); - flags = le32_to_cpu(rhead->flags); - - /* - * if this connection filled our message, drop our reference now, to - * avoid a (safe but slower) revoke later. - */ - if (req->r_con_filling_msg == con && req->r_reply == msg) { - dout(" dropping con_filling_msg ref %p\n", con); - req->r_con_filling_msg = NULL; - ceph_con_put(con); - } - - if (!req->r_got_reply) { - unsigned bytes; - - req->r_result = le32_to_cpu(rhead->result); - bytes = le32_to_cpu(msg->hdr.data_len); - dout("handle_reply result %d bytes %d\n", req->r_result, - bytes); - if (req->r_result == 0) - req->r_result = bytes; - - /* in case this is a write and we need to replay, */ - req->r_reassert_version = rhead->reassert_version; - - req->r_got_reply = 1; - } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) { - dout("handle_reply tid %llu dup ack\n", tid); - mutex_unlock(&osdc->request_mutex); - goto done; - } - - dout("handle_reply tid %llu flags %d\n", tid, flags); - - /* either this is a read, or we got the safe response */ - if (result < 0 || - (flags & CEPH_OSD_FLAG_ONDISK) || - ((flags & CEPH_OSD_FLAG_WRITE) == 0)) - __unregister_request(osdc, req); - - mutex_unlock(&osdc->request_mutex); - - if (req->r_callback) - req->r_callback(req, msg); - else - complete_all(&req->r_completion); - - if (flags & CEPH_OSD_FLAG_ONDISK) { - if (req->r_safe_callback) - req->r_safe_callback(req, msg); - complete_all(&req->r_safe_completion); /* fsync waiter */ - } - -done: - ceph_osdc_put_request(req); - return; - -bad: - pr_err("corrupt osd_op_reply got %d %d expected %d\n", - (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len), - (int)sizeof(*rhead)); - ceph_msg_dump(msg); -} - - -static int __kick_requests(struct ceph_osd_client *osdc, - struct ceph_osd *kickosd) -{ - struct ceph_osd_request *req; - struct rb_node *p, *n; - int needmap = 0; - int err; - - dout("kick_requests osd%d\n", kickosd ? kickosd->o_osd : -1); - if (kickosd) { - err = __reset_osd(osdc, kickosd); - if (err == -EAGAIN) - return 1; - } else { - for (p = rb_first(&osdc->osds); p; p = n) { - struct ceph_osd *osd = - rb_entry(p, struct ceph_osd, o_node); - - n = rb_next(p); - if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) || - memcmp(&osd->o_con.peer_addr, - ceph_osd_addr(osdc->osdmap, - osd->o_osd), - sizeof(struct ceph_entity_addr)) != 0) - __reset_osd(osdc, osd); - } - } - - for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { - req = rb_entry(p, struct ceph_osd_request, r_node); - - if (req->r_resend) { - dout(" r_resend set on tid %llu\n", req->r_tid); - __cancel_request(req); - goto kick; - } - if (req->r_osd && kickosd == req->r_osd) { - __cancel_request(req); - goto kick; - } - - err = __map_osds(osdc, req); - if (err == 0) - continue; /* no change */ - if (err < 0) { - /* - * FIXME: really, we should set the request - * error and fail if this isn't a 'nofail' - * request, but that's a fair bit more - * complicated to do. So retry! - */ - dout(" setting r_resend on %llu\n", req->r_tid); - req->r_resend = true; - continue; - } - if (req->r_osd == NULL) { - dout("tid %llu maps to no valid osd\n", req->r_tid); - needmap++; /* request a newer map */ - continue; - } - -kick: - dout("kicking %p tid %llu osd%d\n", req, req->r_tid, - req->r_osd ? req->r_osd->o_osd : -1); - req->r_flags |= CEPH_OSD_FLAG_RETRY; - err = __send_request(osdc, req); - if (err) { - dout(" setting r_resend on %llu\n", req->r_tid); - req->r_resend = true; - } - } - - return needmap; -} - -/* - * Resubmit osd requests whose osd or osd address has changed. Request - * a new osd map if osds are down, or we are otherwise unable to determine - * how to direct a request. - * - * Close connections to down osds. - * - * If @who is specified, resubmit requests for that specific osd. - * - * Caller should hold map_sem for read and request_mutex. - */ -static void kick_requests(struct ceph_osd_client *osdc, - struct ceph_osd *kickosd) -{ - int needmap; - - mutex_lock(&osdc->request_mutex); - needmap = __kick_requests(osdc, kickosd); - mutex_unlock(&osdc->request_mutex); - - if (needmap) { - dout("%d requests for down osds, need new map\n", needmap); - ceph_monc_request_next_osdmap(&osdc->client->monc); - } - -} -/* - * Process updated osd map. - * - * The message contains any number of incremental and full maps, normally - * indicating some sort of topology change in the cluster. Kick requests - * off to different OSDs as needed. - */ -void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg) -{ - void *p, *end, *next; - u32 nr_maps, maplen; - u32 epoch; - struct ceph_osdmap *newmap = NULL, *oldmap; - int err; - struct ceph_fsid fsid; - - dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0); - p = msg->front.iov_base; - end = p + msg->front.iov_len; - - /* verify fsid */ - ceph_decode_need(&p, end, sizeof(fsid), bad); - ceph_decode_copy(&p, &fsid, sizeof(fsid)); - if (ceph_check_fsid(osdc->client, &fsid) < 0) - return; - - down_write(&osdc->map_sem); - - /* incremental maps */ - ceph_decode_32_safe(&p, end, nr_maps, bad); - dout(" %d inc maps\n", nr_maps); - while (nr_maps > 0) { - ceph_decode_need(&p, end, 2*sizeof(u32), bad); - epoch = ceph_decode_32(&p); - maplen = ceph_decode_32(&p); - ceph_decode_need(&p, end, maplen, bad); - next = p + maplen; - if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) { - dout("applying incremental map %u len %d\n", - epoch, maplen); - newmap = osdmap_apply_incremental(&p, next, - osdc->osdmap, - osdc->client->msgr); - if (IS_ERR(newmap)) { - err = PTR_ERR(newmap); - goto bad; - } - BUG_ON(!newmap); - if (newmap != osdc->osdmap) { - ceph_osdmap_destroy(osdc->osdmap); - osdc->osdmap = newmap; - } - } else { - dout("ignoring incremental map %u len %d\n", - epoch, maplen); - } - p = next; - nr_maps--; - } - if (newmap) - goto done; - - /* full maps */ - ceph_decode_32_safe(&p, end, nr_maps, bad); - dout(" %d full maps\n", nr_maps); - while (nr_maps) { - ceph_decode_need(&p, end, 2*sizeof(u32), bad); - epoch = ceph_decode_32(&p); - maplen = ceph_decode_32(&p); - ceph_decode_need(&p, end, maplen, bad); - if (nr_maps > 1) { - dout("skipping non-latest full map %u len %d\n", - epoch, maplen); - } else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) { - dout("skipping full map %u len %d, " - "older than our %u\n", epoch, maplen, - osdc->osdmap->epoch); - } else { - dout("taking full map %u len %d\n", epoch, maplen); - newmap = osdmap_decode(&p, p+maplen); - if (IS_ERR(newmap)) { - err = PTR_ERR(newmap); - goto bad; - } - BUG_ON(!newmap); - oldmap = osdc->osdmap; - osdc->osdmap = newmap; - if (oldmap) - ceph_osdmap_destroy(oldmap); - } - p += maplen; - nr_maps--; - } - -done: - downgrade_write(&osdc->map_sem); - ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch); - if (newmap) - kick_requests(osdc, NULL); - up_read(&osdc->map_sem); - wake_up_all(&osdc->client->auth_wq); - return; - -bad: - pr_err("osdc handle_map corrupt msg\n"); - ceph_msg_dump(msg); - up_write(&osdc->map_sem); - return; -} - -/* - * Register request, send initial attempt. - */ -int ceph_osdc_start_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req, - bool nofail) -{ - int rc = 0; - - req->r_request->pages = req->r_pages; - req->r_request->nr_pages = req->r_num_pages; -#ifdef CONFIG_BLOCK - req->r_request->bio = req->r_bio; -#endif - req->r_request->trail = req->r_trail; - - register_request(osdc, req); - - down_read(&osdc->map_sem); - mutex_lock(&osdc->request_mutex); - /* - * a racing kick_requests() may have sent the message for us - * while we dropped request_mutex above, so only send now if - * the request still han't been touched yet. - */ - if (req->r_sent == 0) { - rc = __send_request(osdc, req); - if (rc) { - if (nofail) { - dout("osdc_start_request failed send, " - " marking %lld\n", req->r_tid); - req->r_resend = true; - rc = 0; - } else { - __unregister_request(osdc, req); - } - } - } - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); - return rc; -} - -/* - * wait for a request to complete - */ -int ceph_osdc_wait_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) -{ - int rc; - - rc = wait_for_completion_interruptible(&req->r_completion); - if (rc < 0) { - mutex_lock(&osdc->request_mutex); - __cancel_request(req); - __unregister_request(osdc, req); - mutex_unlock(&osdc->request_mutex); - dout("wait_request tid %llu canceled/timed out\n", req->r_tid); - return rc; - } - - dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result); - return req->r_result; -} - -/* - * sync - wait for all in-flight requests to flush. avoid starvation. - */ -void ceph_osdc_sync(struct ceph_osd_client *osdc) -{ - struct ceph_osd_request *req; - u64 last_tid, next_tid = 0; - - mutex_lock(&osdc->request_mutex); - last_tid = osdc->last_tid; - while (1) { - req = __lookup_request_ge(osdc, next_tid); - if (!req) - break; - if (req->r_tid > last_tid) - break; - - next_tid = req->r_tid + 1; - if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0) - continue; - - ceph_osdc_get_request(req); - mutex_unlock(&osdc->request_mutex); - dout("sync waiting on tid %llu (last is %llu)\n", - req->r_tid, last_tid); - wait_for_completion(&req->r_safe_completion); - mutex_lock(&osdc->request_mutex); - ceph_osdc_put_request(req); - } - mutex_unlock(&osdc->request_mutex); - dout("sync done (thru tid %llu)\n", last_tid); -} - -/* - * init, shutdown - */ -int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client) -{ - int err; - - dout("init\n"); - osdc->client = client; - osdc->osdmap = NULL; - init_rwsem(&osdc->map_sem); - init_completion(&osdc->map_waiters); - osdc->last_requested_map = 0; - mutex_init(&osdc->request_mutex); - osdc->last_tid = 0; - osdc->osds = RB_ROOT; - INIT_LIST_HEAD(&osdc->osd_lru); - osdc->requests = RB_ROOT; - INIT_LIST_HEAD(&osdc->req_lru); - osdc->num_requests = 0; - INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout); - INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout); - - schedule_delayed_work(&osdc->osds_timeout_work, - round_jiffies_relative(osdc->client->mount_args->osd_idle_ttl * HZ)); - - err = -ENOMEM; - osdc->req_mempool = mempool_create_kmalloc_pool(10, - sizeof(struct ceph_osd_request)); - if (!osdc->req_mempool) - goto out; - - err = ceph_msgpool_init(&osdc->msgpool_op, OSD_OP_FRONT_LEN, 10, true, - "osd_op"); - if (err < 0) - goto out_mempool; - err = ceph_msgpool_init(&osdc->msgpool_op_reply, - OSD_OPREPLY_FRONT_LEN, 10, true, - "osd_op_reply"); - if (err < 0) - goto out_msgpool; - return 0; - -out_msgpool: - ceph_msgpool_destroy(&osdc->msgpool_op); -out_mempool: - mempool_destroy(osdc->req_mempool); -out: - return err; -} - -void ceph_osdc_stop(struct ceph_osd_client *osdc) -{ - cancel_delayed_work_sync(&osdc->timeout_work); - cancel_delayed_work_sync(&osdc->osds_timeout_work); - if (osdc->osdmap) { - ceph_osdmap_destroy(osdc->osdmap); - osdc->osdmap = NULL; - } - remove_old_osds(osdc, 1); - mempool_destroy(osdc->req_mempool); - ceph_msgpool_destroy(&osdc->msgpool_op); - ceph_msgpool_destroy(&osdc->msgpool_op_reply); -} - -/* - * Read some contiguous pages. If we cross a stripe boundary, shorten - * *plen. Return number of bytes read, or error. - */ -int ceph_osdc_readpages(struct ceph_osd_client *osdc, - struct ceph_vino vino, struct ceph_file_layout *layout, - u64 off, u64 *plen, - u32 truncate_seq, u64 truncate_size, - struct page **pages, int num_pages) -{ - struct ceph_osd_request *req; - int rc = 0; - - dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino, - vino.snap, off, *plen); - req = ceph_osdc_new_request(osdc, layout, vino, off, plen, - CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ, - NULL, 0, truncate_seq, truncate_size, NULL, - false, 1); - if (!req) - return -ENOMEM; - - /* it may be a short read due to an object boundary */ - req->r_pages = pages; - - dout("readpages final extent is %llu~%llu (%d pages)\n", - off, *plen, req->r_num_pages); - - rc = ceph_osdc_start_request(osdc, req, false); - if (!rc) - rc = ceph_osdc_wait_request(osdc, req); - - ceph_osdc_put_request(req); - dout("readpages result %d\n", rc); - return rc; -} - -/* - * do a synchronous write on N pages - */ -int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino, - struct ceph_file_layout *layout, - struct ceph_snap_context *snapc, - u64 off, u64 len, - u32 truncate_seq, u64 truncate_size, - struct timespec *mtime, - struct page **pages, int num_pages, - int flags, int do_sync, bool nofail) -{ - struct ceph_osd_request *req; - int rc = 0; - - BUG_ON(vino.snap != CEPH_NOSNAP); - req = ceph_osdc_new_request(osdc, layout, vino, off, &len, - CEPH_OSD_OP_WRITE, - flags | CEPH_OSD_FLAG_ONDISK | - CEPH_OSD_FLAG_WRITE, - snapc, do_sync, - truncate_seq, truncate_size, mtime, - nofail, 1); - if (!req) - return -ENOMEM; - - /* it may be a short write due to an object boundary */ - req->r_pages = pages; - dout("writepages %llu~%llu (%d pages)\n", off, len, - req->r_num_pages); - - rc = ceph_osdc_start_request(osdc, req, nofail); - if (!rc) - rc = ceph_osdc_wait_request(osdc, req); - - ceph_osdc_put_request(req); - if (rc == 0) - rc = len; - dout("writepages result %d\n", rc); - return rc; -} - -/* - * handle incoming message - */ -static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) -{ - struct ceph_osd *osd = con->private; - struct ceph_osd_client *osdc; - int type = le16_to_cpu(msg->hdr.type); - - if (!osd) - goto out; - osdc = osd->o_osdc; - - switch (type) { - case CEPH_MSG_OSD_MAP: - ceph_osdc_handle_map(osdc, msg); - break; - case CEPH_MSG_OSD_OPREPLY: - handle_reply(osdc, msg, con); - break; - - default: - pr_err("received unknown message type %d %s\n", type, - ceph_msg_type_name(type)); - } -out: - ceph_msg_put(msg); -} - -/* - * lookup and return message for incoming reply. set up reply message - * pages. - */ -static struct ceph_msg *get_reply(struct ceph_connection *con, - struct ceph_msg_header *hdr, - int *skip) -{ - struct ceph_osd *osd = con->private; - struct ceph_osd_client *osdc = osd->o_osdc; - struct ceph_msg *m; - struct ceph_osd_request *req; - int front = le32_to_cpu(hdr->front_len); - int data_len = le32_to_cpu(hdr->data_len); - u64 tid; - - tid = le64_to_cpu(hdr->tid); - mutex_lock(&osdc->request_mutex); - req = __lookup_request(osdc, tid); - if (!req) { - *skip = 1; - m = NULL; - pr_info("get_reply unknown tid %llu from osd%d\n", tid, - osd->o_osd); - goto out; - } - - if (req->r_con_filling_msg) { - dout("get_reply revoking msg %p from old con %p\n", - req->r_reply, req->r_con_filling_msg); - ceph_con_revoke_message(req->r_con_filling_msg, req->r_reply); - ceph_con_put(req->r_con_filling_msg); - req->r_con_filling_msg = NULL; - } - - if (front > req->r_reply->front.iov_len) { - pr_warning("get_reply front %d > preallocated %d\n", - front, (int)req->r_reply->front.iov_len); - m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS); - if (!m) - goto out; - ceph_msg_put(req->r_reply); - req->r_reply = m; - } - m = ceph_msg_get(req->r_reply); - - if (data_len > 0) { - unsigned data_off = le16_to_cpu(hdr->data_off); - int want = calc_pages_for(data_off & ~PAGE_MASK, data_len); - - if (unlikely(req->r_num_pages < want)) { - pr_warning("tid %lld reply %d > expected %d pages\n", - tid, want, m->nr_pages); - *skip = 1; - ceph_msg_put(m); - m = NULL; - goto out; - } - m->pages = req->r_pages; - m->nr_pages = req->r_num_pages; -#ifdef CONFIG_BLOCK - m->bio = req->r_bio; -#endif - } - *skip = 0; - req->r_con_filling_msg = ceph_con_get(con); - dout("get_reply tid %lld %p\n", tid, m); - -out: - mutex_unlock(&osdc->request_mutex); - return m; - -} - -static struct ceph_msg *alloc_msg(struct ceph_connection *con, - struct ceph_msg_header *hdr, - int *skip) -{ - struct ceph_osd *osd = con->private; - int type = le16_to_cpu(hdr->type); - int front = le32_to_cpu(hdr->front_len); - - switch (type) { - case CEPH_MSG_OSD_MAP: - return ceph_msg_new(type, front, GFP_NOFS); - case CEPH_MSG_OSD_OPREPLY: - return get_reply(con, hdr, skip); - default: - pr_info("alloc_msg unexpected msg type %d from osd%d\n", type, - osd->o_osd); - *skip = 1; - return NULL; - } -} - -/* - * Wrappers to refcount containing ceph_osd struct - */ -static struct ceph_connection *get_osd_con(struct ceph_connection *con) -{ - struct ceph_osd *osd = con->private; - if (get_osd(osd)) - return con; - return NULL; -} - -static void put_osd_con(struct ceph_connection *con) -{ - struct ceph_osd *osd = con->private; - put_osd(osd); -} - -/* - * authentication - */ -static int get_authorizer(struct ceph_connection *con, - void **buf, int *len, int *proto, - void **reply_buf, int *reply_len, int force_new) -{ - struct ceph_osd *o = con->private; - struct ceph_osd_client *osdc = o->o_osdc; - struct ceph_auth_client *ac = osdc->client->monc.auth; - int ret = 0; - - if (force_new && o->o_authorizer) { - ac->ops->destroy_authorizer(ac, o->o_authorizer); - o->o_authorizer = NULL; - } - if (o->o_authorizer == NULL) { - ret = ac->ops->create_authorizer( - ac, CEPH_ENTITY_TYPE_OSD, - &o->o_authorizer, - &o->o_authorizer_buf, - &o->o_authorizer_buf_len, - &o->o_authorizer_reply_buf, - &o->o_authorizer_reply_buf_len); - if (ret) - return ret; - } - - *proto = ac->protocol; - *buf = o->o_authorizer_buf; - *len = o->o_authorizer_buf_len; - *reply_buf = o->o_authorizer_reply_buf; - *reply_len = o->o_authorizer_reply_buf_len; - return 0; -} - - -static int verify_authorizer_reply(struct ceph_connection *con, int len) -{ - struct ceph_osd *o = con->private; - struct ceph_osd_client *osdc = o->o_osdc; - struct ceph_auth_client *ac = osdc->client->monc.auth; - - return ac->ops->verify_authorizer_reply(ac, o->o_authorizer, len); -} - -static int invalidate_authorizer(struct ceph_connection *con) -{ - struct ceph_osd *o = con->private; - struct ceph_osd_client *osdc = o->o_osdc; - struct ceph_auth_client *ac = osdc->client->monc.auth; - - if (ac->ops->invalidate_authorizer) - ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD); - - return ceph_monc_validate_auth(&osdc->client->monc); -} - -static const struct ceph_connection_operations osd_con_ops = { - .get = get_osd_con, - .put = put_osd_con, - .dispatch = dispatch, - .get_authorizer = get_authorizer, - .verify_authorizer_reply = verify_authorizer_reply, - .invalidate_authorizer = invalidate_authorizer, - .alloc_msg = alloc_msg, - .fault = osd_reset, -}; diff --git a/fs/ceph/osd_client.h b/fs/ceph/osd_client.h deleted file mode 100644 index 65c9c56..0000000 --- a/fs/ceph/osd_client.h +++ /dev/null @@ -1,233 +0,0 @@ -#ifndef _FS_CEPH_OSD_CLIENT_H -#define _FS_CEPH_OSD_CLIENT_H - -#include -#include -#include -#include - -#include "types.h" -#include "osdmap.h" -#include "messenger.h" - -struct ceph_msg; -struct ceph_snap_context; -struct ceph_osd_request; -struct ceph_osd_client; -struct ceph_authorizer; -struct ceph_pagelist; - -/* - * completion callback for async writepages - */ -typedef void (*ceph_osdc_callback_t)(struct ceph_osd_request *, - struct ceph_msg *); - -/* a given osd we're communicating with */ -struct ceph_osd { - atomic_t o_ref; - struct ceph_osd_client *o_osdc; - int o_osd; - int o_incarnation; - struct rb_node o_node; - struct ceph_connection o_con; - struct list_head o_requests; - struct list_head o_osd_lru; - struct ceph_authorizer *o_authorizer; - void *o_authorizer_buf, *o_authorizer_reply_buf; - size_t o_authorizer_buf_len, o_authorizer_reply_buf_len; - unsigned long lru_ttl; - int o_marked_for_keepalive; - struct list_head o_keepalive_item; -}; - -/* an in-flight request */ -struct ceph_osd_request { - u64 r_tid; /* unique for this client */ - struct rb_node r_node; - struct list_head r_req_lru_item; - struct list_head r_osd_item; - struct ceph_osd *r_osd; - struct ceph_pg r_pgid; - int r_pg_osds[CEPH_PG_MAX_SIZE]; - int r_num_pg_osds; - - struct ceph_connection *r_con_filling_msg; - - struct ceph_msg *r_request, *r_reply; - int r_result; - int r_flags; /* any additional flags for the osd */ - u32 r_sent; /* >0 if r_request is sending/sent */ - int r_got_reply; - - struct ceph_osd_client *r_osdc; - struct kref r_kref; - bool r_mempool; - struct completion r_completion, r_safe_completion; - ceph_osdc_callback_t r_callback, r_safe_callback; - struct ceph_eversion r_reassert_version; - struct list_head r_unsafe_item; - - struct inode *r_inode; /* for use by callbacks */ - - char r_oid[40]; /* object name */ - int r_oid_len; - unsigned long r_stamp; /* send OR check time */ - bool r_resend; /* msg send failed, needs retry */ - - struct ceph_file_layout r_file_layout; - struct ceph_snap_context *r_snapc; /* snap context for writes */ - unsigned r_num_pages; /* size of page array (follows) */ - struct page **r_pages; /* pages for data payload */ - int r_pages_from_pool; - int r_own_pages; /* if true, i own page list */ -#ifdef CONFIG_BLOCK - struct bio *r_bio; /* instead of pages */ -#endif - - struct ceph_pagelist *r_trail; /* trailing part of the data */ -}; - -struct ceph_osd_client { - struct ceph_client *client; - - struct ceph_osdmap *osdmap; /* current map */ - struct rw_semaphore map_sem; - struct completion map_waiters; - u64 last_requested_map; - - struct mutex request_mutex; - struct rb_root osds; /* osds */ - struct list_head osd_lru; /* idle osds */ - u64 timeout_tid; /* tid of timeout triggering rq */ - u64 last_tid; /* tid of last request */ - struct rb_root requests; /* pending requests */ - struct list_head req_lru; /* pending requests lru */ - int num_requests; - struct delayed_work timeout_work; - struct delayed_work osds_timeout_work; -#ifdef CONFIG_DEBUG_FS - struct dentry *debugfs_file; -#endif - - mempool_t *req_mempool; - - struct ceph_msgpool msgpool_op; - struct ceph_msgpool msgpool_op_reply; -}; - -struct ceph_osd_req_op { - u16 op; /* CEPH_OSD_OP_* */ - u32 flags; /* CEPH_OSD_FLAG_* */ - union { - struct { - u64 offset, length; - u64 truncate_size; - u32 truncate_seq; - } extent; - struct { - const char *name; - u32 name_len; - const char *val; - u32 value_len; - __u8 cmp_op; /* CEPH_OSD_CMPXATTR_OP_* */ - __u8 cmp_mode; /* CEPH_OSD_CMPXATTR_MODE_* */ - } xattr; - struct { - const char *class_name; - __u8 class_len; - const char *method_name; - __u8 method_len; - __u8 argc; - const char *indata; - u32 indata_len; - } cls; - struct { - u64 cookie, count; - } pgls; - struct { - u64 snapid; - } snap; - }; - u32 payload_len; -}; - -extern int ceph_osdc_init(struct ceph_osd_client *osdc, - struct ceph_client *client); -extern void ceph_osdc_stop(struct ceph_osd_client *osdc); - -extern void ceph_osdc_handle_reply(struct ceph_osd_client *osdc, - struct ceph_msg *msg); -extern void ceph_osdc_handle_map(struct ceph_osd_client *osdc, - struct ceph_msg *msg); - -extern void ceph_calc_raw_layout(struct ceph_osd_client *osdc, - struct ceph_file_layout *layout, - u64 snapid, - u64 off, u64 *plen, u64 *bno, - struct ceph_osd_request *req, - struct ceph_osd_req_op *op); - -extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, - int flags, - struct ceph_snap_context *snapc, - struct ceph_osd_req_op *ops, - bool use_mempool, - gfp_t gfp_flags, - struct page **pages, - struct bio *bio); - -extern void ceph_osdc_build_request(struct ceph_osd_request *req, - u64 off, u64 *plen, - struct ceph_osd_req_op *src_ops, - struct ceph_snap_context *snapc, - struct timespec *mtime, - const char *oid, - int oid_len); - -extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *, - struct ceph_file_layout *layout, - struct ceph_vino vino, - u64 offset, u64 *len, int op, int flags, - struct ceph_snap_context *snapc, - int do_sync, u32 truncate_seq, - u64 truncate_size, - struct timespec *mtime, - bool use_mempool, int num_reply); - -static inline void ceph_osdc_get_request(struct ceph_osd_request *req) -{ - kref_get(&req->r_kref); -} -extern void ceph_osdc_release_request(struct kref *kref); -static inline void ceph_osdc_put_request(struct ceph_osd_request *req) -{ - kref_put(&req->r_kref, ceph_osdc_release_request); -} - -extern int ceph_osdc_start_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req, - bool nofail); -extern int ceph_osdc_wait_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req); -extern void ceph_osdc_sync(struct ceph_osd_client *osdc); - -extern int ceph_osdc_readpages(struct ceph_osd_client *osdc, - struct ceph_vino vino, - struct ceph_file_layout *layout, - u64 off, u64 *plen, - u32 truncate_seq, u64 truncate_size, - struct page **pages, int nr_pages); - -extern int ceph_osdc_writepages(struct ceph_osd_client *osdc, - struct ceph_vino vino, - struct ceph_file_layout *layout, - struct ceph_snap_context *sc, - u64 off, u64 len, - u32 truncate_seq, u64 truncate_size, - struct timespec *mtime, - struct page **pages, int nr_pages, - int flags, int do_sync, bool nofail); - -#endif - diff --git a/fs/ceph/osdmap.c b/fs/ceph/osdmap.c deleted file mode 100644 index 3ccd117..0000000 --- a/fs/ceph/osdmap.c +++ /dev/null @@ -1,1123 +0,0 @@ - -#include "ceph_debug.h" - -#include -#include - -#include "super.h" -#include "osdmap.h" -#include "crush/hash.h" -#include "crush/mapper.h" -#include "decode.h" - -char *ceph_osdmap_state_str(char *str, int len, int state) -{ - int flag = 0; - - if (!len) - goto done; - - *str = '\0'; - if (state) { - if (state & CEPH_OSD_EXISTS) { - snprintf(str, len, "exists"); - flag = 1; - } - if (state & CEPH_OSD_UP) { - snprintf(str, len, "%s%s%s", str, (flag ? ", " : ""), - "up"); - flag = 1; - } - } else { - snprintf(str, len, "doesn't exist"); - } -done: - return str; -} - -/* maps */ - -static int calc_bits_of(unsigned t) -{ - int b = 0; - while (t) { - t = t >> 1; - b++; - } - return b; -} - -/* - * the foo_mask is the smallest value 2^n-1 that is >= foo. - */ -static void calc_pg_masks(struct ceph_pg_pool_info *pi) -{ - pi->pg_num_mask = (1 << calc_bits_of(le32_to_cpu(pi->v.pg_num)-1)) - 1; - pi->pgp_num_mask = - (1 << calc_bits_of(le32_to_cpu(pi->v.pgp_num)-1)) - 1; - pi->lpg_num_mask = - (1 << calc_bits_of(le32_to_cpu(pi->v.lpg_num)-1)) - 1; - pi->lpgp_num_mask = - (1 << calc_bits_of(le32_to_cpu(pi->v.lpgp_num)-1)) - 1; -} - -/* - * decode crush map - */ -static int crush_decode_uniform_bucket(void **p, void *end, - struct crush_bucket_uniform *b) -{ - dout("crush_decode_uniform_bucket %p to %p\n", *p, end); - ceph_decode_need(p, end, (1+b->h.size) * sizeof(u32), bad); - b->item_weight = ceph_decode_32(p); - return 0; -bad: - return -EINVAL; -} - -static int crush_decode_list_bucket(void **p, void *end, - struct crush_bucket_list *b) -{ - int j; - dout("crush_decode_list_bucket %p to %p\n", *p, end); - b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); - if (b->item_weights == NULL) - return -ENOMEM; - b->sum_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); - if (b->sum_weights == NULL) - return -ENOMEM; - ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad); - for (j = 0; j < b->h.size; j++) { - b->item_weights[j] = ceph_decode_32(p); - b->sum_weights[j] = ceph_decode_32(p); - } - return 0; -bad: - return -EINVAL; -} - -static int crush_decode_tree_bucket(void **p, void *end, - struct crush_bucket_tree *b) -{ - int j; - dout("crush_decode_tree_bucket %p to %p\n", *p, end); - ceph_decode_32_safe(p, end, b->num_nodes, bad); - b->node_weights = kcalloc(b->num_nodes, sizeof(u32), GFP_NOFS); - if (b->node_weights == NULL) - return -ENOMEM; - ceph_decode_need(p, end, b->num_nodes * sizeof(u32), bad); - for (j = 0; j < b->num_nodes; j++) - b->node_weights[j] = ceph_decode_32(p); - return 0; -bad: - return -EINVAL; -} - -static int crush_decode_straw_bucket(void **p, void *end, - struct crush_bucket_straw *b) -{ - int j; - dout("crush_decode_straw_bucket %p to %p\n", *p, end); - b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); - if (b->item_weights == NULL) - return -ENOMEM; - b->straws = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); - if (b->straws == NULL) - return -ENOMEM; - ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad); - for (j = 0; j < b->h.size; j++) { - b->item_weights[j] = ceph_decode_32(p); - b->straws[j] = ceph_decode_32(p); - } - return 0; -bad: - return -EINVAL; -} - -static struct crush_map *crush_decode(void *pbyval, void *end) -{ - struct crush_map *c; - int err = -EINVAL; - int i, j; - void **p = &pbyval; - void *start = pbyval; - u32 magic; - - dout("crush_decode %p to %p len %d\n", *p, end, (int)(end - *p)); - - c = kzalloc(sizeof(*c), GFP_NOFS); - if (c == NULL) - return ERR_PTR(-ENOMEM); - - ceph_decode_need(p, end, 4*sizeof(u32), bad); - magic = ceph_decode_32(p); - if (magic != CRUSH_MAGIC) { - pr_err("crush_decode magic %x != current %x\n", - (unsigned)magic, (unsigned)CRUSH_MAGIC); - goto bad; - } - c->max_buckets = ceph_decode_32(p); - c->max_rules = ceph_decode_32(p); - c->max_devices = ceph_decode_32(p); - - c->device_parents = kcalloc(c->max_devices, sizeof(u32), GFP_NOFS); - if (c->device_parents == NULL) - goto badmem; - c->bucket_parents = kcalloc(c->max_buckets, sizeof(u32), GFP_NOFS); - if (c->bucket_parents == NULL) - goto badmem; - - c->buckets = kcalloc(c->max_buckets, sizeof(*c->buckets), GFP_NOFS); - if (c->buckets == NULL) - goto badmem; - c->rules = kcalloc(c->max_rules, sizeof(*c->rules), GFP_NOFS); - if (c->rules == NULL) - goto badmem; - - /* buckets */ - for (i = 0; i < c->max_buckets; i++) { - int size = 0; - u32 alg; - struct crush_bucket *b; - - ceph_decode_32_safe(p, end, alg, bad); - if (alg == 0) { - c->buckets[i] = NULL; - continue; - } - dout("crush_decode bucket %d off %x %p to %p\n", - i, (int)(*p-start), *p, end); - - switch (alg) { - case CRUSH_BUCKET_UNIFORM: - size = sizeof(struct crush_bucket_uniform); - break; - case CRUSH_BUCKET_LIST: - size = sizeof(struct crush_bucket_list); - break; - case CRUSH_BUCKET_TREE: - size = sizeof(struct crush_bucket_tree); - break; - case CRUSH_BUCKET_STRAW: - size = sizeof(struct crush_bucket_straw); - break; - default: - err = -EINVAL; - goto bad; - } - BUG_ON(size == 0); - b = c->buckets[i] = kzalloc(size, GFP_NOFS); - if (b == NULL) - goto badmem; - - ceph_decode_need(p, end, 4*sizeof(u32), bad); - b->id = ceph_decode_32(p); - b->type = ceph_decode_16(p); - b->alg = ceph_decode_8(p); - b->hash = ceph_decode_8(p); - b->weight = ceph_decode_32(p); - b->size = ceph_decode_32(p); - - dout("crush_decode bucket size %d off %x %p to %p\n", - b->size, (int)(*p-start), *p, end); - - b->items = kcalloc(b->size, sizeof(__s32), GFP_NOFS); - if (b->items == NULL) - goto badmem; - b->perm = kcalloc(b->size, sizeof(u32), GFP_NOFS); - if (b->perm == NULL) - goto badmem; - b->perm_n = 0; - - ceph_decode_need(p, end, b->size*sizeof(u32), bad); - for (j = 0; j < b->size; j++) - b->items[j] = ceph_decode_32(p); - - switch (b->alg) { - case CRUSH_BUCKET_UNIFORM: - err = crush_decode_uniform_bucket(p, end, - (struct crush_bucket_uniform *)b); - if (err < 0) - goto bad; - break; - case CRUSH_BUCKET_LIST: - err = crush_decode_list_bucket(p, end, - (struct crush_bucket_list *)b); - if (err < 0) - goto bad; - break; - case CRUSH_BUCKET_TREE: - err = crush_decode_tree_bucket(p, end, - (struct crush_bucket_tree *)b); - if (err < 0) - goto bad; - break; - case CRUSH_BUCKET_STRAW: - err = crush_decode_straw_bucket(p, end, - (struct crush_bucket_straw *)b); - if (err < 0) - goto bad; - break; - } - } - - /* rules */ - dout("rule vec is %p\n", c->rules); - for (i = 0; i < c->max_rules; i++) { - u32 yes; - struct crush_rule *r; - - ceph_decode_32_safe(p, end, yes, bad); - if (!yes) { - dout("crush_decode NO rule %d off %x %p to %p\n", - i, (int)(*p-start), *p, end); - c->rules[i] = NULL; - continue; - } - - dout("crush_decode rule %d off %x %p to %p\n", - i, (int)(*p-start), *p, end); - - /* len */ - ceph_decode_32_safe(p, end, yes, bad); -#if BITS_PER_LONG == 32 - err = -EINVAL; - if (yes > ULONG_MAX / sizeof(struct crush_rule_step)) - goto bad; -#endif - r = c->rules[i] = kmalloc(sizeof(*r) + - yes*sizeof(struct crush_rule_step), - GFP_NOFS); - if (r == NULL) - goto badmem; - dout(" rule %d is at %p\n", i, r); - r->len = yes; - ceph_decode_copy_safe(p, end, &r->mask, 4, bad); /* 4 u8's */ - ceph_decode_need(p, end, r->len*3*sizeof(u32), bad); - for (j = 0; j < r->len; j++) { - r->steps[j].op = ceph_decode_32(p); - r->steps[j].arg1 = ceph_decode_32(p); - r->steps[j].arg2 = ceph_decode_32(p); - } - } - - /* ignore trailing name maps. */ - - dout("crush_decode success\n"); - return c; - -badmem: - err = -ENOMEM; -bad: - dout("crush_decode fail %d\n", err); - crush_destroy(c); - return ERR_PTR(err); -} - -/* - * rbtree of pg_mapping for handling pg_temp (explicit mapping of pgid - * to a set of osds) - */ -static int pgid_cmp(struct ceph_pg l, struct ceph_pg r) -{ - u64 a = *(u64 *)&l; - u64 b = *(u64 *)&r; - - if (a < b) - return -1; - if (a > b) - return 1; - return 0; -} - -static int __insert_pg_mapping(struct ceph_pg_mapping *new, - struct rb_root *root) -{ - struct rb_node **p = &root->rb_node; - struct rb_node *parent = NULL; - struct ceph_pg_mapping *pg = NULL; - int c; - - while (*p) { - parent = *p; - pg = rb_entry(parent, struct ceph_pg_mapping, node); - c = pgid_cmp(new->pgid, pg->pgid); - if (c < 0) - p = &(*p)->rb_left; - else if (c > 0) - p = &(*p)->rb_right; - else - return -EEXIST; - } - - rb_link_node(&new->node, parent, p); - rb_insert_color(&new->node, root); - return 0; -} - -static struct ceph_pg_mapping *__lookup_pg_mapping(struct rb_root *root, - struct ceph_pg pgid) -{ - struct rb_node *n = root->rb_node; - struct ceph_pg_mapping *pg; - int c; - - while (n) { - pg = rb_entry(n, struct ceph_pg_mapping, node); - c = pgid_cmp(pgid, pg->pgid); - if (c < 0) - n = n->rb_left; - else if (c > 0) - n = n->rb_right; - else - return pg; - } - return NULL; -} - -/* - * rbtree of pg pool info - */ -static int __insert_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *new) -{ - struct rb_node **p = &root->rb_node; - struct rb_node *parent = NULL; - struct ceph_pg_pool_info *pi = NULL; - - while (*p) { - parent = *p; - pi = rb_entry(parent, struct ceph_pg_pool_info, node); - if (new->id < pi->id) - p = &(*p)->rb_left; - else if (new->id > pi->id) - p = &(*p)->rb_right; - else - return -EEXIST; - } - - rb_link_node(&new->node, parent, p); - rb_insert_color(&new->node, root); - return 0; -} - -static struct ceph_pg_pool_info *__lookup_pg_pool(struct rb_root *root, int id) -{ - struct ceph_pg_pool_info *pi; - struct rb_node *n = root->rb_node; - - while (n) { - pi = rb_entry(n, struct ceph_pg_pool_info, node); - if (id < pi->id) - n = n->rb_left; - else if (id > pi->id) - n = n->rb_right; - else - return pi; - } - return NULL; -} - -int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name) -{ - struct rb_node *rbp; - - for (rbp = rb_first(&map->pg_pools); rbp; rbp = rb_next(rbp)) { - struct ceph_pg_pool_info *pi = - rb_entry(rbp, struct ceph_pg_pool_info, node); - if (pi->name && strcmp(pi->name, name) == 0) - return pi->id; - } - return -ENOENT; -} - -static void __remove_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *pi) -{ - rb_erase(&pi->node, root); - kfree(pi->name); - kfree(pi); -} - -static int __decode_pool(void **p, void *end, struct ceph_pg_pool_info *pi) -{ - unsigned n, m; - - ceph_decode_copy(p, &pi->v, sizeof(pi->v)); - calc_pg_masks(pi); - - /* num_snaps * snap_info_t */ - n = le32_to_cpu(pi->v.num_snaps); - while (n--) { - ceph_decode_need(p, end, sizeof(u64) + 1 + sizeof(u64) + - sizeof(struct ceph_timespec), bad); - *p += sizeof(u64) + /* key */ - 1 + sizeof(u64) + /* u8, snapid */ - sizeof(struct ceph_timespec); - m = ceph_decode_32(p); /* snap name */ - *p += m; - } - - *p += le32_to_cpu(pi->v.num_removed_snap_intervals) * sizeof(u64) * 2; - return 0; - -bad: - return -EINVAL; -} - -static int __decode_pool_names(void **p, void *end, struct ceph_osdmap *map) -{ - struct ceph_pg_pool_info *pi; - u32 num, len, pool; - - ceph_decode_32_safe(p, end, num, bad); - dout(" %d pool names\n", num); - while (num--) { - ceph_decode_32_safe(p, end, pool, bad); - ceph_decode_32_safe(p, end, len, bad); - dout(" pool %d len %d\n", pool, len); - pi = __lookup_pg_pool(&map->pg_pools, pool); - if (pi) { - kfree(pi->name); - pi->name = kmalloc(len + 1, GFP_NOFS); - if (pi->name) { - memcpy(pi->name, *p, len); - pi->name[len] = '\0'; - dout(" name is %s\n", pi->name); - } - } - *p += len; - } - return 0; - -bad: - return -EINVAL; -} - -/* - * osd map - */ -void ceph_osdmap_destroy(struct ceph_osdmap *map) -{ - dout("osdmap_destroy %p\n", map); - if (map->crush) - crush_destroy(map->crush); - while (!RB_EMPTY_ROOT(&map->pg_temp)) { - struct ceph_pg_mapping *pg = - rb_entry(rb_first(&map->pg_temp), - struct ceph_pg_mapping, node); - rb_erase(&pg->node, &map->pg_temp); - kfree(pg); - } - while (!RB_EMPTY_ROOT(&map->pg_pools)) { - struct ceph_pg_pool_info *pi = - rb_entry(rb_first(&map->pg_pools), - struct ceph_pg_pool_info, node); - __remove_pg_pool(&map->pg_pools, pi); - } - kfree(map->osd_state); - kfree(map->osd_weight); - kfree(map->osd_addr); - kfree(map); -} - -/* - * adjust max osd value. reallocate arrays. - */ -static int osdmap_set_max_osd(struct ceph_osdmap *map, int max) -{ - u8 *state; - struct ceph_entity_addr *addr; - u32 *weight; - - state = kcalloc(max, sizeof(*state), GFP_NOFS); - addr = kcalloc(max, sizeof(*addr), GFP_NOFS); - weight = kcalloc(max, sizeof(*weight), GFP_NOFS); - if (state == NULL || addr == NULL || weight == NULL) { - kfree(state); - kfree(addr); - kfree(weight); - return -ENOMEM; - } - - /* copy old? */ - if (map->osd_state) { - memcpy(state, map->osd_state, map->max_osd*sizeof(*state)); - memcpy(addr, map->osd_addr, map->max_osd*sizeof(*addr)); - memcpy(weight, map->osd_weight, map->max_osd*sizeof(*weight)); - kfree(map->osd_state); - kfree(map->osd_addr); - kfree(map->osd_weight); - } - - map->osd_state = state; - map->osd_weight = weight; - map->osd_addr = addr; - map->max_osd = max; - return 0; -} - -/* - * decode a full map. - */ -struct ceph_osdmap *osdmap_decode(void **p, void *end) -{ - struct ceph_osdmap *map; - u16 version; - u32 len, max, i; - u8 ev; - int err = -EINVAL; - void *start = *p; - struct ceph_pg_pool_info *pi; - - dout("osdmap_decode %p to %p len %d\n", *p, end, (int)(end - *p)); - - map = kzalloc(sizeof(*map), GFP_NOFS); - if (map == NULL) - return ERR_PTR(-ENOMEM); - map->pg_temp = RB_ROOT; - - ceph_decode_16_safe(p, end, version, bad); - if (version > CEPH_OSDMAP_VERSION) { - pr_warning("got unknown v %d > %d of osdmap\n", version, - CEPH_OSDMAP_VERSION); - goto bad; - } - - ceph_decode_need(p, end, 2*sizeof(u64)+6*sizeof(u32), bad); - ceph_decode_copy(p, &map->fsid, sizeof(map->fsid)); - map->epoch = ceph_decode_32(p); - ceph_decode_copy(p, &map->created, sizeof(map->created)); - ceph_decode_copy(p, &map->modified, sizeof(map->modified)); - - ceph_decode_32_safe(p, end, max, bad); - while (max--) { - ceph_decode_need(p, end, 4 + 1 + sizeof(pi->v), bad); - pi = kzalloc(sizeof(*pi), GFP_NOFS); - if (!pi) - goto bad; - pi->id = ceph_decode_32(p); - ev = ceph_decode_8(p); /* encoding version */ - if (ev > CEPH_PG_POOL_VERSION) { - pr_warning("got unknown v %d > %d of ceph_pg_pool\n", - ev, CEPH_PG_POOL_VERSION); - kfree(pi); - goto bad; - } - err = __decode_pool(p, end, pi); - if (err < 0) - goto bad; - __insert_pg_pool(&map->pg_pools, pi); - } - - if (version >= 5 && __decode_pool_names(p, end, map) < 0) - goto bad; - - ceph_decode_32_safe(p, end, map->pool_max, bad); - - ceph_decode_32_safe(p, end, map->flags, bad); - - max = ceph_decode_32(p); - - /* (re)alloc osd arrays */ - err = osdmap_set_max_osd(map, max); - if (err < 0) - goto bad; - dout("osdmap_decode max_osd = %d\n", map->max_osd); - - /* osds */ - err = -EINVAL; - ceph_decode_need(p, end, 3*sizeof(u32) + - map->max_osd*(1 + sizeof(*map->osd_weight) + - sizeof(*map->osd_addr)), bad); - *p += 4; /* skip length field (should match max) */ - ceph_decode_copy(p, map->osd_state, map->max_osd); - - *p += 4; /* skip length field (should match max) */ - for (i = 0; i < map->max_osd; i++) - map->osd_weight[i] = ceph_decode_32(p); - - *p += 4; /* skip length field (should match max) */ - ceph_decode_copy(p, map->osd_addr, map->max_osd*sizeof(*map->osd_addr)); - for (i = 0; i < map->max_osd; i++) - ceph_decode_addr(&map->osd_addr[i]); - - /* pg_temp */ - ceph_decode_32_safe(p, end, len, bad); - for (i = 0; i < len; i++) { - int n, j; - struct ceph_pg pgid; - struct ceph_pg_mapping *pg; - - ceph_decode_need(p, end, sizeof(u32) + sizeof(u64), bad); - ceph_decode_copy(p, &pgid, sizeof(pgid)); - n = ceph_decode_32(p); - ceph_decode_need(p, end, n * sizeof(u32), bad); - err = -ENOMEM; - pg = kmalloc(sizeof(*pg) + n*sizeof(u32), GFP_NOFS); - if (!pg) - goto bad; - pg->pgid = pgid; - pg->len = n; - for (j = 0; j < n; j++) - pg->osds[j] = ceph_decode_32(p); - - err = __insert_pg_mapping(pg, &map->pg_temp); - if (err) - goto bad; - dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid, len); - } - - /* crush */ - ceph_decode_32_safe(p, end, len, bad); - dout("osdmap_decode crush len %d from off 0x%x\n", len, - (int)(*p - start)); - ceph_decode_need(p, end, len, bad); - map->crush = crush_decode(*p, end); - *p += len; - if (IS_ERR(map->crush)) { - err = PTR_ERR(map->crush); - map->crush = NULL; - goto bad; - } - - /* ignore the rest of the map */ - *p = end; - - dout("osdmap_decode done %p %p\n", *p, end); - return map; - -bad: - dout("osdmap_decode fail\n"); - ceph_osdmap_destroy(map); - return ERR_PTR(err); -} - -/* - * decode and apply an incremental map update. - */ -struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end, - struct ceph_osdmap *map, - struct ceph_messenger *msgr) -{ - struct crush_map *newcrush = NULL; - struct ceph_fsid fsid; - u32 epoch = 0; - struct ceph_timespec modified; - u32 len, pool; - __s32 new_pool_max, new_flags, max; - void *start = *p; - int err = -EINVAL; - u16 version; - struct rb_node *rbp; - - ceph_decode_16_safe(p, end, version, bad); - if (version > CEPH_OSDMAP_INC_VERSION) { - pr_warning("got unknown v %d > %d of inc osdmap\n", version, - CEPH_OSDMAP_INC_VERSION); - goto bad; - } - - ceph_decode_need(p, end, sizeof(fsid)+sizeof(modified)+2*sizeof(u32), - bad); - ceph_decode_copy(p, &fsid, sizeof(fsid)); - epoch = ceph_decode_32(p); - BUG_ON(epoch != map->epoch+1); - ceph_decode_copy(p, &modified, sizeof(modified)); - new_pool_max = ceph_decode_32(p); - new_flags = ceph_decode_32(p); - - /* full map? */ - ceph_decode_32_safe(p, end, len, bad); - if (len > 0) { - dout("apply_incremental full map len %d, %p to %p\n", - len, *p, end); - return osdmap_decode(p, min(*p+len, end)); - } - - /* new crush? */ - ceph_decode_32_safe(p, end, len, bad); - if (len > 0) { - dout("apply_incremental new crush map len %d, %p to %p\n", - len, *p, end); - newcrush = crush_decode(*p, min(*p+len, end)); - if (IS_ERR(newcrush)) - return ERR_CAST(newcrush); - *p += len; - } - - /* new flags? */ - if (new_flags >= 0) - map->flags = new_flags; - if (new_pool_max >= 0) - map->pool_max = new_pool_max; - - ceph_decode_need(p, end, 5*sizeof(u32), bad); - - /* new max? */ - max = ceph_decode_32(p); - if (max >= 0) { - err = osdmap_set_max_osd(map, max); - if (err < 0) - goto bad; - } - - map->epoch++; - map->modified = map->modified; - if (newcrush) { - if (map->crush) - crush_destroy(map->crush); - map->crush = newcrush; - newcrush = NULL; - } - - /* new_pool */ - ceph_decode_32_safe(p, end, len, bad); - while (len--) { - __u8 ev; - struct ceph_pg_pool_info *pi; - - ceph_decode_32_safe(p, end, pool, bad); - ceph_decode_need(p, end, 1 + sizeof(pi->v), bad); - ev = ceph_decode_8(p); /* encoding version */ - if (ev > CEPH_PG_POOL_VERSION) { - pr_warning("got unknown v %d > %d of ceph_pg_pool\n", - ev, CEPH_PG_POOL_VERSION); - goto bad; - } - pi = __lookup_pg_pool(&map->pg_pools, pool); - if (!pi) { - pi = kzalloc(sizeof(*pi), GFP_NOFS); - if (!pi) { - err = -ENOMEM; - goto bad; - } - pi->id = pool; - __insert_pg_pool(&map->pg_pools, pi); - } - err = __decode_pool(p, end, pi); - if (err < 0) - goto bad; - } - if (version >= 5 && __decode_pool_names(p, end, map) < 0) - goto bad; - - /* old_pool */ - ceph_decode_32_safe(p, end, len, bad); - while (len--) { - struct ceph_pg_pool_info *pi; - - ceph_decode_32_safe(p, end, pool, bad); - pi = __lookup_pg_pool(&map->pg_pools, pool); - if (pi) - __remove_pg_pool(&map->pg_pools, pi); - } - - /* new_up */ - err = -EINVAL; - ceph_decode_32_safe(p, end, len, bad); - while (len--) { - u32 osd; - struct ceph_entity_addr addr; - ceph_decode_32_safe(p, end, osd, bad); - ceph_decode_copy_safe(p, end, &addr, sizeof(addr), bad); - ceph_decode_addr(&addr); - pr_info("osd%d up\n", osd); - BUG_ON(osd >= map->max_osd); - map->osd_state[osd] |= CEPH_OSD_UP; - map->osd_addr[osd] = addr; - } - - /* new_down */ - ceph_decode_32_safe(p, end, len, bad); - while (len--) { - u32 osd; - ceph_decode_32_safe(p, end, osd, bad); - (*p)++; /* clean flag */ - pr_info("osd%d down\n", osd); - if (osd < map->max_osd) - map->osd_state[osd] &= ~CEPH_OSD_UP; - } - - /* new_weight */ - ceph_decode_32_safe(p, end, len, bad); - while (len--) { - u32 osd, off; - ceph_decode_need(p, end, sizeof(u32)*2, bad); - osd = ceph_decode_32(p); - off = ceph_decode_32(p); - pr_info("osd%d weight 0x%x %s\n", osd, off, - off == CEPH_OSD_IN ? "(in)" : - (off == CEPH_OSD_OUT ? "(out)" : "")); - if (osd < map->max_osd) - map->osd_weight[osd] = off; - } - - /* new_pg_temp */ - rbp = rb_first(&map->pg_temp); - ceph_decode_32_safe(p, end, len, bad); - while (len--) { - struct ceph_pg_mapping *pg; - int j; - struct ceph_pg pgid; - u32 pglen; - ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad); - ceph_decode_copy(p, &pgid, sizeof(pgid)); - pglen = ceph_decode_32(p); - - /* remove any? */ - while (rbp && pgid_cmp(rb_entry(rbp, struct ceph_pg_mapping, - node)->pgid, pgid) <= 0) { - struct ceph_pg_mapping *cur = - rb_entry(rbp, struct ceph_pg_mapping, node); - - rbp = rb_next(rbp); - dout(" removed pg_temp %llx\n", *(u64 *)&cur->pgid); - rb_erase(&cur->node, &map->pg_temp); - kfree(cur); - } - - if (pglen) { - /* insert */ - ceph_decode_need(p, end, pglen*sizeof(u32), bad); - pg = kmalloc(sizeof(*pg) + sizeof(u32)*pglen, GFP_NOFS); - if (!pg) { - err = -ENOMEM; - goto bad; - } - pg->pgid = pgid; - pg->len = pglen; - for (j = 0; j < pglen; j++) - pg->osds[j] = ceph_decode_32(p); - err = __insert_pg_mapping(pg, &map->pg_temp); - if (err) { - kfree(pg); - goto bad; - } - dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid, - pglen); - } - } - while (rbp) { - struct ceph_pg_mapping *cur = - rb_entry(rbp, struct ceph_pg_mapping, node); - - rbp = rb_next(rbp); - dout(" removed pg_temp %llx\n", *(u64 *)&cur->pgid); - rb_erase(&cur->node, &map->pg_temp); - kfree(cur); - } - - /* ignore the rest */ - *p = end; - return map; - -bad: - pr_err("corrupt inc osdmap epoch %d off %d (%p of %p-%p)\n", - epoch, (int)(*p - start), *p, start, end); - print_hex_dump(KERN_DEBUG, "osdmap: ", - DUMP_PREFIX_OFFSET, 16, 1, - start, end - start, true); - if (newcrush) - crush_destroy(newcrush); - return ERR_PTR(err); -} - - - - -/* - * calculate file layout from given offset, length. - * fill in correct oid, logical length, and object extent - * offset, length. - * - * for now, we write only a single su, until we can - * pass a stride back to the caller. - */ -void ceph_calc_file_object_mapping(struct ceph_file_layout *layout, - u64 off, u64 *plen, - u64 *ono, - u64 *oxoff, u64 *oxlen) -{ - u32 osize = le32_to_cpu(layout->fl_object_size); - u32 su = le32_to_cpu(layout->fl_stripe_unit); - u32 sc = le32_to_cpu(layout->fl_stripe_count); - u32 bl, stripeno, stripepos, objsetno; - u32 su_per_object; - u64 t, su_offset; - - dout("mapping %llu~%llu osize %u fl_su %u\n", off, *plen, - osize, su); - su_per_object = osize / su; - dout("osize %u / su %u = su_per_object %u\n", osize, su, - su_per_object); - - BUG_ON((su & ~PAGE_MASK) != 0); - /* bl = *off / su; */ - t = off; - do_div(t, su); - bl = t; - dout("off %llu / su %u = bl %u\n", off, su, bl); - - stripeno = bl / sc; - stripepos = bl % sc; - objsetno = stripeno / su_per_object; - - *ono = objsetno * sc + stripepos; - dout("objset %u * sc %u = ono %u\n", objsetno, sc, (unsigned)*ono); - - /* *oxoff = *off % layout->fl_stripe_unit; # offset in su */ - t = off; - su_offset = do_div(t, su); - *oxoff = su_offset + (stripeno % su_per_object) * su; - - /* - * Calculate the length of the extent being written to the selected - * object. This is the minimum of the full length requested (plen) or - * the remainder of the current stripe being written to. - */ - *oxlen = min_t(u64, *plen, su - su_offset); - *plen = *oxlen; - - dout(" obj extent %llu~%llu\n", *oxoff, *oxlen); -} - -/* - * calculate an object layout (i.e. pgid) from an oid, - * file_layout, and osdmap - */ -int ceph_calc_object_layout(struct ceph_object_layout *ol, - const char *oid, - struct ceph_file_layout *fl, - struct ceph_osdmap *osdmap) -{ - unsigned num, num_mask; - struct ceph_pg pgid; - s32 preferred = (s32)le32_to_cpu(fl->fl_pg_preferred); - int poolid = le32_to_cpu(fl->fl_pg_pool); - struct ceph_pg_pool_info *pool; - unsigned ps; - - BUG_ON(!osdmap); - - pool = __lookup_pg_pool(&osdmap->pg_pools, poolid); - if (!pool) - return -EIO; - ps = ceph_str_hash(pool->v.object_hash, oid, strlen(oid)); - if (preferred >= 0) { - ps += preferred; - num = le32_to_cpu(pool->v.lpg_num); - num_mask = pool->lpg_num_mask; - } else { - num = le32_to_cpu(pool->v.pg_num); - num_mask = pool->pg_num_mask; - } - - pgid.ps = cpu_to_le16(ps); - pgid.preferred = cpu_to_le16(preferred); - pgid.pool = fl->fl_pg_pool; - if (preferred >= 0) - dout("calc_object_layout '%s' pgid %d.%xp%d\n", oid, poolid, ps, - (int)preferred); - else - dout("calc_object_layout '%s' pgid %d.%x\n", oid, poolid, ps); - - ol->ol_pgid = pgid; - ol->ol_stripe_unit = fl->fl_object_stripe_unit; - return 0; -} - -/* - * Calculate raw osd vector for the given pgid. Return pointer to osd - * array, or NULL on failure. - */ -static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid, - int *osds, int *num) -{ - struct ceph_pg_mapping *pg; - struct ceph_pg_pool_info *pool; - int ruleno; - unsigned poolid, ps, pps; - int preferred; - - /* pg_temp? */ - pg = __lookup_pg_mapping(&osdmap->pg_temp, pgid); - if (pg) { - *num = pg->len; - return pg->osds; - } - - /* crush */ - poolid = le32_to_cpu(pgid.pool); - ps = le16_to_cpu(pgid.ps); - preferred = (s16)le16_to_cpu(pgid.preferred); - - /* don't forcefeed bad device ids to crush */ - if (preferred >= osdmap->max_osd || - preferred >= osdmap->crush->max_devices) - preferred = -1; - - pool = __lookup_pg_pool(&osdmap->pg_pools, poolid); - if (!pool) - return NULL; - ruleno = crush_find_rule(osdmap->crush, pool->v.crush_ruleset, - pool->v.type, pool->v.size); - if (ruleno < 0) { - pr_err("no crush rule pool %d ruleset %d type %d size %d\n", - poolid, pool->v.crush_ruleset, pool->v.type, - pool->v.size); - return NULL; - } - - if (preferred >= 0) - pps = ceph_stable_mod(ps, - le32_to_cpu(pool->v.lpgp_num), - pool->lpgp_num_mask); - else - pps = ceph_stable_mod(ps, - le32_to_cpu(pool->v.pgp_num), - pool->pgp_num_mask); - pps += poolid; - *num = crush_do_rule(osdmap->crush, ruleno, pps, osds, - min_t(int, pool->v.size, *num), - preferred, osdmap->osd_weight); - return osds; -} - -/* - * Return acting set for given pgid. - */ -int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid, - int *acting) -{ - int rawosds[CEPH_PG_MAX_SIZE], *osds; - int i, o, num = CEPH_PG_MAX_SIZE; - - osds = calc_pg_raw(osdmap, pgid, rawosds, &num); - if (!osds) - return -1; - - /* primary is first up osd */ - o = 0; - for (i = 0; i < num; i++) - if (ceph_osd_is_up(osdmap, osds[i])) - acting[o++] = osds[i]; - return o; -} - -/* - * Return primary osd for given pgid, or -1 if none. - */ -int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg pgid) -{ - int rawosds[CEPH_PG_MAX_SIZE], *osds; - int i, num = CEPH_PG_MAX_SIZE; - - osds = calc_pg_raw(osdmap, pgid, rawosds, &num); - if (!osds) - return -1; - - /* primary is first up osd */ - for (i = 0; i < num; i++) - if (ceph_osd_is_up(osdmap, osds[i])) - return osds[i]; - return -1; -} diff --git a/fs/ceph/osdmap.h b/fs/ceph/osdmap.h deleted file mode 100644 index a592b21..0000000 --- a/fs/ceph/osdmap.h +++ /dev/null @@ -1,130 +0,0 @@ -#ifndef _FS_CEPH_OSDMAP_H -#define _FS_CEPH_OSDMAP_H - -#include -#include "types.h" -#include "ceph_fs.h" -#include "crush/crush.h" - -/* - * The osd map describes the current membership of the osd cluster and - * specifies the mapping of objects to placement groups and placement - * groups to (sets of) osds. That is, it completely specifies the - * (desired) distribution of all data objects in the system at some - * point in time. - * - * Each map version is identified by an epoch, which increases monotonically. - * - * The map can be updated either via an incremental map (diff) describing - * the change between two successive epochs, or as a fully encoded map. - */ -struct ceph_pg_pool_info { - struct rb_node node; - int id; - struct ceph_pg_pool v; - int pg_num_mask, pgp_num_mask, lpg_num_mask, lpgp_num_mask; - char *name; -}; - -struct ceph_pg_mapping { - struct rb_node node; - struct ceph_pg pgid; - int len; - int osds[]; -}; - -struct ceph_osdmap { - struct ceph_fsid fsid; - u32 epoch; - u32 mkfs_epoch; - struct ceph_timespec created, modified; - - u32 flags; /* CEPH_OSDMAP_* */ - - u32 max_osd; /* size of osd_state, _offload, _addr arrays */ - u8 *osd_state; /* CEPH_OSD_* */ - u32 *osd_weight; /* 0 = failed, 0x10000 = 100% normal */ - struct ceph_entity_addr *osd_addr; - - struct rb_root pg_temp; - struct rb_root pg_pools; - u32 pool_max; - - /* the CRUSH map specifies the mapping of placement groups to - * the list of osds that store+replicate them. */ - struct crush_map *crush; -}; - -/* - * file layout helpers - */ -#define ceph_file_layout_su(l) ((__s32)le32_to_cpu((l).fl_stripe_unit)) -#define ceph_file_layout_stripe_count(l) \ - ((__s32)le32_to_cpu((l).fl_stripe_count)) -#define ceph_file_layout_object_size(l) ((__s32)le32_to_cpu((l).fl_object_size)) -#define ceph_file_layout_cas_hash(l) ((__s32)le32_to_cpu((l).fl_cas_hash)) -#define ceph_file_layout_object_su(l) \ - ((__s32)le32_to_cpu((l).fl_object_stripe_unit)) -#define ceph_file_layout_pg_preferred(l) \ - ((__s32)le32_to_cpu((l).fl_pg_preferred)) -#define ceph_file_layout_pg_pool(l) \ - ((__s32)le32_to_cpu((l).fl_pg_pool)) - -static inline unsigned ceph_file_layout_stripe_width(struct ceph_file_layout *l) -{ - return le32_to_cpu(l->fl_stripe_unit) * - le32_to_cpu(l->fl_stripe_count); -} - -/* "period" == bytes before i start on a new set of objects */ -static inline unsigned ceph_file_layout_period(struct ceph_file_layout *l) -{ - return le32_to_cpu(l->fl_object_size) * - le32_to_cpu(l->fl_stripe_count); -} - - -static inline int ceph_osd_is_up(struct ceph_osdmap *map, int osd) -{ - return (osd < map->max_osd) && (map->osd_state[osd] & CEPH_OSD_UP); -} - -static inline bool ceph_osdmap_flag(struct ceph_osdmap *map, int flag) -{ - return map && (map->flags & flag); -} - -extern char *ceph_osdmap_state_str(char *str, int len, int state); - -static inline struct ceph_entity_addr *ceph_osd_addr(struct ceph_osdmap *map, - int osd) -{ - if (osd >= map->max_osd) - return NULL; - return &map->osd_addr[osd]; -} - -extern struct ceph_osdmap *osdmap_decode(void **p, void *end); -extern struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end, - struct ceph_osdmap *map, - struct ceph_messenger *msgr); -extern void ceph_osdmap_destroy(struct ceph_osdmap *map); - -/* calculate mapping of a file extent to an object */ -extern void ceph_calc_file_object_mapping(struct ceph_file_layout *layout, - u64 off, u64 *plen, - u64 *bno, u64 *oxoff, u64 *oxlen); - -/* calculate mapping of object to a placement group */ -extern int ceph_calc_object_layout(struct ceph_object_layout *ol, - const char *oid, - struct ceph_file_layout *fl, - struct ceph_osdmap *osdmap); -extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid, - int *acting); -extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, - struct ceph_pg pgid); - -extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name); - -#endif diff --git a/fs/ceph/pagelist.c b/fs/ceph/pagelist.c deleted file mode 100644 index 326e1c0..0000000 --- a/fs/ceph/pagelist.c +++ /dev/null @@ -1,63 +0,0 @@ - -#include -#include -#include - -#include "pagelist.h" - -static void ceph_pagelist_unmap_tail(struct ceph_pagelist *pl) -{ - struct page *page = list_entry(pl->head.prev, struct page, - lru); - kunmap(page); -} - -int ceph_pagelist_release(struct ceph_pagelist *pl) -{ - if (pl->mapped_tail) - ceph_pagelist_unmap_tail(pl); - - while (!list_empty(&pl->head)) { - struct page *page = list_first_entry(&pl->head, struct page, - lru); - list_del(&page->lru); - __free_page(page); - } - return 0; -} - -static int ceph_pagelist_addpage(struct ceph_pagelist *pl) -{ - struct page *page = __page_cache_alloc(GFP_NOFS); - if (!page) - return -ENOMEM; - pl->room += PAGE_SIZE; - list_add_tail(&page->lru, &pl->head); - if (pl->mapped_tail) - ceph_pagelist_unmap_tail(pl); - pl->mapped_tail = kmap(page); - return 0; -} - -int ceph_pagelist_append(struct ceph_pagelist *pl, const void *buf, size_t len) -{ - while (pl->room < len) { - size_t bit = pl->room; - int ret; - - memcpy(pl->mapped_tail + (pl->length & ~PAGE_CACHE_MASK), - buf, bit); - pl->length += bit; - pl->room -= bit; - buf += bit; - len -= bit; - ret = ceph_pagelist_addpage(pl); - if (ret) - return ret; - } - - memcpy(pl->mapped_tail + (pl->length & ~PAGE_CACHE_MASK), buf, len); - pl->length += len; - pl->room -= len; - return 0; -} diff --git a/fs/ceph/pagelist.h b/fs/ceph/pagelist.h deleted file mode 100644 index cc9327a..0000000 --- a/fs/ceph/pagelist.h +++ /dev/null @@ -1,54 +0,0 @@ -#ifndef __FS_CEPH_PAGELIST_H -#define __FS_CEPH_PAGELIST_H - -#include - -struct ceph_pagelist { - struct list_head head; - void *mapped_tail; - size_t length; - size_t room; -}; - -static inline void ceph_pagelist_init(struct ceph_pagelist *pl) -{ - INIT_LIST_HEAD(&pl->head); - pl->mapped_tail = NULL; - pl->length = 0; - pl->room = 0; -} -extern int ceph_pagelist_release(struct ceph_pagelist *pl); - -extern int ceph_pagelist_append(struct ceph_pagelist *pl, const void *d, size_t l); - -static inline int ceph_pagelist_encode_64(struct ceph_pagelist *pl, u64 v) -{ - __le64 ev = cpu_to_le64(v); - return ceph_pagelist_append(pl, &ev, sizeof(ev)); -} -static inline int ceph_pagelist_encode_32(struct ceph_pagelist *pl, u32 v) -{ - __le32 ev = cpu_to_le32(v); - return ceph_pagelist_append(pl, &ev, sizeof(ev)); -} -static inline int ceph_pagelist_encode_16(struct ceph_pagelist *pl, u16 v) -{ - __le16 ev = cpu_to_le16(v); - return ceph_pagelist_append(pl, &ev, sizeof(ev)); -} -static inline int ceph_pagelist_encode_8(struct ceph_pagelist *pl, u8 v) -{ - return ceph_pagelist_append(pl, &v, 1); -} -static inline int ceph_pagelist_encode_string(struct ceph_pagelist *pl, - char *s, size_t len) -{ - int ret = ceph_pagelist_encode_32(pl, len); - if (ret) - return ret; - if (len) - return ceph_pagelist_append(pl, s, len); - return 0; -} - -#endif diff --git a/fs/ceph/rados.h b/fs/ceph/rados.h deleted file mode 100644 index 6d5247f..0000000 --- a/fs/ceph/rados.h +++ /dev/null @@ -1,405 +0,0 @@ -#ifndef CEPH_RADOS_H -#define CEPH_RADOS_H - -/* - * Data types for the Ceph distributed object storage layer RADOS - * (Reliable Autonomic Distributed Object Store). - */ - -#include "msgr.h" - -/* - * osdmap encoding versions - */ -#define CEPH_OSDMAP_INC_VERSION 5 -#define CEPH_OSDMAP_INC_VERSION_EXT 5 -#define CEPH_OSDMAP_VERSION 5 -#define CEPH_OSDMAP_VERSION_EXT 5 - -/* - * fs id - */ -struct ceph_fsid { - unsigned char fsid[16]; -}; - -static inline int ceph_fsid_compare(const struct ceph_fsid *a, - const struct ceph_fsid *b) -{ - return memcmp(a, b, sizeof(*a)); -} - -/* - * ino, object, etc. - */ -typedef __le64 ceph_snapid_t; -#define CEPH_SNAPDIR ((__u64)(-1)) /* reserved for hidden .snap dir */ -#define CEPH_NOSNAP ((__u64)(-2)) /* "head", "live" revision */ -#define CEPH_MAXSNAP ((__u64)(-3)) /* largest valid snapid */ - -struct ceph_timespec { - __le32 tv_sec; - __le32 tv_nsec; -} __attribute__ ((packed)); - - -/* - * object layout - how objects are mapped into PGs - */ -#define CEPH_OBJECT_LAYOUT_HASH 1 -#define CEPH_OBJECT_LAYOUT_LINEAR 2 -#define CEPH_OBJECT_LAYOUT_HASHINO 3 - -/* - * pg layout -- how PGs are mapped onto (sets of) OSDs - */ -#define CEPH_PG_LAYOUT_CRUSH 0 -#define CEPH_PG_LAYOUT_HASH 1 -#define CEPH_PG_LAYOUT_LINEAR 2 -#define CEPH_PG_LAYOUT_HYBRID 3 - -#define CEPH_PG_MAX_SIZE 16 /* max # osds in a single pg */ - -/* - * placement group. - * we encode this into one __le64. - */ -struct ceph_pg { - __le16 preferred; /* preferred primary osd */ - __le16 ps; /* placement seed */ - __le32 pool; /* object pool */ -} __attribute__ ((packed)); - -/* - * pg_pool is a set of pgs storing a pool of objects - * - * pg_num -- base number of pseudorandomly placed pgs - * - * pgp_num -- effective number when calculating pg placement. this - * is used for pg_num increases. new pgs result in data being "split" - * into new pgs. for this to proceed smoothly, new pgs are intiially - * colocated with their parents; that is, pgp_num doesn't increase - * until the new pgs have successfully split. only _then_ are the new - * pgs placed independently. - * - * lpg_num -- localized pg count (per device). replicas are randomly - * selected. - * - * lpgp_num -- as above. - */ -#define CEPH_PG_TYPE_REP 1 -#define CEPH_PG_TYPE_RAID4 2 -#define CEPH_PG_POOL_VERSION 2 -struct ceph_pg_pool { - __u8 type; /* CEPH_PG_TYPE_* */ - __u8 size; /* number of osds in each pg */ - __u8 crush_ruleset; /* crush placement rule */ - __u8 object_hash; /* hash mapping object name to ps */ - __le32 pg_num, pgp_num; /* number of pg's */ - __le32 lpg_num, lpgp_num; /* number of localized pg's */ - __le32 last_change; /* most recent epoch changed */ - __le64 snap_seq; /* seq for per-pool snapshot */ - __le32 snap_epoch; /* epoch of last snap */ - __le32 num_snaps; - __le32 num_removed_snap_intervals; /* if non-empty, NO per-pool snaps */ - __le64 auid; /* who owns the pg */ -} __attribute__ ((packed)); - -/* - * stable_mod func is used to control number of placement groups. - * similar to straight-up modulo, but produces a stable mapping as b - * increases over time. b is the number of bins, and bmask is the - * containing power of 2 minus 1. - * - * b <= bmask and bmask=(2**n)-1 - * e.g., b=12 -> bmask=15, b=123 -> bmask=127 - */ -static inline int ceph_stable_mod(int x, int b, int bmask) -{ - if ((x & bmask) < b) - return x & bmask; - else - return x & (bmask >> 1); -} - -/* - * object layout - how a given object should be stored. - */ -struct ceph_object_layout { - struct ceph_pg ol_pgid; /* raw pg, with _full_ ps precision. */ - __le32 ol_stripe_unit; /* for per-object parity, if any */ -} __attribute__ ((packed)); - -/* - * compound epoch+version, used by storage layer to serialize mutations - */ -struct ceph_eversion { - __le32 epoch; - __le64 version; -} __attribute__ ((packed)); - -/* - * osd map bits - */ - -/* status bits */ -#define CEPH_OSD_EXISTS 1 -#define CEPH_OSD_UP 2 - -/* osd weights. fixed point value: 0x10000 == 1.0 ("in"), 0 == "out" */ -#define CEPH_OSD_IN 0x10000 -#define CEPH_OSD_OUT 0 - - -/* - * osd map flag bits - */ -#define CEPH_OSDMAP_NEARFULL (1<<0) /* sync writes (near ENOSPC) */ -#define CEPH_OSDMAP_FULL (1<<1) /* no data writes (ENOSPC) */ -#define CEPH_OSDMAP_PAUSERD (1<<2) /* pause all reads */ -#define CEPH_OSDMAP_PAUSEWR (1<<3) /* pause all writes */ -#define CEPH_OSDMAP_PAUSEREC (1<<4) /* pause recovery */ - -/* - * osd ops - */ -#define CEPH_OSD_OP_MODE 0xf000 -#define CEPH_OSD_OP_MODE_RD 0x1000 -#define CEPH_OSD_OP_MODE_WR 0x2000 -#define CEPH_OSD_OP_MODE_RMW 0x3000 -#define CEPH_OSD_OP_MODE_SUB 0x4000 - -#define CEPH_OSD_OP_TYPE 0x0f00 -#define CEPH_OSD_OP_TYPE_LOCK 0x0100 -#define CEPH_OSD_OP_TYPE_DATA 0x0200 -#define CEPH_OSD_OP_TYPE_ATTR 0x0300 -#define CEPH_OSD_OP_TYPE_EXEC 0x0400 -#define CEPH_OSD_OP_TYPE_PG 0x0500 - -enum { - /** data **/ - /* read */ - CEPH_OSD_OP_READ = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 1, - CEPH_OSD_OP_STAT = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 2, - - /* fancy read */ - CEPH_OSD_OP_MASKTRUNC = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 4, - - /* write */ - CEPH_OSD_OP_WRITE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 1, - CEPH_OSD_OP_WRITEFULL = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 2, - CEPH_OSD_OP_TRUNCATE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 3, - CEPH_OSD_OP_ZERO = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 4, - CEPH_OSD_OP_DELETE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 5, - - /* fancy write */ - CEPH_OSD_OP_APPEND = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 6, - CEPH_OSD_OP_STARTSYNC = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 7, - CEPH_OSD_OP_SETTRUNC = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 8, - CEPH_OSD_OP_TRIMTRUNC = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 9, - - CEPH_OSD_OP_TMAPUP = CEPH_OSD_OP_MODE_RMW | CEPH_OSD_OP_TYPE_DATA | 10, - CEPH_OSD_OP_TMAPPUT = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 11, - CEPH_OSD_OP_TMAPGET = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 12, - - CEPH_OSD_OP_CREATE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 13, - CEPH_OSD_OP_ROLLBACK= CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 14, - - /** attrs **/ - /* read */ - CEPH_OSD_OP_GETXATTR = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 1, - CEPH_OSD_OP_GETXATTRS = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 2, - CEPH_OSD_OP_CMPXATTR = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 3, - - /* write */ - CEPH_OSD_OP_SETXATTR = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 1, - CEPH_OSD_OP_SETXATTRS = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 2, - CEPH_OSD_OP_RESETXATTRS = CEPH_OSD_OP_MODE_WR|CEPH_OSD_OP_TYPE_ATTR | 3, - CEPH_OSD_OP_RMXATTR = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 4, - - /** subop **/ - CEPH_OSD_OP_PULL = CEPH_OSD_OP_MODE_SUB | 1, - CEPH_OSD_OP_PUSH = CEPH_OSD_OP_MODE_SUB | 2, - CEPH_OSD_OP_BALANCEREADS = CEPH_OSD_OP_MODE_SUB | 3, - CEPH_OSD_OP_UNBALANCEREADS = CEPH_OSD_OP_MODE_SUB | 4, - CEPH_OSD_OP_SCRUB = CEPH_OSD_OP_MODE_SUB | 5, - - /** lock **/ - CEPH_OSD_OP_WRLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 1, - CEPH_OSD_OP_WRUNLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 2, - CEPH_OSD_OP_RDLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 3, - CEPH_OSD_OP_RDUNLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 4, - CEPH_OSD_OP_UPLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 5, - CEPH_OSD_OP_DNLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 6, - - /** exec **/ - CEPH_OSD_OP_CALL = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_EXEC | 1, - - /** pg **/ - CEPH_OSD_OP_PGLS = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_PG | 1, -}; - -static inline int ceph_osd_op_type_lock(int op) -{ - return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_LOCK; -} -static inline int ceph_osd_op_type_data(int op) -{ - return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_DATA; -} -static inline int ceph_osd_op_type_attr(int op) -{ - return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_ATTR; -} -static inline int ceph_osd_op_type_exec(int op) -{ - return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_EXEC; -} -static inline int ceph_osd_op_type_pg(int op) -{ - return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_PG; -} - -static inline int ceph_osd_op_mode_subop(int op) -{ - return (op & CEPH_OSD_OP_MODE) == CEPH_OSD_OP_MODE_SUB; -} -static inline int ceph_osd_op_mode_read(int op) -{ - return (op & CEPH_OSD_OP_MODE) == CEPH_OSD_OP_MODE_RD; -} -static inline int ceph_osd_op_mode_modify(int op) -{ - return (op & CEPH_OSD_OP_MODE) == CEPH_OSD_OP_MODE_WR; -} - -/* - * note that the following tmap stuff is also defined in the ceph librados.h - * any modification here needs to be updated there - */ -#define CEPH_OSD_TMAP_HDR 'h' -#define CEPH_OSD_TMAP_SET 's' -#define CEPH_OSD_TMAP_RM 'r' - -extern const char *ceph_osd_op_name(int op); - - -/* - * osd op flags - * - * An op may be READ, WRITE, or READ|WRITE. - */ -enum { - CEPH_OSD_FLAG_ACK = 1, /* want (or is) "ack" ack */ - CEPH_OSD_FLAG_ONNVRAM = 2, /* want (or is) "onnvram" ack */ - CEPH_OSD_FLAG_ONDISK = 4, /* want (or is) "ondisk" ack */ - CEPH_OSD_FLAG_RETRY = 8, /* resend attempt */ - CEPH_OSD_FLAG_READ = 16, /* op may read */ - CEPH_OSD_FLAG_WRITE = 32, /* op may write */ - CEPH_OSD_FLAG_ORDERSNAP = 64, /* EOLDSNAP if snapc is out of order */ - CEPH_OSD_FLAG_PEERSTAT = 128, /* msg includes osd_peer_stat */ - CEPH_OSD_FLAG_BALANCE_READS = 256, - CEPH_OSD_FLAG_PARALLELEXEC = 512, /* execute op in parallel */ - CEPH_OSD_FLAG_PGOP = 1024, /* pg op, no object */ - CEPH_OSD_FLAG_EXEC = 2048, /* op may exec */ - CEPH_OSD_FLAG_EXEC_PUBLIC = 4096, /* op may exec (public) */ -}; - -enum { - CEPH_OSD_OP_FLAG_EXCL = 1, /* EXCL object create */ -}; - -#define EOLDSNAPC ERESTART /* ORDERSNAP flag set; writer has old snapc*/ -#define EBLACKLISTED ESHUTDOWN /* blacklisted */ - -/* xattr comparison */ -enum { - CEPH_OSD_CMPXATTR_OP_NOP = 0, - CEPH_OSD_CMPXATTR_OP_EQ = 1, - CEPH_OSD_CMPXATTR_OP_NE = 2, - CEPH_OSD_CMPXATTR_OP_GT = 3, - CEPH_OSD_CMPXATTR_OP_GTE = 4, - CEPH_OSD_CMPXATTR_OP_LT = 5, - CEPH_OSD_CMPXATTR_OP_LTE = 6 -}; - -enum { - CEPH_OSD_CMPXATTR_MODE_STRING = 1, - CEPH_OSD_CMPXATTR_MODE_U64 = 2 -}; - -/* - * an individual object operation. each may be accompanied by some data - * payload - */ -struct ceph_osd_op { - __le16 op; /* CEPH_OSD_OP_* */ - __le32 flags; /* CEPH_OSD_FLAG_* */ - union { - struct { - __le64 offset, length; - __le64 truncate_size; - __le32 truncate_seq; - } __attribute__ ((packed)) extent; - struct { - __le32 name_len; - __le32 value_len; - __u8 cmp_op; /* CEPH_OSD_CMPXATTR_OP_* */ - __u8 cmp_mode; /* CEPH_OSD_CMPXATTR_MODE_* */ - } __attribute__ ((packed)) xattr; - struct { - __u8 class_len; - __u8 method_len; - __u8 argc; - __le32 indata_len; - } __attribute__ ((packed)) cls; - struct { - __le64 cookie, count; - } __attribute__ ((packed)) pgls; - struct { - __le64 snapid; - } __attribute__ ((packed)) snap; - }; - __le32 payload_len; -} __attribute__ ((packed)); - -/* - * osd request message header. each request may include multiple - * ceph_osd_op object operations. - */ -struct ceph_osd_request_head { - __le32 client_inc; /* client incarnation */ - struct ceph_object_layout layout; /* pgid */ - __le32 osdmap_epoch; /* client's osdmap epoch */ - - __le32 flags; - - struct ceph_timespec mtime; /* for mutations only */ - struct ceph_eversion reassert_version; /* if we are replaying op */ - - __le32 object_len; /* length of object name */ - - __le64 snapid; /* snapid to read */ - __le64 snap_seq; /* writer's snap context */ - __le32 num_snaps; - - __le16 num_ops; - struct ceph_osd_op ops[]; /* followed by ops[], obj, ticket, snaps */ -} __attribute__ ((packed)); - -struct ceph_osd_reply_head { - __le32 client_inc; /* client incarnation */ - __le32 flags; - struct ceph_object_layout layout; - __le32 osdmap_epoch; - struct ceph_eversion reassert_version; /* for replaying uncommitted */ - - __le32 result; /* result code */ - - __le32 object_len; /* length of object name */ - __le32 num_ops; - struct ceph_osd_op ops[0]; /* ops[], object */ -} __attribute__ ((packed)); - - -#endif diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index 190b6c4..39c243a 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -1,10 +1,12 @@ -#include "ceph_debug.h" +#include #include #include #include "super.h" -#include "decode.h" +#include "mds_client.h" + +#include /* * Snapshots in ceph are driven in large part by cooperation from the @@ -526,7 +528,7 @@ int __ceph_finish_cap_snap(struct ceph_inode_info *ci, struct ceph_cap_snap *capsnap) { struct inode *inode = &ci->vfs_inode; - struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; BUG_ON(capsnap->writing); capsnap->size = inode->i_size; @@ -747,7 +749,7 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, struct ceph_msg *msg) { - struct super_block *sb = mdsc->client->sb; + struct super_block *sb = mdsc->fsc->sb; int mds = session->s_mds; u64 split; int op; diff --git a/fs/ceph/strings.c b/fs/ceph/strings.c new file mode 100644 index 0000000..cd5097d --- /dev/null +++ b/fs/ceph/strings.c @@ -0,0 +1,117 @@ +/* + * Ceph fs string constants + */ +#include +#include + + +const char *ceph_mds_state_name(int s) +{ + switch (s) { + /* down and out */ + case CEPH_MDS_STATE_DNE: return "down:dne"; + case CEPH_MDS_STATE_STOPPED: return "down:stopped"; + /* up and out */ + case CEPH_MDS_STATE_BOOT: return "up:boot"; + case CEPH_MDS_STATE_STANDBY: return "up:standby"; + case CEPH_MDS_STATE_STANDBY_REPLAY: return "up:standby-replay"; + case CEPH_MDS_STATE_CREATING: return "up:creating"; + case CEPH_MDS_STATE_STARTING: return "up:starting"; + /* up and in */ + case CEPH_MDS_STATE_REPLAY: return "up:replay"; + case CEPH_MDS_STATE_RESOLVE: return "up:resolve"; + case CEPH_MDS_STATE_RECONNECT: return "up:reconnect"; + case CEPH_MDS_STATE_REJOIN: return "up:rejoin"; + case CEPH_MDS_STATE_CLIENTREPLAY: return "up:clientreplay"; + case CEPH_MDS_STATE_ACTIVE: return "up:active"; + case CEPH_MDS_STATE_STOPPING: return "up:stopping"; + } + return "???"; +} + +const char *ceph_session_op_name(int op) +{ + switch (op) { + case CEPH_SESSION_REQUEST_OPEN: return "request_open"; + case CEPH_SESSION_OPEN: return "open"; + case CEPH_SESSION_REQUEST_CLOSE: return "request_close"; + case CEPH_SESSION_CLOSE: return "close"; + case CEPH_SESSION_REQUEST_RENEWCAPS: return "request_renewcaps"; + case CEPH_SESSION_RENEWCAPS: return "renewcaps"; + case CEPH_SESSION_STALE: return "stale"; + case CEPH_SESSION_RECALL_STATE: return "recall_state"; + } + return "???"; +} + +const char *ceph_mds_op_name(int op) +{ + switch (op) { + case CEPH_MDS_OP_LOOKUP: return "lookup"; + case CEPH_MDS_OP_LOOKUPHASH: return "lookuphash"; + case CEPH_MDS_OP_LOOKUPPARENT: return "lookupparent"; + case CEPH_MDS_OP_GETATTR: return "getattr"; + case CEPH_MDS_OP_SETXATTR: return "setxattr"; + case CEPH_MDS_OP_SETATTR: return "setattr"; + case CEPH_MDS_OP_RMXATTR: return "rmxattr"; + case CEPH_MDS_OP_READDIR: return "readdir"; + case CEPH_MDS_OP_MKNOD: return "mknod"; + case CEPH_MDS_OP_LINK: return "link"; + case CEPH_MDS_OP_UNLINK: return "unlink"; + case CEPH_MDS_OP_RENAME: return "rename"; + case CEPH_MDS_OP_MKDIR: return "mkdir"; + case CEPH_MDS_OP_RMDIR: return "rmdir"; + case CEPH_MDS_OP_SYMLINK: return "symlink"; + case CEPH_MDS_OP_CREATE: return "create"; + case CEPH_MDS_OP_OPEN: return "open"; + case CEPH_MDS_OP_LOOKUPSNAP: return "lookupsnap"; + case CEPH_MDS_OP_LSSNAP: return "lssnap"; + case CEPH_MDS_OP_MKSNAP: return "mksnap"; + case CEPH_MDS_OP_RMSNAP: return "rmsnap"; + case CEPH_MDS_OP_SETFILELOCK: return "setfilelock"; + case CEPH_MDS_OP_GETFILELOCK: return "getfilelock"; + } + return "???"; +} + +const char *ceph_cap_op_name(int op) +{ + switch (op) { + case CEPH_CAP_OP_GRANT: return "grant"; + case CEPH_CAP_OP_REVOKE: return "revoke"; + case CEPH_CAP_OP_TRUNC: return "trunc"; + case CEPH_CAP_OP_EXPORT: return "export"; + case CEPH_CAP_OP_IMPORT: return "import"; + case CEPH_CAP_OP_UPDATE: return "update"; + case CEPH_CAP_OP_DROP: return "drop"; + case CEPH_CAP_OP_FLUSH: return "flush"; + case CEPH_CAP_OP_FLUSH_ACK: return "flush_ack"; + case CEPH_CAP_OP_FLUSHSNAP: return "flushsnap"; + case CEPH_CAP_OP_FLUSHSNAP_ACK: return "flushsnap_ack"; + case CEPH_CAP_OP_RELEASE: return "release"; + case CEPH_CAP_OP_RENEW: return "renew"; + } + return "???"; +} + +const char *ceph_lease_op_name(int o) +{ + switch (o) { + case CEPH_MDS_LEASE_REVOKE: return "revoke"; + case CEPH_MDS_LEASE_RELEASE: return "release"; + case CEPH_MDS_LEASE_RENEW: return "renew"; + case CEPH_MDS_LEASE_REVOKE_ACK: return "revoke_ack"; + } + return "???"; +} + +const char *ceph_snap_op_name(int o) +{ + switch (o) { + case CEPH_SNAP_OP_UPDATE: return "update"; + case CEPH_SNAP_OP_CREATE: return "create"; + case CEPH_SNAP_OP_DESTROY: return "destroy"; + case CEPH_SNAP_OP_SPLIT: return "split"; + } + return "???"; +} diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 9922628..d6e0e04 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -1,5 +1,5 @@ -#include "ceph_debug.h" +#include #include #include @@ -15,10 +15,13 @@ #include #include -#include "decode.h" #include "super.h" -#include "mon_client.h" -#include "auth.h" +#include "mds_client.h" + +#include +#include +#include +#include /* * Ceph superblock operations @@ -26,36 +29,22 @@ * Handle the basics of mounting, unmounting. */ - -/* - * find filename portion of a path (/foo/bar/baz -> baz) - */ -const char *ceph_file_part(const char *s, int len) -{ - const char *e = s + len; - - while (e != s && *(e-1) != '/') - e--; - return e; -} - - /* * super ops */ static void ceph_put_super(struct super_block *s) { - struct ceph_client *client = ceph_sb_to_client(s); + struct ceph_fs_client *fsc = ceph_sb_to_client(s); dout("put_super\n"); - ceph_mdsc_close_sessions(&client->mdsc); + ceph_mdsc_close_sessions(fsc->mdsc); /* * ensure we release the bdi before put_anon_super releases * the device name. */ - if (s->s_bdi == &client->backing_dev_info) { - bdi_unregister(&client->backing_dev_info); + if (s->s_bdi == &fsc->backing_dev_info) { + bdi_unregister(&fsc->backing_dev_info); s->s_bdi = NULL; } @@ -64,14 +53,14 @@ static void ceph_put_super(struct super_block *s) static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf) { - struct ceph_client *client = ceph_inode_to_client(dentry->d_inode); - struct ceph_monmap *monmap = client->monc.monmap; + struct ceph_fs_client *fsc = ceph_inode_to_client(dentry->d_inode); + struct ceph_monmap *monmap = fsc->client->monc.monmap; struct ceph_statfs st; u64 fsid; int err; dout("statfs\n"); - err = ceph_monc_do_statfs(&client->monc, &st); + err = ceph_monc_do_statfs(&fsc->client->monc, &st); if (err < 0) return err; @@ -104,238 +93,28 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf) static int ceph_sync_fs(struct super_block *sb, int wait) { - struct ceph_client *client = ceph_sb_to_client(sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(sb); if (!wait) { dout("sync_fs (non-blocking)\n"); - ceph_flush_dirty_caps(&client->mdsc); + ceph_flush_dirty_caps(fsc->mdsc); dout("sync_fs (non-blocking) done\n"); return 0; } dout("sync_fs (blocking)\n"); - ceph_osdc_sync(&ceph_sb_to_client(sb)->osdc); - ceph_mdsc_sync(&ceph_sb_to_client(sb)->mdsc); + ceph_osdc_sync(&fsc->client->osdc); + ceph_mdsc_sync(fsc->mdsc); dout("sync_fs (blocking) done\n"); return 0; } -static int default_congestion_kb(void) -{ - int congestion_kb; - - /* - * Copied from NFS - * - * congestion size, scale with available memory. - * - * 64MB: 8192k - * 128MB: 11585k - * 256MB: 16384k - * 512MB: 23170k - * 1GB: 32768k - * 2GB: 46340k - * 4GB: 65536k - * 8GB: 92681k - * 16GB: 131072k - * - * This allows larger machines to have larger/more transfers. - * Limit the default to 256M - */ - congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10); - if (congestion_kb > 256*1024) - congestion_kb = 256*1024; - - return congestion_kb; -} - -/** - * ceph_show_options - Show mount options in /proc/mounts - * @m: seq_file to write to - * @mnt: mount descriptor - */ -static int ceph_show_options(struct seq_file *m, struct vfsmount *mnt) -{ - struct ceph_client *client = ceph_sb_to_client(mnt->mnt_sb); - struct ceph_mount_args *args = client->mount_args; - - if (args->flags & CEPH_OPT_FSID) - seq_printf(m, ",fsid=%pU", &args->fsid); - if (args->flags & CEPH_OPT_NOSHARE) - seq_puts(m, ",noshare"); - if (args->flags & CEPH_OPT_DIRSTAT) - seq_puts(m, ",dirstat"); - if ((args->flags & CEPH_OPT_RBYTES) == 0) - seq_puts(m, ",norbytes"); - if (args->flags & CEPH_OPT_NOCRC) - seq_puts(m, ",nocrc"); - if (args->flags & CEPH_OPT_NOASYNCREADDIR) - seq_puts(m, ",noasyncreaddir"); - - if (args->mount_timeout != CEPH_MOUNT_TIMEOUT_DEFAULT) - seq_printf(m, ",mount_timeout=%d", args->mount_timeout); - if (args->osd_idle_ttl != CEPH_OSD_IDLE_TTL_DEFAULT) - seq_printf(m, ",osd_idle_ttl=%d", args->osd_idle_ttl); - if (args->osd_timeout != CEPH_OSD_TIMEOUT_DEFAULT) - seq_printf(m, ",osdtimeout=%d", args->osd_timeout); - if (args->osd_keepalive_timeout != CEPH_OSD_KEEPALIVE_DEFAULT) - seq_printf(m, ",osdkeepalivetimeout=%d", - args->osd_keepalive_timeout); - if (args->wsize) - seq_printf(m, ",wsize=%d", args->wsize); - if (args->rsize != CEPH_MOUNT_RSIZE_DEFAULT) - seq_printf(m, ",rsize=%d", args->rsize); - if (args->congestion_kb != default_congestion_kb()) - seq_printf(m, ",write_congestion_kb=%d", args->congestion_kb); - if (args->caps_wanted_delay_min != CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT) - seq_printf(m, ",caps_wanted_delay_min=%d", - args->caps_wanted_delay_min); - if (args->caps_wanted_delay_max != CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT) - seq_printf(m, ",caps_wanted_delay_max=%d", - args->caps_wanted_delay_max); - if (args->cap_release_safety != CEPH_CAP_RELEASE_SAFETY_DEFAULT) - seq_printf(m, ",cap_release_safety=%d", - args->cap_release_safety); - if (args->max_readdir != CEPH_MAX_READDIR_DEFAULT) - seq_printf(m, ",readdir_max_entries=%d", args->max_readdir); - if (args->max_readdir_bytes != CEPH_MAX_READDIR_BYTES_DEFAULT) - seq_printf(m, ",readdir_max_bytes=%d", args->max_readdir_bytes); - if (strcmp(args->snapdir_name, CEPH_SNAPDIRNAME_DEFAULT)) - seq_printf(m, ",snapdirname=%s", args->snapdir_name); - if (args->name) - seq_printf(m, ",name=%s", args->name); - if (args->secret) - seq_puts(m, ",secret="); - return 0; -} - -/* - * caches - */ -struct kmem_cache *ceph_inode_cachep; -struct kmem_cache *ceph_cap_cachep; -struct kmem_cache *ceph_dentry_cachep; -struct kmem_cache *ceph_file_cachep; - -static void ceph_inode_init_once(void *foo) -{ - struct ceph_inode_info *ci = foo; - inode_init_once(&ci->vfs_inode); -} - -static int __init init_caches(void) -{ - ceph_inode_cachep = kmem_cache_create("ceph_inode_info", - sizeof(struct ceph_inode_info), - __alignof__(struct ceph_inode_info), - (SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD), - ceph_inode_init_once); - if (ceph_inode_cachep == NULL) - return -ENOMEM; - - ceph_cap_cachep = KMEM_CACHE(ceph_cap, - SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); - if (ceph_cap_cachep == NULL) - goto bad_cap; - - ceph_dentry_cachep = KMEM_CACHE(ceph_dentry_info, - SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); - if (ceph_dentry_cachep == NULL) - goto bad_dentry; - - ceph_file_cachep = KMEM_CACHE(ceph_file_info, - SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); - if (ceph_file_cachep == NULL) - goto bad_file; - - return 0; - -bad_file: - kmem_cache_destroy(ceph_dentry_cachep); -bad_dentry: - kmem_cache_destroy(ceph_cap_cachep); -bad_cap: - kmem_cache_destroy(ceph_inode_cachep); - return -ENOMEM; -} - -static void destroy_caches(void) -{ - kmem_cache_destroy(ceph_inode_cachep); - kmem_cache_destroy(ceph_cap_cachep); - kmem_cache_destroy(ceph_dentry_cachep); - kmem_cache_destroy(ceph_file_cachep); -} - - -/* - * ceph_umount_begin - initiate forced umount. Tear down down the - * mount, skipping steps that may hang while waiting for server(s). - */ -static void ceph_umount_begin(struct super_block *sb) -{ - struct ceph_client *client = ceph_sb_to_client(sb); - - dout("ceph_umount_begin - starting forced umount\n"); - if (!client) - return; - client->mount_state = CEPH_MOUNT_SHUTDOWN; - return; -} - -static const struct super_operations ceph_super_ops = { - .alloc_inode = ceph_alloc_inode, - .destroy_inode = ceph_destroy_inode, - .write_inode = ceph_write_inode, - .sync_fs = ceph_sync_fs, - .put_super = ceph_put_super, - .show_options = ceph_show_options, - .statfs = ceph_statfs, - .umount_begin = ceph_umount_begin, -}; - - -const char *ceph_msg_type_name(int type) -{ - switch (type) { - case CEPH_MSG_SHUTDOWN: return "shutdown"; - case CEPH_MSG_PING: return "ping"; - case CEPH_MSG_AUTH: return "auth"; - case CEPH_MSG_AUTH_REPLY: return "auth_reply"; - case CEPH_MSG_MON_MAP: return "mon_map"; - case CEPH_MSG_MON_GET_MAP: return "mon_get_map"; - case CEPH_MSG_MON_SUBSCRIBE: return "mon_subscribe"; - case CEPH_MSG_MON_SUBSCRIBE_ACK: return "mon_subscribe_ack"; - case CEPH_MSG_STATFS: return "statfs"; - case CEPH_MSG_STATFS_REPLY: return "statfs_reply"; - case CEPH_MSG_MDS_MAP: return "mds_map"; - case CEPH_MSG_CLIENT_SESSION: return "client_session"; - case CEPH_MSG_CLIENT_RECONNECT: return "client_reconnect"; - case CEPH_MSG_CLIENT_REQUEST: return "client_request"; - case CEPH_MSG_CLIENT_REQUEST_FORWARD: return "client_request_forward"; - case CEPH_MSG_CLIENT_REPLY: return "client_reply"; - case CEPH_MSG_CLIENT_CAPS: return "client_caps"; - case CEPH_MSG_CLIENT_CAPRELEASE: return "client_cap_release"; - case CEPH_MSG_CLIENT_SNAP: return "client_snap"; - case CEPH_MSG_CLIENT_LEASE: return "client_lease"; - case CEPH_MSG_OSD_MAP: return "osd_map"; - case CEPH_MSG_OSD_OP: return "osd_op"; - case CEPH_MSG_OSD_OPREPLY: return "osd_opreply"; - default: return "unknown"; - } -} - - /* * mount options */ enum { Opt_wsize, Opt_rsize, - Opt_osdtimeout, - Opt_osdkeepalivetimeout, - Opt_mount_timeout, - Opt_osd_idle_ttl, Opt_caps_wanted_delay_min, Opt_caps_wanted_delay_max, Opt_cap_release_safety, @@ -344,29 +123,19 @@ enum { Opt_congestion_kb, Opt_last_int, /* int args above */ - Opt_fsid, Opt_snapdirname, - Opt_name, - Opt_secret, Opt_last_string, /* string args above */ - Opt_ip, - Opt_noshare, Opt_dirstat, Opt_nodirstat, Opt_rbytes, Opt_norbytes, - Opt_nocrc, Opt_noasyncreaddir, }; -static match_table_t arg_tokens = { +static match_table_t fsopt_tokens = { {Opt_wsize, "wsize=%d"}, {Opt_rsize, "rsize=%d"}, - {Opt_osdtimeout, "osdtimeout=%d"}, - {Opt_osdkeepalivetimeout, "osdkeepalive=%d"}, - {Opt_mount_timeout, "mount_timeout=%d"}, - {Opt_osd_idle_ttl, "osd_idle_ttl=%d"}, {Opt_caps_wanted_delay_min, "caps_wanted_delay_min=%d"}, {Opt_caps_wanted_delay_max, "caps_wanted_delay_max=%d"}, {Opt_cap_release_safety, "cap_release_safety=%d"}, @@ -374,403 +143,459 @@ static match_table_t arg_tokens = { {Opt_readdir_max_bytes, "readdir_max_bytes=%d"}, {Opt_congestion_kb, "write_congestion_kb=%d"}, /* int args above */ - {Opt_fsid, "fsid=%s"}, {Opt_snapdirname, "snapdirname=%s"}, - {Opt_name, "name=%s"}, - {Opt_secret, "secret=%s"}, /* string args above */ - {Opt_ip, "ip=%s"}, - {Opt_noshare, "noshare"}, {Opt_dirstat, "dirstat"}, {Opt_nodirstat, "nodirstat"}, {Opt_rbytes, "rbytes"}, {Opt_norbytes, "norbytes"}, - {Opt_nocrc, "nocrc"}, {Opt_noasyncreaddir, "noasyncreaddir"}, {-1, NULL} }; -static int parse_fsid(const char *str, struct ceph_fsid *fsid) +static int parse_fsopt_token(char *c, void *private) { - int i = 0; - char tmp[3]; - int err = -EINVAL; - int d; - - dout("parse_fsid '%s'\n", str); - tmp[2] = 0; - while (*str && i < 16) { - if (ispunct(*str)) { - str++; - continue; + struct ceph_mount_options *fsopt = private; + substring_t argstr[MAX_OPT_ARGS]; + int token, intval, ret; + + token = match_token((char *)c, fsopt_tokens, argstr); + if (token < 0) + return -EINVAL; + + if (token < Opt_last_int) { + ret = match_int(&argstr[0], &intval); + if (ret < 0) { + pr_err("bad mount option arg (not int) " + "at '%s'\n", c); + return ret; } - if (!isxdigit(str[0]) || !isxdigit(str[1])) - break; - tmp[0] = str[0]; - tmp[1] = str[1]; - if (sscanf(tmp, "%x", &d) < 1) - break; - fsid->fsid[i] = d & 0xff; - i++; - str += 2; + dout("got int token %d val %d\n", token, intval); + } else if (token > Opt_last_int && token < Opt_last_string) { + dout("got string token %d val %s\n", token, + argstr[0].from); + } else { + dout("got token %d\n", token); } - if (i == 16) - err = 0; - dout("parse_fsid ret %d got fsid %pU", err, fsid); - return err; + switch (token) { + case Opt_snapdirname: + kfree(fsopt->snapdir_name); + fsopt->snapdir_name = kstrndup(argstr[0].from, + argstr[0].to-argstr[0].from, + GFP_KERNEL); + if (!fsopt->snapdir_name) + return -ENOMEM; + break; + + /* misc */ + case Opt_wsize: + fsopt->wsize = intval; + break; + case Opt_rsize: + fsopt->rsize = intval; + break; + case Opt_caps_wanted_delay_min: + fsopt->caps_wanted_delay_min = intval; + break; + case Opt_caps_wanted_delay_max: + fsopt->caps_wanted_delay_max = intval; + break; + case Opt_readdir_max_entries: + fsopt->max_readdir = intval; + break; + case Opt_readdir_max_bytes: + fsopt->max_readdir_bytes = intval; + break; + case Opt_congestion_kb: + fsopt->congestion_kb = intval; + break; + case Opt_dirstat: + fsopt->flags |= CEPH_MOUNT_OPT_DIRSTAT; + break; + case Opt_nodirstat: + fsopt->flags &= ~CEPH_MOUNT_OPT_DIRSTAT; + break; + case Opt_rbytes: + fsopt->flags |= CEPH_MOUNT_OPT_RBYTES; + break; + case Opt_norbytes: + fsopt->flags &= ~CEPH_MOUNT_OPT_RBYTES; + break; + case Opt_noasyncreaddir: + fsopt->flags |= CEPH_MOUNT_OPT_NOASYNCREADDIR; + break; + default: + BUG_ON(token); + } + return 0; } -static struct ceph_mount_args *parse_mount_args(int flags, char *options, - const char *dev_name, - const char **path) +static void destroy_mount_options(struct ceph_mount_options *args) { - struct ceph_mount_args *args; - const char *c; - int err = -ENOMEM; - substring_t argstr[MAX_OPT_ARGS]; + dout("destroy_mount_options %p\n", args); + kfree(args->snapdir_name); + kfree(args); +} - args = kzalloc(sizeof(*args), GFP_KERNEL); - if (!args) - return ERR_PTR(-ENOMEM); - args->mon_addr = kcalloc(CEPH_MAX_MON, sizeof(*args->mon_addr), - GFP_KERNEL); - if (!args->mon_addr) - goto out; +static int strcmp_null(const char *s1, const char *s2) +{ + if (!s1 && !s2) + return 0; + if (s1 && !s2) + return -1; + if (!s1 && s2) + return 1; + return strcmp(s1, s2); +} - dout("parse_mount_args %p, dev_name '%s'\n", args, dev_name); - - /* start with defaults */ - args->sb_flags = flags; - args->flags = CEPH_OPT_DEFAULT; - args->osd_timeout = CEPH_OSD_TIMEOUT_DEFAULT; - args->osd_keepalive_timeout = CEPH_OSD_KEEPALIVE_DEFAULT; - args->mount_timeout = CEPH_MOUNT_TIMEOUT_DEFAULT; /* seconds */ - args->osd_idle_ttl = CEPH_OSD_IDLE_TTL_DEFAULT; /* seconds */ - args->caps_wanted_delay_min = CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT; - args->caps_wanted_delay_max = CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT; - args->rsize = CEPH_MOUNT_RSIZE_DEFAULT; - args->snapdir_name = kstrdup(CEPH_SNAPDIRNAME_DEFAULT, GFP_KERNEL); - args->cap_release_safety = CEPH_CAP_RELEASE_SAFETY_DEFAULT; - args->max_readdir = CEPH_MAX_READDIR_DEFAULT; - args->max_readdir_bytes = CEPH_MAX_READDIR_BYTES_DEFAULT; - args->congestion_kb = default_congestion_kb(); - - /* ip1[:port1][,ip2[:port2]...]:/subdir/in/fs */ - err = -EINVAL; - if (!dev_name) - goto out; - *path = strstr(dev_name, ":/"); - if (*path == NULL) { - pr_err("device name is missing path (no :/ in %s)\n", - dev_name); - goto out; - } +static int compare_mount_options(struct ceph_mount_options *new_fsopt, + struct ceph_options *new_opt, + struct ceph_fs_client *fsc) +{ + struct ceph_mount_options *fsopt1 = new_fsopt; + struct ceph_mount_options *fsopt2 = fsc->mount_options; + int ofs = offsetof(struct ceph_mount_options, snapdir_name); + int ret; - /* get mon ip(s) */ - err = ceph_parse_ips(dev_name, *path, args->mon_addr, - CEPH_MAX_MON, &args->num_mon); - if (err < 0) - goto out; + ret = memcmp(fsopt1, fsopt2, ofs); + if (ret) + return ret; + + ret = strcmp_null(fsopt1->snapdir_name, fsopt2->snapdir_name); + if (ret) + return ret; + + return ceph_compare_options(new_opt, fsc->client); +} + +static int parse_mount_options(struct ceph_mount_options **pfsopt, + struct ceph_options **popt, + int flags, char *options, + const char *dev_name, + const char **path) +{ + struct ceph_mount_options *fsopt; + const char *dev_name_end; + int err = -ENOMEM; + + fsopt = kzalloc(sizeof(*fsopt), GFP_KERNEL); + if (!fsopt) + return -ENOMEM; + + dout("parse_mount_options %p, dev_name '%s'\n", fsopt, dev_name); + + fsopt->sb_flags = flags; + fsopt->flags = CEPH_MOUNT_OPT_DEFAULT; + + fsopt->rsize = CEPH_MOUNT_RSIZE_DEFAULT; + fsopt->snapdir_name = kstrdup(CEPH_SNAPDIRNAME_DEFAULT, GFP_KERNEL); + fsopt->cap_release_safety = CEPH_CAP_RELEASE_SAFETY_DEFAULT; + fsopt->max_readdir = CEPH_MAX_READDIR_DEFAULT; + fsopt->max_readdir_bytes = CEPH_MAX_READDIR_BYTES_DEFAULT; + fsopt->congestion_kb = default_congestion_kb(); + + /* ip1[:port1][,ip2[:port2]...]:/subdir/in/fs */ + err = -EINVAL; + if (!dev_name) + goto out; + *path = strstr(dev_name, ":/"); + if (*path == NULL) { + pr_err("device name is missing path (no :/ in %s)\n", + dev_name); + goto out; + } + dev_name_end = *path; + dout("device name '%.*s'\n", (int)(dev_name_end - dev_name), dev_name); /* path on server */ *path += 2; dout("server path '%s'\n", *path); - /* parse mount options */ - while ((c = strsep(&options, ",")) != NULL) { - int token, intval, ret; - if (!*c) - continue; - err = -EINVAL; - token = match_token((char *)c, arg_tokens, argstr); - if (token < 0) { - pr_err("bad mount option at '%s'\n", c); - goto out; - } - if (token < Opt_last_int) { - ret = match_int(&argstr[0], &intval); - if (ret < 0) { - pr_err("bad mount option arg (not int) " - "at '%s'\n", c); - continue; - } - dout("got int token %d val %d\n", token, intval); - } else if (token > Opt_last_int && token < Opt_last_string) { - dout("got string token %d val %s\n", token, - argstr[0].from); - } else { - dout("got token %d\n", token); - } - switch (token) { - case Opt_ip: - err = ceph_parse_ips(argstr[0].from, - argstr[0].to, - &args->my_addr, - 1, NULL); - if (err < 0) - goto out; - args->flags |= CEPH_OPT_MYIP; - break; - - case Opt_fsid: - err = parse_fsid(argstr[0].from, &args->fsid); - if (err == 0) - args->flags |= CEPH_OPT_FSID; - break; - case Opt_snapdirname: - kfree(args->snapdir_name); - args->snapdir_name = kstrndup(argstr[0].from, - argstr[0].to-argstr[0].from, - GFP_KERNEL); - break; - case Opt_name: - args->name = kstrndup(argstr[0].from, - argstr[0].to-argstr[0].from, - GFP_KERNEL); - break; - case Opt_secret: - args->secret = kstrndup(argstr[0].from, - argstr[0].to-argstr[0].from, - GFP_KERNEL); - break; - - /* misc */ - case Opt_wsize: - args->wsize = intval; - break; - case Opt_rsize: - args->rsize = intval; - break; - case Opt_osdtimeout: - args->osd_timeout = intval; - break; - case Opt_osdkeepalivetimeout: - args->osd_keepalive_timeout = intval; - break; - case Opt_osd_idle_ttl: - args->osd_idle_ttl = intval; - break; - case Opt_mount_timeout: - args->mount_timeout = intval; - break; - case Opt_caps_wanted_delay_min: - args->caps_wanted_delay_min = intval; - break; - case Opt_caps_wanted_delay_max: - args->caps_wanted_delay_max = intval; - break; - case Opt_readdir_max_entries: - args->max_readdir = intval; - break; - case Opt_readdir_max_bytes: - args->max_readdir_bytes = intval; - break; - case Opt_congestion_kb: - args->congestion_kb = intval; - break; - - case Opt_noshare: - args->flags |= CEPH_OPT_NOSHARE; - break; - - case Opt_dirstat: - args->flags |= CEPH_OPT_DIRSTAT; - break; - case Opt_nodirstat: - args->flags &= ~CEPH_OPT_DIRSTAT; - break; - case Opt_rbytes: - args->flags |= CEPH_OPT_RBYTES; - break; - case Opt_norbytes: - args->flags &= ~CEPH_OPT_RBYTES; - break; - case Opt_nocrc: - args->flags |= CEPH_OPT_NOCRC; - break; - case Opt_noasyncreaddir: - args->flags |= CEPH_OPT_NOASYNCREADDIR; - break; - - default: - BUG_ON(token); - } - } - return args; + err = ceph_parse_options(popt, options, dev_name, dev_name_end, + parse_fsopt_token, (void *)fsopt); + if (err) + goto out; + + /* success */ + *pfsopt = fsopt; + return 0; out: - kfree(args->mon_addr); - kfree(args); - return ERR_PTR(err); + destroy_mount_options(fsopt); + return err; } -static void destroy_mount_args(struct ceph_mount_args *args) +/** + * ceph_show_options - Show mount options in /proc/mounts + * @m: seq_file to write to + * @mnt: mount descriptor + */ +static int ceph_show_options(struct seq_file *m, struct vfsmount *mnt) { - dout("destroy_mount_args %p\n", args); - kfree(args->snapdir_name); - args->snapdir_name = NULL; - kfree(args->name); - args->name = NULL; - kfree(args->secret); - args->secret = NULL; - kfree(args); + struct ceph_fs_client *fsc = ceph_sb_to_client(mnt->mnt_sb); + struct ceph_mount_options *fsopt = fsc->mount_options; + struct ceph_options *opt = fsc->client->options; + + if (opt->flags & CEPH_OPT_FSID) + seq_printf(m, ",fsid=%pU", &opt->fsid); + if (opt->flags & CEPH_OPT_NOSHARE) + seq_puts(m, ",noshare"); + if (opt->flags & CEPH_OPT_NOCRC) + seq_puts(m, ",nocrc"); + + if (opt->name) + seq_printf(m, ",name=%s", opt->name); + if (opt->secret) + seq_puts(m, ",secret="); + + if (opt->mount_timeout != CEPH_MOUNT_TIMEOUT_DEFAULT) + seq_printf(m, ",mount_timeout=%d", opt->mount_timeout); + if (opt->osd_idle_ttl != CEPH_OSD_IDLE_TTL_DEFAULT) + seq_printf(m, ",osd_idle_ttl=%d", opt->osd_idle_ttl); + if (opt->osd_timeout != CEPH_OSD_TIMEOUT_DEFAULT) + seq_printf(m, ",osdtimeout=%d", opt->osd_timeout); + if (opt->osd_keepalive_timeout != CEPH_OSD_KEEPALIVE_DEFAULT) + seq_printf(m, ",osdkeepalivetimeout=%d", + opt->osd_keepalive_timeout); + + if (fsopt->flags & CEPH_MOUNT_OPT_DIRSTAT) + seq_puts(m, ",dirstat"); + if ((fsopt->flags & CEPH_MOUNT_OPT_RBYTES) == 0) + seq_puts(m, ",norbytes"); + if (fsopt->flags & CEPH_MOUNT_OPT_NOASYNCREADDIR) + seq_puts(m, ",noasyncreaddir"); + + if (fsopt->wsize) + seq_printf(m, ",wsize=%d", fsopt->wsize); + if (fsopt->rsize != CEPH_MOUNT_RSIZE_DEFAULT) + seq_printf(m, ",rsize=%d", fsopt->rsize); + if (fsopt->congestion_kb != default_congestion_kb()) + seq_printf(m, ",write_congestion_kb=%d", fsopt->congestion_kb); + if (fsopt->caps_wanted_delay_min != CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT) + seq_printf(m, ",caps_wanted_delay_min=%d", + fsopt->caps_wanted_delay_min); + if (fsopt->caps_wanted_delay_max != CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT) + seq_printf(m, ",caps_wanted_delay_max=%d", + fsopt->caps_wanted_delay_max); + if (fsopt->cap_release_safety != CEPH_CAP_RELEASE_SAFETY_DEFAULT) + seq_printf(m, ",cap_release_safety=%d", + fsopt->cap_release_safety); + if (fsopt->max_readdir != CEPH_MAX_READDIR_DEFAULT) + seq_printf(m, ",readdir_max_entries=%d", fsopt->max_readdir); + if (fsopt->max_readdir_bytes != CEPH_MAX_READDIR_BYTES_DEFAULT) + seq_printf(m, ",readdir_max_bytes=%d", fsopt->max_readdir_bytes); + if (strcmp(fsopt->snapdir_name, CEPH_SNAPDIRNAME_DEFAULT)) + seq_printf(m, ",snapdirname=%s", fsopt->snapdir_name); + return 0; } /* - * create a fresh client instance + * handle any mon messages the standard library doesn't understand. + * return error if we don't either. */ -static struct ceph_client *ceph_create_client(struct ceph_mount_args *args) +static int extra_mon_dispatch(struct ceph_client *client, struct ceph_msg *msg) { - struct ceph_client *client; + struct ceph_fs_client *fsc = client->private; + int type = le16_to_cpu(msg->hdr.type); + + switch (type) { + case CEPH_MSG_MDS_MAP: + ceph_mdsc_handle_map(fsc->mdsc, msg); + return 0; + + default: + return -1; + } +} + +/* + * create a new fs client + */ +struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, + struct ceph_options *opt) +{ + struct ceph_fs_client *fsc; int err = -ENOMEM; - client = kzalloc(sizeof(*client), GFP_KERNEL); - if (client == NULL) + fsc = kzalloc(sizeof(*fsc), GFP_KERNEL); + if (!fsc) return ERR_PTR(-ENOMEM); - mutex_init(&client->mount_mutex); - - init_waitqueue_head(&client->auth_wq); + fsc->client = ceph_create_client(opt, fsc); + if (IS_ERR(fsc->client)) { + err = PTR_ERR(fsc->client); + goto fail; + } + fsc->client->extra_mon_dispatch = extra_mon_dispatch; + fsc->client->supported_features |= CEPH_FEATURE_FLOCK; + fsc->client->monc.want_mdsmap = 1; - client->sb = NULL; - client->mount_state = CEPH_MOUNT_MOUNTING; - client->mount_args = args; + fsc->mount_options = fsopt; - client->msgr = NULL; + fsc->sb = NULL; + fsc->mount_state = CEPH_MOUNT_MOUNTING; - client->auth_err = 0; - atomic_long_set(&client->writeback_count, 0); + atomic_long_set(&fsc->writeback_count, 0); - err = bdi_init(&client->backing_dev_info); + err = bdi_init(&fsc->backing_dev_info); if (err < 0) - goto fail; + goto fail_client; err = -ENOMEM; - client->wb_wq = create_workqueue("ceph-writeback"); - if (client->wb_wq == NULL) + fsc->wb_wq = create_workqueue("ceph-writeback"); + if (fsc->wb_wq == NULL) goto fail_bdi; - client->pg_inv_wq = create_singlethread_workqueue("ceph-pg-invalid"); - if (client->pg_inv_wq == NULL) + fsc->pg_inv_wq = create_singlethread_workqueue("ceph-pg-invalid"); + if (fsc->pg_inv_wq == NULL) goto fail_wb_wq; - client->trunc_wq = create_singlethread_workqueue("ceph-trunc"); - if (client->trunc_wq == NULL) + fsc->trunc_wq = create_singlethread_workqueue("ceph-trunc"); + if (fsc->trunc_wq == NULL) goto fail_pg_inv_wq; /* set up mempools */ err = -ENOMEM; - client->wb_pagevec_pool = mempool_create_kmalloc_pool(10, - client->mount_args->wsize >> PAGE_CACHE_SHIFT); - if (!client->wb_pagevec_pool) + fsc->wb_pagevec_pool = mempool_create_kmalloc_pool(10, + fsc->mount_options->wsize >> PAGE_CACHE_SHIFT); + if (!fsc->wb_pagevec_pool) goto fail_trunc_wq; /* caps */ - client->min_caps = args->max_readdir; + fsc->min_caps = fsopt->max_readdir; + + return fsc; - /* subsystems */ - err = ceph_monc_init(&client->monc, client); - if (err < 0) - goto fail_mempool; - err = ceph_osdc_init(&client->osdc, client); - if (err < 0) - goto fail_monc; - err = ceph_mdsc_init(&client->mdsc, client); - if (err < 0) - goto fail_osdc; - return client; - -fail_osdc: - ceph_osdc_stop(&client->osdc); -fail_monc: - ceph_monc_stop(&client->monc); -fail_mempool: - mempool_destroy(client->wb_pagevec_pool); fail_trunc_wq: - destroy_workqueue(client->trunc_wq); + destroy_workqueue(fsc->trunc_wq); fail_pg_inv_wq: - destroy_workqueue(client->pg_inv_wq); + destroy_workqueue(fsc->pg_inv_wq); fail_wb_wq: - destroy_workqueue(client->wb_wq); + destroy_workqueue(fsc->wb_wq); fail_bdi: - bdi_destroy(&client->backing_dev_info); + bdi_destroy(&fsc->backing_dev_info); +fail_client: + ceph_destroy_client(fsc->client); fail: - kfree(client); + kfree(fsc); return ERR_PTR(err); } -static void ceph_destroy_client(struct ceph_client *client) +void destroy_fs_client(struct ceph_fs_client *fsc) { - dout("destroy_client %p\n", client); + dout("destroy_fs_client %p\n", fsc); - /* unmount */ - ceph_mdsc_stop(&client->mdsc); - ceph_osdc_stop(&client->osdc); + destroy_workqueue(fsc->wb_wq); + destroy_workqueue(fsc->pg_inv_wq); + destroy_workqueue(fsc->trunc_wq); - /* - * make sure mds and osd connections close out before destroying - * the auth module, which is needed to free those connections' - * ceph_authorizers. - */ - ceph_msgr_flush(); - - ceph_monc_stop(&client->monc); + bdi_destroy(&fsc->backing_dev_info); - ceph_debugfs_client_cleanup(client); - destroy_workqueue(client->wb_wq); - destroy_workqueue(client->pg_inv_wq); - destroy_workqueue(client->trunc_wq); + mempool_destroy(fsc->wb_pagevec_pool); - bdi_destroy(&client->backing_dev_info); + destroy_mount_options(fsc->mount_options); - if (client->msgr) - ceph_messenger_destroy(client->msgr); - mempool_destroy(client->wb_pagevec_pool); + ceph_fs_debugfs_cleanup(fsc); - destroy_mount_args(client->mount_args); + ceph_destroy_client(fsc->client); - kfree(client); - dout("destroy_client %p done\n", client); + kfree(fsc); + dout("destroy_fs_client %p done\n", fsc); } /* - * Initially learn our fsid, or verify an fsid matches. + * caches */ -int ceph_check_fsid(struct ceph_client *client, struct ceph_fsid *fsid) +struct kmem_cache *ceph_inode_cachep; +struct kmem_cache *ceph_cap_cachep; +struct kmem_cache *ceph_dentry_cachep; +struct kmem_cache *ceph_file_cachep; + +static void ceph_inode_init_once(void *foo) { - if (client->have_fsid) { - if (ceph_fsid_compare(&client->fsid, fsid)) { - pr_err("bad fsid, had %pU got %pU", - &client->fsid, fsid); - return -1; - } - } else { - pr_info("client%lld fsid %pU\n", client->monc.auth->global_id, - fsid); - memcpy(&client->fsid, fsid, sizeof(*fsid)); - ceph_debugfs_client_init(client); - client->have_fsid = true; - } + struct ceph_inode_info *ci = foo; + inode_init_once(&ci->vfs_inode); +} + +static int __init init_caches(void) +{ + ceph_inode_cachep = kmem_cache_create("ceph_inode_info", + sizeof(struct ceph_inode_info), + __alignof__(struct ceph_inode_info), + (SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD), + ceph_inode_init_once); + if (ceph_inode_cachep == NULL) + return -ENOMEM; + + ceph_cap_cachep = KMEM_CACHE(ceph_cap, + SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); + if (ceph_cap_cachep == NULL) + goto bad_cap; + + ceph_dentry_cachep = KMEM_CACHE(ceph_dentry_info, + SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); + if (ceph_dentry_cachep == NULL) + goto bad_dentry; + + ceph_file_cachep = KMEM_CACHE(ceph_file_info, + SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); + if (ceph_file_cachep == NULL) + goto bad_file; + return 0; + +bad_file: + kmem_cache_destroy(ceph_dentry_cachep); +bad_dentry: + kmem_cache_destroy(ceph_cap_cachep); +bad_cap: + kmem_cache_destroy(ceph_inode_cachep); + return -ENOMEM; } +static void destroy_caches(void) +{ + kmem_cache_destroy(ceph_inode_cachep); + kmem_cache_destroy(ceph_cap_cachep); + kmem_cache_destroy(ceph_dentry_cachep); + kmem_cache_destroy(ceph_file_cachep); +} + + /* - * true if we have the mon map (and have thus joined the cluster) + * ceph_umount_begin - initiate forced umount. Tear down down the + * mount, skipping steps that may hang while waiting for server(s). */ -static int have_mon_and_osd_map(struct ceph_client *client) +static void ceph_umount_begin(struct super_block *sb) { - return client->monc.monmap && client->monc.monmap->epoch && - client->osdc.osdmap && client->osdc.osdmap->epoch; + struct ceph_fs_client *fsc = ceph_sb_to_client(sb); + + dout("ceph_umount_begin - starting forced umount\n"); + if (!fsc) + return; + fsc->mount_state = CEPH_MOUNT_SHUTDOWN; + return; } +static const struct super_operations ceph_super_ops = { + .alloc_inode = ceph_alloc_inode, + .destroy_inode = ceph_destroy_inode, + .write_inode = ceph_write_inode, + .sync_fs = ceph_sync_fs, + .put_super = ceph_put_super, + .show_options = ceph_show_options, + .statfs = ceph_statfs, + .umount_begin = ceph_umount_begin, +}; + /* * Bootstrap mount by opening the root directory. Note the mount * @started time from caller, and time out if this takes too long. */ -static struct dentry *open_root_dentry(struct ceph_client *client, +static struct dentry *open_root_dentry(struct ceph_fs_client *fsc, const char *path, unsigned long started) { - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_mds_request *req = NULL; int err; struct dentry *root; @@ -784,14 +609,14 @@ static struct dentry *open_root_dentry(struct ceph_client *client, req->r_ino1.ino = CEPH_INO_ROOT; req->r_ino1.snap = CEPH_NOSNAP; req->r_started = started; - req->r_timeout = client->mount_args->mount_timeout * HZ; + req->r_timeout = fsc->client->options->mount_timeout * HZ; req->r_args.getattr.mask = cpu_to_le32(CEPH_STAT_CAP_INODE); req->r_num_caps = 2; err = ceph_mdsc_do_request(mdsc, NULL, req); if (err == 0) { dout("open_root_inode success\n"); if (ceph_ino(req->r_target_inode) == CEPH_INO_ROOT && - client->sb->s_root == NULL) + fsc->sb->s_root == NULL) root = d_alloc_root(req->r_target_inode); else root = d_obtain_alias(req->r_target_inode); @@ -804,105 +629,86 @@ static struct dentry *open_root_dentry(struct ceph_client *client, return root; } + + + /* * mount: join the ceph cluster, and open root directory. */ -static int ceph_mount(struct ceph_client *client, struct vfsmount *mnt, +static int ceph_mount(struct ceph_fs_client *fsc, struct vfsmount *mnt, const char *path) { - struct ceph_entity_addr *myaddr = NULL; int err; - unsigned long timeout = client->mount_args->mount_timeout * HZ; unsigned long started = jiffies; /* note the start time */ struct dentry *root; + int first = 0; /* first vfsmount for this super_block */ dout("mount start\n"); - mutex_lock(&client->mount_mutex); - - /* initialize the messenger */ - if (client->msgr == NULL) { - if (ceph_test_opt(client, MYIP)) - myaddr = &client->mount_args->my_addr; - client->msgr = ceph_messenger_create(myaddr); - if (IS_ERR(client->msgr)) { - err = PTR_ERR(client->msgr); - client->msgr = NULL; - goto out; - } - client->msgr->nocrc = ceph_test_opt(client, NOCRC); - } + mutex_lock(&fsc->client->mount_mutex); - /* open session, and wait for mon, mds, and osd maps */ - err = ceph_monc_open_session(&client->monc); + err = __ceph_open_session(fsc->client, started); if (err < 0) goto out; - while (!have_mon_and_osd_map(client)) { - err = -EIO; - if (timeout && time_after_eq(jiffies, started + timeout)) - goto out; - - /* wait */ - dout("mount waiting for mon_map\n"); - err = wait_event_interruptible_timeout(client->auth_wq, - have_mon_and_osd_map(client) || (client->auth_err < 0), - timeout); - if (err == -EINTR || err == -ERESTARTSYS) - goto out; - if (client->auth_err < 0) { - err = client->auth_err; - goto out; - } - } - dout("mount opening root\n"); - root = open_root_dentry(client, "", started); + root = open_root_dentry(fsc, "", started); if (IS_ERR(root)) { err = PTR_ERR(root); goto out; } - if (client->sb->s_root) + if (fsc->sb->s_root) { dput(root); - else - client->sb->s_root = root; + } else { + fsc->sb->s_root = root; + first = 1; + + err = ceph_fs_debugfs_init(fsc); + if (err < 0) + goto fail; + } if (path[0] == 0) { dget(root); } else { dout("mount opening base mountpoint\n"); - root = open_root_dentry(client, path, started); + root = open_root_dentry(fsc, path, started); if (IS_ERR(root)) { err = PTR_ERR(root); - dput(client->sb->s_root); - client->sb->s_root = NULL; - goto out; + goto fail; } } mnt->mnt_root = root; - mnt->mnt_sb = client->sb; + mnt->mnt_sb = fsc->sb; - client->mount_state = CEPH_MOUNT_MOUNTED; + fsc->mount_state = CEPH_MOUNT_MOUNTED; dout("mount success\n"); err = 0; out: - mutex_unlock(&client->mount_mutex); + mutex_unlock(&fsc->client->mount_mutex); return err; + +fail: + if (first) { + dput(fsc->sb->s_root); + fsc->sb->s_root = NULL; + } + goto out; } static int ceph_set_super(struct super_block *s, void *data) { - struct ceph_client *client = data; + struct ceph_fs_client *fsc = data; int ret; dout("set_super %p data %p\n", s, data); - s->s_flags = client->mount_args->sb_flags; + s->s_flags = fsc->mount_options->sb_flags; s->s_maxbytes = 1ULL << 40; /* temp value until we get mdsmap */ - s->s_fs_info = client; - client->sb = s; + s->s_fs_info = fsc; + fsc->sb = s; s->s_op = &ceph_super_ops; s->s_export_op = &ceph_export_ops; @@ -917,7 +723,7 @@ static int ceph_set_super(struct super_block *s, void *data) fail: s->s_fs_info = NULL; - client->sb = NULL; + fsc->sb = NULL; return ret; } @@ -926,30 +732,23 @@ fail: */ static int ceph_compare_super(struct super_block *sb, void *data) { - struct ceph_client *new = data; - struct ceph_mount_args *args = new->mount_args; - struct ceph_client *other = ceph_sb_to_client(sb); - int i; + struct ceph_fs_client *new = data; + struct ceph_mount_options *fsopt = new->mount_options; + struct ceph_options *opt = new->client->options; + struct ceph_fs_client *other = ceph_sb_to_client(sb); dout("ceph_compare_super %p\n", sb); - if (args->flags & CEPH_OPT_FSID) { - if (ceph_fsid_compare(&args->fsid, &other->fsid)) { - dout("fsid doesn't match\n"); - return 0; - } - } else { - /* do we share (a) monitor? */ - for (i = 0; i < new->monc.monmap->num_mon; i++) - if (ceph_monmap_contains(other->monc.monmap, - &new->monc.monmap->mon_inst[i].addr)) - break; - if (i == new->monc.monmap->num_mon) { - dout("mon ip not part of monmap\n"); - return 0; - } - dout("mon ip matches existing sb %p\n", sb); + + if (compare_mount_options(fsopt, opt, other)) { + dout("monitor(s)/mount options don't match\n"); + return 0; } - if (args->sb_flags != other->mount_args->sb_flags) { + if ((opt->flags & CEPH_OPT_FSID) && + ceph_fsid_compare(&opt->fsid, &other->client->fsid)) { + dout("fsid doesn't match\n"); + return 0; + } + if (fsopt->sb_flags != other->mount_options->sb_flags) { dout("flags differ\n"); return 0; } @@ -961,19 +760,20 @@ static int ceph_compare_super(struct super_block *sb, void *data) */ static atomic_long_t bdi_seq = ATOMIC_LONG_INIT(0); -static int ceph_register_bdi(struct super_block *sb, struct ceph_client *client) +static int ceph_register_bdi(struct super_block *sb, + struct ceph_fs_client *fsc) { int err; /* set ra_pages based on rsize mount option? */ - if (client->mount_args->rsize >= PAGE_CACHE_SIZE) - client->backing_dev_info.ra_pages = - (client->mount_args->rsize + PAGE_CACHE_SIZE - 1) + if (fsc->mount_options->rsize >= PAGE_CACHE_SIZE) + fsc->backing_dev_info.ra_pages = + (fsc->mount_options->rsize + PAGE_CACHE_SIZE - 1) >> PAGE_SHIFT; - err = bdi_register(&client->backing_dev_info, NULL, "ceph-%d", + err = bdi_register(&fsc->backing_dev_info, NULL, "ceph-%d", atomic_long_inc_return(&bdi_seq)); if (!err) - sb->s_bdi = &client->backing_dev_info; + sb->s_bdi = &fsc->backing_dev_info; return err; } @@ -982,46 +782,52 @@ static int ceph_get_sb(struct file_system_type *fs_type, struct vfsmount *mnt) { struct super_block *sb; - struct ceph_client *client; + struct ceph_fs_client *fsc; int err; int (*compare_super)(struct super_block *, void *) = ceph_compare_super; const char *path = NULL; - struct ceph_mount_args *args; + struct ceph_mount_options *fsopt = NULL; + struct ceph_options *opt = NULL; dout("ceph_get_sb\n"); - args = parse_mount_args(flags, data, dev_name, &path); - if (IS_ERR(args)) { - err = PTR_ERR(args); + err = parse_mount_options(&fsopt, &opt, flags, data, dev_name, &path); + if (err < 0) goto out_final; - } /* create client (which we may/may not use) */ - client = ceph_create_client(args); - if (IS_ERR(client)) { - err = PTR_ERR(client); + fsc = create_fs_client(fsopt, opt); + if (IS_ERR(fsc)) { + err = PTR_ERR(fsc); + kfree(fsopt); + kfree(opt); goto out_final; } - if (client->mount_args->flags & CEPH_OPT_NOSHARE) + err = ceph_mdsc_init(fsc); + if (err < 0) + goto out; + + if (ceph_test_opt(fsc->client, NOSHARE)) compare_super = NULL; - sb = sget(fs_type, compare_super, ceph_set_super, client); + sb = sget(fs_type, compare_super, ceph_set_super, fsc); if (IS_ERR(sb)) { err = PTR_ERR(sb); goto out; } - if (ceph_sb_to_client(sb) != client) { - ceph_destroy_client(client); - client = ceph_sb_to_client(sb); - dout("get_sb got existing client %p\n", client); + if (ceph_sb_to_client(sb) != fsc) { + ceph_mdsc_destroy(fsc); + destroy_fs_client(fsc); + fsc = ceph_sb_to_client(sb); + dout("get_sb got existing client %p\n", fsc); } else { - dout("get_sb using new client %p\n", client); - err = ceph_register_bdi(sb, client); + dout("get_sb using new client %p\n", fsc); + err = ceph_register_bdi(sb, fsc); if (err < 0) goto out_splat; } - err = ceph_mount(client, mnt, path); + err = ceph_mount(fsc, mnt, path); if (err < 0) goto out_splat; dout("root %p inode %p ino %llx.%llx\n", mnt->mnt_root, @@ -1029,12 +835,13 @@ static int ceph_get_sb(struct file_system_type *fs_type, return 0; out_splat: - ceph_mdsc_close_sessions(&client->mdsc); + ceph_mdsc_close_sessions(fsc->mdsc); deactivate_locked_super(sb); goto out_final; out: - ceph_destroy_client(client); + ceph_mdsc_destroy(fsc); + destroy_fs_client(fsc); out_final: dout("ceph_get_sb fail %d\n", err); return err; @@ -1042,11 +849,12 @@ out_final: static void ceph_kill_sb(struct super_block *s) { - struct ceph_client *client = ceph_sb_to_client(s); + struct ceph_fs_client *fsc = ceph_sb_to_client(s); dout("kill_sb %p\n", s); - ceph_mdsc_pre_umount(&client->mdsc); + ceph_mdsc_pre_umount(fsc->mdsc); kill_anon_super(s); /* will call put_super after sb is r/o */ - ceph_destroy_client(client); + ceph_mdsc_destroy(fsc); + destroy_fs_client(fsc); } static struct file_system_type ceph_fs_type = { @@ -1062,36 +870,20 @@ static struct file_system_type ceph_fs_type = { static int __init init_ceph(void) { - int ret = 0; - - ret = ceph_debugfs_init(); - if (ret < 0) - goto out; - - ret = ceph_msgr_init(); - if (ret < 0) - goto out_debugfs; - - ret = init_caches(); + int ret = init_caches(); if (ret) - goto out_msgr; + goto out; ret = register_filesystem(&ceph_fs_type); if (ret) goto out_icache; - pr_info("loaded (mon/mds/osd proto %d/%d/%d, osdmap %d/%d %d/%d)\n", - CEPH_MONC_PROTOCOL, CEPH_MDSC_PROTOCOL, CEPH_OSDC_PROTOCOL, - CEPH_OSDMAP_VERSION, CEPH_OSDMAP_VERSION_EXT, - CEPH_OSDMAP_INC_VERSION, CEPH_OSDMAP_INC_VERSION_EXT); + pr_info("loaded (mds proto %d)\n", CEPH_MDSC_PROTOCOL); + return 0; out_icache: destroy_caches(); -out_msgr: - ceph_msgr_exit(); -out_debugfs: - ceph_debugfs_cleanup(); out: return ret; } @@ -1101,8 +893,6 @@ static void __exit exit_ceph(void) dout("exit_ceph\n"); unregister_filesystem(&ceph_fs_type); destroy_caches(); - ceph_msgr_exit(); - ceph_debugfs_cleanup(); } module_init(init_ceph); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index b87638e..e2e9044 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -1,7 +1,7 @@ #ifndef _FS_CEPH_SUPER_H #define _FS_CEPH_SUPER_H -#include "ceph_debug.h" +#include #include #include @@ -14,13 +14,7 @@ #include #include -#include "types.h" -#include "messenger.h" -#include "msgpool.h" -#include "mon_client.h" -#include "mds_client.h" -#include "osd_client.h" -#include "ceph_fs.h" +#include /* f_type in struct statfs */ #define CEPH_SUPER_MAGIC 0x00c36400 @@ -30,42 +24,25 @@ #define CEPH_BLOCK_SHIFT 20 /* 1 MB */ #define CEPH_BLOCK (1 << CEPH_BLOCK_SHIFT) -/* - * Supported features - */ -#define CEPH_FEATURE_SUPPORTED CEPH_FEATURE_NOSRCADDR | CEPH_FEATURE_FLOCK -#define CEPH_FEATURE_REQUIRED CEPH_FEATURE_NOSRCADDR +#define CEPH_MOUNT_OPT_DIRSTAT (1<<4) /* `cat dirname` for stats */ +#define CEPH_MOUNT_OPT_RBYTES (1<<5) /* dir st_bytes = rbytes */ +#define CEPH_MOUNT_OPT_NOASYNCREADDIR (1<<7) /* no dcache readdir */ -/* - * mount options - */ -#define CEPH_OPT_FSID (1<<0) -#define CEPH_OPT_NOSHARE (1<<1) /* don't share client with other sbs */ -#define CEPH_OPT_MYIP (1<<2) /* specified my ip */ -#define CEPH_OPT_DIRSTAT (1<<4) /* funky `cat dirname` for stats */ -#define CEPH_OPT_RBYTES (1<<5) /* dir st_bytes = rbytes */ -#define CEPH_OPT_NOCRC (1<<6) /* no data crc on writes */ -#define CEPH_OPT_NOASYNCREADDIR (1<<7) /* no dcache readdir */ +#define CEPH_MOUNT_OPT_DEFAULT (CEPH_MOUNT_OPT_RBYTES) -#define CEPH_OPT_DEFAULT (CEPH_OPT_RBYTES) +#define ceph_set_mount_opt(fsc, opt) \ + (fsc)->mount_options->flags |= CEPH_MOUNT_OPT_##opt; +#define ceph_test_mount_opt(fsc, opt) \ + (!!((fsc)->mount_options->flags & CEPH_MOUNT_OPT_##opt)) -#define ceph_set_opt(client, opt) \ - (client)->mount_args->flags |= CEPH_OPT_##opt; -#define ceph_test_opt(client, opt) \ - (!!((client)->mount_args->flags & CEPH_OPT_##opt)) +#define CEPH_MAX_READDIR_DEFAULT 1024 +#define CEPH_MAX_READDIR_BYTES_DEFAULT (512*1024) +#define CEPH_SNAPDIRNAME_DEFAULT ".snap" - -struct ceph_mount_args { - int sb_flags; +struct ceph_mount_options { int flags; - struct ceph_fsid fsid; - struct ceph_entity_addr my_addr; - int num_mon; - struct ceph_entity_addr *mon_addr; - int mount_timeout; - int osd_idle_ttl; - int osd_timeout; - int osd_keepalive_timeout; + int sb_flags; + int wsize; int rsize; /* max readahead */ int congestion_kb; /* max writeback in flight */ @@ -73,82 +50,25 @@ struct ceph_mount_args { int cap_release_safety; int max_readdir; /* max readdir result (entires) */ int max_readdir_bytes; /* max readdir result (bytes) */ - char *snapdir_name; /* default ".snap" */ - char *name; - char *secret; -}; - -/* - * defaults - */ -#define CEPH_MOUNT_TIMEOUT_DEFAULT 60 -#define CEPH_OSD_TIMEOUT_DEFAULT 60 /* seconds */ -#define CEPH_OSD_KEEPALIVE_DEFAULT 5 -#define CEPH_OSD_IDLE_TTL_DEFAULT 60 -#define CEPH_MOUNT_RSIZE_DEFAULT (512*1024) /* readahead */ -#define CEPH_MAX_READDIR_DEFAULT 1024 -#define CEPH_MAX_READDIR_BYTES_DEFAULT (512*1024) - -#define CEPH_MSG_MAX_FRONT_LEN (16*1024*1024) -#define CEPH_MSG_MAX_DATA_LEN (16*1024*1024) - -#define CEPH_SNAPDIRNAME_DEFAULT ".snap" -#define CEPH_AUTH_NAME_DEFAULT "guest" -/* - * Delay telling the MDS we no longer want caps, in case we reopen - * the file. Delay a minimum amount of time, even if we send a cap - * message for some other reason. Otherwise, take the oppotunity to - * update the mds to avoid sending another message later. - */ -#define CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT 5 /* cap release delay */ -#define CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT 60 /* cap release delay */ - -#define CEPH_CAP_RELEASE_SAFETY_DEFAULT (CEPH_CAPS_PER_RELEASE * 4) - -/* mount state */ -enum { - CEPH_MOUNT_MOUNTING, - CEPH_MOUNT_MOUNTED, - CEPH_MOUNT_UNMOUNTING, - CEPH_MOUNT_UNMOUNTED, - CEPH_MOUNT_SHUTDOWN, -}; -/* - * subtract jiffies - */ -static inline unsigned long time_sub(unsigned long a, unsigned long b) -{ - BUG_ON(time_after(b, a)); - return (long)a - (long)b; -} - -/* - * per-filesystem client state - * - * possibly shared by multiple mount points, if they are - * mounting the same ceph filesystem/cluster. - */ -struct ceph_client { - struct ceph_fsid fsid; - bool have_fsid; + /* + * everything above this point can be memcmp'd; everything below + * is handled in compare_mount_options() + */ - struct mutex mount_mutex; /* serialize mount attempts */ - struct ceph_mount_args *mount_args; + char *snapdir_name; /* default ".snap" */ +}; +struct ceph_fs_client { struct super_block *sb; - unsigned long mount_state; - wait_queue_head_t auth_wq; - - int auth_err; + struct ceph_mount_options *mount_options; + struct ceph_client *client; + unsigned long mount_state; int min_caps; /* min caps i added */ - struct ceph_messenger *msgr; /* messenger instance */ - struct ceph_mon_client monc; - struct ceph_mds_client mdsc; - struct ceph_osd_client osdc; + struct ceph_mds_client *mdsc; /* writeback */ mempool_t *wb_pagevec_pool; @@ -160,14 +80,14 @@ struct ceph_client { struct backing_dev_info backing_dev_info; #ifdef CONFIG_DEBUG_FS - struct dentry *debugfs_monmap; - struct dentry *debugfs_mdsmap, *debugfs_osdmap; - struct dentry *debugfs_dir, *debugfs_dentry_lru, *debugfs_caps; + struct dentry *debugfs_dentry_lru, *debugfs_caps; struct dentry *debugfs_congestion_kb; struct dentry *debugfs_bdi; + struct dentry *debugfs_mdsc, *debugfs_mdsmap; #endif }; + /* * File i/o capability. This tracks shared state with the metadata * server that allows us to cache or writeback attributes or to read @@ -275,6 +195,20 @@ struct ceph_inode_xattr { int should_free_val; }; +/* + * Ceph dentry state + */ +struct ceph_dentry_info { + struct ceph_mds_session *lease_session; + u32 lease_gen, lease_shared_gen; + u32 lease_seq; + unsigned long lease_renew_after, lease_renew_from; + struct list_head lru; + struct dentry *dentry; + u64 time; + u64 offset; +}; + struct ceph_inode_xattrs_info { /* * (still encoded) xattr blob. we avoid the overhead of parsing @@ -296,11 +230,6 @@ struct ceph_inode_xattrs_info { /* * Ceph inode. */ -#define CEPH_I_COMPLETE 1 /* we have complete directory cached */ -#define CEPH_I_NODELAY 4 /* do not delay cap release */ -#define CEPH_I_FLUSH 8 /* do not delay flush of dirty metadata */ -#define CEPH_I_NOFLUSH 16 /* do not flush dirty caps */ - struct ceph_inode_info { struct ceph_vino i_vino; /* ceph ino + snap */ @@ -391,6 +320,63 @@ static inline struct ceph_inode_info *ceph_inode(struct inode *inode) return container_of(inode, struct ceph_inode_info, vfs_inode); } +static inline struct ceph_vino ceph_vino(struct inode *inode) +{ + return ceph_inode(inode)->i_vino; +} + +/* + * ino_t is <64 bits on many architectures, blech. + * + * don't include snap in ino hash, at least for now. + */ +static inline ino_t ceph_vino_to_ino(struct ceph_vino vino) +{ + ino_t ino = (ino_t)vino.ino; /* ^ (vino.snap << 20); */ +#if BITS_PER_LONG == 32 + ino ^= vino.ino >> (sizeof(u64)-sizeof(ino_t)) * 8; + if (!ino) + ino = 1; +#endif + return ino; +} + +/* for printf-style formatting */ +#define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap + +static inline u64 ceph_ino(struct inode *inode) +{ + return ceph_inode(inode)->i_vino.ino; +} +static inline u64 ceph_snap(struct inode *inode) +{ + return ceph_inode(inode)->i_vino.snap; +} + +static inline int ceph_ino_compare(struct inode *inode, void *data) +{ + struct ceph_vino *pvino = (struct ceph_vino *)data; + struct ceph_inode_info *ci = ceph_inode(inode); + return ci->i_vino.ino == pvino->ino && + ci->i_vino.snap == pvino->snap; +} + +static inline struct inode *ceph_find_inode(struct super_block *sb, + struct ceph_vino vino) +{ + ino_t t = ceph_vino_to_ino(vino); + return ilookup5(sb, t, ceph_ino_compare, &vino); +} + + +/* + * Ceph inode. + */ +#define CEPH_I_COMPLETE 1 /* we have complete directory cached */ +#define CEPH_I_NODELAY 4 /* do not delay cap release */ +#define CEPH_I_FLUSH 8 /* do not delay flush of dirty metadata */ +#define CEPH_I_NOFLUSH 16 /* do not flush dirty caps */ + static inline void ceph_i_clear(struct inode *inode, unsigned mask) { struct ceph_inode_info *ci = ceph_inode(inode); @@ -432,20 +418,6 @@ extern u32 ceph_choose_frag(struct ceph_inode_info *ci, u32 v, struct ceph_inode_frag *pfrag, int *found); -/* - * Ceph dentry state - */ -struct ceph_dentry_info { - struct ceph_mds_session *lease_session; - u32 lease_gen, lease_shared_gen; - u32 lease_seq; - unsigned long lease_renew_after, lease_renew_from; - struct list_head lru; - struct dentry *dentry; - u64 time; - u64 offset; -}; - static inline struct ceph_dentry_info *ceph_dentry(struct dentry *dentry) { return (struct ceph_dentry_info *)dentry->d_fsdata; @@ -456,22 +428,6 @@ static inline loff_t ceph_make_fpos(unsigned frag, unsigned off) return ((loff_t)frag << 32) | (loff_t)off; } -/* - * ino_t is <64 bits on many architectures, blech. - * - * don't include snap in ino hash, at least for now. - */ -static inline ino_t ceph_vino_to_ino(struct ceph_vino vino) -{ - ino_t ino = (ino_t)vino.ino; /* ^ (vino.snap << 20); */ -#if BITS_PER_LONG == 32 - ino ^= vino.ino >> (sizeof(u64)-sizeof(ino_t)) * 8; - if (!ino) - ino = 1; -#endif - return ino; -} - static inline int ceph_set_ino_cb(struct inode *inode, void *data) { ceph_inode(inode)->i_vino = *(struct ceph_vino *)data; @@ -479,39 +435,6 @@ static inline int ceph_set_ino_cb(struct inode *inode, void *data) return 0; } -static inline struct ceph_vino ceph_vino(struct inode *inode) -{ - return ceph_inode(inode)->i_vino; -} - -/* for printf-style formatting */ -#define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap - -static inline u64 ceph_ino(struct inode *inode) -{ - return ceph_inode(inode)->i_vino.ino; -} -static inline u64 ceph_snap(struct inode *inode) -{ - return ceph_inode(inode)->i_vino.snap; -} - -static inline int ceph_ino_compare(struct inode *inode, void *data) -{ - struct ceph_vino *pvino = (struct ceph_vino *)data; - struct ceph_inode_info *ci = ceph_inode(inode); - return ci->i_vino.ino == pvino->ino && - ci->i_vino.snap == pvino->snap; -} - -static inline struct inode *ceph_find_inode(struct super_block *sb, - struct ceph_vino vino) -{ - ino_t t = ceph_vino_to_ino(vino); - return ilookup5(sb, t, ceph_ino_compare, &vino); -} - - /* * caps helpers */ @@ -576,18 +499,18 @@ extern int ceph_reserve_caps(struct ceph_mds_client *mdsc, struct ceph_cap_reservation *ctx, int need); extern int ceph_unreserve_caps(struct ceph_mds_client *mdsc, struct ceph_cap_reservation *ctx); -extern void ceph_reservation_status(struct ceph_client *client, +extern void ceph_reservation_status(struct ceph_fs_client *client, int *total, int *avail, int *used, int *reserved, int *min); -static inline struct ceph_client *ceph_inode_to_client(struct inode *inode) +static inline struct ceph_fs_client *ceph_inode_to_client(struct inode *inode) { - return (struct ceph_client *)inode->i_sb->s_fs_info; + return (struct ceph_fs_client *)inode->i_sb->s_fs_info; } -static inline struct ceph_client *ceph_sb_to_client(struct super_block *sb) +static inline struct ceph_fs_client *ceph_sb_to_client(struct super_block *sb) { - return (struct ceph_client *)sb->s_fs_info; + return (struct ceph_fs_client *)sb->s_fs_info; } @@ -617,51 +540,6 @@ struct ceph_file_info { /* - * snapshots - */ - -/* - * A "snap context" is the set of existing snapshots when we - * write data. It is used by the OSD to guide its COW behavior. - * - * The ceph_snap_context is refcounted, and attached to each dirty - * page, indicating which context the dirty data belonged when it was - * dirtied. - */ -struct ceph_snap_context { - atomic_t nref; - u64 seq; - int num_snaps; - u64 snaps[]; -}; - -static inline struct ceph_snap_context * -ceph_get_snap_context(struct ceph_snap_context *sc) -{ - /* - printk("get_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref), - atomic_read(&sc->nref)+1); - */ - if (sc) - atomic_inc(&sc->nref); - return sc; -} - -static inline void ceph_put_snap_context(struct ceph_snap_context *sc) -{ - if (!sc) - return; - /* - printk("put_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref), - atomic_read(&sc->nref)-1); - */ - if (atomic_dec_and_test(&sc->nref)) { - /*printk(" deleting snap_context %p\n", sc);*/ - kfree(sc); - } -} - -/* * A "snap realm" describes a subset of the file hierarchy sharing * the same set of snapshots that apply to it. The realms themselves * are organized into a hierarchy, such that children inherit (some of) @@ -699,16 +577,33 @@ struct ceph_snap_realm { spinlock_t inodes_with_caps_lock; }; - - -/* - * calculate the number of pages a given length and offset map onto, - * if we align the data. - */ -static inline int calc_pages_for(u64 off, u64 len) +static inline int default_congestion_kb(void) { - return ((off+len+PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT) - - (off >> PAGE_CACHE_SHIFT); + int congestion_kb; + + /* + * Copied from NFS + * + * congestion size, scale with available memory. + * + * 64MB: 8192k + * 128MB: 11585k + * 256MB: 16384k + * 512MB: 23170k + * 1GB: 32768k + * 2GB: 46340k + * 4GB: 65536k + * 8GB: 92681k + * 16GB: 131072k + * + * This allows larger machines to have larger/more transfers. + * Limit the default to 256M + */ + congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10); + if (congestion_kb > 256*1024) + congestion_kb = 256*1024; + + return congestion_kb; } @@ -741,16 +636,6 @@ static inline bool __ceph_have_pending_cap_snap(struct ceph_inode_info *ci) ci_item)->writing; } - -/* super.c */ -extern struct kmem_cache *ceph_inode_cachep; -extern struct kmem_cache *ceph_cap_cachep; -extern struct kmem_cache *ceph_dentry_cachep; -extern struct kmem_cache *ceph_file_cachep; - -extern const char *ceph_msg_type_name(int type); -extern int ceph_check_fsid(struct ceph_client *client, struct ceph_fsid *fsid); - /* inode.c */ extern const struct inode_operations ceph_file_iops; @@ -857,12 +742,18 @@ extern int ceph_mmap(struct file *file, struct vm_area_struct *vma); /* file.c */ extern const struct file_operations ceph_file_fops; extern const struct address_space_operations ceph_aops; +extern int ceph_copy_to_page_vector(struct page **pages, + const char *data, + loff_t off, size_t len); +extern int ceph_copy_from_page_vector(struct page **pages, + char *data, + loff_t off, size_t len); +extern struct page **ceph_alloc_page_vector(int num_pages, gfp_t flags); extern int ceph_open(struct inode *inode, struct file *file); extern struct dentry *ceph_lookup_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd, int mode, int locked_dir); extern int ceph_release(struct inode *inode, struct file *filp); -extern void ceph_release_page_vector(struct page **pages, int num_pages); /* dir.c */ extern const struct file_operations ceph_dir_fops; @@ -892,12 +783,6 @@ extern long ceph_ioctl(struct file *file, unsigned int cmd, unsigned long arg); /* export.c */ extern const struct export_operations ceph_export_ops; -/* debugfs.c */ -extern int ceph_debugfs_init(void); -extern void ceph_debugfs_cleanup(void); -extern int ceph_debugfs_client_init(struct ceph_client *client); -extern void ceph_debugfs_client_cleanup(struct ceph_client *client); - /* locks.c */ extern int ceph_lock(struct file *file, int cmd, struct file_lock *fl); extern int ceph_flock(struct file *file, int cmd, struct file_lock *fl); @@ -914,4 +799,8 @@ static inline struct inode *get_dentry_parent_inode(struct dentry *dentry) return NULL; } +/* debugfs.c */ +extern int ceph_fs_debugfs_init(struct ceph_fs_client *client); +extern void ceph_fs_debugfs_cleanup(struct ceph_fs_client *client); + #endif /* _FS_CEPH_SUPER_H */ diff --git a/fs/ceph/types.h b/fs/ceph/types.h deleted file mode 100644 index 28b35a0..0000000 --- a/fs/ceph/types.h +++ /dev/null @@ -1,29 +0,0 @@ -#ifndef _FS_CEPH_TYPES_H -#define _FS_CEPH_TYPES_H - -/* needed before including ceph_fs.h */ -#include -#include -#include -#include - -#include "ceph_fs.h" -#include "ceph_frag.h" -#include "ceph_hash.h" - -/* - * Identify inodes by both their ino AND snapshot id (a u64). - */ -struct ceph_vino { - u64 ino; - u64 snap; -}; - - -/* context for the caps reservation mechanism */ -struct ceph_cap_reservation { - int count; -}; - - -#endif diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 9578af6..70e9199 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -1,6 +1,9 @@ -#include "ceph_debug.h" +#include + #include "super.h" -#include "decode.h" +#include "mds_client.h" + +#include #include #include @@ -620,12 +623,12 @@ out: static int ceph_sync_setxattr(struct dentry *dentry, const char *name, const char *value, size_t size, int flags) { - struct ceph_client *client = ceph_sb_to_client(dentry->d_sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); struct inode *inode = dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); struct inode *parent_inode = dentry->d_parent->d_inode; struct ceph_mds_request *req; - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_mds_client *mdsc = fsc->mdsc; int err; int i, nr_pages; struct page **pages = NULL; @@ -777,8 +780,8 @@ out: static int ceph_send_removexattr(struct dentry *dentry, const char *name) { - struct ceph_client *client = ceph_sb_to_client(dentry->d_sb); - struct ceph_mds_client *mdsc = &client->mdsc; + struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); + struct ceph_mds_client *mdsc = fsc->mdsc; struct inode *inode = dentry->d_inode; struct inode *parent_inode = dentry->d_parent->d_inode; struct ceph_mds_request *req; diff --git a/include/linux/ceph/auth.h b/include/linux/ceph/auth.h new file mode 100644 index 0000000..7fff521 --- /dev/null +++ b/include/linux/ceph/auth.h @@ -0,0 +1,92 @@ +#ifndef _FS_CEPH_AUTH_H +#define _FS_CEPH_AUTH_H + +#include +#include + +/* + * Abstract interface for communicating with the authenticate module. + * There is some handshake that takes place between us and the monitor + * to acquire the necessary keys. These are used to generate an + * 'authorizer' that we use when connecting to a service (mds, osd). + */ + +struct ceph_auth_client; +struct ceph_authorizer; + +struct ceph_auth_client_ops { + const char *name; + + /* + * true if we are authenticated and can connect to + * services. + */ + int (*is_authenticated)(struct ceph_auth_client *ac); + + /* + * true if we should (re)authenticate, e.g., when our tickets + * are getting old and crusty. + */ + int (*should_authenticate)(struct ceph_auth_client *ac); + + /* + * build requests and process replies during monitor + * handshake. if handle_reply returns -EAGAIN, we build + * another request. + */ + int (*build_request)(struct ceph_auth_client *ac, void *buf, void *end); + int (*handle_reply)(struct ceph_auth_client *ac, int result, + void *buf, void *end); + + /* + * Create authorizer for connecting to a service, and verify + * the response to authenticate the service. + */ + int (*create_authorizer)(struct ceph_auth_client *ac, int peer_type, + struct ceph_authorizer **a, + void **buf, size_t *len, + void **reply_buf, size_t *reply_len); + int (*verify_authorizer_reply)(struct ceph_auth_client *ac, + struct ceph_authorizer *a, size_t len); + void (*destroy_authorizer)(struct ceph_auth_client *ac, + struct ceph_authorizer *a); + void (*invalidate_authorizer)(struct ceph_auth_client *ac, + int peer_type); + + /* reset when we (re)connect to a monitor */ + void (*reset)(struct ceph_auth_client *ac); + + void (*destroy)(struct ceph_auth_client *ac); +}; + +struct ceph_auth_client { + u32 protocol; /* CEPH_AUTH_* */ + void *private; /* for use by protocol implementation */ + const struct ceph_auth_client_ops *ops; /* null iff protocol==0 */ + + bool negotiating; /* true if negotiating protocol */ + const char *name; /* entity name */ + u64 global_id; /* our unique id in system */ + const char *secret; /* our secret key */ + unsigned want_keys; /* which services we want */ +}; + +extern struct ceph_auth_client *ceph_auth_init(const char *name, + const char *secret); +extern void ceph_auth_destroy(struct ceph_auth_client *ac); + +extern void ceph_auth_reset(struct ceph_auth_client *ac); + +extern int ceph_auth_build_hello(struct ceph_auth_client *ac, + void *buf, size_t len); +extern int ceph_handle_auth_reply(struct ceph_auth_client *ac, + void *buf, size_t len, + void *reply_buf, size_t reply_len); +extern int ceph_entity_name_encode(const char *name, void **p, void *end); + +extern int ceph_build_auth(struct ceph_auth_client *ac, + void *msg_buf, size_t msg_len); + +extern int ceph_auth_is_authenticated(struct ceph_auth_client *ac); + +#endif diff --git a/include/linux/ceph/buffer.h b/include/linux/ceph/buffer.h new file mode 100644 index 0000000..58d1901 --- /dev/null +++ b/include/linux/ceph/buffer.h @@ -0,0 +1,39 @@ +#ifndef __FS_CEPH_BUFFER_H +#define __FS_CEPH_BUFFER_H + +#include +#include +#include +#include +#include + +/* + * a simple reference counted buffer. + * + * use kmalloc for small sizes (<= one page), vmalloc for larger + * sizes. + */ +struct ceph_buffer { + struct kref kref; + struct kvec vec; + size_t alloc_len; + bool is_vmalloc; +}; + +extern struct ceph_buffer *ceph_buffer_new(size_t len, gfp_t gfp); +extern void ceph_buffer_release(struct kref *kref); + +static inline struct ceph_buffer *ceph_buffer_get(struct ceph_buffer *b) +{ + kref_get(&b->kref); + return b; +} + +static inline void ceph_buffer_put(struct ceph_buffer *b) +{ + kref_put(&b->kref, ceph_buffer_release); +} + +extern int ceph_decode_buffer(struct ceph_buffer **b, void **p, void *end); + +#endif diff --git a/include/linux/ceph/ceph_debug.h b/include/linux/ceph/ceph_debug.h new file mode 100644 index 0000000..aa2e191 --- /dev/null +++ b/include/linux/ceph/ceph_debug.h @@ -0,0 +1,38 @@ +#ifndef _FS_CEPH_DEBUG_H +#define _FS_CEPH_DEBUG_H + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#ifdef CONFIG_CEPH_LIB_PRETTYDEBUG + +/* + * wrap pr_debug to include a filename:lineno prefix on each line. + * this incurs some overhead (kernel size and execution time) due to + * the extra function call at each call site. + */ + +# if defined(DEBUG) || defined(CONFIG_DYNAMIC_DEBUG) +extern const char *ceph_file_part(const char *s, int len); +# define dout(fmt, ...) \ + pr_debug("%.*s %12.12s:%-4d : " fmt, \ + 8 - (int)sizeof(KBUILD_MODNAME), " ", \ + ceph_file_part(__FILE__, sizeof(__FILE__)), \ + __LINE__, ##__VA_ARGS__) +# else +/* faux printk call just to see any compiler warnings. */ +# define dout(fmt, ...) do { \ + if (0) \ + printk(KERN_DEBUG fmt, ##__VA_ARGS__); \ + } while (0) +# endif + +#else + +/* + * or, just wrap pr_debug + */ +# define dout(fmt, ...) pr_debug(" " fmt, ##__VA_ARGS__) + +#endif + +#endif diff --git a/include/linux/ceph/ceph_frag.h b/include/linux/ceph/ceph_frag.h new file mode 100644 index 0000000..5babb8e --- /dev/null +++ b/include/linux/ceph/ceph_frag.h @@ -0,0 +1,109 @@ +#ifndef FS_CEPH_FRAG_H +#define FS_CEPH_FRAG_H + +/* + * "Frags" are a way to describe a subset of a 32-bit number space, + * using a mask and a value to match against that mask. Any given frag + * (subset of the number space) can be partitioned into 2^n sub-frags. + * + * Frags are encoded into a 32-bit word: + * 8 upper bits = "bits" + * 24 lower bits = "value" + * (We could go to 5+27 bits, but who cares.) + * + * We use the _most_ significant bits of the 24 bit value. This makes + * values logically sort. + * + * Unfortunately, because the "bits" field is still in the high bits, we + * can't sort encoded frags numerically. However, it does allow you + * to feed encoded frags as values into frag_contains_value. + */ +static inline __u32 ceph_frag_make(__u32 b, __u32 v) +{ + return (b << 24) | + (v & (0xffffffu << (24-b)) & 0xffffffu); +} +static inline __u32 ceph_frag_bits(__u32 f) +{ + return f >> 24; +} +static inline __u32 ceph_frag_value(__u32 f) +{ + return f & 0xffffffu; +} +static inline __u32 ceph_frag_mask(__u32 f) +{ + return (0xffffffu << (24-ceph_frag_bits(f))) & 0xffffffu; +} +static inline __u32 ceph_frag_mask_shift(__u32 f) +{ + return 24 - ceph_frag_bits(f); +} + +static inline int ceph_frag_contains_value(__u32 f, __u32 v) +{ + return (v & ceph_frag_mask(f)) == ceph_frag_value(f); +} +static inline int ceph_frag_contains_frag(__u32 f, __u32 sub) +{ + /* is sub as specific as us, and contained by us? */ + return ceph_frag_bits(sub) >= ceph_frag_bits(f) && + (ceph_frag_value(sub) & ceph_frag_mask(f)) == ceph_frag_value(f); +} + +static inline __u32 ceph_frag_parent(__u32 f) +{ + return ceph_frag_make(ceph_frag_bits(f) - 1, + ceph_frag_value(f) & (ceph_frag_mask(f) << 1)); +} +static inline int ceph_frag_is_left_child(__u32 f) +{ + return ceph_frag_bits(f) > 0 && + (ceph_frag_value(f) & (0x1000000 >> ceph_frag_bits(f))) == 0; +} +static inline int ceph_frag_is_right_child(__u32 f) +{ + return ceph_frag_bits(f) > 0 && + (ceph_frag_value(f) & (0x1000000 >> ceph_frag_bits(f))) == 1; +} +static inline __u32 ceph_frag_sibling(__u32 f) +{ + return ceph_frag_make(ceph_frag_bits(f), + ceph_frag_value(f) ^ (0x1000000 >> ceph_frag_bits(f))); +} +static inline __u32 ceph_frag_left_child(__u32 f) +{ + return ceph_frag_make(ceph_frag_bits(f)+1, ceph_frag_value(f)); +} +static inline __u32 ceph_frag_right_child(__u32 f) +{ + return ceph_frag_make(ceph_frag_bits(f)+1, + ceph_frag_value(f) | (0x1000000 >> (1+ceph_frag_bits(f)))); +} +static inline __u32 ceph_frag_make_child(__u32 f, int by, int i) +{ + int newbits = ceph_frag_bits(f) + by; + return ceph_frag_make(newbits, + ceph_frag_value(f) | (i << (24 - newbits))); +} +static inline int ceph_frag_is_leftmost(__u32 f) +{ + return ceph_frag_value(f) == 0; +} +static inline int ceph_frag_is_rightmost(__u32 f) +{ + return ceph_frag_value(f) == ceph_frag_mask(f); +} +static inline __u32 ceph_frag_next(__u32 f) +{ + return ceph_frag_make(ceph_frag_bits(f), + ceph_frag_value(f) + (0x1000000 >> ceph_frag_bits(f))); +} + +/* + * comparator to sort frags logically, as when traversing the + * number space in ascending order... + */ +int ceph_frag_compare(__u32 a, __u32 b); + +#endif diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h new file mode 100644 index 0000000..d5619ac --- /dev/null +++ b/include/linux/ceph/ceph_fs.h @@ -0,0 +1,728 @@ +/* + * ceph_fs.h - Ceph constants and data types to share between kernel and + * user space. + * + * Most types in this file are defined as little-endian, and are + * primarily intended to describe data structures that pass over the + * wire or that are stored on disk. + * + * LGPL2 + */ + +#ifndef CEPH_FS_H +#define CEPH_FS_H + +#include "msgr.h" +#include "rados.h" + +/* + * subprotocol versions. when specific messages types or high-level + * protocols change, bump the affected components. we keep rev + * internal cluster protocols separately from the public, + * client-facing protocol. + */ +#define CEPH_OSD_PROTOCOL 8 /* cluster internal */ +#define CEPH_MDS_PROTOCOL 12 /* cluster internal */ +#define CEPH_MON_PROTOCOL 5 /* cluster internal */ +#define CEPH_OSDC_PROTOCOL 24 /* server/client */ +#define CEPH_MDSC_PROTOCOL 32 /* server/client */ +#define CEPH_MONC_PROTOCOL 15 /* server/client */ + + +#define CEPH_INO_ROOT 1 +#define CEPH_INO_CEPH 2 /* hidden .ceph dir */ + +/* arbitrary limit on max # of monitors (cluster of 3 is typical) */ +#define CEPH_MAX_MON 31 + + +/* + * feature bits + */ +#define CEPH_FEATURE_UID (1<<0) +#define CEPH_FEATURE_NOSRCADDR (1<<1) +#define CEPH_FEATURE_MONCLOCKCHECK (1<<2) +#define CEPH_FEATURE_FLOCK (1<<3) + + +/* + * ceph_file_layout - describe data layout for a file/inode + */ +struct ceph_file_layout { + /* file -> object mapping */ + __le32 fl_stripe_unit; /* stripe unit, in bytes. must be multiple + of page size. */ + __le32 fl_stripe_count; /* over this many objects */ + __le32 fl_object_size; /* until objects are this big, then move to + new objects */ + __le32 fl_cas_hash; /* 0 = none; 1 = sha256 */ + + /* pg -> disk layout */ + __le32 fl_object_stripe_unit; /* for per-object parity, if any */ + + /* object -> pg layout */ + __le32 fl_pg_preferred; /* preferred primary for pg (-1 for none) */ + __le32 fl_pg_pool; /* namespace, crush ruleset, rep level */ +} __attribute__ ((packed)); + +#define CEPH_MIN_STRIPE_UNIT 65536 + +int ceph_file_layout_is_valid(const struct ceph_file_layout *layout); + + +/* crypto algorithms */ +#define CEPH_CRYPTO_NONE 0x0 +#define CEPH_CRYPTO_AES 0x1 + +#define CEPH_AES_IV "cephsageyudagreg" + +/* security/authentication protocols */ +#define CEPH_AUTH_UNKNOWN 0x0 +#define CEPH_AUTH_NONE 0x1 +#define CEPH_AUTH_CEPHX 0x2 + +#define CEPH_AUTH_UID_DEFAULT ((__u64) -1) + + +/********************************************* + * message layer + */ + +/* + * message types + */ + +/* misc */ +#define CEPH_MSG_SHUTDOWN 1 +#define CEPH_MSG_PING 2 + +/* client <-> monitor */ +#define CEPH_MSG_MON_MAP 4 +#define CEPH_MSG_MON_GET_MAP 5 +#define CEPH_MSG_STATFS 13 +#define CEPH_MSG_STATFS_REPLY 14 +#define CEPH_MSG_MON_SUBSCRIBE 15 +#define CEPH_MSG_MON_SUBSCRIBE_ACK 16 +#define CEPH_MSG_AUTH 17 +#define CEPH_MSG_AUTH_REPLY 18 + +/* client <-> mds */ +#define CEPH_MSG_MDS_MAP 21 + +#define CEPH_MSG_CLIENT_SESSION 22 +#define CEPH_MSG_CLIENT_RECONNECT 23 + +#define CEPH_MSG_CLIENT_REQUEST 24 +#define CEPH_MSG_CLIENT_REQUEST_FORWARD 25 +#define CEPH_MSG_CLIENT_REPLY 26 +#define CEPH_MSG_CLIENT_CAPS 0x310 +#define CEPH_MSG_CLIENT_LEASE 0x311 +#define CEPH_MSG_CLIENT_SNAP 0x312 +#define CEPH_MSG_CLIENT_CAPRELEASE 0x313 + +/* pool ops */ +#define CEPH_MSG_POOLOP_REPLY 48 +#define CEPH_MSG_POOLOP 49 + + +/* osd */ +#define CEPH_MSG_OSD_MAP 41 +#define CEPH_MSG_OSD_OP 42 +#define CEPH_MSG_OSD_OPREPLY 43 + +/* pool operations */ +enum { + POOL_OP_CREATE = 0x01, + POOL_OP_DELETE = 0x02, + POOL_OP_AUID_CHANGE = 0x03, + POOL_OP_CREATE_SNAP = 0x11, + POOL_OP_DELETE_SNAP = 0x12, + POOL_OP_CREATE_UNMANAGED_SNAP = 0x21, + POOL_OP_DELETE_UNMANAGED_SNAP = 0x22, +}; + +struct ceph_mon_request_header { + __le64 have_version; + __le16 session_mon; + __le64 session_mon_tid; +} __attribute__ ((packed)); + +struct ceph_mon_statfs { + struct ceph_mon_request_header monhdr; + struct ceph_fsid fsid; +} __attribute__ ((packed)); + +struct ceph_statfs { + __le64 kb, kb_used, kb_avail; + __le64 num_objects; +} __attribute__ ((packed)); + +struct ceph_mon_statfs_reply { + struct ceph_fsid fsid; + __le64 version; + struct ceph_statfs st; +} __attribute__ ((packed)); + +const char *ceph_pool_op_name(int op); + +struct ceph_mon_poolop { + struct ceph_mon_request_header monhdr; + struct ceph_fsid fsid; + __le32 pool; + __le32 op; + __le64 auid; + __le64 snapid; + __le32 name_len; +} __attribute__ ((packed)); + +struct ceph_mon_poolop_reply { + struct ceph_mon_request_header monhdr; + struct ceph_fsid fsid; + __le32 reply_code; + __le32 epoch; + char has_data; + char data[0]; +} __attribute__ ((packed)); + +struct ceph_mon_unmanaged_snap { + __le64 snapid; +} __attribute__ ((packed)); + +struct ceph_osd_getmap { + struct ceph_mon_request_header monhdr; + struct ceph_fsid fsid; + __le32 start; +} __attribute__ ((packed)); + +struct ceph_mds_getmap { + struct ceph_mon_request_header monhdr; + struct ceph_fsid fsid; +} __attribute__ ((packed)); + +struct ceph_client_mount { + struct ceph_mon_request_header monhdr; +} __attribute__ ((packed)); + +struct ceph_mon_subscribe_item { + __le64 have_version; __le64 have; + __u8 onetime; +} __attribute__ ((packed)); + +struct ceph_mon_subscribe_ack { + __le32 duration; /* seconds */ + struct ceph_fsid fsid; +} __attribute__ ((packed)); + +/* + * mds states + * > 0 -> in + * <= 0 -> out + */ +#define CEPH_MDS_STATE_DNE 0 /* down, does not exist. */ +#define CEPH_MDS_STATE_STOPPED -1 /* down, once existed, but no subtrees. + empty log. */ +#define CEPH_MDS_STATE_BOOT -4 /* up, boot announcement. */ +#define CEPH_MDS_STATE_STANDBY -5 /* up, idle. waiting for assignment. */ +#define CEPH_MDS_STATE_CREATING -6 /* up, creating MDS instance. */ +#define CEPH_MDS_STATE_STARTING -7 /* up, starting previously stopped mds */ +#define CEPH_MDS_STATE_STANDBY_REPLAY -8 /* up, tailing active node's journal */ + +#define CEPH_MDS_STATE_REPLAY 8 /* up, replaying journal. */ +#define CEPH_MDS_STATE_RESOLVE 9 /* up, disambiguating distributed + operations (import, rename, etc.) */ +#define CEPH_MDS_STATE_RECONNECT 10 /* up, reconnect to clients */ +#define CEPH_MDS_STATE_REJOIN 11 /* up, rejoining distributed cache */ +#define CEPH_MDS_STATE_CLIENTREPLAY 12 /* up, replaying client operations */ +#define CEPH_MDS_STATE_ACTIVE 13 /* up, active */ +#define CEPH_MDS_STATE_STOPPING 14 /* up, but exporting metadata */ + +extern const char *ceph_mds_state_name(int s); + + +/* + * metadata lock types. + * - these are bitmasks.. we can compose them + * - they also define the lock ordering by the MDS + * - a few of these are internal to the mds + */ +#define CEPH_LOCK_DVERSION 1 +#define CEPH_LOCK_DN 2 +#define CEPH_LOCK_ISNAP 16 +#define CEPH_LOCK_IVERSION 32 /* mds internal */ +#define CEPH_LOCK_IFILE 64 +#define CEPH_LOCK_IAUTH 128 +#define CEPH_LOCK_ILINK 256 +#define CEPH_LOCK_IDFT 512 /* dir frag tree */ +#define CEPH_LOCK_INEST 1024 /* mds internal */ +#define CEPH_LOCK_IXATTR 2048 +#define CEPH_LOCK_IFLOCK 4096 /* advisory file locks */ +#define CEPH_LOCK_INO 8192 /* immutable inode bits; not a lock */ + +/* client_session ops */ +enum { + CEPH_SESSION_REQUEST_OPEN, + CEPH_SESSION_OPEN, + CEPH_SESSION_REQUEST_CLOSE, + CEPH_SESSION_CLOSE, + CEPH_SESSION_REQUEST_RENEWCAPS, + CEPH_SESSION_RENEWCAPS, + CEPH_SESSION_STALE, + CEPH_SESSION_RECALL_STATE, +}; + +extern const char *ceph_session_op_name(int op); + +struct ceph_mds_session_head { + __le32 op; + __le64 seq; + struct ceph_timespec stamp; + __le32 max_caps, max_leases; +} __attribute__ ((packed)); + +/* client_request */ +/* + * metadata ops. + * & 0x001000 -> write op + * & 0x010000 -> follow symlink (e.g. stat(), not lstat()). + & & 0x100000 -> use weird ino/path trace + */ +#define CEPH_MDS_OP_WRITE 0x001000 +enum { + CEPH_MDS_OP_LOOKUP = 0x00100, + CEPH_MDS_OP_GETATTR = 0x00101, + CEPH_MDS_OP_LOOKUPHASH = 0x00102, + CEPH_MDS_OP_LOOKUPPARENT = 0x00103, + + CEPH_MDS_OP_SETXATTR = 0x01105, + CEPH_MDS_OP_RMXATTR = 0x01106, + CEPH_MDS_OP_SETLAYOUT = 0x01107, + CEPH_MDS_OP_SETATTR = 0x01108, + CEPH_MDS_OP_SETFILELOCK= 0x01109, + CEPH_MDS_OP_GETFILELOCK= 0x00110, + + CEPH_MDS_OP_MKNOD = 0x01201, + CEPH_MDS_OP_LINK = 0x01202, + CEPH_MDS_OP_UNLINK = 0x01203, + CEPH_MDS_OP_RENAME = 0x01204, + CEPH_MDS_OP_MKDIR = 0x01220, + CEPH_MDS_OP_RMDIR = 0x01221, + CEPH_MDS_OP_SYMLINK = 0x01222, + + CEPH_MDS_OP_CREATE = 0x01301, + CEPH_MDS_OP_OPEN = 0x00302, + CEPH_MDS_OP_READDIR = 0x00305, + + CEPH_MDS_OP_LOOKUPSNAP = 0x00400, + CEPH_MDS_OP_MKSNAP = 0x01400, + CEPH_MDS_OP_RMSNAP = 0x01401, + CEPH_MDS_OP_LSSNAP = 0x00402, +}; + +extern const char *ceph_mds_op_name(int op); + + +#define CEPH_SETATTR_MODE 1 +#define CEPH_SETATTR_UID 2 +#define CEPH_SETATTR_GID 4 +#define CEPH_SETATTR_MTIME 8 +#define CEPH_SETATTR_ATIME 16 +#define CEPH_SETATTR_SIZE 32 +#define CEPH_SETATTR_CTIME 64 + +union ceph_mds_request_args { + struct { + __le32 mask; /* CEPH_CAP_* */ + } __attribute__ ((packed)) getattr; + struct { + __le32 mode; + __le32 uid; + __le32 gid; + struct ceph_timespec mtime; + struct ceph_timespec atime; + __le64 size, old_size; /* old_size needed by truncate */ + __le32 mask; /* CEPH_SETATTR_* */ + } __attribute__ ((packed)) setattr; + struct { + __le32 frag; /* which dir fragment */ + __le32 max_entries; /* how many dentries to grab */ + __le32 max_bytes; + } __attribute__ ((packed)) readdir; + struct { + __le32 mode; + __le32 rdev; + } __attribute__ ((packed)) mknod; + struct { + __le32 mode; + } __attribute__ ((packed)) mkdir; + struct { + __le32 flags; + __le32 mode; + __le32 stripe_unit; /* layout for newly created file */ + __le32 stripe_count; /* ... */ + __le32 object_size; + __le32 file_replication; + __le32 preferred; + } __attribute__ ((packed)) open; + struct { + __le32 flags; + } __attribute__ ((packed)) setxattr; + struct { + struct ceph_file_layout layout; + } __attribute__ ((packed)) setlayout; + struct { + __u8 rule; /* currently fcntl or flock */ + __u8 type; /* shared, exclusive, remove*/ + __le64 pid; /* process id requesting the lock */ + __le64 pid_namespace; + __le64 start; /* initial location to lock */ + __le64 length; /* num bytes to lock from start */ + __u8 wait; /* will caller wait for lock to become available? */ + } __attribute__ ((packed)) filelock_change; +} __attribute__ ((packed)); + +#define CEPH_MDS_FLAG_REPLAY 1 /* this is a replayed op */ +#define CEPH_MDS_FLAG_WANT_DENTRY 2 /* want dentry in reply */ + +struct ceph_mds_request_head { + __le64 oldest_client_tid; + __le32 mdsmap_epoch; /* on client */ + __le32 flags; /* CEPH_MDS_FLAG_* */ + __u8 num_retry, num_fwd; /* count retry, fwd attempts */ + __le16 num_releases; /* # include cap/lease release records */ + __le32 op; /* mds op code */ + __le32 caller_uid, caller_gid; + __le64 ino; /* use this ino for openc, mkdir, mknod, + etc. (if replaying) */ + union ceph_mds_request_args args; +} __attribute__ ((packed)); + +/* cap/lease release record */ +struct ceph_mds_request_release { + __le64 ino, cap_id; /* ino and unique cap id */ + __le32 caps, wanted; /* new issued, wanted */ + __le32 seq, issue_seq, mseq; + __le32 dname_seq; /* if releasing a dentry lease, a */ + __le32 dname_len; /* string follows. */ +} __attribute__ ((packed)); + +/* client reply */ +struct ceph_mds_reply_head { + __le32 op; + __le32 result; + __le32 mdsmap_epoch; + __u8 safe; /* true if committed to disk */ + __u8 is_dentry, is_target; /* true if dentry, target inode records + are included with reply */ +} __attribute__ ((packed)); + +/* one for each node split */ +struct ceph_frag_tree_split { + __le32 frag; /* this frag splits... */ + __le32 by; /* ...by this many bits */ +} __attribute__ ((packed)); + +struct ceph_frag_tree_head { + __le32 nsplits; /* num ceph_frag_tree_split records */ + struct ceph_frag_tree_split splits[]; +} __attribute__ ((packed)); + +/* capability issue, for bundling with mds reply */ +struct ceph_mds_reply_cap { + __le32 caps, wanted; /* caps issued, wanted */ + __le64 cap_id; + __le32 seq, mseq; + __le64 realm; /* snap realm */ + __u8 flags; /* CEPH_CAP_FLAG_* */ +} __attribute__ ((packed)); + +#define CEPH_CAP_FLAG_AUTH 1 /* cap is issued by auth mds */ + +/* inode record, for bundling with mds reply */ +struct ceph_mds_reply_inode { + __le64 ino; + __le64 snapid; + __le32 rdev; + __le64 version; /* inode version */ + __le64 xattr_version; /* version for xattr blob */ + struct ceph_mds_reply_cap cap; /* caps issued for this inode */ + struct ceph_file_layout layout; + struct ceph_timespec ctime, mtime, atime; + __le32 time_warp_seq; + __le64 size, max_size, truncate_size; + __le32 truncate_seq; + __le32 mode, uid, gid; + __le32 nlink; + __le64 files, subdirs, rbytes, rfiles, rsubdirs; /* dir stats */ + struct ceph_timespec rctime; + struct ceph_frag_tree_head fragtree; /* (must be at end of struct) */ +} __attribute__ ((packed)); +/* followed by frag array, then symlink string, then xattr blob */ + +/* reply_lease follows dname, and reply_inode */ +struct ceph_mds_reply_lease { + __le16 mask; /* lease type(s) */ + __le32 duration_ms; /* lease duration */ + __le32 seq; +} __attribute__ ((packed)); + +struct ceph_mds_reply_dirfrag { + __le32 frag; /* fragment */ + __le32 auth; /* auth mds, if this is a delegation point */ + __le32 ndist; /* number of mds' this is replicated on */ + __le32 dist[]; +} __attribute__ ((packed)); + +#define CEPH_LOCK_FCNTL 1 +#define CEPH_LOCK_FLOCK 2 + +#define CEPH_LOCK_SHARED 1 +#define CEPH_LOCK_EXCL 2 +#define CEPH_LOCK_UNLOCK 4 + +struct ceph_filelock { + __le64 start;/* file offset to start lock at */ + __le64 length; /* num bytes to lock; 0 for all following start */ + __le64 client; /* which client holds the lock */ + __le64 pid; /* process id holding the lock on the client */ + __le64 pid_namespace; + __u8 type; /* shared lock, exclusive lock, or unlock */ +} __attribute__ ((packed)); + + +/* file access modes */ +#define CEPH_FILE_MODE_PIN 0 +#define CEPH_FILE_MODE_RD 1 +#define CEPH_FILE_MODE_WR 2 +#define CEPH_FILE_MODE_RDWR 3 /* RD | WR */ +#define CEPH_FILE_MODE_LAZY 4 /* lazy io */ +#define CEPH_FILE_MODE_NUM 8 /* bc these are bit fields.. mostly */ + +int ceph_flags_to_mode(int flags); + + +/* capability bits */ +#define CEPH_CAP_PIN 1 /* no specific capabilities beyond the pin */ + +/* generic cap bits */ +#define CEPH_CAP_GSHARED 1 /* client can reads */ +#define CEPH_CAP_GEXCL 2 /* client can read and update */ +#define CEPH_CAP_GCACHE 4 /* (file) client can cache reads */ +#define CEPH_CAP_GRD 8 /* (file) client can read */ +#define CEPH_CAP_GWR 16 /* (file) client can write */ +#define CEPH_CAP_GBUFFER 32 /* (file) client can buffer writes */ +#define CEPH_CAP_GWREXTEND 64 /* (file) client can extend EOF */ +#define CEPH_CAP_GLAZYIO 128 /* (file) client can perform lazy io */ + +/* per-lock shift */ +#define CEPH_CAP_SAUTH 2 +#define CEPH_CAP_SLINK 4 +#define CEPH_CAP_SXATTR 6 +#define CEPH_CAP_SFILE 8 +#define CEPH_CAP_SFLOCK 20 + +#define CEPH_CAP_BITS 22 + +/* composed values */ +#define CEPH_CAP_AUTH_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SAUTH) +#define CEPH_CAP_AUTH_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SAUTH) +#define CEPH_CAP_LINK_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SLINK) +#define CEPH_CAP_LINK_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SLINK) +#define CEPH_CAP_XATTR_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SXATTR) +#define CEPH_CAP_XATTR_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SXATTR) +#define CEPH_CAP_FILE(x) (x << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_CACHE (CEPH_CAP_GCACHE << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_RD (CEPH_CAP_GRD << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_WR (CEPH_CAP_GWR << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_BUFFER (CEPH_CAP_GBUFFER << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_WREXTEND (CEPH_CAP_GWREXTEND << CEPH_CAP_SFILE) +#define CEPH_CAP_FILE_LAZYIO (CEPH_CAP_GLAZYIO << CEPH_CAP_SFILE) +#define CEPH_CAP_FLOCK_SHARED (CEPH_CAP_GSHARED << CEPH_CAP_SFLOCK) +#define CEPH_CAP_FLOCK_EXCL (CEPH_CAP_GEXCL << CEPH_CAP_SFLOCK) + + +/* cap masks (for getattr) */ +#define CEPH_STAT_CAP_INODE CEPH_CAP_PIN +#define CEPH_STAT_CAP_TYPE CEPH_CAP_PIN /* mode >> 12 */ +#define CEPH_STAT_CAP_SYMLINK CEPH_CAP_PIN +#define CEPH_STAT_CAP_UID CEPH_CAP_AUTH_SHARED +#define CEPH_STAT_CAP_GID CEPH_CAP_AUTH_SHARED +#define CEPH_STAT_CAP_MODE CEPH_CAP_AUTH_SHARED +#define CEPH_STAT_CAP_NLINK CEPH_CAP_LINK_SHARED +#define CEPH_STAT_CAP_LAYOUT CEPH_CAP_FILE_SHARED +#define CEPH_STAT_CAP_MTIME CEPH_CAP_FILE_SHARED +#define CEPH_STAT_CAP_SIZE CEPH_CAP_FILE_SHARED +#define CEPH_STAT_CAP_ATIME CEPH_CAP_FILE_SHARED /* fixme */ +#define CEPH_STAT_CAP_XATTR CEPH_CAP_XATTR_SHARED +#define CEPH_STAT_CAP_INODE_ALL (CEPH_CAP_PIN | \ + CEPH_CAP_AUTH_SHARED | \ + CEPH_CAP_LINK_SHARED | \ + CEPH_CAP_FILE_SHARED | \ + CEPH_CAP_XATTR_SHARED) + +#define CEPH_CAP_ANY_SHARED (CEPH_CAP_AUTH_SHARED | \ + CEPH_CAP_LINK_SHARED | \ + CEPH_CAP_XATTR_SHARED | \ + CEPH_CAP_FILE_SHARED) +#define CEPH_CAP_ANY_RD (CEPH_CAP_ANY_SHARED | CEPH_CAP_FILE_RD | \ + CEPH_CAP_FILE_CACHE) + +#define CEPH_CAP_ANY_EXCL (CEPH_CAP_AUTH_EXCL | \ + CEPH_CAP_LINK_EXCL | \ + CEPH_CAP_XATTR_EXCL | \ + CEPH_CAP_FILE_EXCL) +#define CEPH_CAP_ANY_FILE_WR (CEPH_CAP_FILE_WR | CEPH_CAP_FILE_BUFFER | \ + CEPH_CAP_FILE_EXCL) +#define CEPH_CAP_ANY_WR (CEPH_CAP_ANY_EXCL | CEPH_CAP_ANY_FILE_WR) +#define CEPH_CAP_ANY (CEPH_CAP_ANY_RD | CEPH_CAP_ANY_EXCL | \ + CEPH_CAP_ANY_FILE_WR | CEPH_CAP_FILE_LAZYIO | \ + CEPH_CAP_PIN) + +#define CEPH_CAP_LOCKS (CEPH_LOCK_IFILE | CEPH_LOCK_IAUTH | CEPH_LOCK_ILINK | \ + CEPH_LOCK_IXATTR) + +int ceph_caps_for_mode(int mode); + +enum { + CEPH_CAP_OP_GRANT, /* mds->client grant */ + CEPH_CAP_OP_REVOKE, /* mds->client revoke */ + CEPH_CAP_OP_TRUNC, /* mds->client trunc notify */ + CEPH_CAP_OP_EXPORT, /* mds has exported the cap */ + CEPH_CAP_OP_IMPORT, /* mds has imported the cap */ + CEPH_CAP_OP_UPDATE, /* client->mds update */ + CEPH_CAP_OP_DROP, /* client->mds drop cap bits */ + CEPH_CAP_OP_FLUSH, /* client->mds cap writeback */ + CEPH_CAP_OP_FLUSH_ACK, /* mds->client flushed */ + CEPH_CAP_OP_FLUSHSNAP, /* client->mds flush snapped metadata */ + CEPH_CAP_OP_FLUSHSNAP_ACK, /* mds->client flushed snapped metadata */ + CEPH_CAP_OP_RELEASE, /* client->mds release (clean) cap */ + CEPH_CAP_OP_RENEW, /* client->mds renewal request */ +}; + +extern const char *ceph_cap_op_name(int op); + +/* + * caps message, used for capability callbacks, acks, requests, etc. + */ +struct ceph_mds_caps { + __le32 op; /* CEPH_CAP_OP_* */ + __le64 ino, realm; + __le64 cap_id; + __le32 seq, issue_seq; + __le32 caps, wanted, dirty; /* latest issued/wanted/dirty */ + __le32 migrate_seq; + __le64 snap_follows; + __le32 snap_trace_len; + + /* authlock */ + __le32 uid, gid, mode; + + /* linklock */ + __le32 nlink; + + /* xattrlock */ + __le32 xattr_len; + __le64 xattr_version; + + /* filelock */ + __le64 size, max_size, truncate_size; + __le32 truncate_seq; + struct ceph_timespec mtime, atime, ctime; + struct ceph_file_layout layout; + __le32 time_warp_seq; +} __attribute__ ((packed)); + +/* cap release msg head */ +struct ceph_mds_cap_release { + __le32 num; /* number of cap_items that follow */ +} __attribute__ ((packed)); + +struct ceph_mds_cap_item { + __le64 ino; + __le64 cap_id; + __le32 migrate_seq, seq; +} __attribute__ ((packed)); + +#define CEPH_MDS_LEASE_REVOKE 1 /* mds -> client */ +#define CEPH_MDS_LEASE_RELEASE 2 /* client -> mds */ +#define CEPH_MDS_LEASE_RENEW 3 /* client <-> mds */ +#define CEPH_MDS_LEASE_REVOKE_ACK 4 /* client -> mds */ + +extern const char *ceph_lease_op_name(int o); + +/* lease msg header */ +struct ceph_mds_lease { + __u8 action; /* CEPH_MDS_LEASE_* */ + __le16 mask; /* which lease */ + __le64 ino; + __le64 first, last; /* snap range */ + __le32 seq; + __le32 duration_ms; /* duration of renewal */ +} __attribute__ ((packed)); +/* followed by a __le32+string for dname */ + +/* client reconnect */ +struct ceph_mds_cap_reconnect { + __le64 cap_id; + __le32 wanted; + __le32 issued; + __le64 snaprealm; + __le64 pathbase; /* base ino for our path to this ino */ + __le32 flock_len; /* size of flock state blob, if any */ +} __attribute__ ((packed)); +/* followed by flock blob */ + +struct ceph_mds_cap_reconnect_v1 { + __le64 cap_id; + __le32 wanted; + __le32 issued; + __le64 size; + struct ceph_timespec mtime, atime; + __le64 snaprealm; + __le64 pathbase; /* base ino for our path to this ino */ +} __attribute__ ((packed)); + +struct ceph_mds_snaprealm_reconnect { + __le64 ino; /* snap realm base */ + __le64 seq; /* snap seq for this snap realm */ + __le64 parent; /* parent realm */ +} __attribute__ ((packed)); + +/* + * snaps + */ +enum { + CEPH_SNAP_OP_UPDATE, /* CREATE or DESTROY */ + CEPH_SNAP_OP_CREATE, + CEPH_SNAP_OP_DESTROY, + CEPH_SNAP_OP_SPLIT, +}; + +extern const char *ceph_snap_op_name(int o); + +/* snap msg header */ +struct ceph_mds_snap_head { + __le32 op; /* CEPH_SNAP_OP_* */ + __le64 split; /* ino to split off, if any */ + __le32 num_split_inos; /* # inos belonging to new child realm */ + __le32 num_split_realms; /* # child realms udner new child realm */ + __le32 trace_len; /* size of snap trace blob */ +} __attribute__ ((packed)); +/* followed by split ino list, then split realms, then the trace blob */ + +/* + * encode info about a snaprealm, as viewed by a client + */ +struct ceph_mds_snap_realm { + __le64 ino; /* ino */ + __le64 created; /* snap: when created */ + __le64 parent; /* ino: parent realm */ + __le64 parent_since; /* snap: same parent since */ + __le64 seq; /* snap: version */ + __le32 num_snaps; + __le32 num_prior_parent_snaps; +} __attribute__ ((packed)); +/* followed by my snap list, then prior parent snap list */ + +#endif diff --git a/include/linux/ceph/ceph_hash.h b/include/linux/ceph/ceph_hash.h new file mode 100644 index 0000000..d099c3f --- /dev/null +++ b/include/linux/ceph/ceph_hash.h @@ -0,0 +1,13 @@ +#ifndef FS_CEPH_HASH_H +#define FS_CEPH_HASH_H + +#define CEPH_STR_HASH_LINUX 0x1 /* linux dcache hash */ +#define CEPH_STR_HASH_RJENKINS 0x2 /* robert jenkins' */ + +extern unsigned ceph_str_hash_linux(const char *s, unsigned len); +extern unsigned ceph_str_hash_rjenkins(const char *s, unsigned len); + +extern unsigned ceph_str_hash(int type, const char *s, unsigned len); +extern const char *ceph_str_hash_name(int type); + +#endif diff --git a/include/linux/ceph/debugfs.h b/include/linux/ceph/debugfs.h new file mode 100644 index 0000000..2a79702 --- /dev/null +++ b/include/linux/ceph/debugfs.h @@ -0,0 +1,33 @@ +#ifndef _FS_CEPH_DEBUGFS_H +#define _FS_CEPH_DEBUGFS_H + +#include "ceph_debug.h" +#include "types.h" + +#define CEPH_DEFINE_SHOW_FUNC(name) \ +static int name##_open(struct inode *inode, struct file *file) \ +{ \ + struct seq_file *sf; \ + int ret; \ + \ + ret = single_open(file, name, NULL); \ + sf = file->private_data; \ + sf->private = inode->i_private; \ + return ret; \ +} \ + \ +static const struct file_operations name##_fops = { \ + .open = name##_open, \ + .read = seq_read, \ + .llseek = seq_lseek, \ + .release = single_release, \ +}; + +/* debugfs.c */ +extern int ceph_debugfs_init(void); +extern void ceph_debugfs_cleanup(void); +extern int ceph_debugfs_client_init(struct ceph_client *client); +extern void ceph_debugfs_client_cleanup(struct ceph_client *client); + +#endif + diff --git a/include/linux/ceph/decode.h b/include/linux/ceph/decode.h new file mode 100644 index 0000000..c5b6939 --- /dev/null +++ b/include/linux/ceph/decode.h @@ -0,0 +1,201 @@ +#ifndef __CEPH_DECODE_H +#define __CEPH_DECODE_H + +#include +#include + +#include "types.h" + +/* + * in all cases, + * void **p pointer to position pointer + * void *end pointer to end of buffer (last byte + 1) + */ + +static inline u64 ceph_decode_64(void **p) +{ + u64 v = get_unaligned_le64(*p); + *p += sizeof(u64); + return v; +} +static inline u32 ceph_decode_32(void **p) +{ + u32 v = get_unaligned_le32(*p); + *p += sizeof(u32); + return v; +} +static inline u16 ceph_decode_16(void **p) +{ + u16 v = get_unaligned_le16(*p); + *p += sizeof(u16); + return v; +} +static inline u8 ceph_decode_8(void **p) +{ + u8 v = *(u8 *)*p; + (*p)++; + return v; +} +static inline void ceph_decode_copy(void **p, void *pv, size_t n) +{ + memcpy(pv, *p, n); + *p += n; +} + +/* + * bounds check input. + */ +#define ceph_decode_need(p, end, n, bad) \ + do { \ + if (unlikely(*(p) + (n) > (end))) \ + goto bad; \ + } while (0) + +#define ceph_decode_64_safe(p, end, v, bad) \ + do { \ + ceph_decode_need(p, end, sizeof(u64), bad); \ + v = ceph_decode_64(p); \ + } while (0) +#define ceph_decode_32_safe(p, end, v, bad) \ + do { \ + ceph_decode_need(p, end, sizeof(u32), bad); \ + v = ceph_decode_32(p); \ + } while (0) +#define ceph_decode_16_safe(p, end, v, bad) \ + do { \ + ceph_decode_need(p, end, sizeof(u16), bad); \ + v = ceph_decode_16(p); \ + } while (0) +#define ceph_decode_8_safe(p, end, v, bad) \ + do { \ + ceph_decode_need(p, end, sizeof(u8), bad); \ + v = ceph_decode_8(p); \ + } while (0) + +#define ceph_decode_copy_safe(p, end, pv, n, bad) \ + do { \ + ceph_decode_need(p, end, n, bad); \ + ceph_decode_copy(p, pv, n); \ + } while (0) + +/* + * struct ceph_timespec <-> struct timespec + */ +static inline void ceph_decode_timespec(struct timespec *ts, + const struct ceph_timespec *tv) +{ + ts->tv_sec = le32_to_cpu(tv->tv_sec); + ts->tv_nsec = le32_to_cpu(tv->tv_nsec); +} +static inline void ceph_encode_timespec(struct ceph_timespec *tv, + const struct timespec *ts) +{ + tv->tv_sec = cpu_to_le32(ts->tv_sec); + tv->tv_nsec = cpu_to_le32(ts->tv_nsec); +} + +/* + * sockaddr_storage <-> ceph_sockaddr + */ +static inline void ceph_encode_addr(struct ceph_entity_addr *a) +{ + __be16 ss_family = htons(a->in_addr.ss_family); + a->in_addr.ss_family = *(__u16 *)&ss_family; +} +static inline void ceph_decode_addr(struct ceph_entity_addr *a) +{ + __be16 ss_family = *(__be16 *)&a->in_addr.ss_family; + a->in_addr.ss_family = ntohs(ss_family); + WARN_ON(a->in_addr.ss_family == 512); +} + +/* + * encoders + */ +static inline void ceph_encode_64(void **p, u64 v) +{ + put_unaligned_le64(v, (__le64 *)*p); + *p += sizeof(u64); +} +static inline void ceph_encode_32(void **p, u32 v) +{ + put_unaligned_le32(v, (__le32 *)*p); + *p += sizeof(u32); +} +static inline void ceph_encode_16(void **p, u16 v) +{ + put_unaligned_le16(v, (__le16 *)*p); + *p += sizeof(u16); +} +static inline void ceph_encode_8(void **p, u8 v) +{ + *(u8 *)*p = v; + (*p)++; +} +static inline void ceph_encode_copy(void **p, const void *s, int len) +{ + memcpy(*p, s, len); + *p += len; +} + +/* + * filepath, string encoders + */ +static inline void ceph_encode_filepath(void **p, void *end, + u64 ino, const char *path) +{ + u32 len = path ? strlen(path) : 0; + BUG_ON(*p + sizeof(ino) + sizeof(len) + len > end); + ceph_encode_8(p, 1); + ceph_encode_64(p, ino); + ceph_encode_32(p, len); + if (len) + memcpy(*p, path, len); + *p += len; +} + +static inline void ceph_encode_string(void **p, void *end, + const char *s, u32 len) +{ + BUG_ON(*p + sizeof(len) + len > end); + ceph_encode_32(p, len); + if (len) + memcpy(*p, s, len); + *p += len; +} + +#define ceph_encode_need(p, end, n, bad) \ + do { \ + if (unlikely(*(p) + (n) > (end))) \ + goto bad; \ + } while (0) + +#define ceph_encode_64_safe(p, end, v, bad) \ + do { \ + ceph_encode_need(p, end, sizeof(u64), bad); \ + ceph_encode_64(p, v); \ + } while (0) +#define ceph_encode_32_safe(p, end, v, bad) \ + do { \ + ceph_encode_need(p, end, sizeof(u32), bad); \ + ceph_encode_32(p, v); \ + } while (0) +#define ceph_encode_16_safe(p, end, v, bad) \ + do { \ + ceph_encode_need(p, end, sizeof(u16), bad); \ + ceph_encode_16(p, v); \ + } while (0) + +#define ceph_encode_copy_safe(p, end, pv, n, bad) \ + do { \ + ceph_encode_need(p, end, n, bad); \ + ceph_encode_copy(p, pv, n); \ + } while (0) +#define ceph_encode_string_safe(p, end, s, n, bad) \ + do { \ + ceph_encode_need(p, end, n, bad); \ + ceph_encode_string(p, end, s, n); \ + } while (0) + + +#endif diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h new file mode 100644 index 0000000..f22b2e9 --- /dev/null +++ b/include/linux/ceph/libceph.h @@ -0,0 +1,249 @@ +#ifndef _FS_CEPH_LIBCEPH_H +#define _FS_CEPH_LIBCEPH_H + +#include "ceph_debug.h" + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "types.h" +#include "messenger.h" +#include "msgpool.h" +#include "mon_client.h" +#include "osd_client.h" +#include "ceph_fs.h" + +/* + * Supported features + */ +#define CEPH_FEATURE_SUPPORTED_DEFAULT CEPH_FEATURE_NOSRCADDR +#define CEPH_FEATURE_REQUIRED_DEFAULT CEPH_FEATURE_NOSRCADDR + +/* + * mount options + */ +#define CEPH_OPT_FSID (1<<0) +#define CEPH_OPT_NOSHARE (1<<1) /* don't share client with other sbs */ +#define CEPH_OPT_MYIP (1<<2) /* specified my ip */ +#define CEPH_OPT_NOCRC (1<<3) /* no data crc on writes */ + +#define CEPH_OPT_DEFAULT (0); + +#define ceph_set_opt(client, opt) \ + (client)->options->flags |= CEPH_OPT_##opt; +#define ceph_test_opt(client, opt) \ + (!!((client)->options->flags & CEPH_OPT_##opt)) + +struct ceph_options { + int flags; + struct ceph_fsid fsid; + struct ceph_entity_addr my_addr; + int mount_timeout; + int osd_idle_ttl; + int osd_timeout; + int osd_keepalive_timeout; + + /* + * any type that can't be simply compared or doesn't need need + * to be compared should go beyond this point, + * ceph_compare_options() should be updated accordingly + */ + + struct ceph_entity_addr *mon_addr; /* should be the first + pointer type of args */ + int num_mon; + char *name; + char *secret; +}; + +/* + * defaults + */ +#define CEPH_MOUNT_TIMEOUT_DEFAULT 60 +#define CEPH_OSD_TIMEOUT_DEFAULT 60 /* seconds */ +#define CEPH_OSD_KEEPALIVE_DEFAULT 5 +#define CEPH_OSD_IDLE_TTL_DEFAULT 60 +#define CEPH_MOUNT_RSIZE_DEFAULT (512*1024) /* readahead */ + +#define CEPH_MSG_MAX_FRONT_LEN (16*1024*1024) +#define CEPH_MSG_MAX_DATA_LEN (16*1024*1024) + +#define CEPH_AUTH_NAME_DEFAULT "guest" + +/* + * Delay telling the MDS we no longer want caps, in case we reopen + * the file. Delay a minimum amount of time, even if we send a cap + * message for some other reason. Otherwise, take the oppotunity to + * update the mds to avoid sending another message later. + */ +#define CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT 5 /* cap release delay */ +#define CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT 60 /* cap release delay */ + +#define CEPH_CAP_RELEASE_SAFETY_DEFAULT (CEPH_CAPS_PER_RELEASE * 4) + +/* mount state */ +enum { + CEPH_MOUNT_MOUNTING, + CEPH_MOUNT_MOUNTED, + CEPH_MOUNT_UNMOUNTING, + CEPH_MOUNT_UNMOUNTED, + CEPH_MOUNT_SHUTDOWN, +}; + +/* + * subtract jiffies + */ +static inline unsigned long time_sub(unsigned long a, unsigned long b) +{ + BUG_ON(time_after(b, a)); + return (long)a - (long)b; +} + +struct ceph_mds_client; + +/* + * per client state + * + * possibly shared by multiple mount points, if they are + * mounting the same ceph filesystem/cluster. + */ +struct ceph_client { + struct ceph_fsid fsid; + bool have_fsid; + + void *private; + + struct ceph_options *options; + + struct mutex mount_mutex; /* serialize mount attempts */ + wait_queue_head_t auth_wq; + int auth_err; + + int (*extra_mon_dispatch)(struct ceph_client *, struct ceph_msg *); + + u32 supported_features; + u32 required_features; + + struct ceph_messenger *msgr; /* messenger instance */ + struct ceph_mon_client monc; + struct ceph_osd_client osdc; + +#ifdef CONFIG_DEBUG_FS + struct dentry *debugfs_dir; + struct dentry *debugfs_monmap; + struct dentry *debugfs_osdmap; +#endif +}; + + + +/* + * snapshots + */ + +/* + * A "snap context" is the set of existing snapshots when we + * write data. It is used by the OSD to guide its COW behavior. + * + * The ceph_snap_context is refcounted, and attached to each dirty + * page, indicating which context the dirty data belonged when it was + * dirtied. + */ +struct ceph_snap_context { + atomic_t nref; + u64 seq; + int num_snaps; + u64 snaps[]; +}; + +static inline struct ceph_snap_context * +ceph_get_snap_context(struct ceph_snap_context *sc) +{ + /* + printk("get_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref), + atomic_read(&sc->nref)+1); + */ + if (sc) + atomic_inc(&sc->nref); + return sc; +} + +static inline void ceph_put_snap_context(struct ceph_snap_context *sc) +{ + if (!sc) + return; + /* + printk("put_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref), + atomic_read(&sc->nref)-1); + */ + if (atomic_dec_and_test(&sc->nref)) { + /*printk(" deleting snap_context %p\n", sc);*/ + kfree(sc); + } +} + +/* + * calculate the number of pages a given length and offset map onto, + * if we align the data. + */ +static inline int calc_pages_for(u64 off, u64 len) +{ + return ((off+len+PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT) - + (off >> PAGE_CACHE_SHIFT); +} + +/* ceph_common.c */ +extern const char *ceph_msg_type_name(int type); +extern int ceph_check_fsid(struct ceph_client *client, struct ceph_fsid *fsid); +extern struct kmem_cache *ceph_inode_cachep; +extern struct kmem_cache *ceph_cap_cachep; +extern struct kmem_cache *ceph_dentry_cachep; +extern struct kmem_cache *ceph_file_cachep; + +extern int ceph_parse_options(struct ceph_options **popt, char *options, + const char *dev_name, const char *dev_name_end, + int (*parse_extra_token)(char *c, void *private), + void *private); +extern void ceph_destroy_options(struct ceph_options *opt); +extern int ceph_compare_options(struct ceph_options *new_opt, + struct ceph_client *client); +extern struct ceph_client *ceph_create_client(struct ceph_options *opt, + void *private); +extern u64 ceph_client_id(struct ceph_client *client); +extern void ceph_destroy_client(struct ceph_client *client); +extern int __ceph_open_session(struct ceph_client *client, + unsigned long started); +extern int ceph_open_session(struct ceph_client *client); + +/* pagevec.c */ +extern void ceph_release_page_vector(struct page **pages, int num_pages); + +extern struct page **ceph_get_direct_page_vector(const char __user *data, + int num_pages, + loff_t off, size_t len); +extern void ceph_put_page_vector(struct page **pages, int num_pages); +extern void ceph_release_page_vector(struct page **pages, int num_pages); +extern struct page **ceph_alloc_page_vector(int num_pages, gfp_t flags); +extern int ceph_copy_user_to_page_vector(struct page **pages, + const char __user *data, + loff_t off, size_t len); +extern int ceph_copy_to_page_vector(struct page **pages, + const char *data, + loff_t off, size_t len); +extern int ceph_copy_from_page_vector(struct page **pages, + char *data, + loff_t off, size_t len); +extern int ceph_copy_page_vector_to_user(struct page **pages, char __user *data, + loff_t off, size_t len); +extern void ceph_zero_page_vector_range(int off, int len, struct page **pages); + + +#endif /* _FS_CEPH_SUPER_H */ diff --git a/include/linux/ceph/mdsmap.h b/include/linux/ceph/mdsmap.h new file mode 100644 index 0000000..4c5cb08 --- /dev/null +++ b/include/linux/ceph/mdsmap.h @@ -0,0 +1,62 @@ +#ifndef _FS_CEPH_MDSMAP_H +#define _FS_CEPH_MDSMAP_H + +#include "types.h" + +/* + * mds map - describe servers in the mds cluster. + * + * we limit fields to those the client actually xcares about + */ +struct ceph_mds_info { + u64 global_id; + struct ceph_entity_addr addr; + s32 state; + int num_export_targets; + bool laggy; + u32 *export_targets; +}; + +struct ceph_mdsmap { + u32 m_epoch, m_client_epoch, m_last_failure; + u32 m_root; + u32 m_session_timeout; /* seconds */ + u32 m_session_autoclose; /* seconds */ + u64 m_max_file_size; + u32 m_max_mds; /* size of m_addr, m_state arrays */ + struct ceph_mds_info *m_info; + + /* which object pools file data can be stored in */ + int m_num_data_pg_pools; + u32 *m_data_pg_pools; + u32 m_cas_pg_pool; +}; + +static inline struct ceph_entity_addr * +ceph_mdsmap_get_addr(struct ceph_mdsmap *m, int w) +{ + if (w >= m->m_max_mds) + return NULL; + return &m->m_info[w].addr; +} + +static inline int ceph_mdsmap_get_state(struct ceph_mdsmap *m, int w) +{ + BUG_ON(w < 0); + if (w >= m->m_max_mds) + return CEPH_MDS_STATE_DNE; + return m->m_info[w].state; +} + +static inline bool ceph_mdsmap_is_laggy(struct ceph_mdsmap *m, int w) +{ + if (w >= 0 && w < m->m_max_mds) + return m->m_info[w].laggy; + return false; +} + +extern int ceph_mdsmap_get_random_mds(struct ceph_mdsmap *m); +extern struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end); +extern void ceph_mdsmap_destroy(struct ceph_mdsmap *m); + +#endif diff --git a/include/linux/ceph/messenger.h b/include/linux/ceph/messenger.h new file mode 100644 index 0000000..5956d62 --- /dev/null +++ b/include/linux/ceph/messenger.h @@ -0,0 +1,261 @@ +#ifndef __FS_CEPH_MESSENGER_H +#define __FS_CEPH_MESSENGER_H + +#include +#include +#include +#include +#include +#include +#include + +#include "types.h" +#include "buffer.h" + +struct ceph_msg; +struct ceph_connection; + +extern struct workqueue_struct *ceph_msgr_wq; /* receive work queue */ + +/* + * Ceph defines these callbacks for handling connection events. + */ +struct ceph_connection_operations { + struct ceph_connection *(*get)(struct ceph_connection *); + void (*put)(struct ceph_connection *); + + /* handle an incoming message. */ + void (*dispatch) (struct ceph_connection *con, struct ceph_msg *m); + + /* authorize an outgoing connection */ + int (*get_authorizer) (struct ceph_connection *con, + void **buf, int *len, int *proto, + void **reply_buf, int *reply_len, int force_new); + int (*verify_authorizer_reply) (struct ceph_connection *con, int len); + int (*invalidate_authorizer)(struct ceph_connection *con); + + /* protocol version mismatch */ + void (*bad_proto) (struct ceph_connection *con); + + /* there was some error on the socket (disconnect, whatever) */ + void (*fault) (struct ceph_connection *con); + + /* a remote host as terminated a message exchange session, and messages + * we sent (or they tried to send us) may be lost. */ + void (*peer_reset) (struct ceph_connection *con); + + struct ceph_msg * (*alloc_msg) (struct ceph_connection *con, + struct ceph_msg_header *hdr, + int *skip); +}; + +/* use format string %s%d */ +#define ENTITY_NAME(n) ceph_entity_type_name((n).type), le64_to_cpu((n).num) + +struct ceph_messenger { + struct ceph_entity_inst inst; /* my name+address */ + struct ceph_entity_addr my_enc_addr; + struct page *zero_page; /* used in certain error cases */ + + bool nocrc; + + /* + * the global_seq counts connections i (attempt to) initiate + * in order to disambiguate certain connect race conditions. + */ + u32 global_seq; + spinlock_t global_seq_lock; + + u32 supported_features; + u32 required_features; +}; + +/* + * a single message. it contains a header (src, dest, message type, etc.), + * footer (crc values, mainly), a "front" message body, and possibly a + * data payload (stored in some number of pages). + */ +struct ceph_msg { + struct ceph_msg_header hdr; /* header */ + struct ceph_msg_footer footer; /* footer */ + struct kvec front; /* unaligned blobs of message */ + struct ceph_buffer *middle; + struct page **pages; /* data payload. NOT OWNER. */ + unsigned nr_pages; /* size of page array */ + struct ceph_pagelist *pagelist; /* instead of pages */ + struct list_head list_head; + struct kref kref; + struct bio *bio; /* instead of pages/pagelist */ + struct bio *bio_iter; /* bio iterator */ + int bio_seg; /* current bio segment */ + struct ceph_pagelist *trail; /* the trailing part of the data */ + bool front_is_vmalloc; + bool more_to_follow; + bool needs_out_seq; + int front_max; + + struct ceph_msgpool *pool; +}; + +struct ceph_msg_pos { + int page, page_pos; /* which page; offset in page */ + int data_pos; /* offset in data payload */ + int did_page_crc; /* true if we've calculated crc for current page */ +}; + +/* ceph connection fault delay defaults, for exponential backoff */ +#define BASE_DELAY_INTERVAL (HZ/2) +#define MAX_DELAY_INTERVAL (5 * 60 * HZ) + +/* + * ceph_connection state bit flags + * + * QUEUED and BUSY are used together to ensure that only a single + * thread is currently opening, reading or writing data to the socket. + */ +#define LOSSYTX 0 /* we can close channel or drop messages on errors */ +#define CONNECTING 1 +#define NEGOTIATING 2 +#define KEEPALIVE_PENDING 3 +#define WRITE_PENDING 4 /* we have data ready to send */ +#define QUEUED 5 /* there is work queued on this connection */ +#define BUSY 6 /* work is being done */ +#define STANDBY 8 /* no outgoing messages, socket closed. we keep + * the ceph_connection around to maintain shared + * state with the peer. */ +#define CLOSED 10 /* we've closed the connection */ +#define SOCK_CLOSED 11 /* socket state changed to closed */ +#define OPENING 13 /* open connection w/ (possibly new) peer */ +#define DEAD 14 /* dead, about to kfree */ + +/* + * A single connection with another host. + * + * We maintain a queue of outgoing messages, and some session state to + * ensure that we can preserve the lossless, ordered delivery of + * messages in the case of a TCP disconnect. + */ +struct ceph_connection { + void *private; + atomic_t nref; + + const struct ceph_connection_operations *ops; + + struct ceph_messenger *msgr; + struct socket *sock; + unsigned long state; /* connection state (see flags above) */ + const char *error_msg; /* error message, if any */ + + struct ceph_entity_addr peer_addr; /* peer address */ + struct ceph_entity_name peer_name; /* peer name */ + struct ceph_entity_addr peer_addr_for_me; + unsigned peer_features; + u32 connect_seq; /* identify the most recent connection + attempt for this connection, client */ + u32 peer_global_seq; /* peer's global seq for this connection */ + + int auth_retry; /* true if we need a newer authorizer */ + void *auth_reply_buf; /* where to put the authorizer reply */ + int auth_reply_buf_len; + + struct mutex mutex; + + /* out queue */ + struct list_head out_queue; + struct list_head out_sent; /* sending or sent but unacked */ + u64 out_seq; /* last message queued for send */ + bool out_keepalive_pending; + + u64 in_seq, in_seq_acked; /* last message received, acked */ + + /* connection negotiation temps */ + char in_banner[CEPH_BANNER_MAX_LEN]; + union { + struct { /* outgoing connection */ + struct ceph_msg_connect out_connect; + struct ceph_msg_connect_reply in_reply; + }; + struct { /* incoming */ + struct ceph_msg_connect in_connect; + struct ceph_msg_connect_reply out_reply; + }; + }; + struct ceph_entity_addr actual_peer_addr; + + /* message out temps */ + struct ceph_msg *out_msg; /* sending message (== tail of + out_sent) */ + bool out_msg_done; + struct ceph_msg_pos out_msg_pos; + + struct kvec out_kvec[8], /* sending header/footer data */ + *out_kvec_cur; + int out_kvec_left; /* kvec's left in out_kvec */ + int out_skip; /* skip this many bytes */ + int out_kvec_bytes; /* total bytes left */ + bool out_kvec_is_msg; /* kvec refers to out_msg */ + int out_more; /* there is more data after the kvecs */ + __le64 out_temp_ack; /* for writing an ack */ + + /* message in temps */ + struct ceph_msg_header in_hdr; + struct ceph_msg *in_msg; + struct ceph_msg_pos in_msg_pos; + u32 in_front_crc, in_middle_crc, in_data_crc; /* calculated crc */ + + char in_tag; /* protocol control byte */ + int in_base_pos; /* bytes read */ + __le64 in_temp_ack; /* for reading an ack */ + + struct delayed_work work; /* send|recv work */ + unsigned long delay; /* current delay interval */ +}; + + +extern const char *ceph_pr_addr(const struct sockaddr_storage *ss); +extern int ceph_parse_ips(const char *c, const char *end, + struct ceph_entity_addr *addr, + int max_count, int *count); + + +extern int ceph_msgr_init(void); +extern void ceph_msgr_exit(void); +extern void ceph_msgr_flush(void); + +extern struct ceph_messenger *ceph_messenger_create( + struct ceph_entity_addr *myaddr, + u32 features, u32 required); +extern void ceph_messenger_destroy(struct ceph_messenger *); + +extern void ceph_con_init(struct ceph_messenger *msgr, + struct ceph_connection *con); +extern void ceph_con_open(struct ceph_connection *con, + struct ceph_entity_addr *addr); +extern bool ceph_con_opened(struct ceph_connection *con); +extern void ceph_con_close(struct ceph_connection *con); +extern void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg); +extern void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg); +extern void ceph_con_revoke_message(struct ceph_connection *con, + struct ceph_msg *msg); +extern void ceph_con_keepalive(struct ceph_connection *con); +extern struct ceph_connection *ceph_con_get(struct ceph_connection *con); +extern void ceph_con_put(struct ceph_connection *con); + +extern struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags); +extern void ceph_msg_kfree(struct ceph_msg *m); + + +static inline struct ceph_msg *ceph_msg_get(struct ceph_msg *msg) +{ + kref_get(&msg->kref); + return msg; +} +extern void ceph_msg_last_put(struct kref *kref); +static inline void ceph_msg_put(struct ceph_msg *msg) +{ + kref_put(&msg->kref, ceph_msg_last_put); +} + +extern void ceph_msg_dump(struct ceph_msg *msg); + +#endif diff --git a/include/linux/ceph/mon_client.h b/include/linux/ceph/mon_client.h new file mode 100644 index 0000000..545f859 --- /dev/null +++ b/include/linux/ceph/mon_client.h @@ -0,0 +1,122 @@ +#ifndef _FS_CEPH_MON_CLIENT_H +#define _FS_CEPH_MON_CLIENT_H + +#include +#include +#include + +#include "messenger.h" + +struct ceph_client; +struct ceph_mount_args; +struct ceph_auth_client; + +/* + * The monitor map enumerates the set of all monitors. + */ +struct ceph_monmap { + struct ceph_fsid fsid; + u32 epoch; + u32 num_mon; + struct ceph_entity_inst mon_inst[0]; +}; + +struct ceph_mon_client; +struct ceph_mon_generic_request; + + +/* + * Generic mechanism for resending monitor requests. + */ +typedef void (*ceph_monc_request_func_t)(struct ceph_mon_client *monc, + int newmon); + +/* a pending monitor request */ +struct ceph_mon_request { + struct ceph_mon_client *monc; + struct delayed_work delayed_work; + unsigned long delay; + ceph_monc_request_func_t do_request; +}; + +/* + * ceph_mon_generic_request is being used for the statfs and poolop requests + * which are bening done a bit differently because we need to get data back + * to the caller + */ +struct ceph_mon_generic_request { + struct kref kref; + u64 tid; + struct rb_node node; + int result; + void *buf; + int buf_len; + struct completion completion; + struct ceph_msg *request; /* original request */ + struct ceph_msg *reply; /* and reply */ +}; + +struct ceph_mon_client { + struct ceph_client *client; + struct ceph_monmap *monmap; + + struct mutex mutex; + struct delayed_work delayed_work; + + struct ceph_auth_client *auth; + struct ceph_msg *m_auth, *m_auth_reply, *m_subscribe, *m_subscribe_ack; + int pending_auth; + + bool hunting; + int cur_mon; /* last monitor i contacted */ + unsigned long sub_sent, sub_renew_after; + struct ceph_connection *con; + bool have_fsid; + + /* pending generic requests */ + struct rb_root generic_request_tree; + int num_generic_requests; + u64 last_tid; + + /* mds/osd map */ + int want_mdsmap; + int want_next_osdmap; /* 1 = want, 2 = want+asked */ + u32 have_osdmap, have_mdsmap; + +#ifdef CONFIG_DEBUG_FS + struct dentry *debugfs_file; +#endif +}; + +extern struct ceph_monmap *ceph_monmap_decode(void *p, void *end); +extern int ceph_monmap_contains(struct ceph_monmap *m, + struct ceph_entity_addr *addr); + +extern int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl); +extern void ceph_monc_stop(struct ceph_mon_client *monc); + +/* + * The model here is to indicate that we need a new map of at least + * epoch @want, and also call in when we receive a map. We will + * periodically rerequest the map from the monitor cluster until we + * get what we want. + */ +extern int ceph_monc_got_mdsmap(struct ceph_mon_client *monc, u32 have); +extern int ceph_monc_got_osdmap(struct ceph_mon_client *monc, u32 have); + +extern void ceph_monc_request_next_osdmap(struct ceph_mon_client *monc); + +extern int ceph_monc_do_statfs(struct ceph_mon_client *monc, + struct ceph_statfs *buf); + +extern int ceph_monc_open_session(struct ceph_mon_client *monc); + +extern int ceph_monc_validate_auth(struct ceph_mon_client *monc); + +extern int ceph_monc_create_snapid(struct ceph_mon_client *monc, + u32 pool, u64 *snapid); + +extern int ceph_monc_delete_snapid(struct ceph_mon_client *monc, + u32 pool, u64 snapid); + +#endif diff --git a/include/linux/ceph/msgpool.h b/include/linux/ceph/msgpool.h new file mode 100644 index 0000000..a362605f --- /dev/null +++ b/include/linux/ceph/msgpool.h @@ -0,0 +1,25 @@ +#ifndef _FS_CEPH_MSGPOOL +#define _FS_CEPH_MSGPOOL + +#include +#include "messenger.h" + +/* + * we use memory pools for preallocating messages we may receive, to + * avoid unexpected OOM conditions. + */ +struct ceph_msgpool { + const char *name; + mempool_t *pool; + int front_len; /* preallocated payload size */ +}; + +extern int ceph_msgpool_init(struct ceph_msgpool *pool, + int front_len, int size, bool blocking, + const char *name); +extern void ceph_msgpool_destroy(struct ceph_msgpool *pool); +extern struct ceph_msg *ceph_msgpool_get(struct ceph_msgpool *, + int front_len); +extern void ceph_msgpool_put(struct ceph_msgpool *, struct ceph_msg *); + +#endif diff --git a/include/linux/ceph/msgr.h b/include/linux/ceph/msgr.h new file mode 100644 index 0000000..680d3d6 --- /dev/null +++ b/include/linux/ceph/msgr.h @@ -0,0 +1,175 @@ +#ifndef CEPH_MSGR_H +#define CEPH_MSGR_H + +/* + * Data types for message passing layer used by Ceph. + */ + +#define CEPH_MON_PORT 6789 /* default monitor port */ + +/* + * client-side processes will try to bind to ports in this + * range, simply for the benefit of tools like nmap or wireshark + * that would like to identify the protocol. + */ +#define CEPH_PORT_FIRST 6789 +#define CEPH_PORT_START 6800 /* non-monitors start here */ +#define CEPH_PORT_LAST 6900 + +/* + * tcp connection banner. include a protocol version. and adjust + * whenever the wire protocol changes. try to keep this string length + * constant. + */ +#define CEPH_BANNER "ceph v027" +#define CEPH_BANNER_MAX_LEN 30 + + +/* + * Rollover-safe type and comparator for 32-bit sequence numbers. + * Comparator returns -1, 0, or 1. + */ +typedef __u32 ceph_seq_t; + +static inline __s32 ceph_seq_cmp(__u32 a, __u32 b) +{ + return (__s32)a - (__s32)b; +} + + +/* + * entity_name -- logical name for a process participating in the + * network, e.g. 'mds0' or 'osd3'. + */ +struct ceph_entity_name { + __u8 type; /* CEPH_ENTITY_TYPE_* */ + __le64 num; +} __attribute__ ((packed)); + +#define CEPH_ENTITY_TYPE_MON 0x01 +#define CEPH_ENTITY_TYPE_MDS 0x02 +#define CEPH_ENTITY_TYPE_OSD 0x04 +#define CEPH_ENTITY_TYPE_CLIENT 0x08 +#define CEPH_ENTITY_TYPE_AUTH 0x20 + +#define CEPH_ENTITY_TYPE_ANY 0xFF + +extern const char *ceph_entity_type_name(int type); + +/* + * entity_addr -- network address + */ +struct ceph_entity_addr { + __le32 type; + __le32 nonce; /* unique id for process (e.g. pid) */ + struct sockaddr_storage in_addr; +} __attribute__ ((packed)); + +struct ceph_entity_inst { + struct ceph_entity_name name; + struct ceph_entity_addr addr; +} __attribute__ ((packed)); + + +/* used by message exchange protocol */ +#define CEPH_MSGR_TAG_READY 1 /* server->client: ready for messages */ +#define CEPH_MSGR_TAG_RESETSESSION 2 /* server->client: reset, try again */ +#define CEPH_MSGR_TAG_WAIT 3 /* server->client: wait for racing + incoming connection */ +#define CEPH_MSGR_TAG_RETRY_SESSION 4 /* server->client + cseq: try again + with higher cseq */ +#define CEPH_MSGR_TAG_RETRY_GLOBAL 5 /* server->client + gseq: try again + with higher gseq */ +#define CEPH_MSGR_TAG_CLOSE 6 /* closing pipe */ +#define CEPH_MSGR_TAG_MSG 7 /* message */ +#define CEPH_MSGR_TAG_ACK 8 /* message ack */ +#define CEPH_MSGR_TAG_KEEPALIVE 9 /* just a keepalive byte! */ +#define CEPH_MSGR_TAG_BADPROTOVER 10 /* bad protocol version */ +#define CEPH_MSGR_TAG_BADAUTHORIZER 11 /* bad authorizer */ +#define CEPH_MSGR_TAG_FEATURES 12 /* insufficient features */ + + +/* + * connection negotiation + */ +struct ceph_msg_connect { + __le64 features; /* supported feature bits */ + __le32 host_type; /* CEPH_ENTITY_TYPE_* */ + __le32 global_seq; /* count connections initiated by this host */ + __le32 connect_seq; /* count connections initiated in this session */ + __le32 protocol_version; + __le32 authorizer_protocol; + __le32 authorizer_len; + __u8 flags; /* CEPH_MSG_CONNECT_* */ +} __attribute__ ((packed)); + +struct ceph_msg_connect_reply { + __u8 tag; + __le64 features; /* feature bits for this session */ + __le32 global_seq; + __le32 connect_seq; + __le32 protocol_version; + __le32 authorizer_len; + __u8 flags; +} __attribute__ ((packed)); + +#define CEPH_MSG_CONNECT_LOSSY 1 /* messages i send may be safely dropped */ + + +/* + * message header + */ +struct ceph_msg_header_old { + __le64 seq; /* message seq# for this session */ + __le64 tid; /* transaction id */ + __le16 type; /* message type */ + __le16 priority; /* priority. higher value == higher priority */ + __le16 version; /* version of message encoding */ + + __le32 front_len; /* bytes in main payload */ + __le32 middle_len;/* bytes in middle payload */ + __le32 data_len; /* bytes of data payload */ + __le16 data_off; /* sender: include full offset; + receiver: mask against ~PAGE_MASK */ + + struct ceph_entity_inst src, orig_src; + __le32 reserved; + __le32 crc; /* header crc32c */ +} __attribute__ ((packed)); + +struct ceph_msg_header { + __le64 seq; /* message seq# for this session */ + __le64 tid; /* transaction id */ + __le16 type; /* message type */ + __le16 priority; /* priority. higher value == higher priority */ + __le16 version; /* version of message encoding */ + + __le32 front_len; /* bytes in main payload */ + __le32 middle_len;/* bytes in middle payload */ + __le32 data_len; /* bytes of data payload */ + __le16 data_off; /* sender: include full offset; + receiver: mask against ~PAGE_MASK */ + + struct ceph_entity_name src; + __le32 reserved; + __le32 crc; /* header crc32c */ +} __attribute__ ((packed)); + +#define CEPH_MSG_PRIO_LOW 64 +#define CEPH_MSG_PRIO_DEFAULT 127 +#define CEPH_MSG_PRIO_HIGH 196 +#define CEPH_MSG_PRIO_HIGHEST 255 + +/* + * follows data payload + */ +struct ceph_msg_footer { + __le32 front_crc, middle_crc, data_crc; + __u8 flags; +} __attribute__ ((packed)); + +#define CEPH_MSG_FOOTER_COMPLETE (1<<0) /* msg wasn't aborted */ +#define CEPH_MSG_FOOTER_NOCRC (1<<1) /* no data crc */ + + +#endif diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h new file mode 100644 index 0000000..6c91fb0 --- /dev/null +++ b/include/linux/ceph/osd_client.h @@ -0,0 +1,234 @@ +#ifndef _FS_CEPH_OSD_CLIENT_H +#define _FS_CEPH_OSD_CLIENT_H + +#include +#include +#include +#include + +#include "types.h" +#include "osdmap.h" +#include "messenger.h" + +struct ceph_msg; +struct ceph_snap_context; +struct ceph_osd_request; +struct ceph_osd_client; +struct ceph_authorizer; +struct ceph_pagelist; + +/* + * completion callback for async writepages + */ +typedef void (*ceph_osdc_callback_t)(struct ceph_osd_request *, + struct ceph_msg *); + +/* a given osd we're communicating with */ +struct ceph_osd { + atomic_t o_ref; + struct ceph_osd_client *o_osdc; + int o_osd; + int o_incarnation; + struct rb_node o_node; + struct ceph_connection o_con; + struct list_head o_requests; + struct list_head o_osd_lru; + struct ceph_authorizer *o_authorizer; + void *o_authorizer_buf, *o_authorizer_reply_buf; + size_t o_authorizer_buf_len, o_authorizer_reply_buf_len; + unsigned long lru_ttl; + int o_marked_for_keepalive; + struct list_head o_keepalive_item; +}; + +/* an in-flight request */ +struct ceph_osd_request { + u64 r_tid; /* unique for this client */ + struct rb_node r_node; + struct list_head r_req_lru_item; + struct list_head r_osd_item; + struct ceph_osd *r_osd; + struct ceph_pg r_pgid; + int r_pg_osds[CEPH_PG_MAX_SIZE]; + int r_num_pg_osds; + + struct ceph_connection *r_con_filling_msg; + + struct ceph_msg *r_request, *r_reply; + int r_result; + int r_flags; /* any additional flags for the osd */ + u32 r_sent; /* >0 if r_request is sending/sent */ + int r_got_reply; + + struct ceph_osd_client *r_osdc; + struct kref r_kref; + bool r_mempool; + struct completion r_completion, r_safe_completion; + ceph_osdc_callback_t r_callback, r_safe_callback; + struct ceph_eversion r_reassert_version; + struct list_head r_unsafe_item; + + struct inode *r_inode; /* for use by callbacks */ + void *r_priv; /* ditto */ + + char r_oid[40]; /* object name */ + int r_oid_len; + unsigned long r_stamp; /* send OR check time */ + bool r_resend; /* msg send failed, needs retry */ + + struct ceph_file_layout r_file_layout; + struct ceph_snap_context *r_snapc; /* snap context for writes */ + unsigned r_num_pages; /* size of page array (follows) */ + struct page **r_pages; /* pages for data payload */ + int r_pages_from_pool; + int r_own_pages; /* if true, i own page list */ +#ifdef CONFIG_BLOCK + struct bio *r_bio; /* instead of pages */ +#endif + + struct ceph_pagelist *r_trail; /* trailing part of the data */ +}; + +struct ceph_osd_client { + struct ceph_client *client; + + struct ceph_osdmap *osdmap; /* current map */ + struct rw_semaphore map_sem; + struct completion map_waiters; + u64 last_requested_map; + + struct mutex request_mutex; + struct rb_root osds; /* osds */ + struct list_head osd_lru; /* idle osds */ + u64 timeout_tid; /* tid of timeout triggering rq */ + u64 last_tid; /* tid of last request */ + struct rb_root requests; /* pending requests */ + struct list_head req_lru; /* pending requests lru */ + int num_requests; + struct delayed_work timeout_work; + struct delayed_work osds_timeout_work; +#ifdef CONFIG_DEBUG_FS + struct dentry *debugfs_file; +#endif + + mempool_t *req_mempool; + + struct ceph_msgpool msgpool_op; + struct ceph_msgpool msgpool_op_reply; +}; + +struct ceph_osd_req_op { + u16 op; /* CEPH_OSD_OP_* */ + u32 flags; /* CEPH_OSD_FLAG_* */ + union { + struct { + u64 offset, length; + u64 truncate_size; + u32 truncate_seq; + } extent; + struct { + const char *name; + u32 name_len; + const char *val; + u32 value_len; + __u8 cmp_op; /* CEPH_OSD_CMPXATTR_OP_* */ + __u8 cmp_mode; /* CEPH_OSD_CMPXATTR_MODE_* */ + } xattr; + struct { + const char *class_name; + __u8 class_len; + const char *method_name; + __u8 method_len; + __u8 argc; + const char *indata; + u32 indata_len; + } cls; + struct { + u64 cookie, count; + } pgls; + struct { + u64 snapid; + } snap; + }; + u32 payload_len; +}; + +extern int ceph_osdc_init(struct ceph_osd_client *osdc, + struct ceph_client *client); +extern void ceph_osdc_stop(struct ceph_osd_client *osdc); + +extern void ceph_osdc_handle_reply(struct ceph_osd_client *osdc, + struct ceph_msg *msg); +extern void ceph_osdc_handle_map(struct ceph_osd_client *osdc, + struct ceph_msg *msg); + +extern void ceph_calc_raw_layout(struct ceph_osd_client *osdc, + struct ceph_file_layout *layout, + u64 snapid, + u64 off, u64 *plen, u64 *bno, + struct ceph_osd_request *req, + struct ceph_osd_req_op *op); + +extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, + int flags, + struct ceph_snap_context *snapc, + struct ceph_osd_req_op *ops, + bool use_mempool, + gfp_t gfp_flags, + struct page **pages, + struct bio *bio); + +extern void ceph_osdc_build_request(struct ceph_osd_request *req, + u64 off, u64 *plen, + struct ceph_osd_req_op *src_ops, + struct ceph_snap_context *snapc, + struct timespec *mtime, + const char *oid, + int oid_len); + +extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *, + struct ceph_file_layout *layout, + struct ceph_vino vino, + u64 offset, u64 *len, int op, int flags, + struct ceph_snap_context *snapc, + int do_sync, u32 truncate_seq, + u64 truncate_size, + struct timespec *mtime, + bool use_mempool, int num_reply); + +static inline void ceph_osdc_get_request(struct ceph_osd_request *req) +{ + kref_get(&req->r_kref); +} +extern void ceph_osdc_release_request(struct kref *kref); +static inline void ceph_osdc_put_request(struct ceph_osd_request *req) +{ + kref_put(&req->r_kref, ceph_osdc_release_request); +} + +extern int ceph_osdc_start_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req, + bool nofail); +extern int ceph_osdc_wait_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req); +extern void ceph_osdc_sync(struct ceph_osd_client *osdc); + +extern int ceph_osdc_readpages(struct ceph_osd_client *osdc, + struct ceph_vino vino, + struct ceph_file_layout *layout, + u64 off, u64 *plen, + u32 truncate_seq, u64 truncate_size, + struct page **pages, int nr_pages); + +extern int ceph_osdc_writepages(struct ceph_osd_client *osdc, + struct ceph_vino vino, + struct ceph_file_layout *layout, + struct ceph_snap_context *sc, + u64 off, u64 len, + u32 truncate_seq, u64 truncate_size, + struct timespec *mtime, + struct page **pages, int nr_pages, + int flags, int do_sync, bool nofail); + +#endif + diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h new file mode 100644 index 0000000..ba4c205 --- /dev/null +++ b/include/linux/ceph/osdmap.h @@ -0,0 +1,130 @@ +#ifndef _FS_CEPH_OSDMAP_H +#define _FS_CEPH_OSDMAP_H + +#include +#include "types.h" +#include "ceph_fs.h" +#include + +/* + * The osd map describes the current membership of the osd cluster and + * specifies the mapping of objects to placement groups and placement + * groups to (sets of) osds. That is, it completely specifies the + * (desired) distribution of all data objects in the system at some + * point in time. + * + * Each map version is identified by an epoch, which increases monotonically. + * + * The map can be updated either via an incremental map (diff) describing + * the change between two successive epochs, or as a fully encoded map. + */ +struct ceph_pg_pool_info { + struct rb_node node; + int id; + struct ceph_pg_pool v; + int pg_num_mask, pgp_num_mask, lpg_num_mask, lpgp_num_mask; + char *name; +}; + +struct ceph_pg_mapping { + struct rb_node node; + struct ceph_pg pgid; + int len; + int osds[]; +}; + +struct ceph_osdmap { + struct ceph_fsid fsid; + u32 epoch; + u32 mkfs_epoch; + struct ceph_timespec created, modified; + + u32 flags; /* CEPH_OSDMAP_* */ + + u32 max_osd; /* size of osd_state, _offload, _addr arrays */ + u8 *osd_state; /* CEPH_OSD_* */ + u32 *osd_weight; /* 0 = failed, 0x10000 = 100% normal */ + struct ceph_entity_addr *osd_addr; + + struct rb_root pg_temp; + struct rb_root pg_pools; + u32 pool_max; + + /* the CRUSH map specifies the mapping of placement groups to + * the list of osds that store+replicate them. */ + struct crush_map *crush; +}; + +/* + * file layout helpers + */ +#define ceph_file_layout_su(l) ((__s32)le32_to_cpu((l).fl_stripe_unit)) +#define ceph_file_layout_stripe_count(l) \ + ((__s32)le32_to_cpu((l).fl_stripe_count)) +#define ceph_file_layout_object_size(l) ((__s32)le32_to_cpu((l).fl_object_size)) +#define ceph_file_layout_cas_hash(l) ((__s32)le32_to_cpu((l).fl_cas_hash)) +#define ceph_file_layout_object_su(l) \ + ((__s32)le32_to_cpu((l).fl_object_stripe_unit)) +#define ceph_file_layout_pg_preferred(l) \ + ((__s32)le32_to_cpu((l).fl_pg_preferred)) +#define ceph_file_layout_pg_pool(l) \ + ((__s32)le32_to_cpu((l).fl_pg_pool)) + +static inline unsigned ceph_file_layout_stripe_width(struct ceph_file_layout *l) +{ + return le32_to_cpu(l->fl_stripe_unit) * + le32_to_cpu(l->fl_stripe_count); +} + +/* "period" == bytes before i start on a new set of objects */ +static inline unsigned ceph_file_layout_period(struct ceph_file_layout *l) +{ + return le32_to_cpu(l->fl_object_size) * + le32_to_cpu(l->fl_stripe_count); +} + + +static inline int ceph_osd_is_up(struct ceph_osdmap *map, int osd) +{ + return (osd < map->max_osd) && (map->osd_state[osd] & CEPH_OSD_UP); +} + +static inline bool ceph_osdmap_flag(struct ceph_osdmap *map, int flag) +{ + return map && (map->flags & flag); +} + +extern char *ceph_osdmap_state_str(char *str, int len, int state); + +static inline struct ceph_entity_addr *ceph_osd_addr(struct ceph_osdmap *map, + int osd) +{ + if (osd >= map->max_osd) + return NULL; + return &map->osd_addr[osd]; +} + +extern struct ceph_osdmap *osdmap_decode(void **p, void *end); +extern struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end, + struct ceph_osdmap *map, + struct ceph_messenger *msgr); +extern void ceph_osdmap_destroy(struct ceph_osdmap *map); + +/* calculate mapping of a file extent to an object */ +extern void ceph_calc_file_object_mapping(struct ceph_file_layout *layout, + u64 off, u64 *plen, + u64 *bno, u64 *oxoff, u64 *oxlen); + +/* calculate mapping of object to a placement group */ +extern int ceph_calc_object_layout(struct ceph_object_layout *ol, + const char *oid, + struct ceph_file_layout *fl, + struct ceph_osdmap *osdmap); +extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid, + int *acting); +extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, + struct ceph_pg pgid); + +extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name); + +#endif diff --git a/include/linux/ceph/pagelist.h b/include/linux/ceph/pagelist.h new file mode 100644 index 0000000..cc9327a --- /dev/null +++ b/include/linux/ceph/pagelist.h @@ -0,0 +1,54 @@ +#ifndef __FS_CEPH_PAGELIST_H +#define __FS_CEPH_PAGELIST_H + +#include + +struct ceph_pagelist { + struct list_head head; + void *mapped_tail; + size_t length; + size_t room; +}; + +static inline void ceph_pagelist_init(struct ceph_pagelist *pl) +{ + INIT_LIST_HEAD(&pl->head); + pl->mapped_tail = NULL; + pl->length = 0; + pl->room = 0; +} +extern int ceph_pagelist_release(struct ceph_pagelist *pl); + +extern int ceph_pagelist_append(struct ceph_pagelist *pl, const void *d, size_t l); + +static inline int ceph_pagelist_encode_64(struct ceph_pagelist *pl, u64 v) +{ + __le64 ev = cpu_to_le64(v); + return ceph_pagelist_append(pl, &ev, sizeof(ev)); +} +static inline int ceph_pagelist_encode_32(struct ceph_pagelist *pl, u32 v) +{ + __le32 ev = cpu_to_le32(v); + return ceph_pagelist_append(pl, &ev, sizeof(ev)); +} +static inline int ceph_pagelist_encode_16(struct ceph_pagelist *pl, u16 v) +{ + __le16 ev = cpu_to_le16(v); + return ceph_pagelist_append(pl, &ev, sizeof(ev)); +} +static inline int ceph_pagelist_encode_8(struct ceph_pagelist *pl, u8 v) +{ + return ceph_pagelist_append(pl, &v, 1); +} +static inline int ceph_pagelist_encode_string(struct ceph_pagelist *pl, + char *s, size_t len) +{ + int ret = ceph_pagelist_encode_32(pl, len); + if (ret) + return ret; + if (len) + return ceph_pagelist_append(pl, s, len); + return 0; +} + +#endif diff --git a/include/linux/ceph/rados.h b/include/linux/ceph/rados.h new file mode 100644 index 0000000..6d5247f --- /dev/null +++ b/include/linux/ceph/rados.h @@ -0,0 +1,405 @@ +#ifndef CEPH_RADOS_H +#define CEPH_RADOS_H + +/* + * Data types for the Ceph distributed object storage layer RADOS + * (Reliable Autonomic Distributed Object Store). + */ + +#include "msgr.h" + +/* + * osdmap encoding versions + */ +#define CEPH_OSDMAP_INC_VERSION 5 +#define CEPH_OSDMAP_INC_VERSION_EXT 5 +#define CEPH_OSDMAP_VERSION 5 +#define CEPH_OSDMAP_VERSION_EXT 5 + +/* + * fs id + */ +struct ceph_fsid { + unsigned char fsid[16]; +}; + +static inline int ceph_fsid_compare(const struct ceph_fsid *a, + const struct ceph_fsid *b) +{ + return memcmp(a, b, sizeof(*a)); +} + +/* + * ino, object, etc. + */ +typedef __le64 ceph_snapid_t; +#define CEPH_SNAPDIR ((__u64)(-1)) /* reserved for hidden .snap dir */ +#define CEPH_NOSNAP ((__u64)(-2)) /* "head", "live" revision */ +#define CEPH_MAXSNAP ((__u64)(-3)) /* largest valid snapid */ + +struct ceph_timespec { + __le32 tv_sec; + __le32 tv_nsec; +} __attribute__ ((packed)); + + +/* + * object layout - how objects are mapped into PGs + */ +#define CEPH_OBJECT_LAYOUT_HASH 1 +#define CEPH_OBJECT_LAYOUT_LINEAR 2 +#define CEPH_OBJECT_LAYOUT_HASHINO 3 + +/* + * pg layout -- how PGs are mapped onto (sets of) OSDs + */ +#define CEPH_PG_LAYOUT_CRUSH 0 +#define CEPH_PG_LAYOUT_HASH 1 +#define CEPH_PG_LAYOUT_LINEAR 2 +#define CEPH_PG_LAYOUT_HYBRID 3 + +#define CEPH_PG_MAX_SIZE 16 /* max # osds in a single pg */ + +/* + * placement group. + * we encode this into one __le64. + */ +struct ceph_pg { + __le16 preferred; /* preferred primary osd */ + __le16 ps; /* placement seed */ + __le32 pool; /* object pool */ +} __attribute__ ((packed)); + +/* + * pg_pool is a set of pgs storing a pool of objects + * + * pg_num -- base number of pseudorandomly placed pgs + * + * pgp_num -- effective number when calculating pg placement. this + * is used for pg_num increases. new pgs result in data being "split" + * into new pgs. for this to proceed smoothly, new pgs are intiially + * colocated with their parents; that is, pgp_num doesn't increase + * until the new pgs have successfully split. only _then_ are the new + * pgs placed independently. + * + * lpg_num -- localized pg count (per device). replicas are randomly + * selected. + * + * lpgp_num -- as above. + */ +#define CEPH_PG_TYPE_REP 1 +#define CEPH_PG_TYPE_RAID4 2 +#define CEPH_PG_POOL_VERSION 2 +struct ceph_pg_pool { + __u8 type; /* CEPH_PG_TYPE_* */ + __u8 size; /* number of osds in each pg */ + __u8 crush_ruleset; /* crush placement rule */ + __u8 object_hash; /* hash mapping object name to ps */ + __le32 pg_num, pgp_num; /* number of pg's */ + __le32 lpg_num, lpgp_num; /* number of localized pg's */ + __le32 last_change; /* most recent epoch changed */ + __le64 snap_seq; /* seq for per-pool snapshot */ + __le32 snap_epoch; /* epoch of last snap */ + __le32 num_snaps; + __le32 num_removed_snap_intervals; /* if non-empty, NO per-pool snaps */ + __le64 auid; /* who owns the pg */ +} __attribute__ ((packed)); + +/* + * stable_mod func is used to control number of placement groups. + * similar to straight-up modulo, but produces a stable mapping as b + * increases over time. b is the number of bins, and bmask is the + * containing power of 2 minus 1. + * + * b <= bmask and bmask=(2**n)-1 + * e.g., b=12 -> bmask=15, b=123 -> bmask=127 + */ +static inline int ceph_stable_mod(int x, int b, int bmask) +{ + if ((x & bmask) < b) + return x & bmask; + else + return x & (bmask >> 1); +} + +/* + * object layout - how a given object should be stored. + */ +struct ceph_object_layout { + struct ceph_pg ol_pgid; /* raw pg, with _full_ ps precision. */ + __le32 ol_stripe_unit; /* for per-object parity, if any */ +} __attribute__ ((packed)); + +/* + * compound epoch+version, used by storage layer to serialize mutations + */ +struct ceph_eversion { + __le32 epoch; + __le64 version; +} __attribute__ ((packed)); + +/* + * osd map bits + */ + +/* status bits */ +#define CEPH_OSD_EXISTS 1 +#define CEPH_OSD_UP 2 + +/* osd weights. fixed point value: 0x10000 == 1.0 ("in"), 0 == "out" */ +#define CEPH_OSD_IN 0x10000 +#define CEPH_OSD_OUT 0 + + +/* + * osd map flag bits + */ +#define CEPH_OSDMAP_NEARFULL (1<<0) /* sync writes (near ENOSPC) */ +#define CEPH_OSDMAP_FULL (1<<1) /* no data writes (ENOSPC) */ +#define CEPH_OSDMAP_PAUSERD (1<<2) /* pause all reads */ +#define CEPH_OSDMAP_PAUSEWR (1<<3) /* pause all writes */ +#define CEPH_OSDMAP_PAUSEREC (1<<4) /* pause recovery */ + +/* + * osd ops + */ +#define CEPH_OSD_OP_MODE 0xf000 +#define CEPH_OSD_OP_MODE_RD 0x1000 +#define CEPH_OSD_OP_MODE_WR 0x2000 +#define CEPH_OSD_OP_MODE_RMW 0x3000 +#define CEPH_OSD_OP_MODE_SUB 0x4000 + +#define CEPH_OSD_OP_TYPE 0x0f00 +#define CEPH_OSD_OP_TYPE_LOCK 0x0100 +#define CEPH_OSD_OP_TYPE_DATA 0x0200 +#define CEPH_OSD_OP_TYPE_ATTR 0x0300 +#define CEPH_OSD_OP_TYPE_EXEC 0x0400 +#define CEPH_OSD_OP_TYPE_PG 0x0500 + +enum { + /** data **/ + /* read */ + CEPH_OSD_OP_READ = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 1, + CEPH_OSD_OP_STAT = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 2, + + /* fancy read */ + CEPH_OSD_OP_MASKTRUNC = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 4, + + /* write */ + CEPH_OSD_OP_WRITE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 1, + CEPH_OSD_OP_WRITEFULL = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 2, + CEPH_OSD_OP_TRUNCATE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 3, + CEPH_OSD_OP_ZERO = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 4, + CEPH_OSD_OP_DELETE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 5, + + /* fancy write */ + CEPH_OSD_OP_APPEND = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 6, + CEPH_OSD_OP_STARTSYNC = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 7, + CEPH_OSD_OP_SETTRUNC = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 8, + CEPH_OSD_OP_TRIMTRUNC = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 9, + + CEPH_OSD_OP_TMAPUP = CEPH_OSD_OP_MODE_RMW | CEPH_OSD_OP_TYPE_DATA | 10, + CEPH_OSD_OP_TMAPPUT = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 11, + CEPH_OSD_OP_TMAPGET = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 12, + + CEPH_OSD_OP_CREATE = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 13, + CEPH_OSD_OP_ROLLBACK= CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 14, + + /** attrs **/ + /* read */ + CEPH_OSD_OP_GETXATTR = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 1, + CEPH_OSD_OP_GETXATTRS = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 2, + CEPH_OSD_OP_CMPXATTR = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 3, + + /* write */ + CEPH_OSD_OP_SETXATTR = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 1, + CEPH_OSD_OP_SETXATTRS = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 2, + CEPH_OSD_OP_RESETXATTRS = CEPH_OSD_OP_MODE_WR|CEPH_OSD_OP_TYPE_ATTR | 3, + CEPH_OSD_OP_RMXATTR = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 4, + + /** subop **/ + CEPH_OSD_OP_PULL = CEPH_OSD_OP_MODE_SUB | 1, + CEPH_OSD_OP_PUSH = CEPH_OSD_OP_MODE_SUB | 2, + CEPH_OSD_OP_BALANCEREADS = CEPH_OSD_OP_MODE_SUB | 3, + CEPH_OSD_OP_UNBALANCEREADS = CEPH_OSD_OP_MODE_SUB | 4, + CEPH_OSD_OP_SCRUB = CEPH_OSD_OP_MODE_SUB | 5, + + /** lock **/ + CEPH_OSD_OP_WRLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 1, + CEPH_OSD_OP_WRUNLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 2, + CEPH_OSD_OP_RDLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 3, + CEPH_OSD_OP_RDUNLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 4, + CEPH_OSD_OP_UPLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 5, + CEPH_OSD_OP_DNLOCK = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 6, + + /** exec **/ + CEPH_OSD_OP_CALL = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_EXEC | 1, + + /** pg **/ + CEPH_OSD_OP_PGLS = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_PG | 1, +}; + +static inline int ceph_osd_op_type_lock(int op) +{ + return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_LOCK; +} +static inline int ceph_osd_op_type_data(int op) +{ + return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_DATA; +} +static inline int ceph_osd_op_type_attr(int op) +{ + return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_ATTR; +} +static inline int ceph_osd_op_type_exec(int op) +{ + return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_EXEC; +} +static inline int ceph_osd_op_type_pg(int op) +{ + return (op & CEPH_OSD_OP_TYPE) == CEPH_OSD_OP_TYPE_PG; +} + +static inline int ceph_osd_op_mode_subop(int op) +{ + return (op & CEPH_OSD_OP_MODE) == CEPH_OSD_OP_MODE_SUB; +} +static inline int ceph_osd_op_mode_read(int op) +{ + return (op & CEPH_OSD_OP_MODE) == CEPH_OSD_OP_MODE_RD; +} +static inline int ceph_osd_op_mode_modify(int op) +{ + return (op & CEPH_OSD_OP_MODE) == CEPH_OSD_OP_MODE_WR; +} + +/* + * note that the following tmap stuff is also defined in the ceph librados.h + * any modification here needs to be updated there + */ +#define CEPH_OSD_TMAP_HDR 'h' +#define CEPH_OSD_TMAP_SET 's' +#define CEPH_OSD_TMAP_RM 'r' + +extern const char *ceph_osd_op_name(int op); + + +/* + * osd op flags + * + * An op may be READ, WRITE, or READ|WRITE. + */ +enum { + CEPH_OSD_FLAG_ACK = 1, /* want (or is) "ack" ack */ + CEPH_OSD_FLAG_ONNVRAM = 2, /* want (or is) "onnvram" ack */ + CEPH_OSD_FLAG_ONDISK = 4, /* want (or is) "ondisk" ack */ + CEPH_OSD_FLAG_RETRY = 8, /* resend attempt */ + CEPH_OSD_FLAG_READ = 16, /* op may read */ + CEPH_OSD_FLAG_WRITE = 32, /* op may write */ + CEPH_OSD_FLAG_ORDERSNAP = 64, /* EOLDSNAP if snapc is out of order */ + CEPH_OSD_FLAG_PEERSTAT = 128, /* msg includes osd_peer_stat */ + CEPH_OSD_FLAG_BALANCE_READS = 256, + CEPH_OSD_FLAG_PARALLELEXEC = 512, /* execute op in parallel */ + CEPH_OSD_FLAG_PGOP = 1024, /* pg op, no object */ + CEPH_OSD_FLAG_EXEC = 2048, /* op may exec */ + CEPH_OSD_FLAG_EXEC_PUBLIC = 4096, /* op may exec (public) */ +}; + +enum { + CEPH_OSD_OP_FLAG_EXCL = 1, /* EXCL object create */ +}; + +#define EOLDSNAPC ERESTART /* ORDERSNAP flag set; writer has old snapc*/ +#define EBLACKLISTED ESHUTDOWN /* blacklisted */ + +/* xattr comparison */ +enum { + CEPH_OSD_CMPXATTR_OP_NOP = 0, + CEPH_OSD_CMPXATTR_OP_EQ = 1, + CEPH_OSD_CMPXATTR_OP_NE = 2, + CEPH_OSD_CMPXATTR_OP_GT = 3, + CEPH_OSD_CMPXATTR_OP_GTE = 4, + CEPH_OSD_CMPXATTR_OP_LT = 5, + CEPH_OSD_CMPXATTR_OP_LTE = 6 +}; + +enum { + CEPH_OSD_CMPXATTR_MODE_STRING = 1, + CEPH_OSD_CMPXATTR_MODE_U64 = 2 +}; + +/* + * an individual object operation. each may be accompanied by some data + * payload + */ +struct ceph_osd_op { + __le16 op; /* CEPH_OSD_OP_* */ + __le32 flags; /* CEPH_OSD_FLAG_* */ + union { + struct { + __le64 offset, length; + __le64 truncate_size; + __le32 truncate_seq; + } __attribute__ ((packed)) extent; + struct { + __le32 name_len; + __le32 value_len; + __u8 cmp_op; /* CEPH_OSD_CMPXATTR_OP_* */ + __u8 cmp_mode; /* CEPH_OSD_CMPXATTR_MODE_* */ + } __attribute__ ((packed)) xattr; + struct { + __u8 class_len; + __u8 method_len; + __u8 argc; + __le32 indata_len; + } __attribute__ ((packed)) cls; + struct { + __le64 cookie, count; + } __attribute__ ((packed)) pgls; + struct { + __le64 snapid; + } __attribute__ ((packed)) snap; + }; + __le32 payload_len; +} __attribute__ ((packed)); + +/* + * osd request message header. each request may include multiple + * ceph_osd_op object operations. + */ +struct ceph_osd_request_head { + __le32 client_inc; /* client incarnation */ + struct ceph_object_layout layout; /* pgid */ + __le32 osdmap_epoch; /* client's osdmap epoch */ + + __le32 flags; + + struct ceph_timespec mtime; /* for mutations only */ + struct ceph_eversion reassert_version; /* if we are replaying op */ + + __le32 object_len; /* length of object name */ + + __le64 snapid; /* snapid to read */ + __le64 snap_seq; /* writer's snap context */ + __le32 num_snaps; + + __le16 num_ops; + struct ceph_osd_op ops[]; /* followed by ops[], obj, ticket, snaps */ +} __attribute__ ((packed)); + +struct ceph_osd_reply_head { + __le32 client_inc; /* client incarnation */ + __le32 flags; + struct ceph_object_layout layout; + __le32 osdmap_epoch; + struct ceph_eversion reassert_version; /* for replaying uncommitted */ + + __le32 result; /* result code */ + + __le32 object_len; /* length of object name */ + __le32 num_ops; + struct ceph_osd_op ops[0]; /* ops[], object */ +} __attribute__ ((packed)); + + +#endif diff --git a/include/linux/ceph/types.h b/include/linux/ceph/types.h new file mode 100644 index 0000000..28b35a0 --- /dev/null +++ b/include/linux/ceph/types.h @@ -0,0 +1,29 @@ +#ifndef _FS_CEPH_TYPES_H +#define _FS_CEPH_TYPES_H + +/* needed before including ceph_fs.h */ +#include +#include +#include +#include + +#include "ceph_fs.h" +#include "ceph_frag.h" +#include "ceph_hash.h" + +/* + * Identify inodes by both their ino AND snapshot id (a u64). + */ +struct ceph_vino { + u64 ino; + u64 snap; +}; + + +/* context for the caps reservation mechanism */ +struct ceph_cap_reservation { + int count; +}; + + +#endif diff --git a/include/linux/crush/crush.h b/include/linux/crush/crush.h new file mode 100644 index 0000000..97e435b --- /dev/null +++ b/include/linux/crush/crush.h @@ -0,0 +1,180 @@ +#ifndef CEPH_CRUSH_CRUSH_H +#define CEPH_CRUSH_CRUSH_H + +#include + +/* + * CRUSH is a pseudo-random data distribution algorithm that + * efficiently distributes input values (typically, data objects) + * across a heterogeneous, structured storage cluster. + * + * The algorithm was originally described in detail in this paper + * (although the algorithm has evolved somewhat since then): + * + * http://www.ssrc.ucsc.edu/Papers/weil-sc06.pdf + * + * LGPL2 + */ + + +#define CRUSH_MAGIC 0x00010000ul /* for detecting algorithm revisions */ + + +#define CRUSH_MAX_DEPTH 10 /* max crush hierarchy depth */ +#define CRUSH_MAX_SET 10 /* max size of a mapping result */ + + +/* + * CRUSH uses user-defined "rules" to describe how inputs should be + * mapped to devices. A rule consists of sequence of steps to perform + * to generate the set of output devices. + */ +struct crush_rule_step { + __u32 op; + __s32 arg1; + __s32 arg2; +}; + +/* step op codes */ +enum { + CRUSH_RULE_NOOP = 0, + CRUSH_RULE_TAKE = 1, /* arg1 = value to start with */ + CRUSH_RULE_CHOOSE_FIRSTN = 2, /* arg1 = num items to pick */ + /* arg2 = type */ + CRUSH_RULE_CHOOSE_INDEP = 3, /* same */ + CRUSH_RULE_EMIT = 4, /* no args */ + CRUSH_RULE_CHOOSE_LEAF_FIRSTN = 6, + CRUSH_RULE_CHOOSE_LEAF_INDEP = 7, +}; + +/* + * for specifying choose num (arg1) relative to the max parameter + * passed to do_rule + */ +#define CRUSH_CHOOSE_N 0 +#define CRUSH_CHOOSE_N_MINUS(x) (-(x)) + +/* + * The rule mask is used to describe what the rule is intended for. + * Given a ruleset and size of output set, we search through the + * rule list for a matching rule_mask. + */ +struct crush_rule_mask { + __u8 ruleset; + __u8 type; + __u8 min_size; + __u8 max_size; +}; + +struct crush_rule { + __u32 len; + struct crush_rule_mask mask; + struct crush_rule_step steps[0]; +}; + +#define crush_rule_size(len) (sizeof(struct crush_rule) + \ + (len)*sizeof(struct crush_rule_step)) + + + +/* + * A bucket is a named container of other items (either devices or + * other buckets). Items within a bucket are chosen using one of a + * few different algorithms. The table summarizes how the speed of + * each option measures up against mapping stability when items are + * added or removed. + * + * Bucket Alg Speed Additions Removals + * ------------------------------------------------ + * uniform O(1) poor poor + * list O(n) optimal poor + * tree O(log n) good good + * straw O(n) optimal optimal + */ +enum { + CRUSH_BUCKET_UNIFORM = 1, + CRUSH_BUCKET_LIST = 2, + CRUSH_BUCKET_TREE = 3, + CRUSH_BUCKET_STRAW = 4 +}; +extern const char *crush_bucket_alg_name(int alg); + +struct crush_bucket { + __s32 id; /* this'll be negative */ + __u16 type; /* non-zero; type=0 is reserved for devices */ + __u8 alg; /* one of CRUSH_BUCKET_* */ + __u8 hash; /* which hash function to use, CRUSH_HASH_* */ + __u32 weight; /* 16-bit fixed point */ + __u32 size; /* num items */ + __s32 *items; + + /* + * cached random permutation: used for uniform bucket and for + * the linear search fallback for the other bucket types. + */ + __u32 perm_x; /* @x for which *perm is defined */ + __u32 perm_n; /* num elements of *perm that are permuted/defined */ + __u32 *perm; +}; + +struct crush_bucket_uniform { + struct crush_bucket h; + __u32 item_weight; /* 16-bit fixed point; all items equally weighted */ +}; + +struct crush_bucket_list { + struct crush_bucket h; + __u32 *item_weights; /* 16-bit fixed point */ + __u32 *sum_weights; /* 16-bit fixed point. element i is sum + of weights 0..i, inclusive */ +}; + +struct crush_bucket_tree { + struct crush_bucket h; /* note: h.size is _tree_ size, not number of + actual items */ + __u8 num_nodes; + __u32 *node_weights; +}; + +struct crush_bucket_straw { + struct crush_bucket h; + __u32 *item_weights; /* 16-bit fixed point */ + __u32 *straws; /* 16-bit fixed point */ +}; + + + +/* + * CRUSH map includes all buckets, rules, etc. + */ +struct crush_map { + struct crush_bucket **buckets; + struct crush_rule **rules; + + /* + * Parent pointers to identify the parent bucket a device or + * bucket in the hierarchy. If an item appears more than + * once, this is the _last_ time it appeared (where buckets + * are processed in bucket id order, from -1 on down to + * -max_buckets. + */ + __u32 *bucket_parents; + __u32 *device_parents; + + __s32 max_buckets; + __u32 max_rules; + __s32 max_devices; +}; + + +/* crush.c */ +extern int crush_get_bucket_item_weight(struct crush_bucket *b, int pos); +extern void crush_calc_parents(struct crush_map *map); +extern void crush_destroy_bucket_uniform(struct crush_bucket_uniform *b); +extern void crush_destroy_bucket_list(struct crush_bucket_list *b); +extern void crush_destroy_bucket_tree(struct crush_bucket_tree *b); +extern void crush_destroy_bucket_straw(struct crush_bucket_straw *b); +extern void crush_destroy_bucket(struct crush_bucket *b); +extern void crush_destroy(struct crush_map *map); + +#endif diff --git a/include/linux/crush/hash.h b/include/linux/crush/hash.h new file mode 100644 index 0000000..91e8842 --- /dev/null +++ b/include/linux/crush/hash.h @@ -0,0 +1,17 @@ +#ifndef CEPH_CRUSH_HASH_H +#define CEPH_CRUSH_HASH_H + +#define CRUSH_HASH_RJENKINS1 0 + +#define CRUSH_HASH_DEFAULT CRUSH_HASH_RJENKINS1 + +extern const char *crush_hash_name(int type); + +extern __u32 crush_hash32(int type, __u32 a); +extern __u32 crush_hash32_2(int type, __u32 a, __u32 b); +extern __u32 crush_hash32_3(int type, __u32 a, __u32 b, __u32 c); +extern __u32 crush_hash32_4(int type, __u32 a, __u32 b, __u32 c, __u32 d); +extern __u32 crush_hash32_5(int type, __u32 a, __u32 b, __u32 c, __u32 d, + __u32 e); + +#endif diff --git a/include/linux/crush/mapper.h b/include/linux/crush/mapper.h new file mode 100644 index 0000000..c46b99c --- /dev/null +++ b/include/linux/crush/mapper.h @@ -0,0 +1,20 @@ +#ifndef CEPH_CRUSH_MAPPER_H +#define CEPH_CRUSH_MAPPER_H + +/* + * CRUSH functions for find rules and then mapping an input to an + * output set. + * + * LGPL2 + */ + +#include "crush.h" + +extern int crush_find_rule(struct crush_map *map, int pool, int type, int size); +extern int crush_do_rule(struct crush_map *map, + int ruleno, + int x, int *result, int result_max, + int forcefeed, /* -1 for none */ + __u32 *weights); + +#endif diff --git a/net/Kconfig b/net/Kconfig index e926884..55fd82e 100644 --- a/net/Kconfig +++ b/net/Kconfig @@ -293,6 +293,7 @@ source "net/wimax/Kconfig" source "net/rfkill/Kconfig" source "net/9p/Kconfig" source "net/caif/Kconfig" +source "net/ceph/Kconfig" endif # if NET diff --git a/net/Makefile b/net/Makefile index ea60fbc..6b7bfd7 100644 --- a/net/Makefile +++ b/net/Makefile @@ -68,3 +68,4 @@ obj-$(CONFIG_SYSCTL) += sysctl_net.o endif obj-$(CONFIG_WIMAX) += wimax/ obj-$(CONFIG_DNS_RESOLVER) += dns_resolver/ +obj-$(CONFIG_CEPH_LIB) += ceph/ diff --git a/net/ceph/Kconfig b/net/ceph/Kconfig new file mode 100644 index 0000000..ad42404 --- /dev/null +++ b/net/ceph/Kconfig @@ -0,0 +1,28 @@ +config CEPH_LIB + tristate "Ceph core library (EXPERIMENTAL)" + depends on INET && EXPERIMENTAL + select LIBCRC32C + select CRYPTO_AES + select CRYPTO + default n + help + Choose Y or M here to include cephlib, which provides the + common functionality to both the Ceph filesystem and + to the rados block device (rbd). + + More information at http://ceph.newdream.net/. + + If unsure, say N. + +config CEPH_LIB_PRETTYDEBUG + bool "Include file:line in ceph debug output" + depends on CEPH_LIB + default n + help + If you say Y here, debug output will include a filename and + line to aid debugging. This increases kernel size and slows + execution slightly when debug call sites are enabled (e.g., + via CONFIG_DYNAMIC_DEBUG). + + If unsure, say N. + diff --git a/net/ceph/Makefile b/net/ceph/Makefile new file mode 100644 index 0000000..aab1cab --- /dev/null +++ b/net/ceph/Makefile @@ -0,0 +1,37 @@ +# +# Makefile for CEPH filesystem. +# + +ifneq ($(KERNELRELEASE),) + +obj-$(CONFIG_CEPH_LIB) += libceph.o + +libceph-objs := ceph_common.o messenger.o msgpool.o buffer.o pagelist.o \ + mon_client.o \ + osd_client.o osdmap.o crush/crush.o crush/mapper.o crush/hash.o \ + debugfs.o \ + auth.o auth_none.o \ + crypto.o armor.o \ + auth_x.o \ + ceph_fs.o ceph_strings.o ceph_hash.o \ + pagevec.o + +else +#Otherwise we were called directly from the command +# line; invoke the kernel build system. + +KERNELDIR ?= /lib/modules/$(shell uname -r)/build +PWD := $(shell pwd) + +default: all + +all: + $(MAKE) -C $(KERNELDIR) M=$(PWD) CONFIG_CEPH_LIB=m modules + +modules_install: + $(MAKE) -C $(KERNELDIR) M=$(PWD) CONFIG_CEPH_LIB=m modules_install + +clean: + $(MAKE) -C $(KERNELDIR) M=$(PWD) clean + +endif diff --git a/net/ceph/armor.c b/net/ceph/armor.c new file mode 100644 index 0000000..eb2a666 --- /dev/null +++ b/net/ceph/armor.c @@ -0,0 +1,103 @@ + +#include + +int ceph_armor(char *dst, const char *src, const char *end); +int ceph_unarmor(char *dst, const char *src, const char *end); + +/* + * base64 encode/decode. + */ + +static const char *pem_key = + "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; + +static int encode_bits(int c) +{ + return pem_key[c]; +} + +static int decode_bits(char c) +{ + if (c >= 'A' && c <= 'Z') + return c - 'A'; + if (c >= 'a' && c <= 'z') + return c - 'a' + 26; + if (c >= '0' && c <= '9') + return c - '0' + 52; + if (c == '+') + return 62; + if (c == '/') + return 63; + if (c == '=') + return 0; /* just non-negative, please */ + return -EINVAL; +} + +int ceph_armor(char *dst, const char *src, const char *end) +{ + int olen = 0; + int line = 0; + + while (src < end) { + unsigned char a, b, c; + + a = *src++; + *dst++ = encode_bits(a >> 2); + if (src < end) { + b = *src++; + *dst++ = encode_bits(((a & 3) << 4) | (b >> 4)); + if (src < end) { + c = *src++; + *dst++ = encode_bits(((b & 15) << 2) | + (c >> 6)); + *dst++ = encode_bits(c & 63); + } else { + *dst++ = encode_bits((b & 15) << 2); + *dst++ = '='; + } + } else { + *dst++ = encode_bits(((a & 3) << 4)); + *dst++ = '='; + *dst++ = '='; + } + olen += 4; + line += 4; + if (line == 64) { + line = 0; + *(dst++) = '\n'; + olen++; + } + } + return olen; +} + +int ceph_unarmor(char *dst, const char *src, const char *end) +{ + int olen = 0; + + while (src < end) { + int a, b, c, d; + + if (src < end && src[0] == '\n') + src++; + if (src + 4 > end) + return -EINVAL; + a = decode_bits(src[0]); + b = decode_bits(src[1]); + c = decode_bits(src[2]); + d = decode_bits(src[3]); + if (a < 0 || b < 0 || c < 0 || d < 0) + return -EINVAL; + + *dst++ = (a << 2) | (b >> 4); + if (src[2] == '=') + return olen + 1; + *dst++ = ((b & 15) << 4) | (c >> 2); + if (src[3] == '=') + return olen + 2; + *dst++ = ((c & 3) << 6) | d; + olen += 3; + src += 4; + } + return olen; +} diff --git a/net/ceph/auth.c b/net/ceph/auth.c new file mode 100644 index 0000000..549c1f4 --- /dev/null +++ b/net/ceph/auth.c @@ -0,0 +1,259 @@ +#include + +#include +#include +#include + +#include +#include +#include +#include +#include "auth_none.h" +#include "auth_x.h" + + +/* + * get protocol handler + */ +static u32 supported_protocols[] = { + CEPH_AUTH_NONE, + CEPH_AUTH_CEPHX +}; + +static int ceph_auth_init_protocol(struct ceph_auth_client *ac, int protocol) +{ + switch (protocol) { + case CEPH_AUTH_NONE: + return ceph_auth_none_init(ac); + case CEPH_AUTH_CEPHX: + return ceph_x_init(ac); + default: + return -ENOENT; + } +} + +/* + * setup, teardown. + */ +struct ceph_auth_client *ceph_auth_init(const char *name, const char *secret) +{ + struct ceph_auth_client *ac; + int ret; + + dout("auth_init name '%s' secret '%s'\n", name, secret); + + ret = -ENOMEM; + ac = kzalloc(sizeof(*ac), GFP_NOFS); + if (!ac) + goto out; + + ac->negotiating = true; + if (name) + ac->name = name; + else + ac->name = CEPH_AUTH_NAME_DEFAULT; + dout("auth_init name %s secret %s\n", ac->name, secret); + ac->secret = secret; + return ac; + +out: + return ERR_PTR(ret); +} + +void ceph_auth_destroy(struct ceph_auth_client *ac) +{ + dout("auth_destroy %p\n", ac); + if (ac->ops) + ac->ops->destroy(ac); + kfree(ac); +} + +/* + * Reset occurs when reconnecting to the monitor. + */ +void ceph_auth_reset(struct ceph_auth_client *ac) +{ + dout("auth_reset %p\n", ac); + if (ac->ops && !ac->negotiating) + ac->ops->reset(ac); + ac->negotiating = true; +} + +int ceph_entity_name_encode(const char *name, void **p, void *end) +{ + int len = strlen(name); + + if (*p + 2*sizeof(u32) + len > end) + return -ERANGE; + ceph_encode_32(p, CEPH_ENTITY_TYPE_CLIENT); + ceph_encode_32(p, len); + ceph_encode_copy(p, name, len); + return 0; +} + +/* + * Initiate protocol negotiation with monitor. Include entity name + * and list supported protocols. + */ +int ceph_auth_build_hello(struct ceph_auth_client *ac, void *buf, size_t len) +{ + struct ceph_mon_request_header *monhdr = buf; + void *p = monhdr + 1, *end = buf + len, *lenp; + int i, num; + int ret; + + dout("auth_build_hello\n"); + monhdr->have_version = 0; + monhdr->session_mon = cpu_to_le16(-1); + monhdr->session_mon_tid = 0; + + ceph_encode_32(&p, 0); /* no protocol, yet */ + + lenp = p; + p += sizeof(u32); + + ceph_decode_need(&p, end, 1 + sizeof(u32), bad); + ceph_encode_8(&p, 1); + num = ARRAY_SIZE(supported_protocols); + ceph_encode_32(&p, num); + ceph_decode_need(&p, end, num * sizeof(u32), bad); + for (i = 0; i < num; i++) + ceph_encode_32(&p, supported_protocols[i]); + + ret = ceph_entity_name_encode(ac->name, &p, end); + if (ret < 0) + return ret; + ceph_decode_need(&p, end, sizeof(u64), bad); + ceph_encode_64(&p, ac->global_id); + + ceph_encode_32(&lenp, p - lenp - sizeof(u32)); + return p - buf; + +bad: + return -ERANGE; +} + +static int ceph_build_auth_request(struct ceph_auth_client *ac, + void *msg_buf, size_t msg_len) +{ + struct ceph_mon_request_header *monhdr = msg_buf; + void *p = monhdr + 1; + void *end = msg_buf + msg_len; + int ret; + + monhdr->have_version = 0; + monhdr->session_mon = cpu_to_le16(-1); + monhdr->session_mon_tid = 0; + + ceph_encode_32(&p, ac->protocol); + + ret = ac->ops->build_request(ac, p + sizeof(u32), end); + if (ret < 0) { + pr_err("error %d building auth method %s request\n", ret, + ac->ops->name); + return ret; + } + dout(" built request %d bytes\n", ret); + ceph_encode_32(&p, ret); + return p + ret - msg_buf; +} + +/* + * Handle auth message from monitor. + */ +int ceph_handle_auth_reply(struct ceph_auth_client *ac, + void *buf, size_t len, + void *reply_buf, size_t reply_len) +{ + void *p = buf; + void *end = buf + len; + int protocol; + s32 result; + u64 global_id; + void *payload, *payload_end; + int payload_len; + char *result_msg; + int result_msg_len; + int ret = -EINVAL; + + dout("handle_auth_reply %p %p\n", p, end); + ceph_decode_need(&p, end, sizeof(u32) * 3 + sizeof(u64), bad); + protocol = ceph_decode_32(&p); + result = ceph_decode_32(&p); + global_id = ceph_decode_64(&p); + payload_len = ceph_decode_32(&p); + payload = p; + p += payload_len; + ceph_decode_need(&p, end, sizeof(u32), bad); + result_msg_len = ceph_decode_32(&p); + result_msg = p; + p += result_msg_len; + if (p != end) + goto bad; + + dout(" result %d '%.*s' gid %llu len %d\n", result, result_msg_len, + result_msg, global_id, payload_len); + + payload_end = payload + payload_len; + + if (global_id && ac->global_id != global_id) { + dout(" set global_id %lld -> %lld\n", ac->global_id, global_id); + ac->global_id = global_id; + } + + if (ac->negotiating) { + /* server does not support our protocols? */ + if (!protocol && result < 0) { + ret = result; + goto out; + } + /* set up (new) protocol handler? */ + if (ac->protocol && ac->protocol != protocol) { + ac->ops->destroy(ac); + ac->protocol = 0; + ac->ops = NULL; + } + if (ac->protocol != protocol) { + ret = ceph_auth_init_protocol(ac, protocol); + if (ret) { + pr_err("error %d on auth protocol %d init\n", + ret, protocol); + goto out; + } + } + + ac->negotiating = false; + } + + ret = ac->ops->handle_reply(ac, result, payload, payload_end); + if (ret == -EAGAIN) { + return ceph_build_auth_request(ac, reply_buf, reply_len); + } else if (ret) { + pr_err("auth method '%s' error %d\n", ac->ops->name, ret); + return ret; + } + return 0; + +bad: + pr_err("failed to decode auth msg\n"); +out: + return ret; +} + +int ceph_build_auth(struct ceph_auth_client *ac, + void *msg_buf, size_t msg_len) +{ + if (!ac->protocol) + return ceph_auth_build_hello(ac, msg_buf, msg_len); + BUG_ON(!ac->ops); + if (ac->ops->should_authenticate(ac)) + return ceph_build_auth_request(ac, msg_buf, msg_len); + return 0; +} + +int ceph_auth_is_authenticated(struct ceph_auth_client *ac) +{ + if (!ac->ops) + return 0; + return ac->ops->is_authenticated(ac); +} diff --git a/net/ceph/auth_none.c b/net/ceph/auth_none.c new file mode 100644 index 0000000..214c2bb --- /dev/null +++ b/net/ceph/auth_none.c @@ -0,0 +1,132 @@ + +#include + +#include +#include +#include +#include + +#include +#include + +#include "auth_none.h" + +static void reset(struct ceph_auth_client *ac) +{ + struct ceph_auth_none_info *xi = ac->private; + + xi->starting = true; + xi->built_authorizer = false; +} + +static void destroy(struct ceph_auth_client *ac) +{ + kfree(ac->private); + ac->private = NULL; +} + +static int is_authenticated(struct ceph_auth_client *ac) +{ + struct ceph_auth_none_info *xi = ac->private; + + return !xi->starting; +} + +static int should_authenticate(struct ceph_auth_client *ac) +{ + struct ceph_auth_none_info *xi = ac->private; + + return xi->starting; +} + +/* + * the generic auth code decode the global_id, and we carry no actual + * authenticate state, so nothing happens here. + */ +static int handle_reply(struct ceph_auth_client *ac, int result, + void *buf, void *end) +{ + struct ceph_auth_none_info *xi = ac->private; + + xi->starting = false; + return result; +} + +/* + * build an 'authorizer' with our entity_name and global_id. we can + * reuse a single static copy since it is identical for all services + * we connect to. + */ +static int ceph_auth_none_create_authorizer( + struct ceph_auth_client *ac, int peer_type, + struct ceph_authorizer **a, + void **buf, size_t *len, + void **reply_buf, size_t *reply_len) +{ + struct ceph_auth_none_info *ai = ac->private; + struct ceph_none_authorizer *au = &ai->au; + void *p, *end; + int ret; + + if (!ai->built_authorizer) { + p = au->buf; + end = p + sizeof(au->buf); + ceph_encode_8(&p, 1); + ret = ceph_entity_name_encode(ac->name, &p, end - 8); + if (ret < 0) + goto bad; + ceph_decode_need(&p, end, sizeof(u64), bad2); + ceph_encode_64(&p, ac->global_id); + au->buf_len = p - (void *)au->buf; + ai->built_authorizer = true; + dout("built authorizer len %d\n", au->buf_len); + } + + *a = (struct ceph_authorizer *)au; + *buf = au->buf; + *len = au->buf_len; + *reply_buf = au->reply_buf; + *reply_len = sizeof(au->reply_buf); + return 0; + +bad2: + ret = -ERANGE; +bad: + return ret; +} + +static void ceph_auth_none_destroy_authorizer(struct ceph_auth_client *ac, + struct ceph_authorizer *a) +{ + /* nothing to do */ +} + +static const struct ceph_auth_client_ops ceph_auth_none_ops = { + .name = "none", + .reset = reset, + .destroy = destroy, + .is_authenticated = is_authenticated, + .should_authenticate = should_authenticate, + .handle_reply = handle_reply, + .create_authorizer = ceph_auth_none_create_authorizer, + .destroy_authorizer = ceph_auth_none_destroy_authorizer, +}; + +int ceph_auth_none_init(struct ceph_auth_client *ac) +{ + struct ceph_auth_none_info *xi; + + dout("ceph_auth_none_init %p\n", ac); + xi = kzalloc(sizeof(*xi), GFP_NOFS); + if (!xi) + return -ENOMEM; + + xi->starting = true; + xi->built_authorizer = false; + + ac->protocol = CEPH_AUTH_NONE; + ac->private = xi; + ac->ops = &ceph_auth_none_ops; + return 0; +} + diff --git a/net/ceph/auth_none.h b/net/ceph/auth_none.h new file mode 100644 index 0000000..ed7d088 --- /dev/null +++ b/net/ceph/auth_none.h @@ -0,0 +1,29 @@ +#ifndef _FS_CEPH_AUTH_NONE_H +#define _FS_CEPH_AUTH_NONE_H + +#include +#include + +/* + * null security mode. + * + * we use a single static authorizer that simply encodes our entity name + * and global id. + */ + +struct ceph_none_authorizer { + char buf[128]; + int buf_len; + char reply_buf[0]; +}; + +struct ceph_auth_none_info { + bool starting; + bool built_authorizer; + struct ceph_none_authorizer au; /* we only need one; it's static */ +}; + +extern int ceph_auth_none_init(struct ceph_auth_client *ac); + +#endif + diff --git a/net/ceph/auth_x.c b/net/ceph/auth_x.c new file mode 100644 index 0000000..7fd5dfc --- /dev/null +++ b/net/ceph/auth_x.c @@ -0,0 +1,688 @@ + +#include + +#include +#include +#include +#include + +#include +#include + +#include "crypto.h" +#include "auth_x.h" +#include "auth_x_protocol.h" + +#define TEMP_TICKET_BUF_LEN 256 + +static void ceph_x_validate_tickets(struct ceph_auth_client *ac, int *pneed); + +static int ceph_x_is_authenticated(struct ceph_auth_client *ac) +{ + struct ceph_x_info *xi = ac->private; + int need; + + ceph_x_validate_tickets(ac, &need); + dout("ceph_x_is_authenticated want=%d need=%d have=%d\n", + ac->want_keys, need, xi->have_keys); + return (ac->want_keys & xi->have_keys) == ac->want_keys; +} + +static int ceph_x_should_authenticate(struct ceph_auth_client *ac) +{ + struct ceph_x_info *xi = ac->private; + int need; + + ceph_x_validate_tickets(ac, &need); + dout("ceph_x_should_authenticate want=%d need=%d have=%d\n", + ac->want_keys, need, xi->have_keys); + return need != 0; +} + +static int ceph_x_encrypt_buflen(int ilen) +{ + return sizeof(struct ceph_x_encrypt_header) + ilen + 16 + + sizeof(u32); +} + +static int ceph_x_encrypt(struct ceph_crypto_key *secret, + void *ibuf, int ilen, void *obuf, size_t olen) +{ + struct ceph_x_encrypt_header head = { + .struct_v = 1, + .magic = cpu_to_le64(CEPHX_ENC_MAGIC) + }; + size_t len = olen - sizeof(u32); + int ret; + + ret = ceph_encrypt2(secret, obuf + sizeof(u32), &len, + &head, sizeof(head), ibuf, ilen); + if (ret) + return ret; + ceph_encode_32(&obuf, len); + return len + sizeof(u32); +} + +static int ceph_x_decrypt(struct ceph_crypto_key *secret, + void **p, void *end, void *obuf, size_t olen) +{ + struct ceph_x_encrypt_header head; + size_t head_len = sizeof(head); + int len, ret; + + len = ceph_decode_32(p); + if (*p + len > end) + return -EINVAL; + + dout("ceph_x_decrypt len %d\n", len); + ret = ceph_decrypt2(secret, &head, &head_len, obuf, &olen, + *p, len); + if (ret) + return ret; + if (head.struct_v != 1 || le64_to_cpu(head.magic) != CEPHX_ENC_MAGIC) + return -EPERM; + *p += len; + return olen; +} + +/* + * get existing (or insert new) ticket handler + */ +static struct ceph_x_ticket_handler * +get_ticket_handler(struct ceph_auth_client *ac, int service) +{ + struct ceph_x_ticket_handler *th; + struct ceph_x_info *xi = ac->private; + struct rb_node *parent = NULL, **p = &xi->ticket_handlers.rb_node; + + while (*p) { + parent = *p; + th = rb_entry(parent, struct ceph_x_ticket_handler, node); + if (service < th->service) + p = &(*p)->rb_left; + else if (service > th->service) + p = &(*p)->rb_right; + else + return th; + } + + /* add it */ + th = kzalloc(sizeof(*th), GFP_NOFS); + if (!th) + return ERR_PTR(-ENOMEM); + th->service = service; + rb_link_node(&th->node, parent, p); + rb_insert_color(&th->node, &xi->ticket_handlers); + return th; +} + +static void remove_ticket_handler(struct ceph_auth_client *ac, + struct ceph_x_ticket_handler *th) +{ + struct ceph_x_info *xi = ac->private; + + dout("remove_ticket_handler %p %d\n", th, th->service); + rb_erase(&th->node, &xi->ticket_handlers); + ceph_crypto_key_destroy(&th->session_key); + if (th->ticket_blob) + ceph_buffer_put(th->ticket_blob); + kfree(th); +} + +static int ceph_x_proc_ticket_reply(struct ceph_auth_client *ac, + struct ceph_crypto_key *secret, + void *buf, void *end) +{ + struct ceph_x_info *xi = ac->private; + int num; + void *p = buf; + int ret; + char *dbuf; + char *ticket_buf; + u8 reply_struct_v; + + dbuf = kmalloc(TEMP_TICKET_BUF_LEN, GFP_NOFS); + if (!dbuf) + return -ENOMEM; + + ret = -ENOMEM; + ticket_buf = kmalloc(TEMP_TICKET_BUF_LEN, GFP_NOFS); + if (!ticket_buf) + goto out_dbuf; + + ceph_decode_need(&p, end, 1 + sizeof(u32), bad); + reply_struct_v = ceph_decode_8(&p); + if (reply_struct_v != 1) + goto bad; + num = ceph_decode_32(&p); + dout("%d tickets\n", num); + while (num--) { + int type; + u8 tkt_struct_v, blob_struct_v; + struct ceph_x_ticket_handler *th; + void *dp, *dend; + int dlen; + char is_enc; + struct timespec validity; + struct ceph_crypto_key old_key; + void *tp, *tpend; + struct ceph_timespec new_validity; + struct ceph_crypto_key new_session_key; + struct ceph_buffer *new_ticket_blob; + unsigned long new_expires, new_renew_after; + u64 new_secret_id; + + ceph_decode_need(&p, end, sizeof(u32) + 1, bad); + + type = ceph_decode_32(&p); + dout(" ticket type %d %s\n", type, ceph_entity_type_name(type)); + + tkt_struct_v = ceph_decode_8(&p); + if (tkt_struct_v != 1) + goto bad; + + th = get_ticket_handler(ac, type); + if (IS_ERR(th)) { + ret = PTR_ERR(th); + goto out; + } + + /* blob for me */ + dlen = ceph_x_decrypt(secret, &p, end, dbuf, + TEMP_TICKET_BUF_LEN); + if (dlen <= 0) { + ret = dlen; + goto out; + } + dout(" decrypted %d bytes\n", dlen); + dend = dbuf + dlen; + dp = dbuf; + + tkt_struct_v = ceph_decode_8(&dp); + if (tkt_struct_v != 1) + goto bad; + + memcpy(&old_key, &th->session_key, sizeof(old_key)); + ret = ceph_crypto_key_decode(&new_session_key, &dp, dend); + if (ret) + goto out; + + ceph_decode_copy(&dp, &new_validity, sizeof(new_validity)); + ceph_decode_timespec(&validity, &new_validity); + new_expires = get_seconds() + validity.tv_sec; + new_renew_after = new_expires - (validity.tv_sec / 4); + dout(" expires=%lu renew_after=%lu\n", new_expires, + new_renew_after); + + /* ticket blob for service */ + ceph_decode_8_safe(&p, end, is_enc, bad); + tp = ticket_buf; + if (is_enc) { + /* encrypted */ + dout(" encrypted ticket\n"); + dlen = ceph_x_decrypt(&old_key, &p, end, ticket_buf, + TEMP_TICKET_BUF_LEN); + if (dlen < 0) { + ret = dlen; + goto out; + } + dlen = ceph_decode_32(&tp); + } else { + /* unencrypted */ + ceph_decode_32_safe(&p, end, dlen, bad); + ceph_decode_need(&p, end, dlen, bad); + ceph_decode_copy(&p, ticket_buf, dlen); + } + tpend = tp + dlen; + dout(" ticket blob is %d bytes\n", dlen); + ceph_decode_need(&tp, tpend, 1 + sizeof(u64), bad); + blob_struct_v = ceph_decode_8(&tp); + new_secret_id = ceph_decode_64(&tp); + ret = ceph_decode_buffer(&new_ticket_blob, &tp, tpend); + if (ret) + goto out; + + /* all is well, update our ticket */ + ceph_crypto_key_destroy(&th->session_key); + if (th->ticket_blob) + ceph_buffer_put(th->ticket_blob); + th->session_key = new_session_key; + th->ticket_blob = new_ticket_blob; + th->validity = new_validity; + th->secret_id = new_secret_id; + th->expires = new_expires; + th->renew_after = new_renew_after; + dout(" got ticket service %d (%s) secret_id %lld len %d\n", + type, ceph_entity_type_name(type), th->secret_id, + (int)th->ticket_blob->vec.iov_len); + xi->have_keys |= th->service; + } + + ret = 0; +out: + kfree(ticket_buf); +out_dbuf: + kfree(dbuf); + return ret; + +bad: + ret = -EINVAL; + goto out; +} + +static int ceph_x_build_authorizer(struct ceph_auth_client *ac, + struct ceph_x_ticket_handler *th, + struct ceph_x_authorizer *au) +{ + int maxlen; + struct ceph_x_authorize_a *msg_a; + struct ceph_x_authorize_b msg_b; + void *p, *end; + int ret; + int ticket_blob_len = + (th->ticket_blob ? th->ticket_blob->vec.iov_len : 0); + + dout("build_authorizer for %s %p\n", + ceph_entity_type_name(th->service), au); + + maxlen = sizeof(*msg_a) + sizeof(msg_b) + + ceph_x_encrypt_buflen(ticket_blob_len); + dout(" need len %d\n", maxlen); + if (au->buf && au->buf->alloc_len < maxlen) { + ceph_buffer_put(au->buf); + au->buf = NULL; + } + if (!au->buf) { + au->buf = ceph_buffer_new(maxlen, GFP_NOFS); + if (!au->buf) + return -ENOMEM; + } + au->service = th->service; + + msg_a = au->buf->vec.iov_base; + msg_a->struct_v = 1; + msg_a->global_id = cpu_to_le64(ac->global_id); + msg_a->service_id = cpu_to_le32(th->service); + msg_a->ticket_blob.struct_v = 1; + msg_a->ticket_blob.secret_id = cpu_to_le64(th->secret_id); + msg_a->ticket_blob.blob_len = cpu_to_le32(ticket_blob_len); + if (ticket_blob_len) { + memcpy(msg_a->ticket_blob.blob, th->ticket_blob->vec.iov_base, + th->ticket_blob->vec.iov_len); + } + dout(" th %p secret_id %lld %lld\n", th, th->secret_id, + le64_to_cpu(msg_a->ticket_blob.secret_id)); + + p = msg_a + 1; + p += ticket_blob_len; + end = au->buf->vec.iov_base + au->buf->vec.iov_len; + + get_random_bytes(&au->nonce, sizeof(au->nonce)); + msg_b.struct_v = 1; + msg_b.nonce = cpu_to_le64(au->nonce); + ret = ceph_x_encrypt(&th->session_key, &msg_b, sizeof(msg_b), + p, end - p); + if (ret < 0) + goto out_buf; + p += ret; + au->buf->vec.iov_len = p - au->buf->vec.iov_base; + dout(" built authorizer nonce %llx len %d\n", au->nonce, + (int)au->buf->vec.iov_len); + BUG_ON(au->buf->vec.iov_len > maxlen); + return 0; + +out_buf: + ceph_buffer_put(au->buf); + au->buf = NULL; + return ret; +} + +static int ceph_x_encode_ticket(struct ceph_x_ticket_handler *th, + void **p, void *end) +{ + ceph_decode_need(p, end, 1 + sizeof(u64), bad); + ceph_encode_8(p, 1); + ceph_encode_64(p, th->secret_id); + if (th->ticket_blob) { + const char *buf = th->ticket_blob->vec.iov_base; + u32 len = th->ticket_blob->vec.iov_len; + + ceph_encode_32_safe(p, end, len, bad); + ceph_encode_copy_safe(p, end, buf, len, bad); + } else { + ceph_encode_32_safe(p, end, 0, bad); + } + + return 0; +bad: + return -ERANGE; +} + +static void ceph_x_validate_tickets(struct ceph_auth_client *ac, int *pneed) +{ + int want = ac->want_keys; + struct ceph_x_info *xi = ac->private; + int service; + + *pneed = ac->want_keys & ~(xi->have_keys); + + for (service = 1; service <= want; service <<= 1) { + struct ceph_x_ticket_handler *th; + + if (!(ac->want_keys & service)) + continue; + + if (*pneed & service) + continue; + + th = get_ticket_handler(ac, service); + + if (IS_ERR(th)) { + *pneed |= service; + continue; + } + + if (get_seconds() >= th->renew_after) + *pneed |= service; + if (get_seconds() >= th->expires) + xi->have_keys &= ~service; + } +} + + +static int ceph_x_build_request(struct ceph_auth_client *ac, + void *buf, void *end) +{ + struct ceph_x_info *xi = ac->private; + int need; + struct ceph_x_request_header *head = buf; + int ret; + struct ceph_x_ticket_handler *th = + get_ticket_handler(ac, CEPH_ENTITY_TYPE_AUTH); + + if (IS_ERR(th)) + return PTR_ERR(th); + + ceph_x_validate_tickets(ac, &need); + + dout("build_request want %x have %x need %x\n", + ac->want_keys, xi->have_keys, need); + + if (need & CEPH_ENTITY_TYPE_AUTH) { + struct ceph_x_authenticate *auth = (void *)(head + 1); + void *p = auth + 1; + struct ceph_x_challenge_blob tmp; + char tmp_enc[40]; + u64 *u; + + if (p > end) + return -ERANGE; + + dout(" get_auth_session_key\n"); + head->op = cpu_to_le16(CEPHX_GET_AUTH_SESSION_KEY); + + /* encrypt and hash */ + get_random_bytes(&auth->client_challenge, sizeof(u64)); + tmp.client_challenge = auth->client_challenge; + tmp.server_challenge = cpu_to_le64(xi->server_challenge); + ret = ceph_x_encrypt(&xi->secret, &tmp, sizeof(tmp), + tmp_enc, sizeof(tmp_enc)); + if (ret < 0) + return ret; + + auth->struct_v = 1; + auth->key = 0; + for (u = (u64 *)tmp_enc; u + 1 <= (u64 *)(tmp_enc + ret); u++) + auth->key ^= *(__le64 *)u; + dout(" server_challenge %llx client_challenge %llx key %llx\n", + xi->server_challenge, le64_to_cpu(auth->client_challenge), + le64_to_cpu(auth->key)); + + /* now encode the old ticket if exists */ + ret = ceph_x_encode_ticket(th, &p, end); + if (ret < 0) + return ret; + + return p - buf; + } + + if (need) { + void *p = head + 1; + struct ceph_x_service_ticket_request *req; + + if (p > end) + return -ERANGE; + head->op = cpu_to_le16(CEPHX_GET_PRINCIPAL_SESSION_KEY); + + ret = ceph_x_build_authorizer(ac, th, &xi->auth_authorizer); + if (ret) + return ret; + ceph_encode_copy(&p, xi->auth_authorizer.buf->vec.iov_base, + xi->auth_authorizer.buf->vec.iov_len); + + req = p; + req->keys = cpu_to_le32(need); + p += sizeof(*req); + return p - buf; + } + + return 0; +} + +static int ceph_x_handle_reply(struct ceph_auth_client *ac, int result, + void *buf, void *end) +{ + struct ceph_x_info *xi = ac->private; + struct ceph_x_reply_header *head = buf; + struct ceph_x_ticket_handler *th; + int len = end - buf; + int op; + int ret; + + if (result) + return result; /* XXX hmm? */ + + if (xi->starting) { + /* it's a hello */ + struct ceph_x_server_challenge *sc = buf; + + if (len != sizeof(*sc)) + return -EINVAL; + xi->server_challenge = le64_to_cpu(sc->server_challenge); + dout("handle_reply got server challenge %llx\n", + xi->server_challenge); + xi->starting = false; + xi->have_keys &= ~CEPH_ENTITY_TYPE_AUTH; + return -EAGAIN; + } + + op = le16_to_cpu(head->op); + result = le32_to_cpu(head->result); + dout("handle_reply op %d result %d\n", op, result); + switch (op) { + case CEPHX_GET_AUTH_SESSION_KEY: + /* verify auth key */ + ret = ceph_x_proc_ticket_reply(ac, &xi->secret, + buf + sizeof(*head), end); + break; + + case CEPHX_GET_PRINCIPAL_SESSION_KEY: + th = get_ticket_handler(ac, CEPH_ENTITY_TYPE_AUTH); + if (IS_ERR(th)) + return PTR_ERR(th); + ret = ceph_x_proc_ticket_reply(ac, &th->session_key, + buf + sizeof(*head), end); + break; + + default: + return -EINVAL; + } + if (ret) + return ret; + if (ac->want_keys == xi->have_keys) + return 0; + return -EAGAIN; +} + +static int ceph_x_create_authorizer( + struct ceph_auth_client *ac, int peer_type, + struct ceph_authorizer **a, + void **buf, size_t *len, + void **reply_buf, size_t *reply_len) +{ + struct ceph_x_authorizer *au; + struct ceph_x_ticket_handler *th; + int ret; + + th = get_ticket_handler(ac, peer_type); + if (IS_ERR(th)) + return PTR_ERR(th); + + au = kzalloc(sizeof(*au), GFP_NOFS); + if (!au) + return -ENOMEM; + + ret = ceph_x_build_authorizer(ac, th, au); + if (ret) { + kfree(au); + return ret; + } + + *a = (struct ceph_authorizer *)au; + *buf = au->buf->vec.iov_base; + *len = au->buf->vec.iov_len; + *reply_buf = au->reply_buf; + *reply_len = sizeof(au->reply_buf); + return 0; +} + +static int ceph_x_verify_authorizer_reply(struct ceph_auth_client *ac, + struct ceph_authorizer *a, size_t len) +{ + struct ceph_x_authorizer *au = (void *)a; + struct ceph_x_ticket_handler *th; + int ret = 0; + struct ceph_x_authorize_reply reply; + void *p = au->reply_buf; + void *end = p + sizeof(au->reply_buf); + + th = get_ticket_handler(ac, au->service); + if (IS_ERR(th)) + return PTR_ERR(th); + ret = ceph_x_decrypt(&th->session_key, &p, end, &reply, sizeof(reply)); + if (ret < 0) + return ret; + if (ret != sizeof(reply)) + return -EPERM; + + if (au->nonce + 1 != le64_to_cpu(reply.nonce_plus_one)) + ret = -EPERM; + else + ret = 0; + dout("verify_authorizer_reply nonce %llx got %llx ret %d\n", + au->nonce, le64_to_cpu(reply.nonce_plus_one), ret); + return ret; +} + +static void ceph_x_destroy_authorizer(struct ceph_auth_client *ac, + struct ceph_authorizer *a) +{ + struct ceph_x_authorizer *au = (void *)a; + + ceph_buffer_put(au->buf); + kfree(au); +} + + +static void ceph_x_reset(struct ceph_auth_client *ac) +{ + struct ceph_x_info *xi = ac->private; + + dout("reset\n"); + xi->starting = true; + xi->server_challenge = 0; +} + +static void ceph_x_destroy(struct ceph_auth_client *ac) +{ + struct ceph_x_info *xi = ac->private; + struct rb_node *p; + + dout("ceph_x_destroy %p\n", ac); + ceph_crypto_key_destroy(&xi->secret); + + while ((p = rb_first(&xi->ticket_handlers)) != NULL) { + struct ceph_x_ticket_handler *th = + rb_entry(p, struct ceph_x_ticket_handler, node); + remove_ticket_handler(ac, th); + } + + if (xi->auth_authorizer.buf) + ceph_buffer_put(xi->auth_authorizer.buf); + + kfree(ac->private); + ac->private = NULL; +} + +static void ceph_x_invalidate_authorizer(struct ceph_auth_client *ac, + int peer_type) +{ + struct ceph_x_ticket_handler *th; + + th = get_ticket_handler(ac, peer_type); + if (!IS_ERR(th)) + remove_ticket_handler(ac, th); +} + + +static const struct ceph_auth_client_ops ceph_x_ops = { + .name = "x", + .is_authenticated = ceph_x_is_authenticated, + .should_authenticate = ceph_x_should_authenticate, + .build_request = ceph_x_build_request, + .handle_reply = ceph_x_handle_reply, + .create_authorizer = ceph_x_create_authorizer, + .verify_authorizer_reply = ceph_x_verify_authorizer_reply, + .destroy_authorizer = ceph_x_destroy_authorizer, + .invalidate_authorizer = ceph_x_invalidate_authorizer, + .reset = ceph_x_reset, + .destroy = ceph_x_destroy, +}; + + +int ceph_x_init(struct ceph_auth_client *ac) +{ + struct ceph_x_info *xi; + int ret; + + dout("ceph_x_init %p\n", ac); + ret = -ENOMEM; + xi = kzalloc(sizeof(*xi), GFP_NOFS); + if (!xi) + goto out; + + ret = -EINVAL; + if (!ac->secret) { + pr_err("no secret set (for auth_x protocol)\n"); + goto out_nomem; + } + + ret = ceph_crypto_key_unarmor(&xi->secret, ac->secret); + if (ret) + goto out_nomem; + + xi->starting = true; + xi->ticket_handlers = RB_ROOT; + + ac->protocol = CEPH_AUTH_CEPHX; + ac->private = xi; + ac->ops = &ceph_x_ops; + return 0; + +out_nomem: + kfree(xi); +out: + return ret; +} + + diff --git a/net/ceph/auth_x.h b/net/ceph/auth_x.h new file mode 100644 index 0000000..e02da7a --- /dev/null +++ b/net/ceph/auth_x.h @@ -0,0 +1,50 @@ +#ifndef _FS_CEPH_AUTH_X_H +#define _FS_CEPH_AUTH_X_H + +#include + +#include + +#include "crypto.h" +#include "auth_x_protocol.h" + +/* + * Handle ticket for a single service. + */ +struct ceph_x_ticket_handler { + struct rb_node node; + unsigned service; + + struct ceph_crypto_key session_key; + struct ceph_timespec validity; + + u64 secret_id; + struct ceph_buffer *ticket_blob; + + unsigned long renew_after, expires; +}; + + +struct ceph_x_authorizer { + struct ceph_buffer *buf; + unsigned service; + u64 nonce; + char reply_buf[128]; /* big enough for encrypted blob */ +}; + +struct ceph_x_info { + struct ceph_crypto_key secret; + + bool starting; + u64 server_challenge; + + unsigned have_keys; + struct rb_root ticket_handlers; + + struct ceph_x_authorizer auth_authorizer; +}; + +extern int ceph_x_init(struct ceph_auth_client *ac); + +#endif + diff --git a/net/ceph/auth_x_protocol.h b/net/ceph/auth_x_protocol.h new file mode 100644 index 0000000..671d305 --- /dev/null +++ b/net/ceph/auth_x_protocol.h @@ -0,0 +1,90 @@ +#ifndef __FS_CEPH_AUTH_X_PROTOCOL +#define __FS_CEPH_AUTH_X_PROTOCOL + +#define CEPHX_GET_AUTH_SESSION_KEY 0x0100 +#define CEPHX_GET_PRINCIPAL_SESSION_KEY 0x0200 +#define CEPHX_GET_ROTATING_KEY 0x0400 + +/* common bits */ +struct ceph_x_ticket_blob { + __u8 struct_v; + __le64 secret_id; + __le32 blob_len; + char blob[]; +} __attribute__ ((packed)); + + +/* common request/reply headers */ +struct ceph_x_request_header { + __le16 op; +} __attribute__ ((packed)); + +struct ceph_x_reply_header { + __le16 op; + __le32 result; +} __attribute__ ((packed)); + + +/* authenticate handshake */ + +/* initial hello (no reply header) */ +struct ceph_x_server_challenge { + __u8 struct_v; + __le64 server_challenge; +} __attribute__ ((packed)); + +struct ceph_x_authenticate { + __u8 struct_v; + __le64 client_challenge; + __le64 key; + /* ticket blob */ +} __attribute__ ((packed)); + +struct ceph_x_service_ticket_request { + __u8 struct_v; + __le32 keys; +} __attribute__ ((packed)); + +struct ceph_x_challenge_blob { + __le64 server_challenge; + __le64 client_challenge; +} __attribute__ ((packed)); + + + +/* authorize handshake */ + +/* + * The authorizer consists of two pieces: + * a - service id, ticket blob + * b - encrypted with session key + */ +struct ceph_x_authorize_a { + __u8 struct_v; + __le64 global_id; + __le32 service_id; + struct ceph_x_ticket_blob ticket_blob; +} __attribute__ ((packed)); + +struct ceph_x_authorize_b { + __u8 struct_v; + __le64 nonce; +} __attribute__ ((packed)); + +struct ceph_x_authorize_reply { + __u8 struct_v; + __le64 nonce_plus_one; +} __attribute__ ((packed)); + + +/* + * encyption bundle + */ +#define CEPHX_ENC_MAGIC 0xff009cad8826aa55ull + +struct ceph_x_encrypt_header { + __u8 struct_v; + __le64 magic; +} __attribute__ ((packed)); + +#endif diff --git a/net/ceph/buffer.c b/net/ceph/buffer.c new file mode 100644 index 0000000..53d8abf --- /dev/null +++ b/net/ceph/buffer.c @@ -0,0 +1,68 @@ + +#include + +#include +#include + +#include +#include + +struct ceph_buffer *ceph_buffer_new(size_t len, gfp_t gfp) +{ + struct ceph_buffer *b; + + b = kmalloc(sizeof(*b), gfp); + if (!b) + return NULL; + + b->vec.iov_base = kmalloc(len, gfp | __GFP_NOWARN); + if (b->vec.iov_base) { + b->is_vmalloc = false; + } else { + b->vec.iov_base = __vmalloc(len, gfp, PAGE_KERNEL); + if (!b->vec.iov_base) { + kfree(b); + return NULL; + } + b->is_vmalloc = true; + } + + kref_init(&b->kref); + b->alloc_len = len; + b->vec.iov_len = len; + dout("buffer_new %p\n", b); + return b; +} +EXPORT_SYMBOL(ceph_buffer_new); + +void ceph_buffer_release(struct kref *kref) +{ + struct ceph_buffer *b = container_of(kref, struct ceph_buffer, kref); + + dout("buffer_release %p\n", b); + if (b->vec.iov_base) { + if (b->is_vmalloc) + vfree(b->vec.iov_base); + else + kfree(b->vec.iov_base); + } + kfree(b); +} +EXPORT_SYMBOL(ceph_buffer_release); + +int ceph_decode_buffer(struct ceph_buffer **b, void **p, void *end) +{ + size_t len; + + ceph_decode_need(p, end, sizeof(u32), bad); + len = ceph_decode_32(p); + dout("decode_buffer len %d\n", (int)len); + ceph_decode_need(p, end, len, bad); + *b = ceph_buffer_new(len, GFP_NOFS); + if (!*b) + return -ENOMEM; + ceph_decode_copy(p, (*b)->vec.iov_base, len); + return 0; +bad: + return -EINVAL; +} diff --git a/net/ceph/ceph_common.c b/net/ceph/ceph_common.c new file mode 100644 index 0000000..f6f2eeb --- /dev/null +++ b/net/ceph/ceph_common.c @@ -0,0 +1,529 @@ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + + +#include +#include +#include +#include +#include + + + +/* + * find filename portion of a path (/foo/bar/baz -> baz) + */ +const char *ceph_file_part(const char *s, int len) +{ + const char *e = s + len; + + while (e != s && *(e-1) != '/') + e--; + return e; +} +EXPORT_SYMBOL(ceph_file_part); + +const char *ceph_msg_type_name(int type) +{ + switch (type) { + case CEPH_MSG_SHUTDOWN: return "shutdown"; + case CEPH_MSG_PING: return "ping"; + case CEPH_MSG_AUTH: return "auth"; + case CEPH_MSG_AUTH_REPLY: return "auth_reply"; + case CEPH_MSG_MON_MAP: return "mon_map"; + case CEPH_MSG_MON_GET_MAP: return "mon_get_map"; + case CEPH_MSG_MON_SUBSCRIBE: return "mon_subscribe"; + case CEPH_MSG_MON_SUBSCRIBE_ACK: return "mon_subscribe_ack"; + case CEPH_MSG_STATFS: return "statfs"; + case CEPH_MSG_STATFS_REPLY: return "statfs_reply"; + case CEPH_MSG_MDS_MAP: return "mds_map"; + case CEPH_MSG_CLIENT_SESSION: return "client_session"; + case CEPH_MSG_CLIENT_RECONNECT: return "client_reconnect"; + case CEPH_MSG_CLIENT_REQUEST: return "client_request"; + case CEPH_MSG_CLIENT_REQUEST_FORWARD: return "client_request_forward"; + case CEPH_MSG_CLIENT_REPLY: return "client_reply"; + case CEPH_MSG_CLIENT_CAPS: return "client_caps"; + case CEPH_MSG_CLIENT_CAPRELEASE: return "client_cap_release"; + case CEPH_MSG_CLIENT_SNAP: return "client_snap"; + case CEPH_MSG_CLIENT_LEASE: return "client_lease"; + case CEPH_MSG_OSD_MAP: return "osd_map"; + case CEPH_MSG_OSD_OP: return "osd_op"; + case CEPH_MSG_OSD_OPREPLY: return "osd_opreply"; + default: return "unknown"; + } +} +EXPORT_SYMBOL(ceph_msg_type_name); + +/* + * Initially learn our fsid, or verify an fsid matches. + */ +int ceph_check_fsid(struct ceph_client *client, struct ceph_fsid *fsid) +{ + if (client->have_fsid) { + if (ceph_fsid_compare(&client->fsid, fsid)) { + pr_err("bad fsid, had %pU got %pU", + &client->fsid, fsid); + return -1; + } + } else { + pr_info("client%lld fsid %pU\n", ceph_client_id(client), fsid); + memcpy(&client->fsid, fsid, sizeof(*fsid)); + ceph_debugfs_client_init(client); + client->have_fsid = true; + } + return 0; +} +EXPORT_SYMBOL(ceph_check_fsid); + +static int strcmp_null(const char *s1, const char *s2) +{ + if (!s1 && !s2) + return 0; + if (s1 && !s2) + return -1; + if (!s1 && s2) + return 1; + return strcmp(s1, s2); +} + +int ceph_compare_options(struct ceph_options *new_opt, + struct ceph_client *client) +{ + struct ceph_options *opt1 = new_opt; + struct ceph_options *opt2 = client->options; + int ofs = offsetof(struct ceph_options, mon_addr); + int i; + int ret; + + ret = memcmp(opt1, opt2, ofs); + if (ret) + return ret; + + ret = strcmp_null(opt1->name, opt2->name); + if (ret) + return ret; + + ret = strcmp_null(opt1->secret, opt2->secret); + if (ret) + return ret; + + /* any matching mon ip implies a match */ + for (i = 0; i < opt1->num_mon; i++) { + if (ceph_monmap_contains(client->monc.monmap, + &opt1->mon_addr[i])) + return 0; + } + return -1; +} +EXPORT_SYMBOL(ceph_compare_options); + + +static int parse_fsid(const char *str, struct ceph_fsid *fsid) +{ + int i = 0; + char tmp[3]; + int err = -EINVAL; + int d; + + dout("parse_fsid '%s'\n", str); + tmp[2] = 0; + while (*str && i < 16) { + if (ispunct(*str)) { + str++; + continue; + } + if (!isxdigit(str[0]) || !isxdigit(str[1])) + break; + tmp[0] = str[0]; + tmp[1] = str[1]; + if (sscanf(tmp, "%x", &d) < 1) + break; + fsid->fsid[i] = d & 0xff; + i++; + str += 2; + } + + if (i == 16) + err = 0; + dout("parse_fsid ret %d got fsid %pU", err, fsid); + return err; +} + +/* + * ceph options + */ +enum { + Opt_osdtimeout, + Opt_osdkeepalivetimeout, + Opt_mount_timeout, + Opt_osd_idle_ttl, + Opt_last_int, + /* int args above */ + Opt_fsid, + Opt_name, + Opt_secret, + Opt_ip, + Opt_last_string, + /* string args above */ + Opt_noshare, + Opt_nocrc, +}; + +static match_table_t opt_tokens = { + {Opt_osdtimeout, "osdtimeout=%d"}, + {Opt_osdkeepalivetimeout, "osdkeepalive=%d"}, + {Opt_mount_timeout, "mount_timeout=%d"}, + {Opt_osd_idle_ttl, "osd_idle_ttl=%d"}, + /* int args above */ + {Opt_fsid, "fsid=%s"}, + {Opt_name, "name=%s"}, + {Opt_secret, "secret=%s"}, + {Opt_ip, "ip=%s"}, + /* string args above */ + {Opt_noshare, "noshare"}, + {Opt_nocrc, "nocrc"}, + {-1, NULL} +}; + +void ceph_destroy_options(struct ceph_options *opt) +{ + dout("destroy_options %p\n", opt); + kfree(opt->name); + kfree(opt->secret); + kfree(opt); +} +EXPORT_SYMBOL(ceph_destroy_options); + +int ceph_parse_options(struct ceph_options **popt, char *options, + const char *dev_name, const char *dev_name_end, + int (*parse_extra_token)(char *c, void *private), + void *private) +{ + struct ceph_options *opt; + const char *c; + int err = -ENOMEM; + substring_t argstr[MAX_OPT_ARGS]; + + opt = kzalloc(sizeof(*opt), GFP_KERNEL); + if (!opt) + return err; + opt->mon_addr = kcalloc(CEPH_MAX_MON, sizeof(*opt->mon_addr), + GFP_KERNEL); + if (!opt->mon_addr) + goto out; + + dout("parse_options %p options '%s' dev_name '%s'\n", opt, options, + dev_name); + + /* start with defaults */ + opt->flags = CEPH_OPT_DEFAULT; + opt->osd_timeout = CEPH_OSD_TIMEOUT_DEFAULT; + opt->osd_keepalive_timeout = CEPH_OSD_KEEPALIVE_DEFAULT; + opt->mount_timeout = CEPH_MOUNT_TIMEOUT_DEFAULT; /* seconds */ + opt->osd_idle_ttl = CEPH_OSD_IDLE_TTL_DEFAULT; /* seconds */ + + /* get mon ip(s) */ + /* ip1[:port1][,ip2[:port2]...] */ + err = ceph_parse_ips(dev_name, dev_name_end, opt->mon_addr, + CEPH_MAX_MON, &opt->num_mon); + if (err < 0) + goto out; + + /* parse mount options */ + while ((c = strsep(&options, ",")) != NULL) { + int token, intval, ret; + if (!*c) + continue; + err = -EINVAL; + token = match_token((char *)c, opt_tokens, argstr); + if (token < 0) { + /* extra? */ + err = parse_extra_token((char *)c, private); + if (err < 0) { + pr_err("bad option at '%s'\n", c); + goto out; + } + continue; + } + if (token < Opt_last_int) { + ret = match_int(&argstr[0], &intval); + if (ret < 0) { + pr_err("bad mount option arg (not int) " + "at '%s'\n", c); + continue; + } + dout("got int token %d val %d\n", token, intval); + } else if (token > Opt_last_int && token < Opt_last_string) { + dout("got string token %d val %s\n", token, + argstr[0].from); + } else { + dout("got token %d\n", token); + } + switch (token) { + case Opt_ip: + err = ceph_parse_ips(argstr[0].from, + argstr[0].to, + &opt->my_addr, + 1, NULL); + if (err < 0) + goto out; + opt->flags |= CEPH_OPT_MYIP; + break; + + case Opt_fsid: + err = parse_fsid(argstr[0].from, &opt->fsid); + if (err == 0) + opt->flags |= CEPH_OPT_FSID; + break; + case Opt_name: + opt->name = kstrndup(argstr[0].from, + argstr[0].to-argstr[0].from, + GFP_KERNEL); + break; + case Opt_secret: + opt->secret = kstrndup(argstr[0].from, + argstr[0].to-argstr[0].from, + GFP_KERNEL); + break; + + /* misc */ + case Opt_osdtimeout: + opt->osd_timeout = intval; + break; + case Opt_osdkeepalivetimeout: + opt->osd_keepalive_timeout = intval; + break; + case Opt_osd_idle_ttl: + opt->osd_idle_ttl = intval; + break; + case Opt_mount_timeout: + opt->mount_timeout = intval; + break; + + case Opt_noshare: + opt->flags |= CEPH_OPT_NOSHARE; + break; + + case Opt_nocrc: + opt->flags |= CEPH_OPT_NOCRC; + break; + + default: + BUG_ON(token); + } + } + + /* success */ + *popt = opt; + return 0; + +out: + ceph_destroy_options(opt); + return err; +} +EXPORT_SYMBOL(ceph_parse_options); + +u64 ceph_client_id(struct ceph_client *client) +{ + return client->monc.auth->global_id; +} +EXPORT_SYMBOL(ceph_client_id); + +/* + * create a fresh client instance + */ +struct ceph_client *ceph_create_client(struct ceph_options *opt, void *private) +{ + struct ceph_client *client; + int err = -ENOMEM; + + client = kzalloc(sizeof(*client), GFP_KERNEL); + if (client == NULL) + return ERR_PTR(-ENOMEM); + + client->private = private; + client->options = opt; + + mutex_init(&client->mount_mutex); + init_waitqueue_head(&client->auth_wq); + client->auth_err = 0; + + client->extra_mon_dispatch = NULL; + client->supported_features = CEPH_FEATURE_SUPPORTED_DEFAULT; + client->required_features = CEPH_FEATURE_REQUIRED_DEFAULT; + + client->msgr = NULL; + + /* subsystems */ + err = ceph_monc_init(&client->monc, client); + if (err < 0) + goto fail; + err = ceph_osdc_init(&client->osdc, client); + if (err < 0) + goto fail_monc; + + return client; + +fail_monc: + ceph_monc_stop(&client->monc); +fail: + kfree(client); + return ERR_PTR(err); +} +EXPORT_SYMBOL(ceph_create_client); + +void ceph_destroy_client(struct ceph_client *client) +{ + dout("destroy_client %p\n", client); + + /* unmount */ + ceph_osdc_stop(&client->osdc); + + /* + * make sure mds and osd connections close out before destroying + * the auth module, which is needed to free those connections' + * ceph_authorizers. + */ + ceph_msgr_flush(); + + ceph_monc_stop(&client->monc); + + ceph_debugfs_client_cleanup(client); + + if (client->msgr) + ceph_messenger_destroy(client->msgr); + + ceph_destroy_options(client->options); + + kfree(client); + dout("destroy_client %p done\n", client); +} +EXPORT_SYMBOL(ceph_destroy_client); + +/* + * true if we have the mon map (and have thus joined the cluster) + */ +static int have_mon_and_osd_map(struct ceph_client *client) +{ + return client->monc.monmap && client->monc.monmap->epoch && + client->osdc.osdmap && client->osdc.osdmap->epoch; +} + +/* + * mount: join the ceph cluster, and open root directory. + */ +int __ceph_open_session(struct ceph_client *client, unsigned long started) +{ + struct ceph_entity_addr *myaddr = NULL; + int err; + unsigned long timeout = client->options->mount_timeout * HZ; + + /* initialize the messenger */ + if (client->msgr == NULL) { + if (ceph_test_opt(client, MYIP)) + myaddr = &client->options->my_addr; + client->msgr = ceph_messenger_create(myaddr, + client->supported_features, + client->required_features); + if (IS_ERR(client->msgr)) { + client->msgr = NULL; + return PTR_ERR(client->msgr); + } + client->msgr->nocrc = ceph_test_opt(client, NOCRC); + } + + /* open session, and wait for mon and osd maps */ + err = ceph_monc_open_session(&client->monc); + if (err < 0) + return err; + + while (!have_mon_and_osd_map(client)) { + err = -EIO; + if (timeout && time_after_eq(jiffies, started + timeout)) + return err; + + /* wait */ + dout("mount waiting for mon_map\n"); + err = wait_event_interruptible_timeout(client->auth_wq, + have_mon_and_osd_map(client) || (client->auth_err < 0), + timeout); + if (err == -EINTR || err == -ERESTARTSYS) + return err; + if (client->auth_err < 0) + return client->auth_err; + } + + return 0; +} +EXPORT_SYMBOL(__ceph_open_session); + + +int ceph_open_session(struct ceph_client *client) +{ + int ret; + unsigned long started = jiffies; /* note the start time */ + + dout("open_session start\n"); + mutex_lock(&client->mount_mutex); + + ret = __ceph_open_session(client, started); + + mutex_unlock(&client->mount_mutex); + return ret; +} +EXPORT_SYMBOL(ceph_open_session); + + +static int __init init_ceph_lib(void) +{ + int ret = 0; + + ret = ceph_debugfs_init(); + if (ret < 0) + goto out; + + ret = ceph_msgr_init(); + if (ret < 0) + goto out_debugfs; + + pr_info("loaded (mon/osd proto %d/%d, osdmap %d/%d %d/%d)\n", + CEPH_MONC_PROTOCOL, CEPH_OSDC_PROTOCOL, + CEPH_OSDMAP_VERSION, CEPH_OSDMAP_VERSION_EXT, + CEPH_OSDMAP_INC_VERSION, CEPH_OSDMAP_INC_VERSION_EXT); + + return 0; + +out_debugfs: + ceph_debugfs_cleanup(); +out: + return ret; +} + +static void __exit exit_ceph_lib(void) +{ + dout("exit_ceph_lib\n"); + ceph_msgr_exit(); + ceph_debugfs_cleanup(); +} + +module_init(init_ceph_lib); +module_exit(exit_ceph_lib); + +MODULE_AUTHOR("Sage Weil "); +MODULE_AUTHOR("Yehuda Sadeh "); +MODULE_AUTHOR("Patience Warnick "); +MODULE_DESCRIPTION("Ceph filesystem for Linux"); +MODULE_LICENSE("GPL"); diff --git a/net/ceph/ceph_fs.c b/net/ceph/ceph_fs.c new file mode 100644 index 0000000..a3a3a31 --- /dev/null +++ b/net/ceph/ceph_fs.c @@ -0,0 +1,75 @@ +/* + * Some non-inline ceph helpers + */ +#include +#include + +/* + * return true if @layout appears to be valid + */ +int ceph_file_layout_is_valid(const struct ceph_file_layout *layout) +{ + __u32 su = le32_to_cpu(layout->fl_stripe_unit); + __u32 sc = le32_to_cpu(layout->fl_stripe_count); + __u32 os = le32_to_cpu(layout->fl_object_size); + + /* stripe unit, object size must be non-zero, 64k increment */ + if (!su || (su & (CEPH_MIN_STRIPE_UNIT-1))) + return 0; + if (!os || (os & (CEPH_MIN_STRIPE_UNIT-1))) + return 0; + /* object size must be a multiple of stripe unit */ + if (os < su || os % su) + return 0; + /* stripe count must be non-zero */ + if (!sc) + return 0; + return 1; +} + + +int ceph_flags_to_mode(int flags) +{ + int mode; + +#ifdef O_DIRECTORY /* fixme */ + if ((flags & O_DIRECTORY) == O_DIRECTORY) + return CEPH_FILE_MODE_PIN; +#endif + if ((flags & O_APPEND) == O_APPEND) + flags |= O_WRONLY; + + if ((flags & O_ACCMODE) == O_RDWR) + mode = CEPH_FILE_MODE_RDWR; + else if ((flags & O_ACCMODE) == O_WRONLY) + mode = CEPH_FILE_MODE_WR; + else + mode = CEPH_FILE_MODE_RD; + +#ifdef O_LAZY + if (flags & O_LAZY) + mode |= CEPH_FILE_MODE_LAZY; +#endif + + return mode; +} +EXPORT_SYMBOL(ceph_flags_to_mode); + +int ceph_caps_for_mode(int mode) +{ + int caps = CEPH_CAP_PIN; + + if (mode & CEPH_FILE_MODE_RD) + caps |= CEPH_CAP_FILE_SHARED | + CEPH_CAP_FILE_RD | CEPH_CAP_FILE_CACHE; + if (mode & CEPH_FILE_MODE_WR) + caps |= CEPH_CAP_FILE_EXCL | + CEPH_CAP_FILE_WR | CEPH_CAP_FILE_BUFFER | + CEPH_CAP_AUTH_SHARED | CEPH_CAP_AUTH_EXCL | + CEPH_CAP_XATTR_SHARED | CEPH_CAP_XATTR_EXCL; + if (mode & CEPH_FILE_MODE_LAZY) + caps |= CEPH_CAP_FILE_LAZYIO; + + return caps; +} +EXPORT_SYMBOL(ceph_caps_for_mode); diff --git a/net/ceph/ceph_hash.c b/net/ceph/ceph_hash.c new file mode 100644 index 0000000..815ef88 --- /dev/null +++ b/net/ceph/ceph_hash.c @@ -0,0 +1,118 @@ + +#include + +/* + * Robert Jenkin's hash function. + * http://burtleburtle.net/bob/hash/evahash.html + * This is in the public domain. + */ +#define mix(a, b, c) \ + do { \ + a = a - b; a = a - c; a = a ^ (c >> 13); \ + b = b - c; b = b - a; b = b ^ (a << 8); \ + c = c - a; c = c - b; c = c ^ (b >> 13); \ + a = a - b; a = a - c; a = a ^ (c >> 12); \ + b = b - c; b = b - a; b = b ^ (a << 16); \ + c = c - a; c = c - b; c = c ^ (b >> 5); \ + a = a - b; a = a - c; a = a ^ (c >> 3); \ + b = b - c; b = b - a; b = b ^ (a << 10); \ + c = c - a; c = c - b; c = c ^ (b >> 15); \ + } while (0) + +unsigned ceph_str_hash_rjenkins(const char *str, unsigned length) +{ + const unsigned char *k = (const unsigned char *)str; + __u32 a, b, c; /* the internal state */ + __u32 len; /* how many key bytes still need mixing */ + + /* Set up the internal state */ + len = length; + a = 0x9e3779b9; /* the golden ratio; an arbitrary value */ + b = a; + c = 0; /* variable initialization of internal state */ + + /* handle most of the key */ + while (len >= 12) { + a = a + (k[0] + ((__u32)k[1] << 8) + ((__u32)k[2] << 16) + + ((__u32)k[3] << 24)); + b = b + (k[4] + ((__u32)k[5] << 8) + ((__u32)k[6] << 16) + + ((__u32)k[7] << 24)); + c = c + (k[8] + ((__u32)k[9] << 8) + ((__u32)k[10] << 16) + + ((__u32)k[11] << 24)); + mix(a, b, c); + k = k + 12; + len = len - 12; + } + + /* handle the last 11 bytes */ + c = c + length; + switch (len) { /* all the case statements fall through */ + case 11: + c = c + ((__u32)k[10] << 24); + case 10: + c = c + ((__u32)k[9] << 16); + case 9: + c = c + ((__u32)k[8] << 8); + /* the first byte of c is reserved for the length */ + case 8: + b = b + ((__u32)k[7] << 24); + case 7: + b = b + ((__u32)k[6] << 16); + case 6: + b = b + ((__u32)k[5] << 8); + case 5: + b = b + k[4]; + case 4: + a = a + ((__u32)k[3] << 24); + case 3: + a = a + ((__u32)k[2] << 16); + case 2: + a = a + ((__u32)k[1] << 8); + case 1: + a = a + k[0]; + /* case 0: nothing left to add */ + } + mix(a, b, c); + + return c; +} + +/* + * linux dcache hash + */ +unsigned ceph_str_hash_linux(const char *str, unsigned length) +{ + unsigned long hash = 0; + unsigned char c; + + while (length--) { + c = *str++; + hash = (hash + (c << 4) + (c >> 4)) * 11; + } + return hash; +} + + +unsigned ceph_str_hash(int type, const char *s, unsigned len) +{ + switch (type) { + case CEPH_STR_HASH_LINUX: + return ceph_str_hash_linux(s, len); + case CEPH_STR_HASH_RJENKINS: + return ceph_str_hash_rjenkins(s, len); + default: + return -1; + } +} + +const char *ceph_str_hash_name(int type) +{ + switch (type) { + case CEPH_STR_HASH_LINUX: + return "linux"; + case CEPH_STR_HASH_RJENKINS: + return "rjenkins"; + default: + return "unknown"; + } +} diff --git a/net/ceph/ceph_strings.c b/net/ceph/ceph_strings.c new file mode 100644 index 0000000..3fbda04 --- /dev/null +++ b/net/ceph/ceph_strings.c @@ -0,0 +1,84 @@ +/* + * Ceph string constants + */ +#include +#include + +const char *ceph_entity_type_name(int type) +{ + switch (type) { + case CEPH_ENTITY_TYPE_MDS: return "mds"; + case CEPH_ENTITY_TYPE_OSD: return "osd"; + case CEPH_ENTITY_TYPE_MON: return "mon"; + case CEPH_ENTITY_TYPE_CLIENT: return "client"; + case CEPH_ENTITY_TYPE_AUTH: return "auth"; + default: return "unknown"; + } +} + +const char *ceph_osd_op_name(int op) +{ + switch (op) { + case CEPH_OSD_OP_READ: return "read"; + case CEPH_OSD_OP_STAT: return "stat"; + + case CEPH_OSD_OP_MASKTRUNC: return "masktrunc"; + + case CEPH_OSD_OP_WRITE: return "write"; + case CEPH_OSD_OP_DELETE: return "delete"; + case CEPH_OSD_OP_TRUNCATE: return "truncate"; + case CEPH_OSD_OP_ZERO: return "zero"; + case CEPH_OSD_OP_WRITEFULL: return "writefull"; + case CEPH_OSD_OP_ROLLBACK: return "rollback"; + + case CEPH_OSD_OP_APPEND: return "append"; + case CEPH_OSD_OP_STARTSYNC: return "startsync"; + case CEPH_OSD_OP_SETTRUNC: return "settrunc"; + case CEPH_OSD_OP_TRIMTRUNC: return "trimtrunc"; + + case CEPH_OSD_OP_TMAPUP: return "tmapup"; + case CEPH_OSD_OP_TMAPGET: return "tmapget"; + case CEPH_OSD_OP_TMAPPUT: return "tmapput"; + + case CEPH_OSD_OP_GETXATTR: return "getxattr"; + case CEPH_OSD_OP_GETXATTRS: return "getxattrs"; + case CEPH_OSD_OP_SETXATTR: return "setxattr"; + case CEPH_OSD_OP_SETXATTRS: return "setxattrs"; + case CEPH_OSD_OP_RESETXATTRS: return "resetxattrs"; + case CEPH_OSD_OP_RMXATTR: return "rmxattr"; + case CEPH_OSD_OP_CMPXATTR: return "cmpxattr"; + + case CEPH_OSD_OP_PULL: return "pull"; + case CEPH_OSD_OP_PUSH: return "push"; + case CEPH_OSD_OP_BALANCEREADS: return "balance-reads"; + case CEPH_OSD_OP_UNBALANCEREADS: return "unbalance-reads"; + case CEPH_OSD_OP_SCRUB: return "scrub"; + + case CEPH_OSD_OP_WRLOCK: return "wrlock"; + case CEPH_OSD_OP_WRUNLOCK: return "wrunlock"; + case CEPH_OSD_OP_RDLOCK: return "rdlock"; + case CEPH_OSD_OP_RDUNLOCK: return "rdunlock"; + case CEPH_OSD_OP_UPLOCK: return "uplock"; + case CEPH_OSD_OP_DNLOCK: return "dnlock"; + + case CEPH_OSD_OP_CALL: return "call"; + + case CEPH_OSD_OP_PGLS: return "pgls"; + } + return "???"; +} + + +const char *ceph_pool_op_name(int op) +{ + switch (op) { + case POOL_OP_CREATE: return "create"; + case POOL_OP_DELETE: return "delete"; + case POOL_OP_AUID_CHANGE: return "auid change"; + case POOL_OP_CREATE_SNAP: return "create snap"; + case POOL_OP_DELETE_SNAP: return "delete snap"; + case POOL_OP_CREATE_UNMANAGED_SNAP: return "create unmanaged snap"; + case POOL_OP_DELETE_UNMANAGED_SNAP: return "delete unmanaged snap"; + } + return "???"; +} diff --git a/net/ceph/crush/crush.c b/net/ceph/crush/crush.c new file mode 100644 index 0000000..d6ebb13 --- /dev/null +++ b/net/ceph/crush/crush.c @@ -0,0 +1,151 @@ + +#ifdef __KERNEL__ +# include +#else +# include +# include +# define kfree(x) do { if (x) free(x); } while (0) +# define BUG_ON(x) assert(!(x)) +#endif + +#include + +const char *crush_bucket_alg_name(int alg) +{ + switch (alg) { + case CRUSH_BUCKET_UNIFORM: return "uniform"; + case CRUSH_BUCKET_LIST: return "list"; + case CRUSH_BUCKET_TREE: return "tree"; + case CRUSH_BUCKET_STRAW: return "straw"; + default: return "unknown"; + } +} + +/** + * crush_get_bucket_item_weight - Get weight of an item in given bucket + * @b: bucket pointer + * @p: item index in bucket + */ +int crush_get_bucket_item_weight(struct crush_bucket *b, int p) +{ + if (p >= b->size) + return 0; + + switch (b->alg) { + case CRUSH_BUCKET_UNIFORM: + return ((struct crush_bucket_uniform *)b)->item_weight; + case CRUSH_BUCKET_LIST: + return ((struct crush_bucket_list *)b)->item_weights[p]; + case CRUSH_BUCKET_TREE: + if (p & 1) + return ((struct crush_bucket_tree *)b)->node_weights[p]; + return 0; + case CRUSH_BUCKET_STRAW: + return ((struct crush_bucket_straw *)b)->item_weights[p]; + } + return 0; +} + +/** + * crush_calc_parents - Calculate parent vectors for the given crush map. + * @map: crush_map pointer + */ +void crush_calc_parents(struct crush_map *map) +{ + int i, b, c; + + for (b = 0; b < map->max_buckets; b++) { + if (map->buckets[b] == NULL) + continue; + for (i = 0; i < map->buckets[b]->size; i++) { + c = map->buckets[b]->items[i]; + BUG_ON(c >= map->max_devices || + c < -map->max_buckets); + if (c >= 0) + map->device_parents[c] = map->buckets[b]->id; + else + map->bucket_parents[-1-c] = map->buckets[b]->id; + } + } +} + +void crush_destroy_bucket_uniform(struct crush_bucket_uniform *b) +{ + kfree(b->h.perm); + kfree(b->h.items); + kfree(b); +} + +void crush_destroy_bucket_list(struct crush_bucket_list *b) +{ + kfree(b->item_weights); + kfree(b->sum_weights); + kfree(b->h.perm); + kfree(b->h.items); + kfree(b); +} + +void crush_destroy_bucket_tree(struct crush_bucket_tree *b) +{ + kfree(b->node_weights); + kfree(b); +} + +void crush_destroy_bucket_straw(struct crush_bucket_straw *b) +{ + kfree(b->straws); + kfree(b->item_weights); + kfree(b->h.perm); + kfree(b->h.items); + kfree(b); +} + +void crush_destroy_bucket(struct crush_bucket *b) +{ + switch (b->alg) { + case CRUSH_BUCKET_UNIFORM: + crush_destroy_bucket_uniform((struct crush_bucket_uniform *)b); + break; + case CRUSH_BUCKET_LIST: + crush_destroy_bucket_list((struct crush_bucket_list *)b); + break; + case CRUSH_BUCKET_TREE: + crush_destroy_bucket_tree((struct crush_bucket_tree *)b); + break; + case CRUSH_BUCKET_STRAW: + crush_destroy_bucket_straw((struct crush_bucket_straw *)b); + break; + } +} + +/** + * crush_destroy - Destroy a crush_map + * @map: crush_map pointer + */ +void crush_destroy(struct crush_map *map) +{ + int b; + + /* buckets */ + if (map->buckets) { + for (b = 0; b < map->max_buckets; b++) { + if (map->buckets[b] == NULL) + continue; + crush_destroy_bucket(map->buckets[b]); + } + kfree(map->buckets); + } + + /* rules */ + if (map->rules) { + for (b = 0; b < map->max_rules; b++) + kfree(map->rules[b]); + kfree(map->rules); + } + + kfree(map->bucket_parents); + kfree(map->device_parents); + kfree(map); +} + + diff --git a/net/ceph/crush/hash.c b/net/ceph/crush/hash.c new file mode 100644 index 0000000..5bb63e3 --- /dev/null +++ b/net/ceph/crush/hash.c @@ -0,0 +1,149 @@ + +#include +#include + +/* + * Robert Jenkins' function for mixing 32-bit values + * http://burtleburtle.net/bob/hash/evahash.html + * a, b = random bits, c = input and output + */ +#define crush_hashmix(a, b, c) do { \ + a = a-b; a = a-c; a = a^(c>>13); \ + b = b-c; b = b-a; b = b^(a<<8); \ + c = c-a; c = c-b; c = c^(b>>13); \ + a = a-b; a = a-c; a = a^(c>>12); \ + b = b-c; b = b-a; b = b^(a<<16); \ + c = c-a; c = c-b; c = c^(b>>5); \ + a = a-b; a = a-c; a = a^(c>>3); \ + b = b-c; b = b-a; b = b^(a<<10); \ + c = c-a; c = c-b; c = c^(b>>15); \ + } while (0) + +#define crush_hash_seed 1315423911 + +static __u32 crush_hash32_rjenkins1(__u32 a) +{ + __u32 hash = crush_hash_seed ^ a; + __u32 b = a; + __u32 x = 231232; + __u32 y = 1232; + crush_hashmix(b, x, hash); + crush_hashmix(y, a, hash); + return hash; +} + +static __u32 crush_hash32_rjenkins1_2(__u32 a, __u32 b) +{ + __u32 hash = crush_hash_seed ^ a ^ b; + __u32 x = 231232; + __u32 y = 1232; + crush_hashmix(a, b, hash); + crush_hashmix(x, a, hash); + crush_hashmix(b, y, hash); + return hash; +} + +static __u32 crush_hash32_rjenkins1_3(__u32 a, __u32 b, __u32 c) +{ + __u32 hash = crush_hash_seed ^ a ^ b ^ c; + __u32 x = 231232; + __u32 y = 1232; + crush_hashmix(a, b, hash); + crush_hashmix(c, x, hash); + crush_hashmix(y, a, hash); + crush_hashmix(b, x, hash); + crush_hashmix(y, c, hash); + return hash; +} + +static __u32 crush_hash32_rjenkins1_4(__u32 a, __u32 b, __u32 c, __u32 d) +{ + __u32 hash = crush_hash_seed ^ a ^ b ^ c ^ d; + __u32 x = 231232; + __u32 y = 1232; + crush_hashmix(a, b, hash); + crush_hashmix(c, d, hash); + crush_hashmix(a, x, hash); + crush_hashmix(y, b, hash); + crush_hashmix(c, x, hash); + crush_hashmix(y, d, hash); + return hash; +} + +static __u32 crush_hash32_rjenkins1_5(__u32 a, __u32 b, __u32 c, __u32 d, + __u32 e) +{ + __u32 hash = crush_hash_seed ^ a ^ b ^ c ^ d ^ e; + __u32 x = 231232; + __u32 y = 1232; + crush_hashmix(a, b, hash); + crush_hashmix(c, d, hash); + crush_hashmix(e, x, hash); + crush_hashmix(y, a, hash); + crush_hashmix(b, x, hash); + crush_hashmix(y, c, hash); + crush_hashmix(d, x, hash); + crush_hashmix(y, e, hash); + return hash; +} + + +__u32 crush_hash32(int type, __u32 a) +{ + switch (type) { + case CRUSH_HASH_RJENKINS1: + return crush_hash32_rjenkins1(a); + default: + return 0; + } +} + +__u32 crush_hash32_2(int type, __u32 a, __u32 b) +{ + switch (type) { + case CRUSH_HASH_RJENKINS1: + return crush_hash32_rjenkins1_2(a, b); + default: + return 0; + } +} + +__u32 crush_hash32_3(int type, __u32 a, __u32 b, __u32 c) +{ + switch (type) { + case CRUSH_HASH_RJENKINS1: + return crush_hash32_rjenkins1_3(a, b, c); + default: + return 0; + } +} + +__u32 crush_hash32_4(int type, __u32 a, __u32 b, __u32 c, __u32 d) +{ + switch (type) { + case CRUSH_HASH_RJENKINS1: + return crush_hash32_rjenkins1_4(a, b, c, d); + default: + return 0; + } +} + +__u32 crush_hash32_5(int type, __u32 a, __u32 b, __u32 c, __u32 d, __u32 e) +{ + switch (type) { + case CRUSH_HASH_RJENKINS1: + return crush_hash32_rjenkins1_5(a, b, c, d, e); + default: + return 0; + } +} + +const char *crush_hash_name(int type) +{ + switch (type) { + case CRUSH_HASH_RJENKINS1: + return "rjenkins1"; + default: + return "unknown"; + } +} diff --git a/net/ceph/crush/mapper.c b/net/ceph/crush/mapper.c new file mode 100644 index 0000000..42599e3 --- /dev/null +++ b/net/ceph/crush/mapper.c @@ -0,0 +1,609 @@ + +#ifdef __KERNEL__ +# include +# include +# include +# include +# ifndef dprintk +# define dprintk(args...) +# endif +#else +# include +# include +# include +# include +# define BUG_ON(x) assert(!(x)) +# define dprintk(args...) /* printf(args) */ +# define kmalloc(x, f) malloc(x) +# define kfree(x) free(x) +#endif + +#include +#include + +/* + * Implement the core CRUSH mapping algorithm. + */ + +/** + * crush_find_rule - find a crush_rule id for a given ruleset, type, and size. + * @map: the crush_map + * @ruleset: the storage ruleset id (user defined) + * @type: storage ruleset type (user defined) + * @size: output set size + */ +int crush_find_rule(struct crush_map *map, int ruleset, int type, int size) +{ + int i; + + for (i = 0; i < map->max_rules; i++) { + if (map->rules[i] && + map->rules[i]->mask.ruleset == ruleset && + map->rules[i]->mask.type == type && + map->rules[i]->mask.min_size <= size && + map->rules[i]->mask.max_size >= size) + return i; + } + return -1; +} + + +/* + * bucket choose methods + * + * For each bucket algorithm, we have a "choose" method that, given a + * crush input @x and replica position (usually, position in output set) @r, + * will produce an item in the bucket. + */ + +/* + * Choose based on a random permutation of the bucket. + * + * We used to use some prime number arithmetic to do this, but it + * wasn't very random, and had some other bad behaviors. Instead, we + * calculate an actual random permutation of the bucket members. + * Since this is expensive, we optimize for the r=0 case, which + * captures the vast majority of calls. + */ +static int bucket_perm_choose(struct crush_bucket *bucket, + int x, int r) +{ + unsigned pr = r % bucket->size; + unsigned i, s; + + /* start a new permutation if @x has changed */ + if (bucket->perm_x != x || bucket->perm_n == 0) { + dprintk("bucket %d new x=%d\n", bucket->id, x); + bucket->perm_x = x; + + /* optimize common r=0 case */ + if (pr == 0) { + s = crush_hash32_3(bucket->hash, x, bucket->id, 0) % + bucket->size; + bucket->perm[0] = s; + bucket->perm_n = 0xffff; /* magic value, see below */ + goto out; + } + + for (i = 0; i < bucket->size; i++) + bucket->perm[i] = i; + bucket->perm_n = 0; + } else if (bucket->perm_n == 0xffff) { + /* clean up after the r=0 case above */ + for (i = 1; i < bucket->size; i++) + bucket->perm[i] = i; + bucket->perm[bucket->perm[0]] = 0; + bucket->perm_n = 1; + } + + /* calculate permutation up to pr */ + for (i = 0; i < bucket->perm_n; i++) + dprintk(" perm_choose have %d: %d\n", i, bucket->perm[i]); + while (bucket->perm_n <= pr) { + unsigned p = bucket->perm_n; + /* no point in swapping the final entry */ + if (p < bucket->size - 1) { + i = crush_hash32_3(bucket->hash, x, bucket->id, p) % + (bucket->size - p); + if (i) { + unsigned t = bucket->perm[p + i]; + bucket->perm[p + i] = bucket->perm[p]; + bucket->perm[p] = t; + } + dprintk(" perm_choose swap %d with %d\n", p, p+i); + } + bucket->perm_n++; + } + for (i = 0; i < bucket->size; i++) + dprintk(" perm_choose %d: %d\n", i, bucket->perm[i]); + + s = bucket->perm[pr]; +out: + dprintk(" perm_choose %d sz=%d x=%d r=%d (%d) s=%d\n", bucket->id, + bucket->size, x, r, pr, s); + return bucket->items[s]; +} + +/* uniform */ +static int bucket_uniform_choose(struct crush_bucket_uniform *bucket, + int x, int r) +{ + return bucket_perm_choose(&bucket->h, x, r); +} + +/* list */ +static int bucket_list_choose(struct crush_bucket_list *bucket, + int x, int r) +{ + int i; + + for (i = bucket->h.size-1; i >= 0; i--) { + __u64 w = crush_hash32_4(bucket->h.hash,x, bucket->h.items[i], + r, bucket->h.id); + w &= 0xffff; + dprintk("list_choose i=%d x=%d r=%d item %d weight %x " + "sw %x rand %llx", + i, x, r, bucket->h.items[i], bucket->item_weights[i], + bucket->sum_weights[i], w); + w *= bucket->sum_weights[i]; + w = w >> 16; + /*dprintk(" scaled %llx\n", w);*/ + if (w < bucket->item_weights[i]) + return bucket->h.items[i]; + } + + BUG_ON(1); + return 0; +} + + +/* (binary) tree */ +static int height(int n) +{ + int h = 0; + while ((n & 1) == 0) { + h++; + n = n >> 1; + } + return h; +} + +static int left(int x) +{ + int h = height(x); + return x - (1 << (h-1)); +} + +static int right(int x) +{ + int h = height(x); + return x + (1 << (h-1)); +} + +static int terminal(int x) +{ + return x & 1; +} + +static int bucket_tree_choose(struct crush_bucket_tree *bucket, + int x, int r) +{ + int n, l; + __u32 w; + __u64 t; + + /* start at root */ + n = bucket->num_nodes >> 1; + + while (!terminal(n)) { + /* pick point in [0, w) */ + w = bucket->node_weights[n]; + t = (__u64)crush_hash32_4(bucket->h.hash, x, n, r, + bucket->h.id) * (__u64)w; + t = t >> 32; + + /* descend to the left or right? */ + l = left(n); + if (t < bucket->node_weights[l]) + n = l; + else + n = right(n); + } + + return bucket->h.items[n >> 1]; +} + + +/* straw */ + +static int bucket_straw_choose(struct crush_bucket_straw *bucket, + int x, int r) +{ + int i; + int high = 0; + __u64 high_draw = 0; + __u64 draw; + + for (i = 0; i < bucket->h.size; i++) { + draw = crush_hash32_3(bucket->h.hash, x, bucket->h.items[i], r); + draw &= 0xffff; + draw *= bucket->straws[i]; + if (i == 0 || draw > high_draw) { + high = i; + high_draw = draw; + } + } + return bucket->h.items[high]; +} + +static int crush_bucket_choose(struct crush_bucket *in, int x, int r) +{ + dprintk(" crush_bucket_choose %d x=%d r=%d\n", in->id, x, r); + switch (in->alg) { + case CRUSH_BUCKET_UNIFORM: + return bucket_uniform_choose((struct crush_bucket_uniform *)in, + x, r); + case CRUSH_BUCKET_LIST: + return bucket_list_choose((struct crush_bucket_list *)in, + x, r); + case CRUSH_BUCKET_TREE: + return bucket_tree_choose((struct crush_bucket_tree *)in, + x, r); + case CRUSH_BUCKET_STRAW: + return bucket_straw_choose((struct crush_bucket_straw *)in, + x, r); + default: + BUG_ON(1); + return in->items[0]; + } +} + +/* + * true if device is marked "out" (failed, fully offloaded) + * of the cluster + */ +static int is_out(struct crush_map *map, __u32 *weight, int item, int x) +{ + if (weight[item] >= 0x10000) + return 0; + if (weight[item] == 0) + return 1; + if ((crush_hash32_2(CRUSH_HASH_RJENKINS1, x, item) & 0xffff) + < weight[item]) + return 0; + return 1; +} + +/** + * crush_choose - choose numrep distinct items of given type + * @map: the crush_map + * @bucket: the bucket we are choose an item from + * @x: crush input value + * @numrep: the number of items to choose + * @type: the type of item to choose + * @out: pointer to output vector + * @outpos: our position in that vector + * @firstn: true if choosing "first n" items, false if choosing "indep" + * @recurse_to_leaf: true if we want one device under each item of given type + * @out2: second output vector for leaf items (if @recurse_to_leaf) + */ +static int crush_choose(struct crush_map *map, + struct crush_bucket *bucket, + __u32 *weight, + int x, int numrep, int type, + int *out, int outpos, + int firstn, int recurse_to_leaf, + int *out2) +{ + int rep; + int ftotal, flocal; + int retry_descent, retry_bucket, skip_rep; + struct crush_bucket *in = bucket; + int r; + int i; + int item = 0; + int itemtype; + int collide, reject; + const int orig_tries = 5; /* attempts before we fall back to search */ + + dprintk("CHOOSE%s bucket %d x %d outpos %d numrep %d\n", recurse_to_leaf ? "_LEAF" : "", + bucket->id, x, outpos, numrep); + + for (rep = outpos; rep < numrep; rep++) { + /* keep trying until we get a non-out, non-colliding item */ + ftotal = 0; + skip_rep = 0; + do { + retry_descent = 0; + in = bucket; /* initial bucket */ + + /* choose through intervening buckets */ + flocal = 0; + do { + collide = 0; + retry_bucket = 0; + r = rep; + if (in->alg == CRUSH_BUCKET_UNIFORM) { + /* be careful */ + if (firstn || numrep >= in->size) + /* r' = r + f_total */ + r += ftotal; + else if (in->size % numrep == 0) + /* r'=r+(n+1)*f_local */ + r += (numrep+1) * + (flocal+ftotal); + else + /* r' = r + n*f_local */ + r += numrep * (flocal+ftotal); + } else { + if (firstn) + /* r' = r + f_total */ + r += ftotal; + else + /* r' = r + n*f_local */ + r += numrep * (flocal+ftotal); + } + + /* bucket choose */ + if (in->size == 0) { + reject = 1; + goto reject; + } + if (flocal >= (in->size>>1) && + flocal > orig_tries) + item = bucket_perm_choose(in, x, r); + else + item = crush_bucket_choose(in, x, r); + BUG_ON(item >= map->max_devices); + + /* desired type? */ + if (item < 0) + itemtype = map->buckets[-1-item]->type; + else + itemtype = 0; + dprintk(" item %d type %d\n", item, itemtype); + + /* keep going? */ + if (itemtype != type) { + BUG_ON(item >= 0 || + (-1-item) >= map->max_buckets); + in = map->buckets[-1-item]; + retry_bucket = 1; + continue; + } + + /* collision? */ + for (i = 0; i < outpos; i++) { + if (out[i] == item) { + collide = 1; + break; + } + } + + reject = 0; + if (recurse_to_leaf) { + if (item < 0) { + if (crush_choose(map, + map->buckets[-1-item], + weight, + x, outpos+1, 0, + out2, outpos, + firstn, 0, + NULL) <= outpos) + /* didn't get leaf */ + reject = 1; + } else { + /* we already have a leaf! */ + out2[outpos] = item; + } + } + + if (!reject) { + /* out? */ + if (itemtype == 0) + reject = is_out(map, weight, + item, x); + else + reject = 0; + } + +reject: + if (reject || collide) { + ftotal++; + flocal++; + + if (collide && flocal < 3) + /* retry locally a few times */ + retry_bucket = 1; + else if (flocal < in->size + orig_tries) + /* exhaustive bucket search */ + retry_bucket = 1; + else if (ftotal < 20) + /* then retry descent */ + retry_descent = 1; + else + /* else give up */ + skip_rep = 1; + dprintk(" reject %d collide %d " + "ftotal %d flocal %d\n", + reject, collide, ftotal, + flocal); + } + } while (retry_bucket); + } while (retry_descent); + + if (skip_rep) { + dprintk("skip rep\n"); + continue; + } + + dprintk("CHOOSE got %d\n", item); + out[outpos] = item; + outpos++; + } + + dprintk("CHOOSE returns %d\n", outpos); + return outpos; +} + + +/** + * crush_do_rule - calculate a mapping with the given input and rule + * @map: the crush_map + * @ruleno: the rule id + * @x: hash input + * @result: pointer to result vector + * @result_max: maximum result size + * @force: force initial replica choice; -1 for none + */ +int crush_do_rule(struct crush_map *map, + int ruleno, int x, int *result, int result_max, + int force, __u32 *weight) +{ + int result_len; + int force_context[CRUSH_MAX_DEPTH]; + int force_pos = -1; + int a[CRUSH_MAX_SET]; + int b[CRUSH_MAX_SET]; + int c[CRUSH_MAX_SET]; + int recurse_to_leaf; + int *w; + int wsize = 0; + int *o; + int osize; + int *tmp; + struct crush_rule *rule; + int step; + int i, j; + int numrep; + int firstn; + int rc = -1; + + BUG_ON(ruleno >= map->max_rules); + + rule = map->rules[ruleno]; + result_len = 0; + w = a; + o = b; + + /* + * determine hierarchical context of force, if any. note + * that this may or may not correspond to the specific types + * referenced by the crush rule. + */ + if (force >= 0) { + if (force >= map->max_devices || + map->device_parents[force] == 0) { + /*dprintk("CRUSH: forcefed device dne\n");*/ + rc = -1; /* force fed device dne */ + goto out; + } + if (!is_out(map, weight, force, x)) { + while (1) { + force_context[++force_pos] = force; + if (force >= 0) + force = map->device_parents[force]; + else + force = map->bucket_parents[-1-force]; + if (force == 0) + break; + } + } + } + + for (step = 0; step < rule->len; step++) { + firstn = 0; + switch (rule->steps[step].op) { + case CRUSH_RULE_TAKE: + w[0] = rule->steps[step].arg1; + if (force_pos >= 0) { + BUG_ON(force_context[force_pos] != w[0]); + force_pos--; + } + wsize = 1; + break; + + case CRUSH_RULE_CHOOSE_LEAF_FIRSTN: + case CRUSH_RULE_CHOOSE_FIRSTN: + firstn = 1; + case CRUSH_RULE_CHOOSE_LEAF_INDEP: + case CRUSH_RULE_CHOOSE_INDEP: + BUG_ON(wsize == 0); + + recurse_to_leaf = + rule->steps[step].op == + CRUSH_RULE_CHOOSE_LEAF_FIRSTN || + rule->steps[step].op == + CRUSH_RULE_CHOOSE_LEAF_INDEP; + + /* reset output */ + osize = 0; + + for (i = 0; i < wsize; i++) { + /* + * see CRUSH_N, CRUSH_N_MINUS macros. + * basically, numrep <= 0 means relative to + * the provided result_max + */ + numrep = rule->steps[step].arg1; + if (numrep <= 0) { + numrep += result_max; + if (numrep <= 0) + continue; + } + j = 0; + if (osize == 0 && force_pos >= 0) { + /* skip any intermediate types */ + while (force_pos && + force_context[force_pos] < 0 && + rule->steps[step].arg2 != + map->buckets[-1 - + force_context[force_pos]]->type) + force_pos--; + o[osize] = force_context[force_pos]; + if (recurse_to_leaf) + c[osize] = force_context[0]; + j++; + force_pos--; + } + osize += crush_choose(map, + map->buckets[-1-w[i]], + weight, + x, numrep, + rule->steps[step].arg2, + o+osize, j, + firstn, + recurse_to_leaf, c+osize); + } + + if (recurse_to_leaf) + /* copy final _leaf_ values to output set */ + memcpy(o, c, osize*sizeof(*o)); + + /* swap t and w arrays */ + tmp = o; + o = w; + w = tmp; + wsize = osize; + break; + + + case CRUSH_RULE_EMIT: + for (i = 0; i < wsize && result_len < result_max; i++) { + result[result_len] = w[i]; + result_len++; + } + wsize = 0; + break; + + default: + BUG_ON(1); + } + } + rc = result_len; + +out: + return rc; +} + + diff --git a/net/ceph/crypto.c b/net/ceph/crypto.c new file mode 100644 index 0000000..7b505b0 --- /dev/null +++ b/net/ceph/crypto.c @@ -0,0 +1,412 @@ + +#include + +#include +#include +#include +#include + +#include +#include "crypto.h" + +int ceph_crypto_key_encode(struct ceph_crypto_key *key, void **p, void *end) +{ + if (*p + sizeof(u16) + sizeof(key->created) + + sizeof(u16) + key->len > end) + return -ERANGE; + ceph_encode_16(p, key->type); + ceph_encode_copy(p, &key->created, sizeof(key->created)); + ceph_encode_16(p, key->len); + ceph_encode_copy(p, key->key, key->len); + return 0; +} + +int ceph_crypto_key_decode(struct ceph_crypto_key *key, void **p, void *end) +{ + ceph_decode_need(p, end, 2*sizeof(u16) + sizeof(key->created), bad); + key->type = ceph_decode_16(p); + ceph_decode_copy(p, &key->created, sizeof(key->created)); + key->len = ceph_decode_16(p); + ceph_decode_need(p, end, key->len, bad); + key->key = kmalloc(key->len, GFP_NOFS); + if (!key->key) + return -ENOMEM; + ceph_decode_copy(p, key->key, key->len); + return 0; + +bad: + dout("failed to decode crypto key\n"); + return -EINVAL; +} + +int ceph_crypto_key_unarmor(struct ceph_crypto_key *key, const char *inkey) +{ + int inlen = strlen(inkey); + int blen = inlen * 3 / 4; + void *buf, *p; + int ret; + + dout("crypto_key_unarmor %s\n", inkey); + buf = kmalloc(blen, GFP_NOFS); + if (!buf) + return -ENOMEM; + blen = ceph_unarmor(buf, inkey, inkey+inlen); + if (blen < 0) { + kfree(buf); + return blen; + } + + p = buf; + ret = ceph_crypto_key_decode(key, &p, p + blen); + kfree(buf); + if (ret) + return ret; + dout("crypto_key_unarmor key %p type %d len %d\n", key, + key->type, key->len); + return 0; +} + + + +#define AES_KEY_SIZE 16 + +static struct crypto_blkcipher *ceph_crypto_alloc_cipher(void) +{ + return crypto_alloc_blkcipher("cbc(aes)", 0, CRYPTO_ALG_ASYNC); +} + +static const u8 *aes_iv = (u8 *)CEPH_AES_IV; + +static int ceph_aes_encrypt(const void *key, int key_len, + void *dst, size_t *dst_len, + const void *src, size_t src_len) +{ + struct scatterlist sg_in[2], sg_out[1]; + struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); + struct blkcipher_desc desc = { .tfm = tfm, .flags = 0 }; + int ret; + void *iv; + int ivsize; + size_t zero_padding = (0x10 - (src_len & 0x0f)); + char pad[16]; + + if (IS_ERR(tfm)) + return PTR_ERR(tfm); + + memset(pad, zero_padding, zero_padding); + + *dst_len = src_len + zero_padding; + + crypto_blkcipher_setkey((void *)tfm, key, key_len); + sg_init_table(sg_in, 2); + sg_set_buf(&sg_in[0], src, src_len); + sg_set_buf(&sg_in[1], pad, zero_padding); + sg_init_table(sg_out, 1); + sg_set_buf(sg_out, dst, *dst_len); + iv = crypto_blkcipher_crt(tfm)->iv; + ivsize = crypto_blkcipher_ivsize(tfm); + + memcpy(iv, aes_iv, ivsize); + /* + print_hex_dump(KERN_ERR, "enc key: ", DUMP_PREFIX_NONE, 16, 1, + key, key_len, 1); + print_hex_dump(KERN_ERR, "enc src: ", DUMP_PREFIX_NONE, 16, 1, + src, src_len, 1); + print_hex_dump(KERN_ERR, "enc pad: ", DUMP_PREFIX_NONE, 16, 1, + pad, zero_padding, 1); + */ + ret = crypto_blkcipher_encrypt(&desc, sg_out, sg_in, + src_len + zero_padding); + crypto_free_blkcipher(tfm); + if (ret < 0) + pr_err("ceph_aes_crypt failed %d\n", ret); + /* + print_hex_dump(KERN_ERR, "enc out: ", DUMP_PREFIX_NONE, 16, 1, + dst, *dst_len, 1); + */ + return 0; +} + +static int ceph_aes_encrypt2(const void *key, int key_len, void *dst, + size_t *dst_len, + const void *src1, size_t src1_len, + const void *src2, size_t src2_len) +{ + struct scatterlist sg_in[3], sg_out[1]; + struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); + struct blkcipher_desc desc = { .tfm = tfm, .flags = 0 }; + int ret; + void *iv; + int ivsize; + size_t zero_padding = (0x10 - ((src1_len + src2_len) & 0x0f)); + char pad[16]; + + if (IS_ERR(tfm)) + return PTR_ERR(tfm); + + memset(pad, zero_padding, zero_padding); + + *dst_len = src1_len + src2_len + zero_padding; + + crypto_blkcipher_setkey((void *)tfm, key, key_len); + sg_init_table(sg_in, 3); + sg_set_buf(&sg_in[0], src1, src1_len); + sg_set_buf(&sg_in[1], src2, src2_len); + sg_set_buf(&sg_in[2], pad, zero_padding); + sg_init_table(sg_out, 1); + sg_set_buf(sg_out, dst, *dst_len); + iv = crypto_blkcipher_crt(tfm)->iv; + ivsize = crypto_blkcipher_ivsize(tfm); + + memcpy(iv, aes_iv, ivsize); + /* + print_hex_dump(KERN_ERR, "enc key: ", DUMP_PREFIX_NONE, 16, 1, + key, key_len, 1); + print_hex_dump(KERN_ERR, "enc src1: ", DUMP_PREFIX_NONE, 16, 1, + src1, src1_len, 1); + print_hex_dump(KERN_ERR, "enc src2: ", DUMP_PREFIX_NONE, 16, 1, + src2, src2_len, 1); + print_hex_dump(KERN_ERR, "enc pad: ", DUMP_PREFIX_NONE, 16, 1, + pad, zero_padding, 1); + */ + ret = crypto_blkcipher_encrypt(&desc, sg_out, sg_in, + src1_len + src2_len + zero_padding); + crypto_free_blkcipher(tfm); + if (ret < 0) + pr_err("ceph_aes_crypt2 failed %d\n", ret); + /* + print_hex_dump(KERN_ERR, "enc out: ", DUMP_PREFIX_NONE, 16, 1, + dst, *dst_len, 1); + */ + return 0; +} + +static int ceph_aes_decrypt(const void *key, int key_len, + void *dst, size_t *dst_len, + const void *src, size_t src_len) +{ + struct scatterlist sg_in[1], sg_out[2]; + struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); + struct blkcipher_desc desc = { .tfm = tfm }; + char pad[16]; + void *iv; + int ivsize; + int ret; + int last_byte; + + if (IS_ERR(tfm)) + return PTR_ERR(tfm); + + crypto_blkcipher_setkey((void *)tfm, key, key_len); + sg_init_table(sg_in, 1); + sg_init_table(sg_out, 2); + sg_set_buf(sg_in, src, src_len); + sg_set_buf(&sg_out[0], dst, *dst_len); + sg_set_buf(&sg_out[1], pad, sizeof(pad)); + + iv = crypto_blkcipher_crt(tfm)->iv; + ivsize = crypto_blkcipher_ivsize(tfm); + + memcpy(iv, aes_iv, ivsize); + + /* + print_hex_dump(KERN_ERR, "dec key: ", DUMP_PREFIX_NONE, 16, 1, + key, key_len, 1); + print_hex_dump(KERN_ERR, "dec in: ", DUMP_PREFIX_NONE, 16, 1, + src, src_len, 1); + */ + + ret = crypto_blkcipher_decrypt(&desc, sg_out, sg_in, src_len); + crypto_free_blkcipher(tfm); + if (ret < 0) { + pr_err("ceph_aes_decrypt failed %d\n", ret); + return ret; + } + + if (src_len <= *dst_len) + last_byte = ((char *)dst)[src_len - 1]; + else + last_byte = pad[src_len - *dst_len - 1]; + if (last_byte <= 16 && src_len >= last_byte) { + *dst_len = src_len - last_byte; + } else { + pr_err("ceph_aes_decrypt got bad padding %d on src len %d\n", + last_byte, (int)src_len); + return -EPERM; /* bad padding */ + } + /* + print_hex_dump(KERN_ERR, "dec out: ", DUMP_PREFIX_NONE, 16, 1, + dst, *dst_len, 1); + */ + return 0; +} + +static int ceph_aes_decrypt2(const void *key, int key_len, + void *dst1, size_t *dst1_len, + void *dst2, size_t *dst2_len, + const void *src, size_t src_len) +{ + struct scatterlist sg_in[1], sg_out[3]; + struct crypto_blkcipher *tfm = ceph_crypto_alloc_cipher(); + struct blkcipher_desc desc = { .tfm = tfm }; + char pad[16]; + void *iv; + int ivsize; + int ret; + int last_byte; + + if (IS_ERR(tfm)) + return PTR_ERR(tfm); + + sg_init_table(sg_in, 1); + sg_set_buf(sg_in, src, src_len); + sg_init_table(sg_out, 3); + sg_set_buf(&sg_out[0], dst1, *dst1_len); + sg_set_buf(&sg_out[1], dst2, *dst2_len); + sg_set_buf(&sg_out[2], pad, sizeof(pad)); + + crypto_blkcipher_setkey((void *)tfm, key, key_len); + iv = crypto_blkcipher_crt(tfm)->iv; + ivsize = crypto_blkcipher_ivsize(tfm); + + memcpy(iv, aes_iv, ivsize); + + /* + print_hex_dump(KERN_ERR, "dec key: ", DUMP_PREFIX_NONE, 16, 1, + key, key_len, 1); + print_hex_dump(KERN_ERR, "dec in: ", DUMP_PREFIX_NONE, 16, 1, + src, src_len, 1); + */ + + ret = crypto_blkcipher_decrypt(&desc, sg_out, sg_in, src_len); + crypto_free_blkcipher(tfm); + if (ret < 0) { + pr_err("ceph_aes_decrypt failed %d\n", ret); + return ret; + } + + if (src_len <= *dst1_len) + last_byte = ((char *)dst1)[src_len - 1]; + else if (src_len <= *dst1_len + *dst2_len) + last_byte = ((char *)dst2)[src_len - *dst1_len - 1]; + else + last_byte = pad[src_len - *dst1_len - *dst2_len - 1]; + if (last_byte <= 16 && src_len >= last_byte) { + src_len -= last_byte; + } else { + pr_err("ceph_aes_decrypt got bad padding %d on src len %d\n", + last_byte, (int)src_len); + return -EPERM; /* bad padding */ + } + + if (src_len < *dst1_len) { + *dst1_len = src_len; + *dst2_len = 0; + } else { + *dst2_len = src_len - *dst1_len; + } + /* + print_hex_dump(KERN_ERR, "dec out1: ", DUMP_PREFIX_NONE, 16, 1, + dst1, *dst1_len, 1); + print_hex_dump(KERN_ERR, "dec out2: ", DUMP_PREFIX_NONE, 16, 1, + dst2, *dst2_len, 1); + */ + + return 0; +} + + +int ceph_decrypt(struct ceph_crypto_key *secret, void *dst, size_t *dst_len, + const void *src, size_t src_len) +{ + switch (secret->type) { + case CEPH_CRYPTO_NONE: + if (*dst_len < src_len) + return -ERANGE; + memcpy(dst, src, src_len); + *dst_len = src_len; + return 0; + + case CEPH_CRYPTO_AES: + return ceph_aes_decrypt(secret->key, secret->len, dst, + dst_len, src, src_len); + + default: + return -EINVAL; + } +} + +int ceph_decrypt2(struct ceph_crypto_key *secret, + void *dst1, size_t *dst1_len, + void *dst2, size_t *dst2_len, + const void *src, size_t src_len) +{ + size_t t; + + switch (secret->type) { + case CEPH_CRYPTO_NONE: + if (*dst1_len + *dst2_len < src_len) + return -ERANGE; + t = min(*dst1_len, src_len); + memcpy(dst1, src, t); + *dst1_len = t; + src += t; + src_len -= t; + if (src_len) { + t = min(*dst2_len, src_len); + memcpy(dst2, src, t); + *dst2_len = t; + } + return 0; + + case CEPH_CRYPTO_AES: + return ceph_aes_decrypt2(secret->key, secret->len, + dst1, dst1_len, dst2, dst2_len, + src, src_len); + + default: + return -EINVAL; + } +} + +int ceph_encrypt(struct ceph_crypto_key *secret, void *dst, size_t *dst_len, + const void *src, size_t src_len) +{ + switch (secret->type) { + case CEPH_CRYPTO_NONE: + if (*dst_len < src_len) + return -ERANGE; + memcpy(dst, src, src_len); + *dst_len = src_len; + return 0; + + case CEPH_CRYPTO_AES: + return ceph_aes_encrypt(secret->key, secret->len, dst, + dst_len, src, src_len); + + default: + return -EINVAL; + } +} + +int ceph_encrypt2(struct ceph_crypto_key *secret, void *dst, size_t *dst_len, + const void *src1, size_t src1_len, + const void *src2, size_t src2_len) +{ + switch (secret->type) { + case CEPH_CRYPTO_NONE: + if (*dst_len < src1_len + src2_len) + return -ERANGE; + memcpy(dst, src1, src1_len); + memcpy(dst + src1_len, src2, src2_len); + *dst_len = src1_len + src2_len; + return 0; + + case CEPH_CRYPTO_AES: + return ceph_aes_encrypt2(secret->key, secret->len, dst, dst_len, + src1, src1_len, src2, src2_len); + + default: + return -EINVAL; + } +} diff --git a/net/ceph/crypto.h b/net/ceph/crypto.h new file mode 100644 index 0000000..f9eccac --- /dev/null +++ b/net/ceph/crypto.h @@ -0,0 +1,48 @@ +#ifndef _FS_CEPH_CRYPTO_H +#define _FS_CEPH_CRYPTO_H + +#include +#include + +/* + * cryptographic secret + */ +struct ceph_crypto_key { + int type; + struct ceph_timespec created; + int len; + void *key; +}; + +static inline void ceph_crypto_key_destroy(struct ceph_crypto_key *key) +{ + kfree(key->key); +} + +extern int ceph_crypto_key_encode(struct ceph_crypto_key *key, + void **p, void *end); +extern int ceph_crypto_key_decode(struct ceph_crypto_key *key, + void **p, void *end); +extern int ceph_crypto_key_unarmor(struct ceph_crypto_key *key, const char *in); + +/* crypto.c */ +extern int ceph_decrypt(struct ceph_crypto_key *secret, + void *dst, size_t *dst_len, + const void *src, size_t src_len); +extern int ceph_encrypt(struct ceph_crypto_key *secret, + void *dst, size_t *dst_len, + const void *src, size_t src_len); +extern int ceph_decrypt2(struct ceph_crypto_key *secret, + void *dst1, size_t *dst1_len, + void *dst2, size_t *dst2_len, + const void *src, size_t src_len); +extern int ceph_encrypt2(struct ceph_crypto_key *secret, + void *dst, size_t *dst_len, + const void *src1, size_t src1_len, + const void *src2, size_t src2_len); + +/* armor.c */ +extern int ceph_armor(char *dst, const char *src, const char *end); +extern int ceph_unarmor(char *dst, const char *src, const char *end); + +#endif diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c new file mode 100644 index 0000000..33d0499 --- /dev/null +++ b/net/ceph/debugfs.c @@ -0,0 +1,268 @@ +#include + +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#ifdef CONFIG_DEBUG_FS + +/* + * Implement /sys/kernel/debug/ceph fun + * + * /sys/kernel/debug/ceph/client* - an instance of the ceph client + * .../osdmap - current osdmap + * .../monmap - current monmap + * .../osdc - active osd requests + * .../monc - mon client state + * .../dentry_lru - dump contents of dentry lru + * .../caps - expose cap (reservation) stats + * .../bdi - symlink to ../../bdi/something + */ + +static struct dentry *ceph_debugfs_dir; + +static int monmap_show(struct seq_file *s, void *p) +{ + int i; + struct ceph_client *client = s->private; + + if (client->monc.monmap == NULL) + return 0; + + seq_printf(s, "epoch %d\n", client->monc.monmap->epoch); + for (i = 0; i < client->monc.monmap->num_mon; i++) { + struct ceph_entity_inst *inst = + &client->monc.monmap->mon_inst[i]; + + seq_printf(s, "\t%s%lld\t%s\n", + ENTITY_NAME(inst->name), + ceph_pr_addr(&inst->addr.in_addr)); + } + return 0; +} + +static int osdmap_show(struct seq_file *s, void *p) +{ + int i; + struct ceph_client *client = s->private; + struct rb_node *n; + + if (client->osdc.osdmap == NULL) + return 0; + seq_printf(s, "epoch %d\n", client->osdc.osdmap->epoch); + seq_printf(s, "flags%s%s\n", + (client->osdc.osdmap->flags & CEPH_OSDMAP_NEARFULL) ? + " NEARFULL" : "", + (client->osdc.osdmap->flags & CEPH_OSDMAP_FULL) ? + " FULL" : ""); + for (n = rb_first(&client->osdc.osdmap->pg_pools); n; n = rb_next(n)) { + struct ceph_pg_pool_info *pool = + rb_entry(n, struct ceph_pg_pool_info, node); + seq_printf(s, "pg_pool %d pg_num %d / %d, lpg_num %d / %d\n", + pool->id, pool->v.pg_num, pool->pg_num_mask, + pool->v.lpg_num, pool->lpg_num_mask); + } + for (i = 0; i < client->osdc.osdmap->max_osd; i++) { + struct ceph_entity_addr *addr = + &client->osdc.osdmap->osd_addr[i]; + int state = client->osdc.osdmap->osd_state[i]; + char sb[64]; + + seq_printf(s, "\tosd%d\t%s\t%3d%%\t(%s)\n", + i, ceph_pr_addr(&addr->in_addr), + ((client->osdc.osdmap->osd_weight[i]*100) >> 16), + ceph_osdmap_state_str(sb, sizeof(sb), state)); + } + return 0; +} + +static int monc_show(struct seq_file *s, void *p) +{ + struct ceph_client *client = s->private; + struct ceph_mon_generic_request *req; + struct ceph_mon_client *monc = &client->monc; + struct rb_node *rp; + + mutex_lock(&monc->mutex); + + if (monc->have_mdsmap) + seq_printf(s, "have mdsmap %u\n", (unsigned)monc->have_mdsmap); + if (monc->have_osdmap) + seq_printf(s, "have osdmap %u\n", (unsigned)monc->have_osdmap); + if (monc->want_next_osdmap) + seq_printf(s, "want next osdmap\n"); + + for (rp = rb_first(&monc->generic_request_tree); rp; rp = rb_next(rp)) { + __u16 op; + req = rb_entry(rp, struct ceph_mon_generic_request, node); + op = le16_to_cpu(req->request->hdr.type); + if (op == CEPH_MSG_STATFS) + seq_printf(s, "%lld statfs\n", req->tid); + else + seq_printf(s, "%lld unknown\n", req->tid); + } + + mutex_unlock(&monc->mutex); + return 0; +} + +static int osdc_show(struct seq_file *s, void *pp) +{ + struct ceph_client *client = s->private; + struct ceph_osd_client *osdc = &client->osdc; + struct rb_node *p; + + mutex_lock(&osdc->request_mutex); + for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { + struct ceph_osd_request *req; + struct ceph_osd_request_head *head; + struct ceph_osd_op *op; + int num_ops; + int opcode, olen; + int i; + + req = rb_entry(p, struct ceph_osd_request, r_node); + + seq_printf(s, "%lld\tosd%d\t%d.%x\t", req->r_tid, + req->r_osd ? req->r_osd->o_osd : -1, + le32_to_cpu(req->r_pgid.pool), + le16_to_cpu(req->r_pgid.ps)); + + head = req->r_request->front.iov_base; + op = (void *)(head + 1); + + num_ops = le16_to_cpu(head->num_ops); + olen = le32_to_cpu(head->object_len); + seq_printf(s, "%.*s", olen, + (const char *)(head->ops + num_ops)); + + if (req->r_reassert_version.epoch) + seq_printf(s, "\t%u'%llu", + (unsigned)le32_to_cpu(req->r_reassert_version.epoch), + le64_to_cpu(req->r_reassert_version.version)); + else + seq_printf(s, "\t"); + + for (i = 0; i < num_ops; i++) { + opcode = le16_to_cpu(op->op); + seq_printf(s, "\t%s", ceph_osd_op_name(opcode)); + op++; + } + + seq_printf(s, "\n"); + } + mutex_unlock(&osdc->request_mutex); + return 0; +} + +CEPH_DEFINE_SHOW_FUNC(monmap_show) +CEPH_DEFINE_SHOW_FUNC(osdmap_show) +CEPH_DEFINE_SHOW_FUNC(monc_show) +CEPH_DEFINE_SHOW_FUNC(osdc_show) + +int ceph_debugfs_init(void) +{ + ceph_debugfs_dir = debugfs_create_dir("ceph", NULL); + if (!ceph_debugfs_dir) + return -ENOMEM; + return 0; +} + +void ceph_debugfs_cleanup(void) +{ + debugfs_remove(ceph_debugfs_dir); +} + +int ceph_debugfs_client_init(struct ceph_client *client) +{ + int ret = -ENOMEM; + char name[80]; + + snprintf(name, sizeof(name), "%pU.client%lld", &client->fsid, + client->monc.auth->global_id); + + client->debugfs_dir = debugfs_create_dir(name, ceph_debugfs_dir); + if (!client->debugfs_dir) + goto out; + + client->monc.debugfs_file = debugfs_create_file("monc", + 0600, + client->debugfs_dir, + client, + &monc_show_fops); + if (!client->monc.debugfs_file) + goto out; + + client->osdc.debugfs_file = debugfs_create_file("osdc", + 0600, + client->debugfs_dir, + client, + &osdc_show_fops); + if (!client->osdc.debugfs_file) + goto out; + + client->debugfs_monmap = debugfs_create_file("monmap", + 0600, + client->debugfs_dir, + client, + &monmap_show_fops); + if (!client->debugfs_monmap) + goto out; + + client->debugfs_osdmap = debugfs_create_file("osdmap", + 0600, + client->debugfs_dir, + client, + &osdmap_show_fops); + if (!client->debugfs_osdmap) + goto out; + + return 0; + +out: + ceph_debugfs_client_cleanup(client); + return ret; +} + +void ceph_debugfs_client_cleanup(struct ceph_client *client) +{ + debugfs_remove(client->debugfs_osdmap); + debugfs_remove(client->debugfs_monmap); + debugfs_remove(client->osdc.debugfs_file); + debugfs_remove(client->monc.debugfs_file); + debugfs_remove(client->debugfs_dir); +} + +#else /* CONFIG_DEBUG_FS */ + +int ceph_debugfs_init(void) +{ + return 0; +} + +void ceph_debugfs_cleanup(void) +{ +} + +int ceph_debugfs_client_init(struct ceph_client *client, + int (*module_debugfs_init)(struct ceph_client *)) +{ + return 0; +} + +void ceph_debugfs_client_cleanup(struct ceph_client *client) +{ +} + +#endif /* CONFIG_DEBUG_FS */ + +EXPORT_SYMBOL(ceph_debugfs_init); +EXPORT_SYMBOL(ceph_debugfs_cleanup); diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c new file mode 100644 index 0000000..0e8157e --- /dev/null +++ b/net/ceph/messenger.c @@ -0,0 +1,2453 @@ +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +/* + * Ceph uses the messenger to exchange ceph_msg messages with other + * hosts in the system. The messenger provides ordered and reliable + * delivery. We tolerate TCP disconnects by reconnecting (with + * exponential backoff) in the case of a fault (disconnection, bad + * crc, protocol error). Acks allow sent messages to be discarded by + * the sender. + */ + +/* static tag bytes (protocol control messages) */ +static char tag_msg = CEPH_MSGR_TAG_MSG; +static char tag_ack = CEPH_MSGR_TAG_ACK; +static char tag_keepalive = CEPH_MSGR_TAG_KEEPALIVE; + +#ifdef CONFIG_LOCKDEP +static struct lock_class_key socket_class; +#endif + + +static void queue_con(struct ceph_connection *con); +static void con_work(struct work_struct *); +static void ceph_fault(struct ceph_connection *con); + +/* + * nicely render a sockaddr as a string. + */ +#define MAX_ADDR_STR 20 +#define MAX_ADDR_STR_LEN 60 +static char addr_str[MAX_ADDR_STR][MAX_ADDR_STR_LEN]; +static DEFINE_SPINLOCK(addr_str_lock); +static int last_addr_str; + +const char *ceph_pr_addr(const struct sockaddr_storage *ss) +{ + int i; + char *s; + struct sockaddr_in *in4 = (void *)ss; + struct sockaddr_in6 *in6 = (void *)ss; + + spin_lock(&addr_str_lock); + i = last_addr_str++; + if (last_addr_str == MAX_ADDR_STR) + last_addr_str = 0; + spin_unlock(&addr_str_lock); + s = addr_str[i]; + + switch (ss->ss_family) { + case AF_INET: + snprintf(s, MAX_ADDR_STR_LEN, "%pI4:%u", &in4->sin_addr, + (unsigned int)ntohs(in4->sin_port)); + break; + + case AF_INET6: + snprintf(s, MAX_ADDR_STR_LEN, "[%pI6c]:%u", &in6->sin6_addr, + (unsigned int)ntohs(in6->sin6_port)); + break; + + default: + sprintf(s, "(unknown sockaddr family %d)", (int)ss->ss_family); + } + + return s; +} +EXPORT_SYMBOL(ceph_pr_addr); + +static void encode_my_addr(struct ceph_messenger *msgr) +{ + memcpy(&msgr->my_enc_addr, &msgr->inst.addr, sizeof(msgr->my_enc_addr)); + ceph_encode_addr(&msgr->my_enc_addr); +} + +/* + * work queue for all reading and writing to/from the socket. + */ +struct workqueue_struct *ceph_msgr_wq; + +int ceph_msgr_init(void) +{ + ceph_msgr_wq = create_workqueue("ceph-msgr"); + if (IS_ERR(ceph_msgr_wq)) { + int ret = PTR_ERR(ceph_msgr_wq); + pr_err("msgr_init failed to create workqueue: %d\n", ret); + ceph_msgr_wq = NULL; + return ret; + } + return 0; +} +EXPORT_SYMBOL(ceph_msgr_init); + +void ceph_msgr_exit(void) +{ + destroy_workqueue(ceph_msgr_wq); +} +EXPORT_SYMBOL(ceph_msgr_exit); + +void ceph_msgr_flush(void) +{ + flush_workqueue(ceph_msgr_wq); +} +EXPORT_SYMBOL(ceph_msgr_flush); + + +/* + * socket callback functions + */ + +/* data available on socket, or listen socket received a connect */ +static void ceph_data_ready(struct sock *sk, int count_unused) +{ + struct ceph_connection *con = + (struct ceph_connection *)sk->sk_user_data; + if (sk->sk_state != TCP_CLOSE_WAIT) { + dout("ceph_data_ready on %p state = %lu, queueing work\n", + con, con->state); + queue_con(con); + } +} + +/* socket has buffer space for writing */ +static void ceph_write_space(struct sock *sk) +{ + struct ceph_connection *con = + (struct ceph_connection *)sk->sk_user_data; + + /* only queue to workqueue if there is data we want to write. */ + if (test_bit(WRITE_PENDING, &con->state)) { + dout("ceph_write_space %p queueing write work\n", con); + queue_con(con); + } else { + dout("ceph_write_space %p nothing to write\n", con); + } + + /* since we have our own write_space, clear the SOCK_NOSPACE flag */ + clear_bit(SOCK_NOSPACE, &sk->sk_socket->flags); +} + +/* socket's state has changed */ +static void ceph_state_change(struct sock *sk) +{ + struct ceph_connection *con = + (struct ceph_connection *)sk->sk_user_data; + + dout("ceph_state_change %p state = %lu sk_state = %u\n", + con, con->state, sk->sk_state); + + if (test_bit(CLOSED, &con->state)) + return; + + switch (sk->sk_state) { + case TCP_CLOSE: + dout("ceph_state_change TCP_CLOSE\n"); + case TCP_CLOSE_WAIT: + dout("ceph_state_change TCP_CLOSE_WAIT\n"); + if (test_and_set_bit(SOCK_CLOSED, &con->state) == 0) { + if (test_bit(CONNECTING, &con->state)) + con->error_msg = "connection failed"; + else + con->error_msg = "socket closed"; + queue_con(con); + } + break; + case TCP_ESTABLISHED: + dout("ceph_state_change TCP_ESTABLISHED\n"); + queue_con(con); + break; + } +} + +/* + * set up socket callbacks + */ +static void set_sock_callbacks(struct socket *sock, + struct ceph_connection *con) +{ + struct sock *sk = sock->sk; + sk->sk_user_data = (void *)con; + sk->sk_data_ready = ceph_data_ready; + sk->sk_write_space = ceph_write_space; + sk->sk_state_change = ceph_state_change; +} + + +/* + * socket helpers + */ + +/* + * initiate connection to a remote socket. + */ +static struct socket *ceph_tcp_connect(struct ceph_connection *con) +{ + struct sockaddr_storage *paddr = &con->peer_addr.in_addr; + struct socket *sock; + int ret; + + BUG_ON(con->sock); + ret = sock_create_kern(con->peer_addr.in_addr.ss_family, SOCK_STREAM, + IPPROTO_TCP, &sock); + if (ret) + return ERR_PTR(ret); + con->sock = sock; + sock->sk->sk_allocation = GFP_NOFS; + +#ifdef CONFIG_LOCKDEP + lockdep_set_class(&sock->sk->sk_lock, &socket_class); +#endif + + set_sock_callbacks(sock, con); + + dout("connect %s\n", ceph_pr_addr(&con->peer_addr.in_addr)); + + ret = sock->ops->connect(sock, (struct sockaddr *)paddr, sizeof(*paddr), + O_NONBLOCK); + if (ret == -EINPROGRESS) { + dout("connect %s EINPROGRESS sk_state = %u\n", + ceph_pr_addr(&con->peer_addr.in_addr), + sock->sk->sk_state); + ret = 0; + } + if (ret < 0) { + pr_err("connect %s error %d\n", + ceph_pr_addr(&con->peer_addr.in_addr), ret); + sock_release(sock); + con->sock = NULL; + con->error_msg = "connect error"; + } + + if (ret < 0) + return ERR_PTR(ret); + return sock; +} + +static int ceph_tcp_recvmsg(struct socket *sock, void *buf, size_t len) +{ + struct kvec iov = {buf, len}; + struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL }; + + return kernel_recvmsg(sock, &msg, &iov, 1, len, msg.msg_flags); +} + +/* + * write something. @more is true if caller will be sending more data + * shortly. + */ +static int ceph_tcp_sendmsg(struct socket *sock, struct kvec *iov, + size_t kvlen, size_t len, int more) +{ + struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL }; + + if (more) + msg.msg_flags |= MSG_MORE; + else + msg.msg_flags |= MSG_EOR; /* superfluous, but what the hell */ + + return kernel_sendmsg(sock, &msg, iov, kvlen, len); +} + + +/* + * Shutdown/close the socket for the given connection. + */ +static int con_close_socket(struct ceph_connection *con) +{ + int rc; + + dout("con_close_socket on %p sock %p\n", con, con->sock); + if (!con->sock) + return 0; + set_bit(SOCK_CLOSED, &con->state); + rc = con->sock->ops->shutdown(con->sock, SHUT_RDWR); + sock_release(con->sock); + con->sock = NULL; + clear_bit(SOCK_CLOSED, &con->state); + return rc; +} + +/* + * Reset a connection. Discard all incoming and outgoing messages + * and clear *_seq state. + */ +static void ceph_msg_remove(struct ceph_msg *msg) +{ + list_del_init(&msg->list_head); + ceph_msg_put(msg); +} +static void ceph_msg_remove_list(struct list_head *head) +{ + while (!list_empty(head)) { + struct ceph_msg *msg = list_first_entry(head, struct ceph_msg, + list_head); + ceph_msg_remove(msg); + } +} + +static void reset_connection(struct ceph_connection *con) +{ + /* reset connection, out_queue, msg_ and connect_seq */ + /* discard existing out_queue and msg_seq */ + ceph_msg_remove_list(&con->out_queue); + ceph_msg_remove_list(&con->out_sent); + + if (con->in_msg) { + ceph_msg_put(con->in_msg); + con->in_msg = NULL; + } + + con->connect_seq = 0; + con->out_seq = 0; + if (con->out_msg) { + ceph_msg_put(con->out_msg); + con->out_msg = NULL; + } + con->out_keepalive_pending = false; + con->in_seq = 0; + con->in_seq_acked = 0; +} + +/* + * mark a peer down. drop any open connections. + */ +void ceph_con_close(struct ceph_connection *con) +{ + dout("con_close %p peer %s\n", con, + ceph_pr_addr(&con->peer_addr.in_addr)); + set_bit(CLOSED, &con->state); /* in case there's queued work */ + clear_bit(STANDBY, &con->state); /* avoid connect_seq bump */ + clear_bit(LOSSYTX, &con->state); /* so we retry next connect */ + clear_bit(KEEPALIVE_PENDING, &con->state); + clear_bit(WRITE_PENDING, &con->state); + mutex_lock(&con->mutex); + reset_connection(con); + con->peer_global_seq = 0; + cancel_delayed_work(&con->work); + mutex_unlock(&con->mutex); + queue_con(con); +} +EXPORT_SYMBOL(ceph_con_close); + +/* + * Reopen a closed connection, with a new peer address. + */ +void ceph_con_open(struct ceph_connection *con, struct ceph_entity_addr *addr) +{ + dout("con_open %p %s\n", con, ceph_pr_addr(&addr->in_addr)); + set_bit(OPENING, &con->state); + clear_bit(CLOSED, &con->state); + memcpy(&con->peer_addr, addr, sizeof(*addr)); + con->delay = 0; /* reset backoff memory */ + queue_con(con); +} +EXPORT_SYMBOL(ceph_con_open); + +/* + * return true if this connection ever successfully opened + */ +bool ceph_con_opened(struct ceph_connection *con) +{ + return con->connect_seq > 0; +} + +/* + * generic get/put + */ +struct ceph_connection *ceph_con_get(struct ceph_connection *con) +{ + dout("con_get %p nref = %d -> %d\n", con, + atomic_read(&con->nref), atomic_read(&con->nref) + 1); + if (atomic_inc_not_zero(&con->nref)) + return con; + return NULL; +} + +void ceph_con_put(struct ceph_connection *con) +{ + dout("con_put %p nref = %d -> %d\n", con, + atomic_read(&con->nref), atomic_read(&con->nref) - 1); + BUG_ON(atomic_read(&con->nref) == 0); + if (atomic_dec_and_test(&con->nref)) { + BUG_ON(con->sock); + kfree(con); + } +} + +/* + * initialize a new connection. + */ +void ceph_con_init(struct ceph_messenger *msgr, struct ceph_connection *con) +{ + dout("con_init %p\n", con); + memset(con, 0, sizeof(*con)); + atomic_set(&con->nref, 1); + con->msgr = msgr; + mutex_init(&con->mutex); + INIT_LIST_HEAD(&con->out_queue); + INIT_LIST_HEAD(&con->out_sent); + INIT_DELAYED_WORK(&con->work, con_work); +} +EXPORT_SYMBOL(ceph_con_init); + + +/* + * We maintain a global counter to order connection attempts. Get + * a unique seq greater than @gt. + */ +static u32 get_global_seq(struct ceph_messenger *msgr, u32 gt) +{ + u32 ret; + + spin_lock(&msgr->global_seq_lock); + if (msgr->global_seq < gt) + msgr->global_seq = gt; + ret = ++msgr->global_seq; + spin_unlock(&msgr->global_seq_lock); + return ret; +} + + +/* + * Prepare footer for currently outgoing message, and finish things + * off. Assumes out_kvec* are already valid.. we just add on to the end. + */ +static void prepare_write_message_footer(struct ceph_connection *con, int v) +{ + struct ceph_msg *m = con->out_msg; + + dout("prepare_write_message_footer %p\n", con); + con->out_kvec_is_msg = true; + con->out_kvec[v].iov_base = &m->footer; + con->out_kvec[v].iov_len = sizeof(m->footer); + con->out_kvec_bytes += sizeof(m->footer); + con->out_kvec_left++; + con->out_more = m->more_to_follow; + con->out_msg_done = true; +} + +/* + * Prepare headers for the next outgoing message. + */ +static void prepare_write_message(struct ceph_connection *con) +{ + struct ceph_msg *m; + int v = 0; + + con->out_kvec_bytes = 0; + con->out_kvec_is_msg = true; + con->out_msg_done = false; + + /* Sneak an ack in there first? If we can get it into the same + * TCP packet that's a good thing. */ + if (con->in_seq > con->in_seq_acked) { + con->in_seq_acked = con->in_seq; + con->out_kvec[v].iov_base = &tag_ack; + con->out_kvec[v++].iov_len = 1; + con->out_temp_ack = cpu_to_le64(con->in_seq_acked); + con->out_kvec[v].iov_base = &con->out_temp_ack; + con->out_kvec[v++].iov_len = sizeof(con->out_temp_ack); + con->out_kvec_bytes = 1 + sizeof(con->out_temp_ack); + } + + m = list_first_entry(&con->out_queue, + struct ceph_msg, list_head); + con->out_msg = m; + if (test_bit(LOSSYTX, &con->state)) { + list_del_init(&m->list_head); + } else { + /* put message on sent list */ + ceph_msg_get(m); + list_move_tail(&m->list_head, &con->out_sent); + } + + /* + * only assign outgoing seq # if we haven't sent this message + * yet. if it is requeued, resend with it's original seq. + */ + if (m->needs_out_seq) { + m->hdr.seq = cpu_to_le64(++con->out_seq); + m->needs_out_seq = false; + } + + dout("prepare_write_message %p seq %lld type %d len %d+%d+%d %d pgs\n", + m, con->out_seq, le16_to_cpu(m->hdr.type), + le32_to_cpu(m->hdr.front_len), le32_to_cpu(m->hdr.middle_len), + le32_to_cpu(m->hdr.data_len), + m->nr_pages); + BUG_ON(le32_to_cpu(m->hdr.front_len) != m->front.iov_len); + + /* tag + hdr + front + middle */ + con->out_kvec[v].iov_base = &tag_msg; + con->out_kvec[v++].iov_len = 1; + con->out_kvec[v].iov_base = &m->hdr; + con->out_kvec[v++].iov_len = sizeof(m->hdr); + con->out_kvec[v++] = m->front; + if (m->middle) + con->out_kvec[v++] = m->middle->vec; + con->out_kvec_left = v; + con->out_kvec_bytes += 1 + sizeof(m->hdr) + m->front.iov_len + + (m->middle ? m->middle->vec.iov_len : 0); + con->out_kvec_cur = con->out_kvec; + + /* fill in crc (except data pages), footer */ + con->out_msg->hdr.crc = + cpu_to_le32(crc32c(0, (void *)&m->hdr, + sizeof(m->hdr) - sizeof(m->hdr.crc))); + con->out_msg->footer.flags = CEPH_MSG_FOOTER_COMPLETE; + con->out_msg->footer.front_crc = + cpu_to_le32(crc32c(0, m->front.iov_base, m->front.iov_len)); + if (m->middle) + con->out_msg->footer.middle_crc = + cpu_to_le32(crc32c(0, m->middle->vec.iov_base, + m->middle->vec.iov_len)); + else + con->out_msg->footer.middle_crc = 0; + con->out_msg->footer.data_crc = 0; + dout("prepare_write_message front_crc %u data_crc %u\n", + le32_to_cpu(con->out_msg->footer.front_crc), + le32_to_cpu(con->out_msg->footer.middle_crc)); + + /* is there a data payload? */ + if (le32_to_cpu(m->hdr.data_len) > 0) { + /* initialize page iterator */ + con->out_msg_pos.page = 0; + if (m->pages) + con->out_msg_pos.page_pos = + le16_to_cpu(m->hdr.data_off) & ~PAGE_MASK; + else + con->out_msg_pos.page_pos = 0; + con->out_msg_pos.data_pos = 0; + con->out_msg_pos.did_page_crc = 0; + con->out_more = 1; /* data + footer will follow */ + } else { + /* no, queue up footer too and be done */ + prepare_write_message_footer(con, v); + } + + set_bit(WRITE_PENDING, &con->state); +} + +/* + * Prepare an ack. + */ +static void prepare_write_ack(struct ceph_connection *con) +{ + dout("prepare_write_ack %p %llu -> %llu\n", con, + con->in_seq_acked, con->in_seq); + con->in_seq_acked = con->in_seq; + + con->out_kvec[0].iov_base = &tag_ack; + con->out_kvec[0].iov_len = 1; + con->out_temp_ack = cpu_to_le64(con->in_seq_acked); + con->out_kvec[1].iov_base = &con->out_temp_ack; + con->out_kvec[1].iov_len = sizeof(con->out_temp_ack); + con->out_kvec_left = 2; + con->out_kvec_bytes = 1 + sizeof(con->out_temp_ack); + con->out_kvec_cur = con->out_kvec; + con->out_more = 1; /* more will follow.. eventually.. */ + set_bit(WRITE_PENDING, &con->state); +} + +/* + * Prepare to write keepalive byte. + */ +static void prepare_write_keepalive(struct ceph_connection *con) +{ + dout("prepare_write_keepalive %p\n", con); + con->out_kvec[0].iov_base = &tag_keepalive; + con->out_kvec[0].iov_len = 1; + con->out_kvec_left = 1; + con->out_kvec_bytes = 1; + con->out_kvec_cur = con->out_kvec; + set_bit(WRITE_PENDING, &con->state); +} + +/* + * Connection negotiation. + */ + +static void prepare_connect_authorizer(struct ceph_connection *con) +{ + void *auth_buf; + int auth_len = 0; + int auth_protocol = 0; + + mutex_unlock(&con->mutex); + if (con->ops->get_authorizer) + con->ops->get_authorizer(con, &auth_buf, &auth_len, + &auth_protocol, &con->auth_reply_buf, + &con->auth_reply_buf_len, + con->auth_retry); + mutex_lock(&con->mutex); + + con->out_connect.authorizer_protocol = cpu_to_le32(auth_protocol); + con->out_connect.authorizer_len = cpu_to_le32(auth_len); + + con->out_kvec[con->out_kvec_left].iov_base = auth_buf; + con->out_kvec[con->out_kvec_left].iov_len = auth_len; + con->out_kvec_left++; + con->out_kvec_bytes += auth_len; +} + +/* + * We connected to a peer and are saying hello. + */ +static void prepare_write_banner(struct ceph_messenger *msgr, + struct ceph_connection *con) +{ + int len = strlen(CEPH_BANNER); + + con->out_kvec[0].iov_base = CEPH_BANNER; + con->out_kvec[0].iov_len = len; + con->out_kvec[1].iov_base = &msgr->my_enc_addr; + con->out_kvec[1].iov_len = sizeof(msgr->my_enc_addr); + con->out_kvec_left = 2; + con->out_kvec_bytes = len + sizeof(msgr->my_enc_addr); + con->out_kvec_cur = con->out_kvec; + con->out_more = 0; + set_bit(WRITE_PENDING, &con->state); +} + +static void prepare_write_connect(struct ceph_messenger *msgr, + struct ceph_connection *con, + int after_banner) +{ + unsigned global_seq = get_global_seq(con->msgr, 0); + int proto; + + switch (con->peer_name.type) { + case CEPH_ENTITY_TYPE_MON: + proto = CEPH_MONC_PROTOCOL; + break; + case CEPH_ENTITY_TYPE_OSD: + proto = CEPH_OSDC_PROTOCOL; + break; + case CEPH_ENTITY_TYPE_MDS: + proto = CEPH_MDSC_PROTOCOL; + break; + default: + BUG(); + } + + dout("prepare_write_connect %p cseq=%d gseq=%d proto=%d\n", con, + con->connect_seq, global_seq, proto); + + con->out_connect.features = cpu_to_le64(msgr->supported_features); + con->out_connect.host_type = cpu_to_le32(CEPH_ENTITY_TYPE_CLIENT); + con->out_connect.connect_seq = cpu_to_le32(con->connect_seq); + con->out_connect.global_seq = cpu_to_le32(global_seq); + con->out_connect.protocol_version = cpu_to_le32(proto); + con->out_connect.flags = 0; + + if (!after_banner) { + con->out_kvec_left = 0; + con->out_kvec_bytes = 0; + } + con->out_kvec[con->out_kvec_left].iov_base = &con->out_connect; + con->out_kvec[con->out_kvec_left].iov_len = sizeof(con->out_connect); + con->out_kvec_left++; + con->out_kvec_bytes += sizeof(con->out_connect); + con->out_kvec_cur = con->out_kvec; + con->out_more = 0; + set_bit(WRITE_PENDING, &con->state); + + prepare_connect_authorizer(con); +} + + +/* + * write as much of pending kvecs to the socket as we can. + * 1 -> done + * 0 -> socket full, but more to do + * <0 -> error + */ +static int write_partial_kvec(struct ceph_connection *con) +{ + int ret; + + dout("write_partial_kvec %p %d left\n", con, con->out_kvec_bytes); + while (con->out_kvec_bytes > 0) { + ret = ceph_tcp_sendmsg(con->sock, con->out_kvec_cur, + con->out_kvec_left, con->out_kvec_bytes, + con->out_more); + if (ret <= 0) + goto out; + con->out_kvec_bytes -= ret; + if (con->out_kvec_bytes == 0) + break; /* done */ + while (ret > 0) { + if (ret >= con->out_kvec_cur->iov_len) { + ret -= con->out_kvec_cur->iov_len; + con->out_kvec_cur++; + con->out_kvec_left--; + } else { + con->out_kvec_cur->iov_len -= ret; + con->out_kvec_cur->iov_base += ret; + ret = 0; + break; + } + } + } + con->out_kvec_left = 0; + con->out_kvec_is_msg = false; + ret = 1; +out: + dout("write_partial_kvec %p %d left in %d kvecs ret = %d\n", con, + con->out_kvec_bytes, con->out_kvec_left, ret); + return ret; /* done! */ +} + +#ifdef CONFIG_BLOCK +static void init_bio_iter(struct bio *bio, struct bio **iter, int *seg) +{ + if (!bio) { + *iter = NULL; + *seg = 0; + return; + } + *iter = bio; + *seg = bio->bi_idx; +} + +static void iter_bio_next(struct bio **bio_iter, int *seg) +{ + if (*bio_iter == NULL) + return; + + BUG_ON(*seg >= (*bio_iter)->bi_vcnt); + + (*seg)++; + if (*seg == (*bio_iter)->bi_vcnt) + init_bio_iter((*bio_iter)->bi_next, bio_iter, seg); +} +#endif + +/* + * Write as much message data payload as we can. If we finish, queue + * up the footer. + * 1 -> done, footer is now queued in out_kvec[]. + * 0 -> socket full, but more to do + * <0 -> error + */ +static int write_partial_msg_pages(struct ceph_connection *con) +{ + struct ceph_msg *msg = con->out_msg; + unsigned data_len = le32_to_cpu(msg->hdr.data_len); + size_t len; + int crc = con->msgr->nocrc; + int ret; + int total_max_write; + int in_trail = 0; + size_t trail_len = (msg->trail ? msg->trail->length : 0); + + dout("write_partial_msg_pages %p msg %p page %d/%d offset %d\n", + con, con->out_msg, con->out_msg_pos.page, con->out_msg->nr_pages, + con->out_msg_pos.page_pos); + +#ifdef CONFIG_BLOCK + if (msg->bio && !msg->bio_iter) + init_bio_iter(msg->bio, &msg->bio_iter, &msg->bio_seg); +#endif + + while (data_len > con->out_msg_pos.data_pos) { + struct page *page = NULL; + void *kaddr = NULL; + int max_write = PAGE_SIZE; + int page_shift = 0; + + total_max_write = data_len - trail_len - + con->out_msg_pos.data_pos; + + /* + * if we are calculating the data crc (the default), we need + * to map the page. if our pages[] has been revoked, use the + * zero page. + */ + + /* have we reached the trail part of the data? */ + if (con->out_msg_pos.data_pos >= data_len - trail_len) { + in_trail = 1; + + total_max_write = data_len - con->out_msg_pos.data_pos; + + page = list_first_entry(&msg->trail->head, + struct page, lru); + if (crc) + kaddr = kmap(page); + max_write = PAGE_SIZE; + } else if (msg->pages) { + page = msg->pages[con->out_msg_pos.page]; + if (crc) + kaddr = kmap(page); + } else if (msg->pagelist) { + page = list_first_entry(&msg->pagelist->head, + struct page, lru); + if (crc) + kaddr = kmap(page); +#ifdef CONFIG_BLOCK + } else if (msg->bio) { + struct bio_vec *bv; + + bv = bio_iovec_idx(msg->bio_iter, msg->bio_seg); + page = bv->bv_page; + page_shift = bv->bv_offset; + if (crc) + kaddr = kmap(page) + page_shift; + max_write = bv->bv_len; +#endif + } else { + page = con->msgr->zero_page; + if (crc) + kaddr = page_address(con->msgr->zero_page); + } + len = min_t(int, max_write - con->out_msg_pos.page_pos, + total_max_write); + + if (crc && !con->out_msg_pos.did_page_crc) { + void *base = kaddr + con->out_msg_pos.page_pos; + u32 tmpcrc = le32_to_cpu(con->out_msg->footer.data_crc); + + BUG_ON(kaddr == NULL); + con->out_msg->footer.data_crc = + cpu_to_le32(crc32c(tmpcrc, base, len)); + con->out_msg_pos.did_page_crc = 1; + } + ret = kernel_sendpage(con->sock, page, + con->out_msg_pos.page_pos + page_shift, + len, + MSG_DONTWAIT | MSG_NOSIGNAL | + MSG_MORE); + + if (crc && + (msg->pages || msg->pagelist || msg->bio || in_trail)) + kunmap(page); + + if (ret <= 0) + goto out; + + con->out_msg_pos.data_pos += ret; + con->out_msg_pos.page_pos += ret; + if (ret == len) { + con->out_msg_pos.page_pos = 0; + con->out_msg_pos.page++; + con->out_msg_pos.did_page_crc = 0; + if (in_trail) + list_move_tail(&page->lru, + &msg->trail->head); + else if (msg->pagelist) + list_move_tail(&page->lru, + &msg->pagelist->head); +#ifdef CONFIG_BLOCK + else if (msg->bio) + iter_bio_next(&msg->bio_iter, &msg->bio_seg); +#endif + } + } + + dout("write_partial_msg_pages %p msg %p done\n", con, msg); + + /* prepare and queue up footer, too */ + if (!crc) + con->out_msg->footer.flags |= CEPH_MSG_FOOTER_NOCRC; + con->out_kvec_bytes = 0; + con->out_kvec_left = 0; + con->out_kvec_cur = con->out_kvec; + prepare_write_message_footer(con, 0); + ret = 1; +out: + return ret; +} + +/* + * write some zeros + */ +static int write_partial_skip(struct ceph_connection *con) +{ + int ret; + + while (con->out_skip > 0) { + struct kvec iov = { + .iov_base = page_address(con->msgr->zero_page), + .iov_len = min(con->out_skip, (int)PAGE_CACHE_SIZE) + }; + + ret = ceph_tcp_sendmsg(con->sock, &iov, 1, iov.iov_len, 1); + if (ret <= 0) + goto out; + con->out_skip -= ret; + } + ret = 1; +out: + return ret; +} + +/* + * Prepare to read connection handshake, or an ack. + */ +static void prepare_read_banner(struct ceph_connection *con) +{ + dout("prepare_read_banner %p\n", con); + con->in_base_pos = 0; +} + +static void prepare_read_connect(struct ceph_connection *con) +{ + dout("prepare_read_connect %p\n", con); + con->in_base_pos = 0; +} + +static void prepare_read_ack(struct ceph_connection *con) +{ + dout("prepare_read_ack %p\n", con); + con->in_base_pos = 0; +} + +static void prepare_read_tag(struct ceph_connection *con) +{ + dout("prepare_read_tag %p\n", con); + con->in_base_pos = 0; + con->in_tag = CEPH_MSGR_TAG_READY; +} + +/* + * Prepare to read a message. + */ +static int prepare_read_message(struct ceph_connection *con) +{ + dout("prepare_read_message %p\n", con); + BUG_ON(con->in_msg != NULL); + con->in_base_pos = 0; + con->in_front_crc = con->in_middle_crc = con->in_data_crc = 0; + return 0; +} + + +static int read_partial(struct ceph_connection *con, + int *to, int size, void *object) +{ + *to += size; + while (con->in_base_pos < *to) { + int left = *to - con->in_base_pos; + int have = size - left; + int ret = ceph_tcp_recvmsg(con->sock, object + have, left); + if (ret <= 0) + return ret; + con->in_base_pos += ret; + } + return 1; +} + + +/* + * Read all or part of the connect-side handshake on a new connection + */ +static int read_partial_banner(struct ceph_connection *con) +{ + int ret, to = 0; + + dout("read_partial_banner %p at %d\n", con, con->in_base_pos); + + /* peer's banner */ + ret = read_partial(con, &to, strlen(CEPH_BANNER), con->in_banner); + if (ret <= 0) + goto out; + ret = read_partial(con, &to, sizeof(con->actual_peer_addr), + &con->actual_peer_addr); + if (ret <= 0) + goto out; + ret = read_partial(con, &to, sizeof(con->peer_addr_for_me), + &con->peer_addr_for_me); + if (ret <= 0) + goto out; +out: + return ret; +} + +static int read_partial_connect(struct ceph_connection *con) +{ + int ret, to = 0; + + dout("read_partial_connect %p at %d\n", con, con->in_base_pos); + + ret = read_partial(con, &to, sizeof(con->in_reply), &con->in_reply); + if (ret <= 0) + goto out; + ret = read_partial(con, &to, le32_to_cpu(con->in_reply.authorizer_len), + con->auth_reply_buf); + if (ret <= 0) + goto out; + + dout("read_partial_connect %p tag %d, con_seq = %u, g_seq = %u\n", + con, (int)con->in_reply.tag, + le32_to_cpu(con->in_reply.connect_seq), + le32_to_cpu(con->in_reply.global_seq)); +out: + return ret; + +} + +/* + * Verify the hello banner looks okay. + */ +static int verify_hello(struct ceph_connection *con) +{ + if (memcmp(con->in_banner, CEPH_BANNER, strlen(CEPH_BANNER))) { + pr_err("connect to %s got bad banner\n", + ceph_pr_addr(&con->peer_addr.in_addr)); + con->error_msg = "protocol error, bad banner"; + return -1; + } + return 0; +} + +static bool addr_is_blank(struct sockaddr_storage *ss) +{ + switch (ss->ss_family) { + case AF_INET: + return ((struct sockaddr_in *)ss)->sin_addr.s_addr == 0; + case AF_INET6: + return + ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[0] == 0 && + ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[1] == 0 && + ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[2] == 0 && + ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[3] == 0; + } + return false; +} + +static int addr_port(struct sockaddr_storage *ss) +{ + switch (ss->ss_family) { + case AF_INET: + return ntohs(((struct sockaddr_in *)ss)->sin_port); + case AF_INET6: + return ntohs(((struct sockaddr_in6 *)ss)->sin6_port); + } + return 0; +} + +static void addr_set_port(struct sockaddr_storage *ss, int p) +{ + switch (ss->ss_family) { + case AF_INET: + ((struct sockaddr_in *)ss)->sin_port = htons(p); + case AF_INET6: + ((struct sockaddr_in6 *)ss)->sin6_port = htons(p); + } +} + +/* + * Parse an ip[:port] list into an addr array. Use the default + * monitor port if a port isn't specified. + */ +int ceph_parse_ips(const char *c, const char *end, + struct ceph_entity_addr *addr, + int max_count, int *count) +{ + int i; + const char *p = c; + + dout("parse_ips on '%.*s'\n", (int)(end-c), c); + for (i = 0; i < max_count; i++) { + const char *ipend; + struct sockaddr_storage *ss = &addr[i].in_addr; + struct sockaddr_in *in4 = (void *)ss; + struct sockaddr_in6 *in6 = (void *)ss; + int port; + char delim = ','; + + if (*p == '[') { + delim = ']'; + p++; + } + + memset(ss, 0, sizeof(*ss)); + if (in4_pton(p, end - p, (u8 *)&in4->sin_addr.s_addr, + delim, &ipend)) + ss->ss_family = AF_INET; + else if (in6_pton(p, end - p, (u8 *)&in6->sin6_addr.s6_addr, + delim, &ipend)) + ss->ss_family = AF_INET6; + else + goto bad; + p = ipend; + + if (delim == ']') { + if (*p != ']') { + dout("missing matching ']'\n"); + goto bad; + } + p++; + } + + /* port? */ + if (p < end && *p == ':') { + port = 0; + p++; + while (p < end && *p >= '0' && *p <= '9') { + port = (port * 10) + (*p - '0'); + p++; + } + if (port > 65535 || port == 0) + goto bad; + } else { + port = CEPH_MON_PORT; + } + + addr_set_port(ss, port); + + dout("parse_ips got %s\n", ceph_pr_addr(ss)); + + if (p == end) + break; + if (*p != ',') + goto bad; + p++; + } + + if (p != end) + goto bad; + + if (count) + *count = i + 1; + return 0; + +bad: + pr_err("parse_ips bad ip '%.*s'\n", (int)(end - c), c); + return -EINVAL; +} +EXPORT_SYMBOL(ceph_parse_ips); + +static int process_banner(struct ceph_connection *con) +{ + dout("process_banner on %p\n", con); + + if (verify_hello(con) < 0) + return -1; + + ceph_decode_addr(&con->actual_peer_addr); + ceph_decode_addr(&con->peer_addr_for_me); + + /* + * Make sure the other end is who we wanted. note that the other + * end may not yet know their ip address, so if it's 0.0.0.0, give + * them the benefit of the doubt. + */ + if (memcmp(&con->peer_addr, &con->actual_peer_addr, + sizeof(con->peer_addr)) != 0 && + !(addr_is_blank(&con->actual_peer_addr.in_addr) && + con->actual_peer_addr.nonce == con->peer_addr.nonce)) { + pr_warning("wrong peer, want %s/%d, got %s/%d\n", + ceph_pr_addr(&con->peer_addr.in_addr), + (int)le32_to_cpu(con->peer_addr.nonce), + ceph_pr_addr(&con->actual_peer_addr.in_addr), + (int)le32_to_cpu(con->actual_peer_addr.nonce)); + con->error_msg = "wrong peer at address"; + return -1; + } + + /* + * did we learn our address? + */ + if (addr_is_blank(&con->msgr->inst.addr.in_addr)) { + int port = addr_port(&con->msgr->inst.addr.in_addr); + + memcpy(&con->msgr->inst.addr.in_addr, + &con->peer_addr_for_me.in_addr, + sizeof(con->peer_addr_for_me.in_addr)); + addr_set_port(&con->msgr->inst.addr.in_addr, port); + encode_my_addr(con->msgr); + dout("process_banner learned my addr is %s\n", + ceph_pr_addr(&con->msgr->inst.addr.in_addr)); + } + + set_bit(NEGOTIATING, &con->state); + prepare_read_connect(con); + return 0; +} + +static void fail_protocol(struct ceph_connection *con) +{ + reset_connection(con); + set_bit(CLOSED, &con->state); /* in case there's queued work */ + + mutex_unlock(&con->mutex); + if (con->ops->bad_proto) + con->ops->bad_proto(con); + mutex_lock(&con->mutex); +} + +static int process_connect(struct ceph_connection *con) +{ + u64 sup_feat = con->msgr->supported_features; + u64 req_feat = con->msgr->required_features; + u64 server_feat = le64_to_cpu(con->in_reply.features); + + dout("process_connect on %p tag %d\n", con, (int)con->in_tag); + + switch (con->in_reply.tag) { + case CEPH_MSGR_TAG_FEATURES: + pr_err("%s%lld %s feature set mismatch," + " my %llx < server's %llx, missing %llx\n", + ENTITY_NAME(con->peer_name), + ceph_pr_addr(&con->peer_addr.in_addr), + sup_feat, server_feat, server_feat & ~sup_feat); + con->error_msg = "missing required protocol features"; + fail_protocol(con); + return -1; + + case CEPH_MSGR_TAG_BADPROTOVER: + pr_err("%s%lld %s protocol version mismatch," + " my %d != server's %d\n", + ENTITY_NAME(con->peer_name), + ceph_pr_addr(&con->peer_addr.in_addr), + le32_to_cpu(con->out_connect.protocol_version), + le32_to_cpu(con->in_reply.protocol_version)); + con->error_msg = "protocol version mismatch"; + fail_protocol(con); + return -1; + + case CEPH_MSGR_TAG_BADAUTHORIZER: + con->auth_retry++; + dout("process_connect %p got BADAUTHORIZER attempt %d\n", con, + con->auth_retry); + if (con->auth_retry == 2) { + con->error_msg = "connect authorization failure"; + reset_connection(con); + set_bit(CLOSED, &con->state); + return -1; + } + con->auth_retry = 1; + prepare_write_connect(con->msgr, con, 0); + prepare_read_connect(con); + break; + + case CEPH_MSGR_TAG_RESETSESSION: + /* + * If we connected with a large connect_seq but the peer + * has no record of a session with us (no connection, or + * connect_seq == 0), they will send RESETSESION to indicate + * that they must have reset their session, and may have + * dropped messages. + */ + dout("process_connect got RESET peer seq %u\n", + le32_to_cpu(con->in_connect.connect_seq)); + pr_err("%s%lld %s connection reset\n", + ENTITY_NAME(con->peer_name), + ceph_pr_addr(&con->peer_addr.in_addr)); + reset_connection(con); + prepare_write_connect(con->msgr, con, 0); + prepare_read_connect(con); + + /* Tell ceph about it. */ + mutex_unlock(&con->mutex); + pr_info("reset on %s%lld\n", ENTITY_NAME(con->peer_name)); + if (con->ops->peer_reset) + con->ops->peer_reset(con); + mutex_lock(&con->mutex); + break; + + case CEPH_MSGR_TAG_RETRY_SESSION: + /* + * If we sent a smaller connect_seq than the peer has, try + * again with a larger value. + */ + dout("process_connect got RETRY my seq = %u, peer_seq = %u\n", + le32_to_cpu(con->out_connect.connect_seq), + le32_to_cpu(con->in_connect.connect_seq)); + con->connect_seq = le32_to_cpu(con->in_connect.connect_seq); + prepare_write_connect(con->msgr, con, 0); + prepare_read_connect(con); + break; + + case CEPH_MSGR_TAG_RETRY_GLOBAL: + /* + * If we sent a smaller global_seq than the peer has, try + * again with a larger value. + */ + dout("process_connect got RETRY_GLOBAL my %u peer_gseq %u\n", + con->peer_global_seq, + le32_to_cpu(con->in_connect.global_seq)); + get_global_seq(con->msgr, + le32_to_cpu(con->in_connect.global_seq)); + prepare_write_connect(con->msgr, con, 0); + prepare_read_connect(con); + break; + + case CEPH_MSGR_TAG_READY: + if (req_feat & ~server_feat) { + pr_err("%s%lld %s protocol feature mismatch," + " my required %llx > server's %llx, need %llx\n", + ENTITY_NAME(con->peer_name), + ceph_pr_addr(&con->peer_addr.in_addr), + req_feat, server_feat, req_feat & ~server_feat); + con->error_msg = "missing required protocol features"; + fail_protocol(con); + return -1; + } + clear_bit(CONNECTING, &con->state); + con->peer_global_seq = le32_to_cpu(con->in_reply.global_seq); + con->connect_seq++; + con->peer_features = server_feat; + dout("process_connect got READY gseq %d cseq %d (%d)\n", + con->peer_global_seq, + le32_to_cpu(con->in_reply.connect_seq), + con->connect_seq); + WARN_ON(con->connect_seq != + le32_to_cpu(con->in_reply.connect_seq)); + + if (con->in_reply.flags & CEPH_MSG_CONNECT_LOSSY) + set_bit(LOSSYTX, &con->state); + + prepare_read_tag(con); + break; + + case CEPH_MSGR_TAG_WAIT: + /* + * If there is a connection race (we are opening + * connections to each other), one of us may just have + * to WAIT. This shouldn't happen if we are the + * client. + */ + pr_err("process_connect peer connecting WAIT\n"); + + default: + pr_err("connect protocol error, will retry\n"); + con->error_msg = "protocol error, garbage tag during connect"; + return -1; + } + return 0; +} + + +/* + * read (part of) an ack + */ +static int read_partial_ack(struct ceph_connection *con) +{ + int to = 0; + + return read_partial(con, &to, sizeof(con->in_temp_ack), + &con->in_temp_ack); +} + + +/* + * We can finally discard anything that's been acked. + */ +static void process_ack(struct ceph_connection *con) +{ + struct ceph_msg *m; + u64 ack = le64_to_cpu(con->in_temp_ack); + u64 seq; + + while (!list_empty(&con->out_sent)) { + m = list_first_entry(&con->out_sent, struct ceph_msg, + list_head); + seq = le64_to_cpu(m->hdr.seq); + if (seq > ack) + break; + dout("got ack for seq %llu type %d at %p\n", seq, + le16_to_cpu(m->hdr.type), m); + ceph_msg_remove(m); + } + prepare_read_tag(con); +} + + + + +static int read_partial_message_section(struct ceph_connection *con, + struct kvec *section, + unsigned int sec_len, u32 *crc) +{ + int ret, left; + + BUG_ON(!section); + + while (section->iov_len < sec_len) { + BUG_ON(section->iov_base == NULL); + left = sec_len - section->iov_len; + ret = ceph_tcp_recvmsg(con->sock, (char *)section->iov_base + + section->iov_len, left); + if (ret <= 0) + return ret; + section->iov_len += ret; + if (section->iov_len == sec_len) + *crc = crc32c(0, section->iov_base, + section->iov_len); + } + + return 1; +} + +static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con, + struct ceph_msg_header *hdr, + int *skip); + + +static int read_partial_message_pages(struct ceph_connection *con, + struct page **pages, + unsigned data_len, int datacrc) +{ + void *p; + int ret; + int left; + + left = min((int)(data_len - con->in_msg_pos.data_pos), + (int)(PAGE_SIZE - con->in_msg_pos.page_pos)); + /* (page) data */ + BUG_ON(pages == NULL); + p = kmap(pages[con->in_msg_pos.page]); + ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos, + left); + if (ret > 0 && datacrc) + con->in_data_crc = + crc32c(con->in_data_crc, + p + con->in_msg_pos.page_pos, ret); + kunmap(pages[con->in_msg_pos.page]); + if (ret <= 0) + return ret; + con->in_msg_pos.data_pos += ret; + con->in_msg_pos.page_pos += ret; + if (con->in_msg_pos.page_pos == PAGE_SIZE) { + con->in_msg_pos.page_pos = 0; + con->in_msg_pos.page++; + } + + return ret; +} + +#ifdef CONFIG_BLOCK +static int read_partial_message_bio(struct ceph_connection *con, + struct bio **bio_iter, int *bio_seg, + unsigned data_len, int datacrc) +{ + struct bio_vec *bv = bio_iovec_idx(*bio_iter, *bio_seg); + void *p; + int ret, left; + + if (IS_ERR(bv)) + return PTR_ERR(bv); + + left = min((int)(data_len - con->in_msg_pos.data_pos), + (int)(bv->bv_len - con->in_msg_pos.page_pos)); + + p = kmap(bv->bv_page) + bv->bv_offset; + + ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos, + left); + if (ret > 0 && datacrc) + con->in_data_crc = + crc32c(con->in_data_crc, + p + con->in_msg_pos.page_pos, ret); + kunmap(bv->bv_page); + if (ret <= 0) + return ret; + con->in_msg_pos.data_pos += ret; + con->in_msg_pos.page_pos += ret; + if (con->in_msg_pos.page_pos == bv->bv_len) { + con->in_msg_pos.page_pos = 0; + iter_bio_next(bio_iter, bio_seg); + } + + return ret; +} +#endif + +/* + * read (part of) a message. + */ +static int read_partial_message(struct ceph_connection *con) +{ + struct ceph_msg *m = con->in_msg; + int ret; + int to, left; + unsigned front_len, middle_len, data_len, data_off; + int datacrc = con->msgr->nocrc; + int skip; + u64 seq; + + dout("read_partial_message con %p msg %p\n", con, m); + + /* header */ + while (con->in_base_pos < sizeof(con->in_hdr)) { + left = sizeof(con->in_hdr) - con->in_base_pos; + ret = ceph_tcp_recvmsg(con->sock, + (char *)&con->in_hdr + con->in_base_pos, + left); + if (ret <= 0) + return ret; + con->in_base_pos += ret; + if (con->in_base_pos == sizeof(con->in_hdr)) { + u32 crc = crc32c(0, (void *)&con->in_hdr, + sizeof(con->in_hdr) - sizeof(con->in_hdr.crc)); + if (crc != le32_to_cpu(con->in_hdr.crc)) { + pr_err("read_partial_message bad hdr " + " crc %u != expected %u\n", + crc, con->in_hdr.crc); + return -EBADMSG; + } + } + } + front_len = le32_to_cpu(con->in_hdr.front_len); + if (front_len > CEPH_MSG_MAX_FRONT_LEN) + return -EIO; + middle_len = le32_to_cpu(con->in_hdr.middle_len); + if (middle_len > CEPH_MSG_MAX_DATA_LEN) + return -EIO; + data_len = le32_to_cpu(con->in_hdr.data_len); + if (data_len > CEPH_MSG_MAX_DATA_LEN) + return -EIO; + data_off = le16_to_cpu(con->in_hdr.data_off); + + /* verify seq# */ + seq = le64_to_cpu(con->in_hdr.seq); + if ((s64)seq - (s64)con->in_seq < 1) { + pr_info("skipping %s%lld %s seq %lld, expected %lld\n", + ENTITY_NAME(con->peer_name), + ceph_pr_addr(&con->peer_addr.in_addr), + seq, con->in_seq + 1); + con->in_base_pos = -front_len - middle_len - data_len - + sizeof(m->footer); + con->in_tag = CEPH_MSGR_TAG_READY; + con->in_seq++; + return 0; + } else if ((s64)seq - (s64)con->in_seq > 1) { + pr_err("read_partial_message bad seq %lld expected %lld\n", + seq, con->in_seq + 1); + con->error_msg = "bad message sequence # for incoming message"; + return -EBADMSG; + } + + /* allocate message? */ + if (!con->in_msg) { + dout("got hdr type %d front %d data %d\n", con->in_hdr.type, + con->in_hdr.front_len, con->in_hdr.data_len); + skip = 0; + con->in_msg = ceph_alloc_msg(con, &con->in_hdr, &skip); + if (skip) { + /* skip this message */ + dout("alloc_msg said skip message\n"); + BUG_ON(con->in_msg); + con->in_base_pos = -front_len - middle_len - data_len - + sizeof(m->footer); + con->in_tag = CEPH_MSGR_TAG_READY; + con->in_seq++; + return 0; + } + if (!con->in_msg) { + con->error_msg = + "error allocating memory for incoming message"; + return -ENOMEM; + } + m = con->in_msg; + m->front.iov_len = 0; /* haven't read it yet */ + if (m->middle) + m->middle->vec.iov_len = 0; + + con->in_msg_pos.page = 0; + if (m->pages) + con->in_msg_pos.page_pos = data_off & ~PAGE_MASK; + else + con->in_msg_pos.page_pos = 0; + con->in_msg_pos.data_pos = 0; + } + + /* front */ + ret = read_partial_message_section(con, &m->front, front_len, + &con->in_front_crc); + if (ret <= 0) + return ret; + + /* middle */ + if (m->middle) { + ret = read_partial_message_section(con, &m->middle->vec, + middle_len, + &con->in_middle_crc); + if (ret <= 0) + return ret; + } +#ifdef CONFIG_BLOCK + if (m->bio && !m->bio_iter) + init_bio_iter(m->bio, &m->bio_iter, &m->bio_seg); +#endif + + /* (page) data */ + while (con->in_msg_pos.data_pos < data_len) { + if (m->pages) { + ret = read_partial_message_pages(con, m->pages, + data_len, datacrc); + if (ret <= 0) + return ret; +#ifdef CONFIG_BLOCK + } else if (m->bio) { + + ret = read_partial_message_bio(con, + &m->bio_iter, &m->bio_seg, + data_len, datacrc); + if (ret <= 0) + return ret; +#endif + } else { + BUG_ON(1); + } + } + + /* footer */ + to = sizeof(m->hdr) + sizeof(m->footer); + while (con->in_base_pos < to) { + left = to - con->in_base_pos; + ret = ceph_tcp_recvmsg(con->sock, (char *)&m->footer + + (con->in_base_pos - sizeof(m->hdr)), + left); + if (ret <= 0) + return ret; + con->in_base_pos += ret; + } + dout("read_partial_message got msg %p %d (%u) + %d (%u) + %d (%u)\n", + m, front_len, m->footer.front_crc, middle_len, + m->footer.middle_crc, data_len, m->footer.data_crc); + + /* crc ok? */ + if (con->in_front_crc != le32_to_cpu(m->footer.front_crc)) { + pr_err("read_partial_message %p front crc %u != exp. %u\n", + m, con->in_front_crc, m->footer.front_crc); + return -EBADMSG; + } + if (con->in_middle_crc != le32_to_cpu(m->footer.middle_crc)) { + pr_err("read_partial_message %p middle crc %u != exp %u\n", + m, con->in_middle_crc, m->footer.middle_crc); + return -EBADMSG; + } + if (datacrc && + (m->footer.flags & CEPH_MSG_FOOTER_NOCRC) == 0 && + con->in_data_crc != le32_to_cpu(m->footer.data_crc)) { + pr_err("read_partial_message %p data crc %u != exp. %u\n", m, + con->in_data_crc, le32_to_cpu(m->footer.data_crc)); + return -EBADMSG; + } + + return 1; /* done! */ +} + +/* + * Process message. This happens in the worker thread. The callback should + * be careful not to do anything that waits on other incoming messages or it + * may deadlock. + */ +static void process_message(struct ceph_connection *con) +{ + struct ceph_msg *msg; + + msg = con->in_msg; + con->in_msg = NULL; + + /* if first message, set peer_name */ + if (con->peer_name.type == 0) + con->peer_name = msg->hdr.src; + + con->in_seq++; + mutex_unlock(&con->mutex); + + dout("===== %p %llu from %s%lld %d=%s len %d+%d (%u %u %u) =====\n", + msg, le64_to_cpu(msg->hdr.seq), + ENTITY_NAME(msg->hdr.src), + le16_to_cpu(msg->hdr.type), + ceph_msg_type_name(le16_to_cpu(msg->hdr.type)), + le32_to_cpu(msg->hdr.front_len), + le32_to_cpu(msg->hdr.data_len), + con->in_front_crc, con->in_middle_crc, con->in_data_crc); + con->ops->dispatch(con, msg); + + mutex_lock(&con->mutex); + prepare_read_tag(con); +} + + +/* + * Write something to the socket. Called in a worker thread when the + * socket appears to be writeable and we have something ready to send. + */ +static int try_write(struct ceph_connection *con) +{ + struct ceph_messenger *msgr = con->msgr; + int ret = 1; + + dout("try_write start %p state %lu nref %d\n", con, con->state, + atomic_read(&con->nref)); + +more: + dout("try_write out_kvec_bytes %d\n", con->out_kvec_bytes); + + /* open the socket first? */ + if (con->sock == NULL) { + /* + * if we were STANDBY and are reconnecting _this_ + * connection, bump connect_seq now. Always bump + * global_seq. + */ + if (test_and_clear_bit(STANDBY, &con->state)) + con->connect_seq++; + + prepare_write_banner(msgr, con); + prepare_write_connect(msgr, con, 1); + prepare_read_banner(con); + set_bit(CONNECTING, &con->state); + clear_bit(NEGOTIATING, &con->state); + + BUG_ON(con->in_msg); + con->in_tag = CEPH_MSGR_TAG_READY; + dout("try_write initiating connect on %p new state %lu\n", + con, con->state); + con->sock = ceph_tcp_connect(con); + if (IS_ERR(con->sock)) { + con->sock = NULL; + con->error_msg = "connect error"; + ret = -1; + goto out; + } + } + +more_kvec: + /* kvec data queued? */ + if (con->out_skip) { + ret = write_partial_skip(con); + if (ret <= 0) + goto done; + if (ret < 0) { + dout("try_write write_partial_skip err %d\n", ret); + goto done; + } + } + if (con->out_kvec_left) { + ret = write_partial_kvec(con); + if (ret <= 0) + goto done; + } + + /* msg pages? */ + if (con->out_msg) { + if (con->out_msg_done) { + ceph_msg_put(con->out_msg); + con->out_msg = NULL; /* we're done with this one */ + goto do_next; + } + + ret = write_partial_msg_pages(con); + if (ret == 1) + goto more_kvec; /* we need to send the footer, too! */ + if (ret == 0) + goto done; + if (ret < 0) { + dout("try_write write_partial_msg_pages err %d\n", + ret); + goto done; + } + } + +do_next: + if (!test_bit(CONNECTING, &con->state)) { + /* is anything else pending? */ + if (!list_empty(&con->out_queue)) { + prepare_write_message(con); + goto more; + } + if (con->in_seq > con->in_seq_acked) { + prepare_write_ack(con); + goto more; + } + if (test_and_clear_bit(KEEPALIVE_PENDING, &con->state)) { + prepare_write_keepalive(con); + goto more; + } + } + + /* Nothing to do! */ + clear_bit(WRITE_PENDING, &con->state); + dout("try_write nothing else to write.\n"); +done: + ret = 0; +out: + dout("try_write done on %p\n", con); + return ret; +} + + + +/* + * Read what we can from the socket. + */ +static int try_read(struct ceph_connection *con) +{ + int ret = -1; + + if (!con->sock) + return 0; + + if (test_bit(STANDBY, &con->state)) + return 0; + + dout("try_read start on %p\n", con); + +more: + dout("try_read tag %d in_base_pos %d\n", (int)con->in_tag, + con->in_base_pos); + if (test_bit(CONNECTING, &con->state)) { + if (!test_bit(NEGOTIATING, &con->state)) { + dout("try_read connecting\n"); + ret = read_partial_banner(con); + if (ret <= 0) + goto done; + if (process_banner(con) < 0) { + ret = -1; + goto out; + } + } + ret = read_partial_connect(con); + if (ret <= 0) + goto done; + if (process_connect(con) < 0) { + ret = -1; + goto out; + } + goto more; + } + + if (con->in_base_pos < 0) { + /* + * skipping + discarding content. + * + * FIXME: there must be a better way to do this! + */ + static char buf[1024]; + int skip = min(1024, -con->in_base_pos); + dout("skipping %d / %d bytes\n", skip, -con->in_base_pos); + ret = ceph_tcp_recvmsg(con->sock, buf, skip); + if (ret <= 0) + goto done; + con->in_base_pos += ret; + if (con->in_base_pos) + goto more; + } + if (con->in_tag == CEPH_MSGR_TAG_READY) { + /* + * what's next? + */ + ret = ceph_tcp_recvmsg(con->sock, &con->in_tag, 1); + if (ret <= 0) + goto done; + dout("try_read got tag %d\n", (int)con->in_tag); + switch (con->in_tag) { + case CEPH_MSGR_TAG_MSG: + prepare_read_message(con); + break; + case CEPH_MSGR_TAG_ACK: + prepare_read_ack(con); + break; + case CEPH_MSGR_TAG_CLOSE: + set_bit(CLOSED, &con->state); /* fixme */ + goto done; + default: + goto bad_tag; + } + } + if (con->in_tag == CEPH_MSGR_TAG_MSG) { + ret = read_partial_message(con); + if (ret <= 0) { + switch (ret) { + case -EBADMSG: + con->error_msg = "bad crc"; + ret = -EIO; + goto out; + case -EIO: + con->error_msg = "io error"; + goto out; + default: + goto done; + } + } + if (con->in_tag == CEPH_MSGR_TAG_READY) + goto more; + process_message(con); + goto more; + } + if (con->in_tag == CEPH_MSGR_TAG_ACK) { + ret = read_partial_ack(con); + if (ret <= 0) + goto done; + process_ack(con); + goto more; + } + +done: + ret = 0; +out: + dout("try_read done on %p\n", con); + return ret; + +bad_tag: + pr_err("try_read bad con->in_tag = %d\n", (int)con->in_tag); + con->error_msg = "protocol error, garbage tag"; + ret = -1; + goto out; +} + + +/* + * Atomically queue work on a connection. Bump @con reference to + * avoid races with connection teardown. + * + * There is some trickery going on with QUEUED and BUSY because we + * only want a _single_ thread operating on each connection at any + * point in time, but we want to use all available CPUs. + * + * The worker thread only proceeds if it can atomically set BUSY. It + * clears QUEUED and does it's thing. When it thinks it's done, it + * clears BUSY, then rechecks QUEUED.. if it's set again, it loops + * (tries again to set BUSY). + * + * To queue work, we first set QUEUED, _then_ if BUSY isn't set, we + * try to queue work. If that fails (work is already queued, or BUSY) + * we give up (work also already being done or is queued) but leave QUEUED + * set so that the worker thread will loop if necessary. + */ +static void queue_con(struct ceph_connection *con) +{ + if (test_bit(DEAD, &con->state)) { + dout("queue_con %p ignoring: DEAD\n", + con); + return; + } + + if (!con->ops->get(con)) { + dout("queue_con %p ref count 0\n", con); + return; + } + + set_bit(QUEUED, &con->state); + if (test_bit(BUSY, &con->state)) { + dout("queue_con %p - already BUSY\n", con); + con->ops->put(con); + } else if (!queue_work(ceph_msgr_wq, &con->work.work)) { + dout("queue_con %p - already queued\n", con); + con->ops->put(con); + } else { + dout("queue_con %p\n", con); + } +} + +/* + * Do some work on a connection. Drop a connection ref when we're done. + */ +static void con_work(struct work_struct *work) +{ + struct ceph_connection *con = container_of(work, struct ceph_connection, + work.work); + int backoff = 0; + +more: + if (test_and_set_bit(BUSY, &con->state) != 0) { + dout("con_work %p BUSY already set\n", con); + goto out; + } + dout("con_work %p start, clearing QUEUED\n", con); + clear_bit(QUEUED, &con->state); + + mutex_lock(&con->mutex); + + if (test_bit(CLOSED, &con->state)) { /* e.g. if we are replaced */ + dout("con_work CLOSED\n"); + con_close_socket(con); + goto done; + } + if (test_and_clear_bit(OPENING, &con->state)) { + /* reopen w/ new peer */ + dout("con_work OPENING\n"); + con_close_socket(con); + } + + if (test_and_clear_bit(SOCK_CLOSED, &con->state) || + try_read(con) < 0 || + try_write(con) < 0) { + mutex_unlock(&con->mutex); + backoff = 1; + ceph_fault(con); /* error/fault path */ + goto done_unlocked; + } + +done: + mutex_unlock(&con->mutex); + +done_unlocked: + clear_bit(BUSY, &con->state); + dout("con->state=%lu\n", con->state); + if (test_bit(QUEUED, &con->state)) { + if (!backoff || test_bit(OPENING, &con->state)) { + dout("con_work %p QUEUED reset, looping\n", con); + goto more; + } + dout("con_work %p QUEUED reset, but just faulted\n", con); + clear_bit(QUEUED, &con->state); + } + dout("con_work %p done\n", con); + +out: + con->ops->put(con); +} + + +/* + * Generic error/fault handler. A retry mechanism is used with + * exponential backoff + */ +static void ceph_fault(struct ceph_connection *con) +{ + pr_err("%s%lld %s %s\n", ENTITY_NAME(con->peer_name), + ceph_pr_addr(&con->peer_addr.in_addr), con->error_msg); + dout("fault %p state %lu to peer %s\n", + con, con->state, ceph_pr_addr(&con->peer_addr.in_addr)); + + if (test_bit(LOSSYTX, &con->state)) { + dout("fault on LOSSYTX channel\n"); + goto out; + } + + mutex_lock(&con->mutex); + if (test_bit(CLOSED, &con->state)) + goto out_unlock; + + con_close_socket(con); + + if (con->in_msg) { + ceph_msg_put(con->in_msg); + con->in_msg = NULL; + } + + /* Requeue anything that hasn't been acked */ + list_splice_init(&con->out_sent, &con->out_queue); + + /* If there are no messages in the queue, place the connection + * in a STANDBY state (i.e., don't try to reconnect just yet). */ + if (list_empty(&con->out_queue) && !con->out_keepalive_pending) { + dout("fault setting STANDBY\n"); + set_bit(STANDBY, &con->state); + } else { + /* retry after a delay. */ + if (con->delay == 0) + con->delay = BASE_DELAY_INTERVAL; + else if (con->delay < MAX_DELAY_INTERVAL) + con->delay *= 2; + dout("fault queueing %p delay %lu\n", con, con->delay); + con->ops->get(con); + if (queue_delayed_work(ceph_msgr_wq, &con->work, + round_jiffies_relative(con->delay)) == 0) + con->ops->put(con); + } + +out_unlock: + mutex_unlock(&con->mutex); +out: + /* + * in case we faulted due to authentication, invalidate our + * current tickets so that we can get new ones. + */ + if (con->auth_retry && con->ops->invalidate_authorizer) { + dout("calling invalidate_authorizer()\n"); + con->ops->invalidate_authorizer(con); + } + + if (con->ops->fault) + con->ops->fault(con); +} + + + +/* + * create a new messenger instance + */ +struct ceph_messenger *ceph_messenger_create(struct ceph_entity_addr *myaddr, + u32 supported_features, + u32 required_features) +{ + struct ceph_messenger *msgr; + + msgr = kzalloc(sizeof(*msgr), GFP_KERNEL); + if (msgr == NULL) + return ERR_PTR(-ENOMEM); + + msgr->supported_features = supported_features; + msgr->required_features = required_features; + + spin_lock_init(&msgr->global_seq_lock); + + /* the zero page is needed if a request is "canceled" while the message + * is being written over the socket */ + msgr->zero_page = __page_cache_alloc(GFP_KERNEL | __GFP_ZERO); + if (!msgr->zero_page) { + kfree(msgr); + return ERR_PTR(-ENOMEM); + } + kmap(msgr->zero_page); + + if (myaddr) + msgr->inst.addr = *myaddr; + + /* select a random nonce */ + msgr->inst.addr.type = 0; + get_random_bytes(&msgr->inst.addr.nonce, sizeof(msgr->inst.addr.nonce)); + encode_my_addr(msgr); + + dout("messenger_create %p\n", msgr); + return msgr; +} +EXPORT_SYMBOL(ceph_messenger_create); + +void ceph_messenger_destroy(struct ceph_messenger *msgr) +{ + dout("destroy %p\n", msgr); + kunmap(msgr->zero_page); + __free_page(msgr->zero_page); + kfree(msgr); + dout("destroyed messenger %p\n", msgr); +} +EXPORT_SYMBOL(ceph_messenger_destroy); + +/* + * Queue up an outgoing message on the given connection. + */ +void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg) +{ + if (test_bit(CLOSED, &con->state)) { + dout("con_send %p closed, dropping %p\n", con, msg); + ceph_msg_put(msg); + return; + } + + /* set src+dst */ + msg->hdr.src = con->msgr->inst.name; + + BUG_ON(msg->front.iov_len != le32_to_cpu(msg->hdr.front_len)); + + msg->needs_out_seq = true; + + /* queue */ + mutex_lock(&con->mutex); + BUG_ON(!list_empty(&msg->list_head)); + list_add_tail(&msg->list_head, &con->out_queue); + dout("----- %p to %s%lld %d=%s len %d+%d+%d -----\n", msg, + ENTITY_NAME(con->peer_name), le16_to_cpu(msg->hdr.type), + ceph_msg_type_name(le16_to_cpu(msg->hdr.type)), + le32_to_cpu(msg->hdr.front_len), + le32_to_cpu(msg->hdr.middle_len), + le32_to_cpu(msg->hdr.data_len)); + mutex_unlock(&con->mutex); + + /* if there wasn't anything waiting to send before, queue + * new work */ + if (test_and_set_bit(WRITE_PENDING, &con->state) == 0) + queue_con(con); +} +EXPORT_SYMBOL(ceph_con_send); + +/* + * Revoke a message that was previously queued for send + */ +void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg) +{ + mutex_lock(&con->mutex); + if (!list_empty(&msg->list_head)) { + dout("con_revoke %p msg %p - was on queue\n", con, msg); + list_del_init(&msg->list_head); + ceph_msg_put(msg); + msg->hdr.seq = 0; + } + if (con->out_msg == msg) { + dout("con_revoke %p msg %p - was sending\n", con, msg); + con->out_msg = NULL; + if (con->out_kvec_is_msg) { + con->out_skip = con->out_kvec_bytes; + con->out_kvec_is_msg = false; + } + ceph_msg_put(msg); + msg->hdr.seq = 0; + } + mutex_unlock(&con->mutex); +} + +/* + * Revoke a message that we may be reading data into + */ +void ceph_con_revoke_message(struct ceph_connection *con, struct ceph_msg *msg) +{ + mutex_lock(&con->mutex); + if (con->in_msg && con->in_msg == msg) { + unsigned front_len = le32_to_cpu(con->in_hdr.front_len); + unsigned middle_len = le32_to_cpu(con->in_hdr.middle_len); + unsigned data_len = le32_to_cpu(con->in_hdr.data_len); + + /* skip rest of message */ + dout("con_revoke_pages %p msg %p revoked\n", con, msg); + con->in_base_pos = con->in_base_pos - + sizeof(struct ceph_msg_header) - + front_len - + middle_len - + data_len - + sizeof(struct ceph_msg_footer); + ceph_msg_put(con->in_msg); + con->in_msg = NULL; + con->in_tag = CEPH_MSGR_TAG_READY; + con->in_seq++; + } else { + dout("con_revoke_pages %p msg %p pages %p no-op\n", + con, con->in_msg, msg); + } + mutex_unlock(&con->mutex); +} + +/* + * Queue a keepalive byte to ensure the tcp connection is alive. + */ +void ceph_con_keepalive(struct ceph_connection *con) +{ + if (test_and_set_bit(KEEPALIVE_PENDING, &con->state) == 0 && + test_and_set_bit(WRITE_PENDING, &con->state) == 0) + queue_con(con); +} +EXPORT_SYMBOL(ceph_con_keepalive); + + +/* + * construct a new message with given type, size + * the new msg has a ref count of 1. + */ +struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags) +{ + struct ceph_msg *m; + + m = kmalloc(sizeof(*m), flags); + if (m == NULL) + goto out; + kref_init(&m->kref); + INIT_LIST_HEAD(&m->list_head); + + m->hdr.tid = 0; + m->hdr.type = cpu_to_le16(type); + m->hdr.priority = cpu_to_le16(CEPH_MSG_PRIO_DEFAULT); + m->hdr.version = 0; + m->hdr.front_len = cpu_to_le32(front_len); + m->hdr.middle_len = 0; + m->hdr.data_len = 0; + m->hdr.data_off = 0; + m->hdr.reserved = 0; + m->footer.front_crc = 0; + m->footer.middle_crc = 0; + m->footer.data_crc = 0; + m->footer.flags = 0; + m->front_max = front_len; + m->front_is_vmalloc = false; + m->more_to_follow = false; + m->pool = NULL; + + /* front */ + if (front_len) { + if (front_len > PAGE_CACHE_SIZE) { + m->front.iov_base = __vmalloc(front_len, flags, + PAGE_KERNEL); + m->front_is_vmalloc = true; + } else { + m->front.iov_base = kmalloc(front_len, flags); + } + if (m->front.iov_base == NULL) { + pr_err("msg_new can't allocate %d bytes\n", + front_len); + goto out2; + } + } else { + m->front.iov_base = NULL; + } + m->front.iov_len = front_len; + + /* middle */ + m->middle = NULL; + + /* data */ + m->nr_pages = 0; + m->pages = NULL; + m->pagelist = NULL; + m->bio = NULL; + m->bio_iter = NULL; + m->bio_seg = 0; + m->trail = NULL; + + dout("ceph_msg_new %p front %d\n", m, front_len); + return m; + +out2: + ceph_msg_put(m); +out: + pr_err("msg_new can't create type %d front %d\n", type, front_len); + return NULL; +} +EXPORT_SYMBOL(ceph_msg_new); + +/* + * Allocate "middle" portion of a message, if it is needed and wasn't + * allocated by alloc_msg. This allows us to read a small fixed-size + * per-type header in the front and then gracefully fail (i.e., + * propagate the error to the caller based on info in the front) when + * the middle is too large. + */ +static int ceph_alloc_middle(struct ceph_connection *con, struct ceph_msg *msg) +{ + int type = le16_to_cpu(msg->hdr.type); + int middle_len = le32_to_cpu(msg->hdr.middle_len); + + dout("alloc_middle %p type %d %s middle_len %d\n", msg, type, + ceph_msg_type_name(type), middle_len); + BUG_ON(!middle_len); + BUG_ON(msg->middle); + + msg->middle = ceph_buffer_new(middle_len, GFP_NOFS); + if (!msg->middle) + return -ENOMEM; + return 0; +} + +/* + * Generic message allocator, for incoming messages. + */ +static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con, + struct ceph_msg_header *hdr, + int *skip) +{ + int type = le16_to_cpu(hdr->type); + int front_len = le32_to_cpu(hdr->front_len); + int middle_len = le32_to_cpu(hdr->middle_len); + struct ceph_msg *msg = NULL; + int ret; + + if (con->ops->alloc_msg) { + mutex_unlock(&con->mutex); + msg = con->ops->alloc_msg(con, hdr, skip); + mutex_lock(&con->mutex); + if (!msg || *skip) + return NULL; + } + if (!msg) { + *skip = 0; + msg = ceph_msg_new(type, front_len, GFP_NOFS); + if (!msg) { + pr_err("unable to allocate msg type %d len %d\n", + type, front_len); + return NULL; + } + } + memcpy(&msg->hdr, &con->in_hdr, sizeof(con->in_hdr)); + + if (middle_len && !msg->middle) { + ret = ceph_alloc_middle(con, msg); + if (ret < 0) { + ceph_msg_put(msg); + return NULL; + } + } + + return msg; +} + + +/* + * Free a generically kmalloc'd message. + */ +void ceph_msg_kfree(struct ceph_msg *m) +{ + dout("msg_kfree %p\n", m); + if (m->front_is_vmalloc) + vfree(m->front.iov_base); + else + kfree(m->front.iov_base); + kfree(m); +} + +/* + * Drop a msg ref. Destroy as needed. + */ +void ceph_msg_last_put(struct kref *kref) +{ + struct ceph_msg *m = container_of(kref, struct ceph_msg, kref); + + dout("ceph_msg_put last one on %p\n", m); + WARN_ON(!list_empty(&m->list_head)); + + /* drop middle, data, if any */ + if (m->middle) { + ceph_buffer_put(m->middle); + m->middle = NULL; + } + m->nr_pages = 0; + m->pages = NULL; + + if (m->pagelist) { + ceph_pagelist_release(m->pagelist); + kfree(m->pagelist); + m->pagelist = NULL; + } + + m->trail = NULL; + + if (m->pool) + ceph_msgpool_put(m->pool, m); + else + ceph_msg_kfree(m); +} +EXPORT_SYMBOL(ceph_msg_last_put); + +void ceph_msg_dump(struct ceph_msg *msg) +{ + pr_debug("msg_dump %p (front_max %d nr_pages %d)\n", msg, + msg->front_max, msg->nr_pages); + print_hex_dump(KERN_DEBUG, "header: ", + DUMP_PREFIX_OFFSET, 16, 1, + &msg->hdr, sizeof(msg->hdr), true); + print_hex_dump(KERN_DEBUG, " front: ", + DUMP_PREFIX_OFFSET, 16, 1, + msg->front.iov_base, msg->front.iov_len, true); + if (msg->middle) + print_hex_dump(KERN_DEBUG, "middle: ", + DUMP_PREFIX_OFFSET, 16, 1, + msg->middle->vec.iov_base, + msg->middle->vec.iov_len, true); + print_hex_dump(KERN_DEBUG, "footer: ", + DUMP_PREFIX_OFFSET, 16, 1, + &msg->footer, sizeof(msg->footer), true); +} +EXPORT_SYMBOL(ceph_msg_dump); diff --git a/net/ceph/mon_client.c b/net/ceph/mon_client.c new file mode 100644 index 0000000..8a07939 --- /dev/null +++ b/net/ceph/mon_client.c @@ -0,0 +1,1027 @@ +#include + +#include +#include +#include +#include +#include + +#include +#include +#include + +#include + +/* + * Interact with Ceph monitor cluster. Handle requests for new map + * versions, and periodically resend as needed. Also implement + * statfs() and umount(). + * + * A small cluster of Ceph "monitors" are responsible for managing critical + * cluster configuration and state information. An odd number (e.g., 3, 5) + * of cmon daemons use a modified version of the Paxos part-time parliament + * algorithm to manage the MDS map (mds cluster membership), OSD map, and + * list of clients who have mounted the file system. + * + * We maintain an open, active session with a monitor at all times in order to + * receive timely MDSMap updates. We periodically send a keepalive byte on the + * TCP socket to ensure we detect a failure. If the connection does break, we + * randomly hunt for a new monitor. Once the connection is reestablished, we + * resend any outstanding requests. + */ + +static const struct ceph_connection_operations mon_con_ops; + +static int __validate_auth(struct ceph_mon_client *monc); + +/* + * Decode a monmap blob (e.g., during mount). + */ +struct ceph_monmap *ceph_monmap_decode(void *p, void *end) +{ + struct ceph_monmap *m = NULL; + int i, err = -EINVAL; + struct ceph_fsid fsid; + u32 epoch, num_mon; + u16 version; + u32 len; + + ceph_decode_32_safe(&p, end, len, bad); + ceph_decode_need(&p, end, len, bad); + + dout("monmap_decode %p %p len %d\n", p, end, (int)(end-p)); + + ceph_decode_16_safe(&p, end, version, bad); + + ceph_decode_need(&p, end, sizeof(fsid) + 2*sizeof(u32), bad); + ceph_decode_copy(&p, &fsid, sizeof(fsid)); + epoch = ceph_decode_32(&p); + + num_mon = ceph_decode_32(&p); + ceph_decode_need(&p, end, num_mon*sizeof(m->mon_inst[0]), bad); + + if (num_mon >= CEPH_MAX_MON) + goto bad; + m = kmalloc(sizeof(*m) + sizeof(m->mon_inst[0])*num_mon, GFP_NOFS); + if (m == NULL) + return ERR_PTR(-ENOMEM); + m->fsid = fsid; + m->epoch = epoch; + m->num_mon = num_mon; + ceph_decode_copy(&p, m->mon_inst, num_mon*sizeof(m->mon_inst[0])); + for (i = 0; i < num_mon; i++) + ceph_decode_addr(&m->mon_inst[i].addr); + + dout("monmap_decode epoch %d, num_mon %d\n", m->epoch, + m->num_mon); + for (i = 0; i < m->num_mon; i++) + dout("monmap_decode mon%d is %s\n", i, + ceph_pr_addr(&m->mon_inst[i].addr.in_addr)); + return m; + +bad: + dout("monmap_decode failed with %d\n", err); + kfree(m); + return ERR_PTR(err); +} + +/* + * return true if *addr is included in the monmap. + */ +int ceph_monmap_contains(struct ceph_monmap *m, struct ceph_entity_addr *addr) +{ + int i; + + for (i = 0; i < m->num_mon; i++) + if (memcmp(addr, &m->mon_inst[i].addr, sizeof(*addr)) == 0) + return 1; + return 0; +} + +/* + * Send an auth request. + */ +static void __send_prepared_auth_request(struct ceph_mon_client *monc, int len) +{ + monc->pending_auth = 1; + monc->m_auth->front.iov_len = len; + monc->m_auth->hdr.front_len = cpu_to_le32(len); + ceph_con_revoke(monc->con, monc->m_auth); + ceph_msg_get(monc->m_auth); /* keep our ref */ + ceph_con_send(monc->con, monc->m_auth); +} + +/* + * Close monitor session, if any. + */ +static void __close_session(struct ceph_mon_client *monc) +{ + if (monc->con) { + dout("__close_session closing mon%d\n", monc->cur_mon); + ceph_con_revoke(monc->con, monc->m_auth); + ceph_con_close(monc->con); + monc->cur_mon = -1; + monc->pending_auth = 0; + ceph_auth_reset(monc->auth); + } +} + +/* + * Open a session with a (new) monitor. + */ +static int __open_session(struct ceph_mon_client *monc) +{ + char r; + int ret; + + if (monc->cur_mon < 0) { + get_random_bytes(&r, 1); + monc->cur_mon = r % monc->monmap->num_mon; + dout("open_session num=%d r=%d -> mon%d\n", + monc->monmap->num_mon, r, monc->cur_mon); + monc->sub_sent = 0; + monc->sub_renew_after = jiffies; /* i.e., expired */ + monc->want_next_osdmap = !!monc->want_next_osdmap; + + dout("open_session mon%d opening\n", monc->cur_mon); + monc->con->peer_name.type = CEPH_ENTITY_TYPE_MON; + monc->con->peer_name.num = cpu_to_le64(monc->cur_mon); + ceph_con_open(monc->con, + &monc->monmap->mon_inst[monc->cur_mon].addr); + + /* initiatiate authentication handshake */ + ret = ceph_auth_build_hello(monc->auth, + monc->m_auth->front.iov_base, + monc->m_auth->front_max); + __send_prepared_auth_request(monc, ret); + } else { + dout("open_session mon%d already open\n", monc->cur_mon); + } + return 0; +} + +static bool __sub_expired(struct ceph_mon_client *monc) +{ + return time_after_eq(jiffies, monc->sub_renew_after); +} + +/* + * Reschedule delayed work timer. + */ +static void __schedule_delayed(struct ceph_mon_client *monc) +{ + unsigned delay; + + if (monc->cur_mon < 0 || __sub_expired(monc)) + delay = 10 * HZ; + else + delay = 20 * HZ; + dout("__schedule_delayed after %u\n", delay); + schedule_delayed_work(&monc->delayed_work, delay); +} + +/* + * Send subscribe request for mdsmap and/or osdmap. + */ +static void __send_subscribe(struct ceph_mon_client *monc) +{ + dout("__send_subscribe sub_sent=%u exp=%u want_osd=%d\n", + (unsigned)monc->sub_sent, __sub_expired(monc), + monc->want_next_osdmap); + if ((__sub_expired(monc) && !monc->sub_sent) || + monc->want_next_osdmap == 1) { + struct ceph_msg *msg = monc->m_subscribe; + struct ceph_mon_subscribe_item *i; + void *p, *end; + int num; + + p = msg->front.iov_base; + end = p + msg->front_max; + + num = 1 + !!monc->want_next_osdmap + !!monc->want_mdsmap; + ceph_encode_32(&p, num); + + if (monc->want_next_osdmap) { + dout("__send_subscribe to 'osdmap' %u\n", + (unsigned)monc->have_osdmap); + ceph_encode_string(&p, end, "osdmap", 6); + i = p; + i->have = cpu_to_le64(monc->have_osdmap); + i->onetime = 1; + p += sizeof(*i); + monc->want_next_osdmap = 2; /* requested */ + } + if (monc->want_mdsmap) { + dout("__send_subscribe to 'mdsmap' %u+\n", + (unsigned)monc->have_mdsmap); + ceph_encode_string(&p, end, "mdsmap", 6); + i = p; + i->have = cpu_to_le64(monc->have_mdsmap); + i->onetime = 0; + p += sizeof(*i); + } + ceph_encode_string(&p, end, "monmap", 6); + i = p; + i->have = 0; + i->onetime = 0; + p += sizeof(*i); + + msg->front.iov_len = p - msg->front.iov_base; + msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); + ceph_con_revoke(monc->con, msg); + ceph_con_send(monc->con, ceph_msg_get(msg)); + + monc->sub_sent = jiffies | 1; /* never 0 */ + } +} + +static void handle_subscribe_ack(struct ceph_mon_client *monc, + struct ceph_msg *msg) +{ + unsigned seconds; + struct ceph_mon_subscribe_ack *h = msg->front.iov_base; + + if (msg->front.iov_len < sizeof(*h)) + goto bad; + seconds = le32_to_cpu(h->duration); + + mutex_lock(&monc->mutex); + if (monc->hunting) { + pr_info("mon%d %s session established\n", + monc->cur_mon, + ceph_pr_addr(&monc->con->peer_addr.in_addr)); + monc->hunting = false; + } + dout("handle_subscribe_ack after %d seconds\n", seconds); + monc->sub_renew_after = monc->sub_sent + (seconds >> 1)*HZ - 1; + monc->sub_sent = 0; + mutex_unlock(&monc->mutex); + return; +bad: + pr_err("got corrupt subscribe-ack msg\n"); + ceph_msg_dump(msg); +} + +/* + * Keep track of which maps we have + */ +int ceph_monc_got_mdsmap(struct ceph_mon_client *monc, u32 got) +{ + mutex_lock(&monc->mutex); + monc->have_mdsmap = got; + mutex_unlock(&monc->mutex); + return 0; +} +EXPORT_SYMBOL(ceph_monc_got_mdsmap); + +int ceph_monc_got_osdmap(struct ceph_mon_client *monc, u32 got) +{ + mutex_lock(&monc->mutex); + monc->have_osdmap = got; + monc->want_next_osdmap = 0; + mutex_unlock(&monc->mutex); + return 0; +} + +/* + * Register interest in the next osdmap + */ +void ceph_monc_request_next_osdmap(struct ceph_mon_client *monc) +{ + dout("request_next_osdmap have %u\n", monc->have_osdmap); + mutex_lock(&monc->mutex); + if (!monc->want_next_osdmap) + monc->want_next_osdmap = 1; + if (monc->want_next_osdmap < 2) + __send_subscribe(monc); + mutex_unlock(&monc->mutex); +} + +/* + * + */ +int ceph_monc_open_session(struct ceph_mon_client *monc) +{ + if (!monc->con) { + monc->con = kmalloc(sizeof(*monc->con), GFP_KERNEL); + if (!monc->con) + return -ENOMEM; + ceph_con_init(monc->client->msgr, monc->con); + monc->con->private = monc; + monc->con->ops = &mon_con_ops; + } + + mutex_lock(&monc->mutex); + __open_session(monc); + __schedule_delayed(monc); + mutex_unlock(&monc->mutex); + return 0; +} +EXPORT_SYMBOL(ceph_monc_open_session); + +/* + * The monitor responds with mount ack indicate mount success. The + * included client ticket allows the client to talk to MDSs and OSDs. + */ +static void ceph_monc_handle_map(struct ceph_mon_client *monc, + struct ceph_msg *msg) +{ + struct ceph_client *client = monc->client; + struct ceph_monmap *monmap = NULL, *old = monc->monmap; + void *p, *end; + + mutex_lock(&monc->mutex); + + dout("handle_monmap\n"); + p = msg->front.iov_base; + end = p + msg->front.iov_len; + + monmap = ceph_monmap_decode(p, end); + if (IS_ERR(monmap)) { + pr_err("problem decoding monmap, %d\n", + (int)PTR_ERR(monmap)); + goto out; + } + + if (ceph_check_fsid(monc->client, &monmap->fsid) < 0) { + kfree(monmap); + goto out; + } + + client->monc.monmap = monmap; + kfree(old); + +out: + mutex_unlock(&monc->mutex); + wake_up_all(&client->auth_wq); +} + +/* + * generic requests (e.g., statfs, poolop) + */ +static struct ceph_mon_generic_request *__lookup_generic_req( + struct ceph_mon_client *monc, u64 tid) +{ + struct ceph_mon_generic_request *req; + struct rb_node *n = monc->generic_request_tree.rb_node; + + while (n) { + req = rb_entry(n, struct ceph_mon_generic_request, node); + if (tid < req->tid) + n = n->rb_left; + else if (tid > req->tid) + n = n->rb_right; + else + return req; + } + return NULL; +} + +static void __insert_generic_request(struct ceph_mon_client *monc, + struct ceph_mon_generic_request *new) +{ + struct rb_node **p = &monc->generic_request_tree.rb_node; + struct rb_node *parent = NULL; + struct ceph_mon_generic_request *req = NULL; + + while (*p) { + parent = *p; + req = rb_entry(parent, struct ceph_mon_generic_request, node); + if (new->tid < req->tid) + p = &(*p)->rb_left; + else if (new->tid > req->tid) + p = &(*p)->rb_right; + else + BUG(); + } + + rb_link_node(&new->node, parent, p); + rb_insert_color(&new->node, &monc->generic_request_tree); +} + +static void release_generic_request(struct kref *kref) +{ + struct ceph_mon_generic_request *req = + container_of(kref, struct ceph_mon_generic_request, kref); + + if (req->reply) + ceph_msg_put(req->reply); + if (req->request) + ceph_msg_put(req->request); + + kfree(req); +} + +static void put_generic_request(struct ceph_mon_generic_request *req) +{ + kref_put(&req->kref, release_generic_request); +} + +static void get_generic_request(struct ceph_mon_generic_request *req) +{ + kref_get(&req->kref); +} + +static struct ceph_msg *get_generic_reply(struct ceph_connection *con, + struct ceph_msg_header *hdr, + int *skip) +{ + struct ceph_mon_client *monc = con->private; + struct ceph_mon_generic_request *req; + u64 tid = le64_to_cpu(hdr->tid); + struct ceph_msg *m; + + mutex_lock(&monc->mutex); + req = __lookup_generic_req(monc, tid); + if (!req) { + dout("get_generic_reply %lld dne\n", tid); + *skip = 1; + m = NULL; + } else { + dout("get_generic_reply %lld got %p\n", tid, req->reply); + m = ceph_msg_get(req->reply); + /* + * we don't need to track the connection reading into + * this reply because we only have one open connection + * at a time, ever. + */ + } + mutex_unlock(&monc->mutex); + return m; +} + +static int do_generic_request(struct ceph_mon_client *monc, + struct ceph_mon_generic_request *req) +{ + int err; + + /* register request */ + mutex_lock(&monc->mutex); + req->tid = ++monc->last_tid; + req->request->hdr.tid = cpu_to_le64(req->tid); + __insert_generic_request(monc, req); + monc->num_generic_requests++; + ceph_con_send(monc->con, ceph_msg_get(req->request)); + mutex_unlock(&monc->mutex); + + err = wait_for_completion_interruptible(&req->completion); + + mutex_lock(&monc->mutex); + rb_erase(&req->node, &monc->generic_request_tree); + monc->num_generic_requests--; + mutex_unlock(&monc->mutex); + + if (!err) + err = req->result; + return err; +} + +/* + * statfs + */ +static void handle_statfs_reply(struct ceph_mon_client *monc, + struct ceph_msg *msg) +{ + struct ceph_mon_generic_request *req; + struct ceph_mon_statfs_reply *reply = msg->front.iov_base; + u64 tid = le64_to_cpu(msg->hdr.tid); + + if (msg->front.iov_len != sizeof(*reply)) + goto bad; + dout("handle_statfs_reply %p tid %llu\n", msg, tid); + + mutex_lock(&monc->mutex); + req = __lookup_generic_req(monc, tid); + if (req) { + *(struct ceph_statfs *)req->buf = reply->st; + req->result = 0; + get_generic_request(req); + } + mutex_unlock(&monc->mutex); + if (req) { + complete_all(&req->completion); + put_generic_request(req); + } + return; + +bad: + pr_err("corrupt generic reply, tid %llu\n", tid); + ceph_msg_dump(msg); +} + +/* + * Do a synchronous statfs(). + */ +int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf) +{ + struct ceph_mon_generic_request *req; + struct ceph_mon_statfs *h; + int err; + + req = kzalloc(sizeof(*req), GFP_NOFS); + if (!req) + return -ENOMEM; + + kref_init(&req->kref); + req->buf = buf; + req->buf_len = sizeof(*buf); + init_completion(&req->completion); + + err = -ENOMEM; + req->request = ceph_msg_new(CEPH_MSG_STATFS, sizeof(*h), GFP_NOFS); + if (!req->request) + goto out; + req->reply = ceph_msg_new(CEPH_MSG_STATFS_REPLY, 1024, GFP_NOFS); + if (!req->reply) + goto out; + + /* fill out request */ + h = req->request->front.iov_base; + h->monhdr.have_version = 0; + h->monhdr.session_mon = cpu_to_le16(-1); + h->monhdr.session_mon_tid = 0; + h->fsid = monc->monmap->fsid; + + err = do_generic_request(monc, req); + +out: + kref_put(&req->kref, release_generic_request); + return err; +} +EXPORT_SYMBOL(ceph_monc_do_statfs); + +/* + * pool ops + */ +static int get_poolop_reply_buf(const char *src, size_t src_len, + char *dst, size_t dst_len) +{ + u32 buf_len; + + if (src_len != sizeof(u32) + dst_len) + return -EINVAL; + + buf_len = le32_to_cpu(*(u32 *)src); + if (buf_len != dst_len) + return -EINVAL; + + memcpy(dst, src + sizeof(u32), dst_len); + return 0; +} + +static void handle_poolop_reply(struct ceph_mon_client *monc, + struct ceph_msg *msg) +{ + struct ceph_mon_generic_request *req; + struct ceph_mon_poolop_reply *reply = msg->front.iov_base; + u64 tid = le64_to_cpu(msg->hdr.tid); + + if (msg->front.iov_len < sizeof(*reply)) + goto bad; + dout("handle_poolop_reply %p tid %llu\n", msg, tid); + + mutex_lock(&monc->mutex); + req = __lookup_generic_req(monc, tid); + if (req) { + if (req->buf_len && + get_poolop_reply_buf(msg->front.iov_base + sizeof(*reply), + msg->front.iov_len - sizeof(*reply), + req->buf, req->buf_len) < 0) { + mutex_unlock(&monc->mutex); + goto bad; + } + req->result = le32_to_cpu(reply->reply_code); + get_generic_request(req); + } + mutex_unlock(&monc->mutex); + if (req) { + complete(&req->completion); + put_generic_request(req); + } + return; + +bad: + pr_err("corrupt generic reply, tid %llu\n", tid); + ceph_msg_dump(msg); +} + +/* + * Do a synchronous pool op. + */ +int ceph_monc_do_poolop(struct ceph_mon_client *monc, u32 op, + u32 pool, u64 snapid, + char *buf, int len) +{ + struct ceph_mon_generic_request *req; + struct ceph_mon_poolop *h; + int err; + + req = kzalloc(sizeof(*req), GFP_NOFS); + if (!req) + return -ENOMEM; + + kref_init(&req->kref); + req->buf = buf; + req->buf_len = len; + init_completion(&req->completion); + + err = -ENOMEM; + req->request = ceph_msg_new(CEPH_MSG_POOLOP, sizeof(*h), GFP_NOFS); + if (!req->request) + goto out; + req->reply = ceph_msg_new(CEPH_MSG_POOLOP_REPLY, 1024, GFP_NOFS); + if (!req->reply) + goto out; + + /* fill out request */ + req->request->hdr.version = cpu_to_le16(2); + h = req->request->front.iov_base; + h->monhdr.have_version = 0; + h->monhdr.session_mon = cpu_to_le16(-1); + h->monhdr.session_mon_tid = 0; + h->fsid = monc->monmap->fsid; + h->pool = cpu_to_le32(pool); + h->op = cpu_to_le32(op); + h->auid = 0; + h->snapid = cpu_to_le64(snapid); + h->name_len = 0; + + err = do_generic_request(monc, req); + +out: + kref_put(&req->kref, release_generic_request); + return err; +} + +int ceph_monc_create_snapid(struct ceph_mon_client *monc, + u32 pool, u64 *snapid) +{ + return ceph_monc_do_poolop(monc, POOL_OP_CREATE_UNMANAGED_SNAP, + pool, 0, (char *)snapid, sizeof(*snapid)); + +} +EXPORT_SYMBOL(ceph_monc_create_snapid); + +int ceph_monc_delete_snapid(struct ceph_mon_client *monc, + u32 pool, u64 snapid) +{ + return ceph_monc_do_poolop(monc, POOL_OP_CREATE_UNMANAGED_SNAP, + pool, snapid, 0, 0); + +} + +/* + * Resend pending generic requests. + */ +static void __resend_generic_request(struct ceph_mon_client *monc) +{ + struct ceph_mon_generic_request *req; + struct rb_node *p; + + for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) { + req = rb_entry(p, struct ceph_mon_generic_request, node); + ceph_con_revoke(monc->con, req->request); + ceph_con_send(monc->con, ceph_msg_get(req->request)); + } +} + +/* + * Delayed work. If we haven't mounted yet, retry. Otherwise, + * renew/retry subscription as needed (in case it is timing out, or we + * got an ENOMEM). And keep the monitor connection alive. + */ +static void delayed_work(struct work_struct *work) +{ + struct ceph_mon_client *monc = + container_of(work, struct ceph_mon_client, delayed_work.work); + + dout("monc delayed_work\n"); + mutex_lock(&monc->mutex); + if (monc->hunting) { + __close_session(monc); + __open_session(monc); /* continue hunting */ + } else { + ceph_con_keepalive(monc->con); + + __validate_auth(monc); + + if (monc->auth->ops->is_authenticated(monc->auth)) + __send_subscribe(monc); + } + __schedule_delayed(monc); + mutex_unlock(&monc->mutex); +} + +/* + * On startup, we build a temporary monmap populated with the IPs + * provided by mount(2). + */ +static int build_initial_monmap(struct ceph_mon_client *monc) +{ + struct ceph_options *opt = monc->client->options; + struct ceph_entity_addr *mon_addr = opt->mon_addr; + int num_mon = opt->num_mon; + int i; + + /* build initial monmap */ + monc->monmap = kzalloc(sizeof(*monc->monmap) + + num_mon*sizeof(monc->monmap->mon_inst[0]), + GFP_KERNEL); + if (!monc->monmap) + return -ENOMEM; + for (i = 0; i < num_mon; i++) { + monc->monmap->mon_inst[i].addr = mon_addr[i]; + monc->monmap->mon_inst[i].addr.nonce = 0; + monc->monmap->mon_inst[i].name.type = + CEPH_ENTITY_TYPE_MON; + monc->monmap->mon_inst[i].name.num = cpu_to_le64(i); + } + monc->monmap->num_mon = num_mon; + monc->have_fsid = false; + return 0; +} + +int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl) +{ + int err = 0; + + dout("init\n"); + memset(monc, 0, sizeof(*monc)); + monc->client = cl; + monc->monmap = NULL; + mutex_init(&monc->mutex); + + err = build_initial_monmap(monc); + if (err) + goto out; + + monc->con = NULL; + + /* authentication */ + monc->auth = ceph_auth_init(cl->options->name, + cl->options->secret); + if (IS_ERR(monc->auth)) + return PTR_ERR(monc->auth); + monc->auth->want_keys = + CEPH_ENTITY_TYPE_AUTH | CEPH_ENTITY_TYPE_MON | + CEPH_ENTITY_TYPE_OSD | CEPH_ENTITY_TYPE_MDS; + + /* msgs */ + err = -ENOMEM; + monc->m_subscribe_ack = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE_ACK, + sizeof(struct ceph_mon_subscribe_ack), + GFP_NOFS); + if (!monc->m_subscribe_ack) + goto out_monmap; + + monc->m_subscribe = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE, 96, GFP_NOFS); + if (!monc->m_subscribe) + goto out_subscribe_ack; + + monc->m_auth_reply = ceph_msg_new(CEPH_MSG_AUTH_REPLY, 4096, GFP_NOFS); + if (!monc->m_auth_reply) + goto out_subscribe; + + monc->m_auth = ceph_msg_new(CEPH_MSG_AUTH, 4096, GFP_NOFS); + monc->pending_auth = 0; + if (!monc->m_auth) + goto out_auth_reply; + + monc->cur_mon = -1; + monc->hunting = true; + monc->sub_renew_after = jiffies; + monc->sub_sent = 0; + + INIT_DELAYED_WORK(&monc->delayed_work, delayed_work); + monc->generic_request_tree = RB_ROOT; + monc->num_generic_requests = 0; + monc->last_tid = 0; + + monc->have_mdsmap = 0; + monc->have_osdmap = 0; + monc->want_next_osdmap = 1; + return 0; + +out_auth_reply: + ceph_msg_put(monc->m_auth_reply); +out_subscribe: + ceph_msg_put(monc->m_subscribe); +out_subscribe_ack: + ceph_msg_put(monc->m_subscribe_ack); +out_monmap: + kfree(monc->monmap); +out: + return err; +} +EXPORT_SYMBOL(ceph_monc_init); + +void ceph_monc_stop(struct ceph_mon_client *monc) +{ + dout("stop\n"); + cancel_delayed_work_sync(&monc->delayed_work); + + mutex_lock(&monc->mutex); + __close_session(monc); + if (monc->con) { + monc->con->private = NULL; + monc->con->ops->put(monc->con); + monc->con = NULL; + } + mutex_unlock(&monc->mutex); + + ceph_auth_destroy(monc->auth); + + ceph_msg_put(monc->m_auth); + ceph_msg_put(monc->m_auth_reply); + ceph_msg_put(monc->m_subscribe); + ceph_msg_put(monc->m_subscribe_ack); + + kfree(monc->monmap); +} +EXPORT_SYMBOL(ceph_monc_stop); + +static void handle_auth_reply(struct ceph_mon_client *monc, + struct ceph_msg *msg) +{ + int ret; + int was_auth = 0; + + mutex_lock(&monc->mutex); + if (monc->auth->ops) + was_auth = monc->auth->ops->is_authenticated(monc->auth); + monc->pending_auth = 0; + ret = ceph_handle_auth_reply(monc->auth, msg->front.iov_base, + msg->front.iov_len, + monc->m_auth->front.iov_base, + monc->m_auth->front_max); + if (ret < 0) { + monc->client->auth_err = ret; + wake_up_all(&monc->client->auth_wq); + } else if (ret > 0) { + __send_prepared_auth_request(monc, ret); + } else if (!was_auth && monc->auth->ops->is_authenticated(monc->auth)) { + dout("authenticated, starting session\n"); + + monc->client->msgr->inst.name.type = CEPH_ENTITY_TYPE_CLIENT; + monc->client->msgr->inst.name.num = + cpu_to_le64(monc->auth->global_id); + + __send_subscribe(monc); + __resend_generic_request(monc); + } + mutex_unlock(&monc->mutex); +} + +static int __validate_auth(struct ceph_mon_client *monc) +{ + int ret; + + if (monc->pending_auth) + return 0; + + ret = ceph_build_auth(monc->auth, monc->m_auth->front.iov_base, + monc->m_auth->front_max); + if (ret <= 0) + return ret; /* either an error, or no need to authenticate */ + __send_prepared_auth_request(monc, ret); + return 0; +} + +int ceph_monc_validate_auth(struct ceph_mon_client *monc) +{ + int ret; + + mutex_lock(&monc->mutex); + ret = __validate_auth(monc); + mutex_unlock(&monc->mutex); + return ret; +} +EXPORT_SYMBOL(ceph_monc_validate_auth); + +/* + * handle incoming message + */ +static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) +{ + struct ceph_mon_client *monc = con->private; + int type = le16_to_cpu(msg->hdr.type); + + if (!monc) + return; + + switch (type) { + case CEPH_MSG_AUTH_REPLY: + handle_auth_reply(monc, msg); + break; + + case CEPH_MSG_MON_SUBSCRIBE_ACK: + handle_subscribe_ack(monc, msg); + break; + + case CEPH_MSG_STATFS_REPLY: + handle_statfs_reply(monc, msg); + break; + + case CEPH_MSG_POOLOP_REPLY: + handle_poolop_reply(monc, msg); + break; + + case CEPH_MSG_MON_MAP: + ceph_monc_handle_map(monc, msg); + break; + + case CEPH_MSG_OSD_MAP: + ceph_osdc_handle_map(&monc->client->osdc, msg); + break; + + default: + /* can the chained handler handle it? */ + if (monc->client->extra_mon_dispatch && + monc->client->extra_mon_dispatch(monc->client, msg) == 0) + break; + + pr_err("received unknown message type %d %s\n", type, + ceph_msg_type_name(type)); + } + ceph_msg_put(msg); +} + +/* + * Allocate memory for incoming message + */ +static struct ceph_msg *mon_alloc_msg(struct ceph_connection *con, + struct ceph_msg_header *hdr, + int *skip) +{ + struct ceph_mon_client *monc = con->private; + int type = le16_to_cpu(hdr->type); + int front_len = le32_to_cpu(hdr->front_len); + struct ceph_msg *m = NULL; + + *skip = 0; + + switch (type) { + case CEPH_MSG_MON_SUBSCRIBE_ACK: + m = ceph_msg_get(monc->m_subscribe_ack); + break; + case CEPH_MSG_POOLOP_REPLY: + case CEPH_MSG_STATFS_REPLY: + return get_generic_reply(con, hdr, skip); + case CEPH_MSG_AUTH_REPLY: + m = ceph_msg_get(monc->m_auth_reply); + break; + case CEPH_MSG_MON_MAP: + case CEPH_MSG_MDS_MAP: + case CEPH_MSG_OSD_MAP: + m = ceph_msg_new(type, front_len, GFP_NOFS); + break; + } + + if (!m) { + pr_info("alloc_msg unknown type %d\n", type); + *skip = 1; + } + return m; +} + +/* + * If the monitor connection resets, pick a new monitor and resubmit + * any pending requests. + */ +static void mon_fault(struct ceph_connection *con) +{ + struct ceph_mon_client *monc = con->private; + + if (!monc) + return; + + dout("mon_fault\n"); + mutex_lock(&monc->mutex); + if (!con->private) + goto out; + + if (monc->con && !monc->hunting) + pr_info("mon%d %s session lost, " + "hunting for new mon\n", monc->cur_mon, + ceph_pr_addr(&monc->con->peer_addr.in_addr)); + + __close_session(monc); + if (!monc->hunting) { + /* start hunting */ + monc->hunting = true; + __open_session(monc); + } else { + /* already hunting, let's wait a bit */ + __schedule_delayed(monc); + } +out: + mutex_unlock(&monc->mutex); +} + +static const struct ceph_connection_operations mon_con_ops = { + .get = ceph_con_get, + .put = ceph_con_put, + .dispatch = dispatch, + .fault = mon_fault, + .alloc_msg = mon_alloc_msg, +}; diff --git a/net/ceph/msgpool.c b/net/ceph/msgpool.c new file mode 100644 index 0000000..d5f2d97 --- /dev/null +++ b/net/ceph/msgpool.c @@ -0,0 +1,64 @@ +#include + +#include +#include +#include +#include + +#include + +static void *alloc_fn(gfp_t gfp_mask, void *arg) +{ + struct ceph_msgpool *pool = arg; + void *p; + + p = ceph_msg_new(0, pool->front_len, gfp_mask); + if (!p) + pr_err("msgpool %s alloc failed\n", pool->name); + return p; +} + +static void free_fn(void *element, void *arg) +{ + ceph_msg_put(element); +} + +int ceph_msgpool_init(struct ceph_msgpool *pool, + int front_len, int size, bool blocking, const char *name) +{ + pool->front_len = front_len; + pool->pool = mempool_create(size, alloc_fn, free_fn, pool); + if (!pool->pool) + return -ENOMEM; + pool->name = name; + return 0; +} + +void ceph_msgpool_destroy(struct ceph_msgpool *pool) +{ + mempool_destroy(pool->pool); +} + +struct ceph_msg *ceph_msgpool_get(struct ceph_msgpool *pool, + int front_len) +{ + if (front_len > pool->front_len) { + pr_err("msgpool_get pool %s need front %d, pool size is %d\n", + pool->name, front_len, pool->front_len); + WARN_ON(1); + + /* try to alloc a fresh message */ + return ceph_msg_new(0, front_len, GFP_NOFS); + } + + return mempool_alloc(pool->pool, GFP_NOFS); +} + +void ceph_msgpool_put(struct ceph_msgpool *pool, struct ceph_msg *msg) +{ + /* reset msg front_len; user may have changed it */ + msg->front.iov_len = pool->front_len; + msg->hdr.front_len = cpu_to_le32(pool->front_len); + + kref_init(&msg->kref); /* retake single ref */ +} diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c new file mode 100644 index 0000000..7939199 --- /dev/null +++ b/net/ceph/osd_client.c @@ -0,0 +1,1773 @@ +#include + +#include +#include +#include +#include +#include +#include +#include +#ifdef CONFIG_BLOCK +#include +#endif + +#include +#include +#include +#include +#include +#include + +#define OSD_OP_FRONT_LEN 4096 +#define OSD_OPREPLY_FRONT_LEN 512 + +static const struct ceph_connection_operations osd_con_ops; +static int __kick_requests(struct ceph_osd_client *osdc, + struct ceph_osd *kickosd); + +static void kick_requests(struct ceph_osd_client *osdc, struct ceph_osd *osd); + +static int op_needs_trail(int op) +{ + switch (op) { + case CEPH_OSD_OP_GETXATTR: + case CEPH_OSD_OP_SETXATTR: + case CEPH_OSD_OP_CMPXATTR: + case CEPH_OSD_OP_CALL: + return 1; + default: + return 0; + } +} + +static int op_has_extent(int op) +{ + return (op == CEPH_OSD_OP_READ || + op == CEPH_OSD_OP_WRITE); +} + +void ceph_calc_raw_layout(struct ceph_osd_client *osdc, + struct ceph_file_layout *layout, + u64 snapid, + u64 off, u64 *plen, u64 *bno, + struct ceph_osd_request *req, + struct ceph_osd_req_op *op) +{ + struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; + u64 orig_len = *plen; + u64 objoff, objlen; /* extent in object */ + + reqhead->snapid = cpu_to_le64(snapid); + + /* object extent? */ + ceph_calc_file_object_mapping(layout, off, plen, bno, + &objoff, &objlen); + if (*plen < orig_len) + dout(" skipping last %llu, final file extent %llu~%llu\n", + orig_len - *plen, off, *plen); + + if (op_has_extent(op->op)) { + op->extent.offset = objoff; + op->extent.length = objlen; + } + req->r_num_pages = calc_pages_for(off, *plen); + if (op->op == CEPH_OSD_OP_WRITE) + op->payload_len = *plen; + + dout("calc_layout bno=%llx %llu~%llu (%d pages)\n", + *bno, objoff, objlen, req->r_num_pages); + +} +EXPORT_SYMBOL(ceph_calc_raw_layout); + +/* + * Implement client access to distributed object storage cluster. + * + * All data objects are stored within a cluster/cloud of OSDs, or + * "object storage devices." (Note that Ceph OSDs have _nothing_ to + * do with the T10 OSD extensions to SCSI.) Ceph OSDs are simply + * remote daemons serving up and coordinating consistent and safe + * access to storage. + * + * Cluster membership and the mapping of data objects onto storage devices + * are described by the osd map. + * + * We keep track of pending OSD requests (read, write), resubmit + * requests to different OSDs when the cluster topology/data layout + * change, or retry the affected requests when the communications + * channel with an OSD is reset. + */ + +/* + * calculate the mapping of a file extent onto an object, and fill out the + * request accordingly. shorten extent as necessary if it crosses an + * object boundary. + * + * fill osd op in request message. + */ +static void calc_layout(struct ceph_osd_client *osdc, + struct ceph_vino vino, + struct ceph_file_layout *layout, + u64 off, u64 *plen, + struct ceph_osd_request *req, + struct ceph_osd_req_op *op) +{ + u64 bno; + + ceph_calc_raw_layout(osdc, layout, vino.snap, off, + plen, &bno, req, op); + + sprintf(req->r_oid, "%llx.%08llx", vino.ino, bno); + req->r_oid_len = strlen(req->r_oid); +} + +/* + * requests + */ +void ceph_osdc_release_request(struct kref *kref) +{ + struct ceph_osd_request *req = container_of(kref, + struct ceph_osd_request, + r_kref); + + if (req->r_request) + ceph_msg_put(req->r_request); + if (req->r_reply) + ceph_msg_put(req->r_reply); + if (req->r_con_filling_msg) { + dout("release_request revoking pages %p from con %p\n", + req->r_pages, req->r_con_filling_msg); + ceph_con_revoke_message(req->r_con_filling_msg, + req->r_reply); + ceph_con_put(req->r_con_filling_msg); + } + if (req->r_own_pages) + ceph_release_page_vector(req->r_pages, + req->r_num_pages); +#ifdef CONFIG_BLOCK + if (req->r_bio) + bio_put(req->r_bio); +#endif + ceph_put_snap_context(req->r_snapc); + if (req->r_trail) { + ceph_pagelist_release(req->r_trail); + kfree(req->r_trail); + } + if (req->r_mempool) + mempool_free(req, req->r_osdc->req_mempool); + else + kfree(req); +} +EXPORT_SYMBOL(ceph_osdc_release_request); + +static int get_num_ops(struct ceph_osd_req_op *ops, int *needs_trail) +{ + int i = 0; + + if (needs_trail) + *needs_trail = 0; + while (ops[i].op) { + if (needs_trail && op_needs_trail(ops[i].op)) + *needs_trail = 1; + i++; + } + + return i; +} + +struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, + int flags, + struct ceph_snap_context *snapc, + struct ceph_osd_req_op *ops, + bool use_mempool, + gfp_t gfp_flags, + struct page **pages, + struct bio *bio) +{ + struct ceph_osd_request *req; + struct ceph_msg *msg; + int needs_trail; + int num_op = get_num_ops(ops, &needs_trail); + size_t msg_size = sizeof(struct ceph_osd_request_head); + + msg_size += num_op*sizeof(struct ceph_osd_op); + + if (use_mempool) { + req = mempool_alloc(osdc->req_mempool, gfp_flags); + memset(req, 0, sizeof(*req)); + } else { + req = kzalloc(sizeof(*req), gfp_flags); + } + if (req == NULL) + return NULL; + + req->r_osdc = osdc; + req->r_mempool = use_mempool; + + kref_init(&req->r_kref); + init_completion(&req->r_completion); + init_completion(&req->r_safe_completion); + INIT_LIST_HEAD(&req->r_unsafe_item); + req->r_flags = flags; + + WARN_ON((flags & (CEPH_OSD_FLAG_READ|CEPH_OSD_FLAG_WRITE)) == 0); + + /* create reply message */ + if (use_mempool) + msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0); + else + msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, + OSD_OPREPLY_FRONT_LEN, gfp_flags); + if (!msg) { + ceph_osdc_put_request(req); + return NULL; + } + req->r_reply = msg; + + /* allocate space for the trailing data */ + if (needs_trail) { + req->r_trail = kmalloc(sizeof(struct ceph_pagelist), gfp_flags); + if (!req->r_trail) { + ceph_osdc_put_request(req); + return NULL; + } + ceph_pagelist_init(req->r_trail); + } + /* create request message; allow space for oid */ + msg_size += 40; + if (snapc) + msg_size += sizeof(u64) * snapc->num_snaps; + if (use_mempool) + msg = ceph_msgpool_get(&osdc->msgpool_op, 0); + else + msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags); + if (!msg) { + ceph_osdc_put_request(req); + return NULL; + } + + msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP); + memset(msg->front.iov_base, 0, msg->front.iov_len); + + req->r_request = msg; + req->r_pages = pages; +#ifdef CONFIG_BLOCK + if (bio) { + req->r_bio = bio; + bio_get(req->r_bio); + } +#endif + + return req; +} +EXPORT_SYMBOL(ceph_osdc_alloc_request); + +static void osd_req_encode_op(struct ceph_osd_request *req, + struct ceph_osd_op *dst, + struct ceph_osd_req_op *src) +{ + dst->op = cpu_to_le16(src->op); + + switch (dst->op) { + case CEPH_OSD_OP_READ: + case CEPH_OSD_OP_WRITE: + dst->extent.offset = + cpu_to_le64(src->extent.offset); + dst->extent.length = + cpu_to_le64(src->extent.length); + dst->extent.truncate_size = + cpu_to_le64(src->extent.truncate_size); + dst->extent.truncate_seq = + cpu_to_le32(src->extent.truncate_seq); + break; + + case CEPH_OSD_OP_GETXATTR: + case CEPH_OSD_OP_SETXATTR: + case CEPH_OSD_OP_CMPXATTR: + BUG_ON(!req->r_trail); + + dst->xattr.name_len = cpu_to_le32(src->xattr.name_len); + dst->xattr.value_len = cpu_to_le32(src->xattr.value_len); + dst->xattr.cmp_op = src->xattr.cmp_op; + dst->xattr.cmp_mode = src->xattr.cmp_mode; + ceph_pagelist_append(req->r_trail, src->xattr.name, + src->xattr.name_len); + ceph_pagelist_append(req->r_trail, src->xattr.val, + src->xattr.value_len); + break; + case CEPH_OSD_OP_CALL: + BUG_ON(!req->r_trail); + + dst->cls.class_len = src->cls.class_len; + dst->cls.method_len = src->cls.method_len; + dst->cls.indata_len = cpu_to_le32(src->cls.indata_len); + + ceph_pagelist_append(req->r_trail, src->cls.class_name, + src->cls.class_len); + ceph_pagelist_append(req->r_trail, src->cls.method_name, + src->cls.method_len); + ceph_pagelist_append(req->r_trail, src->cls.indata, + src->cls.indata_len); + break; + case CEPH_OSD_OP_ROLLBACK: + dst->snap.snapid = cpu_to_le64(src->snap.snapid); + break; + case CEPH_OSD_OP_STARTSYNC: + break; + default: + pr_err("unrecognized osd opcode %d\n", dst->op); + WARN_ON(1); + break; + } + dst->payload_len = cpu_to_le32(src->payload_len); +} + +/* + * build new request AND message + * + */ +void ceph_osdc_build_request(struct ceph_osd_request *req, + u64 off, u64 *plen, + struct ceph_osd_req_op *src_ops, + struct ceph_snap_context *snapc, + struct timespec *mtime, + const char *oid, + int oid_len) +{ + struct ceph_msg *msg = req->r_request; + struct ceph_osd_request_head *head; + struct ceph_osd_req_op *src_op; + struct ceph_osd_op *op; + void *p; + int num_op = get_num_ops(src_ops, NULL); + size_t msg_size = sizeof(*head) + num_op*sizeof(*op); + int flags = req->r_flags; + u64 data_len = 0; + int i; + + head = msg->front.iov_base; + op = (void *)(head + 1); + p = (void *)(op + num_op); + + req->r_snapc = ceph_get_snap_context(snapc); + + head->client_inc = cpu_to_le32(1); /* always, for now. */ + head->flags = cpu_to_le32(flags); + if (flags & CEPH_OSD_FLAG_WRITE) + ceph_encode_timespec(&head->mtime, mtime); + head->num_ops = cpu_to_le16(num_op); + + + /* fill in oid */ + head->object_len = cpu_to_le32(oid_len); + memcpy(p, oid, oid_len); + p += oid_len; + + src_op = src_ops; + while (src_op->op) { + osd_req_encode_op(req, op, src_op); + src_op++; + op++; + } + + if (req->r_trail) + data_len += req->r_trail->length; + + if (snapc) { + head->snap_seq = cpu_to_le64(snapc->seq); + head->num_snaps = cpu_to_le32(snapc->num_snaps); + for (i = 0; i < snapc->num_snaps; i++) { + put_unaligned_le64(snapc->snaps[i], p); + p += sizeof(u64); + } + } + + if (flags & CEPH_OSD_FLAG_WRITE) { + req->r_request->hdr.data_off = cpu_to_le16(off); + req->r_request->hdr.data_len = cpu_to_le32(*plen + data_len); + } else if (data_len) { + req->r_request->hdr.data_off = 0; + req->r_request->hdr.data_len = cpu_to_le32(data_len); + } + + BUG_ON(p > msg->front.iov_base + msg->front.iov_len); + msg_size = p - msg->front.iov_base; + msg->front.iov_len = msg_size; + msg->hdr.front_len = cpu_to_le32(msg_size); + return; +} +EXPORT_SYMBOL(ceph_osdc_build_request); + +/* + * build new request AND message, calculate layout, and adjust file + * extent as needed. + * + * if the file was recently truncated, we include information about its + * old and new size so that the object can be updated appropriately. (we + * avoid synchronously deleting truncated objects because it's slow.) + * + * if @do_sync, include a 'startsync' command so that the osd will flush + * data quickly. + */ +struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, + struct ceph_file_layout *layout, + struct ceph_vino vino, + u64 off, u64 *plen, + int opcode, int flags, + struct ceph_snap_context *snapc, + int do_sync, + u32 truncate_seq, + u64 truncate_size, + struct timespec *mtime, + bool use_mempool, int num_reply) +{ + struct ceph_osd_req_op ops[3]; + struct ceph_osd_request *req; + + ops[0].op = opcode; + ops[0].extent.truncate_seq = truncate_seq; + ops[0].extent.truncate_size = truncate_size; + ops[0].payload_len = 0; + + if (do_sync) { + ops[1].op = CEPH_OSD_OP_STARTSYNC; + ops[1].payload_len = 0; + ops[2].op = 0; + } else + ops[1].op = 0; + + req = ceph_osdc_alloc_request(osdc, flags, + snapc, ops, + use_mempool, + GFP_NOFS, NULL, NULL); + if (IS_ERR(req)) + return req; + + /* calculate max write size */ + calc_layout(osdc, vino, layout, off, plen, req, ops); + req->r_file_layout = *layout; /* keep a copy */ + + ceph_osdc_build_request(req, off, plen, ops, + snapc, + mtime, + req->r_oid, req->r_oid_len); + + return req; +} +EXPORT_SYMBOL(ceph_osdc_new_request); + +/* + * We keep osd requests in an rbtree, sorted by ->r_tid. + */ +static void __insert_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *new) +{ + struct rb_node **p = &osdc->requests.rb_node; + struct rb_node *parent = NULL; + struct ceph_osd_request *req = NULL; + + while (*p) { + parent = *p; + req = rb_entry(parent, struct ceph_osd_request, r_node); + if (new->r_tid < req->r_tid) + p = &(*p)->rb_left; + else if (new->r_tid > req->r_tid) + p = &(*p)->rb_right; + else + BUG(); + } + + rb_link_node(&new->r_node, parent, p); + rb_insert_color(&new->r_node, &osdc->requests); +} + +static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc, + u64 tid) +{ + struct ceph_osd_request *req; + struct rb_node *n = osdc->requests.rb_node; + + while (n) { + req = rb_entry(n, struct ceph_osd_request, r_node); + if (tid < req->r_tid) + n = n->rb_left; + else if (tid > req->r_tid) + n = n->rb_right; + else + return req; + } + return NULL; +} + +static struct ceph_osd_request * +__lookup_request_ge(struct ceph_osd_client *osdc, + u64 tid) +{ + struct ceph_osd_request *req; + struct rb_node *n = osdc->requests.rb_node; + + while (n) { + req = rb_entry(n, struct ceph_osd_request, r_node); + if (tid < req->r_tid) { + if (!n->rb_left) + return req; + n = n->rb_left; + } else if (tid > req->r_tid) { + n = n->rb_right; + } else { + return req; + } + } + return NULL; +} + + +/* + * If the osd connection drops, we need to resubmit all requests. + */ +static void osd_reset(struct ceph_connection *con) +{ + struct ceph_osd *osd = con->private; + struct ceph_osd_client *osdc; + + if (!osd) + return; + dout("osd_reset osd%d\n", osd->o_osd); + osdc = osd->o_osdc; + down_read(&osdc->map_sem); + kick_requests(osdc, osd); + up_read(&osdc->map_sem); +} + +/* + * Track open sessions with osds. + */ +static struct ceph_osd *create_osd(struct ceph_osd_client *osdc) +{ + struct ceph_osd *osd; + + osd = kzalloc(sizeof(*osd), GFP_NOFS); + if (!osd) + return NULL; + + atomic_set(&osd->o_ref, 1); + osd->o_osdc = osdc; + INIT_LIST_HEAD(&osd->o_requests); + INIT_LIST_HEAD(&osd->o_osd_lru); + osd->o_incarnation = 1; + + ceph_con_init(osdc->client->msgr, &osd->o_con); + osd->o_con.private = osd; + osd->o_con.ops = &osd_con_ops; + osd->o_con.peer_name.type = CEPH_ENTITY_TYPE_OSD; + + INIT_LIST_HEAD(&osd->o_keepalive_item); + return osd; +} + +static struct ceph_osd *get_osd(struct ceph_osd *osd) +{ + if (atomic_inc_not_zero(&osd->o_ref)) { + dout("get_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref)-1, + atomic_read(&osd->o_ref)); + return osd; + } else { + dout("get_osd %p FAIL\n", osd); + return NULL; + } +} + +static void put_osd(struct ceph_osd *osd) +{ + dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref), + atomic_read(&osd->o_ref) - 1); + if (atomic_dec_and_test(&osd->o_ref)) { + struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth; + + if (osd->o_authorizer) + ac->ops->destroy_authorizer(ac, osd->o_authorizer); + kfree(osd); + } +} + +/* + * remove an osd from our map + */ +static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) +{ + dout("__remove_osd %p\n", osd); + BUG_ON(!list_empty(&osd->o_requests)); + rb_erase(&osd->o_node, &osdc->osds); + list_del_init(&osd->o_osd_lru); + ceph_con_close(&osd->o_con); + put_osd(osd); +} + +static void __move_osd_to_lru(struct ceph_osd_client *osdc, + struct ceph_osd *osd) +{ + dout("__move_osd_to_lru %p\n", osd); + BUG_ON(!list_empty(&osd->o_osd_lru)); + list_add_tail(&osd->o_osd_lru, &osdc->osd_lru); + osd->lru_ttl = jiffies + osdc->client->options->osd_idle_ttl * HZ; +} + +static void __remove_osd_from_lru(struct ceph_osd *osd) +{ + dout("__remove_osd_from_lru %p\n", osd); + if (!list_empty(&osd->o_osd_lru)) + list_del_init(&osd->o_osd_lru); +} + +static void remove_old_osds(struct ceph_osd_client *osdc, int remove_all) +{ + struct ceph_osd *osd, *nosd; + + dout("__remove_old_osds %p\n", osdc); + mutex_lock(&osdc->request_mutex); + list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) { + if (!remove_all && time_before(jiffies, osd->lru_ttl)) + break; + __remove_osd(osdc, osd); + } + mutex_unlock(&osdc->request_mutex); +} + +/* + * reset osd connect + */ +static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) +{ + struct ceph_osd_request *req; + int ret = 0; + + dout("__reset_osd %p osd%d\n", osd, osd->o_osd); + if (list_empty(&osd->o_requests)) { + __remove_osd(osdc, osd); + } else if (memcmp(&osdc->osdmap->osd_addr[osd->o_osd], + &osd->o_con.peer_addr, + sizeof(osd->o_con.peer_addr)) == 0 && + !ceph_con_opened(&osd->o_con)) { + dout(" osd addr hasn't changed and connection never opened," + " letting msgr retry"); + /* touch each r_stamp for handle_timeout()'s benfit */ + list_for_each_entry(req, &osd->o_requests, r_osd_item) + req->r_stamp = jiffies; + ret = -EAGAIN; + } else { + ceph_con_close(&osd->o_con); + ceph_con_open(&osd->o_con, &osdc->osdmap->osd_addr[osd->o_osd]); + osd->o_incarnation++; + } + return ret; +} + +static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new) +{ + struct rb_node **p = &osdc->osds.rb_node; + struct rb_node *parent = NULL; + struct ceph_osd *osd = NULL; + + while (*p) { + parent = *p; + osd = rb_entry(parent, struct ceph_osd, o_node); + if (new->o_osd < osd->o_osd) + p = &(*p)->rb_left; + else if (new->o_osd > osd->o_osd) + p = &(*p)->rb_right; + else + BUG(); + } + + rb_link_node(&new->o_node, parent, p); + rb_insert_color(&new->o_node, &osdc->osds); +} + +static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o) +{ + struct ceph_osd *osd; + struct rb_node *n = osdc->osds.rb_node; + + while (n) { + osd = rb_entry(n, struct ceph_osd, o_node); + if (o < osd->o_osd) + n = n->rb_left; + else if (o > osd->o_osd) + n = n->rb_right; + else + return osd; + } + return NULL; +} + +static void __schedule_osd_timeout(struct ceph_osd_client *osdc) +{ + schedule_delayed_work(&osdc->timeout_work, + osdc->client->options->osd_keepalive_timeout * HZ); +} + +static void __cancel_osd_timeout(struct ceph_osd_client *osdc) +{ + cancel_delayed_work(&osdc->timeout_work); +} + +/* + * Register request, assign tid. If this is the first request, set up + * the timeout event. + */ +static void register_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + mutex_lock(&osdc->request_mutex); + req->r_tid = ++osdc->last_tid; + req->r_request->hdr.tid = cpu_to_le64(req->r_tid); + INIT_LIST_HEAD(&req->r_req_lru_item); + + dout("register_request %p tid %lld\n", req, req->r_tid); + __insert_request(osdc, req); + ceph_osdc_get_request(req); + osdc->num_requests++; + + if (osdc->num_requests == 1) { + dout(" first request, scheduling timeout\n"); + __schedule_osd_timeout(osdc); + } + mutex_unlock(&osdc->request_mutex); +} + +/* + * called under osdc->request_mutex + */ +static void __unregister_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + dout("__unregister_request %p tid %lld\n", req, req->r_tid); + rb_erase(&req->r_node, &osdc->requests); + osdc->num_requests--; + + if (req->r_osd) { + /* make sure the original request isn't in flight. */ + ceph_con_revoke(&req->r_osd->o_con, req->r_request); + + list_del_init(&req->r_osd_item); + if (list_empty(&req->r_osd->o_requests)) + __move_osd_to_lru(osdc, req->r_osd); + req->r_osd = NULL; + } + + ceph_osdc_put_request(req); + + list_del_init(&req->r_req_lru_item); + if (osdc->num_requests == 0) { + dout(" no requests, canceling timeout\n"); + __cancel_osd_timeout(osdc); + } +} + +/* + * Cancel a previously queued request message + */ +static void __cancel_request(struct ceph_osd_request *req) +{ + if (req->r_sent && req->r_osd) { + ceph_con_revoke(&req->r_osd->o_con, req->r_request); + req->r_sent = 0; + } + list_del_init(&req->r_req_lru_item); +} + +/* + * Pick an osd (the first 'up' osd in the pg), allocate the osd struct + * (as needed), and set the request r_osd appropriately. If there is + * no up osd, set r_osd to NULL. + * + * Return 0 if unchanged, 1 if changed, or negative on error. + * + * Caller should hold map_sem for read and request_mutex. + */ +static int __map_osds(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; + struct ceph_pg pgid; + int acting[CEPH_PG_MAX_SIZE]; + int o = -1, num = 0; + int err; + + dout("map_osds %p tid %lld\n", req, req->r_tid); + err = ceph_calc_object_layout(&reqhead->layout, req->r_oid, + &req->r_file_layout, osdc->osdmap); + if (err) + return err; + pgid = reqhead->layout.ol_pgid; + req->r_pgid = pgid; + + err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting); + if (err > 0) { + o = acting[0]; + num = err; + } + + if ((req->r_osd && req->r_osd->o_osd == o && + req->r_sent >= req->r_osd->o_incarnation && + req->r_num_pg_osds == num && + memcmp(req->r_pg_osds, acting, sizeof(acting[0])*num) == 0) || + (req->r_osd == NULL && o == -1)) + return 0; /* no change */ + + dout("map_osds tid %llu pgid %d.%x osd%d (was osd%d)\n", + req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o, + req->r_osd ? req->r_osd->o_osd : -1); + + /* record full pg acting set */ + memcpy(req->r_pg_osds, acting, sizeof(acting[0]) * num); + req->r_num_pg_osds = num; + + if (req->r_osd) { + __cancel_request(req); + list_del_init(&req->r_osd_item); + req->r_osd = NULL; + } + + req->r_osd = __lookup_osd(osdc, o); + if (!req->r_osd && o >= 0) { + err = -ENOMEM; + req->r_osd = create_osd(osdc); + if (!req->r_osd) + goto out; + + dout("map_osds osd %p is osd%d\n", req->r_osd, o); + req->r_osd->o_osd = o; + req->r_osd->o_con.peer_name.num = cpu_to_le64(o); + __insert_osd(osdc, req->r_osd); + + ceph_con_open(&req->r_osd->o_con, &osdc->osdmap->osd_addr[o]); + } + + if (req->r_osd) { + __remove_osd_from_lru(req->r_osd); + list_add(&req->r_osd_item, &req->r_osd->o_requests); + } + err = 1; /* osd or pg changed */ + +out: + return err; +} + +/* + * caller should hold map_sem (for read) and request_mutex + */ +static int __send_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + struct ceph_osd_request_head *reqhead; + int err; + + err = __map_osds(osdc, req); + if (err < 0) + return err; + if (req->r_osd == NULL) { + dout("send_request %p no up osds in pg\n", req); + ceph_monc_request_next_osdmap(&osdc->client->monc); + return 0; + } + + dout("send_request %p tid %llu to osd%d flags %d\n", + req, req->r_tid, req->r_osd->o_osd, req->r_flags); + + reqhead = req->r_request->front.iov_base; + reqhead->osdmap_epoch = cpu_to_le32(osdc->osdmap->epoch); + reqhead->flags |= cpu_to_le32(req->r_flags); /* e.g., RETRY */ + reqhead->reassert_version = req->r_reassert_version; + + req->r_stamp = jiffies; + list_move_tail(&req->r_req_lru_item, &osdc->req_lru); + + ceph_msg_get(req->r_request); /* send consumes a ref */ + ceph_con_send(&req->r_osd->o_con, req->r_request); + req->r_sent = req->r_osd->o_incarnation; + return 0; +} + +/* + * Timeout callback, called every N seconds when 1 or more osd + * requests has been active for more than N seconds. When this + * happens, we ping all OSDs with requests who have timed out to + * ensure any communications channel reset is detected. Reset the + * request timeouts another N seconds in the future as we go. + * Reschedule the timeout event another N seconds in future (unless + * there are no open requests). + */ +static void handle_timeout(struct work_struct *work) +{ + struct ceph_osd_client *osdc = + container_of(work, struct ceph_osd_client, timeout_work.work); + struct ceph_osd_request *req, *last_req = NULL; + struct ceph_osd *osd; + unsigned long timeout = osdc->client->options->osd_timeout * HZ; + unsigned long keepalive = + osdc->client->options->osd_keepalive_timeout * HZ; + unsigned long last_stamp = 0; + struct rb_node *p; + struct list_head slow_osds; + + dout("timeout\n"); + down_read(&osdc->map_sem); + + ceph_monc_request_next_osdmap(&osdc->client->monc); + + mutex_lock(&osdc->request_mutex); + for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { + req = rb_entry(p, struct ceph_osd_request, r_node); + + if (req->r_resend) { + int err; + + dout("osdc resending prev failed %lld\n", req->r_tid); + err = __send_request(osdc, req); + if (err) + dout("osdc failed again on %lld\n", req->r_tid); + else + req->r_resend = false; + continue; + } + } + + /* + * reset osds that appear to be _really_ unresponsive. this + * is a failsafe measure.. we really shouldn't be getting to + * this point if the system is working properly. the monitors + * should mark the osd as failed and we should find out about + * it from an updated osd map. + */ + while (timeout && !list_empty(&osdc->req_lru)) { + req = list_entry(osdc->req_lru.next, struct ceph_osd_request, + r_req_lru_item); + + if (time_before(jiffies, req->r_stamp + timeout)) + break; + + BUG_ON(req == last_req && req->r_stamp == last_stamp); + last_req = req; + last_stamp = req->r_stamp; + + osd = req->r_osd; + BUG_ON(!osd); + pr_warning(" tid %llu timed out on osd%d, will reset osd\n", + req->r_tid, osd->o_osd); + __kick_requests(osdc, osd); + } + + /* + * ping osds that are a bit slow. this ensures that if there + * is a break in the TCP connection we will notice, and reopen + * a connection with that osd (from the fault callback). + */ + INIT_LIST_HEAD(&slow_osds); + list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) { + if (time_before(jiffies, req->r_stamp + keepalive)) + break; + + osd = req->r_osd; + BUG_ON(!osd); + dout(" tid %llu is slow, will send keepalive on osd%d\n", + req->r_tid, osd->o_osd); + list_move_tail(&osd->o_keepalive_item, &slow_osds); + } + while (!list_empty(&slow_osds)) { + osd = list_entry(slow_osds.next, struct ceph_osd, + o_keepalive_item); + list_del_init(&osd->o_keepalive_item); + ceph_con_keepalive(&osd->o_con); + } + + __schedule_osd_timeout(osdc); + mutex_unlock(&osdc->request_mutex); + + up_read(&osdc->map_sem); +} + +static void handle_osds_timeout(struct work_struct *work) +{ + struct ceph_osd_client *osdc = + container_of(work, struct ceph_osd_client, + osds_timeout_work.work); + unsigned long delay = + osdc->client->options->osd_idle_ttl * HZ >> 2; + + dout("osds timeout\n"); + down_read(&osdc->map_sem); + remove_old_osds(osdc, 0); + up_read(&osdc->map_sem); + + schedule_delayed_work(&osdc->osds_timeout_work, + round_jiffies_relative(delay)); +} + +/* + * handle osd op reply. either call the callback if it is specified, + * or do the completion to wake up the waiting thread. + */ +static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg, + struct ceph_connection *con) +{ + struct ceph_osd_reply_head *rhead = msg->front.iov_base; + struct ceph_osd_request *req; + u64 tid; + int numops, object_len, flags; + s32 result; + + tid = le64_to_cpu(msg->hdr.tid); + if (msg->front.iov_len < sizeof(*rhead)) + goto bad; + numops = le32_to_cpu(rhead->num_ops); + object_len = le32_to_cpu(rhead->object_len); + result = le32_to_cpu(rhead->result); + if (msg->front.iov_len != sizeof(*rhead) + object_len + + numops * sizeof(struct ceph_osd_op)) + goto bad; + dout("handle_reply %p tid %llu result %d\n", msg, tid, (int)result); + + /* lookup */ + mutex_lock(&osdc->request_mutex); + req = __lookup_request(osdc, tid); + if (req == NULL) { + dout("handle_reply tid %llu dne\n", tid); + mutex_unlock(&osdc->request_mutex); + return; + } + ceph_osdc_get_request(req); + flags = le32_to_cpu(rhead->flags); + + /* + * if this connection filled our message, drop our reference now, to + * avoid a (safe but slower) revoke later. + */ + if (req->r_con_filling_msg == con && req->r_reply == msg) { + dout(" dropping con_filling_msg ref %p\n", con); + req->r_con_filling_msg = NULL; + ceph_con_put(con); + } + + if (!req->r_got_reply) { + unsigned bytes; + + req->r_result = le32_to_cpu(rhead->result); + bytes = le32_to_cpu(msg->hdr.data_len); + dout("handle_reply result %d bytes %d\n", req->r_result, + bytes); + if (req->r_result == 0) + req->r_result = bytes; + + /* in case this is a write and we need to replay, */ + req->r_reassert_version = rhead->reassert_version; + + req->r_got_reply = 1; + } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) { + dout("handle_reply tid %llu dup ack\n", tid); + mutex_unlock(&osdc->request_mutex); + goto done; + } + + dout("handle_reply tid %llu flags %d\n", tid, flags); + + /* either this is a read, or we got the safe response */ + if (result < 0 || + (flags & CEPH_OSD_FLAG_ONDISK) || + ((flags & CEPH_OSD_FLAG_WRITE) == 0)) + __unregister_request(osdc, req); + + mutex_unlock(&osdc->request_mutex); + + if (req->r_callback) + req->r_callback(req, msg); + else + complete_all(&req->r_completion); + + if (flags & CEPH_OSD_FLAG_ONDISK) { + if (req->r_safe_callback) + req->r_safe_callback(req, msg); + complete_all(&req->r_safe_completion); /* fsync waiter */ + } + +done: + ceph_osdc_put_request(req); + return; + +bad: + pr_err("corrupt osd_op_reply got %d %d expected %d\n", + (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len), + (int)sizeof(*rhead)); + ceph_msg_dump(msg); +} + + +static int __kick_requests(struct ceph_osd_client *osdc, + struct ceph_osd *kickosd) +{ + struct ceph_osd_request *req; + struct rb_node *p, *n; + int needmap = 0; + int err; + + dout("kick_requests osd%d\n", kickosd ? kickosd->o_osd : -1); + if (kickosd) { + err = __reset_osd(osdc, kickosd); + if (err == -EAGAIN) + return 1; + } else { + for (p = rb_first(&osdc->osds); p; p = n) { + struct ceph_osd *osd = + rb_entry(p, struct ceph_osd, o_node); + + n = rb_next(p); + if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) || + memcmp(&osd->o_con.peer_addr, + ceph_osd_addr(osdc->osdmap, + osd->o_osd), + sizeof(struct ceph_entity_addr)) != 0) + __reset_osd(osdc, osd); + } + } + + for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { + req = rb_entry(p, struct ceph_osd_request, r_node); + + if (req->r_resend) { + dout(" r_resend set on tid %llu\n", req->r_tid); + __cancel_request(req); + goto kick; + } + if (req->r_osd && kickosd == req->r_osd) { + __cancel_request(req); + goto kick; + } + + err = __map_osds(osdc, req); + if (err == 0) + continue; /* no change */ + if (err < 0) { + /* + * FIXME: really, we should set the request + * error and fail if this isn't a 'nofail' + * request, but that's a fair bit more + * complicated to do. So retry! + */ + dout(" setting r_resend on %llu\n", req->r_tid); + req->r_resend = true; + continue; + } + if (req->r_osd == NULL) { + dout("tid %llu maps to no valid osd\n", req->r_tid); + needmap++; /* request a newer map */ + continue; + } + +kick: + dout("kicking %p tid %llu osd%d\n", req, req->r_tid, + req->r_osd ? req->r_osd->o_osd : -1); + req->r_flags |= CEPH_OSD_FLAG_RETRY; + err = __send_request(osdc, req); + if (err) { + dout(" setting r_resend on %llu\n", req->r_tid); + req->r_resend = true; + } + } + + return needmap; +} + +/* + * Resubmit osd requests whose osd or osd address has changed. Request + * a new osd map if osds are down, or we are otherwise unable to determine + * how to direct a request. + * + * Close connections to down osds. + * + * If @who is specified, resubmit requests for that specific osd. + * + * Caller should hold map_sem for read and request_mutex. + */ +static void kick_requests(struct ceph_osd_client *osdc, + struct ceph_osd *kickosd) +{ + int needmap; + + mutex_lock(&osdc->request_mutex); + needmap = __kick_requests(osdc, kickosd); + mutex_unlock(&osdc->request_mutex); + + if (needmap) { + dout("%d requests for down osds, need new map\n", needmap); + ceph_monc_request_next_osdmap(&osdc->client->monc); + } + +} +/* + * Process updated osd map. + * + * The message contains any number of incremental and full maps, normally + * indicating some sort of topology change in the cluster. Kick requests + * off to different OSDs as needed. + */ +void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg) +{ + void *p, *end, *next; + u32 nr_maps, maplen; + u32 epoch; + struct ceph_osdmap *newmap = NULL, *oldmap; + int err; + struct ceph_fsid fsid; + + dout("handle_map have %u\n", osdc->osdmap ? osdc->osdmap->epoch : 0); + p = msg->front.iov_base; + end = p + msg->front.iov_len; + + /* verify fsid */ + ceph_decode_need(&p, end, sizeof(fsid), bad); + ceph_decode_copy(&p, &fsid, sizeof(fsid)); + if (ceph_check_fsid(osdc->client, &fsid) < 0) + return; + + down_write(&osdc->map_sem); + + /* incremental maps */ + ceph_decode_32_safe(&p, end, nr_maps, bad); + dout(" %d inc maps\n", nr_maps); + while (nr_maps > 0) { + ceph_decode_need(&p, end, 2*sizeof(u32), bad); + epoch = ceph_decode_32(&p); + maplen = ceph_decode_32(&p); + ceph_decode_need(&p, end, maplen, bad); + next = p + maplen; + if (osdc->osdmap && osdc->osdmap->epoch+1 == epoch) { + dout("applying incremental map %u len %d\n", + epoch, maplen); + newmap = osdmap_apply_incremental(&p, next, + osdc->osdmap, + osdc->client->msgr); + if (IS_ERR(newmap)) { + err = PTR_ERR(newmap); + goto bad; + } + BUG_ON(!newmap); + if (newmap != osdc->osdmap) { + ceph_osdmap_destroy(osdc->osdmap); + osdc->osdmap = newmap; + } + } else { + dout("ignoring incremental map %u len %d\n", + epoch, maplen); + } + p = next; + nr_maps--; + } + if (newmap) + goto done; + + /* full maps */ + ceph_decode_32_safe(&p, end, nr_maps, bad); + dout(" %d full maps\n", nr_maps); + while (nr_maps) { + ceph_decode_need(&p, end, 2*sizeof(u32), bad); + epoch = ceph_decode_32(&p); + maplen = ceph_decode_32(&p); + ceph_decode_need(&p, end, maplen, bad); + if (nr_maps > 1) { + dout("skipping non-latest full map %u len %d\n", + epoch, maplen); + } else if (osdc->osdmap && osdc->osdmap->epoch >= epoch) { + dout("skipping full map %u len %d, " + "older than our %u\n", epoch, maplen, + osdc->osdmap->epoch); + } else { + dout("taking full map %u len %d\n", epoch, maplen); + newmap = osdmap_decode(&p, p+maplen); + if (IS_ERR(newmap)) { + err = PTR_ERR(newmap); + goto bad; + } + BUG_ON(!newmap); + oldmap = osdc->osdmap; + osdc->osdmap = newmap; + if (oldmap) + ceph_osdmap_destroy(oldmap); + } + p += maplen; + nr_maps--; + } + +done: + downgrade_write(&osdc->map_sem); + ceph_monc_got_osdmap(&osdc->client->monc, osdc->osdmap->epoch); + if (newmap) + kick_requests(osdc, NULL); + up_read(&osdc->map_sem); + wake_up_all(&osdc->client->auth_wq); + return; + +bad: + pr_err("osdc handle_map corrupt msg\n"); + ceph_msg_dump(msg); + up_write(&osdc->map_sem); + return; +} + +/* + * Register request, send initial attempt. + */ +int ceph_osdc_start_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req, + bool nofail) +{ + int rc = 0; + + req->r_request->pages = req->r_pages; + req->r_request->nr_pages = req->r_num_pages; +#ifdef CONFIG_BLOCK + req->r_request->bio = req->r_bio; +#endif + req->r_request->trail = req->r_trail; + + register_request(osdc, req); + + down_read(&osdc->map_sem); + mutex_lock(&osdc->request_mutex); + /* + * a racing kick_requests() may have sent the message for us + * while we dropped request_mutex above, so only send now if + * the request still han't been touched yet. + */ + if (req->r_sent == 0) { + rc = __send_request(osdc, req); + if (rc) { + if (nofail) { + dout("osdc_start_request failed send, " + " marking %lld\n", req->r_tid); + req->r_resend = true; + rc = 0; + } else { + __unregister_request(osdc, req); + } + } + } + mutex_unlock(&osdc->request_mutex); + up_read(&osdc->map_sem); + return rc; +} +EXPORT_SYMBOL(ceph_osdc_start_request); + +/* + * wait for a request to complete + */ +int ceph_osdc_wait_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + int rc; + + rc = wait_for_completion_interruptible(&req->r_completion); + if (rc < 0) { + mutex_lock(&osdc->request_mutex); + __cancel_request(req); + __unregister_request(osdc, req); + mutex_unlock(&osdc->request_mutex); + dout("wait_request tid %llu canceled/timed out\n", req->r_tid); + return rc; + } + + dout("wait_request tid %llu result %d\n", req->r_tid, req->r_result); + return req->r_result; +} +EXPORT_SYMBOL(ceph_osdc_wait_request); + +/* + * sync - wait for all in-flight requests to flush. avoid starvation. + */ +void ceph_osdc_sync(struct ceph_osd_client *osdc) +{ + struct ceph_osd_request *req; + u64 last_tid, next_tid = 0; + + mutex_lock(&osdc->request_mutex); + last_tid = osdc->last_tid; + while (1) { + req = __lookup_request_ge(osdc, next_tid); + if (!req) + break; + if (req->r_tid > last_tid) + break; + + next_tid = req->r_tid + 1; + if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0) + continue; + + ceph_osdc_get_request(req); + mutex_unlock(&osdc->request_mutex); + dout("sync waiting on tid %llu (last is %llu)\n", + req->r_tid, last_tid); + wait_for_completion(&req->r_safe_completion); + mutex_lock(&osdc->request_mutex); + ceph_osdc_put_request(req); + } + mutex_unlock(&osdc->request_mutex); + dout("sync done (thru tid %llu)\n", last_tid); +} +EXPORT_SYMBOL(ceph_osdc_sync); + +/* + * init, shutdown + */ +int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client) +{ + int err; + + dout("init\n"); + osdc->client = client; + osdc->osdmap = NULL; + init_rwsem(&osdc->map_sem); + init_completion(&osdc->map_waiters); + osdc->last_requested_map = 0; + mutex_init(&osdc->request_mutex); + osdc->last_tid = 0; + osdc->osds = RB_ROOT; + INIT_LIST_HEAD(&osdc->osd_lru); + osdc->requests = RB_ROOT; + INIT_LIST_HEAD(&osdc->req_lru); + osdc->num_requests = 0; + INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout); + INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout); + + schedule_delayed_work(&osdc->osds_timeout_work, + round_jiffies_relative(osdc->client->options->osd_idle_ttl * HZ)); + + err = -ENOMEM; + osdc->req_mempool = mempool_create_kmalloc_pool(10, + sizeof(struct ceph_osd_request)); + if (!osdc->req_mempool) + goto out; + + err = ceph_msgpool_init(&osdc->msgpool_op, OSD_OP_FRONT_LEN, 10, true, + "osd_op"); + if (err < 0) + goto out_mempool; + err = ceph_msgpool_init(&osdc->msgpool_op_reply, + OSD_OPREPLY_FRONT_LEN, 10, true, + "osd_op_reply"); + if (err < 0) + goto out_msgpool; + return 0; + +out_msgpool: + ceph_msgpool_destroy(&osdc->msgpool_op); +out_mempool: + mempool_destroy(osdc->req_mempool); +out: + return err; +} +EXPORT_SYMBOL(ceph_osdc_init); + +void ceph_osdc_stop(struct ceph_osd_client *osdc) +{ + cancel_delayed_work_sync(&osdc->timeout_work); + cancel_delayed_work_sync(&osdc->osds_timeout_work); + if (osdc->osdmap) { + ceph_osdmap_destroy(osdc->osdmap); + osdc->osdmap = NULL; + } + remove_old_osds(osdc, 1); + mempool_destroy(osdc->req_mempool); + ceph_msgpool_destroy(&osdc->msgpool_op); + ceph_msgpool_destroy(&osdc->msgpool_op_reply); +} +EXPORT_SYMBOL(ceph_osdc_stop); + +/* + * Read some contiguous pages. If we cross a stripe boundary, shorten + * *plen. Return number of bytes read, or error. + */ +int ceph_osdc_readpages(struct ceph_osd_client *osdc, + struct ceph_vino vino, struct ceph_file_layout *layout, + u64 off, u64 *plen, + u32 truncate_seq, u64 truncate_size, + struct page **pages, int num_pages) +{ + struct ceph_osd_request *req; + int rc = 0; + + dout("readpages on ino %llx.%llx on %llu~%llu\n", vino.ino, + vino.snap, off, *plen); + req = ceph_osdc_new_request(osdc, layout, vino, off, plen, + CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ, + NULL, 0, truncate_seq, truncate_size, NULL, + false, 1); + if (!req) + return -ENOMEM; + + /* it may be a short read due to an object boundary */ + req->r_pages = pages; + + dout("readpages final extent is %llu~%llu (%d pages)\n", + off, *plen, req->r_num_pages); + + rc = ceph_osdc_start_request(osdc, req, false); + if (!rc) + rc = ceph_osdc_wait_request(osdc, req); + + ceph_osdc_put_request(req); + dout("readpages result %d\n", rc); + return rc; +} +EXPORT_SYMBOL(ceph_osdc_readpages); + +/* + * do a synchronous write on N pages + */ +int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino, + struct ceph_file_layout *layout, + struct ceph_snap_context *snapc, + u64 off, u64 len, + u32 truncate_seq, u64 truncate_size, + struct timespec *mtime, + struct page **pages, int num_pages, + int flags, int do_sync, bool nofail) +{ + struct ceph_osd_request *req; + int rc = 0; + + BUG_ON(vino.snap != CEPH_NOSNAP); + req = ceph_osdc_new_request(osdc, layout, vino, off, &len, + CEPH_OSD_OP_WRITE, + flags | CEPH_OSD_FLAG_ONDISK | + CEPH_OSD_FLAG_WRITE, + snapc, do_sync, + truncate_seq, truncate_size, mtime, + nofail, 1); + if (!req) + return -ENOMEM; + + /* it may be a short write due to an object boundary */ + req->r_pages = pages; + dout("writepages %llu~%llu (%d pages)\n", off, len, + req->r_num_pages); + + rc = ceph_osdc_start_request(osdc, req, nofail); + if (!rc) + rc = ceph_osdc_wait_request(osdc, req); + + ceph_osdc_put_request(req); + if (rc == 0) + rc = len; + dout("writepages result %d\n", rc); + return rc; +} +EXPORT_SYMBOL(ceph_osdc_writepages); + +/* + * handle incoming message + */ +static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) +{ + struct ceph_osd *osd = con->private; + struct ceph_osd_client *osdc; + int type = le16_to_cpu(msg->hdr.type); + + if (!osd) + goto out; + osdc = osd->o_osdc; + + switch (type) { + case CEPH_MSG_OSD_MAP: + ceph_osdc_handle_map(osdc, msg); + break; + case CEPH_MSG_OSD_OPREPLY: + handle_reply(osdc, msg, con); + break; + + default: + pr_err("received unknown message type %d %s\n", type, + ceph_msg_type_name(type)); + } +out: + ceph_msg_put(msg); +} + +/* + * lookup and return message for incoming reply. set up reply message + * pages. + */ +static struct ceph_msg *get_reply(struct ceph_connection *con, + struct ceph_msg_header *hdr, + int *skip) +{ + struct ceph_osd *osd = con->private; + struct ceph_osd_client *osdc = osd->o_osdc; + struct ceph_msg *m; + struct ceph_osd_request *req; + int front = le32_to_cpu(hdr->front_len); + int data_len = le32_to_cpu(hdr->data_len); + u64 tid; + + tid = le64_to_cpu(hdr->tid); + mutex_lock(&osdc->request_mutex); + req = __lookup_request(osdc, tid); + if (!req) { + *skip = 1; + m = NULL; + pr_info("get_reply unknown tid %llu from osd%d\n", tid, + osd->o_osd); + goto out; + } + + if (req->r_con_filling_msg) { + dout("get_reply revoking msg %p from old con %p\n", + req->r_reply, req->r_con_filling_msg); + ceph_con_revoke_message(req->r_con_filling_msg, req->r_reply); + ceph_con_put(req->r_con_filling_msg); + req->r_con_filling_msg = NULL; + } + + if (front > req->r_reply->front.iov_len) { + pr_warning("get_reply front %d > preallocated %d\n", + front, (int)req->r_reply->front.iov_len); + m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS); + if (!m) + goto out; + ceph_msg_put(req->r_reply); + req->r_reply = m; + } + m = ceph_msg_get(req->r_reply); + + if (data_len > 0) { + unsigned data_off = le16_to_cpu(hdr->data_off); + int want = calc_pages_for(data_off & ~PAGE_MASK, data_len); + + if (unlikely(req->r_num_pages < want)) { + pr_warning("tid %lld reply %d > expected %d pages\n", + tid, want, m->nr_pages); + *skip = 1; + ceph_msg_put(m); + m = NULL; + goto out; + } + m->pages = req->r_pages; + m->nr_pages = req->r_num_pages; +#ifdef CONFIG_BLOCK + m->bio = req->r_bio; +#endif + } + *skip = 0; + req->r_con_filling_msg = ceph_con_get(con); + dout("get_reply tid %lld %p\n", tid, m); + +out: + mutex_unlock(&osdc->request_mutex); + return m; + +} + +static struct ceph_msg *alloc_msg(struct ceph_connection *con, + struct ceph_msg_header *hdr, + int *skip) +{ + struct ceph_osd *osd = con->private; + int type = le16_to_cpu(hdr->type); + int front = le32_to_cpu(hdr->front_len); + + switch (type) { + case CEPH_MSG_OSD_MAP: + return ceph_msg_new(type, front, GFP_NOFS); + case CEPH_MSG_OSD_OPREPLY: + return get_reply(con, hdr, skip); + default: + pr_info("alloc_msg unexpected msg type %d from osd%d\n", type, + osd->o_osd); + *skip = 1; + return NULL; + } +} + +/* + * Wrappers to refcount containing ceph_osd struct + */ +static struct ceph_connection *get_osd_con(struct ceph_connection *con) +{ + struct ceph_osd *osd = con->private; + if (get_osd(osd)) + return con; + return NULL; +} + +static void put_osd_con(struct ceph_connection *con) +{ + struct ceph_osd *osd = con->private; + put_osd(osd); +} + +/* + * authentication + */ +static int get_authorizer(struct ceph_connection *con, + void **buf, int *len, int *proto, + void **reply_buf, int *reply_len, int force_new) +{ + struct ceph_osd *o = con->private; + struct ceph_osd_client *osdc = o->o_osdc; + struct ceph_auth_client *ac = osdc->client->monc.auth; + int ret = 0; + + if (force_new && o->o_authorizer) { + ac->ops->destroy_authorizer(ac, o->o_authorizer); + o->o_authorizer = NULL; + } + if (o->o_authorizer == NULL) { + ret = ac->ops->create_authorizer( + ac, CEPH_ENTITY_TYPE_OSD, + &o->o_authorizer, + &o->o_authorizer_buf, + &o->o_authorizer_buf_len, + &o->o_authorizer_reply_buf, + &o->o_authorizer_reply_buf_len); + if (ret) + return ret; + } + + *proto = ac->protocol; + *buf = o->o_authorizer_buf; + *len = o->o_authorizer_buf_len; + *reply_buf = o->o_authorizer_reply_buf; + *reply_len = o->o_authorizer_reply_buf_len; + return 0; +} + + +static int verify_authorizer_reply(struct ceph_connection *con, int len) +{ + struct ceph_osd *o = con->private; + struct ceph_osd_client *osdc = o->o_osdc; + struct ceph_auth_client *ac = osdc->client->monc.auth; + + return ac->ops->verify_authorizer_reply(ac, o->o_authorizer, len); +} + +static int invalidate_authorizer(struct ceph_connection *con) +{ + struct ceph_osd *o = con->private; + struct ceph_osd_client *osdc = o->o_osdc; + struct ceph_auth_client *ac = osdc->client->monc.auth; + + if (ac->ops->invalidate_authorizer) + ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_OSD); + + return ceph_monc_validate_auth(&osdc->client->monc); +} + +static const struct ceph_connection_operations osd_con_ops = { + .get = get_osd_con, + .put = put_osd_con, + .dispatch = dispatch, + .get_authorizer = get_authorizer, + .verify_authorizer_reply = verify_authorizer_reply, + .invalidate_authorizer = invalidate_authorizer, + .alloc_msg = alloc_msg, + .fault = osd_reset, +}; diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c new file mode 100644 index 0000000..d73f3f6 --- /dev/null +++ b/net/ceph/osdmap.c @@ -0,0 +1,1128 @@ + +#include + +#include +#include +#include + +#include +#include +#include +#include +#include + +char *ceph_osdmap_state_str(char *str, int len, int state) +{ + int flag = 0; + + if (!len) + goto done; + + *str = '\0'; + if (state) { + if (state & CEPH_OSD_EXISTS) { + snprintf(str, len, "exists"); + flag = 1; + } + if (state & CEPH_OSD_UP) { + snprintf(str, len, "%s%s%s", str, (flag ? ", " : ""), + "up"); + flag = 1; + } + } else { + snprintf(str, len, "doesn't exist"); + } +done: + return str; +} + +/* maps */ + +static int calc_bits_of(unsigned t) +{ + int b = 0; + while (t) { + t = t >> 1; + b++; + } + return b; +} + +/* + * the foo_mask is the smallest value 2^n-1 that is >= foo. + */ +static void calc_pg_masks(struct ceph_pg_pool_info *pi) +{ + pi->pg_num_mask = (1 << calc_bits_of(le32_to_cpu(pi->v.pg_num)-1)) - 1; + pi->pgp_num_mask = + (1 << calc_bits_of(le32_to_cpu(pi->v.pgp_num)-1)) - 1; + pi->lpg_num_mask = + (1 << calc_bits_of(le32_to_cpu(pi->v.lpg_num)-1)) - 1; + pi->lpgp_num_mask = + (1 << calc_bits_of(le32_to_cpu(pi->v.lpgp_num)-1)) - 1; +} + +/* + * decode crush map + */ +static int crush_decode_uniform_bucket(void **p, void *end, + struct crush_bucket_uniform *b) +{ + dout("crush_decode_uniform_bucket %p to %p\n", *p, end); + ceph_decode_need(p, end, (1+b->h.size) * sizeof(u32), bad); + b->item_weight = ceph_decode_32(p); + return 0; +bad: + return -EINVAL; +} + +static int crush_decode_list_bucket(void **p, void *end, + struct crush_bucket_list *b) +{ + int j; + dout("crush_decode_list_bucket %p to %p\n", *p, end); + b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); + if (b->item_weights == NULL) + return -ENOMEM; + b->sum_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); + if (b->sum_weights == NULL) + return -ENOMEM; + ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad); + for (j = 0; j < b->h.size; j++) { + b->item_weights[j] = ceph_decode_32(p); + b->sum_weights[j] = ceph_decode_32(p); + } + return 0; +bad: + return -EINVAL; +} + +static int crush_decode_tree_bucket(void **p, void *end, + struct crush_bucket_tree *b) +{ + int j; + dout("crush_decode_tree_bucket %p to %p\n", *p, end); + ceph_decode_32_safe(p, end, b->num_nodes, bad); + b->node_weights = kcalloc(b->num_nodes, sizeof(u32), GFP_NOFS); + if (b->node_weights == NULL) + return -ENOMEM; + ceph_decode_need(p, end, b->num_nodes * sizeof(u32), bad); + for (j = 0; j < b->num_nodes; j++) + b->node_weights[j] = ceph_decode_32(p); + return 0; +bad: + return -EINVAL; +} + +static int crush_decode_straw_bucket(void **p, void *end, + struct crush_bucket_straw *b) +{ + int j; + dout("crush_decode_straw_bucket %p to %p\n", *p, end); + b->item_weights = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); + if (b->item_weights == NULL) + return -ENOMEM; + b->straws = kcalloc(b->h.size, sizeof(u32), GFP_NOFS); + if (b->straws == NULL) + return -ENOMEM; + ceph_decode_need(p, end, 2 * b->h.size * sizeof(u32), bad); + for (j = 0; j < b->h.size; j++) { + b->item_weights[j] = ceph_decode_32(p); + b->straws[j] = ceph_decode_32(p); + } + return 0; +bad: + return -EINVAL; +} + +static struct crush_map *crush_decode(void *pbyval, void *end) +{ + struct crush_map *c; + int err = -EINVAL; + int i, j; + void **p = &pbyval; + void *start = pbyval; + u32 magic; + + dout("crush_decode %p to %p len %d\n", *p, end, (int)(end - *p)); + + c = kzalloc(sizeof(*c), GFP_NOFS); + if (c == NULL) + return ERR_PTR(-ENOMEM); + + ceph_decode_need(p, end, 4*sizeof(u32), bad); + magic = ceph_decode_32(p); + if (magic != CRUSH_MAGIC) { + pr_err("crush_decode magic %x != current %x\n", + (unsigned)magic, (unsigned)CRUSH_MAGIC); + goto bad; + } + c->max_buckets = ceph_decode_32(p); + c->max_rules = ceph_decode_32(p); + c->max_devices = ceph_decode_32(p); + + c->device_parents = kcalloc(c->max_devices, sizeof(u32), GFP_NOFS); + if (c->device_parents == NULL) + goto badmem; + c->bucket_parents = kcalloc(c->max_buckets, sizeof(u32), GFP_NOFS); + if (c->bucket_parents == NULL) + goto badmem; + + c->buckets = kcalloc(c->max_buckets, sizeof(*c->buckets), GFP_NOFS); + if (c->buckets == NULL) + goto badmem; + c->rules = kcalloc(c->max_rules, sizeof(*c->rules), GFP_NOFS); + if (c->rules == NULL) + goto badmem; + + /* buckets */ + for (i = 0; i < c->max_buckets; i++) { + int size = 0; + u32 alg; + struct crush_bucket *b; + + ceph_decode_32_safe(p, end, alg, bad); + if (alg == 0) { + c->buckets[i] = NULL; + continue; + } + dout("crush_decode bucket %d off %x %p to %p\n", + i, (int)(*p-start), *p, end); + + switch (alg) { + case CRUSH_BUCKET_UNIFORM: + size = sizeof(struct crush_bucket_uniform); + break; + case CRUSH_BUCKET_LIST: + size = sizeof(struct crush_bucket_list); + break; + case CRUSH_BUCKET_TREE: + size = sizeof(struct crush_bucket_tree); + break; + case CRUSH_BUCKET_STRAW: + size = sizeof(struct crush_bucket_straw); + break; + default: + err = -EINVAL; + goto bad; + } + BUG_ON(size == 0); + b = c->buckets[i] = kzalloc(size, GFP_NOFS); + if (b == NULL) + goto badmem; + + ceph_decode_need(p, end, 4*sizeof(u32), bad); + b->id = ceph_decode_32(p); + b->type = ceph_decode_16(p); + b->alg = ceph_decode_8(p); + b->hash = ceph_decode_8(p); + b->weight = ceph_decode_32(p); + b->size = ceph_decode_32(p); + + dout("crush_decode bucket size %d off %x %p to %p\n", + b->size, (int)(*p-start), *p, end); + + b->items = kcalloc(b->size, sizeof(__s32), GFP_NOFS); + if (b->items == NULL) + goto badmem; + b->perm = kcalloc(b->size, sizeof(u32), GFP_NOFS); + if (b->perm == NULL) + goto badmem; + b->perm_n = 0; + + ceph_decode_need(p, end, b->size*sizeof(u32), bad); + for (j = 0; j < b->size; j++) + b->items[j] = ceph_decode_32(p); + + switch (b->alg) { + case CRUSH_BUCKET_UNIFORM: + err = crush_decode_uniform_bucket(p, end, + (struct crush_bucket_uniform *)b); + if (err < 0) + goto bad; + break; + case CRUSH_BUCKET_LIST: + err = crush_decode_list_bucket(p, end, + (struct crush_bucket_list *)b); + if (err < 0) + goto bad; + break; + case CRUSH_BUCKET_TREE: + err = crush_decode_tree_bucket(p, end, + (struct crush_bucket_tree *)b); + if (err < 0) + goto bad; + break; + case CRUSH_BUCKET_STRAW: + err = crush_decode_straw_bucket(p, end, + (struct crush_bucket_straw *)b); + if (err < 0) + goto bad; + break; + } + } + + /* rules */ + dout("rule vec is %p\n", c->rules); + for (i = 0; i < c->max_rules; i++) { + u32 yes; + struct crush_rule *r; + + ceph_decode_32_safe(p, end, yes, bad); + if (!yes) { + dout("crush_decode NO rule %d off %x %p to %p\n", + i, (int)(*p-start), *p, end); + c->rules[i] = NULL; + continue; + } + + dout("crush_decode rule %d off %x %p to %p\n", + i, (int)(*p-start), *p, end); + + /* len */ + ceph_decode_32_safe(p, end, yes, bad); +#if BITS_PER_LONG == 32 + err = -EINVAL; + if (yes > ULONG_MAX / sizeof(struct crush_rule_step)) + goto bad; +#endif + r = c->rules[i] = kmalloc(sizeof(*r) + + yes*sizeof(struct crush_rule_step), + GFP_NOFS); + if (r == NULL) + goto badmem; + dout(" rule %d is at %p\n", i, r); + r->len = yes; + ceph_decode_copy_safe(p, end, &r->mask, 4, bad); /* 4 u8's */ + ceph_decode_need(p, end, r->len*3*sizeof(u32), bad); + for (j = 0; j < r->len; j++) { + r->steps[j].op = ceph_decode_32(p); + r->steps[j].arg1 = ceph_decode_32(p); + r->steps[j].arg2 = ceph_decode_32(p); + } + } + + /* ignore trailing name maps. */ + + dout("crush_decode success\n"); + return c; + +badmem: + err = -ENOMEM; +bad: + dout("crush_decode fail %d\n", err); + crush_destroy(c); + return ERR_PTR(err); +} + +/* + * rbtree of pg_mapping for handling pg_temp (explicit mapping of pgid + * to a set of osds) + */ +static int pgid_cmp(struct ceph_pg l, struct ceph_pg r) +{ + u64 a = *(u64 *)&l; + u64 b = *(u64 *)&r; + + if (a < b) + return -1; + if (a > b) + return 1; + return 0; +} + +static int __insert_pg_mapping(struct ceph_pg_mapping *new, + struct rb_root *root) +{ + struct rb_node **p = &root->rb_node; + struct rb_node *parent = NULL; + struct ceph_pg_mapping *pg = NULL; + int c; + + while (*p) { + parent = *p; + pg = rb_entry(parent, struct ceph_pg_mapping, node); + c = pgid_cmp(new->pgid, pg->pgid); + if (c < 0) + p = &(*p)->rb_left; + else if (c > 0) + p = &(*p)->rb_right; + else + return -EEXIST; + } + + rb_link_node(&new->node, parent, p); + rb_insert_color(&new->node, root); + return 0; +} + +static struct ceph_pg_mapping *__lookup_pg_mapping(struct rb_root *root, + struct ceph_pg pgid) +{ + struct rb_node *n = root->rb_node; + struct ceph_pg_mapping *pg; + int c; + + while (n) { + pg = rb_entry(n, struct ceph_pg_mapping, node); + c = pgid_cmp(pgid, pg->pgid); + if (c < 0) + n = n->rb_left; + else if (c > 0) + n = n->rb_right; + else + return pg; + } + return NULL; +} + +/* + * rbtree of pg pool info + */ +static int __insert_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *new) +{ + struct rb_node **p = &root->rb_node; + struct rb_node *parent = NULL; + struct ceph_pg_pool_info *pi = NULL; + + while (*p) { + parent = *p; + pi = rb_entry(parent, struct ceph_pg_pool_info, node); + if (new->id < pi->id) + p = &(*p)->rb_left; + else if (new->id > pi->id) + p = &(*p)->rb_right; + else + return -EEXIST; + } + + rb_link_node(&new->node, parent, p); + rb_insert_color(&new->node, root); + return 0; +} + +static struct ceph_pg_pool_info *__lookup_pg_pool(struct rb_root *root, int id) +{ + struct ceph_pg_pool_info *pi; + struct rb_node *n = root->rb_node; + + while (n) { + pi = rb_entry(n, struct ceph_pg_pool_info, node); + if (id < pi->id) + n = n->rb_left; + else if (id > pi->id) + n = n->rb_right; + else + return pi; + } + return NULL; +} + +int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name) +{ + struct rb_node *rbp; + + for (rbp = rb_first(&map->pg_pools); rbp; rbp = rb_next(rbp)) { + struct ceph_pg_pool_info *pi = + rb_entry(rbp, struct ceph_pg_pool_info, node); + if (pi->name && strcmp(pi->name, name) == 0) + return pi->id; + } + return -ENOENT; +} +EXPORT_SYMBOL(ceph_pg_poolid_by_name); + +static void __remove_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *pi) +{ + rb_erase(&pi->node, root); + kfree(pi->name); + kfree(pi); +} + +static int __decode_pool(void **p, void *end, struct ceph_pg_pool_info *pi) +{ + unsigned n, m; + + ceph_decode_copy(p, &pi->v, sizeof(pi->v)); + calc_pg_masks(pi); + + /* num_snaps * snap_info_t */ + n = le32_to_cpu(pi->v.num_snaps); + while (n--) { + ceph_decode_need(p, end, sizeof(u64) + 1 + sizeof(u64) + + sizeof(struct ceph_timespec), bad); + *p += sizeof(u64) + /* key */ + 1 + sizeof(u64) + /* u8, snapid */ + sizeof(struct ceph_timespec); + m = ceph_decode_32(p); /* snap name */ + *p += m; + } + + *p += le32_to_cpu(pi->v.num_removed_snap_intervals) * sizeof(u64) * 2; + return 0; + +bad: + return -EINVAL; +} + +static int __decode_pool_names(void **p, void *end, struct ceph_osdmap *map) +{ + struct ceph_pg_pool_info *pi; + u32 num, len, pool; + + ceph_decode_32_safe(p, end, num, bad); + dout(" %d pool names\n", num); + while (num--) { + ceph_decode_32_safe(p, end, pool, bad); + ceph_decode_32_safe(p, end, len, bad); + dout(" pool %d len %d\n", pool, len); + pi = __lookup_pg_pool(&map->pg_pools, pool); + if (pi) { + kfree(pi->name); + pi->name = kmalloc(len + 1, GFP_NOFS); + if (pi->name) { + memcpy(pi->name, *p, len); + pi->name[len] = '\0'; + dout(" name is %s\n", pi->name); + } + } + *p += len; + } + return 0; + +bad: + return -EINVAL; +} + +/* + * osd map + */ +void ceph_osdmap_destroy(struct ceph_osdmap *map) +{ + dout("osdmap_destroy %p\n", map); + if (map->crush) + crush_destroy(map->crush); + while (!RB_EMPTY_ROOT(&map->pg_temp)) { + struct ceph_pg_mapping *pg = + rb_entry(rb_first(&map->pg_temp), + struct ceph_pg_mapping, node); + rb_erase(&pg->node, &map->pg_temp); + kfree(pg); + } + while (!RB_EMPTY_ROOT(&map->pg_pools)) { + struct ceph_pg_pool_info *pi = + rb_entry(rb_first(&map->pg_pools), + struct ceph_pg_pool_info, node); + __remove_pg_pool(&map->pg_pools, pi); + } + kfree(map->osd_state); + kfree(map->osd_weight); + kfree(map->osd_addr); + kfree(map); +} + +/* + * adjust max osd value. reallocate arrays. + */ +static int osdmap_set_max_osd(struct ceph_osdmap *map, int max) +{ + u8 *state; + struct ceph_entity_addr *addr; + u32 *weight; + + state = kcalloc(max, sizeof(*state), GFP_NOFS); + addr = kcalloc(max, sizeof(*addr), GFP_NOFS); + weight = kcalloc(max, sizeof(*weight), GFP_NOFS); + if (state == NULL || addr == NULL || weight == NULL) { + kfree(state); + kfree(addr); + kfree(weight); + return -ENOMEM; + } + + /* copy old? */ + if (map->osd_state) { + memcpy(state, map->osd_state, map->max_osd*sizeof(*state)); + memcpy(addr, map->osd_addr, map->max_osd*sizeof(*addr)); + memcpy(weight, map->osd_weight, map->max_osd*sizeof(*weight)); + kfree(map->osd_state); + kfree(map->osd_addr); + kfree(map->osd_weight); + } + + map->osd_state = state; + map->osd_weight = weight; + map->osd_addr = addr; + map->max_osd = max; + return 0; +} + +/* + * decode a full map. + */ +struct ceph_osdmap *osdmap_decode(void **p, void *end) +{ + struct ceph_osdmap *map; + u16 version; + u32 len, max, i; + u8 ev; + int err = -EINVAL; + void *start = *p; + struct ceph_pg_pool_info *pi; + + dout("osdmap_decode %p to %p len %d\n", *p, end, (int)(end - *p)); + + map = kzalloc(sizeof(*map), GFP_NOFS); + if (map == NULL) + return ERR_PTR(-ENOMEM); + map->pg_temp = RB_ROOT; + + ceph_decode_16_safe(p, end, version, bad); + if (version > CEPH_OSDMAP_VERSION) { + pr_warning("got unknown v %d > %d of osdmap\n", version, + CEPH_OSDMAP_VERSION); + goto bad; + } + + ceph_decode_need(p, end, 2*sizeof(u64)+6*sizeof(u32), bad); + ceph_decode_copy(p, &map->fsid, sizeof(map->fsid)); + map->epoch = ceph_decode_32(p); + ceph_decode_copy(p, &map->created, sizeof(map->created)); + ceph_decode_copy(p, &map->modified, sizeof(map->modified)); + + ceph_decode_32_safe(p, end, max, bad); + while (max--) { + ceph_decode_need(p, end, 4 + 1 + sizeof(pi->v), bad); + pi = kzalloc(sizeof(*pi), GFP_NOFS); + if (!pi) + goto bad; + pi->id = ceph_decode_32(p); + ev = ceph_decode_8(p); /* encoding version */ + if (ev > CEPH_PG_POOL_VERSION) { + pr_warning("got unknown v %d > %d of ceph_pg_pool\n", + ev, CEPH_PG_POOL_VERSION); + kfree(pi); + goto bad; + } + err = __decode_pool(p, end, pi); + if (err < 0) + goto bad; + __insert_pg_pool(&map->pg_pools, pi); + } + + if (version >= 5 && __decode_pool_names(p, end, map) < 0) + goto bad; + + ceph_decode_32_safe(p, end, map->pool_max, bad); + + ceph_decode_32_safe(p, end, map->flags, bad); + + max = ceph_decode_32(p); + + /* (re)alloc osd arrays */ + err = osdmap_set_max_osd(map, max); + if (err < 0) + goto bad; + dout("osdmap_decode max_osd = %d\n", map->max_osd); + + /* osds */ + err = -EINVAL; + ceph_decode_need(p, end, 3*sizeof(u32) + + map->max_osd*(1 + sizeof(*map->osd_weight) + + sizeof(*map->osd_addr)), bad); + *p += 4; /* skip length field (should match max) */ + ceph_decode_copy(p, map->osd_state, map->max_osd); + + *p += 4; /* skip length field (should match max) */ + for (i = 0; i < map->max_osd; i++) + map->osd_weight[i] = ceph_decode_32(p); + + *p += 4; /* skip length field (should match max) */ + ceph_decode_copy(p, map->osd_addr, map->max_osd*sizeof(*map->osd_addr)); + for (i = 0; i < map->max_osd; i++) + ceph_decode_addr(&map->osd_addr[i]); + + /* pg_temp */ + ceph_decode_32_safe(p, end, len, bad); + for (i = 0; i < len; i++) { + int n, j; + struct ceph_pg pgid; + struct ceph_pg_mapping *pg; + + ceph_decode_need(p, end, sizeof(u32) + sizeof(u64), bad); + ceph_decode_copy(p, &pgid, sizeof(pgid)); + n = ceph_decode_32(p); + ceph_decode_need(p, end, n * sizeof(u32), bad); + err = -ENOMEM; + pg = kmalloc(sizeof(*pg) + n*sizeof(u32), GFP_NOFS); + if (!pg) + goto bad; + pg->pgid = pgid; + pg->len = n; + for (j = 0; j < n; j++) + pg->osds[j] = ceph_decode_32(p); + + err = __insert_pg_mapping(pg, &map->pg_temp); + if (err) + goto bad; + dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid, len); + } + + /* crush */ + ceph_decode_32_safe(p, end, len, bad); + dout("osdmap_decode crush len %d from off 0x%x\n", len, + (int)(*p - start)); + ceph_decode_need(p, end, len, bad); + map->crush = crush_decode(*p, end); + *p += len; + if (IS_ERR(map->crush)) { + err = PTR_ERR(map->crush); + map->crush = NULL; + goto bad; + } + + /* ignore the rest of the map */ + *p = end; + + dout("osdmap_decode done %p %p\n", *p, end); + return map; + +bad: + dout("osdmap_decode fail\n"); + ceph_osdmap_destroy(map); + return ERR_PTR(err); +} + +/* + * decode and apply an incremental map update. + */ +struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end, + struct ceph_osdmap *map, + struct ceph_messenger *msgr) +{ + struct crush_map *newcrush = NULL; + struct ceph_fsid fsid; + u32 epoch = 0; + struct ceph_timespec modified; + u32 len, pool; + __s32 new_pool_max, new_flags, max; + void *start = *p; + int err = -EINVAL; + u16 version; + struct rb_node *rbp; + + ceph_decode_16_safe(p, end, version, bad); + if (version > CEPH_OSDMAP_INC_VERSION) { + pr_warning("got unknown v %d > %d of inc osdmap\n", version, + CEPH_OSDMAP_INC_VERSION); + goto bad; + } + + ceph_decode_need(p, end, sizeof(fsid)+sizeof(modified)+2*sizeof(u32), + bad); + ceph_decode_copy(p, &fsid, sizeof(fsid)); + epoch = ceph_decode_32(p); + BUG_ON(epoch != map->epoch+1); + ceph_decode_copy(p, &modified, sizeof(modified)); + new_pool_max = ceph_decode_32(p); + new_flags = ceph_decode_32(p); + + /* full map? */ + ceph_decode_32_safe(p, end, len, bad); + if (len > 0) { + dout("apply_incremental full map len %d, %p to %p\n", + len, *p, end); + return osdmap_decode(p, min(*p+len, end)); + } + + /* new crush? */ + ceph_decode_32_safe(p, end, len, bad); + if (len > 0) { + dout("apply_incremental new crush map len %d, %p to %p\n", + len, *p, end); + newcrush = crush_decode(*p, min(*p+len, end)); + if (IS_ERR(newcrush)) + return ERR_CAST(newcrush); + *p += len; + } + + /* new flags? */ + if (new_flags >= 0) + map->flags = new_flags; + if (new_pool_max >= 0) + map->pool_max = new_pool_max; + + ceph_decode_need(p, end, 5*sizeof(u32), bad); + + /* new max? */ + max = ceph_decode_32(p); + if (max >= 0) { + err = osdmap_set_max_osd(map, max); + if (err < 0) + goto bad; + } + + map->epoch++; + map->modified = map->modified; + if (newcrush) { + if (map->crush) + crush_destroy(map->crush); + map->crush = newcrush; + newcrush = NULL; + } + + /* new_pool */ + ceph_decode_32_safe(p, end, len, bad); + while (len--) { + __u8 ev; + struct ceph_pg_pool_info *pi; + + ceph_decode_32_safe(p, end, pool, bad); + ceph_decode_need(p, end, 1 + sizeof(pi->v), bad); + ev = ceph_decode_8(p); /* encoding version */ + if (ev > CEPH_PG_POOL_VERSION) { + pr_warning("got unknown v %d > %d of ceph_pg_pool\n", + ev, CEPH_PG_POOL_VERSION); + goto bad; + } + pi = __lookup_pg_pool(&map->pg_pools, pool); + if (!pi) { + pi = kzalloc(sizeof(*pi), GFP_NOFS); + if (!pi) { + err = -ENOMEM; + goto bad; + } + pi->id = pool; + __insert_pg_pool(&map->pg_pools, pi); + } + err = __decode_pool(p, end, pi); + if (err < 0) + goto bad; + } + if (version >= 5 && __decode_pool_names(p, end, map) < 0) + goto bad; + + /* old_pool */ + ceph_decode_32_safe(p, end, len, bad); + while (len--) { + struct ceph_pg_pool_info *pi; + + ceph_decode_32_safe(p, end, pool, bad); + pi = __lookup_pg_pool(&map->pg_pools, pool); + if (pi) + __remove_pg_pool(&map->pg_pools, pi); + } + + /* new_up */ + err = -EINVAL; + ceph_decode_32_safe(p, end, len, bad); + while (len--) { + u32 osd; + struct ceph_entity_addr addr; + ceph_decode_32_safe(p, end, osd, bad); + ceph_decode_copy_safe(p, end, &addr, sizeof(addr), bad); + ceph_decode_addr(&addr); + pr_info("osd%d up\n", osd); + BUG_ON(osd >= map->max_osd); + map->osd_state[osd] |= CEPH_OSD_UP; + map->osd_addr[osd] = addr; + } + + /* new_down */ + ceph_decode_32_safe(p, end, len, bad); + while (len--) { + u32 osd; + ceph_decode_32_safe(p, end, osd, bad); + (*p)++; /* clean flag */ + pr_info("osd%d down\n", osd); + if (osd < map->max_osd) + map->osd_state[osd] &= ~CEPH_OSD_UP; + } + + /* new_weight */ + ceph_decode_32_safe(p, end, len, bad); + while (len--) { + u32 osd, off; + ceph_decode_need(p, end, sizeof(u32)*2, bad); + osd = ceph_decode_32(p); + off = ceph_decode_32(p); + pr_info("osd%d weight 0x%x %s\n", osd, off, + off == CEPH_OSD_IN ? "(in)" : + (off == CEPH_OSD_OUT ? "(out)" : "")); + if (osd < map->max_osd) + map->osd_weight[osd] = off; + } + + /* new_pg_temp */ + rbp = rb_first(&map->pg_temp); + ceph_decode_32_safe(p, end, len, bad); + while (len--) { + struct ceph_pg_mapping *pg; + int j; + struct ceph_pg pgid; + u32 pglen; + ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad); + ceph_decode_copy(p, &pgid, sizeof(pgid)); + pglen = ceph_decode_32(p); + + /* remove any? */ + while (rbp && pgid_cmp(rb_entry(rbp, struct ceph_pg_mapping, + node)->pgid, pgid) <= 0) { + struct ceph_pg_mapping *cur = + rb_entry(rbp, struct ceph_pg_mapping, node); + + rbp = rb_next(rbp); + dout(" removed pg_temp %llx\n", *(u64 *)&cur->pgid); + rb_erase(&cur->node, &map->pg_temp); + kfree(cur); + } + + if (pglen) { + /* insert */ + ceph_decode_need(p, end, pglen*sizeof(u32), bad); + pg = kmalloc(sizeof(*pg) + sizeof(u32)*pglen, GFP_NOFS); + if (!pg) { + err = -ENOMEM; + goto bad; + } + pg->pgid = pgid; + pg->len = pglen; + for (j = 0; j < pglen; j++) + pg->osds[j] = ceph_decode_32(p); + err = __insert_pg_mapping(pg, &map->pg_temp); + if (err) { + kfree(pg); + goto bad; + } + dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid, + pglen); + } + } + while (rbp) { + struct ceph_pg_mapping *cur = + rb_entry(rbp, struct ceph_pg_mapping, node); + + rbp = rb_next(rbp); + dout(" removed pg_temp %llx\n", *(u64 *)&cur->pgid); + rb_erase(&cur->node, &map->pg_temp); + kfree(cur); + } + + /* ignore the rest */ + *p = end; + return map; + +bad: + pr_err("corrupt inc osdmap epoch %d off %d (%p of %p-%p)\n", + epoch, (int)(*p - start), *p, start, end); + print_hex_dump(KERN_DEBUG, "osdmap: ", + DUMP_PREFIX_OFFSET, 16, 1, + start, end - start, true); + if (newcrush) + crush_destroy(newcrush); + return ERR_PTR(err); +} + + + + +/* + * calculate file layout from given offset, length. + * fill in correct oid, logical length, and object extent + * offset, length. + * + * for now, we write only a single su, until we can + * pass a stride back to the caller. + */ +void ceph_calc_file_object_mapping(struct ceph_file_layout *layout, + u64 off, u64 *plen, + u64 *ono, + u64 *oxoff, u64 *oxlen) +{ + u32 osize = le32_to_cpu(layout->fl_object_size); + u32 su = le32_to_cpu(layout->fl_stripe_unit); + u32 sc = le32_to_cpu(layout->fl_stripe_count); + u32 bl, stripeno, stripepos, objsetno; + u32 su_per_object; + u64 t, su_offset; + + dout("mapping %llu~%llu osize %u fl_su %u\n", off, *plen, + osize, su); + su_per_object = osize / su; + dout("osize %u / su %u = su_per_object %u\n", osize, su, + su_per_object); + + BUG_ON((su & ~PAGE_MASK) != 0); + /* bl = *off / su; */ + t = off; + do_div(t, su); + bl = t; + dout("off %llu / su %u = bl %u\n", off, su, bl); + + stripeno = bl / sc; + stripepos = bl % sc; + objsetno = stripeno / su_per_object; + + *ono = objsetno * sc + stripepos; + dout("objset %u * sc %u = ono %u\n", objsetno, sc, (unsigned)*ono); + + /* *oxoff = *off % layout->fl_stripe_unit; # offset in su */ + t = off; + su_offset = do_div(t, su); + *oxoff = su_offset + (stripeno % su_per_object) * su; + + /* + * Calculate the length of the extent being written to the selected + * object. This is the minimum of the full length requested (plen) or + * the remainder of the current stripe being written to. + */ + *oxlen = min_t(u64, *plen, su - su_offset); + *plen = *oxlen; + + dout(" obj extent %llu~%llu\n", *oxoff, *oxlen); +} +EXPORT_SYMBOL(ceph_calc_file_object_mapping); + +/* + * calculate an object layout (i.e. pgid) from an oid, + * file_layout, and osdmap + */ +int ceph_calc_object_layout(struct ceph_object_layout *ol, + const char *oid, + struct ceph_file_layout *fl, + struct ceph_osdmap *osdmap) +{ + unsigned num, num_mask; + struct ceph_pg pgid; + s32 preferred = (s32)le32_to_cpu(fl->fl_pg_preferred); + int poolid = le32_to_cpu(fl->fl_pg_pool); + struct ceph_pg_pool_info *pool; + unsigned ps; + + BUG_ON(!osdmap); + + pool = __lookup_pg_pool(&osdmap->pg_pools, poolid); + if (!pool) + return -EIO; + ps = ceph_str_hash(pool->v.object_hash, oid, strlen(oid)); + if (preferred >= 0) { + ps += preferred; + num = le32_to_cpu(pool->v.lpg_num); + num_mask = pool->lpg_num_mask; + } else { + num = le32_to_cpu(pool->v.pg_num); + num_mask = pool->pg_num_mask; + } + + pgid.ps = cpu_to_le16(ps); + pgid.preferred = cpu_to_le16(preferred); + pgid.pool = fl->fl_pg_pool; + if (preferred >= 0) + dout("calc_object_layout '%s' pgid %d.%xp%d\n", oid, poolid, ps, + (int)preferred); + else + dout("calc_object_layout '%s' pgid %d.%x\n", oid, poolid, ps); + + ol->ol_pgid = pgid; + ol->ol_stripe_unit = fl->fl_object_stripe_unit; + return 0; +} +EXPORT_SYMBOL(ceph_calc_object_layout); + +/* + * Calculate raw osd vector for the given pgid. Return pointer to osd + * array, or NULL on failure. + */ +static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid, + int *osds, int *num) +{ + struct ceph_pg_mapping *pg; + struct ceph_pg_pool_info *pool; + int ruleno; + unsigned poolid, ps, pps; + int preferred; + + /* pg_temp? */ + pg = __lookup_pg_mapping(&osdmap->pg_temp, pgid); + if (pg) { + *num = pg->len; + return pg->osds; + } + + /* crush */ + poolid = le32_to_cpu(pgid.pool); + ps = le16_to_cpu(pgid.ps); + preferred = (s16)le16_to_cpu(pgid.preferred); + + /* don't forcefeed bad device ids to crush */ + if (preferred >= osdmap->max_osd || + preferred >= osdmap->crush->max_devices) + preferred = -1; + + pool = __lookup_pg_pool(&osdmap->pg_pools, poolid); + if (!pool) + return NULL; + ruleno = crush_find_rule(osdmap->crush, pool->v.crush_ruleset, + pool->v.type, pool->v.size); + if (ruleno < 0) { + pr_err("no crush rule pool %d ruleset %d type %d size %d\n", + poolid, pool->v.crush_ruleset, pool->v.type, + pool->v.size); + return NULL; + } + + if (preferred >= 0) + pps = ceph_stable_mod(ps, + le32_to_cpu(pool->v.lpgp_num), + pool->lpgp_num_mask); + else + pps = ceph_stable_mod(ps, + le32_to_cpu(pool->v.pgp_num), + pool->pgp_num_mask); + pps += poolid; + *num = crush_do_rule(osdmap->crush, ruleno, pps, osds, + min_t(int, pool->v.size, *num), + preferred, osdmap->osd_weight); + return osds; +} + +/* + * Return acting set for given pgid. + */ +int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid, + int *acting) +{ + int rawosds[CEPH_PG_MAX_SIZE], *osds; + int i, o, num = CEPH_PG_MAX_SIZE; + + osds = calc_pg_raw(osdmap, pgid, rawosds, &num); + if (!osds) + return -1; + + /* primary is first up osd */ + o = 0; + for (i = 0; i < num; i++) + if (ceph_osd_is_up(osdmap, osds[i])) + acting[o++] = osds[i]; + return o; +} + +/* + * Return primary osd for given pgid, or -1 if none. + */ +int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg pgid) +{ + int rawosds[CEPH_PG_MAX_SIZE], *osds; + int i, num = CEPH_PG_MAX_SIZE; + + osds = calc_pg_raw(osdmap, pgid, rawosds, &num); + if (!osds) + return -1; + + /* primary is first up osd */ + for (i = 0; i < num; i++) + if (ceph_osd_is_up(osdmap, osds[i])) + return osds[i]; + return -1; +} +EXPORT_SYMBOL(ceph_calc_pg_primary); diff --git a/net/ceph/pagelist.c b/net/ceph/pagelist.c new file mode 100644 index 0000000..3b146cf --- /dev/null +++ b/net/ceph/pagelist.c @@ -0,0 +1,65 @@ + +#include +#include +#include +#include +#include + +static void ceph_pagelist_unmap_tail(struct ceph_pagelist *pl) +{ + struct page *page = list_entry(pl->head.prev, struct page, + lru); + kunmap(page); +} + +int ceph_pagelist_release(struct ceph_pagelist *pl) +{ + if (pl->mapped_tail) + ceph_pagelist_unmap_tail(pl); + + while (!list_empty(&pl->head)) { + struct page *page = list_first_entry(&pl->head, struct page, + lru); + list_del(&page->lru); + __free_page(page); + } + return 0; +} +EXPORT_SYMBOL(ceph_pagelist_release); + +static int ceph_pagelist_addpage(struct ceph_pagelist *pl) +{ + struct page *page = __page_cache_alloc(GFP_NOFS); + if (!page) + return -ENOMEM; + pl->room += PAGE_SIZE; + list_add_tail(&page->lru, &pl->head); + if (pl->mapped_tail) + ceph_pagelist_unmap_tail(pl); + pl->mapped_tail = kmap(page); + return 0; +} + +int ceph_pagelist_append(struct ceph_pagelist *pl, const void *buf, size_t len) +{ + while (pl->room < len) { + size_t bit = pl->room; + int ret; + + memcpy(pl->mapped_tail + (pl->length & ~PAGE_CACHE_MASK), + buf, bit); + pl->length += bit; + pl->room -= bit; + buf += bit; + len -= bit; + ret = ceph_pagelist_addpage(pl); + if (ret) + return ret; + } + + memcpy(pl->mapped_tail + (pl->length & ~PAGE_CACHE_MASK), buf, len); + pl->length += len; + pl->room -= len; + return 0; +} +EXPORT_SYMBOL(ceph_pagelist_append); diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c new file mode 100644 index 0000000..54caf06 --- /dev/null +++ b/net/ceph/pagevec.c @@ -0,0 +1,223 @@ +#include + +#include +#include +#include +#include +#include +#include + +#include + +/* + * build a vector of user pages + */ +struct page **ceph_get_direct_page_vector(const char __user *data, + int num_pages, + loff_t off, size_t len) +{ + struct page **pages; + int rc; + + pages = kmalloc(sizeof(*pages) * num_pages, GFP_NOFS); + if (!pages) + return ERR_PTR(-ENOMEM); + + down_read(¤t->mm->mmap_sem); + rc = get_user_pages(current, current->mm, (unsigned long)data, + num_pages, 0, 0, pages, NULL); + up_read(¤t->mm->mmap_sem); + if (rc < 0) + goto fail; + return pages; + +fail: + kfree(pages); + return ERR_PTR(rc); +} +EXPORT_SYMBOL(ceph_get_direct_page_vector); + +void ceph_put_page_vector(struct page **pages, int num_pages) +{ + int i; + + for (i = 0; i < num_pages; i++) + put_page(pages[i]); + kfree(pages); +} +EXPORT_SYMBOL(ceph_put_page_vector); + +void ceph_release_page_vector(struct page **pages, int num_pages) +{ + int i; + + for (i = 0; i < num_pages; i++) + __free_pages(pages[i], 0); + kfree(pages); +} +EXPORT_SYMBOL(ceph_release_page_vector); + +/* + * allocate a vector new pages + */ +struct page **ceph_alloc_page_vector(int num_pages, gfp_t flags) +{ + struct page **pages; + int i; + + pages = kmalloc(sizeof(*pages) * num_pages, flags); + if (!pages) + return ERR_PTR(-ENOMEM); + for (i = 0; i < num_pages; i++) { + pages[i] = __page_cache_alloc(flags); + if (pages[i] == NULL) { + ceph_release_page_vector(pages, i); + return ERR_PTR(-ENOMEM); + } + } + return pages; +} +EXPORT_SYMBOL(ceph_alloc_page_vector); + +/* + * copy user data into a page vector + */ +int ceph_copy_user_to_page_vector(struct page **pages, + const char __user *data, + loff_t off, size_t len) +{ + int i = 0; + int po = off & ~PAGE_CACHE_MASK; + int left = len; + int l, bad; + + while (left > 0) { + l = min_t(int, PAGE_CACHE_SIZE-po, left); + bad = copy_from_user(page_address(pages[i]) + po, data, l); + if (bad == l) + return -EFAULT; + data += l - bad; + left -= l - bad; + po += l - bad; + if (po == PAGE_CACHE_SIZE) { + po = 0; + i++; + } + } + return len; +} +EXPORT_SYMBOL(ceph_copy_user_to_page_vector); + +int ceph_copy_to_page_vector(struct page **pages, + const char *data, + loff_t off, size_t len) +{ + int i = 0; + size_t po = off & ~PAGE_CACHE_MASK; + size_t left = len; + size_t l; + + while (left > 0) { + l = min_t(size_t, PAGE_CACHE_SIZE-po, left); + memcpy(page_address(pages[i]) + po, data, l); + data += l; + left -= l; + po += l; + if (po == PAGE_CACHE_SIZE) { + po = 0; + i++; + } + } + return len; +} +EXPORT_SYMBOL(ceph_copy_to_page_vector); + +int ceph_copy_from_page_vector(struct page **pages, + char *data, + loff_t off, size_t len) +{ + int i = 0; + size_t po = off & ~PAGE_CACHE_MASK; + size_t left = len; + size_t l; + + while (left > 0) { + l = min_t(size_t, PAGE_CACHE_SIZE-po, left); + memcpy(data, page_address(pages[i]) + po, l); + data += l; + left -= l; + po += l; + if (po == PAGE_CACHE_SIZE) { + po = 0; + i++; + } + } + return len; +} +EXPORT_SYMBOL(ceph_copy_from_page_vector); + +/* + * copy user data from a page vector into a user pointer + */ +int ceph_copy_page_vector_to_user(struct page **pages, + char __user *data, + loff_t off, size_t len) +{ + int i = 0; + int po = off & ~PAGE_CACHE_MASK; + int left = len; + int l, bad; + + while (left > 0) { + l = min_t(int, left, PAGE_CACHE_SIZE-po); + bad = copy_to_user(data, page_address(pages[i]) + po, l); + if (bad == l) + return -EFAULT; + data += l - bad; + left -= l - bad; + if (po) { + po += l - bad; + if (po == PAGE_CACHE_SIZE) + po = 0; + } + i++; + } + return len; +} +EXPORT_SYMBOL(ceph_copy_page_vector_to_user); + +/* + * Zero an extent within a page vector. Offset is relative to the + * start of the first page. + */ +void ceph_zero_page_vector_range(int off, int len, struct page **pages) +{ + int i = off >> PAGE_CACHE_SHIFT; + + off &= ~PAGE_CACHE_MASK; + + dout("zero_page_vector_page %u~%u\n", off, len); + + /* leading partial page? */ + if (off) { + int end = min((int)PAGE_CACHE_SIZE, off + len); + dout("zeroing %d %p head from %d\n", i, pages[i], + (int)off); + zero_user_segment(pages[i], off, end); + len -= (end - off); + i++; + } + while (len >= PAGE_CACHE_SIZE) { + dout("zeroing %d %p len=%d\n", i, pages[i], len); + zero_user_segment(pages[i], 0, PAGE_CACHE_SIZE); + len -= PAGE_CACHE_SIZE; + i++; + } + /* trailing partial page? */ + if (len) { + dout("zeroing %d %p tail to %d\n", i, pages[i], (int)len); + zero_user_segment(pages[i], 0, len); + } +} +EXPORT_SYMBOL(ceph_zero_page_vector_range); + -- cgit v0.10.2 From 602adf400201636e95c3fed9f31fba54a3d7e844 Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Thu, 12 Aug 2010 16:11:25 -0700 Subject: rbd: introduce rados block device (rbd), based on libceph The rados block device (rbd), based on osdblk, creates a block device that is backed by objects stored in the Ceph distributed object storage cluster. Each device consists of a single metadata object and data striped over many data objects. The rbd driver supports read-only snapshots. Signed-off-by: Yehuda Sadeh Signed-off-by: Sage Weil diff --git a/MAINTAINERS b/MAINTAINERS index 48d0565..09497ba 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4807,6 +4807,15 @@ F: fs/qnx4/ F: include/linux/qnx4_fs.h F: include/linux/qnxtypes.h +RADOS BLOCK DEVICE (RBD) +F: include/linux/qnxtypes.h +M: Yehuda Sadeh +M: Sage Weil +M: ceph-devel@vger.kernel.org +S: Supported +F: drivers/block/rbd.c +F: drivers/block/rbd_types.h + RADEON FRAMEBUFFER DISPLAY DRIVER M: Benjamin Herrenschmidt L: linux-fbdev@vger.kernel.org diff --git a/drivers/block/Kconfig b/drivers/block/Kconfig index de27768..4b9359a 100644 --- a/drivers/block/Kconfig +++ b/drivers/block/Kconfig @@ -488,4 +488,21 @@ config BLK_DEV_HD If unsure, say N. +config BLK_DEV_RBD + tristate "Rados block device (RBD)" + depends on INET && EXPERIMENTAL && BLOCK + select CEPH_LIB + select LIBCRC32C + select CRYPTO_AES + select CRYPTO + default n + help + Say Y here if you want include the Rados block device, which stripes + a block device over objects stored in the Ceph distributed object + store. + + More information at http://ceph.newdream.net/. + + If unsure, say N. + endif # BLK_DEV diff --git a/drivers/block/Makefile b/drivers/block/Makefile index aff5ac9..d7f463d 100644 --- a/drivers/block/Makefile +++ b/drivers/block/Makefile @@ -37,5 +37,6 @@ obj-$(CONFIG_BLK_DEV_HD) += hd.o obj-$(CONFIG_XEN_BLKDEV_FRONTEND) += xen-blkfront.o obj-$(CONFIG_BLK_DEV_DRBD) += drbd/ +obj-$(CONFIG_BLK_DEV_RBD) += rbd.o swim_mod-objs := swim.o swim_asm.o diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c new file mode 100644 index 0000000..c42f305 --- /dev/null +++ b/drivers/block/rbd.c @@ -0,0 +1,1843 @@ +/* + rbd.c -- Export ceph rados objects as a Linux block device + + + based on drivers/block/osdblk.c: + + Copyright 2009 Red Hat, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; see the file COPYING. If not, write to + the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. + + + + Instructions for use + -------------------- + + 1) Map a Linux block device to an existing rbd image. + + Usage: [snap name] + + $ echo "192.168.0.1 name=admin rbd foo" > /sys/class/rbd/add + + The snapshot name can be "-" or omitted to map the image read/write. + + 2) List all active blkdev<->object mappings. + + In this example, we have performed step #1 twice, creating two blkdevs, + mapped to two separate rados objects in the rados rbd pool + + $ cat /sys/class/rbd/list + #id major client_name pool name snap KB + 0 254 client4143 rbd foo - 1024000 + + The columns, in order, are: + - blkdev unique id + - blkdev assigned major + - rados client id + - rados pool name + - rados block device name + - mapped snapshot ("-" if none) + - device size in KB + + + 3) Create a snapshot. + + Usage: + + $ echo "0 mysnap" > /sys/class/rbd/snap_create + + + 4) Listing a snapshot. + + $ cat /sys/class/rbd/snaps_list + #id snap KB + 0 - 1024000 (*) + 0 foo 1024000 + + The columns, in order, are: + - blkdev unique id + - snapshot name, '-' means none (active read/write version) + - size of device at time of snapshot + - the (*) indicates this is the active version + + 5) Rollback to snapshot. + + Usage: + + $ echo "0 mysnap" > /sys/class/rbd/snap_rollback + + + 6) Mapping an image using snapshot. + + A snapshot mapping is read-only. This is being done by passing + snap= to the options when adding a device. + + $ echo "192.168.0.1 name=admin,snap=mysnap rbd foo" > /sys/class/rbd/add + + + 7) Remove an active blkdev<->rbd image mapping. + + In this example, we remove the mapping with blkdev unique id 1. + + $ echo 1 > /sys/class/rbd/remove + + + NOTE: The actual creation and deletion of rados objects is outside the scope + of this driver. + + */ + +#include +#include +#include +#include + +#include +#include +#include +#include +#include + +#include "rbd_types.h" + +#define DRV_NAME "rbd" +#define DRV_NAME_LONG "rbd (rados block device)" + +#define RBD_MINORS_PER_MAJOR 256 /* max minors per blkdev */ + +#define RBD_MAX_MD_NAME_LEN (96 + sizeof(RBD_SUFFIX)) +#define RBD_MAX_POOL_NAME_LEN 64 +#define RBD_MAX_SNAP_NAME_LEN 32 +#define RBD_MAX_OPT_LEN 1024 + +#define RBD_SNAP_HEAD_NAME "-" + +#define DEV_NAME_LEN 32 + +/* + * block device image metadata (in-memory version) + */ +struct rbd_image_header { + u64 image_size; + char block_name[32]; + __u8 obj_order; + __u8 crypt_type; + __u8 comp_type; + struct rw_semaphore snap_rwsem; + struct ceph_snap_context *snapc; + size_t snap_names_len; + u64 snap_seq; + u32 total_snaps; + + char *snap_names; + u64 *snap_sizes; +}; + +/* + * an instance of the client. multiple devices may share a client. + */ +struct rbd_client { + struct ceph_client *client; + struct kref kref; + struct list_head node; +}; + +/* + * a single io request + */ +struct rbd_request { + struct request *rq; /* blk layer request */ + struct bio *bio; /* cloned bio */ + struct page **pages; /* list of used pages */ + u64 len; +}; + +/* + * a single device + */ +struct rbd_device { + int id; /* blkdev unique id */ + + int major; /* blkdev assigned major */ + struct gendisk *disk; /* blkdev's gendisk and rq */ + struct request_queue *q; + + struct ceph_client *client; + struct rbd_client *rbd_client; + + char name[DEV_NAME_LEN]; /* blkdev name, e.g. rbd3 */ + + spinlock_t lock; /* queue lock */ + + struct rbd_image_header header; + char obj[RBD_MAX_OBJ_NAME_LEN]; /* rbd image name */ + int obj_len; + char obj_md_name[RBD_MAX_MD_NAME_LEN]; /* hdr nm. */ + char pool_name[RBD_MAX_POOL_NAME_LEN]; + int poolid; + + char snap_name[RBD_MAX_SNAP_NAME_LEN]; + u32 cur_snap; /* index+1 of current snapshot within snap context + 0 - for the head */ + int read_only; + + struct list_head node; +}; + +static spinlock_t node_lock; /* protects client get/put */ + +static struct class *class_rbd; /* /sys/class/rbd */ +static DEFINE_MUTEX(ctl_mutex); /* Serialize open/close/setup/teardown */ +static LIST_HEAD(rbd_dev_list); /* devices */ +static LIST_HEAD(rbd_client_list); /* clients */ + + +static int rbd_open(struct block_device *bdev, fmode_t mode) +{ + struct gendisk *disk = bdev->bd_disk; + struct rbd_device *rbd_dev = disk->private_data; + + set_device_ro(bdev, rbd_dev->read_only); + + if ((mode & FMODE_WRITE) && rbd_dev->read_only) + return -EROFS; + + return 0; +} + +static const struct block_device_operations rbd_bd_ops = { + .owner = THIS_MODULE, + .open = rbd_open, +}; + +/* + * Initialize an rbd client instance. + * We own *opt. + */ +static struct rbd_client *rbd_client_create(struct ceph_options *opt) +{ + struct rbd_client *rbdc; + int ret = -ENOMEM; + + dout("rbd_client_create\n"); + rbdc = kmalloc(sizeof(struct rbd_client), GFP_KERNEL); + if (!rbdc) + goto out_opt; + + kref_init(&rbdc->kref); + INIT_LIST_HEAD(&rbdc->node); + + rbdc->client = ceph_create_client(opt, rbdc); + if (IS_ERR(rbdc->client)) + goto out_rbdc; + + ret = ceph_open_session(rbdc->client); + if (ret < 0) + goto out_err; + + spin_lock(&node_lock); + list_add_tail(&rbdc->node, &rbd_client_list); + spin_unlock(&node_lock); + + dout("rbd_client_create created %p\n", rbdc); + return rbdc; + +out_err: + ceph_destroy_client(rbdc->client); + return ERR_PTR(ret); + +out_rbdc: + kfree(rbdc); +out_opt: + ceph_destroy_options(opt); + return ERR_PTR(-ENOMEM); +} + +/* + * Find a ceph client with specific addr and configuration. + */ +static struct rbd_client *__rbd_client_find(struct ceph_options *opt) +{ + struct rbd_client *client_node; + + if (opt->flags & CEPH_OPT_NOSHARE) + return NULL; + + list_for_each_entry(client_node, &rbd_client_list, node) + if (ceph_compare_options(opt, client_node->client) == 0) + return client_node; + return NULL; +} + +/* + * Get a ceph client with specific addr and configuration, if one does + * not exist create it. + */ +static int rbd_get_client(struct rbd_device *rbd_dev, const char *mon_addr, + char *options) +{ + struct rbd_client *rbdc; + struct ceph_options *opt; + int ret; + + ret = ceph_parse_options(&opt, options, mon_addr, + mon_addr + strlen(mon_addr), NULL, NULL); + if (ret < 0) + return ret; + + spin_lock(&node_lock); + rbdc = __rbd_client_find(opt); + if (rbdc) { + ceph_destroy_options(opt); + + /* using an existing client */ + kref_get(&rbdc->kref); + rbd_dev->rbd_client = rbdc; + rbd_dev->client = rbdc->client; + spin_unlock(&node_lock); + return 0; + } + spin_unlock(&node_lock); + + rbdc = rbd_client_create(opt); + if (IS_ERR(rbdc)) + return PTR_ERR(rbdc); + + rbd_dev->rbd_client = rbdc; + rbd_dev->client = rbdc->client; + return 0; +} + +/* + * Destroy ceph client + */ +static void rbd_client_release(struct kref *kref) +{ + struct rbd_client *rbdc = container_of(kref, struct rbd_client, kref); + + dout("rbd_release_client %p\n", rbdc); + spin_lock(&node_lock); + list_del(&rbdc->node); + spin_unlock(&node_lock); + + ceph_destroy_client(rbdc->client); + kfree(rbdc); +} + +/* + * Drop reference to ceph client node. If it's not referenced anymore, release + * it. + */ +static void rbd_put_client(struct rbd_device *rbd_dev) +{ + kref_put(&rbd_dev->rbd_client->kref, rbd_client_release); + rbd_dev->rbd_client = NULL; + rbd_dev->client = NULL; +} + + +/* + * Create a new header structure, translate header format from the on-disk + * header. + */ +static int rbd_header_from_disk(struct rbd_image_header *header, + struct rbd_image_header_ondisk *ondisk, + int allocated_snaps, + gfp_t gfp_flags) +{ + int i; + u32 snap_count = le32_to_cpu(ondisk->snap_count); + int ret = -ENOMEM; + + init_rwsem(&header->snap_rwsem); + + header->snap_names_len = le64_to_cpu(ondisk->snap_names_len); + header->snapc = kmalloc(sizeof(struct ceph_snap_context) + + snap_count * + sizeof(struct rbd_image_snap_ondisk), + gfp_flags); + if (!header->snapc) + return -ENOMEM; + if (snap_count) { + header->snap_names = kmalloc(header->snap_names_len, + GFP_KERNEL); + if (!header->snap_names) + goto err_snapc; + header->snap_sizes = kmalloc(snap_count * sizeof(u64), + GFP_KERNEL); + if (!header->snap_sizes) + goto err_names; + } else { + header->snap_names = NULL; + header->snap_sizes = NULL; + } + memcpy(header->block_name, ondisk->block_name, + sizeof(ondisk->block_name)); + + header->image_size = le64_to_cpu(ondisk->image_size); + header->obj_order = ondisk->options.order; + header->crypt_type = ondisk->options.crypt_type; + header->comp_type = ondisk->options.comp_type; + + atomic_set(&header->snapc->nref, 1); + header->snap_seq = le64_to_cpu(ondisk->snap_seq); + header->snapc->num_snaps = snap_count; + header->total_snaps = snap_count; + + if (snap_count && + allocated_snaps == snap_count) { + for (i = 0; i < snap_count; i++) { + header->snapc->snaps[i] = + le64_to_cpu(ondisk->snaps[i].id); + header->snap_sizes[i] = + le64_to_cpu(ondisk->snaps[i].image_size); + } + + /* copy snapshot names */ + memcpy(header->snap_names, &ondisk->snaps[i], + header->snap_names_len); + } + + return 0; + +err_names: + kfree(header->snap_names); +err_snapc: + kfree(header->snapc); + return ret; +} + +static int snap_index(struct rbd_image_header *header, int snap_num) +{ + return header->total_snaps - snap_num; +} + +static u64 cur_snap_id(struct rbd_device *rbd_dev) +{ + struct rbd_image_header *header = &rbd_dev->header; + + if (!rbd_dev->cur_snap) + return 0; + + return header->snapc->snaps[snap_index(header, rbd_dev->cur_snap)]; +} + +static int snap_by_name(struct rbd_image_header *header, const char *snap_name, + u64 *seq, u64 *size) +{ + int i; + char *p = header->snap_names; + + for (i = 0; i < header->total_snaps; i++, p += strlen(p) + 1) { + if (strcmp(snap_name, p) == 0) + break; + } + if (i == header->total_snaps) + return -ENOENT; + if (seq) + *seq = header->snapc->snaps[i]; + + if (size) + *size = header->snap_sizes[i]; + + return i; +} + +static int rbd_header_set_snap(struct rbd_device *dev, + const char *snap_name, + u64 *size) +{ + struct rbd_image_header *header = &dev->header; + struct ceph_snap_context *snapc = header->snapc; + int ret = -ENOENT; + + down_write(&header->snap_rwsem); + + if (!snap_name || + !*snap_name || + strcmp(snap_name, "-") == 0 || + strcmp(snap_name, RBD_SNAP_HEAD_NAME) == 0) { + if (header->total_snaps) + snapc->seq = header->snap_seq; + else + snapc->seq = 0; + dev->cur_snap = 0; + dev->read_only = 0; + if (size) + *size = header->image_size; + } else { + ret = snap_by_name(header, snap_name, &snapc->seq, size); + if (ret < 0) + goto done; + + dev->cur_snap = header->total_snaps - ret; + dev->read_only = 1; + } + + ret = 0; +done: + up_write(&header->snap_rwsem); + return ret; +} + +static void rbd_header_free(struct rbd_image_header *header) +{ + kfree(header->snapc); + kfree(header->snap_names); + kfree(header->snap_sizes); +} + +/* + * get the actual striped segment name, offset and length + */ +static u64 rbd_get_segment(struct rbd_image_header *header, + const char *block_name, + u64 ofs, u64 len, + char *seg_name, u64 *segofs) +{ + u64 seg = ofs >> header->obj_order; + + if (seg_name) + snprintf(seg_name, RBD_MAX_SEG_NAME_LEN, + "%s.%012llx", block_name, seg); + + ofs = ofs & ((1 << header->obj_order) - 1); + len = min_t(u64, len, (1 << header->obj_order) - ofs); + + if (segofs) + *segofs = ofs; + + return len; +} + +/* + * bio helpers + */ + +static void bio_chain_put(struct bio *chain) +{ + struct bio *tmp; + + while (chain) { + tmp = chain; + chain = chain->bi_next; + bio_put(tmp); + } +} + +/* + * zeros a bio chain, starting at specific offset + */ +static void zero_bio_chain(struct bio *chain, int start_ofs) +{ + struct bio_vec *bv; + unsigned long flags; + void *buf; + int i; + int pos = 0; + + while (chain) { + bio_for_each_segment(bv, chain, i) { + if (pos + bv->bv_len > start_ofs) { + int remainder = max(start_ofs - pos, 0); + buf = bvec_kmap_irq(bv, &flags); + memset(buf + remainder, 0, + bv->bv_len - remainder); + bvec_kunmap_irq(bv, &flags); + } + pos += bv->bv_len; + } + + chain = chain->bi_next; + } +} + +/* + * bio_chain_clone - clone a chain of bios up to a certain length. + * might return a bio_pair that will need to be released. + */ +static struct bio *bio_chain_clone(struct bio **old, struct bio **next, + struct bio_pair **bp, + int len, gfp_t gfpmask) +{ + struct bio *tmp, *old_chain = *old, *new_chain = NULL, *tail = NULL; + int total = 0; + + if (*bp) { + bio_pair_release(*bp); + *bp = NULL; + } + + while (old_chain && (total < len)) { + tmp = bio_kmalloc(gfpmask, old_chain->bi_max_vecs); + if (!tmp) + goto err_out; + + if (total + old_chain->bi_size > len) { + struct bio_pair *bp; + + /* + * this split can only happen with a single paged bio, + * split_bio will BUG_ON if this is not the case + */ + dout("bio_chain_clone split! total=%d remaining=%d" + "bi_size=%d\n", + (int)total, (int)len-total, + (int)old_chain->bi_size); + + /* split the bio. We'll release it either in the next + call, or it will have to be released outside */ + bp = bio_split(old_chain, (len - total) / 512ULL); + if (!bp) + goto err_out; + + __bio_clone(tmp, &bp->bio1); + + *next = &bp->bio2; + } else { + __bio_clone(tmp, old_chain); + *next = old_chain->bi_next; + } + + tmp->bi_bdev = NULL; + gfpmask &= ~__GFP_WAIT; + tmp->bi_next = NULL; + + if (!new_chain) { + new_chain = tail = tmp; + } else { + tail->bi_next = tmp; + tail = tmp; + } + old_chain = old_chain->bi_next; + + total += tmp->bi_size; + } + + BUG_ON(total < len); + + if (tail) + tail->bi_next = NULL; + + *old = old_chain; + + return new_chain; + +err_out: + dout("bio_chain_clone with err\n"); + bio_chain_put(new_chain); + return NULL; +} + +/* + * helpers for osd request op vectors. + */ +static int rbd_create_rw_ops(struct ceph_osd_req_op **ops, + int num_ops, + int opcode, + u32 payload_len) +{ + *ops = kzalloc(sizeof(struct ceph_osd_req_op) * (num_ops + 1), + GFP_NOIO); + if (!*ops) + return -ENOMEM; + (*ops)[0].op = opcode; + /* + * op extent offset and length will be set later on + * in calc_raw_layout() + */ + (*ops)[0].payload_len = payload_len; + return 0; +} + +static void rbd_destroy_ops(struct ceph_osd_req_op *ops) +{ + kfree(ops); +} + +/* + * Send ceph osd request + */ +static int rbd_do_request(struct request *rq, + struct rbd_device *dev, + struct ceph_snap_context *snapc, + u64 snapid, + const char *obj, u64 ofs, u64 len, + struct bio *bio, + struct page **pages, + int num_pages, + int flags, + struct ceph_osd_req_op *ops, + int num_reply, + void (*rbd_cb)(struct ceph_osd_request *req, + struct ceph_msg *msg)) +{ + struct ceph_osd_request *req; + struct ceph_file_layout *layout; + int ret; + u64 bno; + struct timespec mtime = CURRENT_TIME; + struct rbd_request *req_data; + struct ceph_osd_request_head *reqhead; + struct rbd_image_header *header = &dev->header; + + ret = -ENOMEM; + req_data = kzalloc(sizeof(*req_data), GFP_NOIO); + if (!req_data) + goto done; + + dout("rbd_do_request len=%lld ofs=%lld\n", len, ofs); + + down_read(&header->snap_rwsem); + + req = ceph_osdc_alloc_request(&dev->client->osdc, flags, + snapc, + ops, + false, + GFP_NOIO, pages, bio); + if (IS_ERR(req)) { + up_read(&header->snap_rwsem); + ret = PTR_ERR(req); + goto done_pages; + } + + req->r_callback = rbd_cb; + + req_data->rq = rq; + req_data->bio = bio; + req_data->pages = pages; + req_data->len = len; + + req->r_priv = req_data; + + reqhead = req->r_request->front.iov_base; + reqhead->snapid = cpu_to_le64(CEPH_NOSNAP); + + strncpy(req->r_oid, obj, sizeof(req->r_oid)); + req->r_oid_len = strlen(req->r_oid); + + layout = &req->r_file_layout; + memset(layout, 0, sizeof(*layout)); + layout->fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); + layout->fl_stripe_count = cpu_to_le32(1); + layout->fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); + layout->fl_pg_preferred = cpu_to_le32(-1); + layout->fl_pg_pool = cpu_to_le32(dev->poolid); + ceph_calc_raw_layout(&dev->client->osdc, layout, snapid, + ofs, &len, &bno, req, ops); + + ceph_osdc_build_request(req, ofs, &len, + ops, + snapc, + &mtime, + req->r_oid, req->r_oid_len); + up_read(&header->snap_rwsem); + + ret = ceph_osdc_start_request(&dev->client->osdc, req, false); + if (ret < 0) + goto done_err; + + if (!rbd_cb) { + ret = ceph_osdc_wait_request(&dev->client->osdc, req); + ceph_osdc_put_request(req); + } + return ret; + +done_err: + bio_chain_put(req_data->bio); + ceph_osdc_put_request(req); +done_pages: + kfree(req_data); +done: + if (rq) + blk_end_request(rq, ret, len); + return ret; +} + +/* + * Ceph osd op callback + */ +static void rbd_req_cb(struct ceph_osd_request *req, struct ceph_msg *msg) +{ + struct rbd_request *req_data = req->r_priv; + struct ceph_osd_reply_head *replyhead; + struct ceph_osd_op *op; + __s32 rc; + u64 bytes; + int read_op; + + /* parse reply */ + replyhead = msg->front.iov_base; + WARN_ON(le32_to_cpu(replyhead->num_ops) == 0); + op = (void *)(replyhead + 1); + rc = le32_to_cpu(replyhead->result); + bytes = le64_to_cpu(op->extent.length); + read_op = (le32_to_cpu(op->op) == CEPH_OSD_OP_READ); + + dout("rbd_req_cb bytes=%lld readop=%d rc=%d\n", bytes, read_op, rc); + + if (rc == -ENOENT && read_op) { + zero_bio_chain(req_data->bio, 0); + rc = 0; + } else if (rc == 0 && read_op && bytes < req_data->len) { + zero_bio_chain(req_data->bio, bytes); + bytes = req_data->len; + } + + blk_end_request(req_data->rq, rc, bytes); + + if (req_data->bio) + bio_chain_put(req_data->bio); + + ceph_osdc_put_request(req); + kfree(req_data); +} + +/* + * Do a synchronous ceph osd operation + */ +static int rbd_req_sync_op(struct rbd_device *dev, + struct ceph_snap_context *snapc, + u64 snapid, + int opcode, + int flags, + struct ceph_osd_req_op *orig_ops, + int num_reply, + const char *obj, + u64 ofs, u64 len, + char *buf) +{ + int ret; + struct page **pages; + int num_pages; + struct ceph_osd_req_op *ops = orig_ops; + u32 payload_len; + + num_pages = calc_pages_for(ofs , len); + pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL); + if (!pages) + return -ENOMEM; + + if (!orig_ops) { + payload_len = (flags & CEPH_OSD_FLAG_WRITE ? len : 0); + ret = rbd_create_rw_ops(&ops, 1, opcode, payload_len); + if (ret < 0) + goto done; + + if ((flags & CEPH_OSD_FLAG_WRITE) && buf) { + ret = ceph_copy_to_page_vector(pages, buf, ofs, len); + if (ret < 0) + goto done_ops; + } + } + + ret = rbd_do_request(NULL, dev, snapc, snapid, + obj, ofs, len, NULL, + pages, num_pages, + flags, + ops, + 2, + NULL); + if (ret < 0) + goto done_ops; + + if ((flags & CEPH_OSD_FLAG_READ) && buf) + ret = ceph_copy_from_page_vector(pages, buf, ofs, ret); + +done_ops: + if (!orig_ops) + rbd_destroy_ops(ops); +done: + ceph_release_page_vector(pages, num_pages); + return ret; +} + +/* + * Do an asynchronous ceph osd operation + */ +static int rbd_do_op(struct request *rq, + struct rbd_device *rbd_dev , + struct ceph_snap_context *snapc, + u64 snapid, + int opcode, int flags, int num_reply, + u64 ofs, u64 len, + struct bio *bio) +{ + char *seg_name; + u64 seg_ofs; + u64 seg_len; + int ret; + struct ceph_osd_req_op *ops; + u32 payload_len; + + seg_name = kmalloc(RBD_MAX_SEG_NAME_LEN + 1, GFP_NOIO); + if (!seg_name) + return -ENOMEM; + + seg_len = rbd_get_segment(&rbd_dev->header, + rbd_dev->header.block_name, + ofs, len, + seg_name, &seg_ofs); + if (seg_len < 0) + return seg_len; + + payload_len = (flags & CEPH_OSD_FLAG_WRITE ? seg_len : 0); + + ret = rbd_create_rw_ops(&ops, 1, opcode, payload_len); + if (ret < 0) + goto done; + + /* we've taken care of segment sizes earlier when we + cloned the bios. We should never have a segment + truncated at this point */ + BUG_ON(seg_len < len); + + ret = rbd_do_request(rq, rbd_dev, snapc, snapid, + seg_name, seg_ofs, seg_len, + bio, + NULL, 0, + flags, + ops, + num_reply, + rbd_req_cb); +done: + kfree(seg_name); + return ret; +} + +/* + * Request async osd write + */ +static int rbd_req_write(struct request *rq, + struct rbd_device *rbd_dev, + struct ceph_snap_context *snapc, + u64 ofs, u64 len, + struct bio *bio) +{ + return rbd_do_op(rq, rbd_dev, snapc, CEPH_NOSNAP, + CEPH_OSD_OP_WRITE, + CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK, + 2, + ofs, len, bio); +} + +/* + * Request async osd read + */ +static int rbd_req_read(struct request *rq, + struct rbd_device *rbd_dev, + u64 snapid, + u64 ofs, u64 len, + struct bio *bio) +{ + return rbd_do_op(rq, rbd_dev, NULL, + (snapid ? snapid : CEPH_NOSNAP), + CEPH_OSD_OP_READ, + CEPH_OSD_FLAG_READ, + 2, + ofs, len, bio); +} + +/* + * Request sync osd read + */ +static int rbd_req_sync_read(struct rbd_device *dev, + struct ceph_snap_context *snapc, + u64 snapid, + const char *obj, + u64 ofs, u64 len, + char *buf) +{ + return rbd_req_sync_op(dev, NULL, + (snapid ? snapid : CEPH_NOSNAP), + CEPH_OSD_OP_READ, + CEPH_OSD_FLAG_READ, + NULL, + 1, obj, ofs, len, buf); +} + +/* + * Request sync osd read + */ +static int rbd_req_sync_rollback_obj(struct rbd_device *dev, + u64 snapid, + const char *obj) +{ + struct ceph_osd_req_op *ops; + int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_ROLLBACK, 0); + if (ret < 0) + return ret; + + ops[0].snap.snapid = snapid; + + ret = rbd_req_sync_op(dev, NULL, + CEPH_NOSNAP, + 0, + CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK, + ops, + 1, obj, 0, 0, NULL); + + rbd_destroy_ops(ops); + + if (ret < 0) + return ret; + + return ret; +} + +/* + * Request sync osd read + */ +static int rbd_req_sync_exec(struct rbd_device *dev, + const char *obj, + const char *cls, + const char *method, + const char *data, + int len) +{ + struct ceph_osd_req_op *ops; + int cls_len = strlen(cls); + int method_len = strlen(method); + int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_CALL, + cls_len + method_len + len); + if (ret < 0) + return ret; + + ops[0].cls.class_name = cls; + ops[0].cls.class_len = (__u8)cls_len; + ops[0].cls.method_name = method; + ops[0].cls.method_len = (__u8)method_len; + ops[0].cls.argc = 0; + ops[0].cls.indata = data; + ops[0].cls.indata_len = len; + + ret = rbd_req_sync_op(dev, NULL, + CEPH_NOSNAP, + 0, + CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK, + ops, + 1, obj, 0, 0, NULL); + + rbd_destroy_ops(ops); + + dout("cls_exec returned %d\n", ret); + return ret; +} + +/* + * block device queue callback + */ +static void rbd_rq_fn(struct request_queue *q) +{ + struct rbd_device *rbd_dev = q->queuedata; + struct request *rq; + struct bio_pair *bp = NULL; + + rq = blk_fetch_request(q); + + while (1) { + struct bio *bio; + struct bio *rq_bio, *next_bio = NULL; + bool do_write; + int size, op_size = 0; + u64 ofs; + + /* peek at request from block layer */ + if (!rq) + break; + + dout("fetched request\n"); + + /* filter out block requests we don't understand */ + if ((rq->cmd_type != REQ_TYPE_FS)) { + __blk_end_request_all(rq, 0); + goto next; + } + + /* deduce our operation (read, write) */ + do_write = (rq_data_dir(rq) == WRITE); + + size = blk_rq_bytes(rq); + ofs = blk_rq_pos(rq) * 512ULL; + rq_bio = rq->bio; + if (do_write && rbd_dev->read_only) { + __blk_end_request_all(rq, -EROFS); + goto next; + } + + spin_unlock_irq(q->queue_lock); + + dout("%s 0x%x bytes at 0x%llx\n", + do_write ? "write" : "read", + size, blk_rq_pos(rq) * 512ULL); + + do { + /* a bio clone to be passed down to OSD req */ + dout("rq->bio->bi_vcnt=%d\n", rq->bio->bi_vcnt); + op_size = rbd_get_segment(&rbd_dev->header, + rbd_dev->header.block_name, + ofs, size, + NULL, NULL); + bio = bio_chain_clone(&rq_bio, &next_bio, &bp, + op_size, GFP_ATOMIC); + if (!bio) { + spin_lock_irq(q->queue_lock); + __blk_end_request_all(rq, -ENOMEM); + goto next; + } + + /* init OSD command: write or read */ + if (do_write) + rbd_req_write(rq, rbd_dev, + rbd_dev->header.snapc, + ofs, + op_size, bio); + else + rbd_req_read(rq, rbd_dev, + cur_snap_id(rbd_dev), + ofs, + op_size, bio); + + size -= op_size; + ofs += op_size; + + rq_bio = next_bio; + } while (size > 0); + + if (bp) + bio_pair_release(bp); + + spin_lock_irq(q->queue_lock); +next: + rq = blk_fetch_request(q); + } +} + +/* + * a queue callback. Makes sure that we don't create a bio that spans across + * multiple osd objects. One exception would be with a single page bios, + * which we handle later at bio_chain_clone + */ +static int rbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bmd, + struct bio_vec *bvec) +{ + struct rbd_device *rbd_dev = q->queuedata; + unsigned int chunk_sectors = 1 << (rbd_dev->header.obj_order - 9); + sector_t sector = bmd->bi_sector + get_start_sect(bmd->bi_bdev); + unsigned int bio_sectors = bmd->bi_size >> 9; + int max; + + max = (chunk_sectors - ((sector & (chunk_sectors - 1)) + + bio_sectors)) << 9; + if (max < 0) + max = 0; /* bio_add cannot handle a negative return */ + if (max <= bvec->bv_len && bio_sectors == 0) + return bvec->bv_len; + return max; +} + +static void rbd_free_disk(struct rbd_device *rbd_dev) +{ + struct gendisk *disk = rbd_dev->disk; + + if (!disk) + return; + + rbd_header_free(&rbd_dev->header); + + if (disk->flags & GENHD_FL_UP) + del_gendisk(disk); + if (disk->queue) + blk_cleanup_queue(disk->queue); + put_disk(disk); +} + +/* + * reload the ondisk the header + */ +static int rbd_read_header(struct rbd_device *rbd_dev, + struct rbd_image_header *header) +{ + ssize_t rc; + struct rbd_image_header_ondisk *dh; + int snap_count = 0; + u64 snap_names_len = 0; + + while (1) { + int len = sizeof(*dh) + + snap_count * sizeof(struct rbd_image_snap_ondisk) + + snap_names_len; + + rc = -ENOMEM; + dh = kmalloc(len, GFP_KERNEL); + if (!dh) + return -ENOMEM; + + rc = rbd_req_sync_read(rbd_dev, + NULL, CEPH_NOSNAP, + rbd_dev->obj_md_name, + 0, len, + (char *)dh); + if (rc < 0) + goto out_dh; + + rc = rbd_header_from_disk(header, dh, snap_count, GFP_KERNEL); + if (rc < 0) + goto out_dh; + + if (snap_count != header->total_snaps) { + snap_count = header->total_snaps; + snap_names_len = header->snap_names_len; + rbd_header_free(header); + kfree(dh); + continue; + } + break; + } + +out_dh: + kfree(dh); + return rc; +} + +/* + * create a snapshot + */ +static int rbd_header_add_snap(struct rbd_device *dev, + const char *snap_name, + gfp_t gfp_flags) +{ + int name_len = strlen(snap_name); + u64 new_snapid; + int ret; + void *data, *data_start, *data_end; + + /* we should create a snapshot only if we're pointing at the head */ + if (dev->cur_snap) + return -EINVAL; + + ret = ceph_monc_create_snapid(&dev->client->monc, dev->poolid, + &new_snapid); + dout("created snapid=%lld\n", new_snapid); + if (ret < 0) + return ret; + + data = kmalloc(name_len + 16, gfp_flags); + if (!data) + return -ENOMEM; + + data_start = data; + data_end = data + name_len + 16; + + ceph_encode_string_safe(&data, data_end, snap_name, name_len, bad); + ceph_encode_64_safe(&data, data_end, new_snapid, bad); + + ret = rbd_req_sync_exec(dev, dev->obj_md_name, "rbd", "snap_add", + data_start, data - data_start); + + kfree(data_start); + + if (ret < 0) + return ret; + + dev->header.snapc->seq = new_snapid; + + return 0; +bad: + return -ERANGE; +} + +/* + * only read the first part of the ondisk header, without the snaps info + */ +static int rbd_update_snaps(struct rbd_device *rbd_dev) +{ + int ret; + struct rbd_image_header h; + u64 snap_seq; + + ret = rbd_read_header(rbd_dev, &h); + if (ret < 0) + return ret; + + down_write(&rbd_dev->header.snap_rwsem); + + snap_seq = rbd_dev->header.snapc->seq; + + kfree(rbd_dev->header.snapc); + kfree(rbd_dev->header.snap_names); + kfree(rbd_dev->header.snap_sizes); + + rbd_dev->header.total_snaps = h.total_snaps; + rbd_dev->header.snapc = h.snapc; + rbd_dev->header.snap_names = h.snap_names; + rbd_dev->header.snap_sizes = h.snap_sizes; + rbd_dev->header.snapc->seq = snap_seq; + + up_write(&rbd_dev->header.snap_rwsem); + + return 0; +} + +static int rbd_init_disk(struct rbd_device *rbd_dev) +{ + struct gendisk *disk; + struct request_queue *q; + int rc; + u64 total_size = 0; + + /* contact OSD, request size info about the object being mapped */ + rc = rbd_read_header(rbd_dev, &rbd_dev->header); + if (rc) + return rc; + + rc = rbd_header_set_snap(rbd_dev, rbd_dev->snap_name, &total_size); + if (rc) + return rc; + + /* create gendisk info */ + rc = -ENOMEM; + disk = alloc_disk(RBD_MINORS_PER_MAJOR); + if (!disk) + goto out; + + sprintf(disk->disk_name, DRV_NAME "%d", rbd_dev->id); + disk->major = rbd_dev->major; + disk->first_minor = 0; + disk->fops = &rbd_bd_ops; + disk->private_data = rbd_dev; + + /* init rq */ + rc = -ENOMEM; + q = blk_init_queue(rbd_rq_fn, &rbd_dev->lock); + if (!q) + goto out_disk; + blk_queue_merge_bvec(q, rbd_merge_bvec); + disk->queue = q; + + q->queuedata = rbd_dev; + + rbd_dev->disk = disk; + rbd_dev->q = q; + + /* finally, announce the disk to the world */ + set_capacity(disk, total_size / 512ULL); + add_disk(disk); + + pr_info("%s: added with size 0x%llx\n", + disk->disk_name, (unsigned long long)total_size); + return 0; + +out_disk: + put_disk(disk); +out: + return rc; +} + +/******************************************************************** + * /sys/class/rbd/ + * add map rados objects to blkdev + * remove unmap rados objects + * list show mappings + *******************************************************************/ + +static void class_rbd_release(struct class *cls) +{ + kfree(cls); +} + +static ssize_t class_rbd_list(struct class *c, + struct class_attribute *attr, + char *data) +{ + int n = 0; + struct list_head *tmp; + int max = PAGE_SIZE; + + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); + + n += snprintf(data, max, + "#id\tmajor\tclient_name\tpool\tname\tsnap\tKB\n"); + + list_for_each(tmp, &rbd_dev_list) { + struct rbd_device *rbd_dev; + + rbd_dev = list_entry(tmp, struct rbd_device, node); + n += snprintf(data+n, max-n, + "%d\t%d\tclient%lld\t%s\t%s\t%s\t%lld\n", + rbd_dev->id, + rbd_dev->major, + ceph_client_id(rbd_dev->client), + rbd_dev->pool_name, + rbd_dev->obj, rbd_dev->snap_name, + rbd_dev->header.image_size >> 10); + if (n == max) + break; + } + + mutex_unlock(&ctl_mutex); + return n; +} + +static ssize_t class_rbd_add(struct class *c, + struct class_attribute *attr, + const char *buf, size_t count) +{ + struct ceph_osd_client *osdc; + struct rbd_device *rbd_dev; + ssize_t rc = -ENOMEM; + int irc, new_id = 0; + struct list_head *tmp; + char *mon_dev_name; + char *options; + + if (!try_module_get(THIS_MODULE)) + return -ENODEV; + + mon_dev_name = kmalloc(RBD_MAX_OPT_LEN, GFP_KERNEL); + if (!mon_dev_name) + goto err_out_mod; + + options = kmalloc(RBD_MAX_OPT_LEN, GFP_KERNEL); + if (!options) + goto err_mon_dev; + + /* new rbd_device object */ + rbd_dev = kzalloc(sizeof(*rbd_dev), GFP_KERNEL); + if (!rbd_dev) + goto err_out_opt; + + /* static rbd_device initialization */ + spin_lock_init(&rbd_dev->lock); + INIT_LIST_HEAD(&rbd_dev->node); + + /* generate unique id: find highest unique id, add one */ + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); + + list_for_each(tmp, &rbd_dev_list) { + struct rbd_device *rbd_dev; + + rbd_dev = list_entry(tmp, struct rbd_device, node); + if (rbd_dev->id >= new_id) + new_id = rbd_dev->id + 1; + } + + rbd_dev->id = new_id; + + /* add to global list */ + list_add_tail(&rbd_dev->node, &rbd_dev_list); + + /* parse add command */ + if (sscanf(buf, "%" __stringify(RBD_MAX_OPT_LEN) "s " + "%" __stringify(RBD_MAX_OPT_LEN) "s " + "%" __stringify(RBD_MAX_POOL_NAME_LEN) "s " + "%" __stringify(RBD_MAX_OBJ_NAME_LEN) "s" + "%" __stringify(RBD_MAX_SNAP_NAME_LEN) "s", + mon_dev_name, options, rbd_dev->pool_name, + rbd_dev->obj, rbd_dev->snap_name) < 4) { + rc = -EINVAL; + goto err_out_slot; + } + + if (rbd_dev->snap_name[0] == 0) + rbd_dev->snap_name[0] = '-'; + + rbd_dev->obj_len = strlen(rbd_dev->obj); + snprintf(rbd_dev->obj_md_name, sizeof(rbd_dev->obj_md_name), "%s%s", + rbd_dev->obj, RBD_SUFFIX); + + /* initialize rest of new object */ + snprintf(rbd_dev->name, DEV_NAME_LEN, DRV_NAME "%d", rbd_dev->id); + rc = rbd_get_client(rbd_dev, mon_dev_name, options); + if (rc < 0) + goto err_out_slot; + + mutex_unlock(&ctl_mutex); + + /* pick the pool */ + osdc = &rbd_dev->client->osdc; + rc = ceph_pg_poolid_by_name(osdc->osdmap, rbd_dev->pool_name); + if (rc < 0) + goto err_out_client; + rbd_dev->poolid = rc; + + /* register our block device */ + irc = register_blkdev(0, rbd_dev->name); + if (irc < 0) { + rc = irc; + goto err_out_client; + } + rbd_dev->major = irc; + + /* set up and announce blkdev mapping */ + rc = rbd_init_disk(rbd_dev); + if (rc) + goto err_out_blkdev; + + return count; + +err_out_blkdev: + unregister_blkdev(rbd_dev->major, rbd_dev->name); +err_out_client: + rbd_put_client(rbd_dev); + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); +err_out_slot: + list_del_init(&rbd_dev->node); + mutex_unlock(&ctl_mutex); + + kfree(rbd_dev); +err_out_opt: + kfree(options); +err_mon_dev: + kfree(mon_dev_name); +err_out_mod: + dout("Error adding device %s\n", buf); + module_put(THIS_MODULE); + return rc; +} + +static struct rbd_device *__rbd_get_dev(unsigned long id) +{ + struct list_head *tmp; + struct rbd_device *rbd_dev; + + list_for_each(tmp, &rbd_dev_list) { + rbd_dev = list_entry(tmp, struct rbd_device, node); + if (rbd_dev->id == id) + return rbd_dev; + } + return NULL; +} + +static ssize_t class_rbd_remove(struct class *c, + struct class_attribute *attr, + const char *buf, + size_t count) +{ + struct rbd_device *rbd_dev = NULL; + int target_id, rc; + unsigned long ul; + + rc = strict_strtoul(buf, 10, &ul); + if (rc) + return rc; + + /* convert to int; abort if we lost anything in the conversion */ + target_id = (int) ul; + if (target_id != ul) + return -EINVAL; + + /* remove object from list immediately */ + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); + + rbd_dev = __rbd_get_dev(target_id); + if (rbd_dev) + list_del_init(&rbd_dev->node); + + mutex_unlock(&ctl_mutex); + + if (!rbd_dev) + return -ENOENT; + + rbd_put_client(rbd_dev); + + /* clean up and free blkdev */ + rbd_free_disk(rbd_dev); + unregister_blkdev(rbd_dev->major, rbd_dev->name); + kfree(rbd_dev); + + /* release module ref */ + module_put(THIS_MODULE); + + return count; +} + +static ssize_t class_rbd_snaps_list(struct class *c, + struct class_attribute *attr, + char *data) +{ + struct rbd_device *rbd_dev = NULL; + struct list_head *tmp; + struct rbd_image_header *header; + int i, n = 0, max = PAGE_SIZE; + int ret; + + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); + + n += snprintf(data, max, "#id\tsnap\tKB\n"); + + list_for_each(tmp, &rbd_dev_list) { + char *names, *p; + struct ceph_snap_context *snapc; + + rbd_dev = list_entry(tmp, struct rbd_device, node); + header = &rbd_dev->header; + + down_read(&header->snap_rwsem); + + names = header->snap_names; + snapc = header->snapc; + + n += snprintf(data + n, max - n, "%d\t%s\t%lld%s\n", + rbd_dev->id, RBD_SNAP_HEAD_NAME, + header->image_size >> 10, + (!rbd_dev->cur_snap ? " (*)" : "")); + if (n == max) + break; + + p = names; + for (i = 0; i < header->total_snaps; i++, p += strlen(p) + 1) { + n += snprintf(data + n, max - n, "%d\t%s\t%lld%s\n", + rbd_dev->id, p, header->snap_sizes[i] >> 10, + (rbd_dev->cur_snap && + (snap_index(header, i) == rbd_dev->cur_snap) ? + " (*)" : "")); + if (n == max) + break; + } + + up_read(&header->snap_rwsem); + } + + + ret = n; + mutex_unlock(&ctl_mutex); + return ret; +} + +static ssize_t class_rbd_snaps_refresh(struct class *c, + struct class_attribute *attr, + const char *buf, + size_t count) +{ + struct rbd_device *rbd_dev = NULL; + int target_id, rc; + unsigned long ul; + int ret = count; + + rc = strict_strtoul(buf, 10, &ul); + if (rc) + return rc; + + /* convert to int; abort if we lost anything in the conversion */ + target_id = (int) ul; + if (target_id != ul) + return -EINVAL; + + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); + + rbd_dev = __rbd_get_dev(target_id); + if (!rbd_dev) { + ret = -ENOENT; + goto done; + } + + rc = rbd_update_snaps(rbd_dev); + if (rc < 0) + ret = rc; + +done: + mutex_unlock(&ctl_mutex); + return ret; +} + +static ssize_t class_rbd_snap_create(struct class *c, + struct class_attribute *attr, + const char *buf, + size_t count) +{ + struct rbd_device *rbd_dev = NULL; + int target_id, ret; + char *name; + + name = kmalloc(RBD_MAX_SNAP_NAME_LEN + 1, GFP_KERNEL); + if (!name) + return -ENOMEM; + + /* parse snaps add command */ + if (sscanf(buf, "%d " + "%" __stringify(RBD_MAX_SNAP_NAME_LEN) "s", + &target_id, + name) != 2) { + ret = -EINVAL; + goto done; + } + + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); + + rbd_dev = __rbd_get_dev(target_id); + if (!rbd_dev) { + ret = -ENOENT; + goto done_unlock; + } + + ret = rbd_header_add_snap(rbd_dev, + name, GFP_KERNEL); + if (ret < 0) + goto done_unlock; + + ret = rbd_update_snaps(rbd_dev); + if (ret < 0) + goto done_unlock; + + ret = count; +done_unlock: + mutex_unlock(&ctl_mutex); +done: + kfree(name); + return ret; +} + +static ssize_t class_rbd_rollback(struct class *c, + struct class_attribute *attr, + const char *buf, + size_t count) +{ + struct rbd_device *rbd_dev = NULL; + int target_id, ret; + u64 snapid; + char snap_name[RBD_MAX_SNAP_NAME_LEN]; + u64 cur_ofs; + char *seg_name; + + /* parse snaps add command */ + if (sscanf(buf, "%d " + "%" __stringify(RBD_MAX_SNAP_NAME_LEN) "s", + &target_id, + snap_name) != 2) { + return -EINVAL; + } + + ret = -ENOMEM; + seg_name = kmalloc(RBD_MAX_SEG_NAME_LEN + 1, GFP_NOIO); + if (!seg_name) + return ret; + + mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING); + + rbd_dev = __rbd_get_dev(target_id); + if (!rbd_dev) { + ret = -ENOENT; + goto done_unlock; + } + + ret = snap_by_name(&rbd_dev->header, snap_name, &snapid, NULL); + if (ret < 0) + goto done_unlock; + + dout("snapid=%lld\n", snapid); + + cur_ofs = 0; + while (cur_ofs < rbd_dev->header.image_size) { + cur_ofs += rbd_get_segment(&rbd_dev->header, + rbd_dev->obj, + cur_ofs, (u64)-1, + seg_name, NULL); + dout("seg_name=%s\n", seg_name); + + ret = rbd_req_sync_rollback_obj(rbd_dev, snapid, seg_name); + if (ret < 0) + pr_warning("could not roll back obj %s err=%d\n", + seg_name, ret); + } + + ret = rbd_update_snaps(rbd_dev); + if (ret < 0) + goto done_unlock; + + ret = count; + +done_unlock: + mutex_unlock(&ctl_mutex); + kfree(seg_name); + + return ret; +} + +static struct class_attribute class_rbd_attrs[] = { + __ATTR(add, 0200, NULL, class_rbd_add), + __ATTR(remove, 0200, NULL, class_rbd_remove), + __ATTR(list, 0444, class_rbd_list, NULL), + __ATTR(snaps_refresh, 0200, NULL, class_rbd_snaps_refresh), + __ATTR(snap_create, 0200, NULL, class_rbd_snap_create), + __ATTR(snaps_list, 0444, class_rbd_snaps_list, NULL), + __ATTR(snap_rollback, 0200, NULL, class_rbd_rollback), + __ATTR_NULL +}; + +/* + * create control files in sysfs + * /sys/class/rbd/... + */ +static int rbd_sysfs_init(void) +{ + int ret = -ENOMEM; + + class_rbd = kzalloc(sizeof(*class_rbd), GFP_KERNEL); + if (!class_rbd) + goto out; + + class_rbd->name = DRV_NAME; + class_rbd->owner = THIS_MODULE; + class_rbd->class_release = class_rbd_release; + class_rbd->class_attrs = class_rbd_attrs; + + ret = class_register(class_rbd); + if (ret) + goto out_class; + return 0; + +out_class: + kfree(class_rbd); + class_rbd = NULL; + pr_err(DRV_NAME ": failed to create class rbd\n"); +out: + return ret; +} + +static void rbd_sysfs_cleanup(void) +{ + if (class_rbd) + class_destroy(class_rbd); + class_rbd = NULL; +} + +int __init rbd_init(void) +{ + int rc; + + rc = rbd_sysfs_init(); + if (rc) + return rc; + spin_lock_init(&node_lock); + pr_info("loaded " DRV_NAME_LONG "\n"); + return 0; +} + +void __exit rbd_exit(void) +{ + rbd_sysfs_cleanup(); +} + +module_init(rbd_init); +module_exit(rbd_exit); + +MODULE_AUTHOR("Sage Weil "); +MODULE_AUTHOR("Yehuda Sadeh "); +MODULE_DESCRIPTION("rados block device"); + +/* following authorship retained from original osdblk.c */ +MODULE_AUTHOR("Jeff Garzik "); + +MODULE_LICENSE("GPL"); diff --git a/drivers/block/rbd_types.h b/drivers/block/rbd_types.h new file mode 100644 index 0000000..fc6c678 --- /dev/null +++ b/drivers/block/rbd_types.h @@ -0,0 +1,73 @@ +/* + * Ceph - scalable distributed file system + * + * Copyright (C) 2004-2010 Sage Weil + * + * This is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License version 2.1, as published by the Free Software + * Foundation. See file COPYING. + * + */ + +#ifndef CEPH_RBD_TYPES_H +#define CEPH_RBD_TYPES_H + +#include + +/* + * rbd image 'foo' consists of objects + * foo.rbd - image metadata + * foo.00000000 + * foo.00000001 + * ... - data + */ + +#define RBD_SUFFIX ".rbd" +#define RBD_DIRECTORY "rbd_directory" +#define RBD_INFO "rbd_info" + +#define RBD_DEFAULT_OBJ_ORDER 22 /* 4MB */ +#define RBD_MIN_OBJ_ORDER 16 +#define RBD_MAX_OBJ_ORDER 30 + +#define RBD_MAX_OBJ_NAME_LEN 96 +#define RBD_MAX_SEG_NAME_LEN 128 + +#define RBD_COMP_NONE 0 +#define RBD_CRYPT_NONE 0 + +#define RBD_HEADER_TEXT "<<< Rados Block Device Image >>>\n" +#define RBD_HEADER_SIGNATURE "RBD" +#define RBD_HEADER_VERSION "001.005" + +struct rbd_info { + __le64 max_id; +} __attribute__ ((packed)); + +struct rbd_image_snap_ondisk { + __le64 id; + __le64 image_size; +} __attribute__((packed)); + +struct rbd_image_header_ondisk { + char text[40]; + char block_name[24]; + char signature[4]; + char version[8]; + struct { + __u8 order; + __u8 crypt_type; + __u8 comp_type; + __u8 unused; + } __attribute__((packed)) options; + __le64 image_size; + __le64 snap_seq; + __le32 snap_count; + __le32 reserved; + __le64 snap_names_len; + struct rbd_image_snap_ondisk snaps[0]; +} __attribute__((packed)); + + +#endif diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c index 33d0499..27d4ea31 100644 --- a/net/ceph/debugfs.c +++ b/net/ceph/debugfs.c @@ -252,8 +252,7 @@ void ceph_debugfs_cleanup(void) { } -int ceph_debugfs_client_init(struct ceph_client *client, - int (*module_debugfs_init)(struct ceph_client *)) +int ceph_debugfs_client_init(struct ceph_client *client) { return 0; } -- cgit v0.10.2 From 4c32f5dda5ffe23687a55da1538b7cc426710d1a Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Tue, 24 Aug 2010 16:27:36 -0700 Subject: ceph: do not hide .snap in root directory Snaps in the root directory are now supported by the MDS, and harmless on older versions. Signed-off-by: Sage Weil diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index a8d2aac..4958252 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -493,7 +493,6 @@ struct dentry *ceph_finish_lookup(struct ceph_mds_request *req, /* .snap dir? */ if (err == -ENOENT && - ceph_vino(parent).ino != CEPH_INO_ROOT && /* no .snap in root dir */ strcmp(dentry->d_name.name, fsc->mount_options->snapdir_name) == 0) { struct inode *inode = ceph_get_snapdir(parent); -- cgit v0.10.2 From 93afd449aa3c0430ef409c13e1cb2b3f0458fc10 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Fri, 17 Sep 2010 08:38:25 -0700 Subject: ceph: only invalidate on check_caps if we actually have pages The i_rdcache_gen value only implies we MAY have cached pages; actually check the mapping to see if it's worth bothering with an invalidate. Signed-off-by: Sage Weil diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 3cff67c..3c03460 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1534,7 +1534,7 @@ retry_locked: */ if ((!is_delayed || mdsc->stopping) && ci->i_wrbuffer_ref == 0 && /* no dirty pages... */ - ci->i_rdcache_gen && /* may have cached pages */ + inode->i_data.nrpages && /* have cached pages */ (file_wanted == 0 || /* no open files */ (revoking & (CEPH_CAP_FILE_CACHE| CEPH_CAP_FILE_LAZYIO))) && /* or revoking cache */ -- cgit v0.10.2 From 18a38193efcaac1fb3c94ad8fa04bb117850a3c2 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Fri, 17 Sep 2010 10:46:44 -0700 Subject: ceph: use mapping->nrpages to determine if mapping is empty This is simpler and faster. Signed-off-by: Sage Weil diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 3c03460..98ab13e 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1417,17 +1417,6 @@ static int __mark_caps_flushing(struct inode *inode, /* * try to invalidate mapping pages without blocking. */ -static int mapping_is_empty(struct address_space *mapping) -{ - struct page *page = find_get_page(mapping, 0); - - if (!page) - return 1; - - put_page(page); - return 0; -} - static int try_nonblocking_invalidate(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); @@ -1437,7 +1426,7 @@ static int try_nonblocking_invalidate(struct inode *inode) invalidate_mapping_pages(&inode->i_data, 0, -1); spin_lock(&inode->i_lock); - if (mapping_is_empty(&inode->i_data) && + if (inode->i_data.nrpages == 0 && invalidating_gen == ci->i_rdcache_gen) { /* success. */ dout("try_nonblocking_invalidate %p success\n", inode); -- cgit v0.10.2 From ac0b74d8a1ced8ea86147467daf06b15b130dd94 Mon Sep 17 00:00:00 2001 From: Greg Farnum Date: Fri, 17 Sep 2010 10:10:55 -0700 Subject: ceph: add pagelist_reserve, pagelist_truncate, pagelist_set_cursor These facilitate preallocation of pages so that we can encode into the pagelist in an atomic context. Signed-off-by: Greg Farnum Signed-off-by: Sage Weil diff --git a/include/linux/ceph/pagelist.h b/include/linux/ceph/pagelist.h index cc9327a..9660d6b 100644 --- a/include/linux/ceph/pagelist.h +++ b/include/linux/ceph/pagelist.h @@ -8,6 +8,14 @@ struct ceph_pagelist { void *mapped_tail; size_t length; size_t room; + struct list_head free_list; + size_t num_pages_free; +}; + +struct ceph_pagelist_cursor { + struct ceph_pagelist *pl; /* pagelist, for error checking */ + struct list_head *page_lru; /* page in list */ + size_t room; /* room remaining to reset to */ }; static inline void ceph_pagelist_init(struct ceph_pagelist *pl) @@ -16,11 +24,24 @@ static inline void ceph_pagelist_init(struct ceph_pagelist *pl) pl->mapped_tail = NULL; pl->length = 0; pl->room = 0; + INIT_LIST_HEAD(&pl->free_list); + pl->num_pages_free = 0; } + extern int ceph_pagelist_release(struct ceph_pagelist *pl); extern int ceph_pagelist_append(struct ceph_pagelist *pl, const void *d, size_t l); +extern int ceph_pagelist_reserve(struct ceph_pagelist *pl, size_t space); + +extern int ceph_pagelist_free_reserve(struct ceph_pagelist *pl); + +extern void ceph_pagelist_set_cursor(struct ceph_pagelist *pl, + struct ceph_pagelist_cursor *c); + +extern int ceph_pagelist_truncate(struct ceph_pagelist *pl, + struct ceph_pagelist_cursor *c); + static inline int ceph_pagelist_encode_64(struct ceph_pagelist *pl, u64 v) { __le64 ev = cpu_to_le64(v); diff --git a/net/ceph/pagelist.c b/net/ceph/pagelist.c index 3b146cf..b8cbc45 100644 --- a/net/ceph/pagelist.c +++ b/net/ceph/pagelist.c @@ -7,35 +7,42 @@ static void ceph_pagelist_unmap_tail(struct ceph_pagelist *pl) { - struct page *page = list_entry(pl->head.prev, struct page, - lru); - kunmap(page); + if (pl->mapped_tail) { + struct page *page = list_entry(pl->head.prev, struct page, lru); + kunmap(page); + pl->mapped_tail = NULL; + } } int ceph_pagelist_release(struct ceph_pagelist *pl) { - if (pl->mapped_tail) - ceph_pagelist_unmap_tail(pl); - + ceph_pagelist_unmap_tail(pl); while (!list_empty(&pl->head)) { struct page *page = list_first_entry(&pl->head, struct page, lru); list_del(&page->lru); __free_page(page); } + ceph_pagelist_free_reserve(pl); return 0; } EXPORT_SYMBOL(ceph_pagelist_release); static int ceph_pagelist_addpage(struct ceph_pagelist *pl) { - struct page *page = __page_cache_alloc(GFP_NOFS); + struct page *page; + + if (!pl->num_pages_free) { + page = __page_cache_alloc(GFP_NOFS); + } else { + page = list_first_entry(&pl->free_list, struct page, lru); + list_del(&page->lru); + } if (!page) return -ENOMEM; pl->room += PAGE_SIZE; + ceph_pagelist_unmap_tail(pl); list_add_tail(&page->lru, &pl->head); - if (pl->mapped_tail) - ceph_pagelist_unmap_tail(pl); pl->mapped_tail = kmap(page); return 0; } @@ -63,3 +70,84 @@ int ceph_pagelist_append(struct ceph_pagelist *pl, const void *buf, size_t len) return 0; } EXPORT_SYMBOL(ceph_pagelist_append); + +/** + * Allocate enough pages for a pagelist to append the given amount + * of data without without allocating. + * Returns: 0 on success, -ENOMEM on error. + */ +int ceph_pagelist_reserve(struct ceph_pagelist *pl, size_t space) +{ + if (space <= pl->room) + return 0; + space -= pl->room; + space = (space + PAGE_SIZE - 1) >> PAGE_SHIFT; /* conv to num pages */ + + while (space > pl->num_pages_free) { + struct page *page = __page_cache_alloc(GFP_NOFS); + if (!page) + return -ENOMEM; + list_add_tail(&page->lru, &pl->free_list); + ++pl->num_pages_free; + } + return 0; +} +EXPORT_SYMBOL(ceph_pagelist_reserve); + +/** + * Free any pages that have been preallocated. + */ +int ceph_pagelist_free_reserve(struct ceph_pagelist *pl) +{ + while (!list_empty(&pl->free_list)) { + struct page *page = list_first_entry(&pl->free_list, + struct page, lru); + list_del(&page->lru); + __free_page(page); + --pl->num_pages_free; + } + BUG_ON(pl->num_pages_free); + return 0; +} +EXPORT_SYMBOL(ceph_pagelist_free_reserve); + +/** + * Create a truncation point. + */ +void ceph_pagelist_set_cursor(struct ceph_pagelist *pl, + struct ceph_pagelist_cursor *c) +{ + c->pl = pl; + c->page_lru = pl->head.prev; + c->room = pl->room; +} +EXPORT_SYMBOL(ceph_pagelist_set_cursor); + +/** + * Truncate a pagelist to the given point. Move extra pages to reserve. + * This won't sleep. + * Returns: 0 on success, + * -EINVAL if the pagelist doesn't match the trunc point pagelist + */ +int ceph_pagelist_truncate(struct ceph_pagelist *pl, + struct ceph_pagelist_cursor *c) +{ + struct page *page; + + if (pl != c->pl) + return -EINVAL; + ceph_pagelist_unmap_tail(pl); + while (pl->head.prev != c->page_lru) { + page = list_entry(pl->head.prev, struct page, lru); + list_del(&page->lru); /* remove from pagelist */ + list_add_tail(&page->lru, &pl->free_list); /* add to reserve */ + ++pl->num_pages_free; + } + pl->room = c->room; + if (!list_empty(&pl->head)) { + page = list_entry(pl->head.prev, struct page, lru); + pl->mapped_tail = kmap(page); + } + return 0; +} +EXPORT_SYMBOL(ceph_pagelist_truncate); -- cgit v0.10.2 From fca4451acfdcf894154e4809529ca28a09db88ff Mon Sep 17 00:00:00 2001 From: Greg Farnum Date: Fri, 17 Sep 2010 10:24:02 -0700 Subject: ceph: preallocate flock state without locks held When the lock_kernel() turns into lock_flocks() and a spinlock, we won't be able to do allocations with the lock held. Preallocate space without the lock, and retry if the lock state changes out from underneath us. Signed-off-by: Greg Farnum Signed-off-by: Sage Weil diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c index 087afe9..40abde9 100644 --- a/fs/ceph/locks.c +++ b/fs/ceph/locks.c @@ -181,8 +181,9 @@ void ceph_count_locks(struct inode *inode, int *fcntl_count, int *flock_count) * Encode the flock and fcntl locks for the given inode into the pagelist. * Format is: #fcntl locks, sequential fcntl locks, #flock locks, * sequential flock locks. - * Must be called with BLK already held, and the lock numbers should have - * been gathered under the same lock holding window. + * Must be called with lock_flocks() already held. + * If we encounter more of a specific lock type than expected, + * we return the value 1. */ int ceph_encode_locks(struct inode *inode, struct ceph_pagelist *pagelist, int num_fcntl_locks, int num_flock_locks) @@ -190,6 +191,8 @@ int ceph_encode_locks(struct inode *inode, struct ceph_pagelist *pagelist, struct file_lock *lock; struct ceph_filelock cephlock; int err = 0; + int seen_fcntl = 0; + int seen_flock = 0; dout("encoding %d flock and %d fcntl locks", num_flock_locks, num_fcntl_locks); @@ -198,6 +201,11 @@ int ceph_encode_locks(struct inode *inode, struct ceph_pagelist *pagelist, goto fail; for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { if (lock->fl_flags & FL_POSIX) { + ++seen_fcntl; + if (seen_fcntl > num_fcntl_locks) { + err = -ENOSPC; + goto fail; + } err = lock_to_ceph_filelock(lock, &cephlock); if (err) goto fail; @@ -213,6 +221,11 @@ int ceph_encode_locks(struct inode *inode, struct ceph_pagelist *pagelist, goto fail; for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { if (lock->fl_flags & FL_FLOCK) { + ++seen_flock; + if (seen_flock > num_flock_locks) { + err = -ENOSPC; + goto fail; + } err = lock_to_ceph_filelock(lock, &cephlock); if (err) goto fail; diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 3356823..fbfc298 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -2365,19 +2365,35 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, if (recon_state->flock) { int num_fcntl_locks, num_flock_locks; - - lock_kernel(); - ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks); - rec.v2.flock_len = (2*sizeof(u32) + - (num_fcntl_locks+num_flock_locks) * - sizeof(struct ceph_filelock)); - - err = ceph_pagelist_append(pagelist, &rec, reclen); - if (!err) - err = ceph_encode_locks(inode, pagelist, - num_fcntl_locks, - num_flock_locks); - unlock_kernel(); + struct ceph_pagelist_cursor trunc_point; + + ceph_pagelist_set_cursor(pagelist, &trunc_point); + do { + lock_kernel(); + ceph_count_locks(inode, &num_fcntl_locks, + &num_flock_locks); + rec.v2.flock_len = (2*sizeof(u32) + + (num_fcntl_locks+num_flock_locks) * + sizeof(struct ceph_filelock)); + unlock_kernel(); + + /* pre-alloc pagelist */ + ceph_pagelist_truncate(pagelist, &trunc_point); + err = ceph_pagelist_append(pagelist, &rec, reclen); + if (!err) + err = ceph_pagelist_reserve(pagelist, + rec.v2.flock_len); + + /* encode locks */ + if (!err) { + lock_kernel(); + err = ceph_encode_locks(inode, + pagelist, + num_fcntl_locks, + num_flock_locks); + unlock_kernel(); + } + } while (err == -ENOSPC); } else { err = ceph_pagelist_append(pagelist, &rec, reclen); } -- cgit v0.10.2 From 496e59553c51ce18acc836de070106b583926b87 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Wed, 22 Sep 2010 19:57:10 -0700 Subject: ceph: switch from BKL to lock_flocks() Switch from using the BKL explicitly to the new lock_flocks() interface. Eventually this will turn into a spinlock. Signed-off-by: Sage Weil diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index fbfc298..3142b15 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1,11 +1,12 @@ #include +#include #include #include #include -#include #include #include +#include #include "super.h" #include "mds_client.h" @@ -2369,13 +2370,13 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, ceph_pagelist_set_cursor(pagelist, &trunc_point); do { - lock_kernel(); + lock_flocks(); ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks); rec.v2.flock_len = (2*sizeof(u32) + (num_fcntl_locks+num_flock_locks) * sizeof(struct ceph_filelock)); - unlock_kernel(); + unlock_flocks(); /* pre-alloc pagelist */ ceph_pagelist_truncate(pagelist, &trunc_point); @@ -2386,12 +2387,12 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, /* encode locks */ if (!err) { - lock_kernel(); + lock_flocks(); err = ceph_encode_locks(inode, pagelist, num_fcntl_locks, num_flock_locks); - unlock_kernel(); + unlock_flocks(); } } while (err == -ENOSPC); } else { -- cgit v0.10.2 From 28f259b7cd78eb29d38b7ae6b475d656e08fd348 Mon Sep 17 00:00:00 2001 From: Vasiliy Kulikov Date: Sun, 26 Sep 2010 12:59:37 +0400 Subject: block: rbd: fixed may leaks rbd_client_create() doesn't free rbdc, this leads to many leaks. seg_len in rbd_do_op() is unsigned, so (seg_len < 0) makes no sense. Also if fixed check fails then seg_name is leaked. Signed-off-by: Vasiliy Kulikov Signed-off-by: Yehuda Sadeh diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index c42f305..83b15dd 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -241,6 +241,7 @@ static struct rbd_client *rbd_client_create(struct ceph_options *opt) rbdc->client = ceph_create_client(opt, rbdc); if (IS_ERR(rbdc->client)) goto out_rbdc; + opt = NULL; /* Now rbdc->client is responsible for opt */ ret = ceph_open_session(rbdc->client); if (ret < 0) @@ -255,13 +256,12 @@ static struct rbd_client *rbd_client_create(struct ceph_options *opt) out_err: ceph_destroy_client(rbdc->client); - return ERR_PTR(ret); - out_rbdc: kfree(rbdc); out_opt: - ceph_destroy_options(opt); - return ERR_PTR(-ENOMEM); + if (opt) + ceph_destroy_options(opt); + return ERR_PTR(ret); } /* @@ -889,8 +889,10 @@ static int rbd_do_op(struct request *rq, rbd_dev->header.block_name, ofs, len, seg_name, &seg_ofs); - if (seg_len < 0) - return seg_len; + if ((s64)seg_len < 0) { + ret = seg_len; + goto done; + } payload_len = (flags & CEPH_OSD_FLAG_WRITE ? seg_len : 0); -- cgit v0.10.2 From f4cf3deef4c474381e8fee2e6099d49edd9105cb Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Mon, 27 Sep 2010 10:51:53 -0700 Subject: block: rbd: removing unnecessary test rbd_get_segment() can't return a negative value, we don't need to check the return output. Signed-off-by: Yehuda Sadeh diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 83b15dd..1ac87f1 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -889,10 +889,6 @@ static int rbd_do_op(struct request *rq, rbd_dev->header.block_name, ofs, len, seg_name, &seg_ofs); - if ((s64)seg_len < 0) { - ret = seg_len; - goto done; - } payload_len = (flags & CEPH_OSD_FLAG_WRITE ? seg_len : 0); -- cgit v0.10.2 From 6f453ed6c07dbed83b368269c9c0fb170866ee71 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Tue, 28 Sep 2010 09:53:10 -0700 Subject: ceph: fix debugfs warnings Include "super.h" outside of CONFIG_DEBUG_FS to eliminate a compiler warning: fs/ceph/debugfs.c:266: warning: 'struct ceph_fs_client' declared inside parameter list fs/ceph/debugfs.c:266: warning: its scope is only this definition or declaration, which is probably not what you want fs/ceph/debugfs.c:271: warning: 'struct ceph_fs_client' declared inside parameter list Signed-off-by: Randy Dunlap Signed-off-by: Yehuda Sadeh diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c index 94fe1d2..7ae1b3d 100644 --- a/fs/ceph/debugfs.c +++ b/fs/ceph/debugfs.c @@ -12,9 +12,10 @@ #include #include +#include "super.h" + #ifdef CONFIG_DEBUG_FS -#include "super.h" #include "mds_client.h" static int mdsmap_show(struct seq_file *s, void *p) -- cgit v0.10.2 From 010e3b48fca57920557d2b80b83f8b2899fb5d1e Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Thu, 30 Sep 2010 11:58:31 -0700 Subject: ceph: don't crash when passed bad mount options This only happened when parse_extra_token was not passed to ceph_parse_option() (hence, only happened in rbd). Signed-off-by: Yehuda Sadeh diff --git a/net/ceph/ceph_common.c b/net/ceph/ceph_common.c index f6f2eeb..f3e4a13 100644 --- a/net/ceph/ceph_common.c +++ b/net/ceph/ceph_common.c @@ -249,7 +249,7 @@ int ceph_parse_options(struct ceph_options **popt, char *options, continue; err = -EINVAL; token = match_token((char *)c, opt_tokens, argstr); - if (token < 0) { + if (token < 0 && parse_extra_token) { /* extra? */ err = parse_extra_token((char *)c, private); if (err < 0) { -- cgit v0.10.2 From 571dba52a34015a5a7aa5d480a86936878444a6f Mon Sep 17 00:00:00 2001 From: Greg Farnum Date: Fri, 24 Sep 2010 14:56:40 -0700 Subject: ceph: add CEPH_MDS_OP_SETDIRLAYOUT and associated ioctl. Signed-off-by: Sage Weil diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 899578b..8888c9ba 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -92,6 +92,68 @@ static long ceph_ioctl_set_layout(struct file *file, void __user *arg) } /* + * Set a layout policy on a directory inode. All items in the tree + * rooted at this inode will inherit this layout on creation, + * (It doesn't apply retroactively ) + * unless a subdirectory has its own layout policy. + */ +static long ceph_ioctl_set_layout_policy (struct file *file, void __user *arg) +{ + struct inode *inode = file->f_dentry->d_inode; + struct ceph_mds_request *req; + struct ceph_ioctl_layout l; + int err, i; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; + + /* copy and validate */ + if (copy_from_user(&l, arg, sizeof(l))) + return -EFAULT; + + if ((l.object_size & ~PAGE_MASK) || + (l.stripe_unit & ~PAGE_MASK) || + !l.stripe_unit || + (l.object_size && + (unsigned)l.object_size % (unsigned)l.stripe_unit)) + return -EINVAL; + + /* make sure it's a valid data pool */ + if (l.data_pool > 0) { + mutex_lock(&mdsc->mutex); + err = -EINVAL; + for (i = 0; i < mdsc->mdsmap->m_num_data_pg_pools; i++) + if (mdsc->mdsmap->m_data_pg_pools[i] == l.data_pool) { + err = 0; + break; + } + mutex_unlock(&mdsc->mutex); + if (err) + return err; + } + + req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_SETDIRLAYOUT, + USE_AUTH_MDS); + + if (IS_ERR(req)) + return PTR_ERR(req); + req->r_inode = igrab(inode); + + req->r_args.setlayout.layout.fl_stripe_unit = + cpu_to_le32(l.stripe_unit); + req->r_args.setlayout.layout.fl_stripe_count = + cpu_to_le32(l.stripe_count); + req->r_args.setlayout.layout.fl_object_size = + cpu_to_le32(l.object_size); + req->r_args.setlayout.layout.fl_pg_pool = + cpu_to_le32(l.data_pool); + req->r_args.setlayout.layout.fl_pg_preferred = + cpu_to_le32(l.preferred_osd); + + err = ceph_mdsc_do_request(mdsc, inode, req); + ceph_mdsc_put_request(req); + return err; +} + +/* * Return object name, size/offset information, and location (OSD * number, network address) for a given file offset. */ @@ -177,11 +239,15 @@ long ceph_ioctl(struct file *file, unsigned int cmd, unsigned long arg) case CEPH_IOC_SET_LAYOUT: return ceph_ioctl_set_layout(file, (void __user *)arg); + case CEPH_IOC_SET_LAYOUT_POLICY: + return ceph_ioctl_set_layout_policy(file, (void __user *)arg); + case CEPH_IOC_GET_DATALOC: return ceph_ioctl_get_dataloc(file, (void __user *)arg); case CEPH_IOC_LAZYIO: return ceph_ioctl_lazyio(file); } + return -ENOTTY; } diff --git a/fs/ceph/ioctl.h b/fs/ceph/ioctl.h index 88451a3..a6ce54e 100644 --- a/fs/ceph/ioctl.h +++ b/fs/ceph/ioctl.h @@ -4,7 +4,7 @@ #include #include -#define CEPH_IOCTL_MAGIC 0x97 +#define CEPH_IOCTL_MAGIC 0x98 /* just use u64 to align sanely on all archs */ struct ceph_ioctl_layout { @@ -17,6 +17,8 @@ struct ceph_ioctl_layout { struct ceph_ioctl_layout) #define CEPH_IOC_SET_LAYOUT _IOW(CEPH_IOCTL_MAGIC, 2, \ struct ceph_ioctl_layout) +#define CEPH_IOC_SET_LAYOUT_POLICY _IOW(CEPH_IOCTL_MAGIC, 5, \ + struct ceph_ioctl_layout) /* * Extract identity, address of the OSD and object storing a given diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index d5619ac..c3c74ae 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -299,6 +299,7 @@ enum { CEPH_MDS_OP_SETATTR = 0x01108, CEPH_MDS_OP_SETFILELOCK= 0x01109, CEPH_MDS_OP_GETFILELOCK= 0x00110, + CEPH_MDS_OP_SETDIRLAYOUT=0x0110a, CEPH_MDS_OP_MKNOD = 0x01201, CEPH_MDS_OP_LINK = 0x01202, -- cgit v0.10.2 From 240634e9b3a8ae2a9e2b7effeea3b9a5b30d68fb Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Tue, 5 Oct 2010 12:03:23 -0700 Subject: ceph: fix num_pages_free accounting in pagelist Decrement the free page counter when removing a page from the free_list. Signed-off-by: Sage Weil diff --git a/net/ceph/pagelist.c b/net/ceph/pagelist.c index b8cbc45..13cb409 100644 --- a/net/ceph/pagelist.c +++ b/net/ceph/pagelist.c @@ -37,6 +37,7 @@ static int ceph_pagelist_addpage(struct ceph_pagelist *pl) } else { page = list_first_entry(&pl->free_list, struct page, lru); list_del(&page->lru); + --pl->num_pages_free; } if (!page) return -ENOMEM; -- cgit v0.10.2 From b8d0638a98aa4a42ff322234b882487cd74e5c52 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Mon, 11 Oct 2010 21:14:23 +0200 Subject: rbd: null vs ERR_PTR ceph_alloc_page_vector() returns ERR_PTR(-ENOMEM) on errors. Signed-off-by: Dan Carpenter Signed-off-by: Sage Weil diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 1ac87f1..52f9420 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -826,8 +826,8 @@ static int rbd_req_sync_op(struct rbd_device *dev, num_pages = calc_pages_for(ofs , len); pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL); - if (!pages) - return -ENOMEM; + if (IS_ERR(pages)) + return PTR_ERR(pages); if (!orig_ops) { payload_len = (flags & CEPH_OSD_FLAG_WRITE ? len : 0); -- cgit v0.10.2 From 85b5aaa624aac568b8a3a88dbe4de6628c7cc527 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Mon, 11 Oct 2010 21:15:11 +0200 Subject: rbd: passing wrong variable to bvec_kunmap_irq() We should be passing "buf" here insead of "bv". This is tricky because it's not the same as kmap() and kunmap(). GCC does warn about it if you compile on i386 with CONFIG_HIGHMEM. Signed-off-by: Dan Carpenter Signed-off-by: Sage Weil diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 52f9420..6ec9d53 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -554,7 +554,7 @@ static void zero_bio_chain(struct bio *chain, int start_ofs) buf = bvec_kmap_irq(bv, &flags); memset(buf + remainder, 0, bv->bv_len - remainder); - bvec_kunmap_irq(bv, &flags); + bvec_kunmap_irq(buf, &flags); } pos += bv->bv_len; } -- cgit v0.10.2 From 61413c2f594e6b63db2b14c70c2e7d8cf02f9c00 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Sun, 17 Oct 2010 21:55:21 +0200 Subject: fs/ceph/xattr.c: Use kmemdup Convert a sequence of kmalloc and memcpy to use kmemdup. The semantic patch that performs this transformation is: (http://coccinelle.lip6.fr/) // @@ expression a,flag,len; expression arg,e1,e2; statement S; @@ a = - \(kmalloc\|kzalloc\)(len,flag) + kmemdup(arg,len,flag) <... when != a if (a == NULL || ...) S ...> - memcpy(a,arg,len+1); // Signed-off-by: Julia Lawall Signed-off-by: Sage Weil diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 70e9199..6e12a6b 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -716,10 +716,9 @@ int ceph_setxattr(struct dentry *dentry, const char *name, /* preallocate memory for xattr name, value, index node */ err = -ENOMEM; - newname = kmalloc(name_len + 1, GFP_NOFS); + newname = kmemdup(name, name_len + 1, GFP_NOFS); if (!newname) goto out; - memcpy(newname, name, name_len + 1); if (val_len) { newval = kmalloc(val_len + 1, GFP_NOFS); -- cgit v0.10.2 From efa4c1206eaff047c474af2136748a58eb8cc33b Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Mon, 18 Oct 2010 14:04:31 -0700 Subject: ceph: do not carry i_lock for readdir from dcache We were taking dcache_lock inside of i_lock, which introduces a dependency not found elsewhere in the kernel, complicationg the vfs locking scalability work. Since we don't actually need it here anyway, remove it. We only need i_lock to test for the I_COMPLETE flag, so be careful to do so without dcache_lock held. Signed-off-by: Sage Weil diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 4958252..e0a2dc6 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -95,10 +95,7 @@ static unsigned fpos_off(loff_t p) */ static int __dcache_readdir(struct file *filp, void *dirent, filldir_t filldir) - __releases(inode->i_lock) - __acquires(inode->i_lock) { - struct inode *inode = filp->f_dentry->d_inode; struct ceph_file_info *fi = filp->private_data; struct dentry *parent = filp->f_dentry; struct inode *dir = parent->d_inode; @@ -154,7 +151,6 @@ more: atomic_inc(&dentry->d_count); spin_unlock(&dcache_lock); - spin_unlock(&inode->i_lock); dout(" %llu (%llu) dentry %p %.*s %p\n", di->offset, filp->f_pos, dentry, dentry->d_name.len, dentry->d_name.name, dentry->d_inode); @@ -172,35 +168,30 @@ more: } else { dput(last); } - last = NULL; } - - spin_lock(&inode->i_lock); - spin_lock(&dcache_lock); - last = dentry; if (err < 0) - goto out_unlock; + goto out; - p = p->prev; filp->f_pos++; /* make sure a dentry wasn't dropped while we didn't have dcache_lock */ - if ((ceph_inode(dir)->i_ceph_flags & CEPH_I_COMPLETE)) - goto more; - dout(" lost I_COMPLETE on %p; falling back to mds\n", dir); - err = -EAGAIN; + if (!ceph_i_test(dir, CEPH_I_COMPLETE)) { + dout(" lost I_COMPLETE on %p; falling back to mds\n", dir); + err = -EAGAIN; + goto out; + } + + spin_lock(&dcache_lock); + p = p->prev; /* advance to next dentry */ + goto more; out_unlock: spin_unlock(&dcache_lock); - - if (last) { - spin_unlock(&inode->i_lock); +out: + if (last) dput(last); - spin_lock(&inode->i_lock); - } - return err; } @@ -272,13 +263,13 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir) ceph_snap(inode) != CEPH_SNAPDIR && (ci->i_ceph_flags & CEPH_I_COMPLETE) && __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) { + spin_unlock(&inode->i_lock); err = __dcache_readdir(filp, dirent, filldir); - if (err != -EAGAIN) { - spin_unlock(&inode->i_lock); + if (err != -EAGAIN) return err; - } + } else { + spin_unlock(&inode->i_lock); } - spin_unlock(&inode->i_lock); if (fi->dentry) { err = note_last_dentry(fi, fi->dentry->d_name.name, fi->dentry->d_name.len); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index e2e9044..1886294 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -400,8 +400,9 @@ static inline bool ceph_i_test(struct inode *inode, unsigned mask) struct ceph_inode_info *ci = ceph_inode(inode); bool r; - smp_mb(); + spin_lock(&inode->i_lock); r = (ci->i_ceph_flags & mask) == mask; + spin_unlock(&inode->i_lock); return r; } -- cgit v0.10.2 From 66f2b061546974b96b7b238a92ce89a87ecf0754 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Wed, 20 Oct 2010 15:55:35 -0700 Subject: x86, mm: Enable ARCH_DMA_ADDR_T_64BIT with X86_64 || HIGHMEM64G Set CONFIG_ARCH_DMA_ADDR_T_64BIT when we set dma_addr_t to 64 bits in ; this allows Kconfig decisions based on this property. Signed-off-by: FUJITA Tomonori LKML-Reference: <201010202255.o9KMtZXu009370@imap1.linux-foundation.org> Acked-by: "H. Peter Anvin" Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..2924f4e 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1148,6 +1148,9 @@ config X86_PAE config ARCH_PHYS_ADDR_T_64BIT def_bool X86_64 || X86_PAE +config ARCH_DMA_ADDR_T_64BIT + def_bool X86_64 || HIGHMEM64G + config DIRECT_GBPAGES bool "Enable 1GB pages for kernel pagetables" if EMBEDDED default y -- cgit v0.10.2 From 68eda8f59081c74a51d037cc29893bd7c9b3c2d8 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Sun, 8 Aug 2010 00:17:51 +0200 Subject: tomoyo: cleanup. don't store bogus pointer If domain is NULL then &domain->list is a bogus address. Let's leave head->r.domain NULL instead of saving an unusable pointer. This is just a cleanup. The current code always checks head->r.eof before dereferencing head->r.domain. Signed-off-by: Dan Carpenter Acked-by: Tetsuo Handa diff --git a/security/tomoyo/common.c b/security/tomoyo/common.c index c668b44..e0a1059 100644 --- a/security/tomoyo/common.c +++ b/security/tomoyo/common.c @@ -768,8 +768,10 @@ static bool tomoyo_select_one(struct tomoyo_io_buffer *head, const char *data) return true; /* Do nothing if open(O_WRONLY). */ memset(&head->r, 0, sizeof(head->r)); head->r.print_this_domain_only = true; - head->r.eof = !domain; - head->r.domain = &domain->list; + if (domain) + head->r.domain = &domain->list; + else + head->r.eof = 1; tomoyo_io_printf(head, "# select %s\n", data); if (domain && domain->is_deleted) tomoyo_io_printf(head, "# This is a deleted domain.\n"); -- cgit v0.10.2 From daa6d83a2863c28197b0c7dabfdf1e0606760b78 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Tue, 3 Aug 2010 15:26:05 -0400 Subject: selinux: type_bounds_sanity_check has a meaningless variable declaration type is not used at all, stop declaring and assigning it. Signed-off-by: Eric Paris Acked-by: Stephen Smalley Signed-off-by: James Morris diff --git a/security/selinux/ss/policydb.c b/security/selinux/ss/policydb.c index 3a29704..aa5a2fd 100644 --- a/security/selinux/ss/policydb.c +++ b/security/selinux/ss/policydb.c @@ -1624,11 +1624,11 @@ static int role_bounds_sanity_check(void *key, void *datum, void *datap) static int type_bounds_sanity_check(void *key, void *datum, void *datap) { - struct type_datum *upper, *type; + struct type_datum *upper; struct policydb *p = datap; int depth = 0; - upper = type = datum; + upper = datum; while (upper->bounds) { if (++depth == POLICYDB_BOUNDS_MAXDEPTH) { printk(KERN_ERR "SELinux: type %s: " -- cgit v0.10.2 From 065d78a0603cc6f8d288e96dbf761b96984b634f Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Sat, 28 Aug 2010 14:58:44 +0900 Subject: LSM: Fix security_module_enable() error. We can set default LSM module to DAC (which means "enable no LSM module"). If default LSM module was set to DAC, security_module_enable() must return 0 unless overridden via boot time parameter. Signed-off-by: Tetsuo Handa Acked-by: Serge E. Hallyn Signed-off-by: James Morris diff --git a/security/security.c b/security/security.c index c53949f..43b6463 100644 --- a/security/security.c +++ b/security/security.c @@ -89,20 +89,12 @@ __setup("security=", choose_lsm); * Return true if: * -The passed LSM is the one chosen by user at boot time, * -or the passed LSM is configured as the default and the user did not - * choose an alternate LSM at boot time, - * -or there is no default LSM set and the user didn't specify a - * specific LSM and we're the first to ask for registration permission, - * -or the passed LSM is currently loaded. + * choose an alternate LSM at boot time. * Otherwise, return false. */ int __init security_module_enable(struct security_operations *ops) { - if (!*chosen_lsm) - strncpy(chosen_lsm, ops->name, SECURITY_NAME_MAX); - else if (strncmp(ops->name, chosen_lsm, SECURITY_NAME_MAX)) - return 0; - - return 1; + return !strcmp(ops->name, chosen_lsm); } /** -- cgit v0.10.2 From 4b04a7cfc5ccb573ca3752429c81d37f8dd2f7c6 Mon Sep 17 00:00:00 2001 From: Yong Zhang Date: Sat, 28 Aug 2010 10:25:09 +0800 Subject: .gitignore: ignore apparmor/rlim_names.h Signed-off-by: Yong Zhang Signed-off-by: John Johansen Signed-off-by: James Morris diff --git a/security/apparmor/.gitignore b/security/apparmor/.gitignore index 0a0a99f..4d995ae 100644 --- a/security/apparmor/.gitignore +++ b/security/apparmor/.gitignore @@ -3,3 +3,4 @@ # af_names.h capability_names.h +rlim_names.h -- cgit v0.10.2 From 119041672592d1890d89dd8f194bd0919d801dc8 Mon Sep 17 00:00:00 2001 From: KaiGai Kohei Date: Tue, 14 Sep 2010 18:28:39 +0900 Subject: selinux: fast status update interface (/selinux/status) This patch provides a new /selinux/status entry which allows applications read-only mmap(2). This region reflects selinux_kernel_status structure in kernel space. struct selinux_kernel_status { u32 length; /* length of this structure */ u32 sequence; /* sequence number of seqlock logic */ u32 enforcing; /* current setting of enforcing mode */ u32 policyload; /* times of policy reloaded */ u32 deny_unknown; /* current setting of deny_unknown */ }; When userspace object manager caches access control decisions provided by SELinux, it needs to invalidate the cache on policy reload and setenforce to keep consistency. However, the applications need to check the kernel state for each accesses on userspace avc, or launch a background worker process. In heuristic, frequency of invalidation is much less than frequency of making access control decision, so it is annoying to invoke a system call to check we don't need to invalidate the userspace cache. If we can use a background worker thread, it allows to receive invalidation messages from the kernel. But it requires us an invasive coding toward the base application in some cases; E.g, when we provide a feature performing with SELinux as a plugin module, it is unwelcome manner to launch its own worker thread from the module. If we could map /selinux/status to process memory space, application can know updates of selinux status; policy reload or setenforce. A typical application checks selinux_kernel_status::sequence when it tries to reference userspace avc. If it was changed from the last time when it checked userspace avc, it means something was updated in the kernel space. Then, the application can reset userspace avc or update current enforcing mode, without any system call invocations. This sequence number is updated according to the seqlock logic, so we need to wait for a while if it is odd number. Signed-off-by: KaiGai Kohei Acked-by: Eric Paris -- security/selinux/include/security.h | 21 ++++++ security/selinux/selinuxfs.c | 56 +++++++++++++++ security/selinux/ss/Makefile | 2 +- security/selinux/ss/services.c | 3 + security/selinux/ss/status.c | 129 +++++++++++++++++++++++++++++++++++ 5 files changed, 210 insertions(+), 1 deletions(-) Signed-off-by: James Morris diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h index 1f7c249..e390e31 100644 --- a/security/selinux/include/security.h +++ b/security/selinux/include/security.h @@ -191,5 +191,26 @@ static inline int security_netlbl_sid_to_secattr(u32 sid, const char *security_get_initial_sid_context(u32 sid); +/* + * status notifier using mmap interface + */ +extern struct page *selinux_kernel_status_page(void); + +#define SELINUX_KERNEL_STATUS_VERSION 1 +struct selinux_kernel_status +{ + u32 version; /* version number of thie structure */ + u32 sequence; /* sequence number of seqlock logic */ + u32 enforcing; /* current setting of enforcing mode */ + u32 policyload; /* times of policy reloaded */ + u32 deny_unknown; /* current setting of deny_unknown */ + /* + * The version > 0 supports above members. + */ +} __attribute__((packed)); + +extern void selinux_status_update_setenforce(int enforcing); +extern void selinux_status_update_policyload(int seqno); + #endif /* _SELINUX_SECURITY_H_ */ diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index 79a1bb6..a2e7a85 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -110,6 +110,7 @@ enum sel_inos { SEL_COMPAT_NET, /* whether to use old compat network packet controls */ SEL_REJECT_UNKNOWN, /* export unknown reject handling to userspace */ SEL_DENY_UNKNOWN, /* export unknown deny handling to userspace */ + SEL_STATUS, /* export current status using mmap() */ SEL_INO_NEXT, /* The next inode number to use */ }; @@ -171,6 +172,7 @@ static ssize_t sel_write_enforce(struct file *file, const char __user *buf, if (selinux_enforcing) avc_ss_reset(0); selnl_notify_setenforce(selinux_enforcing); + selinux_status_update_setenforce(selinux_enforcing); } length = count; out: @@ -205,6 +207,59 @@ static const struct file_operations sel_handle_unknown_ops = { .llseek = generic_file_llseek, }; +static int sel_open_handle_status(struct inode *inode, struct file *filp) +{ + struct page *status = selinux_kernel_status_page(); + + if (!status) + return -ENOMEM; + + filp->private_data = status; + + return 0; +} + +static ssize_t sel_read_handle_status(struct file *filp, char __user *buf, + size_t count, loff_t *ppos) +{ + struct page *status = filp->private_data; + + BUG_ON(!status); + + return simple_read_from_buffer(buf, count, ppos, + page_address(status), + sizeof(struct selinux_kernel_status)); +} + +static int sel_mmap_handle_status(struct file *filp, + struct vm_area_struct *vma) +{ + struct page *status = filp->private_data; + unsigned long size = vma->vm_end - vma->vm_start; + + BUG_ON(!status); + + /* only allows one page from the head */ + if (vma->vm_pgoff > 0 || size != PAGE_SIZE) + return -EIO; + /* disallow writable mapping */ + if (vma->vm_flags & VM_WRITE) + return -EPERM; + /* disallow mprotect() turns it into writable */ + vma->vm_flags &= ~VM_MAYWRITE; + + return remap_pfn_range(vma, vma->vm_start, + page_to_pfn(status), + size, vma->vm_page_prot); +} + +static const struct file_operations sel_handle_status_ops = { + .open = sel_open_handle_status, + .read = sel_read_handle_status, + .mmap = sel_mmap_handle_status, + .llseek = generic_file_llseek, +}; + #ifdef CONFIG_SECURITY_SELINUX_DISABLE static ssize_t sel_write_disable(struct file *file, const char __user *buf, size_t count, loff_t *ppos) @@ -1612,6 +1667,7 @@ static int sel_fill_super(struct super_block *sb, void *data, int silent) [SEL_CHECKREQPROT] = {"checkreqprot", &sel_checkreqprot_ops, S_IRUGO|S_IWUSR}, [SEL_REJECT_UNKNOWN] = {"reject_unknown", &sel_handle_unknown_ops, S_IRUGO}, [SEL_DENY_UNKNOWN] = {"deny_unknown", &sel_handle_unknown_ops, S_IRUGO}, + [SEL_STATUS] = {"status", &sel_handle_status_ops, S_IRUGO}, /* last one */ {""} }; ret = simple_fill_super(sb, SELINUX_MAGIC, selinux_files); diff --git a/security/selinux/ss/Makefile b/security/selinux/ss/Makefile index 15d4e62..974e11c 100644 --- a/security/selinux/ss/Makefile +++ b/security/selinux/ss/Makefile @@ -5,5 +5,5 @@ EXTRA_CFLAGS += -Isecurity/selinux -Isecurity/selinux/include obj-y := ss.o -ss-y := ebitmap.o hashtab.o symtab.o sidtab.o avtab.o policydb.o services.o conditional.o mls.o +ss-y := ebitmap.o hashtab.o symtab.o sidtab.o avtab.o policydb.o services.o conditional.o mls.o status.o diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c index 9ea2fec..494ff52 100644 --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -1791,6 +1791,7 @@ int security_load_policy(void *data, size_t len) selinux_complete_init(); avc_ss_reset(seqno); selnl_notify_policyload(seqno); + selinux_status_update_policyload(seqno); selinux_netlbl_cache_invalidate(); selinux_xfrm_notify_policyload(); return 0; @@ -1870,6 +1871,7 @@ int security_load_policy(void *data, size_t len) avc_ss_reset(seqno); selnl_notify_policyload(seqno); + selinux_status_update_policyload(seqno); selinux_netlbl_cache_invalidate(); selinux_xfrm_notify_policyload(); @@ -2374,6 +2376,7 @@ out: if (!rc) { avc_ss_reset(seqno); selnl_notify_policyload(seqno); + selinux_status_update_policyload(seqno); selinux_xfrm_notify_policyload(); } return rc; diff --git a/security/selinux/ss/status.c b/security/selinux/ss/status.c new file mode 100644 index 0000000..5d9b225 --- /dev/null +++ b/security/selinux/ss/status.c @@ -0,0 +1,129 @@ +/* + * mmap based event notifications for SELinux + * + * Author: KaiGai Kohei + * + * Copyright (C) 2010 NEC corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2, + * as published by the Free Software Foundation. + */ +#include +#include +#include +#include +#include "avc.h" +#include "services.h" + +/* + * The selinux_status_page shall be exposed to userspace applications + * using mmap interface on /selinux/status. + * It enables to notify applications a few events that will cause reset + * of userspace access vector without context switching. + * + * The selinux_kernel_status structure on the head of status page is + * protected from concurrent accesses using seqlock logic, so userspace + * application should reference the status page according to the seqlock + * logic. + * + * Typically, application checks status->sequence at the head of access + * control routine. If it is odd-number, kernel is updating the status, + * so please wait for a moment. If it is changed from the last sequence + * number, it means something happen, so application will reset userspace + * avc, if needed. + * In most cases, application shall confirm the kernel status is not + * changed without any system call invocations. + */ +static struct page *selinux_status_page = NULL; +static DEFINE_MUTEX(selinux_status_lock); + +/* + * selinux_kernel_status_page + * + * It returns a reference to selinux_status_page. If the status page is + * not allocated yet, it also tries to allocate it at the first time. + */ +struct page *selinux_kernel_status_page(void) +{ + struct selinux_kernel_status *status; + struct page *result = NULL; + + mutex_lock(&selinux_status_lock); + if (!selinux_status_page) + { + selinux_status_page = alloc_page(GFP_KERNEL|__GFP_ZERO); + if (selinux_status_page) + { + status = page_address(selinux_status_page); + + status->version = SELINUX_KERNEL_STATUS_VERSION; + status->sequence = 0; + status->enforcing = selinux_enforcing; + /* + * NOTE: the next policyload event shall set + * a positive value on the status->policyload, + * although it may not be 1, but never zero. + * So, application can know it was updated. + */ + status->policyload = 0; + status->deny_unknown = !security_get_allow_unknown(); + } + } + result = selinux_status_page; + mutex_unlock(&selinux_status_lock); + + return result; +} + +/* + * selinux_status_update_setenforce + * + * It updates status of the current enforcing/permissive mode. + */ +void selinux_status_update_setenforce(int enforcing) +{ + struct selinux_kernel_status *status; + + mutex_lock(&selinux_status_lock); + if (selinux_status_page) + { + status = page_address(selinux_status_page); + + status->sequence++; + smp_wmb(); + + status->enforcing = enforcing; + + smp_wmb(); + status->sequence++; + } + mutex_unlock(&selinux_status_lock); +} + +/* + * selinux_status_update_policyload + * + * It updates status of the times of policy reloaded, and current + * setting of deny_unknown. + */ +void selinux_status_update_policyload(int seqno) +{ + struct selinux_kernel_status *status; + + mutex_lock(&selinux_status_lock); + if (selinux_status_page) + { + status = page_address(selinux_status_page); + + status->sequence++; + smp_wmb(); + + status->policyload = seqno; + status->deny_unknown = !security_get_allow_unknown(); + + smp_wmb(); + status->sequence++; + } + mutex_unlock(&selinux_status_lock); +} -- cgit v0.10.2 From ceba72a68d17ee36ef24a71b80dde39ee934ece8 Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Mon, 9 Aug 2010 17:34:25 -0400 Subject: selinux: fix parallel compile error Selinux has an autogenerated file, "flask.h" which is included by two other selinux files. The current makefile has a single dependency on the first object file in the selinux-y list, assuming that will get flask.h generated before anyone looks for it, but that assumption breaks down in a "make -jN" situation and you get: selinux/selinuxfs.c:35: fatal error: flask.h: No such file or directory compilation terminated. remake[9]: *** [security/selinux/selinuxfs.o] Error 1 Since flask.h is included by security.h which in turn is included nearly everywhere, make the dependency apply to all of the selinux-y list of objs. Signed-off-by: Paul Gortmaker Signed-off-by: James Morris diff --git a/security/selinux/Makefile b/security/selinux/Makefile index 58d80f3..c6fa30e 100644 --- a/security/selinux/Makefile +++ b/security/selinux/Makefile @@ -20,7 +20,7 @@ selinux-$(CONFIG_NETLABEL) += netlabel.o EXTRA_CFLAGS += -Isecurity/selinux -Isecurity/selinux/include -$(obj)/avc.o: $(obj)/flask.h +$(addprefix $(obj)/,$(selinux-y)): $(obj)/flask.h quiet_cmd_flask = GEN $(obj)/flask.h $(obj)/av_permissions.h cmd_flask = scripts/selinux/genheaders/genheaders $(obj)/flask.h $(obj)/av_permissions.h -- cgit v0.10.2 From 60272da0341e9eaa136e1dc072bfef72c995d851 Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Wed, 15 Sep 2010 20:14:53 -0400 Subject: selinux: really fix dependency causing parallel compile failure. While the previous change to the selinux Makefile reduced the window significantly for this failure, it is still possible to see a compile failure where cpp starts processing selinux files before the auto generated flask.h file is completed. This is easily reproduced by adding the following temporary change to expose the issue everytime: - cmd_flask = scripts/selinux/genheaders/genheaders ... + cmd_flask = sleep 30 ; scripts/selinux/genheaders/genheaders ... This failure happens because the creation of the object files in the ss subdir also depends on flask.h. So simply incorporate them into the parent Makefile, as the ss/Makefile really doesn't do anything unique. With this change, compiling of all selinux files is dependent on completion of the header file generation, and this test case with the "sleep 30" now confirms it is functioning as expected. Signed-off-by: Paul Gortmaker Signed-off-by: James Morris diff --git a/security/selinux/Makefile b/security/selinux/Makefile index c6fa30e..a30f218 100644 --- a/security/selinux/Makefile +++ b/security/selinux/Makefile @@ -2,17 +2,12 @@ # Makefile for building the SELinux module as part of the kernel tree. # -obj-$(CONFIG_SECURITY_SELINUX) := selinux.o ss/ - -selinux-y := avc.o \ - hooks.o \ - selinuxfs.o \ - netlink.o \ - nlmsgtab.o \ - netif.o \ - netnode.o \ - netport.o \ - exports.o +obj-$(CONFIG_SECURITY_SELINUX) := selinux.o + +selinux-y := avc.o hooks.o selinuxfs.o netlink.o nlmsgtab.o netif.o \ + netnode.o netport.o exports.o \ + ss/ebitmap.o ss/hashtab.o ss/symtab.o ss/sidtab.o ss/avtab.o \ + ss/policydb.o ss/services.o ss/conditional.o ss/mls.o ss/status.o selinux-$(CONFIG_SECURITY_NETWORK_XFRM) += xfrm.o diff --git a/security/selinux/ss/Makefile b/security/selinux/ss/Makefile deleted file mode 100644 index 974e11c..0000000 --- a/security/selinux/ss/Makefile +++ /dev/null @@ -1,9 +0,0 @@ -# -# Makefile for building the SELinux security server as part of the kernel tree. -# - -EXTRA_CFLAGS += -Isecurity/selinux -Isecurity/selinux/include -obj-y := ss.o - -ss-y := ebitmap.o hashtab.o symtab.o sidtab.o avtab.o policydb.o services.o conditional.o mls.o status.o - -- cgit v0.10.2 From 8b0c543e5cb1e47a54d3ea791b8a03b9c8a715db Mon Sep 17 00:00:00 2001 From: matt mooney Date: Wed, 22 Sep 2010 23:50:06 -0700 Subject: selinux: change to new flag variable Replace EXTRA_CFLAGS with ccflags-y. Signed-off-by: matt mooney Signed-off-by: James Morris diff --git a/security/selinux/Makefile b/security/selinux/Makefile index a30f218..ad5cd76 100644 --- a/security/selinux/Makefile +++ b/security/selinux/Makefile @@ -13,7 +13,7 @@ selinux-$(CONFIG_SECURITY_NETWORK_XFRM) += xfrm.o selinux-$(CONFIG_NETLABEL) += netlabel.o -EXTRA_CFLAGS += -Isecurity/selinux -Isecurity/selinux/include +ccflags-y := -Isecurity/selinux -Isecurity/selinux/include $(addprefix $(obj)/,$(selinux-y)): $(obj)/flask.h -- cgit v0.10.2 From 36f7f28416c97dbb725154930066d115b4447e17 Mon Sep 17 00:00:00 2001 From: KaiGai Kohei Date: Thu, 30 Sep 2010 11:49:55 +0900 Subject: selinux: fix up style problem on /selinux/status This patch fixes up coding-style problem at this commit: 4f27a7d49789b04404eca26ccde5f527231d01d5 selinux: fast status update interface (/selinux/status) Signed-off-by: KaiGai Kohei Signed-off-by: James Morris diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h index e390e31..4b66f19 100644 --- a/security/selinux/include/security.h +++ b/security/selinux/include/security.h @@ -197,8 +197,7 @@ const char *security_get_initial_sid_context(u32 sid); extern struct page *selinux_kernel_status_page(void); #define SELINUX_KERNEL_STATUS_VERSION 1 -struct selinux_kernel_status -{ +struct selinux_kernel_status { u32 version; /* version number of thie structure */ u32 sequence; /* sequence number of seqlock logic */ u32 enforcing; /* current setting of enforcing mode */ diff --git a/security/selinux/ss/status.c b/security/selinux/ss/status.c index 5d9b225..d982365 100644 --- a/security/selinux/ss/status.c +++ b/security/selinux/ss/status.c @@ -35,7 +35,7 @@ * In most cases, application shall confirm the kernel status is not * changed without any system call invocations. */ -static struct page *selinux_status_page = NULL; +static struct page *selinux_status_page; static DEFINE_MUTEX(selinux_status_lock); /* @@ -50,11 +50,10 @@ struct page *selinux_kernel_status_page(void) struct page *result = NULL; mutex_lock(&selinux_status_lock); - if (!selinux_status_page) - { + if (!selinux_status_page) { selinux_status_page = alloc_page(GFP_KERNEL|__GFP_ZERO); - if (selinux_status_page) - { + + if (selinux_status_page) { status = page_address(selinux_status_page); status->version = SELINUX_KERNEL_STATUS_VERSION; @@ -86,8 +85,7 @@ void selinux_status_update_setenforce(int enforcing) struct selinux_kernel_status *status; mutex_lock(&selinux_status_lock); - if (selinux_status_page) - { + if (selinux_status_page) { status = page_address(selinux_status_page); status->sequence++; @@ -112,8 +110,7 @@ void selinux_status_update_policyload(int seqno) struct selinux_kernel_status *status; mutex_lock(&selinux_status_lock); - if (selinux_status_page) - { + if (selinux_status_page) { status = page_address(selinux_status_page); status->sequence++; -- cgit v0.10.2 From 9b3056cca09529d34af2d81305b2a9c6b622ca1b Mon Sep 17 00:00:00 2001 From: Dmitry Torokhov Date: Fri, 1 Oct 2010 14:16:39 -0700 Subject: tpm: change 'tpm_suspend_pcr' to be module parameter Fix the following warning: drivers/char/tpm/tpm.c:1085: warning: `tpm_suspend_setup' defined but not used and make the workaround operable in case when TPM is compiled as a module. As a side-effect the option will be called tpm.suspend_pcr. Signed-off-by: Dmitry Torokhov Cc: Rajiv Andrade Cc: David Safford Cc: James Morris Cc: Debora Velarde Signed-off-by: Andrew Morton Signed-off-by: James Morris diff --git a/drivers/char/tpm/tpm.c b/drivers/char/tpm/tpm.c index 05ad4a1..7c41335 100644 --- a/drivers/char/tpm/tpm.c +++ b/drivers/char/tpm/tpm.c @@ -47,6 +47,16 @@ enum tpm_duration { #define TPM_MAX_PROTECTED_ORDINAL 12 #define TPM_PROTECTED_ORDINAL_MASK 0xFF +/* + * Bug workaround - some TPM's don't flush the most + * recently changed pcr on suspend, so force the flush + * with an extend to the selected _unused_ non-volatile pcr. + */ +static int tpm_suspend_pcr; +module_param_named(suspend_pcr, tpm_suspend_pcr, uint, 0644); +MODULE_PARM_DESC(suspend_pcr, + "PCR to use for dummy writes to faciltate flush on suspend."); + static LIST_HEAD(tpm_chip_list); static DEFINE_SPINLOCK(driver_lock); static DECLARE_BITMAP(dev_mask, TPM_NUM_DEVICES); @@ -1077,18 +1087,6 @@ static struct tpm_input_header savestate_header = { .ordinal = TPM_ORD_SAVESTATE }; -/* Bug workaround - some TPM's don't flush the most - * recently changed pcr on suspend, so force the flush - * with an extend to the selected _unused_ non-volatile pcr. - */ -static int tpm_suspend_pcr; -static int __init tpm_suspend_setup(char *str) -{ - get_option(&str, &tpm_suspend_pcr); - return 1; -} -__setup("tpm_suspend_pcr=", tpm_suspend_setup); - /* * We are about to suspend. Save the TPM state * so that it can be restored. -- cgit v0.10.2 From b0ae19811375031ae3b3fecc65b702a9c6e5cc28 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 15 Oct 2010 04:21:18 +0900 Subject: security: remove unused parameter from security_task_setscheduler() All security modules shouldn't change sched_param parameter of security_task_setscheduler(). This is not only meaningless, but also make a harmful result if caller pass a static variable. This patch remove policy and sched_param parameter from security_task_setscheduler() becuase none of security module is using it. Cc: James Morris Signed-off-by: KOSAKI Motohiro Signed-off-by: James Morris diff --git a/arch/mips/kernel/mips-mt-fpaff.c b/arch/mips/kernel/mips-mt-fpaff.c index 2340f11..9a526ba 100644 --- a/arch/mips/kernel/mips-mt-fpaff.c +++ b/arch/mips/kernel/mips-mt-fpaff.c @@ -103,7 +103,7 @@ asmlinkage long mipsmt_sys_sched_setaffinity(pid_t pid, unsigned int len, if (!check_same_owner(p) && !capable(CAP_SYS_NICE)) goto out_unlock; - retval = security_task_setscheduler(p, 0, NULL); + retval = security_task_setscheduler(p) if (retval) goto out_unlock; diff --git a/include/linux/security.h b/include/linux/security.h index a22219a..294a0b2 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -74,7 +74,7 @@ extern int cap_file_mmap(struct file *file, unsigned long reqprot, extern int cap_task_fix_setuid(struct cred *new, const struct cred *old, int flags); extern int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5); -extern int cap_task_setscheduler(struct task_struct *p, int policy, struct sched_param *lp); +extern int cap_task_setscheduler(struct task_struct *p); extern int cap_task_setioprio(struct task_struct *p, int ioprio); extern int cap_task_setnice(struct task_struct *p, int nice); extern int cap_syslog(int type, bool from_file); @@ -1501,8 +1501,7 @@ struct security_operations { int (*task_getioprio) (struct task_struct *p); int (*task_setrlimit) (struct task_struct *p, unsigned int resource, struct rlimit *new_rlim); - int (*task_setscheduler) (struct task_struct *p, int policy, - struct sched_param *lp); + int (*task_setscheduler) (struct task_struct *p); int (*task_getscheduler) (struct task_struct *p); int (*task_movememory) (struct task_struct *p); int (*task_kill) (struct task_struct *p, @@ -1752,8 +1751,7 @@ int security_task_setioprio(struct task_struct *p, int ioprio); int security_task_getioprio(struct task_struct *p); int security_task_setrlimit(struct task_struct *p, unsigned int resource, struct rlimit *new_rlim); -int security_task_setscheduler(struct task_struct *p, - int policy, struct sched_param *lp); +int security_task_setscheduler(struct task_struct *p); int security_task_getscheduler(struct task_struct *p); int security_task_movememory(struct task_struct *p); int security_task_kill(struct task_struct *p, struct siginfo *info, @@ -2320,11 +2318,9 @@ static inline int security_task_setrlimit(struct task_struct *p, return 0; } -static inline int security_task_setscheduler(struct task_struct *p, - int policy, - struct sched_param *lp) +static inline int security_task_setscheduler(struct task_struct *p) { - return cap_task_setscheduler(p, policy, lp); + return cap_task_setscheduler(p); } static inline int security_task_getscheduler(struct task_struct *p) diff --git a/kernel/cpuset.c b/kernel/cpuset.c index b23c097..51b143e 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1397,7 +1397,7 @@ static int cpuset_can_attach(struct cgroup_subsys *ss, struct cgroup *cont, if (tsk->flags & PF_THREAD_BOUND) return -EINVAL; - ret = security_task_setscheduler(tsk, 0, NULL); + ret = security_task_setscheduler(tsk); if (ret) return ret; if (threadgroup) { @@ -1405,7 +1405,7 @@ static int cpuset_can_attach(struct cgroup_subsys *ss, struct cgroup *cont, rcu_read_lock(); list_for_each_entry_rcu(c, &tsk->thread_group, thread_group) { - ret = security_task_setscheduler(c, 0, NULL); + ret = security_task_setscheduler(c); if (ret) { rcu_read_unlock(); return ret; diff --git a/kernel/sched.c b/kernel/sched.c index dc85ceb..df6579d 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4645,7 +4645,7 @@ recheck: } if (user) { - retval = security_task_setscheduler(p, policy, param); + retval = security_task_setscheduler(p); if (retval) return retval; } @@ -4887,7 +4887,7 @@ long sched_setaffinity(pid_t pid, const struct cpumask *in_mask) if (!check_same_owner(p) && !capable(CAP_SYS_NICE)) goto out_unlock; - retval = security_task_setscheduler(p, 0, NULL); + retval = security_task_setscheduler(p); if (retval) goto out_unlock; diff --git a/security/commoncap.c b/security/commoncap.c index 9d172e6..5e632b4 100644 --- a/security/commoncap.c +++ b/security/commoncap.c @@ -719,14 +719,11 @@ static int cap_safe_nice(struct task_struct *p) /** * cap_task_setscheduler - Detemine if scheduler policy change is permitted * @p: The task to affect - * @policy: The policy to effect - * @lp: The parameters to the scheduling policy * * Detemine if the requested scheduler policy change is permitted for the * specified task, returning 0 if permission is granted, -ve if denied. */ -int cap_task_setscheduler(struct task_struct *p, int policy, - struct sched_param *lp) +int cap_task_setscheduler(struct task_struct *p) { return cap_safe_nice(p); } diff --git a/security/security.c b/security/security.c index 43b6463..1cbcdfa 100644 --- a/security/security.c +++ b/security/security.c @@ -778,10 +778,9 @@ int security_task_setrlimit(struct task_struct *p, unsigned int resource, return security_ops->task_setrlimit(p, resource, new_rlim); } -int security_task_setscheduler(struct task_struct *p, - int policy, struct sched_param *lp) +int security_task_setscheduler(struct task_struct *p) { - return security_ops->task_setscheduler(p, policy, lp); + return security_ops->task_setscheduler(p); } int security_task_getscheduler(struct task_struct *p) diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 4796ddd..db2b331 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -3354,11 +3354,11 @@ static int selinux_task_setrlimit(struct task_struct *p, unsigned int resource, return 0; } -static int selinux_task_setscheduler(struct task_struct *p, int policy, struct sched_param *lp) +static int selinux_task_setscheduler(struct task_struct *p) { int rc; - rc = cap_task_setscheduler(p, policy, lp); + rc = cap_task_setscheduler(p); if (rc) return rc; diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index c448d57..174aec4 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -1281,12 +1281,11 @@ static int smack_task_getioprio(struct task_struct *p) * * Return 0 if read access is permitted */ -static int smack_task_setscheduler(struct task_struct *p, int policy, - struct sched_param *lp) +static int smack_task_setscheduler(struct task_struct *p) { int rc; - rc = cap_task_setscheduler(p, policy, lp); + rc = cap_task_setscheduler(p); if (rc == 0) rc = smk_curacc_on_task(p, MAY_WRITE); return rc; -- cgit v0.10.2 From 9f1c1d426b0402b25cd0d7ca719ffc8e20e46d5f Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Fri, 8 Oct 2010 14:43:22 +0900 Subject: TOMOYO: Print URL information before panic(). Configuration files for TOMOYO 2.3 are not compatible with TOMOYO 2.2. But current panic() message is too unfriendly and is confusing users. Signed-off-by: Tetsuo Handa Reviewed-by: KOSAKI Motohiro Signed-off-by: James Morris diff --git a/security/tomoyo/common.c b/security/tomoyo/common.c index e0a1059..7556315 100644 --- a/security/tomoyo/common.c +++ b/security/tomoyo/common.c @@ -2053,13 +2053,22 @@ void tomoyo_check_profile(void) const u8 profile = domain->profile; if (tomoyo_profile_ptr[profile]) continue; + printk(KERN_ERR "You need to define profile %u before using it.\n", + profile); + printk(KERN_ERR "Please see http://tomoyo.sourceforge.jp/2.3/ " + "for more information.\n"); panic("Profile %u (used by '%s') not defined.\n", profile, domain->domainname->name); } tomoyo_read_unlock(idx); - if (tomoyo_profile_version != 20090903) + if (tomoyo_profile_version != 20090903) { + printk(KERN_ERR "You need to install userland programs for " + "TOMOYO 2.3 and initialize policy configuration.\n"); + printk(KERN_ERR "Please see http://tomoyo.sourceforge.jp/2.3/ " + "for more information.\n"); panic("Profile version %u is not supported.\n", tomoyo_profile_version); + } printk(KERN_INFO "TOMOYO: 2.3.0\n"); printk(KERN_INFO "Mandatory Access Control activated.\n"); } -- cgit v0.10.2 From 3ed02ada2a5e695e2fbb5e4a0008cfcb0f50feaa Mon Sep 17 00:00:00 2001 From: John Johansen Date: Sat, 9 Oct 2010 00:47:53 -0700 Subject: AppArmor: Ensure the size of the copy is < the buffer allocated to hold it Actually I think in this case the appropriate thing to do is to BUG as there is currently a case (remove) where the alloc_size needs to be larger than the copy_size, and if copy_size is ever greater than alloc_size there is a mistake in the caller code. Signed-off-by: John Johansen Acked-by: Kees Cook Signed-off-by: James Morris diff --git a/security/apparmor/apparmorfs.c b/security/apparmor/apparmorfs.c index 7320331..544ff58 100644 --- a/security/apparmor/apparmorfs.c +++ b/security/apparmor/apparmorfs.c @@ -29,7 +29,7 @@ * aa_simple_write_to_buffer - common routine for getting policy from user * @op: operation doing the user buffer copy * @userbuf: user buffer to copy data from (NOT NULL) - * @alloc_size: size of user buffer + * @alloc_size: size of user buffer (REQUIRES: @alloc_size >= @copy_size) * @copy_size: size of data to copy from user buffer * @pos: position write is at in the file (NOT NULL) * @@ -42,6 +42,8 @@ static char *aa_simple_write_to_buffer(int op, const char __user *userbuf, { char *data; + BUG_ON(copy_size > alloc_size); + if (*pos != 0) /* only writes from pos 0, that is complete writes */ return ERR_PTR(-ESPIPE); -- cgit v0.10.2 From 15714f7b58011cf3948cab2988abea560240c74f Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Tue, 12 Oct 2010 11:40:08 -0400 Subject: secmark: do not return early if there was no error Commit 4a5a5c73 attempted to pass decent error messages back to userspace for netfilter errors. In xt_SECMARK.c however the patch screwed up and returned on 0 (aka no error) early and didn't finish setting up secmark. This results in a kernel BUG if you use SECMARK. Signed-off-by: Eric Paris Acked-by: Paul Moore Signed-off-by: James Morris diff --git a/net/netfilter/xt_SECMARK.c b/net/netfilter/xt_SECMARK.c index 23b2d6c..364ad16 100644 --- a/net/netfilter/xt_SECMARK.c +++ b/net/netfilter/xt_SECMARK.c @@ -101,7 +101,7 @@ static int secmark_tg_check(const struct xt_tgchk_param *par) switch (info->mode) { case SECMARK_MODE_SEL: err = checkentry_selinux(info); - if (err <= 0) + if (err) return err; break; -- cgit v0.10.2 From 2606fd1fa5710205b23ee859563502aa18362447 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 16:24:41 -0400 Subject: secmark: make secmark object handling generic Right now secmark has lots of direct selinux calls. Use all LSM calls and remove all SELinux specific knowledge. The only SELinux specific knowledge we leave is the mode. The only point is to make sure that other LSMs at least test this generic code before they assume it works. (They may also have to make changes if they do not represent labels as strings) Signed-off-by: Eric Paris Acked-by: Paul Moore Acked-by: Patrick McHardy Signed-off-by: James Morris diff --git a/include/linux/netfilter/xt_SECMARK.h b/include/linux/netfilter/xt_SECMARK.h index 6fcd344..989092b 100644 --- a/include/linux/netfilter/xt_SECMARK.h +++ b/include/linux/netfilter/xt_SECMARK.h @@ -11,18 +11,12 @@ * packets are being marked for. */ #define SECMARK_MODE_SEL 0x01 /* SELinux */ -#define SECMARK_SELCTX_MAX 256 - -struct xt_secmark_target_selinux_info { - __u32 selsid; - char selctx[SECMARK_SELCTX_MAX]; -}; +#define SECMARK_SECCTX_MAX 256 struct xt_secmark_target_info { __u8 mode; - union { - struct xt_secmark_target_selinux_info sel; - } u; + __u32 secid; + char secctx[SECMARK_SECCTX_MAX]; }; #endif /*_XT_SECMARK_H_target */ diff --git a/include/linux/security.h b/include/linux/security.h index 294a0b2..d70adc3 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -959,6 +959,12 @@ static inline void security_free_mnt_opts(struct security_mnt_opts *opts) * Sets the new child socket's sid to the openreq sid. * @inet_conn_established: * Sets the connection's peersid to the secmark on skb. + * @secmark_relabel_packet: + * check if the process should be allowed to relabel packets to the given secid + * @security_secmark_refcount_inc + * tells the LSM to increment the number of secmark labeling rules loaded + * @security_secmark_refcount_dec + * tells the LSM to decrement the number of secmark labeling rules loaded * @req_classify_flow: * Sets the flow's sid to the openreq sid. * @tun_dev_create: @@ -1593,6 +1599,9 @@ struct security_operations { struct request_sock *req); void (*inet_csk_clone) (struct sock *newsk, const struct request_sock *req); void (*inet_conn_established) (struct sock *sk, struct sk_buff *skb); + int (*secmark_relabel_packet) (u32 secid); + void (*secmark_refcount_inc) (void); + void (*secmark_refcount_dec) (void); void (*req_classify_flow) (const struct request_sock *req, struct flowi *fl); int (*tun_dev_create)(void); void (*tun_dev_post_create)(struct sock *sk); @@ -2547,6 +2556,9 @@ void security_inet_csk_clone(struct sock *newsk, const struct request_sock *req); void security_inet_conn_established(struct sock *sk, struct sk_buff *skb); +int security_secmark_relabel_packet(u32 secid); +void security_secmark_refcount_inc(void); +void security_secmark_refcount_dec(void); int security_tun_dev_create(void); void security_tun_dev_post_create(struct sock *sk); int security_tun_dev_attach(struct sock *sk); @@ -2701,6 +2713,19 @@ static inline void security_inet_conn_established(struct sock *sk, { } +static inline int security_secmark_relabel_packet(u32 secid) +{ + return 0; +} + +static inline void security_secmark_refcount_inc(void) +{ +} + +static inline void security_secmark_refcount_dec(void) +{ +} + static inline int security_tun_dev_create(void) { return 0; diff --git a/include/linux/selinux.h b/include/linux/selinux.h index 82e0f26..44f4596 100644 --- a/include/linux/selinux.h +++ b/include/linux/selinux.h @@ -21,74 +21,11 @@ struct kern_ipc_perm; #ifdef CONFIG_SECURITY_SELINUX /** - * selinux_string_to_sid - map a security context string to a security ID - * @str: the security context string to be mapped - * @sid: ID value returned via this. - * - * Returns 0 if successful, with the SID stored in sid. A value - * of zero for sid indicates no SID could be determined (but no error - * occurred). - */ -int selinux_string_to_sid(char *str, u32 *sid); - -/** - * selinux_secmark_relabel_packet_permission - secmark permission check - * @sid: SECMARK ID value to be applied to network packet - * - * Returns 0 if the current task is allowed to set the SECMARK label of - * packets with the supplied security ID. Note that it is implicit that - * the packet is always being relabeled from the default unlabeled value, - * and that the access control decision is made in the AVC. - */ -int selinux_secmark_relabel_packet_permission(u32 sid); - -/** - * selinux_secmark_refcount_inc - increments the secmark use counter - * - * SELinux keeps track of the current SECMARK targets in use so it knows - * when to apply SECMARK label access checks to network packets. This - * function incements this reference count to indicate that a new SECMARK - * target has been configured. - */ -void selinux_secmark_refcount_inc(void); - -/** - * selinux_secmark_refcount_dec - decrements the secmark use counter - * - * SELinux keeps track of the current SECMARK targets in use so it knows - * when to apply SECMARK label access checks to network packets. This - * function decements this reference count to indicate that one of the - * existing SECMARK targets has been removed/flushed. - */ -void selinux_secmark_refcount_dec(void); - -/** * selinux_is_enabled - is SELinux enabled? */ bool selinux_is_enabled(void); #else -static inline int selinux_string_to_sid(const char *str, u32 *sid) -{ - *sid = 0; - return 0; -} - -static inline int selinux_secmark_relabel_packet_permission(u32 sid) -{ - return 0; -} - -static inline void selinux_secmark_refcount_inc(void) -{ - return; -} - -static inline void selinux_secmark_refcount_dec(void) -{ - return; -} - static inline bool selinux_is_enabled(void) { return false; diff --git a/net/netfilter/xt_CT.c b/net/netfilter/xt_CT.c index 0cb6053..782e519 100644 --- a/net/netfilter/xt_CT.c +++ b/net/netfilter/xt_CT.c @@ -9,7 +9,6 @@ #include #include #include -#include #include #include #include diff --git a/net/netfilter/xt_SECMARK.c b/net/netfilter/xt_SECMARK.c index 364ad16..9faf5e0 100644 --- a/net/netfilter/xt_SECMARK.c +++ b/net/netfilter/xt_SECMARK.c @@ -14,8 +14,8 @@ */ #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt #include +#include #include -#include #include #include @@ -39,9 +39,8 @@ secmark_tg(struct sk_buff *skb, const struct xt_action_param *par) switch (mode) { case SECMARK_MODE_SEL: - secmark = info->u.sel.selsid; + secmark = info->secid; break; - default: BUG(); } @@ -50,33 +49,33 @@ secmark_tg(struct sk_buff *skb, const struct xt_action_param *par) return XT_CONTINUE; } -static int checkentry_selinux(struct xt_secmark_target_info *info) +static int checkentry_lsm(struct xt_secmark_target_info *info) { int err; - struct xt_secmark_target_selinux_info *sel = &info->u.sel; - sel->selctx[SECMARK_SELCTX_MAX - 1] = '\0'; + info->secctx[SECMARK_SECCTX_MAX - 1] = '\0'; + info->secid = 0; - err = selinux_string_to_sid(sel->selctx, &sel->selsid); + err = security_secctx_to_secid(info->secctx, strlen(info->secctx), + &info->secid); if (err) { if (err == -EINVAL) - pr_info("invalid SELinux context \'%s\'\n", - sel->selctx); + pr_info("invalid security context \'%s\'\n", info->secctx); return err; } - if (!sel->selsid) { - pr_info("unable to map SELinux context \'%s\'\n", sel->selctx); + if (!info->secid) { + pr_info("unable to map security context \'%s\'\n", info->secctx); return -ENOENT; } - err = selinux_secmark_relabel_packet_permission(sel->selsid); + err = security_secmark_relabel_packet(info->secid); if (err) { pr_info("unable to obtain relabeling permission\n"); return err; } - selinux_secmark_refcount_inc(); + security_secmark_refcount_inc(); return 0; } @@ -100,16 +99,16 @@ static int secmark_tg_check(const struct xt_tgchk_param *par) switch (info->mode) { case SECMARK_MODE_SEL: - err = checkentry_selinux(info); - if (err) - return err; break; - default: pr_info("invalid mode: %hu\n", info->mode); return -EINVAL; } + err = checkentry_lsm(info); + if (err) + return err; + if (!mode) mode = info->mode; return 0; @@ -119,7 +118,7 @@ static void secmark_tg_destroy(const struct xt_tgdtor_param *par) { switch (mode) { case SECMARK_MODE_SEL: - selinux_secmark_refcount_dec(); + security_secmark_refcount_dec(); } } diff --git a/security/capability.c b/security/capability.c index 95a6599..30ae00f 100644 --- a/security/capability.c +++ b/security/capability.c @@ -677,7 +677,18 @@ static void cap_inet_conn_established(struct sock *sk, struct sk_buff *skb) { } +static int cap_secmark_relabel_packet(u32 secid) +{ + return 0; +} +static void cap_secmark_refcount_inc(void) +{ +} + +static void cap_secmark_refcount_dec(void) +{ +} static void cap_req_classify_flow(const struct request_sock *req, struct flowi *fl) @@ -777,7 +788,8 @@ static int cap_secid_to_secctx(u32 secid, char **secdata, u32 *seclen) static int cap_secctx_to_secid(const char *secdata, u32 seclen, u32 *secid) { - return -EOPNOTSUPP; + *secid = 0; + return 0; } static void cap_release_secctx(char *secdata, u32 seclen) @@ -1018,6 +1030,9 @@ void __init security_fixup_ops(struct security_operations *ops) set_to_cap_if_null(ops, inet_conn_request); set_to_cap_if_null(ops, inet_csk_clone); set_to_cap_if_null(ops, inet_conn_established); + set_to_cap_if_null(ops, secmark_relabel_packet); + set_to_cap_if_null(ops, secmark_refcount_inc); + set_to_cap_if_null(ops, secmark_refcount_dec); set_to_cap_if_null(ops, req_classify_flow); set_to_cap_if_null(ops, tun_dev_create); set_to_cap_if_null(ops, tun_dev_post_create); diff --git a/security/security.c b/security/security.c index 1cbcdfa..b50f472 100644 --- a/security/security.c +++ b/security/security.c @@ -1136,6 +1136,24 @@ void security_inet_conn_established(struct sock *sk, security_ops->inet_conn_established(sk, skb); } +int security_secmark_relabel_packet(u32 secid) +{ + return security_ops->secmark_relabel_packet(secid); +} +EXPORT_SYMBOL(security_secmark_relabel_packet); + +void security_secmark_refcount_inc(void) +{ + security_ops->secmark_refcount_inc(); +} +EXPORT_SYMBOL(security_secmark_refcount_inc); + +void security_secmark_refcount_dec(void) +{ + security_ops->secmark_refcount_dec(); +} +EXPORT_SYMBOL(security_secmark_refcount_dec); + int security_tun_dev_create(void) { return security_ops->tun_dev_create(); diff --git a/security/selinux/exports.c b/security/selinux/exports.c index c0a454a..9066438 100644 --- a/security/selinux/exports.c +++ b/security/selinux/exports.c @@ -11,58 +11,9 @@ * it under the terms of the GNU General Public License version 2, * as published by the Free Software Foundation. */ -#include -#include #include -#include -#include -#include -#include #include "security.h" -#include "objsec.h" - -/* SECMARK reference count */ -extern atomic_t selinux_secmark_refcount; - -int selinux_string_to_sid(char *str, u32 *sid) -{ - if (selinux_enabled) - return security_context_to_sid(str, strlen(str), sid); - else { - *sid = 0; - return 0; - } -} -EXPORT_SYMBOL_GPL(selinux_string_to_sid); - -int selinux_secmark_relabel_packet_permission(u32 sid) -{ - if (selinux_enabled) { - const struct task_security_struct *__tsec; - u32 tsid; - - __tsec = current_security(); - tsid = __tsec->sid; - - return avc_has_perm(tsid, sid, SECCLASS_PACKET, - PACKET__RELABELTO, NULL); - } - return 0; -} -EXPORT_SYMBOL_GPL(selinux_secmark_relabel_packet_permission); - -void selinux_secmark_refcount_inc(void) -{ - atomic_inc(&selinux_secmark_refcount); -} -EXPORT_SYMBOL_GPL(selinux_secmark_refcount_inc); - -void selinux_secmark_refcount_dec(void) -{ - atomic_dec(&selinux_secmark_refcount); -} -EXPORT_SYMBOL_GPL(selinux_secmark_refcount_dec); bool selinux_is_enabled(void) { diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index db2b331..d9154cf 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -4279,6 +4279,27 @@ static void selinux_inet_conn_established(struct sock *sk, struct sk_buff *skb) selinux_skb_peerlbl_sid(skb, family, &sksec->peer_sid); } +static int selinux_secmark_relabel_packet(u32 sid) +{ + const struct task_security_struct *__tsec; + u32 tsid; + + __tsec = current_security(); + tsid = __tsec->sid; + + return avc_has_perm(tsid, sid, SECCLASS_PACKET, PACKET__RELABELTO, NULL); +} + +static void selinux_secmark_refcount_inc(void) +{ + atomic_inc(&selinux_secmark_refcount); +} + +static void selinux_secmark_refcount_dec(void) +{ + atomic_dec(&selinux_secmark_refcount); +} + static void selinux_req_classify_flow(const struct request_sock *req, struct flowi *fl) { @@ -5533,6 +5554,9 @@ static struct security_operations selinux_ops = { .inet_conn_request = selinux_inet_conn_request, .inet_csk_clone = selinux_inet_csk_clone, .inet_conn_established = selinux_inet_conn_established, + .secmark_relabel_packet = selinux_secmark_relabel_packet, + .secmark_refcount_inc = selinux_secmark_refcount_inc, + .secmark_refcount_dec = selinux_secmark_refcount_dec, .req_classify_flow = selinux_req_classify_flow, .tun_dev_create = selinux_tun_dev_create, .tun_dev_post_create = selinux_tun_dev_post_create, diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h index 4b66f19..611a526 100644 --- a/security/selinux/include/security.h +++ b/security/selinux/include/security.h @@ -9,6 +9,7 @@ #define _SELINUX_SECURITY_H_ #include +#include #include "flask.h" #define SECSID_NULL 0x00000000 /* unspecified SID */ -- cgit v0.10.2 From d5630b9d276bd389299ffea620b7c340ab19bcf5 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 16:24:48 -0400 Subject: security: secid_to_secctx returns len when data is NULL With the (long ago) interface change to have the secid_to_secctx functions do the string allocation instead of having the caller do the allocation we lost the ability to query the security server for the length of the upcoming string. The SECMARK code would like to allocate a netlink skb with enough length to hold the string but it is just too unclean to do the string allocation twice or to do the allocation the first time and hold onto the string and slen. This patch adds the ability to call security_secid_to_secctx() with a NULL data pointer and it will just set the slen pointer. Signed-off-by: Eric Paris Reviewed-by: Paul Moore Signed-off-by: James Morris diff --git a/include/linux/security.h b/include/linux/security.h index d70adc3..b8246a8 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -1285,9 +1285,13 @@ static inline void security_free_mnt_opts(struct security_mnt_opts *opts) * Return 0 if permission is granted. * * @secid_to_secctx: - * Convert secid to security context. + * Convert secid to security context. If secdata is NULL the length of + * the result will be returned in seclen, but no secdata will be returned. + * This does mean that the length could change between calls to check the + * length and the next call which actually allocates and returns the secdata. * @secid contains the security ID. * @secdata contains the pointer that stores the converted security context. + * @seclen pointer which contains the length of the data * @secctx_to_secid: * Convert security context to secid. * @secid contains the pointer to the generated security ID. diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c index 494ff52..60964d7 100644 --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -991,7 +991,8 @@ static int context_struct_to_string(struct context *context, char **scontext, u3 { char *scontextp; - *scontext = NULL; + if (scontext) + *scontext = NULL; *scontext_len = 0; if (context->len) { @@ -1008,6 +1009,9 @@ static int context_struct_to_string(struct context *context, char **scontext, u3 *scontext_len += strlen(policydb.p_type_val_to_name[context->type - 1]) + 1; *scontext_len += mls_compute_context_len(context); + if (!scontext) + return 0; + /* Allocate space for the context; caller must free this space. */ scontextp = kmalloc(*scontext_len, GFP_ATOMIC); if (!scontextp) @@ -1047,7 +1051,8 @@ static int security_sid_to_context_core(u32 sid, char **scontext, struct context *context; int rc = 0; - *scontext = NULL; + if (scontext) + *scontext = NULL; *scontext_len = 0; if (!ss_initialized) { @@ -1055,6 +1060,8 @@ static int security_sid_to_context_core(u32 sid, char **scontext, char *scontextp; *scontext_len = strlen(initial_sid_to_string[sid]) + 1; + if (!scontext) + goto out; scontextp = kmalloc(*scontext_len, GFP_ATOMIC); if (!scontextp) { rc = -ENOMEM; diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index 174aec4..bc39f40 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -3004,7 +3004,8 @@ static int smack_secid_to_secctx(u32 secid, char **secdata, u32 *seclen) { char *sp = smack_from_secid(secid); - *secdata = sp; + if (secdata) + *secdata = sp; *seclen = strlen(sp); return 0; } -- cgit v0.10.2 From 1cc63249adfa957b34ca51effdee90ff8261d63f Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 16:24:54 -0400 Subject: conntrack: export lsm context rather than internal secid via netlink The conntrack code can export the internal secid to userspace. These are dynamic, can change on lsm changes, and have no meaning in userspace. We should instead be sending lsm contexts to userspace instead. This patch sends the secctx (rather than secid) to userspace over the netlink socket. We use a new field CTA_SECCTX and stop using the the old CTA_SECMARK field since it did not send particularly useful information. Signed-off-by: Eric Paris Reviewed-by: Paul Moore Acked-by: Patrick McHardy Signed-off-by: James Morris diff --git a/include/linux/netfilter/nfnetlink_conntrack.h b/include/linux/netfilter/nfnetlink_conntrack.h index 9ed534c..70cd060 100644 --- a/include/linux/netfilter/nfnetlink_conntrack.h +++ b/include/linux/netfilter/nfnetlink_conntrack.h @@ -39,8 +39,9 @@ enum ctattr_type { CTA_TUPLE_MASTER, CTA_NAT_SEQ_ADJ_ORIG, CTA_NAT_SEQ_ADJ_REPLY, - CTA_SECMARK, + CTA_SECMARK, /* obsolete */ CTA_ZONE, + CTA_SECCTX, __CTA_MAX }; #define CTA_MAX (__CTA_MAX - 1) @@ -172,4 +173,11 @@ enum ctattr_help { }; #define CTA_HELP_MAX (__CTA_HELP_MAX - 1) +enum ctattr_secctx { + CTA_SECCTX_UNSPEC, + CTA_SECCTX_NAME, + __CTA_SECCTX_MAX +}; +#define CTA_SECCTX_MAX (__CTA_SECCTX_MAX - 1) + #endif /* _IPCONNTRACK_NETLINK_H */ diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 5bae1cd..b3c6285 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include @@ -245,16 +246,31 @@ nla_put_failure: #ifdef CONFIG_NF_CONNTRACK_SECMARK static inline int -ctnetlink_dump_secmark(struct sk_buff *skb, const struct nf_conn *ct) +ctnetlink_dump_secctx(struct sk_buff *skb, const struct nf_conn *ct) { - NLA_PUT_BE32(skb, CTA_SECMARK, htonl(ct->secmark)); - return 0; + struct nlattr *nest_secctx; + int len, ret; + char *secctx; + + ret = security_secid_to_secctx(ct->secmark, &secctx, &len); + if (ret) + return ret; + + ret = -1; + nest_secctx = nla_nest_start(skb, CTA_SECCTX | NLA_F_NESTED); + if (!nest_secctx) + goto nla_put_failure; + NLA_PUT_STRING(skb, CTA_SECCTX_NAME, secctx); + nla_nest_end(skb, nest_secctx); + + ret = 0; nla_put_failure: - return -1; + security_release_secctx(secctx, len); + return ret; } #else -#define ctnetlink_dump_secmark(a, b) (0) +#define ctnetlink_dump_secctx(a, b) (0) #endif #define master_tuple(ct) &(ct->master->tuplehash[IP_CT_DIR_ORIGINAL].tuple) @@ -391,7 +407,7 @@ ctnetlink_fill_info(struct sk_buff *skb, u32 pid, u32 seq, ctnetlink_dump_protoinfo(skb, ct) < 0 || ctnetlink_dump_helpinfo(skb, ct) < 0 || ctnetlink_dump_mark(skb, ct) < 0 || - ctnetlink_dump_secmark(skb, ct) < 0 || + ctnetlink_dump_secctx(skb, ct) < 0 || ctnetlink_dump_id(skb, ct) < 0 || ctnetlink_dump_use(skb, ct) < 0 || ctnetlink_dump_master(skb, ct) < 0 || @@ -437,6 +453,17 @@ ctnetlink_counters_size(const struct nf_conn *ct) ; } +#ifdef CONFIG_NF_CONNTRACK_SECMARK +static int ctnetlink_nlmsg_secctx_size(const struct nf_conn *ct) +{ + int len; + + security_secid_to_secctx(ct->secmark, NULL, &len); + + return sizeof(char) * len; +} +#endif + static inline size_t ctnetlink_nlmsg_size(const struct nf_conn *ct) { @@ -453,7 +480,8 @@ ctnetlink_nlmsg_size(const struct nf_conn *ct) + nla_total_size(0) /* CTA_HELP */ + nla_total_size(NF_CT_HELPER_NAME_LEN) /* CTA_HELP_NAME */ #ifdef CONFIG_NF_CONNTRACK_SECMARK - + nla_total_size(sizeof(u_int32_t)) /* CTA_SECMARK */ + + nla_total_size(0) /* CTA_SECCTX */ + + nla_total_size(ctnetlink_nlmsg_secctx_size(ct)) /* CTA_SECCTX_NAME */ #endif #ifdef CONFIG_NF_NAT_NEEDED + 2 * nla_total_size(0) /* CTA_NAT_SEQ_ADJ_ORIG|REPL */ @@ -554,11 +582,9 @@ ctnetlink_conntrack_event(unsigned int events, struct nf_ct_event *item) && ctnetlink_dump_helpinfo(skb, ct) < 0) goto nla_put_failure; -#ifdef CONFIG_NF_CONNTRACK_SECMARK if ((events & (1 << IPCT_SECMARK) || ct->secmark) - && ctnetlink_dump_secmark(skb, ct) < 0) + && ctnetlink_dump_secctx(skb, ct) < 0) goto nla_put_failure; -#endif if (events & (1 << IPCT_RELATED) && ctnetlink_dump_master(skb, ct) < 0) -- cgit v0.10.2 From 1ae4de0cdf855305765592647025bde55e85e451 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 16:25:00 -0400 Subject: secmark: export secctx, drop secmark in procfs The current secmark code exports a secmark= field which just indicates if there is special labeling on a packet or not. We drop this field as it isn't particularly useful and instead export a new field secctx= which is the actual human readable text label. Signed-off-by: Eric Paris Acked-by: Patrick McHardy Signed-off-by: James Morris diff --git a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c index 244f7cb..37f8adb 100644 --- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c +++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include @@ -87,6 +88,29 @@ static void ct_seq_stop(struct seq_file *s, void *v) rcu_read_unlock(); } +#ifdef CONFIG_NF_CONNTRACK_SECMARK +static int ct_show_secctx(struct seq_file *s, const struct nf_conn *ct) +{ + int ret; + u32 len; + char *secctx; + + ret = security_secid_to_secctx(ct->secmark, &secctx, &len); + if (ret) + return ret; + + ret = seq_printf(s, "secctx=%s ", secctx); + + security_release_secctx(secctx, len); + return ret; +} +#else +static inline int ct_show_secctx(struct seq_file *s, const struct nf_conn *ct) +{ + return 0; +} +#endif + static int ct_seq_show(struct seq_file *s, void *v) { struct nf_conntrack_tuple_hash *hash = v; @@ -148,10 +172,8 @@ static int ct_seq_show(struct seq_file *s, void *v) goto release; #endif -#ifdef CONFIG_NF_CONNTRACK_SECMARK - if (seq_printf(s, "secmark=%u ", ct->secmark)) + if (ct_show_secctx(s, ct)) goto release; -#endif if (seq_printf(s, "use=%u\n", atomic_read(&ct->ct_general.use))) goto release; diff --git a/net/netfilter/nf_conntrack_standalone.c b/net/netfilter/nf_conntrack_standalone.c index eb973fc..0fb6570 100644 --- a/net/netfilter/nf_conntrack_standalone.c +++ b/net/netfilter/nf_conntrack_standalone.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #ifdef CONFIG_SYSCTL #include @@ -108,6 +109,29 @@ static void ct_seq_stop(struct seq_file *s, void *v) rcu_read_unlock(); } +#ifdef CONFIG_NF_CONNTRACK_SECMARK +static int ct_show_secctx(struct seq_file *s, const struct nf_conn *ct) +{ + int ret; + u32 len; + char *secctx; + + ret = security_secid_to_secctx(ct->secmark, &secctx, &len); + if (ret) + return ret; + + ret = seq_printf(s, "secctx=%s ", secctx); + + security_release_secctx(secctx, len); + return ret; +} +#else +static inline int ct_show_secctx(struct seq_file *s, const struct nf_conn *ct) +{ + return 0; +} +#endif + /* return 0 on success, 1 in case of error */ static int ct_seq_show(struct seq_file *s, void *v) { @@ -168,10 +192,8 @@ static int ct_seq_show(struct seq_file *s, void *v) goto release; #endif -#ifdef CONFIG_NF_CONNTRACK_SECMARK - if (seq_printf(s, "secmark=%u ", ct->secmark)) + if (ct_show_secctx(s, ct)) goto release; -#endif #ifdef CONFIG_NF_CONNTRACK_ZONES if (seq_printf(s, "zone=%u ", nf_ct_zone(ct))) -- cgit v0.10.2 From 686a0f3d71203bbfcc186900bbb8ac2cfc3d803c Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 17:50:02 -0400 Subject: kernel: rounddown helper function The roundup() helper function will round a given value up to a multiple of another given value. aka roundup(11, 7) would give 14 = 7 * 2. This new function does the opposite. It will round a given number down to the nearest multiple of the second number: rounddown(11, 7) would give 7. I need this in some future SELinux code and can carry the macro myself, but figured I would put it in the core kernel so others might find and use it if need be. Signed-off-by: Eric Paris Signed-off-by: James Morris diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 2b0a35e..6d6eea7 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -59,6 +59,12 @@ extern const char linux_proc_banner[]; #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f)) #define DIV_ROUND_UP(n,d) (((n) + (d) - 1) / (d)) #define roundup(x, y) ((((x) + ((y) - 1)) / (y)) * (y)) +#define rounddown(x, y) ( \ +{ \ + typeof(x) __x = (x); \ + __x - (__x % (y)); \ +} \ +) #define DIV_ROUND_CLOSEST(x, divisor)( \ { \ typeof(divisor) __divisor = divisor; \ -- cgit v0.10.2 From b28efd54d9d5c8005a29cd8782335beb9daaa32d Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 17:50:08 -0400 Subject: kernel: roundup should only reference arguments once Currently the roundup macro references it's arguments more than one time. This patch changes it so it will only use its arguments once. Suggested-by: Andrew Morton Signed-off-by: Eric Paris Signed-off-by: James Morris diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 6d6eea7..1759ba5 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -58,7 +58,12 @@ extern const char linux_proc_banner[]; #define FIELD_SIZEOF(t, f) (sizeof(((t*)0)->f)) #define DIV_ROUND_UP(n,d) (((n) + (d) - 1) / (d)) -#define roundup(x, y) ((((x) + ((y) - 1)) / (y)) * (y)) +#define roundup(x, y) ( \ +{ \ + typeof(y) __y = y; \ + (((x) + (__y - 1)) / __y) * __y; \ +} \ +) #define rounddown(x, y) ( \ { \ typeof(x) __x = (x); \ -- cgit v0.10.2 From 4419aae1f4f380a3fba0f4f12ffbbbdf3f267c51 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 17:50:14 -0400 Subject: SELinux: deterministic ordering of range transition rules Range transition rules are placed in the hash table in an (almost) arbitrary order. This patch inserts them in a fixed order to make policy retrival more predictable. Signed-off-by: Eric Paris Signed-off-by: James Morris diff --git a/security/selinux/ss/policydb.c b/security/selinux/ss/policydb.c index aa5a2fd..97fb0cf 100644 --- a/security/selinux/ss/policydb.c +++ b/security/selinux/ss/policydb.c @@ -185,9 +185,19 @@ static u32 rangetr_hash(struct hashtab *h, const void *k) static int rangetr_cmp(struct hashtab *h, const void *k1, const void *k2) { const struct range_trans *key1 = k1, *key2 = k2; - return (key1->source_type != key2->source_type || - key1->target_type != key2->target_type || - key1->target_class != key2->target_class); + int v; + + v = key1->source_type - key2->source_type; + if (v) + return v; + + v = key1->target_type - key2->target_type; + if (v) + return v; + + v = key1->target_class - key2->target_class; + + return v; } /* -- cgit v0.10.2 From 00d85c83ac52e2c1a66397f1abc589f80c543425 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 17:50:19 -0400 Subject: SELinux: drop useless (and incorrect) AVTAB_MAX_SIZE AVTAB_MAX_SIZE was a define which was supposed to be used in userspace to define a maximally sized avtab when userspace wasn't sure how big of a table it needed. It doesn't make sense in the kernel since we always know our table sizes. The only place it is used we have a more appropiately named define called AVTAB_MAX_HASH_BUCKETS, use that instead. Signed-off-by: Eric Paris Signed-off-by: James Morris diff --git a/security/selinux/ss/avtab.c b/security/selinux/ss/avtab.c index 929480c..77a917c 100644 --- a/security/selinux/ss/avtab.c +++ b/security/selinux/ss/avtab.c @@ -266,8 +266,8 @@ int avtab_alloc(struct avtab *h, u32 nrules) if (shift > 2) shift = shift - 2; nslot = 1 << shift; - if (nslot > MAX_AVTAB_SIZE) - nslot = MAX_AVTAB_SIZE; + if (nslot > MAX_AVTAB_HASH_BUCKETS) + nslot = MAX_AVTAB_HASH_BUCKETS; mask = nslot - 1; h->htable = kcalloc(nslot, sizeof(*(h->htable)), GFP_KERNEL); diff --git a/security/selinux/ss/avtab.h b/security/selinux/ss/avtab.h index cd4f734..32bd8b3 100644 --- a/security/selinux/ss/avtab.h +++ b/security/selinux/ss/avtab.h @@ -85,7 +85,6 @@ void avtab_cache_destroy(void); #define MAX_AVTAB_HASH_BITS 11 #define MAX_AVTAB_HASH_BUCKETS (1 << MAX_AVTAB_HASH_BITS) #define MAX_AVTAB_HASH_MASK (MAX_AVTAB_HASH_BUCKETS-1) -#define MAX_AVTAB_SIZE MAX_AVTAB_HASH_BUCKETS #endif /* _SS_AVTAB_H_ */ -- cgit v0.10.2 From cee74f47a6baba0ac457e87687fdcf0abd599f0a Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 17:50:25 -0400 Subject: SELinux: allow userspace to read policy back out of the kernel There is interest in being able to see what the actual policy is that was loaded into the kernel. The patch creates a new selinuxfs file /selinux/policy which can be read by userspace. The actual policy that is loaded into the kernel will be written back out to userspace. Signed-off-by: Eric Paris Signed-off-by: James Morris diff --git a/security/selinux/include/classmap.h b/security/selinux/include/classmap.h index b4c9eb4..8858d2b 100644 --- a/security/selinux/include/classmap.h +++ b/security/selinux/include/classmap.h @@ -17,7 +17,7 @@ struct security_class_mapping secclass_map[] = { { "compute_av", "compute_create", "compute_member", "check_context", "load_policy", "compute_relabel", "compute_user", "setenforce", "setbool", "setsecparam", - "setcheckreqprot", NULL } }, + "setcheckreqprot", "read_policy", NULL } }, { "process", { "fork", "transition", "sigchld", "sigkill", "sigstop", "signull", "signal", "ptrace", "getsched", "setsched", diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h index 611a526..671273e 100644 --- a/security/selinux/include/security.h +++ b/security/selinux/include/security.h @@ -83,6 +83,8 @@ extern int selinux_policycap_openperm; int security_mls_enabled(void); int security_load_policy(void *data, size_t len); +int security_read_policy(void **data, ssize_t *len); +size_t security_policydb_len(void); int security_policycap_supported(unsigned int req_cap); diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index a2e7a85..8eb102c 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -68,6 +68,8 @@ static int *bool_pending_values; static struct dentry *class_dir; static unsigned long last_class_ino; +static char policy_opened; + /* global data for policy capabilities */ static struct dentry *policycap_dir; @@ -111,6 +113,7 @@ enum sel_inos { SEL_REJECT_UNKNOWN, /* export unknown reject handling to userspace */ SEL_DENY_UNKNOWN, /* export unknown deny handling to userspace */ SEL_STATUS, /* export current status using mmap() */ + SEL_POLICY, /* allow userspace to read the in kernel policy */ SEL_INO_NEXT, /* The next inode number to use */ }; @@ -351,6 +354,97 @@ static const struct file_operations sel_mls_ops = { .llseek = generic_file_llseek, }; +struct policy_load_memory { + size_t len; + void *data; +}; + +static int sel_open_policy(struct inode *inode, struct file *filp) +{ + struct policy_load_memory *plm = NULL; + int rc; + + BUG_ON(filp->private_data); + + mutex_lock(&sel_mutex); + + rc = task_has_security(current, SECURITY__READ_POLICY); + if (rc) + goto err; + + rc = -EBUSY; + if (policy_opened) + goto err; + + rc = -ENOMEM; + plm = kzalloc(sizeof(*plm), GFP_KERNEL); + if (!plm) + goto err; + + if (i_size_read(inode) != security_policydb_len()) { + mutex_lock(&inode->i_mutex); + i_size_write(inode, security_policydb_len()); + mutex_unlock(&inode->i_mutex); + } + + rc = security_read_policy(&plm->data, &plm->len); + if (rc) + goto err; + + policy_opened = 1; + + filp->private_data = plm; + + mutex_unlock(&sel_mutex); + + return 0; +err: + mutex_unlock(&sel_mutex); + + if (plm) + vfree(plm->data); + kfree(plm); + return rc; +} + +static int sel_release_policy(struct inode *inode, struct file *filp) +{ + struct policy_load_memory *plm = filp->private_data; + + BUG_ON(!plm); + + policy_opened = 0; + + vfree(plm->data); + kfree(plm); + + return 0; +} + +static ssize_t sel_read_policy(struct file *filp, char __user *buf, + size_t count, loff_t *ppos) +{ + struct policy_load_memory *plm = filp->private_data; + int ret; + + mutex_lock(&sel_mutex); + + ret = task_has_security(current, SECURITY__READ_POLICY); + if (ret) + goto out; + + ret = simple_read_from_buffer(buf, count, ppos, plm->data, plm->len); +out: + mutex_unlock(&sel_mutex); + return ret; +} + +static const struct file_operations sel_policy_ops = { + .open = sel_open_policy, + .read = sel_read_policy, + .release = sel_release_policy, +}; + static ssize_t sel_write_load(struct file *file, const char __user *buf, size_t count, loff_t *ppos) @@ -1668,6 +1762,7 @@ static int sel_fill_super(struct super_block *sb, void *data, int silent) [SEL_REJECT_UNKNOWN] = {"reject_unknown", &sel_handle_unknown_ops, S_IRUGO}, [SEL_DENY_UNKNOWN] = {"deny_unknown", &sel_handle_unknown_ops, S_IRUGO}, [SEL_STATUS] = {"status", &sel_handle_status_ops, S_IRUGO}, + [SEL_POLICY] = {"policy", &sel_policy_ops, S_IRUSR}, /* last one */ {""} }; ret = simple_fill_super(sb, SELINUX_MAGIC, selinux_files); diff --git a/security/selinux/ss/avtab.c b/security/selinux/ss/avtab.c index 77a917c..a3dd9fa 100644 --- a/security/selinux/ss/avtab.c +++ b/security/selinux/ss/avtab.c @@ -501,6 +501,48 @@ bad: goto out; } +int avtab_write_item(struct policydb *p, struct avtab_node *cur, void *fp) +{ + __le16 buf16[4]; + __le32 buf32[1]; + int rc; + + buf16[0] = cpu_to_le16(cur->key.source_type); + buf16[1] = cpu_to_le16(cur->key.target_type); + buf16[2] = cpu_to_le16(cur->key.target_class); + buf16[3] = cpu_to_le16(cur->key.specified); + rc = put_entry(buf16, sizeof(u16), 4, fp); + if (rc) + return rc; + buf32[0] = cpu_to_le32(cur->datum.data); + rc = put_entry(buf32, sizeof(u32), 1, fp); + if (rc) + return rc; + return 0; +} + +int avtab_write(struct policydb *p, struct avtab *a, void *fp) +{ + unsigned int i; + int rc = 0; + struct avtab_node *cur; + __le32 buf[1]; + + buf[0] = cpu_to_le32(a->nel); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + for (i = 0; i < a->nslot; i++) { + for (cur = a->htable[i]; cur; cur = cur->next) { + rc = avtab_write_item(p, cur, fp); + if (rc) + return rc; + } + } + + return rc; +} void avtab_cache_init(void) { avtab_node_cachep = kmem_cache_create("avtab_node", diff --git a/security/selinux/ss/avtab.h b/security/selinux/ss/avtab.h index 32bd8b3..dff0c75 100644 --- a/security/selinux/ss/avtab.h +++ b/security/selinux/ss/avtab.h @@ -71,6 +71,8 @@ int avtab_read_item(struct avtab *a, void *fp, struct policydb *pol, void *p); int avtab_read(struct avtab *a, void *fp, struct policydb *pol); +int avtab_write_item(struct policydb *p, struct avtab_node *cur, void *fp); +int avtab_write(struct policydb *p, struct avtab *a, void *fp); struct avtab_node *avtab_insert_nonunique(struct avtab *h, struct avtab_key *key, struct avtab_datum *datum); diff --git a/security/selinux/ss/conditional.c b/security/selinux/ss/conditional.c index c91e150..655fe1c 100644 --- a/security/selinux/ss/conditional.c +++ b/security/selinux/ss/conditional.c @@ -490,6 +490,129 @@ err: return rc; } +int cond_write_bool(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct cond_bool_datum *booldatum = datum; + struct policy_data *pd = ptr; + void *fp = pd->fp; + __le32 buf[3]; + u32 len; + int rc; + + len = strlen(key); + buf[0] = cpu_to_le32(booldatum->value); + buf[1] = cpu_to_le32(booldatum->state); + buf[2] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 3, fp); + if (rc) + return rc; + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + return 0; +} + +/* + * cond_write_cond_av_list doesn't write out the av_list nodes. + * Instead it writes out the key/value pairs from the avtab. This + * is necessary because there is no way to uniquely identifying rules + * in the avtab so it is not possible to associate individual rules + * in the avtab with a conditional without saving them as part of + * the conditional. This means that the avtab with the conditional + * rules will not be saved but will be rebuilt on policy load. + */ +static int cond_write_av_list(struct policydb *p, + struct cond_av_list *list, struct policy_file *fp) +{ + __le32 buf[1]; + struct cond_av_list *cur_list; + u32 len; + int rc; + + len = 0; + for (cur_list = list; cur_list != NULL; cur_list = cur_list->next) + len++; + + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + if (len == 0) + return 0; + + for (cur_list = list; cur_list != NULL; cur_list = cur_list->next) { + rc = avtab_write_item(p, cur_list->node, fp); + if (rc) + return rc; + } + + return 0; +} + +int cond_write_node(struct policydb *p, struct cond_node *node, + struct policy_file *fp) +{ + struct cond_expr *cur_expr; + __le32 buf[2]; + int rc; + u32 len = 0; + + buf[0] = cpu_to_le32(node->cur_state); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + for (cur_expr = node->expr; cur_expr != NULL; cur_expr = cur_expr->next) + len++; + + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + for (cur_expr = node->expr; cur_expr != NULL; cur_expr = cur_expr->next) { + buf[0] = cpu_to_le32(cur_expr->expr_type); + buf[1] = cpu_to_le32(cur_expr->bool); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + } + + rc = cond_write_av_list(p, node->true_list, fp); + if (rc) + return rc; + rc = cond_write_av_list(p, node->false_list, fp); + if (rc) + return rc; + + return 0; +} + +int cond_write_list(struct policydb *p, struct cond_node *list, void *fp) +{ + struct cond_node *cur; + u32 len; + __le32 buf[1]; + int rc; + + len = 0; + for (cur = list; cur != NULL; cur = cur->next) + len++; + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + for (cur = list; cur != NULL; cur = cur->next) { + rc = cond_write_node(p, cur, fp); + if (rc) + return rc; + } + + return 0; +} /* Determine whether additional permissions are granted by the conditional * av table, and if so, add them to the result */ diff --git a/security/selinux/ss/conditional.h b/security/selinux/ss/conditional.h index 53ddb01..3f209c6 100644 --- a/security/selinux/ss/conditional.h +++ b/security/selinux/ss/conditional.h @@ -69,6 +69,8 @@ int cond_index_bool(void *key, void *datum, void *datap); int cond_read_bool(struct policydb *p, struct hashtab *h, void *fp); int cond_read_list(struct policydb *p, void *fp); +int cond_write_bool(void *key, void *datum, void *ptr); +int cond_write_list(struct policydb *p, struct cond_node *list, void *fp); void cond_compute_av(struct avtab *ctab, struct avtab_key *key, struct av_decision *avd); diff --git a/security/selinux/ss/ebitmap.c b/security/selinux/ss/ebitmap.c index 04b6145..d42951f 100644 --- a/security/selinux/ss/ebitmap.c +++ b/security/selinux/ss/ebitmap.c @@ -22,6 +22,8 @@ #include "ebitmap.h" #include "policydb.h" +#define BITS_PER_U64 (sizeof(u64) * 8) + int ebitmap_cmp(struct ebitmap *e1, struct ebitmap *e2) { struct ebitmap_node *n1, *n2; @@ -363,10 +365,10 @@ int ebitmap_read(struct ebitmap *e, void *fp) e->highbit = le32_to_cpu(buf[1]); count = le32_to_cpu(buf[2]); - if (mapunit != sizeof(u64) * 8) { + if (mapunit != BITS_PER_U64) { printk(KERN_ERR "SELinux: ebitmap: map size %u does not " "match my size %Zd (high bit was %d)\n", - mapunit, sizeof(u64) * 8, e->highbit); + mapunit, BITS_PER_U64, e->highbit); goto bad; } @@ -446,3 +448,78 @@ bad: ebitmap_destroy(e); goto out; } + +int ebitmap_write(struct ebitmap *e, void *fp) +{ + struct ebitmap_node *n; + u32 count; + __le32 buf[3]; + u64 map; + int bit, last_bit, last_startbit, rc; + + buf[0] = cpu_to_le32(BITS_PER_U64); + + count = 0; + last_bit = 0; + last_startbit = -1; + ebitmap_for_each_positive_bit(e, n, bit) { + if (rounddown(bit, (int)BITS_PER_U64) > last_startbit) { + count++; + last_startbit = rounddown(bit, BITS_PER_U64); + } + last_bit = roundup(bit + 1, BITS_PER_U64); + } + buf[1] = cpu_to_le32(last_bit); + buf[2] = cpu_to_le32(count); + + rc = put_entry(buf, sizeof(u32), 3, fp); + if (rc) + return rc; + + map = 0; + last_startbit = INT_MIN; + ebitmap_for_each_positive_bit(e, n, bit) { + if (rounddown(bit, (int)BITS_PER_U64) > last_startbit) { + __le64 buf64[1]; + + /* this is the very first bit */ + if (!map) { + last_startbit = rounddown(bit, BITS_PER_U64); + map = (u64)1 << (bit - last_startbit); + continue; + } + + /* write the last node */ + buf[0] = cpu_to_le32(last_startbit); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + buf64[0] = cpu_to_le64(map); + rc = put_entry(buf64, sizeof(u64), 1, fp); + if (rc) + return rc; + + /* set up for the next node */ + map = 0; + last_startbit = rounddown(bit, BITS_PER_U64); + } + map |= (u64)1 << (bit - last_startbit); + } + /* write the last node */ + if (map) { + __le64 buf64[1]; + + /* write the last node */ + buf[0] = cpu_to_le32(last_startbit); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + buf64[0] = cpu_to_le64(map); + rc = put_entry(buf64, sizeof(u64), 1, fp); + if (rc) + return rc; + } + return 0; +} diff --git a/security/selinux/ss/ebitmap.h b/security/selinux/ss/ebitmap.h index f283b43..1f4e93c 100644 --- a/security/selinux/ss/ebitmap.h +++ b/security/selinux/ss/ebitmap.h @@ -123,6 +123,7 @@ int ebitmap_get_bit(struct ebitmap *e, unsigned long bit); int ebitmap_set_bit(struct ebitmap *e, unsigned long bit, int value); void ebitmap_destroy(struct ebitmap *e); int ebitmap_read(struct ebitmap *e, void *fp); +int ebitmap_write(struct ebitmap *e, void *fp); #ifdef CONFIG_NETLABEL int ebitmap_netlbl_export(struct ebitmap *ebmap, diff --git a/security/selinux/ss/policydb.c b/security/selinux/ss/policydb.c index 97fb0cf..94f630d 100644 --- a/security/selinux/ss/policydb.c +++ b/security/selinux/ss/policydb.c @@ -37,6 +37,7 @@ #include "policydb.h" #include "conditional.h" #include "mls.h" +#include "services.h" #define _DEBUG_HASHES @@ -2316,3 +2317,843 @@ bad: policydb_destroy(p); goto out; } + +/* + * Write a MLS level structure to a policydb binary + * representation file. + */ +static int mls_write_level(struct mls_level *l, void *fp) +{ + __le32 buf[1]; + int rc; + + buf[0] = cpu_to_le32(l->sens); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + rc = ebitmap_write(&l->cat, fp); + if (rc) + return rc; + + return 0; +} + +/* + * Write a MLS range structure to a policydb binary + * representation file. + */ +static int mls_write_range_helper(struct mls_range *r, void *fp) +{ + __le32 buf[3]; + size_t items; + int rc, eq; + + eq = mls_level_eq(&r->level[1], &r->level[0]); + + if (eq) + items = 2; + else + items = 3; + buf[0] = cpu_to_le32(items-1); + buf[1] = cpu_to_le32(r->level[0].sens); + if (!eq) + buf[2] = cpu_to_le32(r->level[1].sens); + + BUG_ON(items > (sizeof(buf)/sizeof(buf[0]))); + + rc = put_entry(buf, sizeof(u32), items, fp); + if (rc) + return rc; + + rc = ebitmap_write(&r->level[0].cat, fp); + if (rc) + return rc; + if (!eq) { + rc = ebitmap_write(&r->level[1].cat, fp); + if (rc) + return rc; + } + + return 0; +} + +static int sens_write(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct level_datum *levdatum = datum; + struct policy_data *pd = ptr; + void *fp = pd->fp; + __le32 buf[2]; + size_t len; + int rc; + + len = strlen(key); + buf[0] = cpu_to_le32(len); + buf[1] = cpu_to_le32(levdatum->isalias); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + rc = mls_write_level(levdatum->level, fp); + if (rc) + return rc; + + return 0; +} + +static int cat_write(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct cat_datum *catdatum = datum; + struct policy_data *pd = ptr; + void *fp = pd->fp; + __le32 buf[3]; + size_t len; + int rc; + + len = strlen(key); + buf[0] = cpu_to_le32(len); + buf[1] = cpu_to_le32(catdatum->value); + buf[2] = cpu_to_le32(catdatum->isalias); + rc = put_entry(buf, sizeof(u32), 3, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + return 0; +} + +static int role_trans_write(struct role_trans *r, void *fp) +{ + struct role_trans *tr; + u32 buf[3]; + size_t nel; + int rc; + + nel = 0; + for (tr = r; tr; tr = tr->next) + nel++; + buf[0] = cpu_to_le32(nel); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + for (tr = r; tr; tr = tr->next) { + buf[0] = cpu_to_le32(tr->role); + buf[1] = cpu_to_le32(tr->type); + buf[2] = cpu_to_le32(tr->new_role); + rc = put_entry(buf, sizeof(u32), 3, fp); + if (rc) + return rc; + } + + return 0; +} + +static int role_allow_write(struct role_allow *r, void *fp) +{ + struct role_allow *ra; + u32 buf[2]; + size_t nel; + int rc; + + nel = 0; + for (ra = r; ra; ra = ra->next) + nel++; + buf[0] = cpu_to_le32(nel); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + for (ra = r; ra; ra = ra->next) { + buf[0] = cpu_to_le32(ra->role); + buf[1] = cpu_to_le32(ra->new_role); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + } + return 0; +} + +/* + * Write a security context structure + * to a policydb binary representation file. + */ +static int context_write(struct policydb *p, struct context *c, + void *fp) +{ + int rc; + __le32 buf[3]; + + buf[0] = cpu_to_le32(c->user); + buf[1] = cpu_to_le32(c->role); + buf[2] = cpu_to_le32(c->type); + + rc = put_entry(buf, sizeof(u32), 3, fp); + if (rc) + return rc; + + rc = mls_write_range_helper(&c->range, fp); + if (rc) + return rc; + + return 0; +} + +/* + * The following *_write functions are used to + * write the symbol data to a policy database + * binary representation file. + */ + +static int perm_write(void *vkey, void *datum, void *fp) +{ + char *key = vkey; + struct perm_datum *perdatum = datum; + __le32 buf[2]; + size_t len; + int rc; + + len = strlen(key); + buf[0] = cpu_to_le32(len); + buf[1] = cpu_to_le32(perdatum->value); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + return 0; +} + +static int common_write(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct common_datum *comdatum = datum; + struct policy_data *pd = ptr; + void *fp = pd->fp; + __le32 buf[4]; + size_t len; + int rc; + + len = strlen(key); + buf[0] = cpu_to_le32(len); + buf[1] = cpu_to_le32(comdatum->value); + buf[2] = cpu_to_le32(comdatum->permissions.nprim); + buf[3] = cpu_to_le32(comdatum->permissions.table->nel); + rc = put_entry(buf, sizeof(u32), 4, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + rc = hashtab_map(comdatum->permissions.table, perm_write, fp); + if (rc) + return rc; + + return 0; +} + +static int write_cons_helper(struct policydb *p, struct constraint_node *node, + void *fp) +{ + struct constraint_node *c; + struct constraint_expr *e; + __le32 buf[3]; + u32 nel; + int rc; + + for (c = node; c; c = c->next) { + nel = 0; + for (e = c->expr; e; e = e->next) + nel++; + buf[0] = cpu_to_le32(c->permissions); + buf[1] = cpu_to_le32(nel); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + for (e = c->expr; e; e = e->next) { + buf[0] = cpu_to_le32(e->expr_type); + buf[1] = cpu_to_le32(e->attr); + buf[2] = cpu_to_le32(e->op); + rc = put_entry(buf, sizeof(u32), 3, fp); + if (rc) + return rc; + + switch (e->expr_type) { + case CEXPR_NAMES: + rc = ebitmap_write(&e->names, fp); + if (rc) + return rc; + break; + default: + break; + } + } + } + + return 0; +} + +static int class_write(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct class_datum *cladatum = datum; + struct policy_data *pd = ptr; + void *fp = pd->fp; + struct policydb *p = pd->p; + struct constraint_node *c; + __le32 buf[6]; + u32 ncons; + size_t len, len2; + int rc; + + len = strlen(key); + if (cladatum->comkey) + len2 = strlen(cladatum->comkey); + else + len2 = 0; + + ncons = 0; + for (c = cladatum->constraints; c; c = c->next) + ncons++; + + buf[0] = cpu_to_le32(len); + buf[1] = cpu_to_le32(len2); + buf[2] = cpu_to_le32(cladatum->value); + buf[3] = cpu_to_le32(cladatum->permissions.nprim); + if (cladatum->permissions.table) + buf[4] = cpu_to_le32(cladatum->permissions.table->nel); + else + buf[4] = 0; + buf[5] = cpu_to_le32(ncons); + rc = put_entry(buf, sizeof(u32), 6, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + if (cladatum->comkey) { + rc = put_entry(cladatum->comkey, 1, len2, fp); + if (rc) + return rc; + } + + rc = hashtab_map(cladatum->permissions.table, perm_write, fp); + if (rc) + return rc; + + rc = write_cons_helper(p, cladatum->constraints, fp); + if (rc) + return rc; + + /* write out the validatetrans rule */ + ncons = 0; + for (c = cladatum->validatetrans; c; c = c->next) + ncons++; + + buf[0] = cpu_to_le32(ncons); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + rc = write_cons_helper(p, cladatum->validatetrans, fp); + if (rc) + return rc; + + return 0; +} + +static int role_write(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct role_datum *role = datum; + struct policy_data *pd = ptr; + void *fp = pd->fp; + struct policydb *p = pd->p; + __le32 buf[3]; + size_t items, len; + int rc; + + len = strlen(key); + items = 0; + buf[items++] = cpu_to_le32(len); + buf[items++] = cpu_to_le32(role->value); + if (p->policyvers >= POLICYDB_VERSION_BOUNDARY) + buf[items++] = cpu_to_le32(role->bounds); + + BUG_ON(items > (sizeof(buf)/sizeof(buf[0]))); + + rc = put_entry(buf, sizeof(u32), items, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + rc = ebitmap_write(&role->dominates, fp); + if (rc) + return rc; + + rc = ebitmap_write(&role->types, fp); + if (rc) + return rc; + + return 0; +} + +static int type_write(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct type_datum *typdatum = datum; + struct policy_data *pd = ptr; + struct policydb *p = pd->p; + void *fp = pd->fp; + __le32 buf[4]; + int rc; + size_t items, len; + + len = strlen(key); + items = 0; + buf[items++] = cpu_to_le32(len); + buf[items++] = cpu_to_le32(typdatum->value); + if (p->policyvers >= POLICYDB_VERSION_BOUNDARY) { + u32 properties = 0; + + if (typdatum->primary) + properties |= TYPEDATUM_PROPERTY_PRIMARY; + + if (typdatum->attribute) + properties |= TYPEDATUM_PROPERTY_ATTRIBUTE; + + buf[items++] = cpu_to_le32(properties); + buf[items++] = cpu_to_le32(typdatum->bounds); + } else { + buf[items++] = cpu_to_le32(typdatum->primary); + } + BUG_ON(items > (sizeof(buf) / sizeof(buf[0]))); + rc = put_entry(buf, sizeof(u32), items, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + return 0; +} + +static int user_write(void *vkey, void *datum, void *ptr) +{ + char *key = vkey; + struct user_datum *usrdatum = datum; + struct policy_data *pd = ptr; + struct policydb *p = pd->p; + void *fp = pd->fp; + __le32 buf[3]; + size_t items, len; + int rc; + + len = strlen(key); + items = 0; + buf[items++] = cpu_to_le32(len); + buf[items++] = cpu_to_le32(usrdatum->value); + if (p->policyvers >= POLICYDB_VERSION_BOUNDARY) + buf[items++] = cpu_to_le32(usrdatum->bounds); + BUG_ON(items > (sizeof(buf) / sizeof(buf[0]))); + rc = put_entry(buf, sizeof(u32), items, fp); + if (rc) + return rc; + + rc = put_entry(key, 1, len, fp); + if (rc) + return rc; + + rc = ebitmap_write(&usrdatum->roles, fp); + if (rc) + return rc; + + rc = mls_write_range_helper(&usrdatum->range, fp); + if (rc) + return rc; + + rc = mls_write_level(&usrdatum->dfltlevel, fp); + if (rc) + return rc; + + return 0; +} + +static int (*write_f[SYM_NUM]) (void *key, void *datum, + void *datap) = +{ + common_write, + class_write, + role_write, + type_write, + user_write, + cond_write_bool, + sens_write, + cat_write, +}; + +static int ocontext_write(struct policydb *p, struct policydb_compat_info *info, + void *fp) +{ + unsigned int i, j, rc; + size_t nel, len; + __le32 buf[3]; + u32 nodebuf[8]; + struct ocontext *c; + for (i = 0; i < info->ocon_num; i++) { + nel = 0; + for (c = p->ocontexts[i]; c; c = c->next) + nel++; + buf[0] = cpu_to_le32(nel); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + for (c = p->ocontexts[i]; c; c = c->next) { + switch (i) { + case OCON_ISID: + buf[0] = cpu_to_le32(c->sid[0]); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + rc = context_write(p, &c->context[0], fp); + if (rc) + return rc; + break; + case OCON_FS: + case OCON_NETIF: + len = strlen(c->u.name); + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + rc = put_entry(c->u.name, 1, len, fp); + if (rc) + return rc; + rc = context_write(p, &c->context[0], fp); + if (rc) + return rc; + rc = context_write(p, &c->context[1], fp); + if (rc) + return rc; + break; + case OCON_PORT: + buf[0] = cpu_to_le32(c->u.port.protocol); + buf[1] = cpu_to_le32(c->u.port.low_port); + buf[2] = cpu_to_le32(c->u.port.high_port); + rc = put_entry(buf, sizeof(u32), 3, fp); + if (rc) + return rc; + rc = context_write(p, &c->context[0], fp); + if (rc) + return rc; + break; + case OCON_NODE: + nodebuf[0] = c->u.node.addr; /* network order */ + nodebuf[1] = c->u.node.mask; /* network order */ + rc = put_entry(nodebuf, sizeof(u32), 2, fp); + if (rc) + return rc; + rc = context_write(p, &c->context[0], fp); + if (rc) + return rc; + break; + case OCON_FSUSE: + buf[0] = cpu_to_le32(c->v.behavior); + len = strlen(c->u.name); + buf[1] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + rc = put_entry(c->u.name, 1, len, fp); + if (rc) + return rc; + rc = context_write(p, &c->context[0], fp); + if (rc) + return rc; + break; + case OCON_NODE6: + for (j = 0; j < 4; j++) + nodebuf[j] = c->u.node6.addr[j]; /* network order */ + for (j = 0; j < 4; j++) + nodebuf[j + 4] = c->u.node6.mask[j]; /* network order */ + rc = put_entry(nodebuf, sizeof(u32), 8, fp); + if (rc) + return rc; + rc = context_write(p, &c->context[0], fp); + if (rc) + return rc; + break; + } + } + } + return 0; +} + +static int genfs_write(struct policydb *p, void *fp) +{ + struct genfs *genfs; + struct ocontext *c; + size_t len; + __le32 buf[1]; + int rc; + + len = 0; + for (genfs = p->genfs; genfs; genfs = genfs->next) + len++; + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + for (genfs = p->genfs; genfs; genfs = genfs->next) { + len = strlen(genfs->fstype); + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + rc = put_entry(genfs->fstype, 1, len, fp); + if (rc) + return rc; + len = 0; + for (c = genfs->head; c; c = c->next) + len++; + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + for (c = genfs->head; c; c = c->next) { + len = strlen(c->u.name); + buf[0] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + rc = put_entry(c->u.name, 1, len, fp); + if (rc) + return rc; + buf[0] = cpu_to_le32(c->v.sclass); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + rc = context_write(p, &c->context[0], fp); + if (rc) + return rc; + } + } + return 0; +} + +static int range_count(void *key, void *data, void *ptr) +{ + int *cnt = ptr; + *cnt = *cnt + 1; + + return 0; +} + +static int range_write_helper(void *key, void *data, void *ptr) +{ + __le32 buf[2]; + struct range_trans *rt = key; + struct mls_range *r = data; + struct policy_data *pd = ptr; + void *fp = pd->fp; + struct policydb *p = pd->p; + int rc; + + buf[0] = cpu_to_le32(rt->source_type); + buf[1] = cpu_to_le32(rt->target_type); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + if (p->policyvers >= POLICYDB_VERSION_RANGETRANS) { + buf[0] = cpu_to_le32(rt->target_class); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + } + rc = mls_write_range_helper(r, fp); + if (rc) + return rc; + + return 0; +} + +static int range_write(struct policydb *p, void *fp) +{ + size_t nel; + __le32 buf[1]; + int rc; + struct policy_data pd; + + pd.p = p; + pd.fp = fp; + + /* count the number of entries in the hashtab */ + nel = 0; + rc = hashtab_map(p->range_tr, range_count, &nel); + if (rc) + return rc; + + buf[0] = cpu_to_le32(nel); + rc = put_entry(buf, sizeof(u32), 1, fp); + if (rc) + return rc; + + /* actually write all of the entries */ + rc = hashtab_map(p->range_tr, range_write_helper, &pd); + if (rc) + return rc; + + return 0; +} + +/* + * Write the configuration data in a policy database + * structure to a policy database binary representation + * file. + */ +int policydb_write(struct policydb *p, void *fp) +{ + unsigned int i, num_syms; + int rc; + __le32 buf[4]; + u32 config; + size_t len; + struct policydb_compat_info *info; + + /* + * refuse to write policy older than compressed avtab + * to simplify the writer. There are other tests dropped + * since we assume this throughout the writer code. Be + * careful if you ever try to remove this restriction + */ + if (p->policyvers < POLICYDB_VERSION_AVTAB) { + printk(KERN_ERR "SELinux: refusing to write policy version %d." + " Because it is less than version %d\n", p->policyvers, + POLICYDB_VERSION_AVTAB); + return -EINVAL; + } + + config = 0; + if (p->mls_enabled) + config |= POLICYDB_CONFIG_MLS; + + if (p->reject_unknown) + config |= REJECT_UNKNOWN; + if (p->allow_unknown) + config |= ALLOW_UNKNOWN; + + /* Write the magic number and string identifiers. */ + buf[0] = cpu_to_le32(POLICYDB_MAGIC); + len = strlen(POLICYDB_STRING); + buf[1] = cpu_to_le32(len); + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + rc = put_entry(POLICYDB_STRING, 1, len, fp); + if (rc) + return rc; + + /* Write the version, config, and table sizes. */ + info = policydb_lookup_compat(p->policyvers); + if (!info) { + printk(KERN_ERR "SELinux: compatibility lookup failed for policy " + "version %d", p->policyvers); + return rc; + } + + buf[0] = cpu_to_le32(p->policyvers); + buf[1] = cpu_to_le32(config); + buf[2] = cpu_to_le32(info->sym_num); + buf[3] = cpu_to_le32(info->ocon_num); + + rc = put_entry(buf, sizeof(u32), 4, fp); + if (rc) + return rc; + + if (p->policyvers >= POLICYDB_VERSION_POLCAP) { + rc = ebitmap_write(&p->policycaps, fp); + if (rc) + return rc; + } + + if (p->policyvers >= POLICYDB_VERSION_PERMISSIVE) { + rc = ebitmap_write(&p->permissive_map, fp); + if (rc) + return rc; + } + + num_syms = info->sym_num; + for (i = 0; i < num_syms; i++) { + struct policy_data pd; + + pd.fp = fp; + pd.p = p; + + buf[0] = cpu_to_le32(p->symtab[i].nprim); + buf[1] = cpu_to_le32(p->symtab[i].table->nel); + + rc = put_entry(buf, sizeof(u32), 2, fp); + if (rc) + return rc; + rc = hashtab_map(p->symtab[i].table, write_f[i], &pd); + if (rc) + return rc; + } + + rc = avtab_write(p, &p->te_avtab, fp); + if (rc) + return rc; + + rc = cond_write_list(p, p->cond_list, fp); + if (rc) + return rc; + + rc = role_trans_write(p->role_tr, fp); + if (rc) + return rc; + + rc = role_allow_write(p->role_allow, fp); + if (rc) + return rc; + + rc = ocontext_write(p, info, fp); + if (rc) + return rc; + + rc = genfs_write(p, fp); + if (rc) + return rc; + + rc = range_write(p, fp); + if (rc) + return rc; + + for (i = 0; i < p->p_types.nprim; i++) { + struct ebitmap *e = flex_array_get(p->type_attr_map_array, i); + + BUG_ON(!e); + rc = ebitmap_write(e, fp); + if (rc) + return rc; + } + + return 0; +} diff --git a/security/selinux/ss/policydb.h b/security/selinux/ss/policydb.h index 310e944..95d3d7d 100644 --- a/security/selinux/ss/policydb.h +++ b/security/selinux/ss/policydb.h @@ -254,6 +254,9 @@ struct policydb { struct ebitmap permissive_map; + /* length of this policy when it was loaded */ + size_t len; + unsigned int policyvers; unsigned int reject_unknown : 1; @@ -270,6 +273,7 @@ extern int policydb_class_isvalid(struct policydb *p, unsigned int class); extern int policydb_type_isvalid(struct policydb *p, unsigned int type); extern int policydb_role_isvalid(struct policydb *p, unsigned int role); extern int policydb_read(struct policydb *p, void *fp); +extern int policydb_write(struct policydb *p, void *fp); #define PERM_SYMTAB_SIZE 32 @@ -290,6 +294,11 @@ struct policy_file { size_t len; }; +struct policy_data { + struct policydb *p; + void *fp; +}; + static inline int next_entry(void *buf, struct policy_file *fp, size_t bytes) { if (bytes > fp->len) @@ -301,6 +310,17 @@ static inline int next_entry(void *buf, struct policy_file *fp, size_t bytes) return 0; } +static inline int put_entry(void *buf, size_t bytes, int num, struct policy_file *fp) +{ + size_t len = bytes * num; + + memcpy(fp->data, buf, len); + fp->data += len; + fp->len -= len; + + return 0; +} + extern u16 string_to_security_class(struct policydb *p, const char *name); extern u32 string_to_av_perm(struct policydb *p, u16 tclass, const char *name); diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c index 60964d7..7565d16 100644 --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -1776,6 +1776,7 @@ int security_load_policy(void *data, size_t len) return rc; } + policydb.len = len; rc = selinux_set_mapping(&policydb, secclass_map, ¤t_mapping, ¤t_mapping_size); @@ -1812,6 +1813,7 @@ int security_load_policy(void *data, size_t len) if (rc) return rc; + newpolicydb.len = len; /* If switching between different policy types, log MLS status */ if (policydb.mls_enabled && !newpolicydb.mls_enabled) printk(KERN_INFO "SELinux: Disabling MLS support...\n"); @@ -1892,6 +1894,17 @@ err: } +size_t security_policydb_len(void) +{ + size_t len; + + read_lock(&policy_rwlock); + len = policydb.len; + read_unlock(&policy_rwlock); + + return len; +} + /** * security_port_sid - Obtain the SID for a port. * @protocol: protocol number @@ -3139,3 +3152,38 @@ netlbl_sid_to_secattr_failure: return rc; } #endif /* CONFIG_NETLABEL */ + +/** + * security_read_policy - read the policy. + * @data: binary policy data + * @len: length of data in bytes + * + */ +int security_read_policy(void **data, ssize_t *len) +{ + int rc; + struct policy_file fp; + + if (!ss_initialized) + return -EINVAL; + + *len = security_policydb_len(); + + *data = vmalloc(*len); + if (!*data) + return -ENOMEM; + + fp.data = *data; + fp.len = *len; + + read_lock(&policy_rwlock); + rc = policydb_write(&policydb, &fp); + read_unlock(&policy_rwlock); + + if (rc) + return rc; + + *len = (unsigned long)fp.data - (unsigned long)*data; + return 0; + +} -- cgit v0.10.2 From 845ca30fe9691f1bab7cfbf30b6d11c944eb4abd Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 13 Oct 2010 17:50:31 -0400 Subject: selinux: implement mmap on /selinux/policy /selinux/policy allows a user to copy the policy back out of the kernel. This patch allows userspace to actually mmap that file and use it directly. Signed-off-by: Eric Paris Signed-off-by: James Morris diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index 8eb102c..87e0556 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -439,9 +439,53 @@ out: return ret; } +static int sel_mmap_policy_fault(struct vm_area_struct *vma, + struct vm_fault *vmf) +{ + struct policy_load_memory *plm = vma->vm_file->private_data; + unsigned long offset; + struct page *page; + + if (vmf->flags & (FAULT_FLAG_MKWRITE | FAULT_FLAG_WRITE)) + return VM_FAULT_SIGBUS; + + offset = vmf->pgoff << PAGE_SHIFT; + if (offset >= roundup(plm->len, PAGE_SIZE)) + return VM_FAULT_SIGBUS; + + page = vmalloc_to_page(plm->data + offset); + get_page(page); + + vmf->page = page; + + return 0; +} + +static struct vm_operations_struct sel_mmap_policy_ops = { + .fault = sel_mmap_policy_fault, + .page_mkwrite = sel_mmap_policy_fault, +}; + +int sel_mmap_policy(struct file *filp, struct vm_area_struct *vma) +{ + if (vma->vm_flags & VM_SHARED) { + /* do not allow mprotect to make mapping writable */ + vma->vm_flags &= ~VM_MAYWRITE; + + if (vma->vm_flags & VM_WRITE) + return -EACCES; + } + + vma->vm_flags |= VM_RESERVED; + vma->vm_ops = &sel_mmap_policy_ops; + + return 0; +} + static const struct file_operations sel_policy_ops = { .open = sel_open_policy, .read = sel_read_policy, + .mmap = sel_mmap_policy, .release = sel_release_policy, }; diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c index 7565d16..3a1739b 100644 --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -3169,7 +3169,7 @@ int security_read_policy(void **data, ssize_t *len) *len = security_policydb_len(); - *data = vmalloc(*len); + *data = vmalloc_user(*len); if (!*data) return -ENOMEM; -- cgit v0.10.2 From ff660c80d00b52287f1f67ee6c115dc0057bcdde Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Tue, 19 Oct 2010 18:17:32 -0400 Subject: secmark: fix config problem when CONFIG_NF_CONNTRACK_SECMARK is not set When CONFIG_NF_CONNTRACK_SECMARK is not set we accidentally attempt to use the secmark fielf of struct nf_conn. Problem is when that config isn't set the field doesn't exist. whoops. Wrap the incorrect usage in the config. Signed-off-by: Eric Paris Signed-off-by: James Morris diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index b3c6285..146476c 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -582,9 +582,11 @@ ctnetlink_conntrack_event(unsigned int events, struct nf_ct_event *item) && ctnetlink_dump_helpinfo(skb, ct) < 0) goto nla_put_failure; +#ifdef CONFIG_NF_CONNTRACK_SECMARK if ((events & (1 << IPCT_SECMARK) || ct->secmark) && ctnetlink_dump_secctx(skb, ct) < 0) goto nla_put_failure; +#endif if (events & (1 << IPCT_RELATED) && ctnetlink_dump_master(skb, ct) < 0) -- cgit v0.10.2 From f0d3d9894e43fc68d47948e2c6f03e32da88b799 Mon Sep 17 00:00:00 2001 From: Stephen Rothwell Date: Wed, 20 Oct 2010 16:08:00 +1100 Subject: selinux: include vmalloc.h for vmalloc_user Include vmalloc.h for vmalloc_user (fixes ppc build warning). Acked-by: Eric Paris Signed-off-by: James Morris diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c index 3a1739b..223c1ff 100644 --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -51,6 +51,7 @@ #include #include #include +#include #include #include "flask.h" -- cgit v0.10.2 From 02238959944ce031f066f21e541a14933aca6575 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:40 +0530 Subject: virtio: console: Reset vdev before removing device The virtqueues should be disabled before attempting to remove the device. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 0f69c5e..076d035 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -1605,6 +1605,9 @@ static void virtcons_remove(struct virtio_device *vdev) portdev = vdev->priv; + /* Disable interrupts for vqs */ + vdev->config->reset(vdev); + /* Finish up work that's lined up */ cancel_work_sync(&portdev->control_work); list_for_each_entry_safe(port, port2, &portdev->ports, list) -- cgit v0.10.2 From 96eb872b2a041b1536ccc6ae2fa87eb28f6e2bb2 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:41 +0530 Subject: virtio: console: Remove control vq data only if using multiport support If a portdev isn't using multiport support, it won't have any control vq data to remove. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 076d035..e15dbe7 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -1600,8 +1600,6 @@ static void virtcons_remove(struct virtio_device *vdev) { struct ports_device *portdev; struct port *port, *port2; - struct port_buffer *buf; - unsigned int len; portdev = vdev->priv; @@ -1615,11 +1613,16 @@ static void virtcons_remove(struct virtio_device *vdev) unregister_chrdev(portdev->chr_major, "virtio-portsdev"); - while ((buf = virtqueue_get_buf(portdev->c_ivq, &len))) - free_buf(buf); + if (use_multiport(portdev)) { + struct port_buffer *buf; + unsigned int len; - while ((buf = virtqueue_detach_unused_buf(portdev->c_ivq))) - free_buf(buf); + while ((buf = virtqueue_get_buf(portdev->c_ivq, &len))) + free_buf(buf); + + while ((buf = virtqueue_detach_unused_buf(portdev->c_ivq))) + free_buf(buf); + } vdev->config->del_vqs(vdev); kfree(portdev->in_vqs); -- cgit v0.10.2 From 84ec06c59a14d0941dd58ca6793b24a7e86b3b85 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:42 +0530 Subject: virtio: console: Check if portdev is valid in send_control_msg() A portdev may have been hot-unplugged while a port was open()ed. Skip sending control messages when the portdev isn't valid. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index e15dbe7..e1d382b 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -410,7 +410,10 @@ static ssize_t __send_control_msg(struct ports_device *portdev, u32 port_id, static ssize_t send_control_msg(struct port *port, unsigned int event, unsigned int value) { - return __send_control_msg(port->portdev, port->id, event, value); + /* Did the port get unplugged before userspace closed it? */ + if (port->portdev) + return __send_control_msg(port->portdev, port->id, event, value); + return 0; } /* Callers must take the port->outvq_lock */ -- cgit v0.10.2 From 3709ea7ae7d698b428576c2db0bbb6e08a18cf12 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:43 +0530 Subject: virtio: console: Un-block reads on chardev close If a chardev is closed, any blocked read / poll calls should just return and not attempt to use other state. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index e1d382b..92f1f65 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -528,6 +528,10 @@ static ssize_t fill_readbuf(struct port *port, char *out_buf, size_t out_count, /* The condition that must be true for polling to end */ static bool will_read_block(struct port *port) { + if (!port->guest_connected) { + /* Port got hot-unplugged. Let's exit. */ + return false; + } return !port_has_data(port) && port->host_connected; } -- cgit v0.10.2 From 8529a504273d4efa6bb004dcd6ef28fe67b64ae9 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:44 +0530 Subject: virtio: console: Unblock poll on port hot-unplug When a port is hot-unplugged while an app is blocked on poll(), unblock the poll() and return. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 92f1f65..47b710c 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -663,6 +663,10 @@ static unsigned int port_fops_poll(struct file *filp, poll_table *wait) port = filp->private_data; poll_wait(filp, &port->waitqueue, wait); + if (!port->guest_connected) { + /* Port got unplugged */ + return POLLHUP; + } ret = 0; if (!will_read_block(port)) ret |= POLLIN | POLLRDNORM; -- cgit v0.10.2 From b3dddb9e6ddab74327f5557c1a6640ea0f56ad1c Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:45 +0530 Subject: virtio: console: Make read() return -ENODEV on hot-unplug When a port is hot-unplugged while an app was blocked on a read() call, the call was unblocked but would not get an error returned. Return -ENODEV to ensure the app knows the port has gone away. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 47b710c..e17ecf5 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -582,6 +582,9 @@ static ssize_t port_fops_read(struct file *filp, char __user *ubuf, if (ret < 0) return ret; } + /* Port got hot-unplugged. */ + if (!port->guest_connected) + return -ENODEV; /* * We could've received a disconnection message while we were * waiting for more data. -- cgit v0.10.2 From f4028119714e452f9b49377ec55e0ed1e5d1dfa4 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:46 +0530 Subject: virtio: console: Make write() return -ENODEV on hot-unplug When a port is hot-unplugged while an app was blocked on a write() call, the call was unblocked but would not get an error returned. Return -ENODEV to ensure the app knows the port has gone away. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index e17ecf5..70f1c38 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -626,6 +626,9 @@ static ssize_t port_fops_write(struct file *filp, const char __user *ubuf, if (ret < 0) return ret; } + /* Port got hot-unplugged. */ + if (!port->guest_connected) + return -ENODEV; count = min((size_t)(32 * 1024), count); -- cgit v0.10.2 From 7a2853178dfba9553d58f356113f47fd582e9cc6 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:47 +0530 Subject: virtio: console: remove_port() should return void When a port is removed, we have to assume the port is gone. So a success/failure return value doesn't make sense. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 70f1c38..21b6213 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -1131,7 +1131,7 @@ fail: } /* Remove all port-specific data. */ -static int remove_port(struct port *port) +static void remove_port(struct port *port) { struct port_buffer *buf; @@ -1181,7 +1181,6 @@ static int remove_port(struct port *port) debugfs_remove(port->debugfs_file); kfree(port); - return 0; } /* Any private messages that the Host and Guest want to share */ -- cgit v0.10.2 From 8ad37e83c8dc413f92b10c3d9bdeabe9237f521d Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:48 +0530 Subject: virtio: console: open: Use a common path for error handling Just re-arrange code for future patches. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 21b6213..2f4f0b2 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -711,6 +711,7 @@ static int port_fops_open(struct inode *inode, struct file *filp) { struct cdev *cdev = inode->i_cdev; struct port *port; + int ret; port = container_of(cdev, struct port, cdev); filp->private_data = port; @@ -719,14 +720,17 @@ static int port_fops_open(struct inode *inode, struct file *filp) * Don't allow opening of console port devices -- that's done * via /dev/hvc */ - if (is_console_port(port)) - return -ENXIO; + if (is_console_port(port)) { + ret = -ENXIO; + goto out; + } /* Allow only one process to open a particular port at a time */ spin_lock_irq(&port->inbuf_lock); if (port->guest_connected) { spin_unlock_irq(&port->inbuf_lock); - return -EMFILE; + ret = -EMFILE; + goto out; } port->guest_connected = true; @@ -745,6 +749,8 @@ static int port_fops_open(struct inode *inode, struct file *filp) send_control_msg(filp->private_data, VIRTIO_CONSOLE_PORT_OPEN, 1); return 0; +out: + return ret; } /* -- cgit v0.10.2 From 6bdf2afd02ae12bf8ac93e6d14c4b4dfef7c4c59 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:11:49 +0530 Subject: virtio: console: Add a list of portdevs that are active The virtio_console.c driver is capable of handling multiple devices at a time. Maintain a list of devices for future traversal. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 2f4f0b2..b70fe96 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -48,6 +48,9 @@ struct ports_driver_data { /* Used for exporting per-port information to debugfs */ struct dentry *debugfs_dir; + /* List of all the devices we're handling */ + struct list_head portdevs; + /* Number of devices this driver is handling */ unsigned int index; @@ -108,6 +111,9 @@ struct port_buffer { * ports for that device (vdev->priv). */ struct ports_device { + /* Next portdev in the list, head is in the pdrvdata struct */ + struct list_head list; + /* * Workqueue handlers where we process deferred work after * notification @@ -1599,6 +1605,10 @@ static int __devinit virtcons_probe(struct virtio_device *vdev) add_port(portdev, 0); } + spin_lock_irq(&pdrvdata_lock); + list_add_tail(&portdev->list, &pdrvdata.portdevs); + spin_unlock_irq(&pdrvdata_lock); + __send_control_msg(portdev, VIRTIO_CONSOLE_BAD_ID, VIRTIO_CONSOLE_DEVICE_READY, 1); return 0; @@ -1625,6 +1635,10 @@ static void virtcons_remove(struct virtio_device *vdev) portdev = vdev->priv; + spin_lock_irq(&pdrvdata_lock); + list_del(&portdev->list); + spin_unlock_irq(&pdrvdata_lock); + /* Disable interrupts for vqs */ vdev->config->reset(vdev); /* Finish up work that's lined up */ @@ -1691,6 +1705,7 @@ static int __init init(void) PTR_ERR(pdrvdata.debugfs_dir)); } INIT_LIST_HEAD(&pdrvdata.consoles); + INIT_LIST_HEAD(&pdrvdata.portdevs); return register_virtio_driver(&virtio_console); } -- cgit v0.10.2 From 04950cdf071b6e5aa4794c93ad3e3ce8a1c4aa8c Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:20:58 +0530 Subject: virtio: console: Add a find_port_by_devt() function To convert to using cdev as a pointer to avoid kref troubles, we have to use a different method to get to a port from an inode than the current container_of method. Add find_port_by_devt() that looks up all portdevs and ports with those portdevs to find the right port. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index b70fe96..0c8f2aa 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -227,6 +227,41 @@ out: return port; } +static struct port *find_port_by_devt_in_portdev(struct ports_device *portdev, + dev_t dev) +{ + struct port *port; + unsigned long flags; + + spin_lock_irqsave(&portdev->ports_lock, flags); + list_for_each_entry(port, &portdev->ports, list) + if (port->cdev.dev == dev) + goto out; + port = NULL; +out: + spin_unlock_irqrestore(&portdev->ports_lock, flags); + + return port; +} + +static struct port *find_port_by_devt(dev_t dev) +{ + struct ports_device *portdev; + struct port *port; + unsigned long flags; + + spin_lock_irqsave(&pdrvdata_lock, flags); + list_for_each_entry(portdev, &pdrvdata.portdevs, list) { + port = find_port_by_devt_in_portdev(portdev, dev); + if (port) + goto out; + } + port = NULL; +out: + spin_unlock_irqrestore(&pdrvdata_lock, flags); + return port; +} + static struct port *find_port_by_id(struct ports_device *portdev, u32 id) { struct port *port; @@ -719,7 +754,7 @@ static int port_fops_open(struct inode *inode, struct file *filp) struct port *port; int ret; - port = container_of(cdev, struct port, cdev); + port = find_port_by_devt(cdev->dev); filp->private_data = port; /* -- cgit v0.10.2 From d22a69892bd8f29e3096f6f54c2c00d8aec2e796 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:20:59 +0530 Subject: virtio: console: Use cdev_alloc() instead of cdev_init() This moves to using cdev on the heap instead of it being embedded in the ports struct. This helps individual refcounting and will allow us to properly remove cdev structs after hot-unplugs and close operations. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 0c8f2aa..8a9c140 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -184,7 +184,7 @@ struct port { struct console cons; /* Each port associates with a separate char device */ - struct cdev cdev; + struct cdev *cdev; struct device *dev; /* A waitqueue for poll() or blocking read operations */ @@ -235,7 +235,7 @@ static struct port *find_port_by_devt_in_portdev(struct ports_device *portdev, spin_lock_irqsave(&portdev->ports_lock, flags); list_for_each_entry(port, &portdev->ports, list) - if (port->cdev.dev == dev) + if (port->cdev->dev == dev) goto out; port = NULL; out: @@ -1096,14 +1096,20 @@ static int add_port(struct ports_device *portdev, u32 id) port->in_vq = portdev->in_vqs[port->id]; port->out_vq = portdev->out_vqs[port->id]; - cdev_init(&port->cdev, &port_fops); + port->cdev = cdev_alloc(); + if (!port->cdev) { + dev_err(&port->portdev->vdev->dev, "Error allocating cdev\n"); + err = -ENOMEM; + goto free_port; + } + port->cdev->ops = &port_fops; devt = MKDEV(portdev->chr_major, id); - err = cdev_add(&port->cdev, devt, 1); + err = cdev_add(port->cdev, devt, 1); if (err < 0) { dev_err(&port->portdev->vdev->dev, "Error %d adding cdev for port %u\n", err, id); - goto free_port; + goto free_cdev; } port->dev = device_create(pdrvdata.class, &port->portdev->vdev->dev, devt, port, "vport%up%u", @@ -1168,7 +1174,7 @@ free_inbufs: free_device: device_destroy(pdrvdata.class, port->dev->devt); free_cdev: - cdev_del(&port->cdev); + cdev_del(port->cdev); free_port: kfree(port); fail: @@ -1212,7 +1218,7 @@ static void remove_port(struct port *port) } sysfs_remove_group(&port->dev->kobj, &port_attribute_group); device_destroy(pdrvdata.class, port->dev->devt); - cdev_del(&port->cdev); + cdev_del(port->cdev); /* Remove unused data this port might have received. */ discard_port_data(port); -- cgit v0.10.2 From b353a6b821627053f82b4e7b907e824cb7a6879c Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:38:29 +0530 Subject: virtio: console: Add reference counting for port struct When a port got hot-unplugged, when a port was open, any file operation after the unplugging resulted in a crash. This is fixed by ref-counting the port structure, and releasing it only when the file is closed. This splits the unplug operation in two parts: first marks the port as unavailable, removes all the buffers in the vqs and removes the port from the per-device list of ports. The second stage, invoked when all references drop to zero, releases the chardev and frees all other memory. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 8a9c140..288701c 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -187,6 +187,9 @@ struct port { struct cdev *cdev; struct device *dev; + /* Reference-counting to handle port hot-unplugs and file operations */ + struct kref kref; + /* A waitqueue for poll() or blocking read operations */ wait_queue_head_t waitqueue; @@ -725,6 +728,8 @@ static unsigned int port_fops_poll(struct file *filp, poll_table *wait) return ret; } +static void remove_port(struct kref *kref); + static int port_fops_release(struct inode *inode, struct file *filp) { struct port *port; @@ -745,6 +750,16 @@ static int port_fops_release(struct inode *inode, struct file *filp) reclaim_consumed_buffers(port); spin_unlock_irq(&port->outvq_lock); + /* + * Locks aren't necessary here as a port can't be opened after + * unplug, and if a port isn't unplugged, a kref would already + * exist for the port. Plus, taking ports_lock here would + * create a dependency on other locks taken by functions + * inside remove_port if we're the last holder of the port, + * creating many problems. + */ + kref_put(&port->kref, remove_port); + return 0; } @@ -757,6 +772,11 @@ static int port_fops_open(struct inode *inode, struct file *filp) port = find_port_by_devt(cdev->dev); filp->private_data = port; + /* Prevent against a port getting hot-unplugged at the same time */ + spin_lock_irq(&port->portdev->ports_lock); + kref_get(&port->kref); + spin_unlock_irq(&port->portdev->ports_lock); + /* * Don't allow opening of console port devices -- that's done * via /dev/hvc @@ -791,6 +811,7 @@ static int port_fops_open(struct inode *inode, struct file *filp) return 0; out: + kref_put(&port->kref, remove_port); return ret; } @@ -1079,6 +1100,7 @@ static int add_port(struct ports_device *portdev, u32 id) err = -ENOMEM; goto fail; } + kref_init(&port->kref); port->portdev = portdev; port->id = id; @@ -1183,22 +1205,43 @@ fail: return err; } -/* Remove all port-specific data. */ -static void remove_port(struct port *port) +/* No users remain, remove all port-specific data. */ +static void remove_port(struct kref *kref) +{ + struct port *port; + + port = container_of(kref, struct port, kref); + + sysfs_remove_group(&port->dev->kobj, &port_attribute_group); + device_destroy(pdrvdata.class, port->dev->devt); + cdev_del(port->cdev); + + kfree(port->name); + + debugfs_remove(port->debugfs_file); + + kfree(port); +} + +/* + * Port got unplugged. Remove port from portdev's list and drop the + * kref reference. If no userspace has this port opened, it will + * result in immediate removal the port. + */ +static void unplug_port(struct port *port) { struct port_buffer *buf; + spin_lock_irq(&port->portdev->ports_lock); + list_del(&port->list); + spin_unlock_irq(&port->portdev->ports_lock); + if (port->guest_connected) { port->guest_connected = false; port->host_connected = false; wake_up_interruptible(&port->waitqueue); - send_control_msg(port, VIRTIO_CONSOLE_PORT_OPEN, 0); } - spin_lock_irq(&port->portdev->ports_lock); - list_del(&port->list); - spin_unlock_irq(&port->portdev->ports_lock); - if (is_console_port(port)) { spin_lock_irq(&pdrvdata_lock); list_del(&port->cons.list); @@ -1216,9 +1259,6 @@ static void remove_port(struct port *port) hvc_remove(port->cons.hvc); #endif } - sysfs_remove_group(&port->dev->kobj, &port_attribute_group); - device_destroy(pdrvdata.class, port->dev->devt); - cdev_del(port->cdev); /* Remove unused data this port might have received. */ discard_port_data(port); @@ -1229,11 +1269,19 @@ static void remove_port(struct port *port) while ((buf = virtqueue_detach_unused_buf(port->in_vq))) free_buf(buf); - kfree(port->name); - - debugfs_remove(port->debugfs_file); + /* + * We should just assume the device itself has gone off -- + * else a close on an open port later will try to send out a + * control message. + */ + port->portdev = NULL; - kfree(port); + /* + * Locks around here are not necessary - a port can't be + * opened after we removed the port struct from ports_list + * above. + */ + kref_put(&port->kref, remove_port); } /* Any private messages that the Host and Guest want to share */ @@ -1272,7 +1320,7 @@ static void handle_control_message(struct ports_device *portdev, add_port(portdev, cpkt->id); break; case VIRTIO_CONSOLE_PORT_REMOVE: - remove_port(port); + unplug_port(port); break; case VIRTIO_CONSOLE_CONSOLE_PORT: if (!cpkt->value) @@ -1686,7 +1734,7 @@ static void virtcons_remove(struct virtio_device *vdev) cancel_work_sync(&portdev->control_work); list_for_each_entry_safe(port, port2, &portdev->ports, list) - remove_port(port); + unplug_port(port); unregister_chrdev(portdev->chr_major, "virtio-portsdev"); -- cgit v0.10.2 From e062013c7d22e40ee634b818d28fd615db36998e Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:38:30 +0530 Subject: virtio: console: Reference counting portdev structs is not needed Explain in a comment why there's no need to reference-count the portdev struct: when a device is yanked out, we can't do anything more with it anyway so just give up doing anything more with the data or the vqs and exit cleanly. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 288701c..c84486b 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -1738,6 +1738,14 @@ static void virtcons_remove(struct virtio_device *vdev) unregister_chrdev(portdev->chr_major, "virtio-portsdev"); + /* + * When yanking out a device, we immediately lose the + * (device-side) queues. So there's no point in keeping the + * guest side around till we drop our final reference. This + * also means that any ports which are in an open state will + * have to just stop using the port, as the vqs are going + * away. + */ if (use_multiport(portdev)) { struct port_buffer *buf; unsigned int len; -- cgit v0.10.2 From 3eae0adea949d8fdd8fa3e5301192901219d2c64 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:47:52 +0530 Subject: virtio: console: Send SIGIO to processes that request it for host events A process can request for SIGIO on host connect / disconnect events using the O_ASYNC file flag using fcntl(). If that's requested, and if the guest-side connection for the port is open, any host-side open/close events for that port will raise a SIGIO. The process can then use poll() within the signal handler to find out which port triggered the signal. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index c84486b..c4ca436 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -196,6 +196,9 @@ struct port { /* The 'name' of the port that we expose via sysfs properties */ char *name; + /* We can notify apps of host connect / disconnect events via SIGIO */ + struct fasync_struct *async_queue; + /* The 'id' to identify the port with the Host */ u32 id; @@ -815,6 +818,14 @@ out: return ret; } +static int port_fops_fasync(int fd, struct file *filp, int mode) +{ + struct port *port; + + port = filp->private_data; + return fasync_helper(fd, filp, mode, &port->async_queue); +} + /* * The file operations that we support: programs in the guest can open * a console device, read from it, write to it, poll for data and @@ -828,6 +839,7 @@ static const struct file_operations port_fops = { .write = port_fops_write, .poll = port_fops_poll, .release = port_fops_release, + .fasync = port_fops_fasync, }; /* @@ -1086,6 +1098,12 @@ static unsigned int fill_queue(struct virtqueue *vq, spinlock_t *lock) return nr_added_bufs; } +static void send_sigio_to_port(struct port *port) +{ + if (port->async_queue && port->guest_connected) + kill_fasync(&port->async_queue, SIGIO, POLL_OUT); +} + static int add_port(struct ports_device *portdev, u32 id) { char debugfs_name[16]; @@ -1108,6 +1126,7 @@ static int add_port(struct ports_device *portdev, u32 id) port->name = NULL; port->inbuf = NULL; port->cons.hvc = NULL; + port->async_queue = NULL; port->cons.ws.ws_row = port->cons.ws.ws_col = 0; @@ -1362,6 +1381,12 @@ static void handle_control_message(struct ports_device *portdev, spin_lock_irq(&port->outvq_lock); reclaim_consumed_buffers(port); spin_unlock_irq(&port->outvq_lock); + + /* + * If the guest is connected, it'll be interested in + * knowing the host connection state changed. + */ + send_sigio_to_port(port); break; case VIRTIO_CONSOLE_PORT_NAME: /* -- cgit v0.10.2 From 55f6bcce3691f68476a530daa6666b66c43420a8 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:47:53 +0530 Subject: virtio: console: Send SIGIO on new data arrival on ports Send a SIGIO signal when new data arrives on a port. This is sent only when the process has requested for the signal to be sent using fcntl(). Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index c4ca436..6d5c579 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -1483,6 +1483,9 @@ static void in_intr(struct virtqueue *vq) wake_up_interruptible(&port->waitqueue); + /* Send a SIGIO indicating new data in case the process asked for it */ + send_sigio_to_port(port); + if (is_console_port(port) && hvc_poll(port->cons.hvc)) hvc_kick(); } -- cgit v0.10.2 From a461e11e7b8ca2705889bcf9582f6a8f84884bd2 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 2 Sep 2010 18:47:54 +0530 Subject: virtio: console: Send SIGIO in case of port unplug If a port has registered for SIGIO signals, let the application know that the port is getting unplugged. Signed-off-by: Amit Shah Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 6d5c579..dc42d56 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -1259,6 +1259,9 @@ static void unplug_port(struct port *port) port->guest_connected = false; port->host_connected = false; wake_up_interruptible(&port->waitqueue); + + /* Let the app know the port is going down. */ + send_sigio_to_port(port); } if (is_console_port(port)) { -- cgit v0.10.2 From 299fb61c08c2fcd1bb6d3a4e87e53dc368475416 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 16 Sep 2010 14:43:09 +0530 Subject: virtio: console: Disable lseek(2) for port file operations The ports are char devices; do not have seeking capabilities. Calling nonseekable_open() from the fops_open() call and setting the llseek fops pointer to no_llseek ensures an lseek() call from userspace returns -ESPIPE. Signed-off-by: Amit Shah CC: Arnd Bergmann Signed-off-by: Rusty Russell diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index dc42d56..6c1b676 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -809,6 +809,8 @@ static int port_fops_open(struct inode *inode, struct file *filp) reclaim_consumed_buffers(port); spin_unlock_irq(&port->outvq_lock); + nonseekable_open(inode, filp); + /* Notify host of port being opened */ send_control_msg(filp->private_data, VIRTIO_CONSOLE_PORT_OPEN, 1); @@ -840,6 +842,7 @@ static const struct file_operations port_fops = { .poll = port_fops_poll, .release = port_fops_release, .fasync = port_fops_fasync, + .llseek = no_llseek, }; /* -- cgit v0.10.2 From fe5a50a10c7b3bf38e47045eadc6b87a5c157f44 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 15 Sep 2010 01:27:23 +0200 Subject: virtio_blk: remove BKL leftovers Remove the BKL usage added in "block: push down BKL into .locked_ioctl". Virtio-blk doesn't use the BKL for anything, and doesn't implement any ioctl command by itself, but only uses the generic scsi_cmd_ioctl which is fine without the BKL. Signed-off-by: Christoph Hellwig Signed-off-by: Rusty Russell diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c index 1101e25..8320490 100644 --- a/drivers/block/virtio_blk.c +++ b/drivers/block/virtio_blk.c @@ -2,7 +2,6 @@ #include #include #include -#include #include #include #include @@ -222,8 +221,8 @@ static int virtblk_get_id(struct gendisk *disk, char *id_str) return err; } -static int virtblk_locked_ioctl(struct block_device *bdev, fmode_t mode, - unsigned cmd, unsigned long data) +static int virtblk_ioctl(struct block_device *bdev, fmode_t mode, + unsigned int cmd, unsigned long data) { struct gendisk *disk = bdev->bd_disk; struct virtio_blk *vblk = disk->private_data; @@ -238,18 +237,6 @@ static int virtblk_locked_ioctl(struct block_device *bdev, fmode_t mode, (void __user *)data); } -static int virtblk_ioctl(struct block_device *bdev, fmode_t mode, - unsigned int cmd, unsigned long param) -{ - int ret; - - lock_kernel(); - ret = virtblk_locked_ioctl(bdev, mode, cmd, param); - unlock_kernel(); - - return ret; -} - /* We provide getgeo only to please some old bootloader/partitioning tools */ static int virtblk_getgeo(struct block_device *bd, struct hd_geometry *geo) { -- cgit v0.10.2 From 2aeb66d3036dbafc297ac553a257a40283dadb3e Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Thu, 21 Oct 2010 00:15:00 -0700 Subject: x86-32, percpu: Correct the ordering of the percpu readmostly section Checkin c957ef2c59e952803766ddc22e89981ab534606f had inconsistent ordering of .data..percpu..page_aligned and .data..percpu..readmostly; the still-broken version affected x86-32 at least. The page aligned version really must be page aligned... Signed-off-by: H. Peter Anvin LKML-Reference: <1287544022.4571.7.camel@sli10-conroe.sh.intel.com> Cc: Shaohua Li Cc: Eric Dumazet diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index d7e7b21..1457b81 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -706,8 +706,8 @@ VMLINUX_SYMBOL(__per_cpu_start) = .; \ *(.data..percpu..first) \ . = ALIGN(PAGE_SIZE); \ - *(.data..percpu..readmostly) \ *(.data..percpu..page_aligned) \ + *(.data..percpu..readmostly) \ *(.data..percpu) \ *(.data..percpu..shared_aligned) \ VMLINUX_SYMBOL(__per_cpu_end) = .; \ -- cgit v0.10.2 From ca755e0a49ff1272efff0b3bfdf3f1e0b0fc5d57 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 1 Sep 2010 16:32:20 +0200 Subject: EDAC: Fix error return We should return a negative value when we cannot get the toplevel edac sysfs class. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c index 8aad94d..aa93ad8 100644 --- a/drivers/edac/edac_mc_sysfs.c +++ b/drivers/edac/edac_mc_sysfs.c @@ -1011,7 +1011,7 @@ void edac_remove_sysfs_mci_device(struct mem_ctl_info *mci) */ int edac_sysfs_setup_mc_kset(void) { - int err = 0; + int err = -EINVAL; struct sysdev_class *edac_class; debugf1("%s()\n", __func__); -- cgit v0.10.2 From c9f281fd96b29367363ee232021c030d025c52a8 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 18 Aug 2010 18:21:42 +0200 Subject: EDAC, MCE: Add HW_ERR prefix .. so that the user knows what she's looking at there in dmesg. Also, fix a minor cosmetic output inconsistency. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c index 9014df6..564fe7c 100644 --- a/drivers/edac/edac_mce_amd.c +++ b/drivers/edac/edac_mce_amd.c @@ -133,7 +133,7 @@ static void amd_decode_dc_mce(u64 mc0_status) u32 ec = mc0_status & 0xffff; u32 xec = (mc0_status >> 16) & 0xf; - pr_emerg("Data Cache Error"); + pr_emerg(HW_ERR "Data Cache Error: "); if (xec == 1 && TLB_ERROR(ec)) pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); @@ -168,7 +168,7 @@ static void amd_decode_dc_mce(u64 mc0_status) return; wrong_dc_mce: - pr_warning("Corrupted DC MCE info?\n"); + pr_emerg(HW_ERR "Corrupted DC MCE info?\n"); } static void amd_decode_ic_mce(u64 mc1_status) @@ -176,7 +176,7 @@ static void amd_decode_ic_mce(u64 mc1_status) u32 ec = mc1_status & 0xffff; u32 xec = (mc1_status >> 16) & 0xf; - pr_emerg("Instruction Cache Error"); + pr_emerg(HW_ERR "Instruction Cache Error"); if (xec == 1 && TLB_ERROR(ec)) pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); @@ -225,7 +225,7 @@ static void amd_decode_ic_mce(u64 mc1_status) return; wrong_ic_mce: - pr_warning("Corrupted IC MCE info?\n"); + pr_emerg(HW_ERR "Corrupted IC MCE info?\n"); } static void amd_decode_bu_mce(u64 mc2_status) @@ -233,7 +233,7 @@ static void amd_decode_bu_mce(u64 mc2_status) u32 ec = mc2_status & 0xffff; u32 xec = (mc2_status >> 16) & 0xf; - pr_emerg("Bus Unit Error"); + pr_emerg(HW_ERR "Bus Unit Error"); if (xec == 0x1) pr_cont(" in the write data buffers.\n"); @@ -267,7 +267,7 @@ static void amd_decode_bu_mce(u64 mc2_status) return; wrong_bu_mce: - pr_warning("Corrupted BU MCE info?\n"); + pr_emerg(HW_ERR "Corrupted BU MCE info?\n"); } static void amd_decode_ls_mce(u64 mc3_status) @@ -275,7 +275,7 @@ static void amd_decode_ls_mce(u64 mc3_status) u32 ec = mc3_status & 0xffff; u32 xec = (mc3_status >> 16) & 0xf; - pr_emerg("Load Store Error"); + pr_emerg(HW_ERR "Load Store Error"); if (xec == 0x0) { u8 rrrr = (ec >> 4) & 0xf; @@ -288,7 +288,7 @@ static void amd_decode_ls_mce(u64 mc3_status) return; wrong_ls_mce: - pr_warning("Corrupted LS MCE info?\n"); + pr_emerg(HW_ERR "Corrupted LS MCE info?\n"); } void amd_decode_nb_mce(int node_id, struct err_regs *regs, int handle_errors) @@ -304,7 +304,7 @@ void amd_decode_nb_mce(int node_id, struct err_regs *regs, int handle_errors) if (TLB_ERROR(ec) && !report_gart_errors) return; - pr_emerg("Northbridge Error, node %d", node_id); + pr_emerg(HW_ERR "Northbridge Error, node %d", node_id); /* * F10h, revD can disable ErrCpu[3:0] so check that first and also the @@ -323,7 +323,7 @@ void amd_decode_nb_mce(int node_id, struct err_regs *regs, int handle_errors) pr_cont("\n"); } - pr_emerg("%s.\n", EXT_ERR_MSG(regs->nbsl)); + pr_emerg(HW_ERR "%s.\n", EXT_ERR_MSG(regs->nbsl)); if (BUS_ERROR(ec) && nb_bus_decoder) nb_bus_decoder(node_id, regs); @@ -334,26 +334,26 @@ static void amd_decode_fr_mce(u64 mc5_status) { /* we have only one error signature so match all fields at once. */ if ((mc5_status & 0xffff) == 0x0f0f) - pr_emerg(" FR Error: CPU Watchdog timer expire.\n"); + pr_emerg(HW_ERR " FR Error: CPU Watchdog timer expire.\n"); else - pr_warning("Corrupted FR MCE info?\n"); + pr_emerg(HW_ERR "Corrupted FR MCE info?\n"); } static inline void amd_decode_err_code(unsigned int ec) { if (TLB_ERROR(ec)) { - pr_emerg("Transaction: %s, Cache Level %s\n", + pr_emerg(HW_ERR "Transaction: %s, Cache Level: %s\n", TT_MSG(ec), LL_MSG(ec)); } else if (MEM_ERROR(ec)) { - pr_emerg("Transaction: %s, Type: %s, Cache Level: %s", + pr_emerg(HW_ERR "Transaction: %s, Type: %s, Cache Level: %s", RRRR_MSG(ec), TT_MSG(ec), LL_MSG(ec)); } else if (BUS_ERROR(ec)) { - pr_emerg("Transaction type: %s(%s), %s, Cache Level: %s, " + pr_emerg(HW_ERR "Transaction type: %s(%s), %s, Cache Level: %s, " "Participating Processor: %s\n", RRRR_MSG(ec), II_MSG(ec), TO_MSG(ec), LL_MSG(ec), PP_MSG(ec)); } else - pr_warning("Huh? Unknown MCE error 0x%x\n", ec); + pr_emerg(HW_ERR "Huh? Unknown MCE error 0x%x\n", ec); } static int amd_decode_mce(struct notifier_block *nb, unsigned long val, @@ -363,7 +363,7 @@ static int amd_decode_mce(struct notifier_block *nb, unsigned long val, struct err_regs regs; int node, ecc; - pr_emerg("MC%d_STATUS: ", m->bank); + pr_emerg(HW_ERR "MC%d_STATUS: ", m->bank); pr_cont("%sorrected error, other errors lost: %s, " "CPU context corrupt: %s", -- cgit v0.10.2 From 0ee8efa8f4672ce35ee370291c0f21d7b87b1e3f Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Mon, 30 Aug 2010 12:34:19 +0200 Subject: EDAC, MCE: Remove unused function parameter Remove remains from previous functionality. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/amd64_edac_dbg.c b/drivers/edac/amd64_edac_dbg.c index 59cf2cf..22ef3fe 100644 --- a/drivers/edac/amd64_edac_dbg.c +++ b/drivers/edac/amd64_edac_dbg.c @@ -24,7 +24,7 @@ static ssize_t amd64_nbea_store(struct mem_ctl_info *mci, const char *data, /* Process the Mapping request */ /* TODO: Add race prevention */ - amd_decode_nb_mce(pvt->mc_node_id, &pvt->ctl_error_info, 1); + amd_decode_nb_mce(pvt->mc_node_id, &pvt->ctl_error_info); return count; } diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c index 564fe7c..765d7fb 100644 --- a/drivers/edac/edac_mce_amd.c +++ b/drivers/edac/edac_mce_amd.c @@ -291,13 +291,10 @@ wrong_ls_mce: pr_emerg(HW_ERR "Corrupted LS MCE info?\n"); } -void amd_decode_nb_mce(int node_id, struct err_regs *regs, int handle_errors) +void amd_decode_nb_mce(int node_id, struct err_regs *regs) { u32 ec = ERROR_CODE(regs->nbsl); - if (!handle_errors) - return; - /* * GART TLB error reporting is disabled by default. Bail out early. */ @@ -402,7 +399,7 @@ static int amd_decode_mce(struct notifier_block *nb, unsigned long val, regs.nbeah = (u32)(m->addr >> 32); node = amd_get_nb_id(m->extcpu); - amd_decode_nb_mce(node, ®s, 1); + amd_decode_nb_mce(node, ®s); break; case 5: diff --git a/drivers/edac/edac_mce_amd.h b/drivers/edac/edac_mce_amd.h index df23ee0..8920133 100644 --- a/drivers/edac/edac_mce_amd.h +++ b/drivers/edac/edac_mce_amd.h @@ -64,6 +64,6 @@ struct err_regs { void amd_report_gart_errors(bool); void amd_register_ecc_decoder(void (*f)(int, struct err_regs *)); void amd_unregister_ecc_decoder(void (*f)(int, struct err_regs *)); -void amd_decode_nb_mce(int, struct err_regs *, int); +void amd_decode_nb_mce(int, struct err_regs *); #endif /* _EDAC_MCE_AMD_H */ -- cgit v0.10.2 From 6337583d7dc0dced36ab98dd63de2389c95c22d9 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Mon, 6 Sep 2010 18:13:39 +0200 Subject: EDAC, MCE: Sanitize error codes Clean up error codes names, shorten to mnemonics, add RRRR boundary checking. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c index 765d7fb..d0e850e 100644 --- a/drivers/edac/edac_mce_amd.c +++ b/drivers/edac/edac_mce_amd.c @@ -30,62 +30,31 @@ EXPORT_SYMBOL_GPL(amd_unregister_ecc_decoder); * string representation for the different MCA reported error types, see F3x48 * or MSR0000_0411. */ -const char *tt_msgs[] = { /* transaction type */ - "instruction", - "data", - "generic", - "reserved" -}; + +/* transaction type */ +const char *tt_msgs[] = { "INSN", "DATA", "GEN", "RESV" }; EXPORT_SYMBOL_GPL(tt_msgs); -const char *ll_msgs[] = { /* cache level */ - "L0", - "L1", - "L2", - "L3/generic" -}; +/* cache level */ +const char *ll_msgs[] = { "RESV", "L1", "L2", "L3/GEN" }; EXPORT_SYMBOL_GPL(ll_msgs); +/* memory transaction type */ const char *rrrr_msgs[] = { - "generic", - "generic read", - "generic write", - "data read", - "data write", - "inst fetch", - "prefetch", - "evict", - "snoop", - "reserved RRRR= 9", - "reserved RRRR= 10", - "reserved RRRR= 11", - "reserved RRRR= 12", - "reserved RRRR= 13", - "reserved RRRR= 14", - "reserved RRRR= 15" + "GEN", "RD", "WR", "DRD", "DWR", "IRD", "PRF", "EV", "SNP" }; EXPORT_SYMBOL_GPL(rrrr_msgs); -const char *pp_msgs[] = { /* participating processor */ - "local node originated (SRC)", - "local node responded to request (RES)", - "local node observed as 3rd party (OBS)", - "generic" -}; +/* participating processor */ +const char *pp_msgs[] = { "SRC", "RES", "OBS", "GEN" }; EXPORT_SYMBOL_GPL(pp_msgs); -const char *to_msgs[] = { - "no timeout", - "timed out" -}; +/* request timeout */ +const char *to_msgs[] = { "no timeout", "timed out" }; EXPORT_SYMBOL_GPL(to_msgs); -const char *ii_msgs[] = { /* memory or i/o */ - "mem access", - "reserved", - "i/o access", - "generic" -}; +/* memory or i/o */ +const char *ii_msgs[] = { "MEM", "RESV", "IO", "GEN" }; EXPORT_SYMBOL_GPL(ii_msgs); /* @@ -336,16 +305,16 @@ static void amd_decode_fr_mce(u64 mc5_status) pr_emerg(HW_ERR "Corrupted FR MCE info?\n"); } -static inline void amd_decode_err_code(unsigned int ec) +static inline void amd_decode_err_code(u16 ec) { if (TLB_ERROR(ec)) { pr_emerg(HW_ERR "Transaction: %s, Cache Level: %s\n", TT_MSG(ec), LL_MSG(ec)); } else if (MEM_ERROR(ec)) { - pr_emerg(HW_ERR "Transaction: %s, Type: %s, Cache Level: %s", + pr_emerg(HW_ERR "Transaction: %s, Type: %s, Cache Level: %s\n", RRRR_MSG(ec), TT_MSG(ec), LL_MSG(ec)); } else if (BUS_ERROR(ec)) { - pr_emerg(HW_ERR "Transaction type: %s(%s), %s, Cache Level: %s, " + pr_emerg(HW_ERR "Transaction: %s (%s), %s, Cache Level: %s, " "Participating Processor: %s\n", RRRR_MSG(ec), II_MSG(ec), TO_MSG(ec), LL_MSG(ec), PP_MSG(ec)); diff --git a/drivers/edac/edac_mce_amd.h b/drivers/edac/edac_mce_amd.h index 8920133..2ee499d 100644 --- a/drivers/edac/edac_mce_amd.h +++ b/drivers/edac/edac_mce_amd.h @@ -20,13 +20,14 @@ #define II_MSG(x) ii_msgs[II(x)] #define LL(x) (((x) >> 0) & 0x3) #define LL_MSG(x) ll_msgs[LL(x)] -#define RRRR(x) (((x) >> 4) & 0xf) -#define RRRR_MSG(x) rrrr_msgs[RRRR(x)] #define TO(x) (((x) >> 8) & 0x1) #define TO_MSG(x) to_msgs[TO(x)] #define PP(x) (((x) >> 9) & 0x3) #define PP_MSG(x) pp_msgs[PP(x)] +#define RRRR(x) (((x) >> 4) & 0xf) +#define RRRR_MSG(x) ((RRRR(x) < 9) ? rrrr_msgs[RRRR(x)] : "Wrong R4!") + #define K8_NBSH 0x4C #define K8_NBSH_VALID_BIT BIT(31) -- cgit v0.10.2 From 7cfd4a87441f5ca3018fdd1f7ad67e8a73a05dc2 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 1 Sep 2010 14:45:20 +0200 Subject: EDAC, MCE: Pass complete MCE info to decoders ... instead of the MCi_STATUS info only for improved handling of certain types of errors later. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c index e7d5d6b..76f7cc0 100644 --- a/drivers/edac/amd64_edac.c +++ b/drivers/edac/amd64_edac.c @@ -2073,11 +2073,18 @@ static inline void __amd64_decode_bus_error(struct mem_ctl_info *mci, amd64_handle_ue(mci, info); } -void amd64_decode_bus_error(int node_id, struct err_regs *regs) +void amd64_decode_bus_error(int node_id, struct mce *m, u32 nbcfg) { struct mem_ctl_info *mci = mci_lookup[node_id]; + struct err_regs regs; - __amd64_decode_bus_error(mci, regs); + regs.nbsl = (u32) m->status; + regs.nbsh = (u32)(m->status >> 32); + regs.nbeal = (u32) m->addr; + regs.nbeah = (u32)(m->addr >> 32); + regs.nbcfg = nbcfg; + + __amd64_decode_bus_error(mci, ®s); /* * Check the UE bit of the NB status high register, if set generate some @@ -2086,7 +2093,7 @@ void amd64_decode_bus_error(int node_id, struct err_regs *regs) * * FIXME: this should go somewhere else, if at all. */ - if (regs->nbsh & K8_NBSH_UC_ERR && !report_gart_errors) + if (regs.nbsh & K8_NBSH_UC_ERR && !report_gart_errors) edac_mc_handle_ue_no_info(mci, "UE bit is set"); } diff --git a/drivers/edac/amd64_edac_dbg.c b/drivers/edac/amd64_edac_dbg.c index 22ef3fe..f6d5695 100644 --- a/drivers/edac/amd64_edac_dbg.c +++ b/drivers/edac/amd64_edac_dbg.c @@ -10,11 +10,14 @@ static ssize_t amd64_nbea_store(struct mem_ctl_info *mci, const char *data, size_t count) { struct amd64_pvt *pvt = mci->pvt_info; - unsigned long long value; + u64 value; int ret = 0; + struct mce m; ret = strict_strtoull(data, 16, &value); if (ret != -EINVAL) { + struct err_regs *regs = &pvt->ctl_error_info; + debugf0("received NBEA= 0x%llx\n", value); /* place the value into the virtual error packet */ @@ -22,9 +25,12 @@ static ssize_t amd64_nbea_store(struct mem_ctl_info *mci, const char *data, value >>= 32; pvt->ctl_error_info.nbeah = (u32) value; + m.addr = value; + m.status = regs->nbsl | ((u64)regs->nbsh << 32); + /* Process the Mapping request */ /* TODO: Add race prevention */ - amd_decode_nb_mce(pvt->mc_node_id, &pvt->ctl_error_info); + amd_decode_nb_mce(pvt->mc_node_id, &m, regs->nbcfg); return count; } diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c index d0e850e..6cfa881 100644 --- a/drivers/edac/edac_mce_amd.c +++ b/drivers/edac/edac_mce_amd.c @@ -2,7 +2,7 @@ #include "edac_mce_amd.h" static bool report_gart_errors; -static void (*nb_bus_decoder)(int node_id, struct err_regs *regs); +static void (*nb_bus_decoder)(int node_id, struct mce *m, u32 nbcfg); void amd_report_gart_errors(bool v) { @@ -10,13 +10,13 @@ void amd_report_gart_errors(bool v) } EXPORT_SYMBOL_GPL(amd_report_gart_errors); -void amd_register_ecc_decoder(void (*f)(int, struct err_regs *)) +void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)) { nb_bus_decoder = f; } EXPORT_SYMBOL_GPL(amd_register_ecc_decoder); -void amd_unregister_ecc_decoder(void (*f)(int, struct err_regs *)) +void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)) { if (nb_bus_decoder) { WARN_ON(nb_bus_decoder != f); @@ -97,17 +97,17 @@ const char *ext_msgs[] = { }; EXPORT_SYMBOL_GPL(ext_msgs); -static void amd_decode_dc_mce(u64 mc0_status) +static void amd_decode_dc_mce(struct mce *m) { - u32 ec = mc0_status & 0xffff; - u32 xec = (mc0_status >> 16) & 0xf; + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; pr_emerg(HW_ERR "Data Cache Error: "); if (xec == 1 && TLB_ERROR(ec)) pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); else if (xec == 0) { - if (mc0_status & (1ULL << 40)) + if (m->status & (1ULL << 40)) pr_cont(" during Data Scrub.\n"); else if (TLB_ERROR(ec)) pr_cont(": %s TLB parity error.\n", LL_MSG(ec)); @@ -140,10 +140,10 @@ wrong_dc_mce: pr_emerg(HW_ERR "Corrupted DC MCE info?\n"); } -static void amd_decode_ic_mce(u64 mc1_status) +static void amd_decode_ic_mce(struct mce *m) { - u32 ec = mc1_status & 0xffff; - u32 xec = (mc1_status >> 16) & 0xf; + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; pr_emerg(HW_ERR "Instruction Cache Error"); @@ -154,7 +154,7 @@ static void amd_decode_ic_mce(u64 mc1_status) pr_cont(": %s TLB Parity error.\n", LL_MSG(ec)); else if (BUS_ERROR(ec)) { if (boot_cpu_data.x86 == 0xf && - (mc1_status & (1ULL << 58))) + (m->status & BIT(58))) pr_cont(" during system linefill.\n"); else pr_cont(" during attempted NB data read.\n"); @@ -197,10 +197,10 @@ wrong_ic_mce: pr_emerg(HW_ERR "Corrupted IC MCE info?\n"); } -static void amd_decode_bu_mce(u64 mc2_status) +static void amd_decode_bu_mce(struct mce *m) { - u32 ec = mc2_status & 0xffff; - u32 xec = (mc2_status >> 16) & 0xf; + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; pr_emerg(HW_ERR "Bus Unit Error"); @@ -239,10 +239,10 @@ wrong_bu_mce: pr_emerg(HW_ERR "Corrupted BU MCE info?\n"); } -static void amd_decode_ls_mce(u64 mc3_status) +static void amd_decode_ls_mce(struct mce *m) { - u32 ec = mc3_status & 0xffff; - u32 xec = (mc3_status >> 16) & 0xf; + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; pr_emerg(HW_ERR "Load Store Error"); @@ -260,9 +260,11 @@ wrong_ls_mce: pr_emerg(HW_ERR "Corrupted LS MCE info?\n"); } -void amd_decode_nb_mce(int node_id, struct err_regs *regs) +void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg) { - u32 ec = ERROR_CODE(regs->nbsl); + u32 ec = m->status & 0xffff; + u32 nbsh = (u32)(m->status >> 32); + u32 nbsl = (u32)m->status; /* * GART TLB error reporting is disabled by default. Bail out early. @@ -278,10 +280,10 @@ void amd_decode_nb_mce(int node_id, struct err_regs *regs) */ if ((boot_cpu_data.x86 == 0x10) && (boot_cpu_data.x86_model > 7)) { - if (regs->nbsh & K8_NBSH_ERR_CPU_VAL) - pr_cont(", core: %u\n", (u8)(regs->nbsh & 0xf)); + if (nbsh & K8_NBSH_ERR_CPU_VAL) + pr_cont(", core: %u\n", (u8)(nbsh & 0xf)); } else { - u8 assoc_cpus = regs->nbsh & 0xf; + u8 assoc_cpus = nbsh & 0xf; if (assoc_cpus > 0) pr_cont(", core: %d", fls(assoc_cpus) - 1); @@ -289,17 +291,17 @@ void amd_decode_nb_mce(int node_id, struct err_regs *regs) pr_cont("\n"); } - pr_emerg(HW_ERR "%s.\n", EXT_ERR_MSG(regs->nbsl)); + pr_emerg(HW_ERR "%s.\n", EXT_ERR_MSG(nbsl)); if (BUS_ERROR(ec) && nb_bus_decoder) - nb_bus_decoder(node_id, regs); + nb_bus_decoder(node_id, m, nbcfg); } EXPORT_SYMBOL_GPL(amd_decode_nb_mce); -static void amd_decode_fr_mce(u64 mc5_status) +static void amd_decode_fr_mce(struct mce *m) { /* we have only one error signature so match all fields at once. */ - if ((mc5_status & 0xffff) == 0x0f0f) + if ((m->status & 0xffff) == 0x0f0f) pr_emerg(HW_ERR " FR Error: CPU Watchdog timer expire.\n"); else pr_emerg(HW_ERR "Corrupted FR MCE info?\n"); @@ -326,7 +328,6 @@ static int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data) { struct mce *m = (struct mce *)data; - struct err_regs regs; int node, ecc; pr_emerg(HW_ERR "MC%d_STATUS: ", m->bank); @@ -346,33 +347,28 @@ static int amd_decode_mce(struct notifier_block *nb, unsigned long val, switch (m->bank) { case 0: - amd_decode_dc_mce(m->status); + amd_decode_dc_mce(m); break; case 1: - amd_decode_ic_mce(m->status); + amd_decode_ic_mce(m); break; case 2: - amd_decode_bu_mce(m->status); + amd_decode_bu_mce(m); break; case 3: - amd_decode_ls_mce(m->status); + amd_decode_ls_mce(m); break; case 4: - regs.nbsl = (u32) m->status; - regs.nbsh = (u32)(m->status >> 32); - regs.nbeal = (u32) m->addr; - regs.nbeah = (u32)(m->addr >> 32); - node = amd_get_nb_id(m->extcpu); - - amd_decode_nb_mce(node, ®s); + node = amd_get_nb_id(m->extcpu); + amd_decode_nb_mce(node, m, 0); break; case 5: - amd_decode_fr_mce(m->status); + amd_decode_fr_mce(m); break; default: diff --git a/drivers/edac/edac_mce_amd.h b/drivers/edac/edac_mce_amd.h index 2ee499d..0fba0e76 100644 --- a/drivers/edac/edac_mce_amd.h +++ b/drivers/edac/edac_mce_amd.h @@ -63,8 +63,8 @@ struct err_regs { void amd_report_gart_errors(bool); -void amd_register_ecc_decoder(void (*f)(int, struct err_regs *)); -void amd_unregister_ecc_decoder(void (*f)(int, struct err_regs *)); -void amd_decode_nb_mce(int, struct err_regs *); +void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)); +void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)); +void amd_decode_nb_mce(int, struct mce *, u32); #endif /* _EDAC_MCE_AMD_H */ -- cgit v0.10.2 From 30e1f7a8122145f44f45c95366e27b6bb0b08428 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 2 Sep 2010 17:26:48 +0200 Subject: EDAC: Export edac sysfs class to users. Move toplevel sysfs class to the stub and make it available to non-modularized code too. Add proper refcounting of its users and move the registration functionality into the reference counting routines. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/edac_device_sysfs.c b/drivers/edac/edac_device_sysfs.c index 0709681..2941dca 100644 --- a/drivers/edac/edac_device_sysfs.c +++ b/drivers/edac/edac_device_sysfs.c @@ -13,6 +13,7 @@ #include #include #include +#include #include "edac_core.h" #include "edac_module.h" @@ -235,7 +236,7 @@ int edac_device_register_sysfs_main_kobj(struct edac_device_ctl_info *edac_dev) debugf1("%s()\n", __func__); /* get the /sys/devices/system/edac reference */ - edac_class = edac_get_edac_class(); + edac_class = edac_get_sysfs_class(); if (edac_class == NULL) { debugf1("%s() no edac_class error\n", __func__); err = -ENODEV; @@ -255,7 +256,7 @@ int edac_device_register_sysfs_main_kobj(struct edac_device_ctl_info *edac_dev) if (!try_module_get(edac_dev->owner)) { err = -ENODEV; - goto err_out; + goto err_mod_get; } /* register */ @@ -282,6 +283,9 @@ int edac_device_register_sysfs_main_kobj(struct edac_device_ctl_info *edac_dev) err_kobj_reg: module_put(edac_dev->owner); +err_mod_get: + edac_put_sysfs_class(); + err_out: return err; } @@ -290,12 +294,11 @@ err_out: * edac_device_unregister_sysfs_main_kobj: * the '..../edac/' kobject */ -void edac_device_unregister_sysfs_main_kobj( - struct edac_device_ctl_info *edac_dev) +void edac_device_unregister_sysfs_main_kobj(struct edac_device_ctl_info *dev) { debugf0("%s()\n", __func__); debugf4("%s() name of kobject is: %s\n", - __func__, kobject_name(&edac_dev->kobj)); + __func__, kobject_name(&dev->kobj)); /* * Unregister the edac device's kobject and @@ -304,7 +307,8 @@ void edac_device_unregister_sysfs_main_kobj( * a) module_put() this module * b) 'kfree' the memory */ - kobject_put(&edac_dev->kobj); + kobject_put(&dev->kobj); + edac_put_sysfs_class(); } /* edac_dev -> instance information */ diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c index aa93ad8..a413586 100644 --- a/drivers/edac/edac_mc_sysfs.c +++ b/drivers/edac/edac_mc_sysfs.c @@ -11,6 +11,7 @@ #include #include +#include #include #include "edac_core.h" @@ -1017,7 +1018,7 @@ int edac_sysfs_setup_mc_kset(void) debugf1("%s()\n", __func__); /* get the /sys/devices/system/edac class reference */ - edac_class = edac_get_edac_class(); + edac_class = edac_get_sysfs_class(); if (edac_class == NULL) { debugf1("%s() no edac_class error=%d\n", __func__, err); goto fail_out; @@ -1028,15 +1029,16 @@ int edac_sysfs_setup_mc_kset(void) if (!mc_kset) { err = -ENOMEM; debugf1("%s() Failed to register '.../edac/mc'\n", __func__); - goto fail_out; + goto fail_kset; } debugf1("%s() Registered '.../edac/mc' kobject\n", __func__); return 0; +fail_kset: + edac_put_sysfs_class(); - /* error unwind stack */ fail_out: return err; } @@ -1049,5 +1051,6 @@ fail_out: void edac_sysfs_teardown_mc_kset(void) { kset_unregister(mc_kset); + edac_put_sysfs_class(); } diff --git a/drivers/edac/edac_module.c b/drivers/edac/edac_module.c index 7e1374a..be4b075 100644 --- a/drivers/edac/edac_module.c +++ b/drivers/edac/edac_module.c @@ -27,15 +27,6 @@ EXPORT_SYMBOL_GPL(edac_debug_level); struct workqueue_struct *edac_workqueue; /* - * sysfs object: /sys/devices/system/edac - * need to export to other files in this modules - */ -static struct sysdev_class edac_class = { - .name = "edac", -}; -static int edac_class_valid; - -/* * edac_op_state_to_string() */ char *edac_op_state_to_string(int opstate) @@ -55,60 +46,6 @@ char *edac_op_state_to_string(int opstate) } /* - * edac_get_edac_class() - * - * return pointer to the edac class of 'edac' - */ -struct sysdev_class *edac_get_edac_class(void) -{ - struct sysdev_class *classptr = NULL; - - if (edac_class_valid) - classptr = &edac_class; - - return classptr; -} - -/* - * edac_register_sysfs_edac_name() - * - * register the 'edac' into /sys/devices/system - * - * return: - * 0 success - * !0 error - */ -static int edac_register_sysfs_edac_name(void) -{ - int err; - - /* create the /sys/devices/system/edac directory */ - err = sysdev_class_register(&edac_class); - - if (err) { - debugf1("%s() error=%d\n", __func__, err); - return err; - } - - edac_class_valid = 1; - return 0; -} - -/* - * sysdev_class_unregister() - * - * unregister the 'edac' from /sys/devices/system - */ -static void edac_unregister_sysfs_edac_name(void) -{ - /* only if currently registered, then unregister it */ - if (edac_class_valid) - sysdev_class_unregister(&edac_class); - - edac_class_valid = 0; -} - -/* * edac_workqueue_setup * initialize the edac work queue for polling operations */ @@ -154,21 +91,11 @@ static int __init edac_init(void) edac_pci_clear_parity_errors(); /* - * perform the registration of the /sys/devices/system/edac class object - */ - if (edac_register_sysfs_edac_name()) { - edac_printk(KERN_ERR, EDAC_MC, - "Error initializing 'edac' kobject\n"); - err = -ENODEV; - goto error; - } - - /* * now set up the mc_kset under the edac class object */ err = edac_sysfs_setup_mc_kset(); if (err) - goto sysfs_setup_fail; + goto error; /* Setup/Initialize the workq for this core */ err = edac_workqueue_setup(); @@ -183,9 +110,6 @@ static int __init edac_init(void) workq_fail: edac_sysfs_teardown_mc_kset(); -sysfs_setup_fail: - edac_unregister_sysfs_edac_name(); - error: return err; } @@ -201,7 +125,6 @@ static void __exit edac_exit(void) /* tear down the various subsystems */ edac_workqueue_teardown(); edac_sysfs_teardown_mc_kset(); - edac_unregister_sysfs_edac_name(); } /* diff --git a/drivers/edac/edac_module.h b/drivers/edac/edac_module.h index 233d479..17aabb7 100644 --- a/drivers/edac/edac_module.h +++ b/drivers/edac/edac_module.h @@ -42,7 +42,6 @@ extern void edac_device_unregister_sysfs_main_kobj( struct edac_device_ctl_info *edac_dev); extern int edac_device_create_sysfs(struct edac_device_ctl_info *edac_dev); extern void edac_device_remove_sysfs(struct edac_device_ctl_info *edac_dev); -extern struct sysdev_class *edac_get_edac_class(void); /* edac core workqueue: single CPU mode */ extern struct workqueue_struct *edac_workqueue; diff --git a/drivers/edac/edac_pci_sysfs.c b/drivers/edac/edac_pci_sysfs.c index c39697d..023b01c 100644 --- a/drivers/edac/edac_pci_sysfs.c +++ b/drivers/edac/edac_pci_sysfs.c @@ -7,7 +7,7 @@ * */ #include -#include +#include #include #include @@ -354,7 +354,7 @@ static int edac_pci_main_kobj_setup(void) /* First time, so create the main kobject and its * controls and atributes */ - edac_class = edac_get_edac_class(); + edac_class = edac_get_sysfs_class(); if (edac_class == NULL) { debugf1("%s() no edac_class\n", __func__); err = -ENODEV; @@ -368,7 +368,7 @@ static int edac_pci_main_kobj_setup(void) if (!try_module_get(THIS_MODULE)) { debugf1("%s() try_module_get() failed\n", __func__); err = -ENODEV; - goto decrement_count_fail; + goto mod_get_fail; } edac_pci_top_main_kobj = kzalloc(sizeof(struct kobject), GFP_KERNEL); @@ -403,6 +403,9 @@ kobject_init_and_add_fail: kzalloc_fail: module_put(THIS_MODULE); +mod_get_fail: + edac_put_sysfs_class(); + decrement_count_fail: /* if are on this error exit, nothing to tear down */ atomic_dec(&edac_pci_sysfs_refcount); @@ -429,6 +432,7 @@ static void edac_pci_main_kobj_teardown(void) __func__); kobject_put(edac_pci_top_main_kobj); } + edac_put_sysfs_class(); } /* diff --git a/drivers/edac/edac_stub.c b/drivers/edac/edac_stub.c index 20b428a..aab9707 100644 --- a/drivers/edac/edac_stub.c +++ b/drivers/edac/edac_stub.c @@ -3,10 +3,13 @@ * * Author: Dave Jiang * - * 2007 (c) MontaVista Software, Inc. This file is licensed under - * the terms of the GNU General Public License version 2. This program - * is licensed "as is" without any warranty of any kind, whether express - * or implied. + * 2007 (c) MontaVista Software, Inc. + * 2010 (c) Advanced Micro Devices Inc. + * Borislav Petkov + * + * This file is licensed under the terms of the GNU General Public + * License version 2. This program is licensed "as is" without any + * warranty of any kind, whether express or implied. * */ #include @@ -23,6 +26,8 @@ EXPORT_SYMBOL_GPL(edac_handlers); int edac_err_assert = 0; EXPORT_SYMBOL_GPL(edac_err_assert); +static atomic_t edac_class_valid = ATOMIC_INIT(0); + /* * called to determine if there is an EDAC driver interested in * knowing an event (such as NMI) occurred @@ -44,3 +49,41 @@ void edac_atomic_assert_error(void) edac_err_assert++; } EXPORT_SYMBOL_GPL(edac_atomic_assert_error); + +/* + * sysfs object: /sys/devices/system/edac + * need to export to other files + */ +struct sysdev_class edac_class = { + .name = "edac", +}; +EXPORT_SYMBOL_GPL(edac_class); + +/* return pointer to the 'edac' node in sysfs */ +struct sysdev_class *edac_get_sysfs_class(void) +{ + int err = 0; + + if (atomic_read(&edac_class_valid)) + goto out; + + /* create the /sys/devices/system/edac directory */ + err = sysdev_class_register(&edac_class); + if (err) { + printk(KERN_ERR "Error registering toplevel EDAC sysfs dir\n"); + return NULL; + } + +out: + atomic_inc(&edac_class_valid); + return &edac_class; +} +EXPORT_SYMBOL_GPL(edac_get_sysfs_class); + +void edac_put_sysfs_class(void) +{ + /* last user unregisters it */ + if (atomic_dec_and_test(&edac_class_valid)) + sysdev_class_unregister(&edac_class); +} +EXPORT_SYMBOL_GPL(edac_put_sysfs_class); diff --git a/include/linux/edac.h b/include/linux/edac.h index 7cf92e8..36c6644 100644 --- a/include/linux/edac.h +++ b/include/linux/edac.h @@ -13,6 +13,7 @@ #define _LINUX_EDAC_H_ #include +#include #define EDAC_OPSTATE_INVAL -1 #define EDAC_OPSTATE_POLL 0 @@ -22,9 +23,12 @@ extern int edac_op_state; extern int edac_err_assert; extern atomic_t edac_handlers; +extern struct sysdev_class edac_class; extern int edac_handler_set(void); extern void edac_atomic_assert_error(void); +extern struct sysdev_class *edac_get_sysfs_class(void); +extern void edac_put_sysfs_class(void); static inline void opstate_init(void) { -- cgit v0.10.2 From 9cdeb404a1870c5022915e576dbdc3cde21af5bf Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 2 Sep 2010 18:33:24 +0200 Subject: EDAC, MCE: Rework MCE injection Add sysfs injection facilities for testing of the MCE decoding code. Remove large parts of amd64_edac_dbg.c, as a result, which did only NB MCE injection anyway and the new injection code supports that functionality already. Add an injection module so that MCE decoding code in production kernels like those in RHEL and SLES can be tested. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index 70bb350..3bb3a67 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -39,7 +39,7 @@ config EDAC_DEBUG there're four debug levels (x=0,1,2,3 from low to high). Usually you should select 'N'. - config EDAC_DECODE_MCE +config EDAC_DECODE_MCE tristate "Decode MCEs in human-readable form (only on AMD for now)" depends on CPU_SUP_AMD && X86_MCE default y @@ -51,6 +51,16 @@ config EDAC_DEBUG which occur really early upon boot, before the module infrastructure has been initialized. +config EDAC_MCE_INJ + tristate "Simple MCE injection interface over /sysfs" + depends on EDAC_DECODE_MCE + default n + help + This is a simple interface to inject MCEs over /sysfs and test + the MCE decoding code in EDAC. + + This is currently AMD-only. + config EDAC_MM_EDAC tristate "Main Memory EDAC (Error Detection And Correction) reporting" help @@ -72,7 +82,7 @@ config EDAC_AMD64 Families of Memory Controllers (K8, F10h and F11h) config EDAC_AMD64_ERROR_INJECTION - bool "Sysfs Error Injection facilities" + bool "Sysfs HW Error injection facilities" depends on EDAC_AMD64 help Recent Opterons (Family 10h and later) provide for Memory Error diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile index ca6b1bb..5c38ad3 100644 --- a/drivers/edac/Makefile +++ b/drivers/edac/Makefile @@ -17,6 +17,8 @@ ifdef CONFIG_PCI edac_core-objs += edac_pci.o edac_pci_sysfs.o endif +obj-$(CONFIG_EDAC_MCE_INJ) += mce_amd_inj.o + obj-$(CONFIG_EDAC_DECODE_MCE) += edac_mce_amd.o obj-$(CONFIG_EDAC_AMD76X) += amd76x_edac.o diff --git a/drivers/edac/amd64_edac.h b/drivers/edac/amd64_edac.h index 613b938..67d9ceb 100644 --- a/drivers/edac/amd64_edac.h +++ b/drivers/edac/amd64_edac.h @@ -486,7 +486,7 @@ extern const char *ext_msgs[32]; extern const char *htlink_msgs[8]; #ifdef CONFIG_EDAC_DEBUG -#define NUM_DBG_ATTRS 9 +#define NUM_DBG_ATTRS 5 #else #define NUM_DBG_ATTRS 0 #endif diff --git a/drivers/edac/amd64_edac_dbg.c b/drivers/edac/amd64_edac_dbg.c index f6d5695..e356228 100644 --- a/drivers/edac/amd64_edac_dbg.c +++ b/drivers/edac/amd64_edac_dbg.c @@ -1,173 +1,16 @@ #include "amd64_edac.h" -/* - * accept a hex value and store it into the virtual error register file, field: - * nbeal and nbeah. Assume virtual error values have already been set for: NBSL, - * NBSH and NBCFG. Then proceed to map the error values to a MC, CSROW and - * CHANNEL - */ -static ssize_t amd64_nbea_store(struct mem_ctl_info *mci, const char *data, - size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - u64 value; - int ret = 0; - struct mce m; - - ret = strict_strtoull(data, 16, &value); - if (ret != -EINVAL) { - struct err_regs *regs = &pvt->ctl_error_info; - - debugf0("received NBEA= 0x%llx\n", value); - - /* place the value into the virtual error packet */ - pvt->ctl_error_info.nbeal = (u32) value; - value >>= 32; - pvt->ctl_error_info.nbeah = (u32) value; - - m.addr = value; - m.status = regs->nbsl | ((u64)regs->nbsh << 32); - - /* Process the Mapping request */ - /* TODO: Add race prevention */ - amd_decode_nb_mce(pvt->mc_node_id, &m, regs->nbcfg); - - return count; - } - return ret; +#define EDAC_DCT_ATTR_SHOW(reg) \ +static ssize_t amd64_##reg##_show(struct mem_ctl_info *mci, char *data) \ +{ \ + struct amd64_pvt *pvt = mci->pvt_info; \ + return sprintf(data, "0x%016llx\n", (u64)pvt->reg); \ } -/* display back what the last NBEA (MCA NB Address (MC4_ADDR)) was written */ -static ssize_t amd64_nbea_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - u64 value; - - value = pvt->ctl_error_info.nbeah; - value <<= 32; - value |= pvt->ctl_error_info.nbeal; - - return sprintf(data, "%llx\n", value); -} - -/* store the NBSL (MCA NB Status Low (MC4_STATUS)) value user desires */ -static ssize_t amd64_nbsl_store(struct mem_ctl_info *mci, const char *data, - size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - unsigned long value; - int ret = 0; - - ret = strict_strtoul(data, 16, &value); - if (ret != -EINVAL) { - debugf0("received NBSL= 0x%lx\n", value); - - pvt->ctl_error_info.nbsl = (u32) value; - - return count; - } - return ret; -} - -/* display back what the last NBSL value written */ -static ssize_t amd64_nbsl_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - u32 value; - - value = pvt->ctl_error_info.nbsl; - - return sprintf(data, "%x\n", value); -} - -/* store the NBSH (MCA NB Status High) value user desires */ -static ssize_t amd64_nbsh_store(struct mem_ctl_info *mci, const char *data, - size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - unsigned long value; - int ret = 0; - - ret = strict_strtoul(data, 16, &value); - if (ret != -EINVAL) { - debugf0("received NBSH= 0x%lx\n", value); - - pvt->ctl_error_info.nbsh = (u32) value; - - return count; - } - return ret; -} - -/* display back what the last NBSH value written */ -static ssize_t amd64_nbsh_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - u32 value; - - value = pvt->ctl_error_info.nbsh; - - return sprintf(data, "%x\n", value); -} - -/* accept and store the NBCFG (MCA NB Configuration) value user desires */ -static ssize_t amd64_nbcfg_store(struct mem_ctl_info *mci, - const char *data, size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - unsigned long value; - int ret = 0; - - ret = strict_strtoul(data, 16, &value); - if (ret != -EINVAL) { - debugf0("received NBCFG= 0x%lx\n", value); - - pvt->ctl_error_info.nbcfg = (u32) value; - - return count; - } - return ret; -} - -/* various show routines for the controls of a MCI */ -static ssize_t amd64_nbcfg_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%x\n", pvt->ctl_error_info.nbcfg); -} - - -static ssize_t amd64_dhar_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%x\n", pvt->dhar); -} - - -static ssize_t amd64_dbam_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%x\n", pvt->dbam0); -} - - -static ssize_t amd64_topmem_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%llx\n", pvt->top_mem); -} - - -static ssize_t amd64_topmem2_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%llx\n", pvt->top_mem2); -} +EDAC_DCT_ATTR_SHOW(dhar); +EDAC_DCT_ATTR_SHOW(dbam0); +EDAC_DCT_ATTR_SHOW(top_mem); +EDAC_DCT_ATTR_SHOW(top_mem2); static ssize_t amd64_hole_show(struct mem_ctl_info *mci, char *data) { @@ -188,38 +31,6 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { { .attr = { - .name = "nbea_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbea_show, - .store = amd64_nbea_store, - }, - { - .attr = { - .name = "nbsl_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbsl_show, - .store = amd64_nbsl_store, - }, - { - .attr = { - .name = "nbsh_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbsh_show, - .store = amd64_nbsh_store, - }, - { - .attr = { - .name = "nbcfg_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbcfg_show, - .store = amd64_nbcfg_store, - }, - { - .attr = { .name = "dhar", .mode = (S_IRUGO) }, @@ -231,7 +42,7 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { .name = "dbam", .mode = (S_IRUGO) }, - .show = amd64_dbam_show, + .show = amd64_dbam0_show, .store = NULL, }, { @@ -239,7 +50,7 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { .name = "topmem", .mode = (S_IRUGO) }, - .show = amd64_topmem_show, + .show = amd64_top_mem_show, .store = NULL, }, { @@ -247,7 +58,7 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { .name = "topmem2", .mode = (S_IRUGO) }, - .show = amd64_topmem2_show, + .show = amd64_top_mem2_show, .store = NULL, }, { diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c index 6cfa881..c75c47b 100644 --- a/drivers/edac/edac_mce_amd.c +++ b/drivers/edac/edac_mce_amd.c @@ -324,8 +324,7 @@ static inline void amd_decode_err_code(u16 ec) pr_emerg(HW_ERR "Huh? Unknown MCE error 0x%x\n", ec); } -static int amd_decode_mce(struct notifier_block *nb, unsigned long val, - void *data) +int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data) { struct mce *m = (struct mce *)data; int node, ecc; @@ -379,6 +378,7 @@ static int amd_decode_mce(struct notifier_block *nb, unsigned long val, return NOTIFY_STOP; } +EXPORT_SYMBOL_GPL(amd_decode_mce); static struct notifier_block amd_mce_dec_nb = { .notifier_call = amd_decode_mce, diff --git a/drivers/edac/edac_mce_amd.h b/drivers/edac/edac_mce_amd.h index 0fba0e76..2712a90 100644 --- a/drivers/edac/edac_mce_amd.h +++ b/drivers/edac/edac_mce_amd.h @@ -1,6 +1,8 @@ #ifndef _EDAC_MCE_AMD_H #define _EDAC_MCE_AMD_H +#include + #include #define ERROR_CODE(x) ((x) & 0xffff) @@ -61,10 +63,10 @@ struct err_regs { u32 nbeal; }; - void amd_report_gart_errors(bool); void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)); void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)); void amd_decode_nb_mce(int, struct mce *, u32); +int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data); #endif /* _EDAC_MCE_AMD_H */ diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c new file mode 100644 index 0000000..0e4f2dc --- /dev/null +++ b/drivers/edac/mce_amd_inj.c @@ -0,0 +1,171 @@ +/* + * A simple MCE injection facility for testing the MCE decoding code. This + * driver should be built as module so that it can be loaded on production + * kernels for testing purposes. + * + * This file may be distributed under the terms of the GNU General Public + * License version 2. + * + * Copyright (c) 2010: Borislav Petkov + * Advanced Micro Devices Inc. + */ + +#include +#include +#include +#include + +#include "edac_mce_amd.h" + +struct edac_mce_attr { + struct attribute attr; + ssize_t (*show) (struct kobject *kobj, struct edac_mce_attr *attr, char *buf); + ssize_t (*store)(struct kobject *kobj, struct edac_mce_attr *attr, + const char *buf, size_t count); +}; + +#define EDAC_MCE_ATTR(_name, _mode, _show, _store) \ +static struct edac_mce_attr mce_attr_##_name = __ATTR(_name, _mode, _show, _store) + +static struct kobject *mce_kobj; + +/* + * Collect all the MCi_XXX settings + */ +static struct mce i_mce; + +#define MCE_INJECT_STORE(reg) \ +static ssize_t edac_inject_##reg##_store(struct kobject *kobj, \ + struct edac_mce_attr *attr, \ + const char *data, size_t count)\ +{ \ + int ret = 0; \ + unsigned long value; \ + \ + ret = strict_strtoul(data, 16, &value); \ + if (ret < 0) \ + printk(KERN_ERR "Error writing MCE " #reg " field.\n"); \ + \ + i_mce.reg = value; \ + \ + return count; \ +} + +MCE_INJECT_STORE(status); +MCE_INJECT_STORE(misc); +MCE_INJECT_STORE(addr); + +#define MCE_INJECT_SHOW(reg) \ +static ssize_t edac_inject_##reg##_show(struct kobject *kobj, \ + struct edac_mce_attr *attr, \ + char *buf) \ +{ \ + return sprintf(buf, "0x%016llx\n", i_mce.reg); \ +} + +MCE_INJECT_SHOW(status); +MCE_INJECT_SHOW(misc); +MCE_INJECT_SHOW(addr); + +EDAC_MCE_ATTR(status, 0644, edac_inject_status_show, edac_inject_status_store); +EDAC_MCE_ATTR(misc, 0644, edac_inject_misc_show, edac_inject_misc_store); +EDAC_MCE_ATTR(addr, 0644, edac_inject_addr_show, edac_inject_addr_store); + +/* + * This denotes into which bank we're injecting and triggers + * the injection, at the same time. + */ +static ssize_t edac_inject_bank_store(struct kobject *kobj, + struct edac_mce_attr *attr, + const char *data, size_t count) +{ + int ret = 0; + unsigned long value; + + ret = strict_strtoul(data, 10, &value); + if (ret < 0) { + printk(KERN_ERR "Invalid bank value!\n"); + return -EINVAL; + } + + if (value > 5) { + printk(KERN_ERR "Non-existant MCE bank: %lu\n", value); + return -EINVAL; + } + + i_mce.bank = value; + + amd_decode_mce(NULL, 0, &i_mce); + + return count; +} + +static ssize_t edac_inject_bank_show(struct kobject *kobj, + struct edac_mce_attr *attr, char *buf) +{ + return sprintf(buf, "%d\n", i_mce.bank); +} + +EDAC_MCE_ATTR(bank, 0644, edac_inject_bank_show, edac_inject_bank_store); + +static struct edac_mce_attr *sysfs_attrs[] = { &mce_attr_status, &mce_attr_misc, + &mce_attr_addr, &mce_attr_bank +}; + +static int __init edac_init_mce_inject(void) +{ + struct sysdev_class *edac_class = NULL; + int i, err = 0; + + edac_class = edac_get_sysfs_class(); + if (!edac_class) + return -EINVAL; + + mce_kobj = kobject_create_and_add("mce", &edac_class->kset.kobj); + if (!mce_kobj) { + printk(KERN_ERR "Error creating a mce kset.\n"); + err = -ENOMEM; + goto err_mce_kobj; + } + + for (i = 0; i < ARRAY_SIZE(sysfs_attrs); i++) { + err = sysfs_create_file(mce_kobj, &sysfs_attrs[i]->attr); + if (err) { + printk(KERN_ERR "Error creating %s in sysfs.\n", + sysfs_attrs[i]->attr.name); + goto err_sysfs_create; + } + } + return 0; + +err_sysfs_create: + while (i-- >= 0) + sysfs_remove_file(mce_kobj, &sysfs_attrs[i]->attr); + + kobject_del(mce_kobj); + +err_mce_kobj: + edac_put_sysfs_class(); + + return err; +} + +static void __exit edac_exit_mce_inject(void) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(sysfs_attrs); i++) + sysfs_remove_file(mce_kobj, &sysfs_attrs[i]->attr); + + kobject_del(mce_kobj); + + edac_put_sysfs_class(); +} + +module_init(edac_init_mce_inject); +module_exit(edac_exit_mce_inject); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Borislav Petkov "); +MODULE_AUTHOR("AMD Inc."); +MODULE_DESCRIPTION("MCE injection facility for testing MCE decoding"); -- cgit v0.10.2 From 47ca08a40b043815134d489e21870b53276f1a4a Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Mon, 27 Sep 2010 15:30:39 +0200 Subject: EDAC, MCE: Rename files Drop "edac_" string from the filenames since they're prefixed with edac/ in their pathname anyway. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile index 5c38ad3..32c7bc9 100644 --- a/drivers/edac/Makefile +++ b/drivers/edac/Makefile @@ -19,6 +19,7 @@ endif obj-$(CONFIG_EDAC_MCE_INJ) += mce_amd_inj.o +edac_mce_amd-objs := mce_amd.o obj-$(CONFIG_EDAC_DECODE_MCE) += edac_mce_amd.o obj-$(CONFIG_EDAC_AMD76X) += amd76x_edac.o diff --git a/drivers/edac/amd64_edac.h b/drivers/edac/amd64_edac.h index 67d9ceb..13e1d6f 100644 --- a/drivers/edac/amd64_edac.h +++ b/drivers/edac/amd64_edac.h @@ -72,7 +72,7 @@ #include #include #include "edac_core.h" -#include "edac_mce_amd.h" +#include "mce_amd.h" #define amd64_printk(level, fmt, arg...) \ edac_printk(level, "amd64", fmt, ##arg) diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c deleted file mode 100644 index c75c47b..0000000 --- a/drivers/edac/edac_mce_amd.c +++ /dev/null @@ -1,414 +0,0 @@ -#include -#include "edac_mce_amd.h" - -static bool report_gart_errors; -static void (*nb_bus_decoder)(int node_id, struct mce *m, u32 nbcfg); - -void amd_report_gart_errors(bool v) -{ - report_gart_errors = v; -} -EXPORT_SYMBOL_GPL(amd_report_gart_errors); - -void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)) -{ - nb_bus_decoder = f; -} -EXPORT_SYMBOL_GPL(amd_register_ecc_decoder); - -void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)) -{ - if (nb_bus_decoder) { - WARN_ON(nb_bus_decoder != f); - - nb_bus_decoder = NULL; - } -} -EXPORT_SYMBOL_GPL(amd_unregister_ecc_decoder); - -/* - * string representation for the different MCA reported error types, see F3x48 - * or MSR0000_0411. - */ - -/* transaction type */ -const char *tt_msgs[] = { "INSN", "DATA", "GEN", "RESV" }; -EXPORT_SYMBOL_GPL(tt_msgs); - -/* cache level */ -const char *ll_msgs[] = { "RESV", "L1", "L2", "L3/GEN" }; -EXPORT_SYMBOL_GPL(ll_msgs); - -/* memory transaction type */ -const char *rrrr_msgs[] = { - "GEN", "RD", "WR", "DRD", "DWR", "IRD", "PRF", "EV", "SNP" -}; -EXPORT_SYMBOL_GPL(rrrr_msgs); - -/* participating processor */ -const char *pp_msgs[] = { "SRC", "RES", "OBS", "GEN" }; -EXPORT_SYMBOL_GPL(pp_msgs); - -/* request timeout */ -const char *to_msgs[] = { "no timeout", "timed out" }; -EXPORT_SYMBOL_GPL(to_msgs); - -/* memory or i/o */ -const char *ii_msgs[] = { "MEM", "RESV", "IO", "GEN" }; -EXPORT_SYMBOL_GPL(ii_msgs); - -/* - * Map the 4 or 5 (family-specific) bits of Extended Error code to the - * string table. - */ -const char *ext_msgs[] = { - "K8 ECC error", /* 0_0000b */ - "CRC error on link", /* 0_0001b */ - "Sync error packets on link", /* 0_0010b */ - "Master Abort during link operation", /* 0_0011b */ - "Target Abort during link operation", /* 0_0100b */ - "Invalid GART PTE entry during table walk", /* 0_0101b */ - "Unsupported atomic RMW command received", /* 0_0110b */ - "WDT error: NB transaction timeout", /* 0_0111b */ - "ECC/ChipKill ECC error", /* 0_1000b */ - "SVM DEV Error", /* 0_1001b */ - "Link Data error", /* 0_1010b */ - "Link/L3/Probe Filter Protocol error", /* 0_1011b */ - "NB Internal Arrays Parity error", /* 0_1100b */ - "DRAM Address/Control Parity error", /* 0_1101b */ - "Link Transmission error", /* 0_1110b */ - "GART/DEV Table Walk Data error" /* 0_1111b */ - "Res 0x100 error", /* 1_0000b */ - "Res 0x101 error", /* 1_0001b */ - "Res 0x102 error", /* 1_0010b */ - "Res 0x103 error", /* 1_0011b */ - "Res 0x104 error", /* 1_0100b */ - "Res 0x105 error", /* 1_0101b */ - "Res 0x106 error", /* 1_0110b */ - "Res 0x107 error", /* 1_0111b */ - "Res 0x108 error", /* 1_1000b */ - "Res 0x109 error", /* 1_1001b */ - "Res 0x10A error", /* 1_1010b */ - "Res 0x10B error", /* 1_1011b */ - "ECC error in L3 Cache Data", /* 1_1100b */ - "L3 Cache Tag error", /* 1_1101b */ - "L3 Cache LRU Parity error", /* 1_1110b */ - "Probe Filter error" /* 1_1111b */ -}; -EXPORT_SYMBOL_GPL(ext_msgs); - -static void amd_decode_dc_mce(struct mce *m) -{ - u32 ec = m->status & 0xffff; - u32 xec = (m->status >> 16) & 0xf; - - pr_emerg(HW_ERR "Data Cache Error: "); - - if (xec == 1 && TLB_ERROR(ec)) - pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); - else if (xec == 0) { - if (m->status & (1ULL << 40)) - pr_cont(" during Data Scrub.\n"); - else if (TLB_ERROR(ec)) - pr_cont(": %s TLB parity error.\n", LL_MSG(ec)); - else if (MEM_ERROR(ec)) { - u8 ll = ec & 0x3; - u8 tt = (ec >> 2) & 0x3; - u8 rrrr = (ec >> 4) & 0xf; - - /* see F10h BKDG (31116), Table 92. */ - if (ll == 0x1) { - if (tt != 0x1) - goto wrong_dc_mce; - - pr_cont(": Data/Tag %s error.\n", RRRR_MSG(ec)); - - } else if (ll == 0x2 && rrrr == 0x3) - pr_cont(" during L1 linefill from L2.\n"); - else - goto wrong_dc_mce; - } else if (BUS_ERROR(ec) && boot_cpu_data.x86 == 0xf) - pr_cont(" during system linefill.\n"); - else - goto wrong_dc_mce; - } else - goto wrong_dc_mce; - - return; - -wrong_dc_mce: - pr_emerg(HW_ERR "Corrupted DC MCE info?\n"); -} - -static void amd_decode_ic_mce(struct mce *m) -{ - u32 ec = m->status & 0xffff; - u32 xec = (m->status >> 16) & 0xf; - - pr_emerg(HW_ERR "Instruction Cache Error"); - - if (xec == 1 && TLB_ERROR(ec)) - pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); - else if (xec == 0) { - if (TLB_ERROR(ec)) - pr_cont(": %s TLB Parity error.\n", LL_MSG(ec)); - else if (BUS_ERROR(ec)) { - if (boot_cpu_data.x86 == 0xf && - (m->status & BIT(58))) - pr_cont(" during system linefill.\n"); - else - pr_cont(" during attempted NB data read.\n"); - } else if (MEM_ERROR(ec)) { - u8 ll = ec & 0x3; - u8 rrrr = (ec >> 4) & 0xf; - - if (ll == 0x2) - pr_cont(" during a linefill from L2.\n"); - else if (ll == 0x1) { - - switch (rrrr) { - case 0x5: - pr_cont(": Parity error during " - "data load.\n"); - break; - - case 0x7: - pr_cont(": Copyback Parity/Victim" - " error.\n"); - break; - - case 0x8: - pr_cont(": Tag Snoop error.\n"); - break; - - default: - goto wrong_ic_mce; - break; - } - } - } else - goto wrong_ic_mce; - } else - goto wrong_ic_mce; - - return; - -wrong_ic_mce: - pr_emerg(HW_ERR "Corrupted IC MCE info?\n"); -} - -static void amd_decode_bu_mce(struct mce *m) -{ - u32 ec = m->status & 0xffff; - u32 xec = (m->status >> 16) & 0xf; - - pr_emerg(HW_ERR "Bus Unit Error"); - - if (xec == 0x1) - pr_cont(" in the write data buffers.\n"); - else if (xec == 0x3) - pr_cont(" in the victim data buffers.\n"); - else if (xec == 0x2 && MEM_ERROR(ec)) - pr_cont(": %s error in the L2 cache tags.\n", RRRR_MSG(ec)); - else if (xec == 0x0) { - if (TLB_ERROR(ec)) - pr_cont(": %s error in a Page Descriptor Cache or " - "Guest TLB.\n", TT_MSG(ec)); - else if (BUS_ERROR(ec)) - pr_cont(": %s/ECC error in data read from NB: %s.\n", - RRRR_MSG(ec), PP_MSG(ec)); - else if (MEM_ERROR(ec)) { - u8 rrrr = (ec >> 4) & 0xf; - - if (rrrr >= 0x7) - pr_cont(": %s error during data copyback.\n", - RRRR_MSG(ec)); - else if (rrrr <= 0x1) - pr_cont(": %s parity/ECC error during data " - "access from L2.\n", RRRR_MSG(ec)); - else - goto wrong_bu_mce; - } else - goto wrong_bu_mce; - } else - goto wrong_bu_mce; - - return; - -wrong_bu_mce: - pr_emerg(HW_ERR "Corrupted BU MCE info?\n"); -} - -static void amd_decode_ls_mce(struct mce *m) -{ - u32 ec = m->status & 0xffff; - u32 xec = (m->status >> 16) & 0xf; - - pr_emerg(HW_ERR "Load Store Error"); - - if (xec == 0x0) { - u8 rrrr = (ec >> 4) & 0xf; - - if (!BUS_ERROR(ec) || (rrrr != 0x3 && rrrr != 0x4)) - goto wrong_ls_mce; - - pr_cont(" during %s.\n", RRRR_MSG(ec)); - } - return; - -wrong_ls_mce: - pr_emerg(HW_ERR "Corrupted LS MCE info?\n"); -} - -void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg) -{ - u32 ec = m->status & 0xffff; - u32 nbsh = (u32)(m->status >> 32); - u32 nbsl = (u32)m->status; - - /* - * GART TLB error reporting is disabled by default. Bail out early. - */ - if (TLB_ERROR(ec) && !report_gart_errors) - return; - - pr_emerg(HW_ERR "Northbridge Error, node %d", node_id); - - /* - * F10h, revD can disable ErrCpu[3:0] so check that first and also the - * value encoding has changed so interpret those differently - */ - if ((boot_cpu_data.x86 == 0x10) && - (boot_cpu_data.x86_model > 7)) { - if (nbsh & K8_NBSH_ERR_CPU_VAL) - pr_cont(", core: %u\n", (u8)(nbsh & 0xf)); - } else { - u8 assoc_cpus = nbsh & 0xf; - - if (assoc_cpus > 0) - pr_cont(", core: %d", fls(assoc_cpus) - 1); - - pr_cont("\n"); - } - - pr_emerg(HW_ERR "%s.\n", EXT_ERR_MSG(nbsl)); - - if (BUS_ERROR(ec) && nb_bus_decoder) - nb_bus_decoder(node_id, m, nbcfg); -} -EXPORT_SYMBOL_GPL(amd_decode_nb_mce); - -static void amd_decode_fr_mce(struct mce *m) -{ - /* we have only one error signature so match all fields at once. */ - if ((m->status & 0xffff) == 0x0f0f) - pr_emerg(HW_ERR " FR Error: CPU Watchdog timer expire.\n"); - else - pr_emerg(HW_ERR "Corrupted FR MCE info?\n"); -} - -static inline void amd_decode_err_code(u16 ec) -{ - if (TLB_ERROR(ec)) { - pr_emerg(HW_ERR "Transaction: %s, Cache Level: %s\n", - TT_MSG(ec), LL_MSG(ec)); - } else if (MEM_ERROR(ec)) { - pr_emerg(HW_ERR "Transaction: %s, Type: %s, Cache Level: %s\n", - RRRR_MSG(ec), TT_MSG(ec), LL_MSG(ec)); - } else if (BUS_ERROR(ec)) { - pr_emerg(HW_ERR "Transaction: %s (%s), %s, Cache Level: %s, " - "Participating Processor: %s\n", - RRRR_MSG(ec), II_MSG(ec), TO_MSG(ec), LL_MSG(ec), - PP_MSG(ec)); - } else - pr_emerg(HW_ERR "Huh? Unknown MCE error 0x%x\n", ec); -} - -int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data) -{ - struct mce *m = (struct mce *)data; - int node, ecc; - - pr_emerg(HW_ERR "MC%d_STATUS: ", m->bank); - - pr_cont("%sorrected error, other errors lost: %s, " - "CPU context corrupt: %s", - ((m->status & MCI_STATUS_UC) ? "Unc" : "C"), - ((m->status & MCI_STATUS_OVER) ? "yes" : "no"), - ((m->status & MCI_STATUS_PCC) ? "yes" : "no")); - - /* do the two bits[14:13] together */ - ecc = (m->status >> 45) & 0x3; - if (ecc) - pr_cont(", %sECC Error", ((ecc == 2) ? "C" : "U")); - - pr_cont("\n"); - - switch (m->bank) { - case 0: - amd_decode_dc_mce(m); - break; - - case 1: - amd_decode_ic_mce(m); - break; - - case 2: - amd_decode_bu_mce(m); - break; - - case 3: - amd_decode_ls_mce(m); - break; - - case 4: - node = amd_get_nb_id(m->extcpu); - amd_decode_nb_mce(node, m, 0); - break; - - case 5: - amd_decode_fr_mce(m); - break; - - default: - break; - } - - amd_decode_err_code(m->status & 0xffff); - - return NOTIFY_STOP; -} -EXPORT_SYMBOL_GPL(amd_decode_mce); - -static struct notifier_block amd_mce_dec_nb = { - .notifier_call = amd_decode_mce, -}; - -static int __init mce_amd_init(void) -{ - /* - * We can decode MCEs for K8, F10h and F11h CPUs: - */ - if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) - return 0; - - if (boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) - return 0; - - atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb); - - return 0; -} -early_initcall(mce_amd_init); - -#ifdef MODULE -static void __exit mce_amd_exit(void) -{ - atomic_notifier_chain_unregister(&x86_mce_decoder_chain, &amd_mce_dec_nb); -} - -MODULE_DESCRIPTION("AMD MCE decoder"); -MODULE_ALIAS("edac-mce-amd"); -MODULE_LICENSE("GPL"); -module_exit(mce_amd_exit); -#endif diff --git a/drivers/edac/edac_mce_amd.h b/drivers/edac/edac_mce_amd.h deleted file mode 100644 index 2712a90..0000000 --- a/drivers/edac/edac_mce_amd.h +++ /dev/null @@ -1,72 +0,0 @@ -#ifndef _EDAC_MCE_AMD_H -#define _EDAC_MCE_AMD_H - -#include - -#include - -#define ERROR_CODE(x) ((x) & 0xffff) -#define EXT_ERROR_CODE(x) (((x) >> 16) & 0x1f) -#define EXT_ERR_MSG(x) ext_msgs[EXT_ERROR_CODE(x)] - -#define LOW_SYNDROME(x) (((x) >> 15) & 0xff) -#define HIGH_SYNDROME(x) (((x) >> 24) & 0xff) - -#define TLB_ERROR(x) (((x) & 0xFFF0) == 0x0010) -#define MEM_ERROR(x) (((x) & 0xFF00) == 0x0100) -#define BUS_ERROR(x) (((x) & 0xF800) == 0x0800) - -#define TT(x) (((x) >> 2) & 0x3) -#define TT_MSG(x) tt_msgs[TT(x)] -#define II(x) (((x) >> 2) & 0x3) -#define II_MSG(x) ii_msgs[II(x)] -#define LL(x) (((x) >> 0) & 0x3) -#define LL_MSG(x) ll_msgs[LL(x)] -#define TO(x) (((x) >> 8) & 0x1) -#define TO_MSG(x) to_msgs[TO(x)] -#define PP(x) (((x) >> 9) & 0x3) -#define PP_MSG(x) pp_msgs[PP(x)] - -#define RRRR(x) (((x) >> 4) & 0xf) -#define RRRR_MSG(x) ((RRRR(x) < 9) ? rrrr_msgs[RRRR(x)] : "Wrong R4!") - -#define K8_NBSH 0x4C - -#define K8_NBSH_VALID_BIT BIT(31) -#define K8_NBSH_OVERFLOW BIT(30) -#define K8_NBSH_UC_ERR BIT(29) -#define K8_NBSH_ERR_EN BIT(28) -#define K8_NBSH_MISCV BIT(27) -#define K8_NBSH_VALID_ERROR_ADDR BIT(26) -#define K8_NBSH_PCC BIT(25) -#define K8_NBSH_ERR_CPU_VAL BIT(24) -#define K8_NBSH_CECC BIT(14) -#define K8_NBSH_UECC BIT(13) -#define K8_NBSH_ERR_SCRUBER BIT(8) - -extern const char *tt_msgs[]; -extern const char *ll_msgs[]; -extern const char *rrrr_msgs[]; -extern const char *pp_msgs[]; -extern const char *to_msgs[]; -extern const char *ii_msgs[]; -extern const char *ext_msgs[]; - -/* - * relevant NB regs - */ -struct err_regs { - u32 nbcfg; - u32 nbsh; - u32 nbsl; - u32 nbeah; - u32 nbeal; -}; - -void amd_report_gart_errors(bool); -void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)); -void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)); -void amd_decode_nb_mce(int, struct mce *, u32); -int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data); - -#endif /* _EDAC_MCE_AMD_H */ diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c new file mode 100644 index 0000000..5eb8042 --- /dev/null +++ b/drivers/edac/mce_amd.c @@ -0,0 +1,414 @@ +#include +#include "mce_amd.h" + +static bool report_gart_errors; +static void (*nb_bus_decoder)(int node_id, struct mce *m, u32 nbcfg); + +void amd_report_gart_errors(bool v) +{ + report_gart_errors = v; +} +EXPORT_SYMBOL_GPL(amd_report_gart_errors); + +void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)) +{ + nb_bus_decoder = f; +} +EXPORT_SYMBOL_GPL(amd_register_ecc_decoder); + +void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)) +{ + if (nb_bus_decoder) { + WARN_ON(nb_bus_decoder != f); + + nb_bus_decoder = NULL; + } +} +EXPORT_SYMBOL_GPL(amd_unregister_ecc_decoder); + +/* + * string representation for the different MCA reported error types, see F3x48 + * or MSR0000_0411. + */ + +/* transaction type */ +const char *tt_msgs[] = { "INSN", "DATA", "GEN", "RESV" }; +EXPORT_SYMBOL_GPL(tt_msgs); + +/* cache level */ +const char *ll_msgs[] = { "RESV", "L1", "L2", "L3/GEN" }; +EXPORT_SYMBOL_GPL(ll_msgs); + +/* memory transaction type */ +const char *rrrr_msgs[] = { + "GEN", "RD", "WR", "DRD", "DWR", "IRD", "PRF", "EV", "SNP" +}; +EXPORT_SYMBOL_GPL(rrrr_msgs); + +/* participating processor */ +const char *pp_msgs[] = { "SRC", "RES", "OBS", "GEN" }; +EXPORT_SYMBOL_GPL(pp_msgs); + +/* request timeout */ +const char *to_msgs[] = { "no timeout", "timed out" }; +EXPORT_SYMBOL_GPL(to_msgs); + +/* memory or i/o */ +const char *ii_msgs[] = { "MEM", "RESV", "IO", "GEN" }; +EXPORT_SYMBOL_GPL(ii_msgs); + +/* + * Map the 4 or 5 (family-specific) bits of Extended Error code to the + * string table. + */ +const char *ext_msgs[] = { + "K8 ECC error", /* 0_0000b */ + "CRC error on link", /* 0_0001b */ + "Sync error packets on link", /* 0_0010b */ + "Master Abort during link operation", /* 0_0011b */ + "Target Abort during link operation", /* 0_0100b */ + "Invalid GART PTE entry during table walk", /* 0_0101b */ + "Unsupported atomic RMW command received", /* 0_0110b */ + "WDT error: NB transaction timeout", /* 0_0111b */ + "ECC/ChipKill ECC error", /* 0_1000b */ + "SVM DEV Error", /* 0_1001b */ + "Link Data error", /* 0_1010b */ + "Link/L3/Probe Filter Protocol error", /* 0_1011b */ + "NB Internal Arrays Parity error", /* 0_1100b */ + "DRAM Address/Control Parity error", /* 0_1101b */ + "Link Transmission error", /* 0_1110b */ + "GART/DEV Table Walk Data error" /* 0_1111b */ + "Res 0x100 error", /* 1_0000b */ + "Res 0x101 error", /* 1_0001b */ + "Res 0x102 error", /* 1_0010b */ + "Res 0x103 error", /* 1_0011b */ + "Res 0x104 error", /* 1_0100b */ + "Res 0x105 error", /* 1_0101b */ + "Res 0x106 error", /* 1_0110b */ + "Res 0x107 error", /* 1_0111b */ + "Res 0x108 error", /* 1_1000b */ + "Res 0x109 error", /* 1_1001b */ + "Res 0x10A error", /* 1_1010b */ + "Res 0x10B error", /* 1_1011b */ + "ECC error in L3 Cache Data", /* 1_1100b */ + "L3 Cache Tag error", /* 1_1101b */ + "L3 Cache LRU Parity error", /* 1_1110b */ + "Probe Filter error" /* 1_1111b */ +}; +EXPORT_SYMBOL_GPL(ext_msgs); + +static void amd_decode_dc_mce(struct mce *m) +{ + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; + + pr_emerg(HW_ERR "Data Cache Error: "); + + if (xec == 1 && TLB_ERROR(ec)) + pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); + else if (xec == 0) { + if (m->status & (1ULL << 40)) + pr_cont(" during Data Scrub.\n"); + else if (TLB_ERROR(ec)) + pr_cont(": %s TLB parity error.\n", LL_MSG(ec)); + else if (MEM_ERROR(ec)) { + u8 ll = ec & 0x3; + u8 tt = (ec >> 2) & 0x3; + u8 rrrr = (ec >> 4) & 0xf; + + /* see F10h BKDG (31116), Table 92. */ + if (ll == 0x1) { + if (tt != 0x1) + goto wrong_dc_mce; + + pr_cont(": Data/Tag %s error.\n", RRRR_MSG(ec)); + + } else if (ll == 0x2 && rrrr == 0x3) + pr_cont(" during L1 linefill from L2.\n"); + else + goto wrong_dc_mce; + } else if (BUS_ERROR(ec) && boot_cpu_data.x86 == 0xf) + pr_cont(" during system linefill.\n"); + else + goto wrong_dc_mce; + } else + goto wrong_dc_mce; + + return; + +wrong_dc_mce: + pr_emerg(HW_ERR "Corrupted DC MCE info?\n"); +} + +static void amd_decode_ic_mce(struct mce *m) +{ + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; + + pr_emerg(HW_ERR "Instruction Cache Error"); + + if (xec == 1 && TLB_ERROR(ec)) + pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); + else if (xec == 0) { + if (TLB_ERROR(ec)) + pr_cont(": %s TLB Parity error.\n", LL_MSG(ec)); + else if (BUS_ERROR(ec)) { + if (boot_cpu_data.x86 == 0xf && + (m->status & BIT(58))) + pr_cont(" during system linefill.\n"); + else + pr_cont(" during attempted NB data read.\n"); + } else if (MEM_ERROR(ec)) { + u8 ll = ec & 0x3; + u8 rrrr = (ec >> 4) & 0xf; + + if (ll == 0x2) + pr_cont(" during a linefill from L2.\n"); + else if (ll == 0x1) { + + switch (rrrr) { + case 0x5: + pr_cont(": Parity error during " + "data load.\n"); + break; + + case 0x7: + pr_cont(": Copyback Parity/Victim" + " error.\n"); + break; + + case 0x8: + pr_cont(": Tag Snoop error.\n"); + break; + + default: + goto wrong_ic_mce; + break; + } + } + } else + goto wrong_ic_mce; + } else + goto wrong_ic_mce; + + return; + +wrong_ic_mce: + pr_emerg(HW_ERR "Corrupted IC MCE info?\n"); +} + +static void amd_decode_bu_mce(struct mce *m) +{ + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; + + pr_emerg(HW_ERR "Bus Unit Error"); + + if (xec == 0x1) + pr_cont(" in the write data buffers.\n"); + else if (xec == 0x3) + pr_cont(" in the victim data buffers.\n"); + else if (xec == 0x2 && MEM_ERROR(ec)) + pr_cont(": %s error in the L2 cache tags.\n", RRRR_MSG(ec)); + else if (xec == 0x0) { + if (TLB_ERROR(ec)) + pr_cont(": %s error in a Page Descriptor Cache or " + "Guest TLB.\n", TT_MSG(ec)); + else if (BUS_ERROR(ec)) + pr_cont(": %s/ECC error in data read from NB: %s.\n", + RRRR_MSG(ec), PP_MSG(ec)); + else if (MEM_ERROR(ec)) { + u8 rrrr = (ec >> 4) & 0xf; + + if (rrrr >= 0x7) + pr_cont(": %s error during data copyback.\n", + RRRR_MSG(ec)); + else if (rrrr <= 0x1) + pr_cont(": %s parity/ECC error during data " + "access from L2.\n", RRRR_MSG(ec)); + else + goto wrong_bu_mce; + } else + goto wrong_bu_mce; + } else + goto wrong_bu_mce; + + return; + +wrong_bu_mce: + pr_emerg(HW_ERR "Corrupted BU MCE info?\n"); +} + +static void amd_decode_ls_mce(struct mce *m) +{ + u32 ec = m->status & 0xffff; + u32 xec = (m->status >> 16) & 0xf; + + pr_emerg(HW_ERR "Load Store Error"); + + if (xec == 0x0) { + u8 rrrr = (ec >> 4) & 0xf; + + if (!BUS_ERROR(ec) || (rrrr != 0x3 && rrrr != 0x4)) + goto wrong_ls_mce; + + pr_cont(" during %s.\n", RRRR_MSG(ec)); + } + return; + +wrong_ls_mce: + pr_emerg(HW_ERR "Corrupted LS MCE info?\n"); +} + +void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg) +{ + u32 ec = m->status & 0xffff; + u32 nbsh = (u32)(m->status >> 32); + u32 nbsl = (u32)m->status; + + /* + * GART TLB error reporting is disabled by default. Bail out early. + */ + if (TLB_ERROR(ec) && !report_gart_errors) + return; + + pr_emerg(HW_ERR "Northbridge Error, node %d", node_id); + + /* + * F10h, revD can disable ErrCpu[3:0] so check that first and also the + * value encoding has changed so interpret those differently + */ + if ((boot_cpu_data.x86 == 0x10) && + (boot_cpu_data.x86_model > 7)) { + if (nbsh & K8_NBSH_ERR_CPU_VAL) + pr_cont(", core: %u\n", (u8)(nbsh & 0xf)); + } else { + u8 assoc_cpus = nbsh & 0xf; + + if (assoc_cpus > 0) + pr_cont(", core: %d", fls(assoc_cpus) - 1); + + pr_cont("\n"); + } + + pr_emerg(HW_ERR "%s.\n", EXT_ERR_MSG(nbsl)); + + if (BUS_ERROR(ec) && nb_bus_decoder) + nb_bus_decoder(node_id, m, nbcfg); +} +EXPORT_SYMBOL_GPL(amd_decode_nb_mce); + +static void amd_decode_fr_mce(struct mce *m) +{ + /* we have only one error signature so match all fields at once. */ + if ((m->status & 0xffff) == 0x0f0f) + pr_emerg(HW_ERR " FR Error: CPU Watchdog timer expire.\n"); + else + pr_emerg(HW_ERR "Corrupted FR MCE info?\n"); +} + +static inline void amd_decode_err_code(u16 ec) +{ + if (TLB_ERROR(ec)) { + pr_emerg(HW_ERR "Transaction: %s, Cache Level: %s\n", + TT_MSG(ec), LL_MSG(ec)); + } else if (MEM_ERROR(ec)) { + pr_emerg(HW_ERR "Transaction: %s, Type: %s, Cache Level: %s\n", + RRRR_MSG(ec), TT_MSG(ec), LL_MSG(ec)); + } else if (BUS_ERROR(ec)) { + pr_emerg(HW_ERR "Transaction: %s (%s), %s, Cache Level: %s, " + "Participating Processor: %s\n", + RRRR_MSG(ec), II_MSG(ec), TO_MSG(ec), LL_MSG(ec), + PP_MSG(ec)); + } else + pr_emerg(HW_ERR "Huh? Unknown MCE error 0x%x\n", ec); +} + +int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data) +{ + struct mce *m = (struct mce *)data; + int node, ecc; + + pr_emerg(HW_ERR "MC%d_STATUS: ", m->bank); + + pr_cont("%sorrected error, other errors lost: %s, " + "CPU context corrupt: %s", + ((m->status & MCI_STATUS_UC) ? "Unc" : "C"), + ((m->status & MCI_STATUS_OVER) ? "yes" : "no"), + ((m->status & MCI_STATUS_PCC) ? "yes" : "no")); + + /* do the two bits[14:13] together */ + ecc = (m->status >> 45) & 0x3; + if (ecc) + pr_cont(", %sECC Error", ((ecc == 2) ? "C" : "U")); + + pr_cont("\n"); + + switch (m->bank) { + case 0: + amd_decode_dc_mce(m); + break; + + case 1: + amd_decode_ic_mce(m); + break; + + case 2: + amd_decode_bu_mce(m); + break; + + case 3: + amd_decode_ls_mce(m); + break; + + case 4: + node = amd_get_nb_id(m->extcpu); + amd_decode_nb_mce(node, m, 0); + break; + + case 5: + amd_decode_fr_mce(m); + break; + + default: + break; + } + + amd_decode_err_code(m->status & 0xffff); + + return NOTIFY_STOP; +} +EXPORT_SYMBOL_GPL(amd_decode_mce); + +static struct notifier_block amd_mce_dec_nb = { + .notifier_call = amd_decode_mce, +}; + +static int __init mce_amd_init(void) +{ + /* + * We can decode MCEs for K8, F10h and F11h CPUs: + */ + if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) + return 0; + + if (boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) + return 0; + + atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb); + + return 0; +} +early_initcall(mce_amd_init); + +#ifdef MODULE +static void __exit mce_amd_exit(void) +{ + atomic_notifier_chain_unregister(&x86_mce_decoder_chain, &amd_mce_dec_nb); +} + +MODULE_DESCRIPTION("AMD MCE decoder"); +MODULE_ALIAS("edac-mce-amd"); +MODULE_LICENSE("GPL"); +module_exit(mce_amd_exit); +#endif diff --git a/drivers/edac/mce_amd.h b/drivers/edac/mce_amd.h new file mode 100644 index 0000000..2712a90 --- /dev/null +++ b/drivers/edac/mce_amd.h @@ -0,0 +1,72 @@ +#ifndef _EDAC_MCE_AMD_H +#define _EDAC_MCE_AMD_H + +#include + +#include + +#define ERROR_CODE(x) ((x) & 0xffff) +#define EXT_ERROR_CODE(x) (((x) >> 16) & 0x1f) +#define EXT_ERR_MSG(x) ext_msgs[EXT_ERROR_CODE(x)] + +#define LOW_SYNDROME(x) (((x) >> 15) & 0xff) +#define HIGH_SYNDROME(x) (((x) >> 24) & 0xff) + +#define TLB_ERROR(x) (((x) & 0xFFF0) == 0x0010) +#define MEM_ERROR(x) (((x) & 0xFF00) == 0x0100) +#define BUS_ERROR(x) (((x) & 0xF800) == 0x0800) + +#define TT(x) (((x) >> 2) & 0x3) +#define TT_MSG(x) tt_msgs[TT(x)] +#define II(x) (((x) >> 2) & 0x3) +#define II_MSG(x) ii_msgs[II(x)] +#define LL(x) (((x) >> 0) & 0x3) +#define LL_MSG(x) ll_msgs[LL(x)] +#define TO(x) (((x) >> 8) & 0x1) +#define TO_MSG(x) to_msgs[TO(x)] +#define PP(x) (((x) >> 9) & 0x3) +#define PP_MSG(x) pp_msgs[PP(x)] + +#define RRRR(x) (((x) >> 4) & 0xf) +#define RRRR_MSG(x) ((RRRR(x) < 9) ? rrrr_msgs[RRRR(x)] : "Wrong R4!") + +#define K8_NBSH 0x4C + +#define K8_NBSH_VALID_BIT BIT(31) +#define K8_NBSH_OVERFLOW BIT(30) +#define K8_NBSH_UC_ERR BIT(29) +#define K8_NBSH_ERR_EN BIT(28) +#define K8_NBSH_MISCV BIT(27) +#define K8_NBSH_VALID_ERROR_ADDR BIT(26) +#define K8_NBSH_PCC BIT(25) +#define K8_NBSH_ERR_CPU_VAL BIT(24) +#define K8_NBSH_CECC BIT(14) +#define K8_NBSH_UECC BIT(13) +#define K8_NBSH_ERR_SCRUBER BIT(8) + +extern const char *tt_msgs[]; +extern const char *ll_msgs[]; +extern const char *rrrr_msgs[]; +extern const char *pp_msgs[]; +extern const char *to_msgs[]; +extern const char *ii_msgs[]; +extern const char *ext_msgs[]; + +/* + * relevant NB regs + */ +struct err_regs { + u32 nbcfg; + u32 nbsh; + u32 nbsl; + u32 nbeah; + u32 nbeal; +}; + +void amd_report_gart_errors(bool); +void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)); +void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)); +void amd_decode_nb_mce(int, struct mce *, u32); +int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data); + +#endif /* _EDAC_MCE_AMD_H */ diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index 0e4f2dc..8d0688f 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -15,7 +15,7 @@ #include #include -#include "edac_mce_amd.h" +#include "mce_amd.h" struct edac_mce_attr { struct attribute attr; -- cgit v0.10.2 From 888ab8e6eb2e41179cdc8edf5d0abd1cce0f0370 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 18 Aug 2010 15:11:35 +0200 Subject: EDAC, MCE: Adjust DC decoders to F14h Add a per-family data cache decoders. Since there is a certain overlap between the different DC MCE signatures, reuse functionality between the families as far as possible. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 5eb8042..33985aa 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -1,6 +1,10 @@ #include +#include + #include "mce_amd.h" +static struct amd_decoder_ops *fam_ops; + static bool report_gart_errors; static void (*nb_bus_decoder)(int node_id, struct mce *m, u32 nbcfg); @@ -97,41 +101,116 @@ const char *ext_msgs[] = { }; EXPORT_SYMBOL_GPL(ext_msgs); -static void amd_decode_dc_mce(struct mce *m) +static bool f10h_dc_mce(u16 ec) { - u32 ec = m->status & 0xffff; - u32 xec = (m->status >> 16) & 0xf; + u8 r4 = (ec >> 4) & 0xf; + bool ret = false; - pr_emerg(HW_ERR "Data Cache Error: "); + if (r4 == R4_GEN) { + pr_cont("during data scrub.\n"); + return true; + } - if (xec == 1 && TLB_ERROR(ec)) - pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); - else if (xec == 0) { - if (m->status & (1ULL << 40)) - pr_cont(" during Data Scrub.\n"); - else if (TLB_ERROR(ec)) - pr_cont(": %s TLB parity error.\n", LL_MSG(ec)); - else if (MEM_ERROR(ec)) { - u8 ll = ec & 0x3; - u8 tt = (ec >> 2) & 0x3; - u8 rrrr = (ec >> 4) & 0xf; + if (MEM_ERROR(ec)) { + u8 ll = ec & 0x3; + ret = true; - /* see F10h BKDG (31116), Table 92. */ - if (ll == 0x1) { - if (tt != 0x1) - goto wrong_dc_mce; + if (ll == LL_L2) + pr_cont("during L1 linefill from L2.\n"); + else if (ll == LL_L1) + pr_cont("Data/Tag %s error.\n", RRRR_MSG(ec)); + else + ret = false; + } + return ret; +} - pr_cont(": Data/Tag %s error.\n", RRRR_MSG(ec)); +static bool k8_dc_mce(u16 ec) +{ + if (BUS_ERROR(ec)) { + pr_cont("during system linefill.\n"); + return true; + } - } else if (ll == 0x2 && rrrr == 0x3) - pr_cont(" during L1 linefill from L2.\n"); - else - goto wrong_dc_mce; - } else if (BUS_ERROR(ec) && boot_cpu_data.x86 == 0xf) - pr_cont(" during system linefill.\n"); + return f10h_dc_mce(ec); +} + +static bool f14h_dc_mce(u16 ec) +{ + u8 r4 = (ec >> 4) & 0xf; + u8 ll = ec & 0x3; + u8 tt = (ec >> 2) & 0x3; + u8 ii = tt; + bool ret = true; + + if (MEM_ERROR(ec)) { + + if (tt != TT_DATA || ll != LL_L1) + return false; + + switch (r4) { + case R4_DRD: + case R4_DWR: + pr_cont("Data/Tag parity error due to %s.\n", + (r4 == R4_DRD ? "load/hw prf" : "store")); + break; + case R4_EVICT: + pr_cont("Copyback parity error on a tag miss.\n"); + break; + case R4_SNOOP: + pr_cont("Tag parity error during snoop.\n"); + break; + default: + ret = false; + } + } else if (BUS_ERROR(ec)) { + + if ((ii != II_MEM && ii != II_IO) || ll != LL_LG) + return false; + + pr_cont("System read data error on a "); + + switch (r4) { + case R4_RD: + pr_cont("TLB reload.\n"); + break; + case R4_DWR: + pr_cont("store.\n"); + break; + case R4_DRD: + pr_cont("load.\n"); + break; + default: + ret = false; + } + } else { + ret = false; + } + + return ret; +} + +static void amd_decode_dc_mce(struct mce *m) +{ + u16 ec = m->status & 0xffff; + u8 xec = (m->status >> 16) & 0xf; + + pr_emerg(HW_ERR "Data Cache Error: "); + + /* TLB error signatures are the same across families */ + if (TLB_ERROR(ec)) { + u8 tt = (ec >> 2) & 0x3; + + if (tt == TT_DATA) { + pr_cont("%s TLB %s.\n", LL_MSG(ec), + (xec ? "multimatch" : "parity error")); + return; + } else goto wrong_dc_mce; - } else + } + + if (!fam_ops->dc_mce(ec)) goto wrong_dc_mce; return; @@ -395,6 +474,30 @@ static int __init mce_amd_init(void) if (boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) return 0; + fam_ops = kzalloc(sizeof(struct amd_decoder_ops), GFP_KERNEL); + if (!fam_ops) + return -ENOMEM; + + switch (boot_cpu_data.x86) { + case 0xf: + fam_ops->dc_mce = k8_dc_mce; + break; + + case 0x10: + fam_ops->dc_mce = f10h_dc_mce; + break; + + case 0x14: + fam_ops->dc_mce = f14h_dc_mce; + break; + + default: + printk(KERN_WARNING "Huh? What family is that: %d?!\n", + boot_cpu_data.x86); + kfree(fam_ops); + return -EINVAL; + } + atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb); return 0; @@ -405,6 +508,7 @@ early_initcall(mce_amd_init); static void __exit mce_amd_exit(void) { atomic_notifier_chain_unregister(&x86_mce_decoder_chain, &amd_mce_dec_nb); + kfree(fam_ops); } MODULE_DESCRIPTION("AMD MCE decoder"); diff --git a/drivers/edac/mce_amd.h b/drivers/edac/mce_amd.h index 2712a90..85985c2 100644 --- a/drivers/edac/mce_amd.h +++ b/drivers/edac/mce_amd.h @@ -44,6 +44,39 @@ #define K8_NBSH_UECC BIT(13) #define K8_NBSH_ERR_SCRUBER BIT(8) +enum tt_ids { + TT_INSTR = 0, + TT_DATA, + TT_GEN, + TT_RESV, +}; + +enum ll_ids { + LL_RESV = 0, + LL_L1, + LL_L2, + LL_LG, +}; + +enum ii_ids { + II_MEM = 0, + II_RESV, + II_IO, + II_GEN, +}; + +enum rrrr_ids { + R4_GEN = 0, + R4_RD, + R4_WR, + R4_DRD, + R4_DWR, + R4_IRD, + R4_PREF, + R4_EVICT, + R4_SNOOP, +}; + extern const char *tt_msgs[]; extern const char *ll_msgs[]; extern const char *rrrr_msgs[]; @@ -63,6 +96,13 @@ struct err_regs { u32 nbeal; }; +/* + * per-family decoder ops + */ +struct amd_decoder_ops { + bool (*dc_mce)(u16); +}; + void amd_report_gart_errors(bool); void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32)); void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32)); -- cgit v0.10.2 From dd53bce4e8987f6848840d42bbeead5221eff308 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 26 Aug 2010 19:05:49 +0200 Subject: EDAC, MCE: Adjust IC decoders to F14h Add support for IC MCEs for F14h CPUs. K8 and F10h are almost identical so use one function for both. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 33985aa..60d5d9f 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -219,61 +219,80 @@ wrong_dc_mce: pr_emerg(HW_ERR "Corrupted DC MCE info?\n"); } -static void amd_decode_ic_mce(struct mce *m) +static bool k8_ic_mce(u16 ec) { - u32 ec = m->status & 0xffff; - u32 xec = (m->status >> 16) & 0xf; + u8 ll = ec & 0x3; + u8 r4 = (ec >> 4) & 0xf; + bool ret = true; - pr_emerg(HW_ERR "Instruction Cache Error"); + if (!MEM_ERROR(ec)) + return false; - if (xec == 1 && TLB_ERROR(ec)) - pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); - else if (xec == 0) { - if (TLB_ERROR(ec)) - pr_cont(": %s TLB Parity error.\n", LL_MSG(ec)); - else if (BUS_ERROR(ec)) { - if (boot_cpu_data.x86 == 0xf && - (m->status & BIT(58))) - pr_cont(" during system linefill.\n"); - else - pr_cont(" during attempted NB data read.\n"); - } else if (MEM_ERROR(ec)) { - u8 ll = ec & 0x3; - u8 rrrr = (ec >> 4) & 0xf; + if (ll == 0x2) + pr_cont("during a linefill from L2.\n"); + else if (ll == 0x1) { + switch (r4) { + case R4_IRD: + pr_cont("Parity error during data load.\n"); + break; - if (ll == 0x2) - pr_cont(" during a linefill from L2.\n"); - else if (ll == 0x1) { - - switch (rrrr) { - case 0x5: - pr_cont(": Parity error during " - "data load.\n"); - break; - - case 0x7: - pr_cont(": Copyback Parity/Victim" - " error.\n"); - break; - - case 0x8: - pr_cont(": Tag Snoop error.\n"); - break; - - default: - goto wrong_ic_mce; - break; - } - } - } else - goto wrong_ic_mce; + case R4_EVICT: + pr_cont("Copyback Parity/Victim error.\n"); + break; + + case R4_SNOOP: + pr_cont("Tag Snoop error.\n"); + break; + + default: + ret = false; + break; + } } else - goto wrong_ic_mce; + ret = false; - return; + return ret; +} + +static bool f14h_ic_mce(u16 ec) +{ + u8 ll = ec & 0x3; + u8 tt = (ec >> 2) & 0x3; + u8 r4 = (ec >> 4) & 0xf; + bool ret = true; -wrong_ic_mce: - pr_emerg(HW_ERR "Corrupted IC MCE info?\n"); + if (MEM_ERROR(ec)) { + if (tt != 0 || ll != 1) + ret = false; + + if (r4 == R4_IRD) + pr_cont("Data/tag array parity error for a tag hit.\n"); + else if (r4 == R4_SNOOP) + pr_cont("Tag error during snoop/victimization.\n"); + else + ret = false; + } + return ret; +} + +static void amd_decode_ic_mce(struct mce *m) +{ + u16 ec = m->status & 0xffff; + u8 xec = (m->status >> 16) & 0xf; + + pr_emerg(HW_ERR "Instruction Cache Error: "); + + if (TLB_ERROR(ec)) + pr_cont("%s TLB %s.\n", LL_MSG(ec), + (xec ? "multimatch" : "parity error")); + else if (BUS_ERROR(ec)) { + bool k8 = (boot_cpu_data.x86 == 0xf && (m->status & BIT(58))); + + pr_cont("during %s.\n", (k8 ? "system linefill" : "NB data read")); + } else if (fam_ops->ic_mce(ec)) + ; + else + pr_emerg(HW_ERR "Corrupted IC MCE info?\n"); } static void amd_decode_bu_mce(struct mce *m) @@ -481,14 +500,17 @@ static int __init mce_amd_init(void) switch (boot_cpu_data.x86) { case 0xf: fam_ops->dc_mce = k8_dc_mce; + fam_ops->ic_mce = k8_ic_mce; break; case 0x10: fam_ops->dc_mce = f10h_dc_mce; + fam_ops->ic_mce = k8_ic_mce; break; case 0x14: fam_ops->dc_mce = f14h_dc_mce; + fam_ops->ic_mce = f14h_ic_mce; break; default: diff --git a/drivers/edac/mce_amd.h b/drivers/edac/mce_amd.h index 85985c2..dc81dba 100644 --- a/drivers/edac/mce_amd.h +++ b/drivers/edac/mce_amd.h @@ -101,6 +101,7 @@ struct err_regs { */ struct amd_decoder_ops { bool (*dc_mce)(u16); + bool (*ic_mce)(u16); }; void amd_report_gart_errors(bool); -- cgit v0.10.2 From ded506232865e8e932bc21c87f48170d50db4d97 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 27 Aug 2010 17:03:34 +0200 Subject: EDAC, MCE: Warn about LS MCEs on F14h F14h CPUs do not generate LS MCEs so exit early and warn the user in case this path is ever hit that something else might be going haywire. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 60d5d9f..3c16167 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -339,19 +339,27 @@ wrong_bu_mce: static void amd_decode_ls_mce(struct mce *m) { - u32 ec = m->status & 0xffff; - u32 xec = (m->status >> 16) & 0xf; + u16 ec = m->status & 0xffff; + u8 xec = (m->status >> 16) & 0xf; + + if (boot_cpu_data.x86 == 0x14) { + pr_emerg("You shouldn't be seeing an LS MCE on this cpu family," + " please report on LKML.\n"); + return; + } pr_emerg(HW_ERR "Load Store Error"); if (xec == 0x0) { - u8 rrrr = (ec >> 4) & 0xf; + u8 r4 = (ec >> 4) & 0xf; - if (!BUS_ERROR(ec) || (rrrr != 0x3 && rrrr != 0x4)) + if (!BUS_ERROR(ec) || (r4 != R4_DRD && r4 != R4_DWR)) goto wrong_ls_mce; pr_cont(" during %s.\n", RRRR_MSG(ec)); - } + } else + goto wrong_ls_mce; + return; wrong_ls_mce: -- cgit v0.10.2 From 5ce88f6ea6bef929f59f9468413f922c9a486fa4 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Tue, 31 Aug 2010 18:28:08 +0200 Subject: EDAC, MCE: Complete NB MCE decoders Add support for decoding F14h BU MCEs and improve decoding of the remaining families. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/amd64_edac.h b/drivers/edac/amd64_edac.h index 13e1d6f..044aee4 100644 --- a/drivers/edac/amd64_edac.h +++ b/drivers/edac/amd64_edac.h @@ -482,7 +482,6 @@ extern const char *rrrr_msgs[16]; extern const char *to_msgs[2]; extern const char *pp_msgs[4]; extern const char *ii_msgs[4]; -extern const char *ext_msgs[32]; extern const char *htlink_msgs[8]; #ifdef CONFIG_EDAC_DEBUG diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 3c16167..d8d1c9d 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -5,6 +5,8 @@ static struct amd_decoder_ops *fam_ops; +static u8 nb_err_cpumask = 0xf; + static bool report_gart_errors; static void (*nb_bus_decoder)(int node_id, struct mce *m, u32 nbcfg); @@ -61,45 +63,16 @@ EXPORT_SYMBOL_GPL(to_msgs); const char *ii_msgs[] = { "MEM", "RESV", "IO", "GEN" }; EXPORT_SYMBOL_GPL(ii_msgs); -/* - * Map the 4 or 5 (family-specific) bits of Extended Error code to the - * string table. - */ -const char *ext_msgs[] = { - "K8 ECC error", /* 0_0000b */ - "CRC error on link", /* 0_0001b */ - "Sync error packets on link", /* 0_0010b */ - "Master Abort during link operation", /* 0_0011b */ - "Target Abort during link operation", /* 0_0100b */ - "Invalid GART PTE entry during table walk", /* 0_0101b */ - "Unsupported atomic RMW command received", /* 0_0110b */ - "WDT error: NB transaction timeout", /* 0_0111b */ - "ECC/ChipKill ECC error", /* 0_1000b */ - "SVM DEV Error", /* 0_1001b */ - "Link Data error", /* 0_1010b */ - "Link/L3/Probe Filter Protocol error", /* 0_1011b */ - "NB Internal Arrays Parity error", /* 0_1100b */ - "DRAM Address/Control Parity error", /* 0_1101b */ - "Link Transmission error", /* 0_1110b */ - "GART/DEV Table Walk Data error" /* 0_1111b */ - "Res 0x100 error", /* 1_0000b */ - "Res 0x101 error", /* 1_0001b */ - "Res 0x102 error", /* 1_0010b */ - "Res 0x103 error", /* 1_0011b */ - "Res 0x104 error", /* 1_0100b */ - "Res 0x105 error", /* 1_0101b */ - "Res 0x106 error", /* 1_0110b */ - "Res 0x107 error", /* 1_0111b */ - "Res 0x108 error", /* 1_1000b */ - "Res 0x109 error", /* 1_1001b */ - "Res 0x10A error", /* 1_1010b */ - "Res 0x10B error", /* 1_1011b */ - "ECC error in L3 Cache Data", /* 1_1100b */ - "L3 Cache Tag error", /* 1_1101b */ - "L3 Cache LRU Parity error", /* 1_1110b */ - "Probe Filter error" /* 1_1111b */ +static const char *f10h_nb_mce_desc[] = { + "HT link data error", + "Protocol error (link, L3, probe filter, etc.)", + "Parity error in NB-internal arrays", + "Link Retry due to IO link transmission error", + "L3 ECC data cache error", + "ECC error in L3 cache tag", + "L3 LRU parity bits error", + "ECC Error in the Probe Filter directory" }; -EXPORT_SYMBOL_GPL(ext_msgs); static bool f10h_dc_mce(u16 ec) { @@ -366,19 +339,97 @@ wrong_ls_mce: pr_emerg(HW_ERR "Corrupted LS MCE info?\n"); } +static bool k8_nb_mce(u16 ec, u8 xec) +{ + bool ret = true; + + switch (xec) { + case 0x1: + pr_cont("CRC error detected on HT link.\n"); + break; + + case 0x5: + pr_cont("Invalid GART PTE entry during GART table walk.\n"); + break; + + case 0x6: + pr_cont("Unsupported atomic RMW received from an IO link.\n"); + break; + + case 0x0: + case 0x8: + pr_cont("DRAM ECC error detected on the NB.\n"); + break; + + case 0xd: + pr_cont("Parity error on the DRAM addr/ctl signals.\n"); + break; + + default: + ret = false; + break; + } + + return ret; +} + +static bool f10h_nb_mce(u16 ec, u8 xec) +{ + bool ret = true; + u8 offset = 0; + + if (k8_nb_mce(ec, xec)) + return true; + + switch(xec) { + case 0xa ... 0xc: + offset = 10; + break; + + case 0xe: + offset = 11; + break; + + case 0xf: + if (TLB_ERROR(ec)) + pr_cont("GART Table Walk data error.\n"); + else if (BUS_ERROR(ec)) + pr_cont("DMA Exclusion Vector Table Walk error.\n"); + else + ret = false; + + goto out; + break; + + case 0x1c ... 0x1f: + offset = 24; + break; + + default: + ret = false; + + goto out; + break; + } + + pr_cont("%s.\n", f10h_nb_mce_desc[xec - offset]); + +out: + return ret; +} + +static bool f14h_nb_mce(u16 ec, u8 xec) +{ + return false; +} + void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg) { - u32 ec = m->status & 0xffff; + u8 xec = (m->status >> 16) & 0x1f; + u16 ec = m->status & 0xffff; u32 nbsh = (u32)(m->status >> 32); - u32 nbsl = (u32)m->status; - - /* - * GART TLB error reporting is disabled by default. Bail out early. - */ - if (TLB_ERROR(ec) && !report_gart_errors) - return; - pr_emerg(HW_ERR "Northbridge Error, node %d", node_id); + pr_emerg(HW_ERR "Northbridge Error, node %d: ", node_id); /* * F10h, revD can disable ErrCpu[3:0] so check that first and also the @@ -387,20 +438,50 @@ void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg) if ((boot_cpu_data.x86 == 0x10) && (boot_cpu_data.x86_model > 7)) { if (nbsh & K8_NBSH_ERR_CPU_VAL) - pr_cont(", core: %u\n", (u8)(nbsh & 0xf)); + pr_cont(", core: %u", (u8)(nbsh & nb_err_cpumask)); } else { - u8 assoc_cpus = nbsh & 0xf; + u8 assoc_cpus = nbsh & nb_err_cpumask; if (assoc_cpus > 0) pr_cont(", core: %d", fls(assoc_cpus) - 1); + } - pr_cont("\n"); + switch (xec) { + case 0x2: + pr_cont("Sync error (sync packets on HT link detected).\n"); + return; + + case 0x3: + pr_cont("HT Master abort.\n"); + return; + + case 0x4: + pr_cont("HT Target abort.\n"); + return; + + case 0x7: + pr_cont("NB Watchdog timeout.\n"); + return; + + case 0x9: + pr_cont("SVM DMA Exclusion Vector error.\n"); + return; + + default: + break; } - pr_emerg(HW_ERR "%s.\n", EXT_ERR_MSG(nbsl)); + if (!fam_ops->nb_mce(ec, xec)) + goto wrong_nb_mce; + + if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10) + if ((xec == 0x8 || xec == 0x0) && nb_bus_decoder) + nb_bus_decoder(node_id, m, nbcfg); - if (BUS_ERROR(ec) && nb_bus_decoder) - nb_bus_decoder(node_id, m, nbcfg); + return; + +wrong_nb_mce: + pr_emerg(HW_ERR "Corrupted NB MCE info?\n"); } EXPORT_SYMBOL_GPL(amd_decode_nb_mce); @@ -430,11 +511,30 @@ static inline void amd_decode_err_code(u16 ec) pr_emerg(HW_ERR "Huh? Unknown MCE error 0x%x\n", ec); } +/* + * Filter out unwanted MCE signatures here. + */ +static bool amd_filter_mce(struct mce *m) +{ + u8 xec = (m->status >> 16) & 0x1f; + + /* + * NB GART TLB error reporting is disabled by default. + */ + if (m->bank == 4 && xec == 0x5 && !report_gart_errors) + return true; + + return false; +} + int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data) { struct mce *m = (struct mce *)data; int node, ecc; + if (amd_filter_mce(m)) + return NOTIFY_STOP; + pr_emerg(HW_ERR "MC%d_STATUS: ", m->bank); pr_cont("%sorrected error, other errors lost: %s, " @@ -509,16 +609,20 @@ static int __init mce_amd_init(void) case 0xf: fam_ops->dc_mce = k8_dc_mce; fam_ops->ic_mce = k8_ic_mce; + fam_ops->nb_mce = k8_nb_mce; break; case 0x10: fam_ops->dc_mce = f10h_dc_mce; fam_ops->ic_mce = k8_ic_mce; + fam_ops->nb_mce = f10h_nb_mce; break; case 0x14: + nb_err_cpumask = 0x3; fam_ops->dc_mce = f14h_dc_mce; fam_ops->ic_mce = f14h_ic_mce; + fam_ops->nb_mce = f14h_nb_mce; break; default: diff --git a/drivers/edac/mce_amd.h b/drivers/edac/mce_amd.h index dc81dba..0d0637d 100644 --- a/drivers/edac/mce_amd.h +++ b/drivers/edac/mce_amd.h @@ -7,7 +7,6 @@ #define ERROR_CODE(x) ((x) & 0xffff) #define EXT_ERROR_CODE(x) (((x) >> 16) & 0x1f) -#define EXT_ERR_MSG(x) ext_msgs[EXT_ERROR_CODE(x)] #define LOW_SYNDROME(x) (((x) >> 15) & 0xff) #define HIGH_SYNDROME(x) (((x) >> 24) & 0xff) @@ -83,7 +82,6 @@ extern const char *rrrr_msgs[]; extern const char *pp_msgs[]; extern const char *to_msgs[]; extern const char *ii_msgs[]; -extern const char *ext_msgs[]; /* * relevant NB regs @@ -102,6 +100,7 @@ struct err_regs { struct amd_decoder_ops { bool (*dc_mce)(u16); bool (*ic_mce)(u16); + bool (*nb_mce)(u16, u8); }; void amd_report_gart_errors(bool); -- cgit v0.10.2 From fe4ea2623bec3e595f8e77a8514307c389c096ae Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Tue, 31 Aug 2010 18:38:24 +0200 Subject: EDAC, MCE: Fix FR MCEs decoding Those are N/A on K8, so don't decode them there. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index d8d1c9d..83b7b5f 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -487,11 +487,17 @@ EXPORT_SYMBOL_GPL(amd_decode_nb_mce); static void amd_decode_fr_mce(struct mce *m) { + if (boot_cpu_data.x86 == 0xf) + goto wrong_fr_mce; + /* we have only one error signature so match all fields at once. */ - if ((m->status & 0xffff) == 0x0f0f) - pr_emerg(HW_ERR " FR Error: CPU Watchdog timer expire.\n"); - else - pr_emerg(HW_ERR "Corrupted FR MCE info?\n"); + if ((m->status & 0xffff) == 0x0f0f) { + pr_emerg(HW_ERR "FR Error: CPU Watchdog timer expire.\n"); + return; + } + +wrong_fr_mce: + pr_emerg(HW_ERR "Corrupted FR MCE info?\n"); } static inline void amd_decode_err_code(u16 ec) -- cgit v0.10.2 From 9530d608ef0e1f76b7fd82bb92645062292fc009 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Mon, 6 Sep 2010 15:05:45 +0200 Subject: EDAC, MCE: Enable MCE decoding on F14h Now that all decoders have been taught about F14h, models < 0x10 MCEs, enable decoding on this family of CPUs. Also, issue a short informational message upon boot that MCE decoding gets enabled. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 83b7b5f..f233c5f 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -598,13 +598,12 @@ static struct notifier_block amd_mce_dec_nb = { static int __init mce_amd_init(void) { - /* - * We can decode MCEs for K8, F10h and F11h CPUs: - */ if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) return 0; - if (boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) + if (boot_cpu_data.x86 != 0xf && + boot_cpu_data.x86 != 0x10 && + (boot_cpu_data.x86 != 0x14 || boot_cpu_data.x86_model > 0xf)) return 0; fam_ops = kzalloc(sizeof(struct amd_decoder_ops), GFP_KERNEL); @@ -638,6 +637,8 @@ static int __init mce_amd_init(void) return -EINVAL; } + pr_info("MCE: In-kernel MCE decoding enabled.\n"); + atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb); return 0; -- cgit v0.10.2 From f0157b3afd2ec6331245768a785487249a3c9734 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Tue, 5 Oct 2010 19:07:16 +0200 Subject: EDAC, MCE: Add support for F11h MCEs F11h has almost the same MCE signatures as K8 except DRAM ECC and MC5 bank errors. Reuse functionality from the other families. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index f233c5f..9fa61ee 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -358,6 +358,9 @@ static bool k8_nb_mce(u16 ec, u8 xec) case 0x0: case 0x8: + if (boot_cpu_data.x86 == 0x11) + return false; + pr_cont("DRAM ECC error detected on the NB.\n"); break; @@ -487,7 +490,8 @@ EXPORT_SYMBOL_GPL(amd_decode_nb_mce); static void amd_decode_fr_mce(struct mce *m) { - if (boot_cpu_data.x86 == 0xf) + if (boot_cpu_data.x86 == 0xf || + boot_cpu_data.x86 == 0x11) goto wrong_fr_mce; /* we have only one error signature so match all fields at once. */ @@ -601,8 +605,7 @@ static int __init mce_amd_init(void) if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) return 0; - if (boot_cpu_data.x86 != 0xf && - boot_cpu_data.x86 != 0x10 && + if ((boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) && (boot_cpu_data.x86 != 0x14 || boot_cpu_data.x86_model > 0xf)) return 0; @@ -623,6 +626,12 @@ static int __init mce_amd_init(void) fam_ops->nb_mce = f10h_nb_mce; break; + case 0x11: + fam_ops->dc_mce = k8_dc_mce; + fam_ops->ic_mce = k8_ic_mce; + fam_ops->nb_mce = f10h_nb_mce; + break; + case 0x14: nb_err_cpumask = 0x3; fam_ops->dc_mce = f14h_dc_mce; -- cgit v0.10.2 From 9be0bb1072e3544934e0ac20f184e50805aecf9c Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 16 Sep 2010 15:08:14 +0200 Subject: EDAC, MCE: Add F12h DC MCE decoder F12h DC MCE signatures are a subset of F10h's so reuse them. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 9fa61ee..1f895df 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -74,16 +74,10 @@ static const char *f10h_nb_mce_desc[] = { "ECC Error in the Probe Filter directory" }; -static bool f10h_dc_mce(u16 ec) +static bool f12h_dc_mce(u16 ec) { - u8 r4 = (ec >> 4) & 0xf; bool ret = false; - if (r4 == R4_GEN) { - pr_cont("during data scrub.\n"); - return true; - } - if (MEM_ERROR(ec)) { u8 ll = ec & 0x3; ret = true; @@ -98,6 +92,18 @@ static bool f10h_dc_mce(u16 ec) return ret; } +static bool f10h_dc_mce(u16 ec) +{ + u8 r4 = (ec >> 4) & 0xf; + u8 ll = ec & 0x3; + + if (r4 == R4_GEN && ll == LL_L1) { + pr_cont("during data scrub.\n"); + return true; + } + return f12h_dc_mce(ec); +} + static bool k8_dc_mce(u16 ec) { if (BUS_ERROR(ec)) { @@ -632,6 +638,10 @@ static int __init mce_amd_init(void) fam_ops->nb_mce = f10h_nb_mce; break; + case 0x12: + fam_ops->dc_mce = f12h_dc_mce; + break; + case 0x14: nb_err_cpumask = 0x3; fam_ops->dc_mce = f14h_dc_mce; -- cgit v0.10.2 From e7281eb37da045abac5bd795d1169fc2e3eeea49 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 16 Sep 2010 16:45:22 +0200 Subject: EDAC, MCE: Add F12h IC MCE decoder ... which is the same as for K8 and F10h. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 1f895df..fc3712f 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -640,6 +640,7 @@ static int __init mce_amd_init(void) case 0x12: fam_ops->dc_mce = f12h_dc_mce; + fam_ops->ic_mce = k8_ic_mce; break; case 0x14: -- cgit v0.10.2 From cb9d5ecdff66197f65a6be8032ccc1ebf7199684 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 16 Sep 2010 17:36:12 +0200 Subject: EDAC, MCE: Add F12h NB MCE decoder F12h is completely covered by the generic path. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index fc3712f..6501392 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -427,7 +427,7 @@ out: return ret; } -static bool f14h_nb_mce(u16 ec, u8 xec) +static bool nb_noop_mce(u16 ec, u8 xec) { return false; } @@ -641,13 +641,14 @@ static int __init mce_amd_init(void) case 0x12: fam_ops->dc_mce = f12h_dc_mce; fam_ops->ic_mce = k8_ic_mce; + fam_ops->nb_mce = nb_noop_mce; break; case 0x14: nb_err_cpumask = 0x3; fam_ops->dc_mce = f14h_dc_mce; fam_ops->ic_mce = f14h_ic_mce; - fam_ops->nb_mce = f14h_nb_mce; + fam_ops->nb_mce = nb_noop_mce; break; default: -- cgit v0.10.2 From fda7561f438aeddf074e2db0890e283195aa7779 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 22 Sep 2010 16:12:03 +0200 Subject: EDAC, MCE: Enable MCE decoding on F12h Turn on MCE decoding on F12h. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 6501392..7f74f0f 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -611,7 +611,7 @@ static int __init mce_amd_init(void) if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) return 0; - if ((boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) && + if ((boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x12) && (boot_cpu_data.x86 != 0x14 || boot_cpu_data.x86_model > 0xf)) return 0; -- cgit v0.10.2 From cf1d2200dbc214c26a116c4d0c75b7cf27bb19b6 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 15 Oct 2010 15:20:18 +0200 Subject: EDAC, MCE: Add a BIT_64() macro Add a macro for 64-bit vectors to use when accessing MSR contents. Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.h b/drivers/edac/mce_amd.h index 0d0637d..35f6e0e 100644 --- a/drivers/edac/mce_amd.h +++ b/drivers/edac/mce_amd.h @@ -5,6 +5,8 @@ #include +#define BIT_64(n) (U64_C(1) << (n)) + #define ERROR_CODE(x) ((x) & 0xffff) #define EXT_ERROR_CODE(x) (((x) >> 16) & 0x1f) -- cgit v0.10.2 From 525906bc898d712f21e5bfcfc85ab0e517e3d086 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 15 Oct 2010 15:27:02 +0200 Subject: EDAC, MCE: Fix shift warning on 32-bit Fix drivers/edac/mce_amd.c:262: warning: left shift count >= width of type on 32-bit builds. Reported-by: Randy Dunlap Signed-off-by: Borislav Petkov diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 7f74f0f..c018109 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -265,7 +265,7 @@ static void amd_decode_ic_mce(struct mce *m) pr_cont("%s TLB %s.\n", LL_MSG(ec), (xec ? "multimatch" : "parity error")); else if (BUS_ERROR(ec)) { - bool k8 = (boot_cpu_data.x86 == 0xf && (m->status & BIT(58))); + bool k8 = (boot_cpu_data.x86 == 0xf && (m->status & BIT_64(58))); pr_cont("during %s.\n", (k8 ? "system linefill" : "NB data read")); } else if (fam_ops->ic_mce(ec)) -- cgit v0.10.2