From 5fc509bc2bd6dddd4107eaf90680cd76cfc2ffed Mon Sep 17 00:00:00 2001 From: Boris Ostrovsky Date: Wed, 3 Aug 2016 13:22:27 -0400 Subject: xen/x86: Move irq allocation from Xen smp_op.cpu_up() Commit ce0d3c0a6fb1 ("genirq: Revert sparse irq locking around __cpu_up() and move it to x86 for now") reverted irq locking introduced by commit a89941816726 ("hotplug: Prevent alloc/free of irq descriptors during cpu up/down") because of Xen allocating irqs in both of its cpu_up ops. We can move those allocations into CPU notifiers so that original patch can be reinstated. Signed-off-by: Boris Ostrovsky Signed-off-by: David Vrabel diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 8ffb089..c7f6b1f9 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -140,6 +140,8 @@ RESERVE_BRK(shared_info_page_brk, PAGE_SIZE); __read_mostly int xen_have_vector_callback; EXPORT_SYMBOL_GPL(xen_have_vector_callback); +static struct notifier_block xen_cpu_notifier; + /* * Point at some empty memory to start with. We map the real shared_info * page as soon as fixmap is up and running. @@ -1627,6 +1629,7 @@ asmlinkage __visible void __init xen_start_kernel(void) xen_initial_gdt = &per_cpu(gdt_page, 0); xen_smp_init(); + register_cpu_notifier(&xen_cpu_notifier); #ifdef CONFIG_ACPI_NUMA /* @@ -1820,21 +1823,53 @@ static void __init init_hvm_pv_info(void) xen_domain_type = XEN_HVM_DOMAIN; } -static int xen_hvm_cpu_notify(struct notifier_block *self, unsigned long action, - void *hcpu) +static int xen_cpu_notify(struct notifier_block *self, unsigned long action, + void *hcpu) { int cpu = (long)hcpu; + int rc; + switch (action) { case CPU_UP_PREPARE: - if (cpu_acpi_id(cpu) != U32_MAX) - per_cpu(xen_vcpu_id, cpu) = cpu_acpi_id(cpu); - else - per_cpu(xen_vcpu_id, cpu) = cpu; - xen_vcpu_setup(cpu); - if (xen_have_vector_callback) { - if (xen_feature(XENFEAT_hvm_safe_pvclock)) - xen_setup_timer(cpu); + if (xen_hvm_domain()) { + /* + * This can happen if CPU was offlined earlier and + * offlining timed out in common_cpu_die(). + */ + if (cpu_report_state(cpu) == CPU_DEAD_FROZEN) { + xen_smp_intr_free(cpu); + xen_uninit_lock_cpu(cpu); + } + + if (cpu_acpi_id(cpu) != U32_MAX) + per_cpu(xen_vcpu_id, cpu) = cpu_acpi_id(cpu); + else + per_cpu(xen_vcpu_id, cpu) = cpu; + xen_vcpu_setup(cpu); } + + if (xen_pv_domain() || + (xen_have_vector_callback && + xen_feature(XENFEAT_hvm_safe_pvclock))) + xen_setup_timer(cpu); + + rc = xen_smp_intr_init(cpu); + if (rc) { + WARN(1, "xen_smp_intr_init() for CPU %d failed: %d\n", + cpu, rc); + return NOTIFY_BAD; + } + + break; + case CPU_ONLINE: + xen_init_lock_cpu(cpu); + break; + case CPU_UP_CANCELED: + xen_smp_intr_free(cpu); + if (xen_pv_domain() || + (xen_have_vector_callback && + xen_feature(XENFEAT_hvm_safe_pvclock))) + xen_teardown_timer(cpu); break; default: break; @@ -1842,8 +1877,8 @@ static int xen_hvm_cpu_notify(struct notifier_block *self, unsigned long action, return NOTIFY_OK; } -static struct notifier_block xen_hvm_cpu_notifier = { - .notifier_call = xen_hvm_cpu_notify, +static struct notifier_block xen_cpu_notifier = { + .notifier_call = xen_cpu_notify, }; #ifdef CONFIG_KEXEC_CORE @@ -1875,7 +1910,7 @@ static void __init xen_hvm_guest_init(void) if (xen_feature(XENFEAT_hvm_callback_vector)) xen_have_vector_callback = 1; xen_hvm_smp_init(); - register_cpu_notifier(&xen_hvm_cpu_notifier); + register_cpu_notifier(&xen_cpu_notifier); xen_unplug_emulated_devices(); x86_init.irqs.intr_init = xen_init_IRQ; xen_hvm_init_time_ops(); diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index 0b4d04c..137afbb 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -115,7 +115,7 @@ asmlinkage __visible void cpu_bringup_and_idle(int cpu) cpu_startup_entry(CPUHP_AP_ONLINE_IDLE); } -static void xen_smp_intr_free(unsigned int cpu) +void xen_smp_intr_free(unsigned int cpu) { if (per_cpu(xen_resched_irq, cpu).irq >= 0) { unbind_from_irqhandler(per_cpu(xen_resched_irq, cpu).irq, NULL); @@ -159,7 +159,7 @@ static void xen_smp_intr_free(unsigned int cpu) per_cpu(xen_pmu_irq, cpu).name = NULL; } }; -static int xen_smp_intr_init(unsigned int cpu) +int xen_smp_intr_init(unsigned int cpu) { int rc; char *resched_name, *callfunc_name, *debug_name, *pmu_name; @@ -475,8 +475,6 @@ static int xen_cpu_up(unsigned int cpu, struct task_struct *idle) common_cpu_up(cpu, idle); xen_setup_runstate_info(cpu); - xen_setup_timer(cpu); - xen_init_lock_cpu(cpu); /* * PV VCPUs are always successfully taken down (see 'while' loop @@ -495,10 +493,6 @@ static int xen_cpu_up(unsigned int cpu, struct task_struct *idle) xen_pmu_init(cpu); - rc = xen_smp_intr_init(cpu); - if (rc) - return rc; - rc = HYPERVISOR_vcpu_op(VCPUOP_up, xen_vcpu_nr(cpu), NULL); BUG_ON(rc); @@ -769,47 +763,12 @@ static void __init xen_hvm_smp_prepare_cpus(unsigned int max_cpus) xen_init_lock_cpu(0); } -static int xen_hvm_cpu_up(unsigned int cpu, struct task_struct *tidle) -{ - int rc; - - /* - * This can happen if CPU was offlined earlier and - * offlining timed out in common_cpu_die(). - */ - if (cpu_report_state(cpu) == CPU_DEAD_FROZEN) { - xen_smp_intr_free(cpu); - xen_uninit_lock_cpu(cpu); - } - - /* - * xen_smp_intr_init() needs to run before native_cpu_up() - * so that IPI vectors are set up on the booting CPU before - * it is marked online in native_cpu_up(). - */ - rc = xen_smp_intr_init(cpu); - WARN_ON(rc); - if (!rc) - rc = native_cpu_up(cpu, tidle); - - /* - * We must initialize the slowpath CPU kicker _after_ the native - * path has executed. If we initialized it before none of the - * unlocker IPI kicks would reach the booting CPU as the booting - * CPU had not set itself 'online' in cpu_online_mask. That mask - * is checked when IPIs are sent (on HVM at least). - */ - xen_init_lock_cpu(cpu); - return rc; -} - void __init xen_hvm_smp_init(void) { if (!xen_have_vector_callback) return; smp_ops.smp_prepare_cpus = xen_hvm_smp_prepare_cpus; smp_ops.smp_send_reschedule = xen_smp_send_reschedule; - smp_ops.cpu_up = xen_hvm_cpu_up; smp_ops.cpu_die = xen_cpu_die; smp_ops.send_call_func_ipi = xen_smp_send_call_function_ipi; smp_ops.send_call_func_single_ipi = xen_smp_send_call_function_single_ipi; diff --git a/arch/x86/xen/smp.h b/arch/x86/xen/smp.h index 963d62a..c5c16dc 100644 --- a/arch/x86/xen/smp.h +++ b/arch/x86/xen/smp.h @@ -1,5 +1,6 @@ #ifndef _XEN_SMP_H +#ifdef CONFIG_SMP extern void xen_send_IPI_mask(const struct cpumask *mask, int vector); extern void xen_send_IPI_mask_allbutself(const struct cpumask *mask, @@ -8,6 +9,18 @@ extern void xen_send_IPI_allbutself(int vector); extern void xen_send_IPI_all(int vector); extern void xen_send_IPI_self(int vector); +extern int xen_smp_intr_init(unsigned int cpu); +extern void xen_smp_intr_free(unsigned int cpu); + +#else /* CONFIG_SMP */ + +static inline int xen_smp_intr_init(unsigned int cpu) +{ + return 0; +} +static inline void xen_smp_intr_free(unsigned int cpu) {} +#endif /* CONFIG_SMP */ + #ifdef CONFIG_XEN_PVH extern void xen_pvh_early_cpu_init(int cpu, bool entry); #else -- cgit v0.10.2 From e1c105a9d75a1e9a44eb2f0b10204a6492057e4d Mon Sep 17 00:00:00 2001 From: Boris Ostrovsky Date: Wed, 3 Aug 2016 13:22:28 -0400 Subject: hotplug: Prevent alloc/free of irq descriptors during cpu up/down (again) Now that Xen no longer allocates irqs in _cpu_up() we can restore commit a89941816726 ("hotplug: Prevent alloc/free of irq descriptors during cpu up/down") Signed-off-by: Boris Ostrovsky Acked-by: Thomas Gleixner CC: x86@kernel.org CC: Thomas Gleixner Signed-off-by: David Vrabel diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 4296beb..26b473d 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1115,17 +1115,8 @@ int native_cpu_up(unsigned int cpu, struct task_struct *tidle) common_cpu_up(cpu, tidle); - /* - * We have to walk the irq descriptors to setup the vector - * space for the cpu which comes online. Prevent irq - * alloc/free across the bringup. - */ - irq_lock_sparse(); - err = do_boot_cpu(apicid, cpu, tidle); - if (err) { - irq_unlock_sparse(); pr_err("do_boot_cpu failed(%d) to wakeup CPU#%u\n", err, cpu); return -EIO; } @@ -1143,8 +1134,6 @@ int native_cpu_up(unsigned int cpu, struct task_struct *tidle) touch_nmi_watchdog(); } - irq_unlock_sparse(); - return 0; } diff --git a/kernel/cpu.c b/kernel/cpu.c index 341bf80..ec12b72 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -349,8 +349,16 @@ static int bringup_cpu(unsigned int cpu) struct task_struct *idle = idle_thread_get(cpu); int ret; + /* + * Some architectures have to walk the irq descriptors to + * setup the vector space for the cpu which comes online. + * Prevent irq alloc/free across the bringup. + */ + irq_lock_sparse(); + /* Arch-specific enabling code. */ ret = __cpu_up(cpu, idle); + irq_unlock_sparse(); if (ret) { cpu_notify(CPU_UP_CANCELED, cpu); return ret; -- cgit v0.10.2 From 5b00b504b13b2f0d1aa73d59cf8984726f19100f Mon Sep 17 00:00:00 2001 From: Juergen Gross Date: Tue, 2 Aug 2016 08:53:36 +0200 Subject: xen: rename xen_pmu_init() in sys-hypervisor.c There are two functions with name xen_pmu_init() in the kernel. Rename the one in drivers/xen/sys-hypervisor.c to avoid shadowing the one in arch/x86/xen/pmu.c To avoid the same problem in future rename some more functions. Signed-off-by: Juergen Gross Signed-off-by: David Vrabel diff --git a/drivers/xen/sys-hypervisor.c b/drivers/xen/sys-hypervisor.c index 6881b3c..84106f9 100644 --- a/drivers/xen/sys-hypervisor.c +++ b/drivers/xen/sys-hypervisor.c @@ -215,7 +215,7 @@ static const struct attribute_group xen_compilation_group = { .attrs = xen_compile_attrs, }; -static int __init xen_compilation_init(void) +static int __init xen_sysfs_compilation_init(void) { return sysfs_create_group(hypervisor_kobj, &xen_compilation_group); } @@ -341,7 +341,7 @@ static const struct attribute_group xen_properties_group = { .attrs = xen_properties_attrs, }; -static int __init xen_properties_init(void) +static int __init xen_sysfs_properties_init(void) { return sysfs_create_group(hypervisor_kobj, &xen_properties_group); } @@ -455,7 +455,7 @@ static const struct attribute_group xen_pmu_group = { .attrs = xen_pmu_attrs, }; -static int __init xen_pmu_init(void) +static int __init xen_sysfs_pmu_init(void) { return sysfs_create_group(hypervisor_kobj, &xen_pmu_group); } @@ -474,18 +474,18 @@ static int __init hyper_sysfs_init(void) ret = xen_sysfs_version_init(); if (ret) goto version_out; - ret = xen_compilation_init(); + ret = xen_sysfs_compilation_init(); if (ret) goto comp_out; ret = xen_sysfs_uuid_init(); if (ret) goto uuid_out; - ret = xen_properties_init(); + ret = xen_sysfs_properties_init(); if (ret) goto prop_out; #ifdef CONFIG_XEN_HAVE_VPMU if (xen_initial_domain()) { - ret = xen_pmu_init(); + ret = xen_sysfs_pmu_init(); if (ret) { sysfs_remove_group(hypervisor_kobj, &xen_properties_group); -- cgit v0.10.2 From 0252937a87e1d46a8261da85cbd99dffe612a2d3 Mon Sep 17 00:00:00 2001 From: Juergen Gross Date: Tue, 2 Aug 2016 09:22:12 +0200 Subject: xen: Make VPMU init message look less scary The default for the Xen hypervisor is to not enable VPMU in order to avoid security issues. In this case the Linux kernel will issue the message "Could not initialize VPMU for cpu 0, error -95" which looks more like an error than a normal state. Change the message to something less scary in case the hypervisor returns EOPNOTSUPP or ENOSYS when trying to activate VPMU. Signed-off-by: Juergen Gross Signed-off-by: David Vrabel diff --git a/arch/x86/xen/pmu.c b/arch/x86/xen/pmu.c index 32bdc2c..b9fc525 100644 --- a/arch/x86/xen/pmu.c +++ b/arch/x86/xen/pmu.c @@ -547,8 +547,11 @@ void xen_pmu_init(int cpu) return; fail: - pr_info_once("Could not initialize VPMU for cpu %d, error %d\n", - cpu, err); + if (err == -EOPNOTSUPP || err == -ENOSYS) + pr_info_once("VPMU disabled by hypervisor.\n"); + else + pr_info_once("Could not initialize VPMU for cpu %d, error %d\n", + cpu, err); free_pages((unsigned long)xenpmu_data, 0); } -- cgit v0.10.2 From 4f0fbdf22e739c94ad4c18c790be014dddaedd28 Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Thu, 25 Aug 2016 13:23:06 +0200 Subject: xen/grant-table: Use kmalloc_array() in arch_gnttab_valloc() * A multiplication for the size determination of a memory allocation indicated that an array data structure should be processed. Thus reuse the corresponding function "kmalloc_array". This issue was detected by using the Coccinelle software. * Replace the specification of a data type by a pointer dereference to make the corresponding size determination a bit safer according to the Linux coding style convention. Signed-off-by: Markus Elfring Reviewed-by: Juergen Gross Signed-off-by: David Vrabel diff --git a/arch/x86/xen/grant-table.c b/arch/x86/xen/grant-table.c index de4144c..809b6c8 100644 --- a/arch/x86/xen/grant-table.c +++ b/arch/x86/xen/grant-table.c @@ -89,7 +89,7 @@ void arch_gnttab_unmap(void *shared, unsigned long nr_gframes) static int arch_gnttab_valloc(struct gnttab_vm_area *area, unsigned nr_frames) { - area->ptes = kmalloc(sizeof(pte_t *) * nr_frames, GFP_KERNEL); + area->ptes = kmalloc_array(nr_frames, sizeof(*area->ptes), GFP_KERNEL); if (area->ptes == NULL) return -ENOMEM; -- cgit v0.10.2 From 8129554c643b0e1a8336d842cce2f3d595aeeed7 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Mon, 12 Sep 2016 11:20:46 +0100 Subject: x86/xen: add missing \n at end of printk warning message The message is missing a \n, add it. Signed-off-by: Colin Ian King Reviewed-by: Juergen Gross Signed-off-by: David Vrabel diff --git a/arch/x86/xen/platform-pci-unplug.c b/arch/x86/xen/platform-pci-unplug.c index d37a0c7..90d1b83 100644 --- a/arch/x86/xen/platform-pci-unplug.c +++ b/arch/x86/xen/platform-pci-unplug.c @@ -61,7 +61,7 @@ static int check_platform_magic(void) } break; default: - printk(KERN_WARNING "Xen Platform PCI: unknown I/O protocol version"); + printk(KERN_WARNING "Xen Platform PCI: unknown I/O protocol version\n"); return XEN_PLATFORM_ERR_PROTOCOL; } -- cgit v0.10.2 From 4d737042d6c4ee10a632cf94b953169d13955a40 Mon Sep 17 00:00:00 2001 From: Boris Ostrovsky Date: Wed, 7 Sep 2016 13:19:00 -0400 Subject: xen/x86: Convert to hotplug state machine Switch to new CPU hotplug infrastructure. Signed-off-by: Boris Ostrovsky Suggested-by: Sebastian Andrzej Siewior Signed-off-by: David Vrabel diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index c7f6b1f9..a9ba14a 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -140,7 +140,9 @@ RESERVE_BRK(shared_info_page_brk, PAGE_SIZE); __read_mostly int xen_have_vector_callback; EXPORT_SYMBOL_GPL(xen_have_vector_callback); -static struct notifier_block xen_cpu_notifier; +static int xen_cpu_up_prepare(unsigned int cpu); +static int xen_cpu_up_online(unsigned int cpu); +static int xen_cpu_dead(unsigned int cpu); /* * Point at some empty memory to start with. We map the real shared_info @@ -1541,6 +1543,24 @@ static void __init xen_dom0_set_legacy_features(void) x86_platform.legacy.rtc = 1; } +static int xen_cpuhp_setup(void) +{ + int rc; + + rc = cpuhp_setup_state_nocalls(CPUHP_XEN_PREPARE, + "XEN_HVM_GUEST_PREPARE", + xen_cpu_up_prepare, xen_cpu_dead); + if (rc >= 0) { + rc = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, + "XEN_HVM_GUEST_ONLINE", + xen_cpu_up_online, NULL); + if (rc < 0) + cpuhp_remove_state_nocalls(CPUHP_XEN_PREPARE); + } + + return rc >= 0 ? 0 : rc; +} + /* First C function to be called on Xen boot */ asmlinkage __visible void __init xen_start_kernel(void) { @@ -1629,7 +1649,7 @@ asmlinkage __visible void __init xen_start_kernel(void) xen_initial_gdt = &per_cpu(gdt_page, 0); xen_smp_init(); - register_cpu_notifier(&xen_cpu_notifier); + WARN_ON(xen_cpuhp_setup()); #ifdef CONFIG_ACPI_NUMA /* @@ -1823,63 +1843,58 @@ static void __init init_hvm_pv_info(void) xen_domain_type = XEN_HVM_DOMAIN; } -static int xen_cpu_notify(struct notifier_block *self, unsigned long action, - void *hcpu) +static int xen_cpu_up_prepare(unsigned int cpu) { - int cpu = (long)hcpu; int rc; - switch (action) { - case CPU_UP_PREPARE: - if (xen_hvm_domain()) { - /* - * This can happen if CPU was offlined earlier and - * offlining timed out in common_cpu_die(). - */ - if (cpu_report_state(cpu) == CPU_DEAD_FROZEN) { - xen_smp_intr_free(cpu); - xen_uninit_lock_cpu(cpu); - } - - if (cpu_acpi_id(cpu) != U32_MAX) - per_cpu(xen_vcpu_id, cpu) = cpu_acpi_id(cpu); - else - per_cpu(xen_vcpu_id, cpu) = cpu; - xen_vcpu_setup(cpu); + if (xen_hvm_domain()) { + /* + * This can happen if CPU was offlined earlier and + * offlining timed out in common_cpu_die(). + */ + if (cpu_report_state(cpu) == CPU_DEAD_FROZEN) { + xen_smp_intr_free(cpu); + xen_uninit_lock_cpu(cpu); } - if (xen_pv_domain() || - (xen_have_vector_callback && - xen_feature(XENFEAT_hvm_safe_pvclock))) - xen_setup_timer(cpu); + if (cpu_acpi_id(cpu) != U32_MAX) + per_cpu(xen_vcpu_id, cpu) = cpu_acpi_id(cpu); + else + per_cpu(xen_vcpu_id, cpu) = cpu; + xen_vcpu_setup(cpu); + } - rc = xen_smp_intr_init(cpu); - if (rc) { - WARN(1, "xen_smp_intr_init() for CPU %d failed: %d\n", - cpu, rc); - return NOTIFY_BAD; - } + if (xen_pv_domain() || + (xen_have_vector_callback && + xen_feature(XENFEAT_hvm_safe_pvclock))) + xen_setup_timer(cpu); - break; - case CPU_ONLINE: - xen_init_lock_cpu(cpu); - break; - case CPU_UP_CANCELED: - xen_smp_intr_free(cpu); - if (xen_pv_domain() || - (xen_have_vector_callback && - xen_feature(XENFEAT_hvm_safe_pvclock))) - xen_teardown_timer(cpu); - break; - default: - break; + rc = xen_smp_intr_init(cpu); + if (rc) { + WARN(1, "xen_smp_intr_init() for CPU %d failed: %d\n", + cpu, rc); + return rc; } - return NOTIFY_OK; + return 0; } -static struct notifier_block xen_cpu_notifier = { - .notifier_call = xen_cpu_notify, -}; +static int xen_cpu_dead(unsigned int cpu) +{ + xen_smp_intr_free(cpu); + + if (xen_pv_domain() || + (xen_have_vector_callback && + xen_feature(XENFEAT_hvm_safe_pvclock))) + xen_teardown_timer(cpu); + + return 0; +} + +static int xen_cpu_up_online(unsigned int cpu) +{ + xen_init_lock_cpu(cpu); + return 0; +} #ifdef CONFIG_KEXEC_CORE static void xen_hvm_shutdown(void) @@ -1910,7 +1925,7 @@ static void __init xen_hvm_guest_init(void) if (xen_feature(XENFEAT_hvm_callback_vector)) xen_have_vector_callback = 1; xen_hvm_smp_init(); - register_cpu_notifier(&xen_cpu_notifier); + WARN_ON(xen_cpuhp_setup()); xen_unplug_emulated_devices(); x86_init.irqs.intr_init = xen_init_IRQ; xen_hvm_init_time_ops(); diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h index 242bf53..33d352f 100644 --- a/include/linux/cpuhotplug.h +++ b/include/linux/cpuhotplug.h @@ -21,6 +21,7 @@ enum cpuhp_state { CPUHP_X2APIC_PREPARE, CPUHP_SMPCFD_PREPARE, CPUHP_RCUTREE_PREP, + CPUHP_XEN_PREPARE, CPUHP_NOTIFY_PREPARE, CPUHP_TIMERS_DEAD, CPUHP_BRINGUP_CPU, -- cgit v0.10.2 From c8761e2016aa51a2829563b02a0a55913bdb0be8 Mon Sep 17 00:00:00 2001 From: Sebastian Andrzej Siewior Date: Wed, 7 Sep 2016 13:19:01 -0400 Subject: xen/events: Convert to hotplug state machine Install the callbacks via the state machine. Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: Boris Ostrovsky Signed-off-by: David Vrabel diff --git a/drivers/xen/events/events_fifo.c b/drivers/xen/events/events_fifo.c index 266c2c7..7ef27c6 100644 --- a/drivers/xen/events/events_fifo.c +++ b/drivers/xen/events/events_fifo.c @@ -418,30 +418,18 @@ static int evtchn_fifo_alloc_control_block(unsigned cpu) return ret; } -static int evtchn_fifo_cpu_notification(struct notifier_block *self, - unsigned long action, - void *hcpu) +static int xen_evtchn_cpu_prepare(unsigned int cpu) { - int cpu = (long)hcpu; - int ret = 0; - - switch (action) { - case CPU_UP_PREPARE: - if (!per_cpu(cpu_control_block, cpu)) - ret = evtchn_fifo_alloc_control_block(cpu); - break; - case CPU_DEAD: - __evtchn_fifo_handle_events(cpu, true); - break; - default: - break; - } - return ret < 0 ? NOTIFY_BAD : NOTIFY_OK; + if (!per_cpu(cpu_control_block, cpu)) + return evtchn_fifo_alloc_control_block(cpu); + return 0; } -static struct notifier_block evtchn_fifo_cpu_notifier = { - .notifier_call = evtchn_fifo_cpu_notification, -}; +static int xen_evtchn_cpu_dead(unsigned int cpu) +{ + __evtchn_fifo_handle_events(cpu, true); + return 0; +} int __init xen_evtchn_fifo_init(void) { @@ -456,7 +444,9 @@ int __init xen_evtchn_fifo_init(void) evtchn_ops = &evtchn_ops_fifo; - register_cpu_notifier(&evtchn_fifo_cpu_notifier); + cpuhp_setup_state_nocalls(CPUHP_XEN_EVTCHN_PREPARE, + "CPUHP_XEN_EVTCHN_PREPARE", + xen_evtchn_cpu_prepare, xen_evtchn_cpu_dead); out: put_cpu(); return ret; diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h index 33d352f..5f60316 100644 --- a/include/linux/cpuhotplug.h +++ b/include/linux/cpuhotplug.h @@ -22,6 +22,7 @@ enum cpuhp_state { CPUHP_SMPCFD_PREPARE, CPUHP_RCUTREE_PREP, CPUHP_XEN_PREPARE, + CPUHP_XEN_EVTCHN_PREPARE, CPUHP_NOTIFY_PREPARE, CPUHP_TIMERS_DEAD, CPUHP_BRINGUP_CPU, -- cgit v0.10.2 From 72a9b186292d98494f222226cfd24a1621796209 Mon Sep 17 00:00:00 2001 From: KarimAllah Ahmed Date: Fri, 26 Aug 2016 23:55:36 +0200 Subject: xen: Remove event channel notification through Xen PCI platform device Ever since commit 254d1a3f02eb ("xen/pv-on-hvm kexec: shutdown watches from old kernel") using the INTx interrupt from Xen PCI platform device for event channel notification would just lockup the guest during bootup. postcore_initcall now calls xs_reset_watches which will eventually try to read a value from XenStore and will get stuck on read_reply at XenBus forever since the platform driver is not probed yet and its INTx interrupt handler is not registered yet. That means that the guest can not be notified at this moment of any pending event channels and none of the per-event handlers will ever be invoked (including the XenStore one) and the reply will never be picked up by the kernel. The exact stack where things get stuck during xenbus_init: -xenbus_init -xs_init -xs_reset_watches -xenbus_scanf -xenbus_read -xs_single -xs_single -xs_talkv Vector callbacks have always been the favourite event notification mechanism since their introduction in commit 38e20b07efd5 ("x86/xen: event channels delivery on HVM.") and the vector callback feature has always been advertised for quite some time by Xen that's why INTx was broken for several years now without impacting anyone. Luckily this also means that event channel notification through INTx is basically dead-code which can be safely removed without impacting anybody since it has been effectively disabled for more than 4 years with nobody complaining about it (at least as far as I'm aware of). This commit removes event channel notification through Xen PCI platform device. Cc: Boris Ostrovsky Cc: David Vrabel Cc: Juergen Gross Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: x86@kernel.org Cc: Konrad Rzeszutek Wilk Cc: Bjorn Helgaas Cc: Stefano Stabellini Cc: Julien Grall Cc: Vitaly Kuznetsov Cc: Paul Gortmaker Cc: Ross Lagerwall Cc: xen-devel@lists.xenproject.org Cc: linux-kernel@vger.kernel.org Cc: linux-pci@vger.kernel.org Cc: Anthony Liguori Signed-off-by: KarimAllah Ahmed Reviewed-by: Boris Ostrovsky Signed-off-by: David Vrabel diff --git a/arch/x86/include/asm/xen/events.h b/arch/x86/include/asm/xen/events.h index e6911ca..608a79d 100644 --- a/arch/x86/include/asm/xen/events.h +++ b/arch/x86/include/asm/xen/events.h @@ -20,15 +20,4 @@ static inline int xen_irqs_disabled(struct pt_regs *regs) /* No need for a barrier -- XCHG is a barrier on x86. */ #define xchg_xen_ulong(ptr, val) xchg((ptr), (val)) -extern int xen_have_vector_callback; - -/* - * Events delivered via platform PCI interrupts are always - * routed to vcpu 0 and hence cannot be rebound. - */ -static inline bool xen_support_evtchn_rebind(void) -{ - return (!xen_hvm_domain() || xen_have_vector_callback); -} - #endif /* _ASM_X86_XEN_EVENTS_H */ diff --git a/arch/x86/pci/xen.c b/arch/x86/pci/xen.c index 3a483cb..bedfab9 100644 --- a/arch/x86/pci/xen.c +++ b/arch/x86/pci/xen.c @@ -456,7 +456,7 @@ void __init xen_msi_init(void) int __init pci_xen_hvm_init(void) { - if (!xen_have_vector_callback || !xen_feature(XENFEAT_hvm_pirqs)) + if (!xen_feature(XENFEAT_hvm_pirqs)) return 0; #ifdef CONFIG_ACPI diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index a9ba14a..366b6ae 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -137,8 +137,6 @@ struct shared_info xen_dummy_shared_info; void *xen_initial_gdt; RESERVE_BRK(shared_info_page_brk, PAGE_SIZE); -__read_mostly int xen_have_vector_callback; -EXPORT_SYMBOL_GPL(xen_have_vector_callback); static int xen_cpu_up_prepare(unsigned int cpu); static int xen_cpu_up_online(unsigned int cpu); @@ -1524,10 +1522,7 @@ static void __init xen_pvh_early_guest_init(void) if (!xen_feature(XENFEAT_auto_translated_physmap)) return; - if (!xen_feature(XENFEAT_hvm_callback_vector)) - return; - - xen_have_vector_callback = 1; + BUG_ON(!xen_feature(XENFEAT_hvm_callback_vector)); xen_pvh_early_cpu_init(0, false); xen_pvh_set_cr_flags(0); @@ -1864,9 +1859,7 @@ static int xen_cpu_up_prepare(unsigned int cpu) xen_vcpu_setup(cpu); } - if (xen_pv_domain() || - (xen_have_vector_callback && - xen_feature(XENFEAT_hvm_safe_pvclock))) + if (xen_pv_domain() || xen_feature(XENFEAT_hvm_safe_pvclock)) xen_setup_timer(cpu); rc = xen_smp_intr_init(cpu); @@ -1882,9 +1875,7 @@ static int xen_cpu_dead(unsigned int cpu) { xen_smp_intr_free(cpu); - if (xen_pv_domain() || - (xen_have_vector_callback && - xen_feature(XENFEAT_hvm_safe_pvclock))) + if (xen_pv_domain() || xen_feature(XENFEAT_hvm_safe_pvclock)) xen_teardown_timer(cpu); return 0; @@ -1922,8 +1913,8 @@ static void __init xen_hvm_guest_init(void) xen_panic_handler_init(); - if (xen_feature(XENFEAT_hvm_callback_vector)) - xen_have_vector_callback = 1; + BUG_ON(!xen_feature(XENFEAT_hvm_callback_vector)); + xen_hvm_smp_init(); WARN_ON(xen_cpuhp_setup()); xen_unplug_emulated_devices(); @@ -1961,7 +1952,7 @@ bool xen_hvm_need_lapic(void) return false; if (!xen_hvm_domain()) return false; - if (xen_feature(XENFEAT_hvm_pirqs) && xen_have_vector_callback) + if (xen_feature(XENFEAT_hvm_pirqs)) return false; return true; } diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index 137afbb..311acad 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -765,8 +765,6 @@ static void __init xen_hvm_smp_prepare_cpus(unsigned int max_cpus) void __init xen_hvm_smp_init(void) { - if (!xen_have_vector_callback) - return; smp_ops.smp_prepare_cpus = xen_hvm_smp_prepare_cpus; smp_ops.smp_send_reschedule = xen_smp_send_reschedule; smp_ops.cpu_die = xen_cpu_die; diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index 67356d2..33d8f6a 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -432,11 +432,6 @@ static void xen_hvm_setup_cpu_clockevents(void) void __init xen_hvm_init_time_ops(void) { - /* vector callback is needed otherwise we cannot receive interrupts - * on cpu > 0 and at this point we don't know how many cpus are - * available */ - if (!xen_have_vector_callback) - return; if (!xen_feature(XENFEAT_hvm_safe_pvclock)) { printk(KERN_INFO "Xen doesn't support pvclock on HVM," "disable pv timer\n"); diff --git a/drivers/xen/events/events_base.c b/drivers/xen/events/events_base.c index d5dbdb9..9ecfcdc 100644 --- a/drivers/xen/events/events_base.c +++ b/drivers/xen/events/events_base.c @@ -1314,9 +1314,6 @@ static int rebind_irq_to_cpu(unsigned irq, unsigned tcpu) if (!VALID_EVTCHN(evtchn)) return -1; - if (!xen_support_evtchn_rebind()) - return -1; - /* Send future instances of this interrupt to other vcpu. */ bind_vcpu.port = evtchn; bind_vcpu.vcpu = xen_vcpu_nr(tcpu); @@ -1650,20 +1647,15 @@ void xen_callback_vector(void) { int rc; uint64_t callback_via; - if (xen_have_vector_callback) { - callback_via = HVM_CALLBACK_VECTOR(HYPERVISOR_CALLBACK_VECTOR); - rc = xen_set_callback_via(callback_via); - if (rc) { - pr_err("Request for Xen HVM callback vector failed\n"); - xen_have_vector_callback = 0; - return; - } - pr_info("Xen HVM callback vector for event delivery is enabled\n"); - /* in the restore case the vector has already been allocated */ - if (!test_bit(HYPERVISOR_CALLBACK_VECTOR, used_vectors)) - alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR, - xen_hvm_callback_vector); - } + + callback_via = HVM_CALLBACK_VECTOR(HYPERVISOR_CALLBACK_VECTOR); + rc = xen_set_callback_via(callback_via); + BUG_ON(rc); + pr_info("Xen HVM callback vector for event delivery is enabled\n"); + /* in the restore case the vector has already been allocated */ + if (!test_bit(HYPERVISOR_CALLBACK_VECTOR, used_vectors)) + alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR, + xen_hvm_callback_vector); } #else void xen_callback_vector(void) {} diff --git a/drivers/xen/platform-pci.c b/drivers/xen/platform-pci.c index cf96666..b59c9455 100644 --- a/drivers/xen/platform-pci.c +++ b/drivers/xen/platform-pci.c @@ -42,7 +42,6 @@ static unsigned long platform_mmio; static unsigned long platform_mmio_alloc; static unsigned long platform_mmiolen; -static uint64_t callback_via; static unsigned long alloc_xen_mmio(unsigned long len) { @@ -55,51 +54,6 @@ static unsigned long alloc_xen_mmio(unsigned long len) return addr; } -static uint64_t get_callback_via(struct pci_dev *pdev) -{ - u8 pin; - int irq; - - irq = pdev->irq; - if (irq < 16) - return irq; /* ISA IRQ */ - - pin = pdev->pin; - - /* We don't know the GSI. Specify the PCI INTx line instead. */ - return ((uint64_t)0x01 << 56) | /* PCI INTx identifier */ - ((uint64_t)pci_domain_nr(pdev->bus) << 32) | - ((uint64_t)pdev->bus->number << 16) | - ((uint64_t)(pdev->devfn & 0xff) << 8) | - ((uint64_t)(pin - 1) & 3); -} - -static irqreturn_t do_hvm_evtchn_intr(int irq, void *dev_id) -{ - xen_hvm_evtchn_do_upcall(); - return IRQ_HANDLED; -} - -static int xen_allocate_irq(struct pci_dev *pdev) -{ - return request_irq(pdev->irq, do_hvm_evtchn_intr, - IRQF_NOBALANCING | IRQF_TRIGGER_RISING, - "xen-platform-pci", pdev); -} - -static int platform_pci_resume(struct pci_dev *pdev) -{ - int err; - if (xen_have_vector_callback) - return 0; - err = xen_set_callback_via(callback_via); - if (err) { - dev_err(&pdev->dev, "platform_pci_resume failure!\n"); - return err; - } - return 0; -} - static int platform_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) { @@ -138,21 +92,6 @@ static int platform_pci_probe(struct pci_dev *pdev, platform_mmio = mmio_addr; platform_mmiolen = mmio_len; - if (!xen_have_vector_callback) { - ret = xen_allocate_irq(pdev); - if (ret) { - dev_warn(&pdev->dev, "request_irq failed err=%d\n", ret); - goto out; - } - callback_via = get_callback_via(pdev); - ret = xen_set_callback_via(callback_via); - if (ret) { - dev_warn(&pdev->dev, "Unable to set the evtchn callback " - "err=%d\n", ret); - goto out; - } - } - max_nr_gframes = gnttab_max_grant_frames(); grant_frames = alloc_xen_mmio(PAGE_SIZE * max_nr_gframes); ret = gnttab_setup_auto_xlat_frames(grant_frames); @@ -184,9 +123,6 @@ static struct pci_driver platform_driver = { .name = DRV_NAME, .probe = platform_pci_probe, .id_table = platform_pci_tbl, -#ifdef CONFIG_PM - .resume_early = platform_pci_resume, -#endif }; static int __init platform_pci_init(void) diff --git a/include/xen/xen.h b/include/xen/xen.h index 0c0e3ef..f0f0252 100644 --- a/include/xen/xen.h +++ b/include/xen/xen.h @@ -38,8 +38,7 @@ extern enum xen_domain_type xen_domain_type; */ #include #define xen_pvh_domain() (xen_pv_domain() && \ - xen_feature(XENFEAT_auto_translated_physmap) && \ - xen_have_vector_callback) + xen_feature(XENFEAT_auto_translated_physmap)) #else #define xen_pvh_domain() (0) #endif -- cgit v0.10.2 From 1af916b701db1a9905e559e742f45818eb233d12 Mon Sep 17 00:00:00 2001 From: Juergen Gross Date: Thu, 22 Sep 2016 10:45:39 +0200 Subject: xen/pciback: simplify pcistub device handling The Xen pciback driver maintains a list of all its seized devices. There are two functions searching the list for a specific device with basically the same semantics just returning different structures in case of a match. Split out the search function. Signed-off-by: Juergen Gross Signed-off-by: David Vrabel diff --git a/drivers/xen/xen-pciback/pci_stub.c b/drivers/xen/xen-pciback/pci_stub.c index 258b7c3..79a9e4d 100644 --- a/drivers/xen/xen-pciback/pci_stub.c +++ b/drivers/xen/xen-pciback/pci_stub.c @@ -149,13 +149,10 @@ static inline void pcistub_device_put(struct pcistub_device *psdev) kref_put(&psdev->kref, pcistub_device_release); } -static struct pcistub_device *pcistub_device_find(int domain, int bus, - int slot, int func) +static struct pcistub_device *pcistub_device_find_locked(int domain, int bus, + int slot, int func) { - struct pcistub_device *psdev = NULL; - unsigned long flags; - - spin_lock_irqsave(&pcistub_devices_lock, flags); + struct pcistub_device *psdev; list_for_each_entry(psdev, &pcistub_devices, dev_list) { if (psdev->dev != NULL @@ -163,15 +160,25 @@ static struct pcistub_device *pcistub_device_find(int domain, int bus, && bus == psdev->dev->bus->number && slot == PCI_SLOT(psdev->dev->devfn) && func == PCI_FUNC(psdev->dev->devfn)) { - pcistub_device_get(psdev); - goto out; + return psdev; } } - /* didn't find it */ - psdev = NULL; + return NULL; +} + +static struct pcistub_device *pcistub_device_find(int domain, int bus, + int slot, int func) +{ + struct pcistub_device *psdev; + unsigned long flags; + + spin_lock_irqsave(&pcistub_devices_lock, flags); + + psdev = pcistub_device_find_locked(domain, bus, slot, func); + if (psdev) + pcistub_device_get(psdev); -out: spin_unlock_irqrestore(&pcistub_devices_lock, flags); return psdev; } @@ -207,16 +214,9 @@ struct pci_dev *pcistub_get_pci_dev_by_slot(struct xen_pcibk_device *pdev, spin_lock_irqsave(&pcistub_devices_lock, flags); - list_for_each_entry(psdev, &pcistub_devices, dev_list) { - if (psdev->dev != NULL - && domain == pci_domain_nr(psdev->dev->bus) - && bus == psdev->dev->bus->number - && slot == PCI_SLOT(psdev->dev->devfn) - && func == PCI_FUNC(psdev->dev->devfn)) { - found_dev = pcistub_device_get_pci_dev(pdev, psdev); - break; - } - } + psdev = pcistub_device_find_locked(domain, bus, slot, func); + if (psdev) + found_dev = pcistub_device_get_pci_dev(pdev, psdev); spin_unlock_irqrestore(&pcistub_devices_lock, flags); return found_dev; -- cgit v0.10.2 From 9f8bee9c981f5fe7382a0615d117cc128dd22458 Mon Sep 17 00:00:00 2001 From: Juergen Gross Date: Thu, 22 Sep 2016 10:45:40 +0200 Subject: xen/pciback: avoid multiple entries in slot list The Xen pciback driver has a list of all pci devices it is ready to seize. There is no check whether a to be added entry already exists. While this might be no problem in the common case it might confuse those which consume the list via sysfs. Modify the handling of this list by not adding an entry which already exists. As this will be needed later split out the list handling into a separate function. Signed-off-by: Juergen Gross Signed-off-by: David Vrabel diff --git a/drivers/xen/xen-pciback/pci_stub.c b/drivers/xen/xen-pciback/pci_stub.c index 79a9e4d..0179333 100644 --- a/drivers/xen/xen-pciback/pci_stub.c +++ b/drivers/xen/xen-pciback/pci_stub.c @@ -478,6 +478,36 @@ static int __init pcistub_init_devices_late(void) return 0; } +static void pcistub_device_id_add_list(struct pcistub_device_id *new, + int domain, int bus, unsigned int devfn) +{ + struct pcistub_device_id *pci_dev_id; + unsigned long flags; + int found = 0; + + spin_lock_irqsave(&device_ids_lock, flags); + + list_for_each_entry(pci_dev_id, &pcistub_device_ids, slot_list) { + if (pci_dev_id->domain == domain && pci_dev_id->bus == bus && + pci_dev_id->devfn == devfn) { + found = 1; + break; + } + } + + if (!found) { + new->domain = domain; + new->bus = bus; + new->devfn = devfn; + list_add_tail(&new->slot_list, &pcistub_device_ids); + } + + spin_unlock_irqrestore(&device_ids_lock, flags); + + if (found) + kfree(new); +} + static int pcistub_seize(struct pci_dev *dev) { struct pcistub_device *psdev; @@ -1012,7 +1042,6 @@ static inline int str_to_quirk(const char *buf, int *domain, int *bus, int static int pcistub_device_id_add(int domain, int bus, int slot, int func) { struct pcistub_device_id *pci_dev_id; - unsigned long flags; int rc = 0, devfn = PCI_DEVFN(slot, func); if (slot < 0) { @@ -1042,16 +1071,10 @@ static int pcistub_device_id_add(int domain, int bus, int slot, int func) if (!pci_dev_id) return -ENOMEM; - pci_dev_id->domain = domain; - pci_dev_id->bus = bus; - pci_dev_id->devfn = devfn; - pr_debug("wants to seize %04x:%02x:%02x.%d\n", domain, bus, slot, func); - spin_lock_irqsave(&device_ids_lock, flags); - list_add_tail(&pci_dev_id->slot_list, &pcistub_device_ids); - spin_unlock_irqrestore(&device_ids_lock, flags); + pcistub_device_id_add_list(pci_dev_id, domain, bus, devfn); return 0; } -- cgit v0.10.2 From b057878b2aadc7e06280e7e702a36e7adb1bcdf7 Mon Sep 17 00:00:00 2001 From: Juergen Gross Date: Thu, 22 Sep 2016 10:45:41 +0200 Subject: xen/pciback: support driver_override Support the driver_override scheme introduced with commit 782a985d7af2 ("PCI: Introduce new device binding path using pci_dev.driver_override") As pcistub_probe() is called for all devices (it has to check for a match based on the slot address rather than device type) it has to check for driver_override set to "pciback" itself. Up to now for assigning a pci device to pciback you need something like: echo 0000:07:10.0 > /sys/bus/pci/devices/0000\:07\:10.0/driver/unbind echo 0000:07:10.0 > /sys/bus/pci/drivers/pciback/new_slot echo 0000:07:10.0 > /sys/bus/pci/drivers_probe while with the patch you can use the same mechanism as for similar drivers like pci-stub and vfio-pci: echo pciback > /sys/bus/pci/devices/0000\:07\:10.0/driver_override echo 0000:07:10.0 > /sys/bus/pci/devices/0000\:07\:10.0/driver/unbind echo 0000:07:10.0 > /sys/bus/pci/drivers_probe So e.g. libvirt doesn't need special handling for pciback. Signed-off-by: Juergen Gross Signed-off-by: David Vrabel diff --git a/drivers/xen/xen-pciback/pci_stub.c b/drivers/xen/xen-pciback/pci_stub.c index 0179333..6331a95 100644 --- a/drivers/xen/xen-pciback/pci_stub.c +++ b/drivers/xen/xen-pciback/pci_stub.c @@ -25,6 +25,8 @@ #include "conf_space.h" #include "conf_space_quirks.h" +#define PCISTUB_DRIVER_NAME "pciback" + static char *pci_devs_to_hide; wait_queue_head_t xen_pcibk_aer_wait_queue; /*Add sem for sync AER handling and xen_pcibk remove/reconfigue ops, @@ -508,15 +510,18 @@ static void pcistub_device_id_add_list(struct pcistub_device_id *new, kfree(new); } -static int pcistub_seize(struct pci_dev *dev) +static int pcistub_seize(struct pci_dev *dev, + struct pcistub_device_id *pci_dev_id) { struct pcistub_device *psdev; unsigned long flags; int err = 0; psdev = pcistub_device_alloc(dev); - if (!psdev) + if (!psdev) { + kfree(pci_dev_id); return -ENOMEM; + } spin_lock_irqsave(&pcistub_devices_lock, flags); @@ -537,8 +542,12 @@ static int pcistub_seize(struct pci_dev *dev) spin_unlock_irqrestore(&pcistub_devices_lock, flags); - if (err) + if (err) { + kfree(pci_dev_id); pcistub_device_put(psdev); + } else if (pci_dev_id) + pcistub_device_id_add_list(pci_dev_id, pci_domain_nr(dev->bus), + dev->bus->number, dev->devfn); return err; } @@ -547,11 +556,16 @@ static int pcistub_seize(struct pci_dev *dev) * other functions that take the sysfs lock. */ static int pcistub_probe(struct pci_dev *dev, const struct pci_device_id *id) { - int err = 0; + int err = 0, match; + struct pcistub_device_id *pci_dev_id = NULL; dev_dbg(&dev->dev, "probing...\n"); - if (pcistub_match(dev)) { + match = pcistub_match(dev); + + if ((dev->driver_override && + !strcmp(dev->driver_override, PCISTUB_DRIVER_NAME)) || + match) { if (dev->hdr_type != PCI_HEADER_TYPE_NORMAL && dev->hdr_type != PCI_HEADER_TYPE_BRIDGE) { @@ -562,8 +576,16 @@ static int pcistub_probe(struct pci_dev *dev, const struct pci_device_id *id) goto out; } + if (!match) { + pci_dev_id = kmalloc(sizeof(*pci_dev_id), GFP_ATOMIC); + if (!pci_dev_id) { + err = -ENOMEM; + goto out; + } + } + dev_info(&dev->dev, "seizing device\n"); - err = pcistub_seize(dev); + err = pcistub_seize(dev, pci_dev_id); } else /* Didn't find the device */ err = -ENODEV; @@ -975,7 +997,7 @@ static const struct pci_error_handlers xen_pcibk_error_handler = { static struct pci_driver xen_pcibk_pci_driver = { /* The name should be xen_pciback, but until the tools are updated * we will keep it as pciback. */ - .name = "pciback", + .name = PCISTUB_DRIVER_NAME, .id_table = pcistub_ids, .probe = pcistub_probe, .remove = pcistub_remove, -- cgit v0.10.2 From 565fdc6a2afeb14e654a8d9ebc792afb9d2c0019 Mon Sep 17 00:00:00 2001 From: Boris Ostrovsky Date: Sun, 2 Oct 2016 18:45:19 -0400 Subject: xen/x86: Initialize per_cpu(xen_vcpu, 0) a little earlier xen_cpuhp_setup() calls mutex_lock() which, when CONFIG_DEBUG_MUTEXES is defined, ends up calling xen_save_fl(). That routine expects per_cpu(xen_vcpu, 0) to be already initialized. Signed-off-by: Boris Ostrovsky Reported-by: Sander Eikelenboom Signed-off-by: David Vrabel diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 366b6ae..96c2dea 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -1644,7 +1644,6 @@ asmlinkage __visible void __init xen_start_kernel(void) xen_initial_gdt = &per_cpu(gdt_page, 0); xen_smp_init(); - WARN_ON(xen_cpuhp_setup()); #ifdef CONFIG_ACPI_NUMA /* @@ -1658,6 +1657,8 @@ asmlinkage __visible void __init xen_start_kernel(void) possible map and a non-dummy shared_info. */ per_cpu(xen_vcpu, 0) = &HYPERVISOR_shared_info->vcpu_info[0]; + WARN_ON(xen_cpuhp_setup()); + local_irq_disable(); early_boot_irqs_disabled = true; -- cgit v0.10.2 From a6a198bc60e6c980a56eca24d33dc7f29139f8ea Mon Sep 17 00:00:00 2001 From: Boris Ostrovsky Date: Wed, 5 Oct 2016 13:09:33 -0400 Subject: xen/x86: Update topology map for PV VCPUs Early during boot topology_update_package_map() computes logical_pkg_ids for all present processors. Later, when processors are brought up, identify_cpu() updates these values based on phys_pkg_id which is a function of initial_apicid. On PV guests the latter may point to a non-existing node, causing logical_pkg_ids to be set to -1. Intel's RAPL uses logical_pkg_id (as topology_logical_package_id()) to index its arrays and therefore in this case will point to index 65535 (since logical_pkg_id is a u16). This could lead to either a crash or may actually access random memory location. As a workaround, we recompute topology during CPU bringup to reset logical_pkg_id to a valid value. (The reason for initial_apicid being bogus is because it is initial_apicid of the processor from which the guest is launched. This value is CPUID(1).EBX[31:24]) Signed-off-by: Boris Ostrovsky Cc: stable@vger.kernel.org Signed-off-by: David Vrabel diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index 311acad..9fa27ce 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -87,6 +87,12 @@ static void cpu_bringup(void) cpu_data(cpu).x86_max_cores = 1; set_cpu_sibling_map(cpu); + /* + * identify_cpu() may have set logical_pkg_id to -1 due + * to incorrect phys_proc_id. Let's re-comupte it. + */ + topology_update_package_map(apic->cpu_present_to_apicid(cpu), cpu); + xen_setup_cpu_clockevents(); notify_cpu_starting(cpu); -- cgit v0.10.2