From 5dd8c4c3f18b2ebf3f5303c7c273396441db6c6c Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Fri, 8 Nov 2013 10:20:06 -0800 Subject: drm/i915/bdw: Add BDW to ULT macro For what we care about ULT and ULX are interchangeable. We know of 3 types of pciids for these cases. I am not sure if at some point we will need to distinguish ULT and ULX. Cc: Paulo Zanoni Signed-off-by: Ben Widawsky Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 8600c31..51951ef 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1755,8 +1755,13 @@ struct drm_i915_file_private { #define IS_MOBILE(dev) (INTEL_INFO(dev)->is_mobile) #define IS_HSW_EARLY_SDV(dev) (IS_HASWELL(dev) && \ ((dev)->pdev->device & 0xFF00) == 0x0C00) -#define IS_ULT(dev) (IS_HASWELL(dev) && \ +#define IS_BDW_ULT(dev) (IS_BROADWELL(dev) && \ + (((dev)->pdev->device & 0xf) == 0x2 || \ + ((dev)->pdev->device & 0xf) == 0x6 || \ + ((dev)->pdev->device & 0xf) == 0xe)) +#define IS_HSW_ULT(dev) (IS_HASWELL(dev) && \ ((dev)->pdev->device & 0xFF00) == 0x0A00) +#define IS_ULT(dev) (IS_HSW_ULT(dev) || IS_BDW_ULT(dev)) #define IS_HSW_GT3(dev) (IS_HASWELL(dev) && \ ((dev)->pdev->device & 0x00F0) == 0x0020) #define IS_PRELIMINARY_HW(intel_info) ((intel_info)->is_preliminary) -- cgit v0.10.2 From f8e100621b072384ec1180563d933e86e2f7e2d5 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Mon, 11 Nov 2013 11:12:57 +0200 Subject: drm/i915/bdw: GEN8 backlight support Prior to Haswell the CPU control register for backlight (BLC_PWM_CPU_CTL) toggled the PCH baclight pin for us. This made some sense as there was no pin on the CPU. With Haswell came the introduction of a CPU backlight pin, but the interface was still controlled by software with the same mechnism. Behind the scenes, hardware did all the dirty work for us. Broadwell no longer provides this for free. If we want to use the PCH backlight pin [1] then we have to set the override bit BLC_PWM_PCH_CTL1 and program BLC_PWM_PCH_CTL2 for the PWM values. This patch implements that. This patch is compile tested only, and given that I rarely if ever touch this code, careful review is welcome. [1] According to Art, we know of no devices that exist which use the CPU pin (and remember it has existed already on HSW). If such a device does exist, we'll have to handle it properly - this is left as TODO until then. v2: Drop the abstraction prep patch, as a bigger backlight overhaul is in the works, and do just the mimimal bdw enabling now. (by Jani) CC: Art Runyan Signed-off-by: Ben Widawsky Signed-off-by: Jani Nikula Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c index f161ac0..e6f782d 100644 --- a/drivers/gpu/drm/i915/intel_panel.c +++ b/drivers/gpu/drm/i915/intel_panel.c @@ -451,7 +451,9 @@ static u32 intel_panel_get_backlight(struct drm_device *dev, spin_lock_irqsave(&dev_priv->backlight.lock, flags); - if (HAS_PCH_SPLIT(dev)) { + if (IS_BROADWELL(dev)) { + val = I915_READ(BLC_PWM_PCH_CTL2) & BACKLIGHT_DUTY_CYCLE_MASK; + } else if (HAS_PCH_SPLIT(dev)) { val = I915_READ(BLC_PWM_CPU_CTL) & BACKLIGHT_DUTY_CYCLE_MASK; } else { if (IS_VALLEYVIEW(dev)) @@ -479,6 +481,13 @@ static u32 intel_panel_get_backlight(struct drm_device *dev, return val; } +static void intel_bdw_panel_set_backlight(struct drm_device *dev, u32 level) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + u32 val = I915_READ(BLC_PWM_PCH_CTL2) & ~BACKLIGHT_DUTY_CYCLE_MASK; + I915_WRITE(BLC_PWM_PCH_CTL2, val | level); +} + static void intel_pch_panel_set_backlight(struct drm_device *dev, u32 level) { struct drm_i915_private *dev_priv = dev->dev_private; @@ -496,7 +505,9 @@ static void intel_panel_actually_set_backlight(struct drm_device *dev, DRM_DEBUG_DRIVER("set backlight PWM = %d\n", level); level = intel_panel_compute_brightness(dev, pipe, level); - if (HAS_PCH_SPLIT(dev)) + if (IS_BROADWELL(dev)) + return intel_bdw_panel_set_backlight(dev, level); + else if (HAS_PCH_SPLIT(dev)) return intel_pch_panel_set_backlight(dev, level); if (is_backlight_combination_mode(dev)) { @@ -666,7 +677,16 @@ void intel_panel_enable_backlight(struct intel_connector *connector) POSTING_READ(reg); I915_WRITE(reg, tmp | BLM_PWM_ENABLE); - if (HAS_PCH_SPLIT(dev) && + if (IS_BROADWELL(dev)) { + /* + * Broadwell requires PCH override to drive the PCH + * backlight pin. The above will configure the CPU + * backlight pin, which we don't plan to use. + */ + tmp = I915_READ(BLC_PWM_PCH_CTL1); + tmp |= BLM_PCH_OVERRIDE_ENABLE | BLM_PCH_PWM_ENABLE; + I915_WRITE(BLC_PWM_PCH_CTL1, tmp); + } else if (HAS_PCH_SPLIT(dev) && !(dev_priv->quirks & QUIRK_NO_PCH_PWM_ENABLE)) { tmp = I915_READ(BLC_PWM_PCH_CTL1); tmp |= BLM_PCH_PWM_ENABLE; -- cgit v0.10.2 From 935e8de97564a1ea22c3bb30ca2b20b12af7cbfb Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Thu, 7 Nov 2013 21:40:47 -0800 Subject: drm/i915/bdw: Do gen6 style reset for gen8 This patch existed before, but was lost over time. Note that reset is still somewhat problematic in my limited testing (ie. module_reload will not pass) but it can be disabled with a module parameter, and support should be considered preliminary anyway. Signed-off-by: Ben Widawsky Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index f9883ce..6a4f9b6 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -782,6 +782,7 @@ static int gen6_do_reset(struct drm_device *dev) int intel_gpu_reset(struct drm_device *dev) { switch (INTEL_INFO(dev)->gen) { + case 8: case 7: case 6: return gen6_do_reset(dev); case 5: return ironlake_do_reset(dev); -- cgit v0.10.2 From 230f955f73074ec5b1b4389f3f26152eebf54404 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Thu, 7 Nov 2013 21:40:48 -0800 Subject: drm/i915/bdw: Free correct number of ppgtt pages I am unclear how this got messed up in the shuffle, but it did. Cc: Imre Deak Signed-off-by: Ben Widawsky Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 3620a1b..5a3cc31 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -335,8 +335,8 @@ static void gen8_ppgtt_cleanup(struct i915_address_space *vm) kfree(ppgtt->gen8_pt_dma_addr[i]); } - __free_pages(ppgtt->gen8_pt_pages, ppgtt->num_pt_pages << PAGE_SHIFT); - __free_pages(ppgtt->pd_pages, ppgtt->num_pd_pages << PAGE_SHIFT); + __free_pages(ppgtt->gen8_pt_pages, get_order(ppgtt->num_pt_pages << PAGE_SHIFT)); + __free_pages(ppgtt->pd_pages, get_order(ppgtt->num_pd_pages << PAGE_SHIFT)); } /** -- cgit v0.10.2 From eb0d4b75d506050cb369df467b93b92360f5acaf Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Thu, 7 Nov 2013 21:40:50 -0800 Subject: drm/i915/bdw: Add comment about gen8 HWS PGA This confused me some many times that I think it is appropriate to add a small comment to instruct the reader of the code that it is indeed doing what it is supposed to do. Signed-off-by: Ben Widawsky Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index b620337..c2f09d45 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -965,6 +965,7 @@ void intel_ring_setup_status_page(struct intel_ring_buffer *ring) } else if (IS_GEN6(ring->dev)) { mmio = RING_HWS_PGA_GEN6(ring->mmio_base); } else { + /* XXX: gen8 returns to sanity */ mmio = RING_HWS_PGA(ring->mmio_base); } -- cgit v0.10.2 From 3a2ffb65eec6dbda2fd8151894f51c18b42c8d41 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Thu, 7 Nov 2013 21:40:51 -0800 Subject: drm/i915/bdw: Limit GTT to 2GB Because of the way in which we're allocating the pages for the Aliasing PPGTT, we cannot actually successfully alloc enough space for anything greater than 2GB. Instead of a quick hack to fix this, we should defer until we have the real solution in place (allocating much less contiguous space). This wasn't found sooner because we didn't not have any systems supporting more than a 2GB GTT. Signed-off-by: Ben Widawsky Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 5a3cc31..f69bdc7 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -1239,6 +1239,11 @@ static inline unsigned int gen8_get_total_gtt_size(u16 bdw_gmch_ctl) bdw_gmch_ctl &= BDW_GMCH_GGMS_MASK; if (bdw_gmch_ctl) bdw_gmch_ctl = 1 << bdw_gmch_ctl; + if (bdw_gmch_ctl > 4) { + WARN_ON(!i915_preliminary_hw_support); + return 4<<20; + } + return bdw_gmch_ctl << 20; } -- cgit v0.10.2 From 596cc11e7a4a89bf6c45f955402d0bd0c7d51f13 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Mon, 11 Nov 2013 14:46:28 -0800 Subject: drm/i915/bdw: PIPE_[BC] I[ME]R moved to powerwell The pipe B and pipe C interrupt mask and enable registers are now part of the pipe, so disabling the pipe power wells will lost the contests of the registers. Art totally debugged this one! v2: Use the irq_lock to clarify code, and prevent future bugs (Daniel) Cc: Art Runyan Cc: Paulo Zanoni Signed-off-by: Ben Widawsky [danvet: Make sparse happy.] Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 0a07d7c..33a8dbe 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -5684,6 +5684,7 @@ static void __intel_set_power_well(struct drm_device *dev, bool enable) { struct drm_i915_private *dev_priv = dev->dev_private; bool is_enabled, enable_requested; + unsigned long irqflags; uint32_t tmp; tmp = I915_READ(HSW_PWR_WELL_DRIVER); @@ -5701,9 +5702,24 @@ static void __intel_set_power_well(struct drm_device *dev, bool enable) HSW_PWR_WELL_STATE_ENABLED), 20)) DRM_ERROR("Timeout enabling power well\n"); } + + if (IS_BROADWELL(dev)) { + spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + I915_WRITE(GEN8_DE_PIPE_IMR(PIPE_B), + dev_priv->de_irq_mask[PIPE_B]); + I915_WRITE(GEN8_DE_PIPE_IER(PIPE_B), + ~dev_priv->de_irq_mask[PIPE_B] | + GEN8_PIPE_VBLANK); + I915_WRITE(GEN8_DE_PIPE_IMR(PIPE_C), + dev_priv->de_irq_mask[PIPE_C]); + I915_WRITE(GEN8_DE_PIPE_IER(PIPE_C), + ~dev_priv->de_irq_mask[PIPE_C] | + GEN8_PIPE_VBLANK); + POSTING_READ(GEN8_DE_PIPE_IER(PIPE_C)); + spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + } } else { if (enable_requested) { - unsigned long irqflags; enum pipe p; I915_WRITE(HSW_PWR_WELL_DRIVER, 0); -- cgit v0.10.2 From caaa4c804fae7bb654f7d00b35b8583280a9c52c Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Sat, 16 Nov 2013 17:46:02 +1100 Subject: KVM: PPC: Book3S HV: Fix physical address calculations This fixes a bug in kvmppc_do_h_enter() where the physical address for a page can be calculated incorrectly if transparent huge pages (THP) are active. Until THP came along, it was true that if we encountered a large (16M) page in kvmppc_do_h_enter(), then the associated memslot must be 16M aligned for both its guest physical address and the userspace address, and the physical address calculations in kvmppc_do_h_enter() assumed that. With THP, that is no longer true. In the case where we are using MMU notifiers and the page size that we get from the Linux page tables is larger than the page being mapped by the guest, we need to fill in some low-order bits of the physical address. Without THP, these bits would be the same in the guest physical address (gpa) and the host virtual address (hva). With THP, they can be different, and we need to use the bits from hva rather than gpa. In the case where we are not using MMU notifiers, the host physical address we get from the memslot->arch.slot_phys[] array already includes the low-order bits down to the PAGE_SIZE level, even if we are using large pages. Thus we can simplify the calculation in this case to just add in the remaining bits in the case where PAGE_SIZE is 64k and the guest is mapping a 4k page. The same bug exists in kvmppc_book3s_hv_page_fault(). The basic fix is to use psize (the page size from the HPTE) rather than pte_size (the page size from the Linux PTE) when updating the HPTE low word in r. That means that pfn needs to be computed to PAGE_SIZE granularity even if the Linux PTE is a huge page PTE. That can be arranged simply by doing the page_to_pfn() before setting page to the head of the compound page. If psize is less than PAGE_SIZE, then we need to make sure we only update the bits from PAGE_SIZE upwards, in order not to lose any sub-page offset bits in r. On the other hand, if psize is greater than PAGE_SIZE, we need to make sure we don't bring in non-zero low order bits in pfn, hence we mask (pfn << PAGE_SHIFT) with ~(psize - 1). Signed-off-by: Paul Mackerras Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c index f3ff587..47bbeaf 100644 --- a/arch/powerpc/kvm/book3s_64_mmu_hv.c +++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c @@ -665,6 +665,7 @@ int kvmppc_book3s_hv_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu, return -EFAULT; } else { page = pages[0]; + pfn = page_to_pfn(page); if (PageHuge(page)) { page = compound_head(page); pte_size <<= compound_order(page); @@ -689,7 +690,6 @@ int kvmppc_book3s_hv_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu, } rcu_read_unlock_sched(); } - pfn = page_to_pfn(page); } ret = -EFAULT; @@ -707,8 +707,14 @@ int kvmppc_book3s_hv_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu, r = (r & ~(HPTE_R_W|HPTE_R_I|HPTE_R_G)) | HPTE_R_M; } - /* Set the HPTE to point to pfn */ - r = (r & ~(HPTE_R_PP0 - pte_size)) | (pfn << PAGE_SHIFT); + /* + * Set the HPTE to point to pfn. + * Since the pfn is at PAGE_SIZE granularity, make sure we + * don't mask out lower-order bits if psize < PAGE_SIZE. + */ + if (psize < PAGE_SIZE) + psize = PAGE_SIZE; + r = (r & ~(HPTE_R_PP0 - psize)) | ((pfn << PAGE_SHIFT) & ~(psize - 1)); if (hpte_is_writable(r) && !write_ok) r = hpte_make_readonly(r); ret = RESUME_GUEST; diff --git a/arch/powerpc/kvm/book3s_hv_rm_mmu.c b/arch/powerpc/kvm/book3s_hv_rm_mmu.c index 9c51544..fddbf98 100644 --- a/arch/powerpc/kvm/book3s_hv_rm_mmu.c +++ b/arch/powerpc/kvm/book3s_hv_rm_mmu.c @@ -225,6 +225,7 @@ long kvmppc_do_h_enter(struct kvm *kvm, unsigned long flags, is_io = pa & (HPTE_R_I | HPTE_R_W); pte_size = PAGE_SIZE << (pa & KVMPPC_PAGE_ORDER_MASK); pa &= PAGE_MASK; + pa |= gpa & ~PAGE_MASK; } else { /* Translate to host virtual address */ hva = __gfn_to_hva_memslot(memslot, gfn); @@ -238,13 +239,12 @@ long kvmppc_do_h_enter(struct kvm *kvm, unsigned long flags, ptel = hpte_make_readonly(ptel); is_io = hpte_cache_bits(pte_val(pte)); pa = pte_pfn(pte) << PAGE_SHIFT; + pa |= hva & (pte_size - 1); } } if (pte_size < psize) return H_PARAMETER; - if (pa && pte_size > psize) - pa |= gpa & (pte_size - 1); ptel &= ~(HPTE_R_PP0 - psize); ptel |= pa; -- cgit v0.10.2 From f019b7ad76e6bdbc8462cbe17ad5b86a25fcdf24 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Sat, 16 Nov 2013 17:46:03 +1100 Subject: KVM: PPC: Book3S HV: Refine barriers in guest entry/exit Some users have reported instances of the host hanging with secondary threads of a core waiting for the primary thread to exit the guest, and the primary thread stuck in nap mode. This prompted a review of the memory barriers in the guest entry/exit code, and this is the result. Most of these changes are the suggestions of Dean Burdick . The barriers between updating napping_threads and reading the entry_exit_count on the one hand, and updating entry_exit_count and reading napping_threads on the other, need to be isync not lwsync, since we need to ensure that either the napping_threads update or the entry_exit_count update get seen. It is not sufficient to order the load vs. lwarx, as lwsync does; we need to order the load vs. the stwcx., so we need isync. In addition, we need a full sync before sending IPIs to wake other threads from nap, to ensure that the write to the entry_exit_count is visible before the IPI occurs. Signed-off-by: Paul Mackerras Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S index bc8de75..bde28da 100644 --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S @@ -153,7 +153,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_206) 13: b machine_check_fwnmi - /* * We come in here when wakened from nap mode on a secondary hw thread. * Relocation is off and most register values are lost. @@ -224,6 +223,11 @@ kvm_start_guest: /* Clear our vcpu pointer so we don't come back in early */ li r0, 0 std r0, HSTATE_KVM_VCPU(r13) + /* + * Make sure we clear HSTATE_KVM_VCPU(r13) before incrementing + * the nap_count, because once the increment to nap_count is + * visible we could be given another vcpu. + */ lwsync /* Clear any pending IPI - we're an offline thread */ ld r5, HSTATE_XICS_PHYS(r13) @@ -241,7 +245,6 @@ kvm_start_guest: /* increment the nap count and then go to nap mode */ ld r4, HSTATE_KVM_VCORE(r13) addi r4, r4, VCORE_NAP_COUNT - lwsync /* make previous updates visible */ 51: lwarx r3, 0, r4 addi r3, r3, 1 stwcx. r3, 0, r4 @@ -990,14 +993,13 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_201) */ /* Increment the threads-exiting-guest count in the 0xff00 bits of vcore->entry_exit_count */ - lwsync ld r5,HSTATE_KVM_VCORE(r13) addi r6,r5,VCORE_ENTRY_EXIT 41: lwarx r3,0,r6 addi r0,r3,0x100 stwcx. r0,0,r6 bne 41b - lwsync + isync /* order stwcx. vs. reading napping_threads */ /* * At this point we have an interrupt that we have to pass @@ -1030,6 +1032,8 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_201) sld r0,r0,r4 andc. r3,r3,r0 /* no sense IPI'ing ourselves */ beq 43f + /* Order entry/exit update vs. IPIs */ + sync mulli r4,r4,PACA_SIZE /* get paca for thread 0 */ subf r6,r4,r13 42: andi. r0,r3,1 @@ -1638,10 +1642,10 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_206) bge kvm_cede_exit stwcx. r4,0,r6 bne 31b + /* order napping_threads update vs testing entry_exit_count */ + isync li r0,1 stb r0,HSTATE_NAPPING(r13) - /* order napping_threads update vs testing entry_exit_count */ - lwsync mr r4,r3 lwz r7,VCORE_ENTRY_EXIT(r5) cmpwi r7,0x100 -- cgit v0.10.2 From bf3d32e1156c36c88b75960fd2e5457d5d75620b Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Sat, 16 Nov 2013 17:46:04 +1100 Subject: KVM: PPC: Book3S HV: Make tbacct_lock irq-safe Lockdep reported that there is a potential for deadlock because vcpu->arch.tbacct_lock is not irq-safe, and is sometimes taken inside the rq_lock (run-queue lock) in the scheduler, which is taken within interrupts. The lockdep splat looks like: ====================================================== [ INFO: HARDIRQ-safe -> HARDIRQ-unsafe lock order detected ] 3.12.0-rc5-kvm+ #8 Not tainted ------------------------------------------------------ qemu-system-ppc/4803 [HC0[0]:SC0[0]:HE0:SE1] is trying to acquire: (&(&vcpu->arch.tbacct_lock)->rlock){+.+...}, at: [] .kvmppc_core_vcpu_put_hv+0x2c/0xa0 and this task is already holding: (&rq->lock){-.-.-.}, at: [] .__schedule+0x180/0xaa0 which would create a new lock dependency: (&rq->lock){-.-.-.} -> (&(&vcpu->arch.tbacct_lock)->rlock){+.+...} but this new dependency connects a HARDIRQ-irq-safe lock: (&rq->lock){-.-.-.} ... which became HARDIRQ-irq-safe at: [] .lock_acquire+0xbc/0x190 [] ._raw_spin_lock+0x34/0x60 [] .scheduler_tick+0x54/0x180 [] .update_process_times+0x70/0xa0 [] .tick_periodic+0x3c/0xe0 [] .tick_handle_periodic+0x28/0xb0 [] .timer_interrupt+0x120/0x2e0 [] decrementer_common+0x168/0x180 [] .get_page_from_freelist+0x924/0xc10 [] .__alloc_pages_nodemask+0x200/0xba0 [] .alloc_pages_exact_nid+0x68/0x110 [] .page_cgroup_init+0x1e0/0x270 [] .start_kernel+0x3e0/0x4e4 [] .start_here_common+0x20/0x70 to a HARDIRQ-irq-unsafe lock: (&(&vcpu->arch.tbacct_lock)->rlock){+.+...} ... which became HARDIRQ-irq-unsafe at: ... [] .lock_acquire+0xbc/0x190 [] ._raw_spin_lock+0x34/0x60 [] .kvmppc_core_vcpu_load_hv+0x2c/0x100 [] .kvmppc_core_vcpu_load+0x2c/0x40 [] .kvm_arch_vcpu_load+0x10/0x30 [] .vcpu_load+0x64/0xd0 [] .kvm_vcpu_ioctl+0x68/0x730 [] .do_vfs_ioctl+0x4dc/0x7a0 [] .SyS_ioctl+0xc4/0xe0 [] syscall_exit+0x0/0x98 Some users have reported this deadlock occurring in practice, though the reports have been primarily on 3.10.x-based kernels. This fixes the problem by making tbacct_lock be irq-safe. Signed-off-by: Paul Mackerras Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c index 072287f..31d9cfb 100644 --- a/arch/powerpc/kvm/book3s_hv.c +++ b/arch/powerpc/kvm/book3s_hv.c @@ -131,8 +131,9 @@ static void kvmppc_fast_vcpu_kick_hv(struct kvm_vcpu *vcpu) static void kvmppc_core_vcpu_load_hv(struct kvm_vcpu *vcpu, int cpu) { struct kvmppc_vcore *vc = vcpu->arch.vcore; + unsigned long flags; - spin_lock(&vcpu->arch.tbacct_lock); + spin_lock_irqsave(&vcpu->arch.tbacct_lock, flags); if (vc->runner == vcpu && vc->vcore_state != VCORE_INACTIVE && vc->preempt_tb != TB_NIL) { vc->stolen_tb += mftb() - vc->preempt_tb; @@ -143,19 +144,20 @@ static void kvmppc_core_vcpu_load_hv(struct kvm_vcpu *vcpu, int cpu) vcpu->arch.busy_stolen += mftb() - vcpu->arch.busy_preempt; vcpu->arch.busy_preempt = TB_NIL; } - spin_unlock(&vcpu->arch.tbacct_lock); + spin_unlock_irqrestore(&vcpu->arch.tbacct_lock, flags); } static void kvmppc_core_vcpu_put_hv(struct kvm_vcpu *vcpu) { struct kvmppc_vcore *vc = vcpu->arch.vcore; + unsigned long flags; - spin_lock(&vcpu->arch.tbacct_lock); + spin_lock_irqsave(&vcpu->arch.tbacct_lock, flags); if (vc->runner == vcpu && vc->vcore_state != VCORE_INACTIVE) vc->preempt_tb = mftb(); if (vcpu->arch.state == KVMPPC_VCPU_BUSY_IN_HOST) vcpu->arch.busy_preempt = mftb(); - spin_unlock(&vcpu->arch.tbacct_lock); + spin_unlock_irqrestore(&vcpu->arch.tbacct_lock, flags); } static void kvmppc_set_msr_hv(struct kvm_vcpu *vcpu, u64 msr) @@ -486,11 +488,11 @@ static u64 vcore_stolen_time(struct kvmppc_vcore *vc, u64 now) */ if (vc->vcore_state != VCORE_INACTIVE && vc->runner->arch.run_task != current) { - spin_lock(&vc->runner->arch.tbacct_lock); + spin_lock_irq(&vc->runner->arch.tbacct_lock); p = vc->stolen_tb; if (vc->preempt_tb != TB_NIL) p += now - vc->preempt_tb; - spin_unlock(&vc->runner->arch.tbacct_lock); + spin_unlock_irq(&vc->runner->arch.tbacct_lock); } else { p = vc->stolen_tb; } @@ -512,10 +514,10 @@ static void kvmppc_create_dtl_entry(struct kvm_vcpu *vcpu, core_stolen = vcore_stolen_time(vc, now); stolen = core_stolen - vcpu->arch.stolen_logged; vcpu->arch.stolen_logged = core_stolen; - spin_lock(&vcpu->arch.tbacct_lock); + spin_lock_irq(&vcpu->arch.tbacct_lock); stolen += vcpu->arch.busy_stolen; vcpu->arch.busy_stolen = 0; - spin_unlock(&vcpu->arch.tbacct_lock); + spin_unlock_irq(&vcpu->arch.tbacct_lock); if (!dt || !vpa) return; memset(dt, 0, sizeof(struct dtl_entry)); @@ -1115,13 +1117,13 @@ static void kvmppc_remove_runnable(struct kvmppc_vcore *vc, if (vcpu->arch.state != KVMPPC_VCPU_RUNNABLE) return; - spin_lock(&vcpu->arch.tbacct_lock); + spin_lock_irq(&vcpu->arch.tbacct_lock); now = mftb(); vcpu->arch.busy_stolen += vcore_stolen_time(vc, now) - vcpu->arch.stolen_logged; vcpu->arch.busy_preempt = now; vcpu->arch.state = KVMPPC_VCPU_BUSY_IN_HOST; - spin_unlock(&vcpu->arch.tbacct_lock); + spin_unlock_irq(&vcpu->arch.tbacct_lock); --vc->n_runnable; list_del(&vcpu->arch.run_list); } -- cgit v0.10.2 From c9438092cae4a5bdbd146ca1385e85dcd6e847f8 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Sat, 16 Nov 2013 17:46:05 +1100 Subject: KVM: PPC: Book3S HV: Take SRCU read lock around kvm_read_guest() call Running a kernel with CONFIG_PROVE_RCU=y yields the following diagnostic: =============================== [ INFO: suspicious RCU usage. ] 3.12.0-rc5-kvm+ #9 Not tainted ------------------------------- include/linux/kvm_host.h:473 suspicious rcu_dereference_check() usage! other info that might help us debug this: rcu_scheduler_active = 1, debug_locks = 0 1 lock held by qemu-system-ppc/4831: stack backtrace: CPU: 28 PID: 4831 Comm: qemu-system-ppc Not tainted 3.12.0-rc5-kvm+ #9 Call Trace: [c000000be462b2a0] [c00000000001644c] .show_stack+0x7c/0x1f0 (unreliable) [c000000be462b370] [c000000000ad57c0] .dump_stack+0x88/0xb4 [c000000be462b3f0] [c0000000001315e8] .lockdep_rcu_suspicious+0x138/0x180 [c000000be462b480] [c00000000007862c] .gfn_to_memslot+0x13c/0x170 [c000000be462b510] [c00000000007d384] .gfn_to_hva_prot+0x24/0x90 [c000000be462b5a0] [c00000000007d420] .kvm_read_guest_page+0x30/0xd0 [c000000be462b630] [c00000000007d528] .kvm_read_guest+0x68/0x110 [c000000be462b6e0] [c000000000084594] .kvmppc_rtas_hcall+0x34/0x180 [c000000be462b7d0] [c000000000097934] .kvmppc_pseries_do_hcall+0x74/0x830 [c000000be462b880] [c0000000000990e8] .kvmppc_vcpu_run_hv+0xff8/0x15a0 [c000000be462b9e0] [c0000000000839cc] .kvmppc_vcpu_run+0x2c/0x40 [c000000be462ba50] [c0000000000810b4] .kvm_arch_vcpu_ioctl_run+0x54/0x1b0 [c000000be462bae0] [c00000000007b508] .kvm_vcpu_ioctl+0x478/0x730 [c000000be462bca0] [c00000000025532c] .do_vfs_ioctl+0x4dc/0x7a0 [c000000be462bd80] [c0000000002556b4] .SyS_ioctl+0xc4/0xe0 [c000000be462be30] [c000000000009ee4] syscall_exit+0x0/0x98 To fix this, we take the SRCU read lock around the kvmppc_rtas_hcall() call. Signed-off-by: Paul Mackerras Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c index 31d9cfb..b51d5db 100644 --- a/arch/powerpc/kvm/book3s_hv.c +++ b/arch/powerpc/kvm/book3s_hv.c @@ -591,7 +591,9 @@ int kvmppc_pseries_do_hcall(struct kvm_vcpu *vcpu) if (list_empty(&vcpu->kvm->arch.rtas_tokens)) return RESUME_HOST; + idx = srcu_read_lock(&vcpu->kvm->srcu); rc = kvmppc_rtas_hcall(vcpu); + srcu_read_unlock(&vcpu->kvm->srcu, idx); if (rc == -ENOENT) return RESUME_HOST; -- cgit v0.10.2 From 91648ec09c1ef69c4d840ab6dab391bfb452d554 Mon Sep 17 00:00:00 2001 From: pingfan liu Date: Fri, 15 Nov 2013 16:35:00 +0800 Subject: powerpc: kvm: fix rare but potential deadlock scene Since kvmppc_hv_find_lock_hpte() is called from both virtmode and realmode, so it can trigger the deadlock. Suppose the following scene: Two physical cpuM, cpuN, two VM instances A, B, each VM has a group of vcpus. If on cpuM, vcpu_A_1 holds bitlock X (HPTE_V_HVLOCK), then is switched out, and on cpuN, vcpu_A_2 try to lock X in realmode, then cpuN will be caught in realmode for a long time. What makes things even worse if the following happens, On cpuM, bitlockX is hold, on cpuN, Y is hold. vcpu_B_2 try to lock Y on cpuM in realmode vcpu_A_2 try to lock X on cpuN in realmode Oops! deadlock happens Signed-off-by: Liu Ping Fan Reviewed-by: Paul Mackerras CC: stable@vger.kernel.org Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c index 47bbeaf..c5d1484 100644 --- a/arch/powerpc/kvm/book3s_64_mmu_hv.c +++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c @@ -469,11 +469,14 @@ static int kvmppc_mmu_book3s_64_hv_xlate(struct kvm_vcpu *vcpu, gva_t eaddr, slb_v = vcpu->kvm->arch.vrma_slb_v; } + preempt_disable(); /* Find the HPTE in the hash table */ index = kvmppc_hv_find_lock_hpte(kvm, eaddr, slb_v, HPTE_V_VALID | HPTE_V_ABSENT); - if (index < 0) + if (index < 0) { + preempt_enable(); return -ENOENT; + } hptep = (unsigned long *)(kvm->arch.hpt_virt + (index << 4)); v = hptep[0] & ~HPTE_V_HVLOCK; gr = kvm->arch.revmap[index].guest_rpte; @@ -481,6 +484,7 @@ static int kvmppc_mmu_book3s_64_hv_xlate(struct kvm_vcpu *vcpu, gva_t eaddr, /* Unlock the HPTE */ asm volatile("lwsync" : : : "memory"); hptep[0] = v; + preempt_enable(); gpte->eaddr = eaddr; gpte->vpage = ((v & HPTE_V_AVPN) << 4) | ((eaddr >> 12) & 0xfff); diff --git a/arch/powerpc/kvm/book3s_hv_rm_mmu.c b/arch/powerpc/kvm/book3s_hv_rm_mmu.c index fddbf98..1931aa3 100644 --- a/arch/powerpc/kvm/book3s_hv_rm_mmu.c +++ b/arch/powerpc/kvm/book3s_hv_rm_mmu.c @@ -749,6 +749,10 @@ static int slb_base_page_shift[4] = { 20, /* 1M, unsupported */ }; +/* When called from virtmode, this func should be protected by + * preempt_disable(), otherwise, the holding of HPTE_V_HVLOCK + * can trigger deadlock issue. + */ long kvmppc_hv_find_lock_hpte(struct kvm *kvm, gva_t eaddr, unsigned long slb_v, unsigned long valid) { -- cgit v0.10.2 From 6802cdc58d4fe66cffd6cd04ee55e65dd61eeeeb Mon Sep 17 00:00:00 2001 From: Simon Horman Date: Wed, 6 Nov 2013 09:41:09 +0900 Subject: ARM: shmobile: lager: phy fixup needs CONFIG_PHYLIB Do not build the phy fixup unless CONFIG_PHYLIB is enabled. Other than not being useful it is also not possible to link the code under this condition as phy_register_fixup_for_id(), mdiobus_read() and mdiobus_write() are absent. arch/arm/mach-shmobile/built-in.o: In function `lager_ksz8041_fixup': board-lager.c:(.text+0xb8): undefined reference to `mdiobus_read' board-lager.c:(.text+0xd4): undefined reference to `mdiobus_write' arch/arm/mach-shmobile/built-in.o: In function `lager_init': board-lager.c:(.init.text+0xafc): undefined reference to `phy_register_fixup_for_id' This problem was introduced by 48c8b96f21817aad ("ARM: shmobile: Lager: add Micrel KSZ8041 PHY fixup") Cc: Sergei Shtylyov Signed-off-by: Simon Horman diff --git a/arch/arm/mach-shmobile/board-lager.c b/arch/arm/mach-shmobile/board-lager.c index a8d3ce6..e0406fd 100644 --- a/arch/arm/mach-shmobile/board-lager.c +++ b/arch/arm/mach-shmobile/board-lager.c @@ -245,7 +245,9 @@ static void __init lager_init(void) { lager_add_standard_devices(); - phy_register_fixup_for_id("r8a7790-ether-ff:01", lager_ksz8041_fixup); + if (IS_ENABLED(CONFIG_PHYLIB)) + phy_register_fixup_for_id("r8a7790-ether-ff:01", + lager_ksz8041_fixup); } static const char * const lager_boards_compat_dt[] __initconst = { -- cgit v0.10.2 From e55bc55867585e6628359fd5496316576fe58a2f Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Sat, 9 Nov 2013 13:18:01 +0100 Subject: irqchip: renesas-intc-irqpin: Fix register bitfield shift calculation The SENSE register bitfield position is incorrectly computed for SoCs that use 2-bit IRQ sense fields. Fix it. This has been tested on the Marzen (H1) and Bockw (M1) boards. This bug has been present since the renesas-intc-irqpin driver was introduced by 443580486e3b9657 ("irqchip: Renesas INTC External IRQ pin driver") in v3.10-rc1. Signed-off-by: Laurent Pinchart Acked-by: Magnus Damm Tested-by: Simon Horman Signed-off-by: Simon Horman diff --git a/drivers/irqchip/irq-renesas-intc-irqpin.c b/drivers/irqchip/irq-renesas-intc-irqpin.c index 82cec63..3ee78f0 100644 --- a/drivers/irqchip/irq-renesas-intc-irqpin.c +++ b/drivers/irqchip/irq-renesas-intc-irqpin.c @@ -149,8 +149,9 @@ static void intc_irqpin_read_modify_write(struct intc_irqpin_priv *p, static void intc_irqpin_mask_unmask_prio(struct intc_irqpin_priv *p, int irq, int do_mask) { - int bitfield_width = 4; /* PRIO assumed to have fixed bitfield width */ - int shift = (7 - irq) * bitfield_width; /* PRIO assumed to be 32-bit */ + /* The PRIO register is assumed to be 32-bit with fixed 4-bit fields. */ + int bitfield_width = 4; + int shift = 32 - (irq + 1) * bitfield_width; intc_irqpin_read_modify_write(p, INTC_IRQPIN_REG_PRIO, shift, bitfield_width, @@ -159,8 +160,9 @@ static void intc_irqpin_mask_unmask_prio(struct intc_irqpin_priv *p, static int intc_irqpin_set_sense(struct intc_irqpin_priv *p, int irq, int value) { + /* The SENSE register is assumed to be 32-bit. */ int bitfield_width = p->config.sense_bitfield_width; - int shift = (7 - irq) * bitfield_width; /* SENSE assumed to be 32-bit */ + int shift = 32 - (irq + 1) * bitfield_width; dev_dbg(&p->pdev->dev, "sense irq = %d, mode = %d\n", irq, value); -- cgit v0.10.2 From 3088f1085d1c08f31f02db26796555a66cdf7ca1 Mon Sep 17 00:00:00 2001 From: Kishon Vijay Abraham I Date: Mon, 25 Nov 2013 15:31:21 +0530 Subject: usb: dwc3: invoke phy_resume after phy_init usb_phy_set_suspend(phy, 0) is called before usb_phy_init. Fix it. Reported-by: Roger Quadros Signed-off-by: Kishon Vijay Abraham I Signed-off-by: Felipe Balbi diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 74f9cf0..4dd1f7e 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -455,9 +455,6 @@ static int dwc3_probe(struct platform_device *pdev) if (IS_ERR(regs)) return PTR_ERR(regs); - usb_phy_set_suspend(dwc->usb2_phy, 0); - usb_phy_set_suspend(dwc->usb3_phy, 0); - spin_lock_init(&dwc->lock); platform_set_drvdata(pdev, dwc); @@ -488,6 +485,9 @@ static int dwc3_probe(struct platform_device *pdev) goto err0; } + usb_phy_set_suspend(dwc->usb2_phy, 0); + usb_phy_set_suspend(dwc->usb3_phy, 0); + ret = dwc3_event_buffers_setup(dwc); if (ret) { dev_err(dwc->dev, "failed to setup event buffers\n"); -- cgit v0.10.2 From 501fae512584e601d34945efccd35c5e5d759132 Mon Sep 17 00:00:00 2001 From: Kishon Vijay Abraham I Date: Mon, 25 Nov 2013 15:31:22 +0530 Subject: usb: dwc3: power off usb phy in error path usb phy was power'ed on but was never power'ed off in the error path. Fix it. Signed-off-by: Kishon Vijay Abraham I Signed-off-by: Felipe Balbi diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 4dd1f7e..a49217a 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -569,6 +569,8 @@ err2: dwc3_event_buffers_cleanup(dwc); err1: + usb_phy_set_suspend(dwc->usb2_phy, 1); + usb_phy_set_suspend(dwc->usb3_phy, 1); dwc3_core_exit(dwc); err0: -- cgit v0.10.2 From 0b55149033403d78f345fb613d8ad52d16c161fc Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 8 Nov 2013 01:43:17 -0800 Subject: usb: phy: twl6030-usb: signedness bug in twl6030_readb() "ret" needs to be signed for the error handling to work. Signed-off-by: Dan Carpenter Signed-off-by: Felipe Balbi diff --git a/drivers/usb/phy/phy-twl6030-usb.c b/drivers/usb/phy/phy-twl6030-usb.c index 30e8a61..bad57ce 100644 --- a/drivers/usb/phy/phy-twl6030-usb.c +++ b/drivers/usb/phy/phy-twl6030-usb.c @@ -127,7 +127,8 @@ static inline int twl6030_writeb(struct twl6030_usb *twl, u8 module, static inline u8 twl6030_readb(struct twl6030_usb *twl, u8 module, u8 address) { - u8 data, ret = 0; + u8 data; + int ret; ret = twl_i2c_read_u8(module, &data, address); if (ret >= 0) -- cgit v0.10.2 From 23de2278ebc3a2f971ce45ca5e5e35c9d5a74040 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Thu, 21 Nov 2013 14:19:29 +0900 Subject: ARM: shmobile: r8a7790: Fix GPIO resources in DTS The r8a7790 GPIO resources are currently incorrect. Fix that by making them match the English r8a7790 v0.6 data sheet. Tested with GPIO LED using Lager DT reference. This problem has been present since GPIOs were added to the r8a7790 SoC by f98e10c88aa95bf7 ("ARM: shmobile: r8a7790: Add GPIO controller devices to device tree") in v3.12-rc1. Signed-off-by: Magnus Damm Acked-by: Laurent Pinchart Signed-off-by: Simon Horman diff --git a/arch/arm/boot/dts/r8a7790.dtsi b/arch/arm/boot/dts/r8a7790.dtsi index ee845fa..46e1d7e 100644 --- a/arch/arm/boot/dts/r8a7790.dtsi +++ b/arch/arm/boot/dts/r8a7790.dtsi @@ -87,9 +87,9 @@ interrupts = <1 9 0xf04>; }; - gpio0: gpio@ffc40000 { + gpio0: gpio@e6050000 { compatible = "renesas,gpio-r8a7790", "renesas,gpio-rcar"; - reg = <0 0xffc40000 0 0x2c>; + reg = <0 0xe6050000 0 0x50>; interrupt-parent = <&gic>; interrupts = <0 4 0x4>; #gpio-cells = <2>; @@ -99,9 +99,9 @@ interrupt-controller; }; - gpio1: gpio@ffc41000 { + gpio1: gpio@e6051000 { compatible = "renesas,gpio-r8a7790", "renesas,gpio-rcar"; - reg = <0 0xffc41000 0 0x2c>; + reg = <0 0xe6051000 0 0x50>; interrupt-parent = <&gic>; interrupts = <0 5 0x4>; #gpio-cells = <2>; @@ -111,9 +111,9 @@ interrupt-controller; }; - gpio2: gpio@ffc42000 { + gpio2: gpio@e6052000 { compatible = "renesas,gpio-r8a7790", "renesas,gpio-rcar"; - reg = <0 0xffc42000 0 0x2c>; + reg = <0 0xe6052000 0 0x50>; interrupt-parent = <&gic>; interrupts = <0 6 0x4>; #gpio-cells = <2>; @@ -123,9 +123,9 @@ interrupt-controller; }; - gpio3: gpio@ffc43000 { + gpio3: gpio@e6053000 { compatible = "renesas,gpio-r8a7790", "renesas,gpio-rcar"; - reg = <0 0xffc43000 0 0x2c>; + reg = <0 0xe6053000 0 0x50>; interrupt-parent = <&gic>; interrupts = <0 7 0x4>; #gpio-cells = <2>; @@ -135,9 +135,9 @@ interrupt-controller; }; - gpio4: gpio@ffc44000 { + gpio4: gpio@e6054000 { compatible = "renesas,gpio-r8a7790", "renesas,gpio-rcar"; - reg = <0 0xffc44000 0 0x2c>; + reg = <0 0xe6054000 0 0x50>; interrupt-parent = <&gic>; interrupts = <0 8 0x4>; #gpio-cells = <2>; @@ -147,9 +147,9 @@ interrupt-controller; }; - gpio5: gpio@ffc45000 { + gpio5: gpio@e6055000 { compatible = "renesas,gpio-r8a7790", "renesas,gpio-rcar"; - reg = <0 0xffc45000 0 0x2c>; + reg = <0 0xe6055000 0 0x50>; interrupt-parent = <&gic>; interrupts = <0 9 0x4>; #gpio-cells = <2>; -- cgit v0.10.2 From 027476642811f8559cbe00ef6cc54db230e48a20 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Mon, 2 Dec 2013 11:08:06 +0200 Subject: drm/i915: Take modeset locks around intel_modeset_setup_hw_state() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Some lower level things get angry if we don't have modeset locks during intel_modeset_setup_hw_state(). Actually the resume and lid_notify codepaths alreday hold the locks, but the init codepath doesn't, so fix that. Note: This slipped through since we only disable pipes if the plane/pipe linking doesn't match. Which is only relevant on older gen3 mobile machines, if the BIOS fails to set up our preferred linking. Signed-off-by: Ville Syrjälä Cc: stable@vger.kernel.org Tested-and-reported-by: Paul Bolle [danvet: Add note now that I could confirm my theory with the log files Paul Bolle provided.] Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 080f6fd..114db51 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -11046,7 +11046,9 @@ void intel_modeset_gem_init(struct drm_device *dev) intel_setup_overlay(dev); + drm_modeset_lock_all(dev); intel_modeset_setup_hw_state(dev, false); + drm_modeset_unlock_all(dev); } void intel_modeset_cleanup(struct drm_device *dev) -- cgit v0.10.2 From edd5b13313551d6b04acfb90d3db58ed3cf3c814 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 2 Dec 2013 17:39:09 +0000 Subject: drm/i915: Do not clobber config status after a forced restore of hw state We call intel_modeset_setup_hw_state() along two paths, driver load/resume and after a lid event notification. During initialisation of the driver, it is imperative that we reset the config state. This correctly sets up the initial connector statuses and prepares the hardware for a thorough probing. However, during a lid event, we only want to undo the damage caused by the bios by resetting our last known mode. In this cirumstance, we do not want to clobber our desired state. In order to try and keep sanity between the config state and our own tracking, do the drm_mode_config_reset() first along the load/resume paths before reading out the hw state and apply any definite known corrections. v2: "As discussed on irc I don't think we should force the connector state to anything here: Imo connector->status should reflect what we believe to be the true output connection state, whereas connector->encoder reflects whether this connector is wired up to a pipe. And since we no longer reject modeset on disconnected connectors and never nuked the pipe if the connector gets disconnected there's no reason for that - such policy is userspace's job. This regression has been introduced in commit 2e9388923e83bc4e2726f170a984621f1d582e77 Author: Daniel Vetter Date: Thu Oct 11 20:08:24 2012 +0200 drm/i915/crt: explicitly set up HOTPLUG_BITS on resume" so sayeth Daniel. Signed-off-by: Chris Wilson Cc: Daniel Vetter Cc: stable@vger.kernel.org (v3.8 and later) Cc: Paulo Zanoni Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 2e367a1..5b7b7e0 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -651,6 +651,7 @@ static int __i915_drm_thaw(struct drm_device *dev, bool restore_gtt_mappings) intel_modeset_init_hw(dev); drm_modeset_lock_all(dev); + drm_mode_config_reset(dev); intel_modeset_setup_hw_state(dev, true); drm_modeset_unlock_all(dev); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 114db51..8e44b16 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -11036,8 +11036,6 @@ void intel_modeset_setup_hw_state(struct drm_device *dev, } intel_modeset_check_state(dev); - - drm_mode_config_reset(dev); } void intel_modeset_gem_init(struct drm_device *dev) @@ -11047,6 +11045,7 @@ void intel_modeset_gem_init(struct drm_device *dev) intel_setup_overlay(dev); drm_modeset_lock_all(dev); + drm_mode_config_reset(dev); intel_modeset_setup_hw_state(dev, false); drm_modeset_unlock_all(dev); } -- cgit v0.10.2 From 5ae68b413214e847a2b5c6d3c65778482542bc1a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Mon, 2 Dec 2013 11:23:39 +0200 Subject: drm/i915: Skip clock checks on BDW MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We don't have clock state readout support for DDI, so skip the pipe config clock checks on all DDI platforms. Signed-off-by: Ville Syrjälä Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 8e44b16..8b8bde7 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -9135,7 +9135,7 @@ intel_pipe_config_compare(struct drm_device *dev, if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) PIPE_CONF_CHECK_I(pipe_bpp); - if (!IS_HASWELL(dev)) { + if (!HAS_DDI(dev)) { PIPE_CONF_CHECK_CLOCK_FUZZY(adjusted_mode.crtc_clock); PIPE_CONF_CHECK_CLOCK_FUZZY(port_clock); } -- cgit v0.10.2 From c94cae53f9e564484f906a79be5639fc66e8cb02 Mon Sep 17 00:00:00 2001 From: Eric Trudeau Date: Wed, 4 Dec 2013 11:39:33 +0000 Subject: XEN: Grant table address, xen_hvm_resume_frames, is a phys_addr not a pfn From: Eric Trudeau xen_hvm_resume_frames stores the physical address of the grant table. englighten.c was incorrectly setting it as if it was a page frame number. This caused the table to be mapped into the guest at an unexpected physical address. Additionally, a warning is improved to include the grant table address which failed in xen_remap. Signed-off-by: Eric Trudeau Signed-off-by: Stefano Stabellini diff --git a/arch/arm/xen/enlighten.c b/arch/arm/xen/enlighten.c index 83e4f95..6a288c7 100644 --- a/arch/arm/xen/enlighten.c +++ b/arch/arm/xen/enlighten.c @@ -224,10 +224,10 @@ static int __init xen_guest_init(void) } if (of_address_to_resource(node, GRANT_TABLE_PHYSADDR, &res)) return 0; - xen_hvm_resume_frames = res.start >> PAGE_SHIFT; + xen_hvm_resume_frames = res.start; xen_events_irq = irq_of_parse_and_map(node, 0); pr_info("Xen %s support found, events_irq=%d gnttab_frame_pfn=%lx\n", - version, xen_events_irq, xen_hvm_resume_frames); + version, xen_events_irq, (xen_hvm_resume_frames >> PAGE_SHIFT)); xen_domain_type = XEN_HVM_DOMAIN; xen_setup_features(); diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c index 02838719..aa846a4 100644 --- a/drivers/xen/grant-table.c +++ b/drivers/xen/grant-table.c @@ -1176,7 +1176,8 @@ static int gnttab_setup(void) gnttab_shared.addr = xen_remap(xen_hvm_resume_frames, PAGE_SIZE * max_nr_gframes); if (gnttab_shared.addr == NULL) { - pr_warn("Failed to ioremap gnttab share frames!\n"); + pr_warn("Failed to ioremap gnttab share frames (addr=0x%08lx)!\n", + xen_hvm_resume_frames); return -ENOMEM; } } -- cgit v0.10.2 From 7f62b6ee767586ee7e5d12787dbaaaf47a91979a Mon Sep 17 00:00:00 2001 From: Nicolin Chen Date: Wed, 4 Dec 2013 11:18:36 +0800 Subject: ASoC: soc-pcm: Use valid condition for snd_soc_dai_digital_mute() in hw_free() The snd_soc_dai_digital_mute() here will be never executed because we only decrease codec->active in snd_soc_close(). Thus correct it. Signed-off-by: Nicolin Chen Signed-off-by: Mark Brown diff --git a/sound/soc/soc-pcm.c b/sound/soc/soc-pcm.c index 11a90cd..891b9a9 100644 --- a/sound/soc/soc-pcm.c +++ b/sound/soc/soc-pcm.c @@ -600,12 +600,13 @@ static int soc_pcm_hw_free(struct snd_pcm_substream *substream) struct snd_soc_platform *platform = rtd->platform; struct snd_soc_dai *cpu_dai = rtd->cpu_dai; struct snd_soc_dai *codec_dai = rtd->codec_dai; - struct snd_soc_codec *codec = rtd->codec; + bool playback = substream->stream == SNDRV_PCM_STREAM_PLAYBACK; mutex_lock_nested(&rtd->pcm_mutex, rtd->pcm_subclass); /* apply codec digital mute */ - if (!codec->active) + if ((playback && codec_dai->playback_active == 1) || + (!playback && codec_dai->capture_active == 1)) snd_soc_dai_digital_mute(codec_dai, 1, substream->stream); /* free any machine hw params */ -- cgit v0.10.2 From a8f1f100ad994725a8295f6997524c57c72e06f5 Mon Sep 17 00:00:00 2001 From: Bo Shen Date: Wed, 4 Dec 2013 10:37:03 +0800 Subject: ASoC: atmel_ssc_dai: add dai trigger ops According to the SSC specifiation, it should be enabled after DMA is enabled. So, add trigger operation to make sure the right sequence. Signed-off-by: Bo Shen Tested-by: Richard Genoud Signed-off-by: Mark Brown diff --git a/sound/soc/atmel/atmel_ssc_dai.c b/sound/soc/atmel/atmel_ssc_dai.c index 8697ced..1ead3c9 100644 --- a/sound/soc/atmel/atmel_ssc_dai.c +++ b/sound/soc/atmel/atmel_ssc_dai.c @@ -648,7 +648,7 @@ static int atmel_ssc_prepare(struct snd_pcm_substream *substream, dma_params = ssc_p->dma_params[dir]; - ssc_writel(ssc_p->ssc->regs, CR, dma_params->mask->ssc_enable); + ssc_writel(ssc_p->ssc->regs, CR, dma_params->mask->ssc_disable); ssc_writel(ssc_p->ssc->regs, IDR, dma_params->mask->ssc_error); pr_debug("%s enabled SSC_SR=0x%08x\n", @@ -657,6 +657,33 @@ static int atmel_ssc_prepare(struct snd_pcm_substream *substream, return 0; } +static int atmel_ssc_trigger(struct snd_pcm_substream *substream, + int cmd, struct snd_soc_dai *dai) +{ + struct atmel_ssc_info *ssc_p = &ssc_info[dai->id]; + struct atmel_pcm_dma_params *dma_params; + int dir; + + if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) + dir = 0; + else + dir = 1; + + dma_params = ssc_p->dma_params[dir]; + + switch (cmd) { + case SNDRV_PCM_TRIGGER_START: + case SNDRV_PCM_TRIGGER_RESUME: + case SNDRV_PCM_TRIGGER_PAUSE_RELEASE: + ssc_writel(ssc_p->ssc->regs, CR, dma_params->mask->ssc_enable); + break; + default: + ssc_writel(ssc_p->ssc->regs, CR, dma_params->mask->ssc_disable); + break; + } + + return 0; +} #ifdef CONFIG_PM static int atmel_ssc_suspend(struct snd_soc_dai *cpu_dai) @@ -731,6 +758,7 @@ static const struct snd_soc_dai_ops atmel_ssc_dai_ops = { .startup = atmel_ssc_startup, .shutdown = atmel_ssc_shutdown, .prepare = atmel_ssc_prepare, + .trigger = atmel_ssc_trigger, .hw_params = atmel_ssc_hw_params, .set_fmt = atmel_ssc_set_dai_fmt, .set_clkdiv = atmel_ssc_set_dai_clkdiv, -- cgit v0.10.2 From bc567a93502275755492141524935269dcf0ea1b Mon Sep 17 00:00:00 2001 From: Bo Shen Date: Wed, 4 Dec 2013 10:37:04 +0800 Subject: ASoC: sam9x5_wm8731: change to work in DSP A mode Change sam9x5 with wm8731 work in DSP A mode, this will fix the left/right channel swap issue. Signed-off-by: Bo Shen Tested-by: Richard Genoud Signed-off-by: Mark Brown diff --git a/sound/soc/atmel/sam9x5_wm8731.c b/sound/soc/atmel/sam9x5_wm8731.c index 1b37228..7d6a905 100644 --- a/sound/soc/atmel/sam9x5_wm8731.c +++ b/sound/soc/atmel/sam9x5_wm8731.c @@ -109,7 +109,7 @@ static int sam9x5_wm8731_driver_probe(struct platform_device *pdev) dai->stream_name = "WM8731 PCM"; dai->codec_dai_name = "wm8731-hifi"; dai->init = sam9x5_wm8731_init; - dai->dai_fmt = SND_SOC_DAIFMT_I2S | SND_SOC_DAIFMT_NB_NF + dai->dai_fmt = SND_SOC_DAIFMT_DSP_A | SND_SOC_DAIFMT_NB_NF | SND_SOC_DAIFMT_CBM_CFM; ret = snd_soc_of_parse_card_name(card, "atmel,model"); -- cgit v0.10.2 From 0d1430a3f4b7cfd8779b78740a4182321f3ca7f3 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 4 Dec 2013 14:52:06 +0000 Subject: drm/i915: Hold mutex across i915_gem_release Inorder to serialise the closing of the file descriptor and its subsequent release of client requests with i915_gem_free_request(), we need to hold the struct_mutex in i915_gem_release(). Failing to do so has the potential to trigger an OOPS, later with a use-after-free. Testcase: igt/gem_close_race Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=70874 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71029 Reported-by: Eric Anholt Signed-off-by: Chris Wilson Cc: stable@vger.kernel.org Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 0cab2d0..ac9dac9 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -1848,8 +1848,10 @@ void i915_driver_lastclose(struct drm_device * dev) void i915_driver_preclose(struct drm_device * dev, struct drm_file *file_priv) { + mutex_lock(&dev->struct_mutex); i915_gem_context_close(dev, file_priv); i915_gem_release(dev, file_priv); + mutex_unlock(&dev->struct_mutex); } void i915_driver_postclose(struct drm_device *dev, struct drm_file *file) diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 72a3df3..4a05956 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -347,10 +347,8 @@ void i915_gem_context_close(struct drm_device *dev, struct drm_file *file) { struct drm_i915_file_private *file_priv = file->driver_priv; - mutex_lock(&dev->struct_mutex); idr_for_each(&file_priv->context_idr, context_idr_cleanup, NULL); idr_destroy(&file_priv->context_idr); - mutex_unlock(&dev->struct_mutex); } static struct i915_hw_context * -- cgit v0.10.2 From cf30d52e2d11523c42048ab89ed4215b5021526a Mon Sep 17 00:00:00 2001 From: Maria Dimakopoulou Date: Thu, 5 Dec 2013 01:24:37 +0200 Subject: perf/x86: Fix constraint table end marker bug The EVENT_CONSTRAINT_END() macro defines the end marker as a constraint with a weight of zero. This was all fine until we blacklisted the corrupting memory events on Intel IvyBridge. These events are blacklisted by using a counter bitmask of zero. Thus, they also get a constraint weight of zero. The iteration macro: for_each_constraint tests the weight==0. Therefore, it was stopping at the first blacklisted event, i.e., 0xd0. The corrupting events were therefore considered as unconstrained and were scheduled on any of the generic counters. This patch fixes the end marker to have a weight of -1. With this, the blacklisted events get an empty constraint and cannot be scheduled which is what we want for now. Signed-off-by: Maria Dimakopoulou Reviewed-by: Stephane Eranian Cc: peterz@infradead.org Cc: ak@linux.intel.com Cc: jolsa@redhat.com Cc: zheng.z.yan@intel.com Link: http://lkml.kernel.org/r/20131204232437.GA10689@starlight Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h index fd00bb2..c1a8618 100644 --- a/arch/x86/kernel/cpu/perf_event.h +++ b/arch/x86/kernel/cpu/perf_event.h @@ -262,11 +262,20 @@ struct cpu_hw_events { __EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \ HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST_HSW) -#define EVENT_CONSTRAINT_END \ - EVENT_CONSTRAINT(0, 0, 0) +/* + * We define the end marker as having a weight of -1 + * to enable blacklisting of events using a counter bitmask + * of zero and thus a weight of zero. + * The end marker has a weight that cannot possibly be + * obtained from counting the bits in the bitmask. + */ +#define EVENT_CONSTRAINT_END { .weight = -1 } +/* + * Check for end marker with weight == -1 + */ #define for_each_event_constraint(e, c) \ - for ((e) = (c); (e)->weight; (e)++) + for ((e) = (c); (e)->weight != -1; (e)++) /* * Extra registers for specific events. -- cgit v0.10.2 From 75704ecfbb4124139b78b71dd603f05d61abe689 Mon Sep 17 00:00:00 2001 From: Nicolin Chen Date: Wed, 4 Dec 2013 17:22:16 +0800 Subject: ASoC: wm8962: Enable SYSCLK provisonally before fetching generated DSPCLK_DIV DSPCLK_DIV can be only generated correctly after enabling SYSCLK. But if the current bias_level hasn't reached SND_SOC_BIAS_ON, DAPM won't enable SYSCLK, which would cause the calculation result from DSPCLK_DIV invalid since bit DSPCLK_DIV will be finally turned to its true value after DAPM enables SYSCLK while the driver won't calculate it again for the current instance. In this circumstance, a playback which needs non-zero DSPCLK_DIV would be distorted due to unexpected clock frequency resulted from an invalid DSPCLK_DIV value. So this patch provisionally enables the SYSCLK to get a valid DSPCLK_DIV for calculation and then disables it afterward. Signed-off-by: Nicolin Chen Acked-by: Charles Keepax Signed-off-by: Mark Brown diff --git a/sound/soc/codecs/wm8962.c b/sound/soc/codecs/wm8962.c index 543c5c2..0f17ed3 100644 --- a/sound/soc/codecs/wm8962.c +++ b/sound/soc/codecs/wm8962.c @@ -2439,7 +2439,20 @@ static void wm8962_configure_bclk(struct snd_soc_codec *codec) snd_soc_update_bits(codec, WM8962_CLOCKING_4, WM8962_SYSCLK_RATE_MASK, clocking4); + /* DSPCLK_DIV can be only generated correctly after enabling SYSCLK. + * So we here provisionally enable it and then disable it afterward + * if current bias_level hasn't reached SND_SOC_BIAS_ON. + */ + if (codec->dapm.bias_level != SND_SOC_BIAS_ON) + snd_soc_update_bits(codec, WM8962_CLOCKING2, + WM8962_SYSCLK_ENA_MASK, WM8962_SYSCLK_ENA); + dspclk = snd_soc_read(codec, WM8962_CLOCKING1); + + if (codec->dapm.bias_level != SND_SOC_BIAS_ON) + snd_soc_update_bits(codec, WM8962_CLOCKING2, + WM8962_SYSCLK_ENA_MASK, 0); + if (dspclk < 0) { dev_err(codec->dev, "Failed to read DSPCLK: %d\n", dspclk); return; -- cgit v0.10.2 From f742a55231b950ed71ac8cc8beec0944ddda312c Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Fri, 6 Dec 2013 10:17:53 +0100 Subject: drm/i915: fix pm init ordering Shovel a bit more of the the code into the setup function, and call it earlier. Otherwise lockdep is unhappy since we cancel the delayed resume work before it's initialized. While at it also shovel the pc8 setup code into the same functions. I wanted to also ditch the header declaration of the hws pc8 functions, but for unfathomable reasons that stuff is in intel_display.c instead of intel_pm.c. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71980 Tested-by: Guo Jinxian Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index ac9dac9..197db09 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -1490,16 +1490,9 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) spin_lock_init(&dev_priv->uncore.lock); spin_lock_init(&dev_priv->mm.object_stat_lock); mutex_init(&dev_priv->dpio_lock); - mutex_init(&dev_priv->rps.hw_lock); mutex_init(&dev_priv->modeset_restore_lock); - mutex_init(&dev_priv->pc8.lock); - dev_priv->pc8.requirements_met = false; - dev_priv->pc8.gpu_idle = false; - dev_priv->pc8.irqs_disabled = false; - dev_priv->pc8.enabled = false; - dev_priv->pc8.disable_count = 2; /* requirements_met + gpu_idle */ - INIT_DELAYED_WORK(&dev_priv->pc8.enable_work, hsw_enable_pc8_work); + intel_pm_setup(dev); intel_display_crc_init(dev); @@ -1603,7 +1596,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) } intel_irq_init(dev); - intel_pm_init(dev); intel_uncore_sanitize(dev); /* Try to make sure MCHBAR is enabled before poking at it */ diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index ccdbecc..6f04fa4 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1901,9 +1901,7 @@ void i915_queue_hangcheck(struct drm_device *dev); void i915_handle_error(struct drm_device *dev, bool wedged); extern void intel_irq_init(struct drm_device *dev); -extern void intel_pm_init(struct drm_device *dev); extern void intel_hpd_init(struct drm_device *dev); -extern void intel_pm_init(struct drm_device *dev); extern void intel_uncore_sanitize(struct drm_device *dev); extern void intel_uncore_early_sanitize(struct drm_device *dev); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index a18e88b..79f91f2 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -821,6 +821,7 @@ void intel_update_sprite_watermarks(struct drm_plane *plane, uint32_t sprite_width, int pixel_size, bool enabled, bool scaled); void intel_init_pm(struct drm_device *dev); +void intel_pm_setup(struct drm_device *dev); bool intel_fbc_enabled(struct drm_device *dev); void intel_update_fbc(struct drm_device *dev); void intel_gpu_ips_init(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 6e0d5e0..e0dec95 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6130,10 +6130,19 @@ int vlv_freq_opcode(int ddr_freq, int val) return val; } -void intel_pm_init(struct drm_device *dev) +void intel_pm_setup(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; + mutex_init(&dev_priv->rps.hw_lock); + + mutex_init(&dev_priv->pc8.lock); + dev_priv->pc8.requirements_met = false; + dev_priv->pc8.gpu_idle = false; + dev_priv->pc8.irqs_disabled = false; + dev_priv->pc8.enabled = false; + dev_priv->pc8.disable_count = 2; /* requirements_met + gpu_idle */ + INIT_DELAYED_WORK(&dev_priv->pc8.enable_work, hsw_enable_pc8_work); INIT_DELAYED_WORK(&dev_priv->rps.delayed_resume_work, intel_gen6_powersave_work); } -- cgit v0.10.2 From acc240d41ea1ab9c488a79219fb313b5b46265ae Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 5 Dec 2013 15:42:34 +0100 Subject: drm/i915: Fix use-after-free in do_switch MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit So apparently under ridiculous amounts of memory pressure we can get into trouble in do_switch when we try to move the old hw context backing storage object onto the active lists. With list debugging enabled that usually results in us chasing a poisoned pointer - which means we've hit upon a vma that has been removed from all lrus with list_del (and then deallocated, so it's a real use-after free). Ian Lister has done some great callchain chasing and noticed that we can reenter do_switch: i915_gem_do_execbuffer() i915_switch_context() do_switch() from = ring->last_context; i915_gem_object_pin() i915_gem_object_bind_to_gtt() ret = drm_mm_insert_node_in_range_generic(); // If the above call fails then it will try i915_gem_evict_something() // If that fails it will call i915_gem_evict_everything() ... i915_gem_evict_everything() i915_gpu_idle() i915_switch_context(DEFAULT_CONTEXT) Like with everything else where the shrinker or eviction code can invalidate pointers we need to reload relevant state. Note that there's no need to recheck whether a context switch is still required because: - Doing a switch to the same context is harmless (besides wasting a bit of energy). - This can only happen with the default context. But since that one's pinned we'll never call down into evict_everything under normal circumstances. Note that there's a little driver bringup fun involved namely that we could recourse into do_switch for the initial switch. Atm we're fine since we assign the context pointer only after the call to do_switch at driver load or resume time. And in the gpu reset case we skip the entire setup sequence (which might be a bug on its own, but definitely not this one here). Cc'ing stable since apparently ChromeOS guys are seeing this in the wild (and not just on artificial stress tests), see the reference. Note that in upstream code doesn't calle evict_everything directly from evict_something, that's an extension in this product branch. But we can still hit upon this bug (and apparently we do, see the linked backtraces). I've noticed this while trying to construct a testcase for this bug and utterly failed to provoke it. It looks like we need to driver the system squarly into the lowmem wall and provoke the shrinker to evict the context object by doing the last-ditch evict_everything call. Aside: There's currently no means to get a badly-fragmenting hw context object away from a bad spot in the upstream code. We should fix this by at least adding some code to evict_something to handle hw contexts. References: https://code.google.com/p/chromium/issues/detail?id=248191 Reported-by: Ian Lister Cc: Ian Lister Cc: stable@vger.kernel.org Cc: Ben Widawsky Cc: Stéphane Marchesin Cc: Bloomfield, Jon Tested-by: Rafael Barbalho Reviewed-by: Ian Lister Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 4a05956..b0f42b9 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -421,11 +421,21 @@ static int do_switch(struct i915_hw_context *to) if (ret) return ret; - /* Clear this page out of any CPU caches for coherent swap-in/out. Note + /* + * Pin can switch back to the default context if we end up calling into + * evict_everything - as a last ditch gtt defrag effort that also + * switches to the default context. Hence we need to reload from here. + */ + from = ring->last_context; + + /* + * Clear this page out of any CPU caches for coherent swap-in/out. Note * that thanks to write = false in this call and us not setting any gpu * write domains when putting a context object onto the active list * (when switching away from it), this won't block. - * XXX: We need a real interface to do this instead of trickery. */ + * + * XXX: We need a real interface to do this instead of trickery. + */ ret = i915_gem_object_set_to_gtt_domain(to->obj, false); if (ret) { i915_gem_object_unpin(to->obj); -- cgit v0.10.2 From b6497b383f65dd1bc60ea1f5d70e157a268fbd15 Mon Sep 17 00:00:00 2001 From: Ian Campbell Date: Fri, 6 Dec 2013 17:55:56 +0000 Subject: xen: privcmd: do not return pages which we have failed to unmap This failure represents a hypervisor issue, but if it does occur then nothing good can come of returning pages which still refer to a foreign owned page into the general allocation pool. Instead we are forced to leak them. Log that we have done so. The potential for failure only exists for autotranslated guest (e.g. ARM and x86 PVH). Signed-off-by: Ian Campbell Signed-off-by: Stefano Stabellini Reviewed-by: David Vrabel diff --git a/drivers/xen/privcmd.c b/drivers/xen/privcmd.c index 8e74590..569a13b 100644 --- a/drivers/xen/privcmd.c +++ b/drivers/xen/privcmd.c @@ -533,12 +533,17 @@ static void privcmd_close(struct vm_area_struct *vma) { struct page **pages = vma->vm_private_data; int numpgs = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT; + int rc; if (!xen_feature(XENFEAT_auto_translated_physmap) || !numpgs || !pages) return; - xen_unmap_domain_mfn_range(vma, numpgs, pages); - free_xenballooned_pages(numpgs, pages); + rc = xen_unmap_domain_mfn_range(vma, numpgs, pages); + if (rc == 0) + free_xenballooned_pages(numpgs, pages); + else + pr_crit("unable to unmap MFN range: leaking %d pages. rc=%d\n", + numpgs, rc); kfree(pages); } -- cgit v0.10.2 From 851dd02b4f1fdb437586190f87217e3382a736bd Mon Sep 17 00:00:00 2001 From: Chris Ruehl Date: Wed, 4 Dec 2013 10:02:44 +0800 Subject: usb: phy-tegra-usb.c: wrong pointer check for remap UTMI usb: phy-tegra-usb.c: wrong pointer check for remap UTMI A wrong pointer was used to test the result of devm_ioremap() Acked-by: Stephen Warren Reviewed-by: Thierry Reding Signed-off-by: Chris Ruehl Acked-by: Venu Byravarasu Signed-off-by: Felipe Balbi diff --git a/drivers/usb/phy/phy-tegra-usb.c b/drivers/usb/phy/phy-tegra-usb.c index 82232ac..bbe4f8e 100644 --- a/drivers/usb/phy/phy-tegra-usb.c +++ b/drivers/usb/phy/phy-tegra-usb.c @@ -876,7 +876,7 @@ static int utmi_phy_probe(struct tegra_usb_phy *tegra_phy, tegra_phy->pad_regs = devm_ioremap(&pdev->dev, res->start, resource_size(res)); - if (!tegra_phy->regs) { + if (!tegra_phy->pad_regs) { dev_err(&pdev->dev, "Failed to remap UTMI Pad regs\n"); return -ENOMEM; } -- cgit v0.10.2 From a8b14744429f90763f258ad0f69601cdcad610aa Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 4 Dec 2013 09:20:40 -0500 Subject: sysfs: give different locking key to regular and bin files 027a485d12e0 ("sysfs: use a separate locking class for open files depending on mmap") assigned different lockdep key to sysfs_open_file->mutex depending on whether the file implements mmap or not in an attempt to avoid spurious lockdep warning caused by merging of regular and bin file paths. While this restored some of the original behavior of using different locks (at least lockdep is concerned) for the different clases of files. The restoration wasn't full because now the lockdep key assignment depends on whether the file has mmap or not instead of whether it's a regular file or not. This means that bin files which don't implement mmap will get assigned the same lockdep class as regular files. This is problematic because file_operations for bin files still implements the mmap file operation and checking whether the sysfs file actually implements mmap happens in the file operation after grabbing @sysfs_open_file->mutex. We still end up adding locking dependency from mmap locking to sysfs_open_file->mutex to the regular file mutex which triggers spurious circular locking warning. Fix it by restoring the original behavior fully by differentiating lockdep key by whether the file is regular or bin, instead of the existence of mmap. Signed-off-by: Tejun Heo Reported-by: Dave Jones Link: http://lkml.kernel.org/g/20131203184324.GA11320@redhat.com Signed-off-by: Greg Kroah-Hartman diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c index b94f936..35e7d08 100644 --- a/fs/sysfs/file.c +++ b/fs/sysfs/file.c @@ -609,7 +609,7 @@ static int sysfs_open_file(struct inode *inode, struct file *file) struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata; struct kobject *kobj = attr_sd->s_parent->s_dir.kobj; struct sysfs_open_file *of; - bool has_read, has_write, has_mmap; + bool has_read, has_write; int error = -EACCES; /* need attr_sd for attr and ops, its parent for kobj */ @@ -621,7 +621,6 @@ static int sysfs_open_file(struct inode *inode, struct file *file) has_read = battr->read || battr->mmap; has_write = battr->write || battr->mmap; - has_mmap = battr->mmap; } else { const struct sysfs_ops *ops = sysfs_file_ops(attr_sd); @@ -633,7 +632,6 @@ static int sysfs_open_file(struct inode *inode, struct file *file) has_read = ops->show; has_write = ops->store; - has_mmap = false; } /* check perms and supported operations */ @@ -661,9 +659,9 @@ static int sysfs_open_file(struct inode *inode, struct file *file) * open file has a separate mutex, it's okay as long as those don't * happen on the same file. At this point, we can't easily give * each file a separate locking class. Let's differentiate on - * whether the file has mmap or not for now. + * whether the file is bin or not for now. */ - if (has_mmap) + if (sysfs_is_bin(attr_sd)) mutex_init(&of->mutex); else mutex_init(&of->mutex); -- cgit v0.10.2 From 11e6a09fba446ce799a9de51d2d99586024bbdaa Mon Sep 17 00:00:00 2001 From: Jakob Bornecrantz Date: Sun, 8 Dec 2013 23:39:04 -0800 Subject: drm/vmwgfx: Add max surface memory param Userspace uses this to workaround overcommit issues by flushing the command stream early. Signed-off-by: Jakob Bornecrantz Reviewed-by: Thomas Hellstrom diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c index a51f48e..45d5b5a 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c @@ -68,6 +68,9 @@ int vmw_getparam_ioctl(struct drm_device *dev, void *data, SVGA_FIFO_3D_HWVERSION)); break; } + case DRM_VMW_PARAM_MAX_SURF_MEMORY: + param->value = dev_priv->memory_size; + break; default: DRM_ERROR("Illegal vmwgfx get param request: %d\n", param->param); diff --git a/include/uapi/drm/vmwgfx_drm.h b/include/uapi/drm/vmwgfx_drm.h index bcb0912..f854ca4 100644 --- a/include/uapi/drm/vmwgfx_drm.h +++ b/include/uapi/drm/vmwgfx_drm.h @@ -75,6 +75,7 @@ #define DRM_VMW_PARAM_FIFO_CAPS 4 #define DRM_VMW_PARAM_MAX_FB_SIZE 5 #define DRM_VMW_PARAM_FIFO_HW_VERSION 6 +#define DRM_VMW_PARAM_MAX_SURF_MEMORY 7 /** * struct drm_vmw_getparam_arg -- cgit v0.10.2 From d825a04387ff4ce66117306f2862c7cedca5c597 Mon Sep 17 00:00:00 2001 From: Alexander Graf Date: Fri, 29 Nov 2013 02:24:18 +0100 Subject: KVM: PPC: Book3S: PR: Don't clobber our exit handler id We call a C helper to save all svcpu fields into our vcpu. The C ABI states that r12 is considered volatile. However, we keep our exit handler id in r12 currently. So we need to save it away into a non-volatile register instead that definitely does get preserved across the C call. This bug usually didn't hit anyone yet since gcc is smart enough to generate code that doesn't even need r12 which means it stayed identical throughout the call by sheer luck. But we can't rely on that. Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_interrupts.S b/arch/powerpc/kvm/book3s_interrupts.S index f4dd041..5e7cb32 100644 --- a/arch/powerpc/kvm/book3s_interrupts.S +++ b/arch/powerpc/kvm/book3s_interrupts.S @@ -132,9 +132,17 @@ kvm_start_lightweight: * */ + PPC_LL r3, GPR4(r1) /* vcpu pointer */ + + /* + * kvmppc_copy_from_svcpu can clobber volatile registers, save + * the exit handler id to the vcpu and restore it from there later. + */ + stw r12, VCPU_TRAP(r3) + /* Transfer reg values from shadow vcpu back to vcpu struct */ /* On 64-bit, interrupts are still off at this point */ - PPC_LL r3, GPR4(r1) /* vcpu pointer */ + GET_SHADOW_VCPU(r4) bl FUNC(kvmppc_copy_from_svcpu) nop @@ -151,7 +159,6 @@ kvm_start_lightweight: */ ld r3, PACA_SPRG3(r13) mtspr SPRN_SPRG3, r3 - #endif /* CONFIG_PPC_BOOK3S_64 */ /* R7 = vcpu */ @@ -177,7 +184,7 @@ kvm_start_lightweight: PPC_STL r31, VCPU_GPR(R31)(r7) /* Pass the exit number as 3rd argument to kvmppc_handle_exit */ - mr r5, r12 + lwz r5, VCPU_TRAP(r7) /* Restore r3 (kvm_run) and r4 (vcpu) */ REST_2GPRS(3, r1) -- cgit v0.10.2 From c9dad7f9db4ed42de37d3f0ef2b2c0e10d5b6f92 Mon Sep 17 00:00:00 2001 From: Alexander Graf Date: Fri, 29 Nov 2013 02:27:23 +0100 Subject: KVM: PPC: Book3S: PR: Export kvmppc_copy_to|from_svcpu The kvmppc_copy_{to,from}_svcpu functions are publically visible, so we should also export them in a header for others C files to consume. So far we didn't need this because we only called it from asm code. The next patch will introduce a C caller. Signed-off-by: Alexander Graf diff --git a/arch/powerpc/include/asm/kvm_book3s.h b/arch/powerpc/include/asm/kvm_book3s.h index 4a594b7..bc23b1b 100644 --- a/arch/powerpc/include/asm/kvm_book3s.h +++ b/arch/powerpc/include/asm/kvm_book3s.h @@ -192,6 +192,10 @@ extern void kvmppc_load_up_vsx(void); extern u32 kvmppc_alignment_dsisr(struct kvm_vcpu *vcpu, unsigned int inst); extern ulong kvmppc_alignment_dar(struct kvm_vcpu *vcpu, unsigned int inst); extern int kvmppc_h_pr(struct kvm_vcpu *vcpu, unsigned long cmd); +extern void kvmppc_copy_to_svcpu(struct kvmppc_book3s_shadow_vcpu *svcpu, + struct kvm_vcpu *vcpu); +extern void kvmppc_copy_from_svcpu(struct kvm_vcpu *vcpu, + struct kvmppc_book3s_shadow_vcpu *svcpu); static inline struct kvmppc_vcpu_book3s *to_book3s(struct kvm_vcpu *vcpu) { -- cgit v0.10.2 From 40fdd8c88c4a5e9b26bfbed2215ac661f24aef07 Mon Sep 17 00:00:00 2001 From: Alexander Graf Date: Fri, 29 Nov 2013 02:29:00 +0100 Subject: KVM: PPC: Book3S: PR: Make svcpu -> vcpu store preempt savvy As soon as we get back to our "highmem" handler in virtual address space we may get preempted. Today the reason we can get preempted is that we replay interrupts and all the lazy logic thinks we have interrupts enabled. However, it's not hard to make the code interruptible and that way we can enable and handle interrupts even earlier. This fixes random guest crashes that happened with CONFIG_PREEMPT=y for me. Signed-off-by: Alexander Graf diff --git a/arch/powerpc/include/asm/kvm_book3s_asm.h b/arch/powerpc/include/asm/kvm_book3s_asm.h index 0bd9348..412b2f3 100644 --- a/arch/powerpc/include/asm/kvm_book3s_asm.h +++ b/arch/powerpc/include/asm/kvm_book3s_asm.h @@ -106,6 +106,7 @@ struct kvmppc_host_state { }; struct kvmppc_book3s_shadow_vcpu { + bool in_use; ulong gpr[14]; u32 cr; u32 xer; diff --git a/arch/powerpc/kvm/book3s_pr.c b/arch/powerpc/kvm/book3s_pr.c index fe14ca3..5b9e906 100644 --- a/arch/powerpc/kvm/book3s_pr.c +++ b/arch/powerpc/kvm/book3s_pr.c @@ -66,6 +66,7 @@ static void kvmppc_core_vcpu_load_pr(struct kvm_vcpu *vcpu, int cpu) struct kvmppc_book3s_shadow_vcpu *svcpu = svcpu_get(vcpu); memcpy(svcpu->slb, to_book3s(vcpu)->slb_shadow, sizeof(svcpu->slb)); svcpu->slb_max = to_book3s(vcpu)->slb_shadow_max; + svcpu->in_use = 0; svcpu_put(svcpu); #endif vcpu->cpu = smp_processor_id(); @@ -78,6 +79,9 @@ static void kvmppc_core_vcpu_put_pr(struct kvm_vcpu *vcpu) { #ifdef CONFIG_PPC_BOOK3S_64 struct kvmppc_book3s_shadow_vcpu *svcpu = svcpu_get(vcpu); + if (svcpu->in_use) { + kvmppc_copy_from_svcpu(vcpu, svcpu); + } memcpy(to_book3s(vcpu)->slb_shadow, svcpu->slb, sizeof(svcpu->slb)); to_book3s(vcpu)->slb_shadow_max = svcpu->slb_max; svcpu_put(svcpu); @@ -110,12 +114,26 @@ void kvmppc_copy_to_svcpu(struct kvmppc_book3s_shadow_vcpu *svcpu, svcpu->ctr = vcpu->arch.ctr; svcpu->lr = vcpu->arch.lr; svcpu->pc = vcpu->arch.pc; + svcpu->in_use = true; } /* Copy data touched by real-mode code from shadow vcpu back to vcpu */ void kvmppc_copy_from_svcpu(struct kvm_vcpu *vcpu, struct kvmppc_book3s_shadow_vcpu *svcpu) { + /* + * vcpu_put would just call us again because in_use hasn't + * been updated yet. + */ + preempt_disable(); + + /* + * Maybe we were already preempted and synced the svcpu from + * our preempt notifiers. Don't bother touching this svcpu then. + */ + if (!svcpu->in_use) + goto out; + vcpu->arch.gpr[0] = svcpu->gpr[0]; vcpu->arch.gpr[1] = svcpu->gpr[1]; vcpu->arch.gpr[2] = svcpu->gpr[2]; @@ -139,6 +157,10 @@ void kvmppc_copy_from_svcpu(struct kvm_vcpu *vcpu, vcpu->arch.fault_dar = svcpu->fault_dar; vcpu->arch.fault_dsisr = svcpu->fault_dsisr; vcpu->arch.last_inst = svcpu->last_inst; + svcpu->in_use = false; + +out: + preempt_enable(); } static int kvmppc_core_check_requests_pr(struct kvm_vcpu *vcpu) -- cgit v0.10.2 From 3d3319b45eea26df56c53aae1a65adf74c8ab12a Mon Sep 17 00:00:00 2001 From: Alexander Graf Date: Fri, 29 Nov 2013 02:32:31 +0100 Subject: KVM: PPC: Book3S: PR: Enable interrupts earlier Now that the svcpu sync is interrupt aware we can enable interrupts earlier in the exit code path again, moving 32bit and 64bit closer together. While at it, document the fact that we're always executing the exit path with interrupts enabled so that the next person doesn't trap over this. Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_interrupts.S b/arch/powerpc/kvm/book3s_interrupts.S index 5e7cb32..f779450 100644 --- a/arch/powerpc/kvm/book3s_interrupts.S +++ b/arch/powerpc/kvm/book3s_interrupts.S @@ -129,6 +129,7 @@ kvm_start_lightweight: * R12 = exit handler id * R13 = PACA * SVCPU.* = guest * + * MSR.EE = 1 * */ @@ -148,11 +149,6 @@ kvm_start_lightweight: nop #ifdef CONFIG_PPC_BOOK3S_64 - /* Re-enable interrupts */ - ld r3, HSTATE_HOST_MSR(r13) - ori r3, r3, MSR_EE - MTMSR_EERI(r3) - /* * Reload kernel SPRG3 value. * No need to save guest value as usermode can't modify SPRG3. diff --git a/arch/powerpc/kvm/book3s_rmhandlers.S b/arch/powerpc/kvm/book3s_rmhandlers.S index a38c4c9..c3c5231 100644 --- a/arch/powerpc/kvm/book3s_rmhandlers.S +++ b/arch/powerpc/kvm/book3s_rmhandlers.S @@ -153,15 +153,11 @@ _GLOBAL(kvmppc_entry_trampoline) li r6, MSR_IR | MSR_DR andc r6, r5, r6 /* Clear DR and IR in MSR value */ -#ifdef CONFIG_PPC_BOOK3S_32 /* * Set EE in HOST_MSR so that it's enabled when we get into our - * C exit handler function. On 64-bit we delay enabling - * interrupts until we have finished transferring stuff - * to or from the PACA. + * C exit handler function. */ ori r5, r5, MSR_EE -#endif mtsrr0 r7 mtsrr1 r6 RFI -- cgit v0.10.2 From 241bf43321a10815225f477bba96a42285a2da73 Mon Sep 17 00:00:00 2001 From: Stephen Warren Date: Fri, 6 Dec 2013 13:34:50 -0700 Subject: ASoC: tegra: fix uninitialized variables in set_fmt In tegra*_i2s_set_fmt(), in the (fmt == SND_SOC_DAIFMT_CBM_CFM) case, "val" is never assigned to, but left uninitialized. The other case does initialized it. Fix this by initializing val at the start of the function, and only ever ORing into it. Update the handling of "mask" so it works the same way for consistency. Update tegra20_spdif.c to use the same code-style for consistency, even though it doesn't happen to suffer from the same problem at present. Signed-off-by: Stephen Warren Reviewed-by: Thierry Reding Signed-off-by: Mark Brown Fixes: 0f163546a772 ("ASoC: tegra: use regmap more directly") Cc: diff --git a/sound/soc/tegra/tegra20_i2s.c b/sound/soc/tegra/tegra20_i2s.c index 364bf6a9..8c819f8 100644 --- a/sound/soc/tegra/tegra20_i2s.c +++ b/sound/soc/tegra/tegra20_i2s.c @@ -74,7 +74,7 @@ static int tegra20_i2s_set_fmt(struct snd_soc_dai *dai, unsigned int fmt) { struct tegra20_i2s *i2s = snd_soc_dai_get_drvdata(dai); - unsigned int mask, val; + unsigned int mask = 0, val = 0; switch (fmt & SND_SOC_DAIFMT_INV_MASK) { case SND_SOC_DAIFMT_NB_NF: @@ -83,10 +83,10 @@ static int tegra20_i2s_set_fmt(struct snd_soc_dai *dai, return -EINVAL; } - mask = TEGRA20_I2S_CTRL_MASTER_ENABLE; + mask |= TEGRA20_I2S_CTRL_MASTER_ENABLE; switch (fmt & SND_SOC_DAIFMT_MASTER_MASK) { case SND_SOC_DAIFMT_CBS_CFS: - val = TEGRA20_I2S_CTRL_MASTER_ENABLE; + val |= TEGRA20_I2S_CTRL_MASTER_ENABLE; break; case SND_SOC_DAIFMT_CBM_CFM: break; diff --git a/sound/soc/tegra/tegra20_spdif.c b/sound/soc/tegra/tegra20_spdif.c index 08bc693..8c7c102 100644 --- a/sound/soc/tegra/tegra20_spdif.c +++ b/sound/soc/tegra/tegra20_spdif.c @@ -67,15 +67,15 @@ static int tegra20_spdif_hw_params(struct snd_pcm_substream *substream, { struct device *dev = dai->dev; struct tegra20_spdif *spdif = snd_soc_dai_get_drvdata(dai); - unsigned int mask, val; + unsigned int mask = 0, val = 0; int ret, spdifclock; - mask = TEGRA20_SPDIF_CTRL_PACK | - TEGRA20_SPDIF_CTRL_BIT_MODE_MASK; + mask |= TEGRA20_SPDIF_CTRL_PACK | + TEGRA20_SPDIF_CTRL_BIT_MODE_MASK; switch (params_format(params)) { case SNDRV_PCM_FORMAT_S16_LE: - val = TEGRA20_SPDIF_CTRL_PACK | - TEGRA20_SPDIF_CTRL_BIT_MODE_16BIT; + val |= TEGRA20_SPDIF_CTRL_PACK | + TEGRA20_SPDIF_CTRL_BIT_MODE_16BIT; break; default: return -EINVAL; diff --git a/sound/soc/tegra/tegra30_i2s.c b/sound/soc/tegra/tegra30_i2s.c index 231a785..02247fe 100644 --- a/sound/soc/tegra/tegra30_i2s.c +++ b/sound/soc/tegra/tegra30_i2s.c @@ -118,7 +118,7 @@ static int tegra30_i2s_set_fmt(struct snd_soc_dai *dai, unsigned int fmt) { struct tegra30_i2s *i2s = snd_soc_dai_get_drvdata(dai); - unsigned int mask, val; + unsigned int mask = 0, val = 0; switch (fmt & SND_SOC_DAIFMT_INV_MASK) { case SND_SOC_DAIFMT_NB_NF: @@ -127,10 +127,10 @@ static int tegra30_i2s_set_fmt(struct snd_soc_dai *dai, return -EINVAL; } - mask = TEGRA30_I2S_CTRL_MASTER_ENABLE; + mask |= TEGRA30_I2S_CTRL_MASTER_ENABLE; switch (fmt & SND_SOC_DAIFMT_MASTER_MASK) { case SND_SOC_DAIFMT_CBS_CFS: - val = TEGRA30_I2S_CTRL_MASTER_ENABLE; + val |= TEGRA30_I2S_CTRL_MASTER_ENABLE; break; case SND_SOC_DAIFMT_CBM_CFM: break; -- cgit v0.10.2 From 5dfc03f141993c101c626c84d639019e98a4f39c Mon Sep 17 00:00:00 2001 From: Nicolin Chen Date: Fri, 6 Dec 2013 23:38:28 +0800 Subject: ASoC: fsl: imx-wm8962: Don't update bias_level in machine driver If we update it here, the set_bias_level() of Codec driver won't be normally called and we will then miss some essential procedures in set_bias_level() of the Codec driver. Thus drop it. Signed-off-by: Nicolin Chen Signed-off-by: Mark Brown diff --git a/sound/soc/fsl/imx-wm8962.c b/sound/soc/fsl/imx-wm8962.c index 72064e9..72718e1 100644 --- a/sound/soc/fsl/imx-wm8962.c +++ b/sound/soc/fsl/imx-wm8962.c @@ -130,8 +130,6 @@ static int imx_wm8962_set_bias_level(struct snd_soc_card *card, break; } - dapm->bias_level = level; - return 0; } -- cgit v0.10.2 From 6b9f3e65282b3bd7ed77e7b2b1edfe7cfed48115 Mon Sep 17 00:00:00 2001 From: Stephen Warren Date: Tue, 3 Dec 2013 14:26:33 -0700 Subject: ASoC: don't leak on error in snd_dmaengine_pcm_register If snd_dmaengine_pcm_register()'s call to snd_soc_add_platform() fails, all objects allocated during registration are leaked. Fix this by adding error-handling code. Signed-off-by: Stephen Warren Acked-by: Lars-Peter Clausen Signed-off-by: Mark Brown diff --git a/sound/soc/soc-generic-dmaengine-pcm.c b/sound/soc/soc-generic-dmaengine-pcm.c index cbc9c96..41949af 100644 --- a/sound/soc/soc-generic-dmaengine-pcm.c +++ b/sound/soc/soc-generic-dmaengine-pcm.c @@ -305,6 +305,20 @@ static void dmaengine_pcm_request_chan_of(struct dmaengine_pcm *pcm, } } +static void dmaengine_pcm_release_chan(struct dmaengine_pcm *pcm) +{ + unsigned int i; + + for (i = SNDRV_PCM_STREAM_PLAYBACK; i <= SNDRV_PCM_STREAM_CAPTURE; + i++) { + if (!pcm->chan[i]) + continue; + dma_release_channel(pcm->chan[i]); + if (pcm->flags & SND_DMAENGINE_PCM_FLAG_HALF_DUPLEX) + break; + } +} + /** * snd_dmaengine_pcm_register - Register a dmaengine based PCM device * @dev: The parent device for the PCM device @@ -315,6 +329,7 @@ int snd_dmaengine_pcm_register(struct device *dev, const struct snd_dmaengine_pcm_config *config, unsigned int flags) { struct dmaengine_pcm *pcm; + int ret; pcm = kzalloc(sizeof(*pcm), GFP_KERNEL); if (!pcm) @@ -326,11 +341,20 @@ int snd_dmaengine_pcm_register(struct device *dev, dmaengine_pcm_request_chan_of(pcm, dev); if (flags & SND_DMAENGINE_PCM_FLAG_NO_RESIDUE) - return snd_soc_add_platform(dev, &pcm->platform, + ret = snd_soc_add_platform(dev, &pcm->platform, &dmaengine_no_residue_pcm_platform); else - return snd_soc_add_platform(dev, &pcm->platform, + ret = snd_soc_add_platform(dev, &pcm->platform, &dmaengine_pcm_platform); + if (ret) + goto err_free_dma; + + return 0; + +err_free_dma: + dmaengine_pcm_release_chan(pcm); + kfree(pcm); + return ret; } EXPORT_SYMBOL_GPL(snd_dmaengine_pcm_register); @@ -345,7 +369,6 @@ void snd_dmaengine_pcm_unregister(struct device *dev) { struct snd_soc_platform *platform; struct dmaengine_pcm *pcm; - unsigned int i; platform = snd_soc_lookup_platform(dev); if (!platform) @@ -353,15 +376,8 @@ void snd_dmaengine_pcm_unregister(struct device *dev) pcm = soc_platform_to_pcm(platform); - for (i = SNDRV_PCM_STREAM_PLAYBACK; i <= SNDRV_PCM_STREAM_CAPTURE; i++) { - if (pcm->chan[i]) { - dma_release_channel(pcm->chan[i]); - if (pcm->flags & SND_DMAENGINE_PCM_FLAG_HALF_DUPLEX) - break; - } - } - snd_soc_remove_platform(platform); + dmaengine_pcm_release_chan(pcm); kfree(pcm); } EXPORT_SYMBOL_GPL(snd_dmaengine_pcm_unregister); -- cgit v0.10.2 From 4144bc861ed7934d56f16d2acd808d44af0fcc90 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Bj=C3=B8rn=20Mork?= Date: Fri, 29 Nov 2013 20:17:45 +0100 Subject: usb: cdc-wdm: manage_power should always set needs_remote_wakeup MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Cc: stable Reported-by: Oliver Neukum Signed-off-by: Bjørn Mork Acked-by: Oliver Neukum Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/usb/class/cdc-wdm.c b/drivers/usb/class/cdc-wdm.c index 4d38759..0b23a86 100644 --- a/drivers/usb/class/cdc-wdm.c +++ b/drivers/usb/class/cdc-wdm.c @@ -854,13 +854,11 @@ static int wdm_manage_power(struct usb_interface *intf, int on) { /* need autopm_get/put here to ensure the usbcore sees the new value */ int rv = usb_autopm_get_interface(intf); - if (rv < 0) - goto err; intf->needs_remote_wakeup = on; - usb_autopm_put_interface(intf); -err: - return rv; + if (!rv) + usb_autopm_put_interface(intf); + return 0; } static int wdm_probe(struct usb_interface *intf, const struct usb_device_id *id) -- cgit v0.10.2 From 52d0dc7597c89b2ab779f3dcb9b9bf0800dd9218 Mon Sep 17 00:00:00 2001 From: Dmitry Kunilov Date: Tue, 3 Dec 2013 12:11:30 -0800 Subject: usb: serial: zte_ev: move support for ZTE AC2726 from zte_ev back to option ZTE AC2726 EVDO modem drops ppp connection every minute when driven by zte_ev but works fine when driven by option. Move the support for AC2726 back to option driver. Signed-off-by: Dmitry Kunilov Cc: stable Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/usb/serial/option.c b/drivers/usb/serial/option.c index 496b7e39..cc7a241 100644 --- a/drivers/usb/serial/option.c +++ b/drivers/usb/serial/option.c @@ -251,6 +251,7 @@ static void option_instat_callback(struct urb *urb); #define ZTE_PRODUCT_MF628 0x0015 #define ZTE_PRODUCT_MF626 0x0031 #define ZTE_PRODUCT_MC2718 0xffe8 +#define ZTE_PRODUCT_AC2726 0xfff1 #define BENQ_VENDOR_ID 0x04a5 #define BENQ_PRODUCT_H10 0x4068 @@ -1453,6 +1454,7 @@ static const struct usb_device_id option_ids[] = { { USB_VENDOR_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0xff, 0x02, 0x01) }, { USB_VENDOR_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0xff, 0x02, 0x05) }, { USB_VENDOR_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0xff, 0x86, 0x10) }, + { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, ZTE_PRODUCT_AC2726, 0xff, 0xff, 0xff) }, { USB_DEVICE(BENQ_VENDOR_ID, BENQ_PRODUCT_H10) }, { USB_DEVICE(DLINK_VENDOR_ID, DLINK_PRODUCT_DWM_652) }, diff --git a/drivers/usb/serial/zte_ev.c b/drivers/usb/serial/zte_ev.c index fca4c75..eae2c87 100644 --- a/drivers/usb/serial/zte_ev.c +++ b/drivers/usb/serial/zte_ev.c @@ -281,8 +281,7 @@ static const struct usb_device_id id_table[] = { { USB_DEVICE(0x19d2, 0xfffd) }, { USB_DEVICE(0x19d2, 0xfffc) }, { USB_DEVICE(0x19d2, 0xfffb) }, - /* AC2726, AC8710_V3 */ - { USB_DEVICE_AND_INTERFACE_INFO(0x19d2, 0xfff1, 0xff, 0xff, 0xff) }, + /* AC8710_V3 */ { USB_DEVICE(0x19d2, 0xfff6) }, { USB_DEVICE(0x19d2, 0xfff7) }, { USB_DEVICE(0x19d2, 0xfff8) }, -- cgit v0.10.2 From cc5c9eb67f912cb2c349b04063ff9f444affbc59 Mon Sep 17 00:00:00 2001 From: Fabio Estevam Date: Thu, 5 Dec 2013 15:20:49 +0800 Subject: usb: chipidea: host: Only disable the vbus regulator if it is not NULL Commit 40ed51a4b (usb: chipidea: host: add vbus regulator control) introduced a smatch complaint because regulator_disable() is called without checking whether ci->platdata->reg_vbus is not NULL. Fix this by adding the check. This patch is needed for 3.12 stable Cc: stable Reported-by: Dan Carpenter Signed-off-by: Fabio Estevam Signed-off-by: Peter Chen Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/usb/chipidea/host.c b/drivers/usb/chipidea/host.c index 59e6020..526cd77 100644 --- a/drivers/usb/chipidea/host.c +++ b/drivers/usb/chipidea/host.c @@ -88,7 +88,8 @@ static int host_start(struct ci_hdrc *ci) return ret; disable_reg: - regulator_disable(ci->platdata->reg_vbus); + if (ci->platdata->reg_vbus) + regulator_disable(ci->platdata->reg_vbus); put_hcd: usb_put_hcd(hcd); -- cgit v0.10.2 From 5a1e1456fc633da5291285b1fff75d2a7507375b Mon Sep 17 00:00:00 2001 From: Peter Chen Date: Thu, 5 Dec 2013 15:20:50 +0800 Subject: usb: chipidea: fix nobody cared IRQ when booting with host role If we connect Male-A-To-Male-A cable between otg-host and host pc, the ci->vbus_active is set wrongly, and cause the controller run at peripheral mode when we load gadget module (ci_udc_start will be run), but the software runs at host mode due to id = 0. The ehci_irq can't handle suspend (USBi_SLI) interrupt which is enabled for peripheral mode, it causes no one will handle irq error. This patch is needed for 3.12 stable Cc: stable Acked-by: Michael Grzeschik Reported-by: Marc Kleine-Budde Tested-by: Marc Kleine-Budde Signed-off-by: Marc Kleine-Budde Signed-off-by: Peter Chen Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/usb/chipidea/core.c b/drivers/usb/chipidea/core.c index 5d8981c..6e73f8c 100644 --- a/drivers/usb/chipidea/core.c +++ b/drivers/usb/chipidea/core.c @@ -642,6 +642,10 @@ static int ci_hdrc_probe(struct platform_device *pdev) : CI_ROLE_GADGET; } + /* only update vbus status for peripheral */ + if (ci->role == CI_ROLE_GADGET) + ci_handle_vbus_change(ci); + ret = ci_role_start(ci, ci->role); if (ret) { dev_err(dev, "can't start %s role\n", ci_role(ci)->name); diff --git a/drivers/usb/chipidea/udc.c b/drivers/usb/chipidea/udc.c index b34c819..69d20fb 100644 --- a/drivers/usb/chipidea/udc.c +++ b/drivers/usb/chipidea/udc.c @@ -1795,9 +1795,6 @@ static int udc_start(struct ci_hdrc *ci) pm_runtime_no_callbacks(&ci->gadget.dev); pm_runtime_enable(&ci->gadget.dev); - /* Update ci->vbus_active */ - ci_handle_vbus_change(ci); - return retval; destroy_eps: -- cgit v0.10.2 From ad071acb53110c8efd26ff1e5b5d57449b43833b Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 9 Dec 2013 10:37:24 +0000 Subject: drm/i915: Repeat eviction search after idling the GPU With the advent of hw context support, we gained some objects that are pinned for the duration of their request. That is we can make aperture space available by idling the GPU and in the process performing a context switch back to the always-pinned default context. As such, we should not conclude that there is no space in the aperture for the current object until we have unpinned any such context objects. Note that we also have the problem of outstanding pageflips preventing eviction of their framebuffer objects to resolve. Testcase: igt/gem_ctx_exec/eviction Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=72507 Signed-off-by: Chris Wilson Tested-by: lu hua Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index b737653..8f3adc7 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -88,6 +88,7 @@ i915_gem_evict_something(struct drm_device *dev, struct i915_address_space *vm, } else drm_mm_init_scan(&vm->mm, min_size, alignment, cache_level); +search_again: /* First see if there is a large enough contiguous idle region... */ list_for_each_entry(vma, &vm->inactive_list, mm_list) { if (mark_free(vma, &unwind_list)) @@ -115,10 +116,17 @@ none: list_del_init(&vma->exec_list); } - /* We expect the caller to unpin, evict all and try again, or give up. - * So calling i915_gem_evict_vm() is unnecessary. + /* Can we unpin some objects such as idle hw contents, + * or pending flips? */ - return -ENOSPC; + ret = nonblocking ? -ENOSPC : i915_gpu_idle(dev); + if (ret) + return ret; + + /* Only idle the GPU and repeat the search once */ + i915_gem_retire_requests(dev); + nonblocking = true; + goto search_again; found: /* drm_mm doesn't allow any other other operations while -- cgit v0.10.2 From f2fa75cdf8b1e1c44353747a156ec25c7478c07a Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Tue, 10 Dec 2013 09:43:09 +0100 Subject: MAINTAINERS: update GPIO maintainers entry Add Alexandre Courbot as co-maintainer of the GPIO subsystem. Provide a pointer to the GPIO GIT tree. Update the documentation file path. Move around to put people and git tree on top. Acked-by: Alexandre Courbot Signed-off-by: Linus Walleij diff --git a/MAINTAINERS b/MAINTAINERS index 13c15c8..7e053e0 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3765,9 +3765,11 @@ F: include/uapi/linux/gigaset_dev.h GPIO SUBSYSTEM M: Linus Walleij -S: Maintained +M: Alexandre Courbot L: linux-gpio@vger.kernel.org -F: Documentation/gpio.txt +T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio.git +S: Maintained +F: Documentation/gpio/ F: drivers/gpio/ F: include/linux/gpio* F: include/asm-generic/gpio.h -- cgit v0.10.2 From 8620f394c4f9abd13e4fdf927d9c2bbeda74cde7 Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Tue, 26 Nov 2013 02:45:34 +0100 Subject: sh-pfc: Fix PINMUX_GPIO macro Commit 7cbb0e55e27e ("sh-pfc: Don't duplicate argument to PINMUX_GPIO macro") erronesouly modified the PINMUX_GPIO macro in a way that resulted in all pins being named "name". Fix the macro to name the pins correctly. Cc: stable@vger.kernel.org Signed-off-by: Laurent Pinchart Signed-off-by: Linus Walleij diff --git a/drivers/pinctrl/sh-pfc/sh_pfc.h b/drivers/pinctrl/sh-pfc/sh_pfc.h index 11bd0d9..e214295 100644 --- a/drivers/pinctrl/sh-pfc/sh_pfc.h +++ b/drivers/pinctrl/sh-pfc/sh_pfc.h @@ -254,7 +254,7 @@ struct sh_pfc_soc_info { #define PINMUX_GPIO(_pin) \ [GPIO_##_pin] = { \ .pin = (u16)-1, \ - .name = __stringify(name), \ + .name = __stringify(GPIO_##_pin), \ .enum_id = _pin##_DATA, \ } -- cgit v0.10.2 From f5837ec11f8cfa6d53ebc5806582771b2c9988c6 Mon Sep 17 00:00:00 2001 From: Roger Quadros Date: Thu, 5 Dec 2013 11:23:35 +0200 Subject: gpio: twl4030: Fix regression for twl gpio LED output Commit 0b2aa8be introduced a regression that causes failure in setting LED GPO direction to OUT. This causes USB host probe failures for Beagleboard C4. platform usb_phy_gen_xceiv.2: Driver usb_phy_gen_xceiv requests probe deferral hsusb2_vcc: Failed to request enable GPIO510: -22 reg-fixed-voltage reg-fixed-voltage.0.auto: Failed to register regulator: -22 reg-fixed-voltage: probe of reg-fixed-voltage.0.auto failed with error -22 direction_out/direction_in must return 0 if the operation succeeded. Also, don't update direction flag and output data if twl4030_set_gpio_direction() failed inside twl_direction_out(); Cc: stable@vger.kernel.org Signed-off-by: Roger Quadros Acked-by: Tony Lindgren Signed-off-by: Linus Walleij diff --git a/drivers/gpio/gpio-twl4030.c b/drivers/gpio/gpio-twl4030.c index b97d6a6..f999689 100644 --- a/drivers/gpio/gpio-twl4030.c +++ b/drivers/gpio/gpio-twl4030.c @@ -300,7 +300,7 @@ static int twl_direction_in(struct gpio_chip *chip, unsigned offset) if (offset < TWL4030_GPIO_MAX) ret = twl4030_set_gpio_direction(offset, 1); else - ret = -EINVAL; + ret = -EINVAL; /* LED outputs can't be set as input */ if (!ret) priv->direction &= ~BIT(offset); @@ -354,11 +354,20 @@ static void twl_set(struct gpio_chip *chip, unsigned offset, int value) static int twl_direction_out(struct gpio_chip *chip, unsigned offset, int value) { struct gpio_twl4030_priv *priv = to_gpio_twl4030(chip); - int ret = -EINVAL; + int ret = 0; mutex_lock(&priv->mutex); - if (offset < TWL4030_GPIO_MAX) + if (offset < TWL4030_GPIO_MAX) { ret = twl4030_set_gpio_direction(offset, 0); + if (ret) { + mutex_unlock(&priv->mutex); + return ret; + } + } + + /* + * LED gpios i.e. offset >= TWL4030_GPIO_MAX are always output + */ priv->direction |= BIT(offset); mutex_unlock(&priv->mutex); -- cgit v0.10.2 From 09ca27579ee5a1b6f537b82165080b85c1febac8 Mon Sep 17 00:00:00 2001 From: Axel Lin Date: Wed, 20 Nov 2013 10:15:11 +0800 Subject: clocksource: time-efm32: Select CLKSRC_MMIO MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The time-efm32 driver uses the clocksource MMIO functions. Thus it needs to select CLKSRC_MMIO in Kconfig. Signed-off-by: Axel Lin Signed-off-by: Daniel Lezcano Acked-by: Uwe Kleine-König diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig index 5c07a56..634c4d6 100644 --- a/drivers/clocksource/Kconfig +++ b/drivers/clocksource/Kconfig @@ -75,6 +75,7 @@ config CLKSRC_DBX500_PRCMU_SCHED_CLOCK config CLKSRC_EFM32 bool "Clocksource for Energy Micro's EFM32 SoCs" if !ARCH_EFM32 depends on OF && ARM && (ARCH_EFM32 || COMPILE_TEST) + select CLKSRC_MMIO default ARCH_EFM32 help Support to use the timers of EFM32 SoCs as clock source and clock -- cgit v0.10.2 From c813eff078588733a3d1a46c033c2d59d66a263b Mon Sep 17 00:00:00 2001 From: Ezequiel Garcia Date: Tue, 26 Nov 2013 18:20:14 -0300 Subject: clocksource: armada-370-xp: Register sched_clock after the counter reset This commit registers the sched_clock _after_ the counter reset (instead of before). This removes the timestamp 'jump' in kernel log messages. Before this change: [ 0.000000] sched_clock: 32 bits at 25MHz, resolution 40ns, wraps every 171798691800ns [ 0.000000] Initializing Coherency fabric [ 0.000000] Aurora cache controller enabled [ 0.000000] l2x0: 16 ways, CACHE_ID 0x00000100, AUX_CTRL 0x1a696b12, Cache size: 1024 kB [ 163.507447] Calibrating delay loop... 1325.05 BogoMIPS (lpj=662528) [ 163.521419] pid_max: default: 32768 minimum: 301 [ 163.526185] Mount-cache hash table entries: 512 [ 163.531095] CPU: Testing write buffer coherency: ok After this change: [ 0.000000] sched_clock: 32 bits at 25MHz, resolution 40ns, wraps every 171798691800ns [ 0.000000] Initializing Coherency fabric [ 0.000000] Aurora cache controller enabled [ 0.000000] l2x0: 16 ways, CACHE_ID 0x00000100, AUX_CTRL 0x1a696b12, Cache size: 1024 kB [ 0.016849] Calibrating delay loop... 1325.05 BogoMIPS (lpj=662528) [ 0.030820] pid_max: default: 32768 minimum: 301 [ 0.035588] Mount-cache hash table entries: 512 [ 0.040500] CPU: Testing write buffer coherency: ok Signed-off-by: Ezequiel Garcia Signed-off-by: Daniel Lezcano Acked-by: Jason Cooper diff --git a/drivers/clocksource/time-armada-370-xp.c b/drivers/clocksource/time-armada-370-xp.c index d8e47e5..4e7f680 100644 --- a/drivers/clocksource/time-armada-370-xp.c +++ b/drivers/clocksource/time-armada-370-xp.c @@ -256,11 +256,6 @@ static void __init armada_370_xp_timer_common_init(struct device_node *np) ticks_per_jiffy = (timer_clk + HZ / 2) / HZ; /* - * Set scale and timer for sched_clock. - */ - sched_clock_register(armada_370_xp_read_sched_clock, 32, timer_clk); - - /* * Setup free-running clocksource timer (interrupts * disabled). */ @@ -270,6 +265,11 @@ static void __init armada_370_xp_timer_common_init(struct device_node *np) timer_ctrl_clrset(0, TIMER0_EN | TIMER0_RELOAD_EN | TIMER0_DIV(TIMER_DIVIDER_SHIFT)); + /* + * Set scale and timer for sched_clock. + */ + sched_clock_register(armada_370_xp_read_sched_clock, 32, timer_clk); + clocksource_mmio_init(timer_base + TIMER0_VAL_OFF, "armada_370_xp_clocksource", timer_clk, 300, 32, clocksource_mmio_readl_down); -- cgit v0.10.2 From 4c4b053235fa73db1ea241aa5a6b021afb0ed8db Mon Sep 17 00:00:00 2001 From: Thierry Reding Date: Sat, 19 Oct 2013 00:49:48 +0200 Subject: clocksource: clksrc-of: Do not drop unheld reference on device node MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When booting a recent kernel on ARM with OF_DYNAMIC enabled, the kernel warns about the following: [ 0.000000] ERROR: Bad of_node_put() on /timer@50004600 [ 0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.12.0-rc5-next-20131017-00077-gedfd827-dirty #406 [ 0.000000] [] (unwind_backtrace+0x0/0xf4) from [] (show_stack+0x10/0x14) [ 0.000000] [] (show_stack+0x10/0x14) from [] (dump_stack+0x9c/0xc8) [ 0.000000] [] (dump_stack+0x9c/0xc8) from [] (of_node_release+0x90/0x9c) [ 0.000000] [] (of_node_release+0x90/0x9c) from [] (of_find_matching_node_and_match+0x78/0xb4) [ 0.000000] [] (of_find_matching_node_and_match+0x78/0xb4) from [] (clocksource_of_init+0x60/0x70) [ 0.000000] [] (clocksource_of_init+0x60/0x70) from [] (start_kernel+0x1f4/0x33c) [ 0.000000] [] (start_kernel+0x1f4/0x33c) from [<80008074>] (0x80008074) This is caused by clocksource_of_init() dropping a reference on the device node that it never took. The reference taken by the loop is implicitly dropped on subsequent iterations. See the implementation of and the comment on top of the of_find_matching_node_and_match() function for reference (no pun intended). Signed-off-by: Thierry Reding Signed-off-by: Daniel Lezcano Cc: Uwe Kleine-König diff --git a/drivers/clocksource/clksrc-of.c b/drivers/clocksource/clksrc-of.c index 35639cf4..b9ddd9e 100644 --- a/drivers/clocksource/clksrc-of.c +++ b/drivers/clocksource/clksrc-of.c @@ -35,6 +35,5 @@ void __init clocksource_of_init(void) init_func = match->data; init_func(np); - of_node_put(np); } } -- cgit v0.10.2 From 6db50bb67598668c525f12e2f7191f5d03ca46f2 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 2 Dec 2013 09:29:35 +0000 Subject: clocksource: sunxi: Stop timer from ticking before enabling interrupts The sun4i timer can still be ticking when we enable the interrupt. If another timer is actually used (A7 architected timer, for example), odds are that the interrupt will eventually fire with the event_handler pointer being NULL. The obvious fix it to stop the timer before registering the interrupt. Observed and tested on sun7i (cubietruck). Cc: Daniel Lezcano Acked-by: Maxime Ripard Signed-off-by: Marc Zyngier Signed-off-by: Daniel Lezcano diff --git a/drivers/clocksource/sun4i_timer.c b/drivers/clocksource/sun4i_timer.c index 2fb4695..a4f6119 100644 --- a/drivers/clocksource/sun4i_timer.c +++ b/drivers/clocksource/sun4i_timer.c @@ -179,6 +179,9 @@ static void __init sun4i_timer_init(struct device_node *node) writel(TIMER_CTL_CLK_SRC(TIMER_CTL_CLK_SRC_OSC24M), timer_base + TIMER_CTL_REG(0)); + /* Make sure timer is stopped before playing with interrupts */ + sun4i_clkevt_time_stop(0); + ret = setup_irq(irq, &sun4i_timer_irq); if (ret) pr_warn("failed to setup irq %d\n", irq); -- cgit v0.10.2 From 85dc6ee1237c8a4a7742e6abab96a20389b7d682 Mon Sep 17 00:00:00 2001 From: Dinh Nguyen Date: Tue, 10 Dec 2013 19:49:18 +0100 Subject: clocksource: dw_apb_timer_of: Fix read_sched_clock The read_sched_clock should return the ~value because the clock is a countdown implementation. read_sched_clock() should be the same as __apbt_read_clocksource(). Signed-off-by: Dinh Nguyen Signed-off-by: Daniel Lezcano diff --git a/drivers/clocksource/dw_apb_timer_of.c b/drivers/clocksource/dw_apb_timer_of.c index 45ba8ae..b29d7cd 100644 --- a/drivers/clocksource/dw_apb_timer_of.c +++ b/drivers/clocksource/dw_apb_timer_of.c @@ -108,7 +108,7 @@ static void __init add_clocksource(struct device_node *source_timer) static u64 read_sched_clock(void) { - return __raw_readl(sched_io_base); + return ~__raw_readl(sched_io_base); } static const struct of_device_id sptimer_ids[] __initconst = { -- cgit v0.10.2 From 9ab4727c1d41e50b67aecde4bf11879560a3ca78 Mon Sep 17 00:00:00 2001 From: Dinh Nguyen Date: Tue, 10 Dec 2013 19:49:18 +0100 Subject: clocksource: dw_apb_timer_of: Fix support for dts binding "snps,dw-apb-timer" In commit 620f5e1cbf (dts: Rename DW APB timer compatible strings), both "snps,dw-apb-timer-sp" and "snps,dw-apb-timer-osc" were deprecated in place of "snps,dw-apb-timer". But the driver also needs to be udpated in order to support this new binding "snps,dw-apb-timer". Signed-off-by: Dinh Nguyen Signed-off-by: Daniel Lezcano diff --git a/drivers/clocksource/dw_apb_timer_of.c b/drivers/clocksource/dw_apb_timer_of.c index b29d7cd..2a2ea27 100644 --- a/drivers/clocksource/dw_apb_timer_of.c +++ b/drivers/clocksource/dw_apb_timer_of.c @@ -113,7 +113,6 @@ static u64 read_sched_clock(void) static const struct of_device_id sptimer_ids[] __initconst = { { .compatible = "picochip,pc3x2-rtc" }, - { .compatible = "snps,dw-apb-timer-sp" }, { /* Sentinel */ }, }; @@ -151,4 +150,6 @@ static void __init dw_apb_timer_init(struct device_node *timer) num_called++; } CLOCKSOURCE_OF_DECLARE(pc3x2_timer, "picochip,pc3x2-timer", dw_apb_timer_init); -CLOCKSOURCE_OF_DECLARE(apb_timer, "snps,dw-apb-timer-osc", dw_apb_timer_init); +CLOCKSOURCE_OF_DECLARE(apb_timer_osc, "snps,dw-apb-timer-osc", dw_apb_timer_init); +CLOCKSOURCE_OF_DECLARE(apb_timer_sp, "snps,dw-apb-timer-sp", dw_apb_timer_init); +CLOCKSOURCE_OF_DECLARE(apb_timer, "snps,dw-apb-timer", dw_apb_timer_init); -- cgit v0.10.2 From 6962d914f317b119e0db7189199b21ec77a4b3e0 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Mon, 9 Dec 2013 14:53:36 +0100 Subject: xhci: Limit the spurious wakeup fix only to HP machines We've got regression reports that my previous fix for spurious wakeups after S5 on HP Haswell machines leads to the automatic reboot at shutdown on some machines. It turned out that the fix for one side triggers another BIOS bug in other side. So, it's exclusive. Since the original S5 wakeups have been confirmed only on HP machines, it'd be safer to apply it only to limited machines. As a wild guess, limiting to machines with HP PCI SSID should suffice. This patch should be backported to kernels as old as 3.12, that contain the commit 638298dc66ea36623dbc2757a24fc2c4ab41b016 "xhci: Fix spurious wakeups after S5 on Haswell". Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=66171 Cc: stable@vger.kernel.org Signed-off-by: Takashi Iwai Signed-off-by: Sarah Sharp Tested-by: Reported-by: Niklas Schnelle Reported-by: Giorgos Reported-by: diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c index b8dffd5..73f5208 100644 --- a/drivers/usb/host/xhci-pci.c +++ b/drivers/usb/host/xhci-pci.c @@ -128,7 +128,12 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) * any other sleep) on Haswell machines with LPT and LPT-LP * with the new Intel BIOS */ - xhci->quirks |= XHCI_SPURIOUS_WAKEUP; + /* Limit the quirk to only known vendors, as this triggers + * yet another BIOS bug on some other machines + * https://bugzilla.kernel.org/show_bug.cgi?id=66171 + */ + if (pdev->subsystem_vendor == PCI_VENDOR_ID_HP) + xhci->quirks |= XHCI_SPURIOUS_WAKEUP; } if (pdev->vendor == PCI_VENDOR_ID_ETRON && pdev->device == PCI_DEVICE_ID_ASROCK_P67) { -- cgit v0.10.2 From 2f5ae4901acc6fdee939718658fa9f59c4f2217b Mon Sep 17 00:00:00 2001 From: Russell King Date: Sun, 27 Oct 2013 15:26:33 +0000 Subject: DRM: Armada: implement lastclose() for fbhelper Call drm_fb_helper_restore_fbdev_mode() upon last close so that in the event of the X server crashing, we have some kind of mode restored. Reviewed-by: Rob Clark Reviewed-by: Thierry Reding Signed-off-by: Russell King diff --git a/drivers/gpu/drm/armada/armada_drm.h b/drivers/gpu/drm/armada/armada_drm.h index eef09ec..a72cae0 100644 --- a/drivers/gpu/drm/armada/armada_drm.h +++ b/drivers/gpu/drm/armada/armada_drm.h @@ -103,6 +103,7 @@ void armada_drm_queue_unref_work(struct drm_device *, extern const struct drm_mode_config_funcs armada_drm_mode_config_funcs; int armada_fbdev_init(struct drm_device *); +void armada_fbdev_lastclose(struct drm_device *); void armada_fbdev_fini(struct drm_device *); int armada_overlay_plane_create(struct drm_device *, unsigned long); diff --git a/drivers/gpu/drm/armada/armada_drv.c b/drivers/gpu/drm/armada/armada_drv.c index 7aede90..15425d2 100644 --- a/drivers/gpu/drm/armada/armada_drv.c +++ b/drivers/gpu/drm/armada/armada_drv.c @@ -321,6 +321,11 @@ static struct drm_ioctl_desc armada_ioctls[] = { DRM_UNLOCKED), }; +static void armada_drm_lastclose(struct drm_device *dev) +{ + armada_fbdev_lastclose(dev); +} + static const struct file_operations armada_drm_fops = { .owner = THIS_MODULE, .llseek = no_llseek, @@ -337,7 +342,7 @@ static struct drm_driver armada_drm_driver = { .open = NULL, .preclose = NULL, .postclose = NULL, - .lastclose = NULL, + .lastclose = armada_drm_lastclose, .unload = armada_drm_unload, .get_vblank_counter = drm_vblank_count, .enable_vblank = armada_drm_enable_vblank, diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c index dd5ea77..743570e 100644 --- a/drivers/gpu/drm/armada/armada_fbdev.c +++ b/drivers/gpu/drm/armada/armada_fbdev.c @@ -177,6 +177,16 @@ int armada_fbdev_init(struct drm_device *dev) return ret; } +void armada_fbdev_lastclose(struct drm_device *dev) +{ + struct armada_private *priv = dev->dev_private; + + drm_modeset_lock_all(dev); + if (priv->fbdev) + drm_fb_helper_restore_fbdev_mode(priv->fbdev); + drm_modeset_unlock_all(dev); +} + void armada_fbdev_fini(struct drm_device *dev) { struct armada_private *priv = dev->dev_private; -- cgit v0.10.2 From 077acbab67a785b5f5ae60e8932fcf6028072359 Mon Sep 17 00:00:00 2001 From: Russell King Date: Sun, 27 Oct 2013 15:35:27 +0000 Subject: DRM: Armada: destroy framebuffer after helper Destroy the framebuffer only after the helper, since the helper may still be referencing the framebufer at this point. Reviewed-by: Rob Clark Reviewed-by: Thierry Reding Signed-off-by: Russell King diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c index 743570e..b348b16 100644 --- a/drivers/gpu/drm/armada/armada_fbdev.c +++ b/drivers/gpu/drm/armada/armada_fbdev.c @@ -202,11 +202,11 @@ void armada_fbdev_fini(struct drm_device *dev) framebuffer_release(info); } + drm_fb_helper_fini(fbh); + if (fbh->fb) fbh->fb->funcs->destroy(fbh->fb); - drm_fb_helper_fini(fbh); - priv->fbdev = NULL; } } -- cgit v0.10.2 From 7513e09596374bb7fbbecfee945fecb5b357c0e9 Mon Sep 17 00:00:00 2001 From: Russell King Date: Wed, 27 Nov 2013 15:46:55 +0000 Subject: DRM: Armada: fix printing of phys_addr_t/dma_addr_t These can be 64-bit quantities, so fix them up appropriately. Reviewed-by: Rob Clark Reviewed-by: Thierry Reding Signed-off-by: Russell King diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c index b348b16..948cb14 100644 --- a/drivers/gpu/drm/armada/armada_fbdev.c +++ b/drivers/gpu/drm/armada/armada_fbdev.c @@ -105,9 +105,9 @@ static int armada_fb_create(struct drm_fb_helper *fbh, drm_fb_helper_fill_fix(info, dfb->fb.pitches[0], dfb->fb.depth); drm_fb_helper_fill_var(info, fbh, sizes->fb_width, sizes->fb_height); - DRM_DEBUG_KMS("allocated %dx%d %dbpp fb: 0x%08x\n", - dfb->fb.width, dfb->fb.height, - dfb->fb.bits_per_pixel, obj->phys_addr); + DRM_DEBUG_KMS("allocated %dx%d %dbpp fb: 0x%08llx\n", + dfb->fb.width, dfb->fb.height, dfb->fb.bits_per_pixel, + (unsigned long long)obj->phys_addr); return 0; diff --git a/drivers/gpu/drm/armada/armada_gem.c b/drivers/gpu/drm/armada/armada_gem.c index 9f2356b..adc7c2d 100644 --- a/drivers/gpu/drm/armada/armada_gem.c +++ b/drivers/gpu/drm/armada/armada_gem.c @@ -172,8 +172,9 @@ armada_gem_linear_back(struct drm_device *dev, struct armada_gem_object *obj) obj->dev_addr = obj->linear->start; } - DRM_DEBUG_DRIVER("obj %p phys %#x dev %#x\n", - obj, obj->phys_addr, obj->dev_addr); + DRM_DEBUG_DRIVER("obj %p phys %#llx dev %#llx\n", obj, + (unsigned long long)obj->phys_addr, + (unsigned long long)obj->dev_addr); return 0; } -- cgit v0.10.2 From 5cd5268806435b6476184430b2806db988c903f0 Mon Sep 17 00:00:00 2001 From: Russell King Date: Sat, 7 Dec 2013 16:28:39 +0000 Subject: DRM: Armada: prime refcounting bug fix Commit 011c2282c74d changed the way refcounting on imported dma_bufs works, and this hadn't been spotted while forward-porting Armada. Reflect the changes in that commit into the Armada driver. Reviewed-by: Rob Clark Reviewed-by: Thierry Reding Signed-off-by: Russell King diff --git a/drivers/gpu/drm/armada/armada_gem.c b/drivers/gpu/drm/armada/armada_gem.c index adc7c2d..887816f 100644 --- a/drivers/gpu/drm/armada/armada_gem.c +++ b/drivers/gpu/drm/armada/armada_gem.c @@ -558,7 +558,6 @@ armada_gem_prime_import(struct drm_device *dev, struct dma_buf *buf) * refcount on the gem object itself. */ drm_gem_object_reference(obj); - dma_buf_put(buf); return obj; } } @@ -574,6 +573,7 @@ armada_gem_prime_import(struct drm_device *dev, struct dma_buf *buf) } dobj->obj.import_attach = attach; + get_dma_buf(buf); /* * Don't call dma_buf_map_attachment() here - it maps the -- cgit v0.10.2 From c1b1731d2002b93857810aa5da856b43f46bb470 Mon Sep 17 00:00:00 2001 From: Sachin Kamat Date: Fri, 6 Dec 2013 17:51:18 +0530 Subject: drivers: phy: Fix memory leak 'phy' was not being freed upon error in one of the cases. Adjust the 'goto's to fix this. Signed-off-by: Sachin Kamat Signed-off-by: Kishon Vijay Abraham I Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/phy/phy-core.c b/drivers/phy/phy-core.c index 03cf8fb..712b358 100644 --- a/drivers/phy/phy-core.c +++ b/drivers/phy/phy-core.c @@ -453,7 +453,7 @@ struct phy *phy_create(struct device *dev, const struct phy_ops *ops, if (id < 0) { dev_err(dev, "unable to get id\n"); ret = id; - goto err0; + goto err1; } device_initialize(&phy->dev); @@ -468,11 +468,11 @@ struct phy *phy_create(struct device *dev, const struct phy_ops *ops, ret = dev_set_name(&phy->dev, "phy-%s.%d", dev_name(dev), id); if (ret) - goto err1; + goto err2; ret = device_add(&phy->dev); if (ret) - goto err1; + goto err2; if (pm_runtime_enabled(dev)) { pm_runtime_enable(&phy->dev); @@ -481,11 +481,11 @@ struct phy *phy_create(struct device *dev, const struct phy_ops *ops, return phy; -err1: +err2: ida_remove(&phy_ida, phy->id); put_device(&phy->dev); +err1: kfree(phy); - err0: return ERR_PTR(ret); } -- cgit v0.10.2 From 52797d293287f5a98bd686616527b102b077ac39 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 6 Dec 2013 17:51:19 +0530 Subject: drivers: phy: tweaks to phy_create() If this was called with a NULL "dev" then it lead to a NULL dereference when we called dev_WARN(). I have changed it to WARN_ON() so that we get a stack dump and can fix the caller. The rest of this patch is just cleanup like returning directly instead of having do-nothing gotos. Using descriptive labels instead of GW-BASIC style "err0" and "err1". I also flipped the order of put_device() and ida_remove() so they are a mirror reflection of the order they were allocated. Signed-off-by: Dan Carpenter Signed-off-by: Kishon Vijay Abraham I Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/phy/phy-core.c b/drivers/phy/phy-core.c index 712b358..58e0e97 100644 --- a/drivers/phy/phy-core.c +++ b/drivers/phy/phy-core.c @@ -437,23 +437,18 @@ struct phy *phy_create(struct device *dev, const struct phy_ops *ops, int id; struct phy *phy; - if (!dev) { - dev_WARN(dev, "no device provided for PHY\n"); - ret = -EINVAL; - goto err0; - } + if (WARN_ON(!dev)) + return ERR_PTR(-EINVAL); phy = kzalloc(sizeof(*phy), GFP_KERNEL); - if (!phy) { - ret = -ENOMEM; - goto err0; - } + if (!phy) + return ERR_PTR(-ENOMEM); id = ida_simple_get(&phy_ida, 0, 0, GFP_KERNEL); if (id < 0) { dev_err(dev, "unable to get id\n"); ret = id; - goto err1; + goto free_phy; } device_initialize(&phy->dev); @@ -468,11 +463,11 @@ struct phy *phy_create(struct device *dev, const struct phy_ops *ops, ret = dev_set_name(&phy->dev, "phy-%s.%d", dev_name(dev), id); if (ret) - goto err2; + goto put_dev; ret = device_add(&phy->dev); if (ret) - goto err2; + goto put_dev; if (pm_runtime_enabled(dev)) { pm_runtime_enable(&phy->dev); @@ -481,12 +476,11 @@ struct phy *phy_create(struct device *dev, const struct phy_ops *ops, return phy; -err2: - ida_remove(&phy_ida, phy->id); +put_dev: put_device(&phy->dev); -err1: + ida_remove(&phy_ida, phy->id); +free_phy: kfree(phy); -err0: return ERR_PTR(ret); } EXPORT_SYMBOL_GPL(phy_create); -- cgit v0.10.2 From 8820784203ac24279be23a6e3d85d216c46d65ee Mon Sep 17 00:00:00 2001 From: Kishon Vijay Abraham I Date: Fri, 6 Dec 2013 17:51:20 +0530 Subject: phy: kconfig: add depends on "USB_PHY" to OMAP_USB2 and TWL4030_USB Fixes warning: (OMAP_USB2 && TWL4030_USB) selects USB_PHY which has unmet direct dependencies (USB_SUPPORT) that shows up while disabling USB_SUPPORT from menuconfig. Reported-by: Russell King Signed-off-by: Kishon Vijay Abraham I Acked-by: Felipe Balbi Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/phy/Kconfig b/drivers/phy/Kconfig index a344f3d..330ef2d 100644 --- a/drivers/phy/Kconfig +++ b/drivers/phy/Kconfig @@ -24,8 +24,8 @@ config PHY_EXYNOS_MIPI_VIDEO config OMAP_USB2 tristate "OMAP USB2 PHY Driver" depends on ARCH_OMAP2PLUS + depends on USB_PHY select GENERIC_PHY - select USB_PHY select OMAP_CONTROL_USB help Enable this to support the transceiver that is part of SOC. This @@ -36,8 +36,8 @@ config OMAP_USB2 config TWL4030_USB tristate "TWL4030 USB Transceiver Driver" depends on TWL4030_CORE && REGULATOR_TWL4030 && USB_MUSB_OMAP2PLUS + depends on USB_PHY select GENERIC_PHY - select USB_PHY help Enable this to support the USB OTG transceiver on TWL4030 family chips (including the TWL5030 and TPS659x0 devices). -- cgit v0.10.2 From f5f972102d5c12729f0a35fce266b580aaa03f66 Mon Sep 17 00:00:00 2001 From: Scott Wood Date: Fri, 22 Nov 2013 15:52:29 -0600 Subject: powerpc/kvm/booke: Fix build break due to stack frame size warning Commit ce11e48b7fdd256ec68b932a89b397a790566031 ("KVM: PPC: E500: Add userspace debug stub support") added "struct thread_struct" to the stack of kvmppc_vcpu_run(). thread_struct is 1152 bytes on my build, compared to 48 bytes for the recently-introduced "struct debug_reg". Use the latter instead. This fixes the following error: cc1: warnings being treated as errors arch/powerpc/kvm/booke.c: In function 'kvmppc_vcpu_run': arch/powerpc/kvm/booke.c:760:1: error: the frame size of 1424 bytes is larger than 1024 bytes make[2]: *** [arch/powerpc/kvm/booke.o] Error 1 make[1]: *** [arch/powerpc/kvm] Error 2 make[1]: *** Waiting for unfinished jobs.... Signed-off-by: Scott Wood Signed-off-by: Alexander Graf diff --git a/arch/powerpc/include/asm/switch_to.h b/arch/powerpc/include/asm/switch_to.h index 9ee1261..aace905 100644 --- a/arch/powerpc/include/asm/switch_to.h +++ b/arch/powerpc/include/asm/switch_to.h @@ -35,7 +35,7 @@ extern void giveup_vsx(struct task_struct *); extern void enable_kernel_spe(void); extern void giveup_spe(struct task_struct *); extern void load_up_spe(struct task_struct *); -extern void switch_booke_debug_regs(struct thread_struct *new_thread); +extern void switch_booke_debug_regs(struct debug_reg *new_debug); #ifndef CONFIG_SMP extern void discard_lazy_cpu_state(void); diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 75c2d10..83530af 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -339,7 +339,7 @@ static void set_debug_reg_defaults(struct thread_struct *thread) #endif } -static void prime_debug_regs(struct thread_struct *thread) +static void prime_debug_regs(struct debug_reg *debug) { /* * We could have inherited MSR_DE from userspace, since @@ -348,22 +348,22 @@ static void prime_debug_regs(struct thread_struct *thread) */ mtmsr(mfmsr() & ~MSR_DE); - mtspr(SPRN_IAC1, thread->debug.iac1); - mtspr(SPRN_IAC2, thread->debug.iac2); + mtspr(SPRN_IAC1, debug->iac1); + mtspr(SPRN_IAC2, debug->iac2); #if CONFIG_PPC_ADV_DEBUG_IACS > 2 - mtspr(SPRN_IAC3, thread->debug.iac3); - mtspr(SPRN_IAC4, thread->debug.iac4); + mtspr(SPRN_IAC3, debug->iac3); + mtspr(SPRN_IAC4, debug->iac4); #endif - mtspr(SPRN_DAC1, thread->debug.dac1); - mtspr(SPRN_DAC2, thread->debug.dac2); + mtspr(SPRN_DAC1, debug->dac1); + mtspr(SPRN_DAC2, debug->dac2); #if CONFIG_PPC_ADV_DEBUG_DVCS > 0 - mtspr(SPRN_DVC1, thread->debug.dvc1); - mtspr(SPRN_DVC2, thread->debug.dvc2); + mtspr(SPRN_DVC1, debug->dvc1); + mtspr(SPRN_DVC2, debug->dvc2); #endif - mtspr(SPRN_DBCR0, thread->debug.dbcr0); - mtspr(SPRN_DBCR1, thread->debug.dbcr1); + mtspr(SPRN_DBCR0, debug->dbcr0); + mtspr(SPRN_DBCR1, debug->dbcr1); #ifdef CONFIG_BOOKE - mtspr(SPRN_DBCR2, thread->debug.dbcr2); + mtspr(SPRN_DBCR2, debug->dbcr2); #endif } /* @@ -371,11 +371,11 @@ static void prime_debug_regs(struct thread_struct *thread) * debug registers, set the debug registers from the values * stored in the new thread. */ -void switch_booke_debug_regs(struct thread_struct *new_thread) +void switch_booke_debug_regs(struct debug_reg *new_debug) { if ((current->thread.debug.dbcr0 & DBCR0_IDM) - || (new_thread->debug.dbcr0 & DBCR0_IDM)) - prime_debug_regs(new_thread); + || (new_debug->dbcr0 & DBCR0_IDM)) + prime_debug_regs(new_debug); } EXPORT_SYMBOL_GPL(switch_booke_debug_regs); #else /* !CONFIG_PPC_ADV_DEBUG_REGS */ @@ -683,7 +683,7 @@ struct task_struct *__switch_to(struct task_struct *prev, #endif /* CONFIG_SMP */ #ifdef CONFIG_PPC_ADV_DEBUG_REGS - switch_booke_debug_regs(&new->thread); + switch_booke_debug_regs(&new->thread.debug); #else /* * For PPC_BOOK3S_64, we use the hw-breakpoint interfaces that would diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c index 53e65a2..0591e05 100644 --- a/arch/powerpc/kvm/booke.c +++ b/arch/powerpc/kvm/booke.c @@ -681,7 +681,7 @@ int kvmppc_core_check_requests(struct kvm_vcpu *vcpu) int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) { int ret, s; - struct thread_struct thread; + struct debug_reg debug; #ifdef CONFIG_PPC_FPU struct thread_fp_state fp; int fpexc_mode; @@ -723,9 +723,9 @@ int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) #endif /* Switch to guest debug context */ - thread.debug = vcpu->arch.shadow_dbg_reg; - switch_booke_debug_regs(&thread); - thread.debug = current->thread.debug; + debug = vcpu->arch.shadow_dbg_reg; + switch_booke_debug_regs(&debug); + debug = current->thread.debug; current->thread.debug = vcpu->arch.shadow_dbg_reg; kvmppc_fix_ee_before_entry(); @@ -736,8 +736,8 @@ int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) We also get here with interrupts enabled. */ /* Switch back to user space debug context */ - switch_booke_debug_regs(&thread); - current->thread.debug = thread.debug; + switch_booke_debug_regs(&debug); + current->thread.debug = debug; #ifdef CONFIG_PPC_FPU kvmppc_save_guest_fp(vcpu); -- cgit v0.10.2 From f01b3926ee645974f549f4a6921268142047717c Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sun, 8 Dec 2013 16:52:31 +0000 Subject: netfilter: SYNPROXY target: restrict to INPUT/FORWARD Fix a crash in synproxy_send_tcp() when using the SYNPROXY target in the PREROUTING chain caused by missing routing information. Reported-by: Nicki P. Signed-off-by: Patrick McHardy Signed-off-by: Pablo Neira Ayuso diff --git a/net/ipv4/netfilter/ipt_SYNPROXY.c b/net/ipv4/netfilter/ipt_SYNPROXY.c index f13bd91..a313c3f 100644 --- a/net/ipv4/netfilter/ipt_SYNPROXY.c +++ b/net/ipv4/netfilter/ipt_SYNPROXY.c @@ -423,6 +423,7 @@ static void synproxy_tg4_destroy(const struct xt_tgdtor_param *par) static struct xt_target synproxy_tg4_reg __read_mostly = { .name = "SYNPROXY", .family = NFPROTO_IPV4, + .hooks = (1 << NF_INET_LOCAL_IN) | (1 << NF_INET_FORWARD), .target = synproxy_tg4, .targetsize = sizeof(struct xt_synproxy_info), .checkentry = synproxy_tg4_check, diff --git a/net/ipv6/netfilter/ip6t_SYNPROXY.c b/net/ipv6/netfilter/ip6t_SYNPROXY.c index f78f41a..a0d1727 100644 --- a/net/ipv6/netfilter/ip6t_SYNPROXY.c +++ b/net/ipv6/netfilter/ip6t_SYNPROXY.c @@ -446,6 +446,7 @@ static void synproxy_tg6_destroy(const struct xt_tgdtor_param *par) static struct xt_target synproxy_tg6_reg __read_mostly = { .name = "SYNPROXY", .family = NFPROTO_IPV6, + .hooks = (1 << NF_INET_LOCAL_IN) | (1 << NF_INET_FORWARD), .target = synproxy_tg6, .targetsize = sizeof(struct xt_synproxy_info), .checkentry = synproxy_tg6_check, -- cgit v0.10.2 From 6c719faca2aceca72f1bf5b1645c1734ed3e9447 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 10 Dec 2013 13:20:59 +0100 Subject: drm/i915: don't update the dri1 breadcrumb with modesetting The update is horribly racy since it doesn't protect at all against concurrent closing of the master fd. And it can't really since that requires us to grab a mutex. Instead of jumping through hoops and offloading this to a worker thread just block this bit of code for the modesetting driver. Note that the race is fairly easy to hit since we call the breadcrumb function for any interrupt. So the vblank interrupt (which usually keeps going for a bit) is enough. But even if we'd block this and only update the breadcrumb for user interrupts from the CS we could hit this race with kms/gem userspace: If a non-master is waiting somewhere (and hence has interrupts enabled) and the master closes its fd (probably due to crashing). v2: Add a code comment to explain why fixing this for real isn't really worth it. Also improve the commit message a bit. v3: Fix the spelling in the comment. Reported-by: Eugene Shatokhin Cc: Eugene Shatokhin Cc: stable@vger.kernel.org Acked-by: Chris Wilson Tested-by: Eugene Shatokhin Signed-off-by: Daniel Vetter diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 197db09..5c64842 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -83,6 +83,14 @@ void i915_update_dri1_breadcrumb(struct drm_device *dev) drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_master_private *master_priv; + /* + * The dri breadcrumb update races against the drm master disappearing. + * Instead of trying to fix this (this is by far not the only ums issue) + * just don't do the update in kms mode. + */ + if (drm_core_check_feature(dev, DRIVER_MODESET)) + return; + if (dev->primary->master) { master_priv = dev->primary->master->driver_priv; if (master_priv->sarea_priv) -- cgit v0.10.2 From 8e8339a3a1069141985daaa2521ba304509ddecd Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 11 Dec 2013 11:09:53 +0100 Subject: sched: Initialize power_orig for overlapping groups Yinghai reported that he saw a /0 in sg_capacity on his EX parts. Make sure to always initialize power_orig now that we actually use it. Ideally build_sched_domains() -> init_sched_groups_power() would also initialize this; but for some yet unexplained reason some setups seem to miss updates there. Reported-by: Yinghai Lu Tested-by: Yinghai Lu Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-l8ng2m9uml6fhibln8wqpom7@git.kernel.org Signed-off-by: Ingo Molnar diff --git a/kernel/sched/core.c b/kernel/sched/core.c index e85cda2..19af58f 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5112,6 +5112,7 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) * die on a /0 trap. */ sg->sgp->power = SCHED_POWER_SCALE * cpumask_weight(sg_span); + sg->sgp->power_orig = sg->sgp->power; /* * Make sure the first group of this domain contains the -- cgit v0.10.2 From ba1f14fbe70965ae0fb1655a5275a62723f65b77 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 28 Nov 2013 14:26:41 +0100 Subject: sched: Remove PREEMPT_NEED_RESCHED from generic code While hunting a preemption issue with Alexander, Ben noticed that the currently generic PREEMPT_NEED_RESCHED stuff is horribly broken for load-store architectures. We currently rely on the IPI to fold TIF_NEED_RESCHED into PREEMPT_NEED_RESCHED, but when this IPI lands while we already have a load for the preempt-count but before the store, the store will erase the PREEMPT_NEED_RESCHED change. The current preempt-count only works on load-store archs because interrupts are assumed to be completely balanced wrt their preempt_count fiddling; the previous preempt_count load will match the preempt_count state after the interrupt and therefore nothing gets lost. This patch removes the PREEMPT_NEED_RESCHED usage from generic code and pushes it into x86 arch code; the generic code goes back to relying on TIF_NEED_RESCHED. Boot tested on x86_64 and compile tested on ppc64. Reported-by: Benjamin Herrenschmidt Reported-and-Tested-by: Alexander Graf Signed-off-by: Peter Zijlstra Cc: Linus Torvalds Link: http://lkml.kernel.org/r/20131128132641.GP10022@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/preempt.h b/arch/x86/include/asm/preempt.h index 8729723..c8b0519 100644 --- a/arch/x86/include/asm/preempt.h +++ b/arch/x86/include/asm/preempt.h @@ -8,6 +8,12 @@ DECLARE_PER_CPU(int, __preempt_count); /* + * We use the PREEMPT_NEED_RESCHED bit as an inverted NEED_RESCHED such + * that a decrement hitting 0 means we can and should reschedule. + */ +#define PREEMPT_ENABLED (0 + PREEMPT_NEED_RESCHED) + +/* * We mask the PREEMPT_NEED_RESCHED bit so as not to confuse all current users * that think a non-zero value indicates we cannot preempt. */ @@ -74,6 +80,11 @@ static __always_inline void __preempt_count_sub(int val) __this_cpu_add_4(__preempt_count, -val); } +/* + * Because we keep PREEMPT_NEED_RESCHED set when we do _not_ need to reschedule + * a decrement which hits zero means we have no preempt_count and should + * reschedule. + */ static __always_inline bool __preempt_count_dec_and_test(void) { GEN_UNARY_RMWcc("decl", __preempt_count, __percpu_arg(0), "e"); diff --git a/include/asm-generic/preempt.h b/include/asm-generic/preempt.h index ddf2b42..1cd3f5d 100644 --- a/include/asm-generic/preempt.h +++ b/include/asm-generic/preempt.h @@ -3,13 +3,11 @@ #include -/* - * We mask the PREEMPT_NEED_RESCHED bit so as not to confuse all current users - * that think a non-zero value indicates we cannot preempt. - */ +#define PREEMPT_ENABLED (0) + static __always_inline int preempt_count(void) { - return current_thread_info()->preempt_count & ~PREEMPT_NEED_RESCHED; + return current_thread_info()->preempt_count; } static __always_inline int *preempt_count_ptr(void) @@ -17,11 +15,6 @@ static __always_inline int *preempt_count_ptr(void) return ¤t_thread_info()->preempt_count; } -/* - * We now loose PREEMPT_NEED_RESCHED and cause an extra reschedule; however the - * alternative is loosing a reschedule. Better schedule too often -- also this - * should be a very rare operation. - */ static __always_inline void preempt_count_set(int pc) { *preempt_count_ptr() = pc; @@ -41,28 +34,17 @@ static __always_inline void preempt_count_set(int pc) task_thread_info(p)->preempt_count = PREEMPT_ENABLED; \ } while (0) -/* - * We fold the NEED_RESCHED bit into the preempt count such that - * preempt_enable() can decrement and test for needing to reschedule with a - * single instruction. - * - * We invert the actual bit, so that when the decrement hits 0 we know we both - * need to resched (the bit is cleared) and can resched (no preempt count). - */ - static __always_inline void set_preempt_need_resched(void) { - *preempt_count_ptr() &= ~PREEMPT_NEED_RESCHED; } static __always_inline void clear_preempt_need_resched(void) { - *preempt_count_ptr() |= PREEMPT_NEED_RESCHED; } static __always_inline bool test_preempt_need_resched(void) { - return !(*preempt_count_ptr() & PREEMPT_NEED_RESCHED); + return false; } /* @@ -81,7 +63,12 @@ static __always_inline void __preempt_count_sub(int val) static __always_inline bool __preempt_count_dec_and_test(void) { - return !--*preempt_count_ptr(); + /* + * Because of load-store architectures cannot do per-cpu atomic + * operations; we cannot use PREEMPT_NEED_RESCHED because it might get + * lost. + */ + return !--*preempt_count_ptr() && tif_need_resched(); } /* @@ -89,7 +76,7 @@ static __always_inline bool __preempt_count_dec_and_test(void) */ static __always_inline bool should_resched(void) { - return unlikely(!*preempt_count_ptr()); + return unlikely(!preempt_count() && tif_need_resched()); } #ifdef CONFIG_PREEMPT diff --git a/include/linux/sched.h b/include/linux/sched.h index 768b037..96d674b 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -440,8 +440,6 @@ struct task_cputime { .sum_exec_runtime = 0, \ } -#define PREEMPT_ENABLED (PREEMPT_NEED_RESCHED) - #ifdef CONFIG_PREEMPT_COUNT #define PREEMPT_DISABLED (1 + PREEMPT_ENABLED) #else -- cgit v0.10.2 From be5e610c0fd6ef772cafb9e0bd4128134804aef3 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 18 Nov 2013 18:27:06 +0100 Subject: math64: Add mul_u64_u32_shr() Introduce mul_u64_u32_shr() as proposed by Andy a while back; it allows using 64x64->128 muls on 64bit archs and recent GCC which defines __SIZEOF_INT128__ and __int128. (This new method will be used by the scheduler.) Signed-off-by: Peter Zijlstra Cc: fweisbec@gmail.com Cc: Andy Lutomirski Cc: Linus Torvalds Link: http://lkml.kernel.org/n/tip-hxjoeuzmrcaumR0uZwjpe2pv@git.kernel.org Signed-off-by: Ingo Molnar diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index e903c71..0952ecd 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -26,6 +26,7 @@ config X86 select HAVE_AOUT if X86_32 select HAVE_UNSTABLE_SCHED_CLOCK select ARCH_SUPPORTS_NUMA_BALANCING + select ARCH_SUPPORTS_INT128 if X86_64 select ARCH_WANTS_PROT_NUMA_PROT_NONE select HAVE_IDE select HAVE_OPROFILE diff --git a/include/linux/math64.h b/include/linux/math64.h index 69ed5f5..c45c089 100644 --- a/include/linux/math64.h +++ b/include/linux/math64.h @@ -133,4 +133,34 @@ __iter_div_u64_rem(u64 dividend, u32 divisor, u64 *remainder) return ret; } +#if defined(CONFIG_ARCH_SUPPORTS_INT128) && defined(__SIZEOF_INT128__) + +#ifndef mul_u64_u32_shr +static inline u64 mul_u64_u32_shr(u64 a, u32 mul, unsigned int shift) +{ + return (u64)(((unsigned __int128)a * mul) >> shift); +} +#endif /* mul_u64_u32_shr */ + +#else + +#ifndef mul_u64_u32_shr +static inline u64 mul_u64_u32_shr(u64 a, u32 mul, unsigned int shift) +{ + u32 ah, al; + u64 ret; + + al = a; + ah = a >> 32; + + ret = ((u64)al * mul) >> shift; + if (ah) + ret += ((u64)ah * mul) << (32 - shift); + + return ret; +} +#endif /* mul_u64_u32_shr */ + +#endif + #endif /* _LINUX_MATH64_H */ diff --git a/init/Kconfig b/init/Kconfig index 79383d3..4e5d96a 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -809,6 +809,12 @@ config GENERIC_SCHED_CLOCK config ARCH_SUPPORTS_NUMA_BALANCING bool +# +# For architectures that know their GCC __int128 support is sound +# +config ARCH_SUPPORTS_INT128 + bool + # For architectures that (ab)use NUMA to represent different memory regions # all cpu-local but of different latencies, such as SuperH. # -- cgit v0.10.2 From 9dbdb155532395ba000c5d5d187658b0e17e529f Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 18 Nov 2013 18:27:06 +0100 Subject: sched/fair: Rework sched_fair time accounting Christian suffers from a bad BIOS that wrecks his i5's TSC sync. This results in him occasionally seeing time going backwards - which crashes the scheduler ... Most of our time accounting can actually handle that except the most common one; the tick time update of sched_fair. There is a further problem with that code; previously we assumed that because we get a tick every TICK_NSEC our time delta could never exceed 32bits and math was simpler. However, ever since Frederic managed to get NO_HZ_FULL merged; this is no longer the case since now a task can run for a long time indeed without getting a tick. It only takes about ~4.2 seconds to overflow our u32 in nanoseconds. This means we not only need to better deal with time going backwards; but also means we need to be able to deal with large deltas. This patch reworks the entire code and uses mul_u64_u32_shr() as proposed by Andy a long while ago. We express our virtual time scale factor in a u32 multiplier and shift right and the 32bit mul_u64_u32_shr() implementation reduces to a single 32x32->64 multiply if the time delta is still short (common case). For 64bit a 64x64->128 multiply can be used if ARCH_SUPPORTS_INT128. Reported-and-Tested-by: Christian Engelmayer Signed-off-by: Peter Zijlstra Cc: fweisbec@gmail.com Cc: Paul Turner Cc: Stanislaw Gruszka Cc: Andy Lutomirski Cc: Linus Torvalds Cc: Andrew Morton Link: http://lkml.kernel.org/r/20131118172706.GI3866@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar diff --git a/include/linux/sched.h b/include/linux/sched.h index 96d674b..53f97eb 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -930,7 +930,8 @@ struct pipe_inode_info; struct uts_namespace; struct load_weight { - unsigned long weight, inv_weight; + unsigned long weight; + u32 inv_weight; }; struct sched_avg { diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index fd773ad..9030da7 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -178,59 +178,61 @@ void sched_init_granularity(void) update_sysctl(); } -#if BITS_PER_LONG == 32 -# define WMULT_CONST (~0UL) -#else -# define WMULT_CONST (1UL << 32) -#endif - +#define WMULT_CONST (~0U) #define WMULT_SHIFT 32 -/* - * Shift right and round: - */ -#define SRR(x, y) (((x) + (1UL << ((y) - 1))) >> (y)) +static void __update_inv_weight(struct load_weight *lw) +{ + unsigned long w; + + if (likely(lw->inv_weight)) + return; + + w = scale_load_down(lw->weight); + + if (BITS_PER_LONG > 32 && unlikely(w >= WMULT_CONST)) + lw->inv_weight = 1; + else if (unlikely(!w)) + lw->inv_weight = WMULT_CONST; + else + lw->inv_weight = WMULT_CONST / w; +} /* - * delta *= weight / lw + * delta_exec * weight / lw.weight + * OR + * (delta_exec * (weight * lw->inv_weight)) >> WMULT_SHIFT + * + * Either weight := NICE_0_LOAD and lw \e prio_to_wmult[], in which case + * we're guaranteed shift stays positive because inv_weight is guaranteed to + * fit 32 bits, and NICE_0_LOAD gives another 10 bits; therefore shift >= 22. + * + * Or, weight =< lw.weight (because lw.weight is the runqueue weight), thus + * weight/lw.weight <= 1, and therefore our shift will also be positive. */ -static unsigned long -calc_delta_mine(unsigned long delta_exec, unsigned long weight, - struct load_weight *lw) +static u64 __calc_delta(u64 delta_exec, unsigned long weight, struct load_weight *lw) { - u64 tmp; + u64 fact = scale_load_down(weight); + int shift = WMULT_SHIFT; - /* - * weight can be less than 2^SCHED_LOAD_RESOLUTION for task group sched - * entities since MIN_SHARES = 2. Treat weight as 1 if less than - * 2^SCHED_LOAD_RESOLUTION. - */ - if (likely(weight > (1UL << SCHED_LOAD_RESOLUTION))) - tmp = (u64)delta_exec * scale_load_down(weight); - else - tmp = (u64)delta_exec; + __update_inv_weight(lw); - if (!lw->inv_weight) { - unsigned long w = scale_load_down(lw->weight); - - if (BITS_PER_LONG > 32 && unlikely(w >= WMULT_CONST)) - lw->inv_weight = 1; - else if (unlikely(!w)) - lw->inv_weight = WMULT_CONST; - else - lw->inv_weight = WMULT_CONST / w; + if (unlikely(fact >> 32)) { + while (fact >> 32) { + fact >>= 1; + shift--; + } } - /* - * Check whether we'd overflow the 64-bit multiplication: - */ - if (unlikely(tmp > WMULT_CONST)) - tmp = SRR(SRR(tmp, WMULT_SHIFT/2) * lw->inv_weight, - WMULT_SHIFT/2); - else - tmp = SRR(tmp * lw->inv_weight, WMULT_SHIFT); + /* hint to use a 32x32->64 mul */ + fact = (u64)(u32)fact * lw->inv_weight; + + while (fact >> 32) { + fact >>= 1; + shift--; + } - return (unsigned long)min(tmp, (u64)(unsigned long)LONG_MAX); + return mul_u64_u32_shr(delta_exec, fact, shift); } @@ -443,7 +445,7 @@ find_matching_se(struct sched_entity **se, struct sched_entity **pse) #endif /* CONFIG_FAIR_GROUP_SCHED */ static __always_inline -void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, unsigned long delta_exec); +void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec); /************************************************************** * Scheduling class tree data structure manipulation methods: @@ -612,11 +614,10 @@ int sched_proc_update_handler(struct ctl_table *table, int write, /* * delta /= w */ -static inline unsigned long -calc_delta_fair(unsigned long delta, struct sched_entity *se) +static inline u64 calc_delta_fair(u64 delta, struct sched_entity *se) { if (unlikely(se->load.weight != NICE_0_LOAD)) - delta = calc_delta_mine(delta, NICE_0_LOAD, &se->load); + delta = __calc_delta(delta, NICE_0_LOAD, &se->load); return delta; } @@ -665,7 +666,7 @@ static u64 sched_slice(struct cfs_rq *cfs_rq, struct sched_entity *se) update_load_add(&lw, se->load.weight); load = &lw; } - slice = calc_delta_mine(slice, se->load.weight, load); + slice = __calc_delta(slice, se->load.weight, load); } return slice; } @@ -703,47 +704,32 @@ void init_task_runnable_average(struct task_struct *p) #endif /* - * Update the current task's runtime statistics. Skip current tasks that - * are not in our scheduling class. + * Update the current task's runtime statistics. */ -static inline void -__update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr, - unsigned long delta_exec) -{ - unsigned long delta_exec_weighted; - - schedstat_set(curr->statistics.exec_max, - max((u64)delta_exec, curr->statistics.exec_max)); - - curr->sum_exec_runtime += delta_exec; - schedstat_add(cfs_rq, exec_clock, delta_exec); - delta_exec_weighted = calc_delta_fair(delta_exec, curr); - - curr->vruntime += delta_exec_weighted; - update_min_vruntime(cfs_rq); -} - static void update_curr(struct cfs_rq *cfs_rq) { struct sched_entity *curr = cfs_rq->curr; u64 now = rq_clock_task(rq_of(cfs_rq)); - unsigned long delta_exec; + u64 delta_exec; if (unlikely(!curr)) return; - /* - * Get the amount of time the current task was running - * since the last time we changed load (this cannot - * overflow on 32 bits): - */ - delta_exec = (unsigned long)(now - curr->exec_start); - if (!delta_exec) + delta_exec = now - curr->exec_start; + if (unlikely((s64)delta_exec <= 0)) return; - __update_curr(cfs_rq, curr, delta_exec); curr->exec_start = now; + schedstat_set(curr->statistics.exec_max, + max(delta_exec, curr->statistics.exec_max)); + + curr->sum_exec_runtime += delta_exec; + schedstat_add(cfs_rq, exec_clock, delta_exec); + + curr->vruntime += calc_delta_fair(delta_exec, curr); + update_min_vruntime(cfs_rq); + if (entity_is_task(curr)) { struct task_struct *curtask = task_of(curr); @@ -3015,8 +3001,7 @@ static void expire_cfs_rq_runtime(struct cfs_rq *cfs_rq) } } -static void __account_cfs_rq_runtime(struct cfs_rq *cfs_rq, - unsigned long delta_exec) +static void __account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec) { /* dock delta_exec before expiring quota (as it could span periods) */ cfs_rq->runtime_remaining -= delta_exec; @@ -3034,7 +3019,7 @@ static void __account_cfs_rq_runtime(struct cfs_rq *cfs_rq, } static __always_inline -void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, unsigned long delta_exec) +void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec) { if (!cfs_bandwidth_used() || !cfs_rq->runtime_enabled) return; @@ -3574,8 +3559,7 @@ static inline u64 cfs_rq_clock_task(struct cfs_rq *cfs_rq) return rq_clock_task(rq_of(cfs_rq)); } -static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, - unsigned long delta_exec) {} +static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec) {} static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq) {} static void check_enqueue_throttle(struct cfs_rq *cfs_rq) {} static __always_inline void return_cfs_rq_runtime(struct cfs_rq *cfs_rq) {} -- cgit v0.10.2 From 02ab71cdae248533620abefa1d46097581457110 Mon Sep 17 00:00:00 2001 From: Stefano Stabellini Date: Mon, 9 Dec 2013 15:55:11 +0000 Subject: xen/arm64: do not call the swiotlb functions twice On arm64 the dma_map_ops implementation is based on the swiotlb. swiotlb-xen, used by default in dom0 on Xen, is also based on the swiotlb. Avoid calling into the default arm64 dma_map_ops functions from xen_dma_map_page, xen_dma_unmap_page, xen_dma_sync_single_for_cpu, and xen_dma_sync_single_for_device otherwise we end up calling into the swiotlb twice. When arm64 gets a non-swiotlb based implementation of dma_map_ops, we'll probably have to reintroduce dma_map_ops calls in page-coherent.h. Signed-off-by: Stefano Stabellini CC: catalin.marinas@arm.com CC: Will.Deacon@arm.com CC: Ian.Campbell@citrix.com diff --git a/arch/arm64/include/asm/xen/page-coherent.h b/arch/arm64/include/asm/xen/page-coherent.h index 2820f1a..dde3fc9 100644 --- a/arch/arm64/include/asm/xen/page-coherent.h +++ b/arch/arm64/include/asm/xen/page-coherent.h @@ -23,25 +23,21 @@ static inline void xen_dma_map_page(struct device *hwdev, struct page *page, unsigned long offset, size_t size, enum dma_data_direction dir, struct dma_attrs *attrs) { - __generic_dma_ops(hwdev)->map_page(hwdev, page, offset, size, dir, attrs); } static inline void xen_dma_unmap_page(struct device *hwdev, dma_addr_t handle, size_t size, enum dma_data_direction dir, struct dma_attrs *attrs) { - __generic_dma_ops(hwdev)->unmap_page(hwdev, handle, size, dir, attrs); } static inline void xen_dma_sync_single_for_cpu(struct device *hwdev, dma_addr_t handle, size_t size, enum dma_data_direction dir) { - __generic_dma_ops(hwdev)->sync_single_for_cpu(hwdev, handle, size, dir); } static inline void xen_dma_sync_single_for_device(struct device *hwdev, dma_addr_t handle, size_t size, enum dma_data_direction dir) { - __generic_dma_ops(hwdev)->sync_single_for_device(hwdev, handle, size, dir); } #endif /* _ASM_ARM64_XEN_PAGE_COHERENT_H */ -- cgit v0.10.2 From a7892f32cc3534d4cc0e64b245fbf47a8e364652 Mon Sep 17 00:00:00 2001 From: Ian Campbell Date: Wed, 11 Dec 2013 17:02:27 +0000 Subject: arm: xen: foreign mapping PTEs are special. These mappings are in fact special and require special handling in privcmd, which already exists. Failure to mark the PTE as special on arm64 causes all sorts of bad PTE fun. e.g. e.g.: BUG: Bad page map in process xl pte:e0004077b33f53 pmd:4079575003 page:ffffffbce1a2f328 count:1 mapcount:-1 mapping: (null) index:0x0 page flags: 0x4000000000000014(referenced|dirty) addr:0000007fb5259000 vm_flags:040644fa anon_vma: (null) mapping:ffffffc03a6fda58 index:0 vma->vm_ops->fault: privcmd_fault+0x0/0x38 vma->vm_file->f_op->mmap: privcmd_mmap+0x0/0x2c CPU: 0 PID: 2657 Comm: xl Not tainted 3.12.0+ #102 Call trace: [] dump_backtrace+0x0/0x12c [] show_stack+0x14/0x1c [] dump_stack+0x70/0x90 [] print_bad_pte+0x12c/0x1bc [] unmap_single_vma+0x4cc/0x700 [] unmap_vmas+0x68/0xb4 [] unmap_region+0xcc/0x1d4 [] do_munmap+0x218/0x314 [] vm_munmap+0x44/0x64 [] SyS_munmap+0x24/0x34 Where unmap_single_vma contains inlined -> unmap_page_range -> zap_pud_range -> zap_pmd_range -> zap_pte_range -> print_bad_pte. Or: BUG: Bad page state in process xl pfn:4077b4d page:ffffffbce1a2f8d8 count:0 mapcount:-1 mapping: (null) index:0x0 page flags: 0x4000000000000014(referenced|dirty) Modules linked in: CPU: 0 PID: 2657 Comm: xl Tainted: G B 3.12.0+ #102 Call trace: [] dump_backtrace+0x0/0x12c [] show_stack+0x14/0x1c [] dump_stack+0x70/0x90 [] bad_page+0xc4/0x110 [] free_pages_prepare+0xd0/0xd8 [] free_hot_cold_page+0x28/0x178 [] free_hot_cold_page_list+0x38/0x60 [] release_pages+0x190/0x1dc [] unmap_region+0x15c/0x1d4 [] do_munmap+0x218/0x314 [] vm_munmap+0x44/0x64 [] SyS_munmap+0x24/0x34 x86 already gets this correct. 32-bit arm gets away with this because there is not PTE_SPECIAL bit in the PTE there and the vm_normal_page fallback path does the right thing. Signed-off-by: Ian Campbell Signed-off-by: Stefano Stabellini diff --git a/arch/arm/xen/enlighten.c b/arch/arm/xen/enlighten.c index 6a288c7..8550123 100644 --- a/arch/arm/xen/enlighten.c +++ b/arch/arm/xen/enlighten.c @@ -96,7 +96,7 @@ static int remap_pte_fn(pte_t *ptep, pgtable_t token, unsigned long addr, struct remap_data *info = data; struct page *page = info->pages[info->index++]; unsigned long pfn = page_to_pfn(page); - pte_t pte = pfn_pte(pfn, info->prot); + pte_t pte = pte_mkspecial(pfn_pte(pfn, info->prot)); if (map_foreign_page(pfn, info->fgmfn, info->domid)) return -EFAULT; -- cgit v0.10.2 From 2306bfb208b9c403592a0513b08f2e6ad53056d5 Mon Sep 17 00:00:00 2001 From: Eric Seppanen Date: Thu, 21 Nov 2013 14:49:56 -0800 Subject: iscsi-target: return -EINVAL on oversized configfs parameter The iSCSI CHAP auth parameters are already copied with respect for the destination buffer size. Return -EINVAL instead of silently truncating the input. Signed-off-by: Eric Seppanen Signed-off-by: Nicholas Bellinger diff --git a/drivers/target/iscsi/iscsi_target_configfs.c b/drivers/target/iscsi/iscsi_target_configfs.c index e3318ed..1c0088f 100644 --- a/drivers/target/iscsi/iscsi_target_configfs.c +++ b/drivers/target/iscsi/iscsi_target_configfs.c @@ -474,7 +474,8 @@ static ssize_t __iscsi_##prefix##_store_##name( \ \ if (!capable(CAP_SYS_ADMIN)) \ return -EPERM; \ - \ + if (count >= sizeof(auth->name)) \ + return -EINVAL; \ snprintf(auth->name, sizeof(auth->name), "%s", page); \ if (!strncmp("NULL", auth->name, 4)) \ auth->naf_flags &= ~flags; \ -- cgit v0.10.2 From a51d5229d10dd3a337b674ce8603437d2996c5c3 Mon Sep 17 00:00:00 2001 From: Roland Dreier Date: Sat, 23 Nov 2013 10:35:58 -0800 Subject: target: Remove write-only stats fields and lock from struct se_node_acl Commit 04f3b31bff72 ("iscsi-target: Convert iscsi_session statistics to atomic_long_t") removed the updating of these fields in iscsi (the only fabric driver that ever touched these counters), and the core has no way to report or otherwise use the values. Remove the last remnants of these counters. Signed-off-by: Roland Dreier Signed-off-by: Nicholas Bellinger diff --git a/drivers/target/target_core_tpg.c b/drivers/target/target_core_tpg.c index f697f8b..f755712 100644 --- a/drivers/target/target_core_tpg.c +++ b/drivers/target/target_core_tpg.c @@ -278,7 +278,6 @@ struct se_node_acl *core_tpg_check_initiator_node_acl( snprintf(acl->initiatorname, TRANSPORT_IQN_LEN, "%s", initiatorname); acl->se_tpg = tpg; acl->acl_index = scsi_get_new_index(SCSI_AUTH_INTR_INDEX); - spin_lock_init(&acl->stats_lock); acl->dynamic_node_acl = 1; tpg->se_tpg_tfo->set_default_node_attributes(acl); @@ -406,7 +405,6 @@ struct se_node_acl *core_tpg_add_initiator_node_acl( snprintf(acl->initiatorname, TRANSPORT_IQN_LEN, "%s", initiatorname); acl->se_tpg = tpg; acl->acl_index = scsi_get_new_index(SCSI_AUTH_INTR_INDEX); - spin_lock_init(&acl->stats_lock); tpg->se_tpg_tfo->set_default_node_attributes(acl); diff --git a/include/target/target_core_base.h b/include/target/target_core_base.h index 45412a6..9f1dda6 100644 --- a/include/target/target_core_base.h +++ b/include/target/target_core_base.h @@ -517,10 +517,6 @@ struct se_node_acl { u32 acl_index; #define MAX_ACL_TAG_SIZE 64 char acl_tag[MAX_ACL_TAG_SIZE]; - u64 num_cmds; - u64 read_bytes; - u64 write_bytes; - spinlock_t stats_lock; /* Used for PR SPEC_I_PT=1 and REGISTER_AND_MOVE */ atomic_t acl_pr_ref_count; struct se_dev_entry **device_list; -- cgit v0.10.2 From 4454b66cb67f14c33cd70ddcf0ff4985b26324b7 Mon Sep 17 00:00:00 2001 From: Nicholas Bellinger Date: Mon, 25 Nov 2013 14:53:57 -0800 Subject: iscsi-target: Fix-up all zero data-length CDBs with R/W_BIT set This patch changes special case handling for ISCSI_OP_SCSI_CMD where an initiator sends a zero length Expected Data Transfer Length (EDTL), but still sets the WRITE and/or READ flag bits when no payload transfer is requested. Many, many moons ago two special cases where added for an ancient version of ESX that has long since been fixed, so instead of adding a new special case for the reported bug with a Broadcom 57800 NIC, go ahead and always strip off the incorrect WRITE + READ flag bits. Also, avoid sending a reject here, as RFC-3720 does mandate this case be handled without protocol error. Reported-by: Witold Bazakbal <865perl@wp.pl> Tested-by: Witold Bazakbal <865perl@wp.pl> Cc: #3.1+ Signed-off-by: Nicholas Bellinger diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c index d70e911..02182ab 100644 --- a/drivers/target/iscsi/iscsi_target.c +++ b/drivers/target/iscsi/iscsi_target.c @@ -823,24 +823,22 @@ int iscsit_setup_scsi_cmd(struct iscsi_conn *conn, struct iscsi_cmd *cmd, if (((hdr->flags & ISCSI_FLAG_CMD_READ) || (hdr->flags & ISCSI_FLAG_CMD_WRITE)) && !hdr->data_length) { /* - * Vmware ESX v3.0 uses a modified Cisco Initiator (v3.4.2) - * that adds support for RESERVE/RELEASE. There is a bug - * add with this new functionality that sets R/W bits when - * neither CDB carries any READ or WRITE datapayloads. + * From RFC-3720 Section 10.3.1: + * + * "Either or both of R and W MAY be 1 when either the + * Expected Data Transfer Length and/or Bidirectional Read + * Expected Data Transfer Length are 0" + * + * For this case, go ahead and clear the unnecssary bits + * to avoid any confusion with ->data_direction. */ - if ((hdr->cdb[0] == 0x16) || (hdr->cdb[0] == 0x17)) { - hdr->flags &= ~ISCSI_FLAG_CMD_READ; - hdr->flags &= ~ISCSI_FLAG_CMD_WRITE; - goto done; - } + hdr->flags &= ~ISCSI_FLAG_CMD_READ; + hdr->flags &= ~ISCSI_FLAG_CMD_WRITE; - pr_err("ISCSI_FLAG_CMD_READ or ISCSI_FLAG_CMD_WRITE" + pr_warn("ISCSI_FLAG_CMD_READ or ISCSI_FLAG_CMD_WRITE" " set when Expected Data Transfer Length is 0 for" - " CDB: 0x%02x. Bad iSCSI Initiator.\n", hdr->cdb[0]); - return iscsit_add_reject_cmd(cmd, - ISCSI_REASON_BOOKMARK_INVALID, buf); + " CDB: 0x%02x, Fixing up flags\n", hdr->cdb[0]); } -done: if (!(hdr->flags & ISCSI_FLAG_CMD_READ) && !(hdr->flags & ISCSI_FLAG_CMD_WRITE) && (hdr->data_length != 0)) { -- cgit v0.10.2 From 94a7111043d99819cd0a72d9b3174c7054adb2a0 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Tue, 29 Oct 2013 09:56:34 +0800 Subject: iser-target: fix error return code in isert_create_device_ib_res() Fix to return a negative error code from the error handling case instead of 0, as done elsewhere in this function. Signed-off-by: Wei Yongjun Cc: #3.10+ Signed-off-by: Nicholas Bellinger diff --git a/drivers/infiniband/ulp/isert/ib_isert.c b/drivers/infiniband/ulp/isert/ib_isert.c index 6be57c3..78f6e92 100644 --- a/drivers/infiniband/ulp/isert/ib_isert.c +++ b/drivers/infiniband/ulp/isert/ib_isert.c @@ -264,21 +264,29 @@ isert_create_device_ib_res(struct isert_device *device) isert_cq_event_callback, (void *)&cq_desc[i], ISER_MAX_RX_CQ_LEN, i); - if (IS_ERR(device->dev_rx_cq[i])) + if (IS_ERR(device->dev_rx_cq[i])) { + ret = PTR_ERR(device->dev_rx_cq[i]); + device->dev_rx_cq[i] = NULL; goto out_cq; + } device->dev_tx_cq[i] = ib_create_cq(device->ib_device, isert_cq_tx_callback, isert_cq_event_callback, (void *)&cq_desc[i], ISER_MAX_TX_CQ_LEN, i); - if (IS_ERR(device->dev_tx_cq[i])) + if (IS_ERR(device->dev_tx_cq[i])) { + ret = PTR_ERR(device->dev_tx_cq[i]); + device->dev_tx_cq[i] = NULL; goto out_cq; + } - if (ib_req_notify_cq(device->dev_rx_cq[i], IB_CQ_NEXT_COMP)) + ret = ib_req_notify_cq(device->dev_rx_cq[i], IB_CQ_NEXT_COMP); + if (ret) goto out_cq; - if (ib_req_notify_cq(device->dev_tx_cq[i], IB_CQ_NEXT_COMP)) + ret = ib_req_notify_cq(device->dev_tx_cq[i], IB_CQ_NEXT_COMP); + if (ret) goto out_cq; } -- cgit v0.10.2 From 63832aabec12a28a41a221773ab3819d30ba0a67 Mon Sep 17 00:00:00 2001 From: Shivaram Upadhyayula Date: Tue, 10 Dec 2013 16:06:40 +0530 Subject: qla2xxx: Fix schedule_delayed_work() for target timeout calculations This patch fixes two cases in qla_target.c code where the schedule_delayed_work() value was being incorrectly calculated from sess->expires - jiffies. Signed-off-by: Shivaram U Cc: #3.6+ Signed-off-by: Nicholas Bellinger diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c index 5964800..3bb0a1d 100644 --- a/drivers/scsi/qla2xxx/qla_target.c +++ b/drivers/scsi/qla2xxx/qla_target.c @@ -471,7 +471,7 @@ static void qlt_schedule_sess_for_deletion(struct qla_tgt_sess *sess, schedule_delayed_work(&tgt->sess_del_work, 0); else schedule_delayed_work(&tgt->sess_del_work, - jiffies - sess->expires); + sess->expires - jiffies); } /* ha->hardware_lock supposed to be held on entry */ @@ -550,13 +550,14 @@ static void qlt_del_sess_work_fn(struct delayed_work *work) struct scsi_qla_host *vha = tgt->vha; struct qla_hw_data *ha = vha->hw; struct qla_tgt_sess *sess; - unsigned long flags; + unsigned long flags, elapsed; spin_lock_irqsave(&ha->hardware_lock, flags); while (!list_empty(&tgt->del_sess_list)) { sess = list_entry(tgt->del_sess_list.next, typeof(*sess), del_list_entry); - if (time_after_eq(jiffies, sess->expires)) { + elapsed = jiffies; + if (time_after_eq(elapsed, sess->expires)) { qlt_undelete_sess(sess); ql_dbg(ql_dbg_tgt_mgt, vha, 0xf004, @@ -566,7 +567,7 @@ static void qlt_del_sess_work_fn(struct delayed_work *work) ha->tgt.tgt_ops->put_sess(sess); } else { schedule_delayed_work(&tgt->sess_del_work, - jiffies - sess->expires); + sess->expires - elapsed); break; } } -- cgit v0.10.2 From a3adadf3018102c24754e0b53a5515c40fbaff4a Mon Sep 17 00:00:00 2001 From: Eric Leblond Date: Thu, 12 Dec 2013 08:51:59 +0100 Subject: netfilter: nft_reject: fix endianness in dump function The dump function in nft_reject_ipv4 was not converting a u32 field to network order before sending it to userspace, this needs to happen for consistency with other nf_tables and nfnetlink subsystems. Signed-off-by: Eric Leblond Signed-off-by: Pablo Neira Ayuso diff --git a/net/ipv4/netfilter/nft_reject_ipv4.c b/net/ipv4/netfilter/nft_reject_ipv4.c index fff5ba1..4a5e94a 100644 --- a/net/ipv4/netfilter/nft_reject_ipv4.c +++ b/net/ipv4/netfilter/nft_reject_ipv4.c @@ -72,7 +72,7 @@ static int nft_reject_dump(struct sk_buff *skb, const struct nft_expr *expr) { const struct nft_reject *priv = nft_expr_priv(expr); - if (nla_put_be32(skb, NFTA_REJECT_TYPE, priv->type)) + if (nla_put_be32(skb, NFTA_REJECT_TYPE, htonl(priv->type))) goto nla_put_failure; switch (priv->type) { -- cgit v0.10.2 From 8333f0fe133be420ce3fcddfd568c3a559ab274e Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 2 Dec 2013 18:15:51 -0500 Subject: drm/radeon: Fix sideport problems on certain RS690 boards Some RS690 boards with 64MB of sideport memory show up as having 128MB sideport + 256MB of UMA. In this case, just skip the sideport memory and use UMA. This fixes rendering corruption and should improve performance. bug: https://bugs.freedesktop.org/show_bug.cgi?id=35457 Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c index 1c56062..e7dab06 100644 --- a/drivers/gpu/drm/radeon/rs690.c +++ b/drivers/gpu/drm/radeon/rs690.c @@ -162,6 +162,16 @@ static void rs690_mc_init(struct radeon_device *rdev) base = RREG32_MC(R_000100_MCCFG_FB_LOCATION); base = G_000100_MC_FB_START(base) << 16; rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); + /* Some boards seem to be configured for 128MB of sideport memory, + * but really only have 64MB. Just skip the sideport and use + * UMA memory. + */ + if (rdev->mc.igp_sideport_enabled && + (rdev->mc.real_vram_size == (384 * 1024 * 1024))) { + base += 128 * 1024 * 1024; + rdev->mc.real_vram_size -= 128 * 1024 * 1024; + rdev->mc.mc_vram_size = rdev->mc.real_vram_size; + } /* Use K8 direct mapping for fast fb access. */ rdev->fastfb_working = false; -- cgit v0.10.2 From 693e0cb052c607e2d41edf9e9f1fa99ff8c266c1 Mon Sep 17 00:00:00 2001 From: David Henningsson Date: Thu, 12 Dec 2013 09:52:03 +0100 Subject: ALSA: hda - Add enable_msi=0 workaround for four HP machines While enabling these machines, we found we would sometimes lose an interrupt if we change hardware volume during playback, and that disabling msi fixed this issue. (Losing the interrupt caused underruns and crackling audio, as the one second timeout is usually bigger than the period size.) The machines were all machines from HP, running AMD Hudson controller, and Realtek ALC282 codec. Cc: stable@vger.kernel.org BugLink: https://bugs.launchpad.net/bugs/1260225 Signed-off-by: David Henningsson Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c index 27aa140..956871d 100644 --- a/sound/pci/hda/hda_intel.c +++ b/sound/pci/hda/hda_intel.c @@ -3433,6 +3433,10 @@ static void check_probe_mask(struct azx *chip, int dev) * white/black-list for enable_msi */ static struct snd_pci_quirk msi_black_list[] = { + SND_PCI_QUIRK(0x103c, 0x2191, "HP", 0), /* AMD Hudson */ + SND_PCI_QUIRK(0x103c, 0x2192, "HP", 0), /* AMD Hudson */ + SND_PCI_QUIRK(0x103c, 0x21f7, "HP", 0), /* AMD Hudson */ + SND_PCI_QUIRK(0x103c, 0x21fa, "HP", 0), /* AMD Hudson */ SND_PCI_QUIRK(0x1043, 0x81f2, "ASUS", 0), /* Athlon64 X2 + nvidia */ SND_PCI_QUIRK(0x1043, 0x81f6, "ASUS", 0), /* nvidia */ SND_PCI_QUIRK(0x1043, 0x822d, "ASUS", 0), /* Athlon64 X2 + nvidia MCP55 */ -- cgit v0.10.2 From 1522f9c79cd29c8f55120777ddc8ba63b9216509 Mon Sep 17 00:00:00 2001 From: Martin Andersson Date: Sat, 7 Dec 2013 23:22:10 +0100 Subject: drm/radeon/dpm: Fix hwmon crash Commit ec39f64bba3421c2060fcbd1aeb6eec81fe0a42d (drm/radeon/dpm: Convert to use devm_hwmon_register_with_groups) converted one usage of dev_get_drvdata, but there were two more. bug: https://bugs.freedesktop.org/show_bug.cgi?id=72457 Signed-off-by: Martin Andersson Signed-off-by: Alex Deucher diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index dc75bb6..984097b 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -552,8 +552,7 @@ static ssize_t radeon_hwmon_show_temp_thresh(struct device *dev, struct device_attribute *attr, char *buf) { - struct drm_device *ddev = dev_get_drvdata(dev); - struct radeon_device *rdev = ddev->dev_private; + struct radeon_device *rdev = dev_get_drvdata(dev); int hyst = to_sensor_dev_attr(attr)->index; int temp; @@ -580,8 +579,7 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, struct attribute *attr, int index) { struct device *dev = container_of(kobj, struct device, kobj); - struct drm_device *ddev = dev_get_drvdata(dev); - struct radeon_device *rdev = ddev->dev_private; + struct radeon_device *rdev = dev_get_drvdata(dev); /* Skip limit attributes if DPM is not enabled */ if (rdev->pm.pm_method != PM_METHOD_DPM && -- cgit v0.10.2 From 7819678faec569f90c243d877c6f948dfb03f3a9 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 9 Dec 2013 17:38:51 -0500 Subject: drm/radeon/cik: plug in missing blit callback I implemented support for this, but forget to hook up the callback so the driver can actually use it. On asics with a dedicated DMA engine, we use the DMA engine for buffer migration so this is just for testing purposes. Signed-off-by: Alex Deucher diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c index e354ce9..c0425bb 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.c +++ b/drivers/gpu/drm/radeon/radeon_asic.c @@ -2021,7 +2021,7 @@ static struct radeon_asic ci_asic = { .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = NULL, + .blit = &cik_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &cik_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -2122,7 +2122,7 @@ static struct radeon_asic kv_asic = { .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = NULL, + .blit = &cik_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &cik_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, -- cgit v0.10.2 From 1b3abef830db98c11d7f916a483abaf2501f3323 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 10 Dec 2013 17:57:37 +0100 Subject: drm/radeon: fix typo in cik_copy_dma MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Otherwise we end up with a rather strange looking result. Signed-off-by: Christian König Tested-by: Tom Stellard Cc: stable@vger.kernel.org Signed-off-by: Alex Deucher diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c index 0300727..d08b83c 100644 --- a/drivers/gpu/drm/radeon/cik_sdma.c +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -458,7 +458,7 @@ int cik_copy_dma(struct radeon_device *rdev, radeon_ring_write(ring, 0); /* src/dst endian swap */ radeon_ring_write(ring, src_offset & 0xffffffff); radeon_ring_write(ring, upper_32_bits(src_offset) & 0xffffffff); - radeon_ring_write(ring, dst_offset & 0xfffffffc); + radeon_ring_write(ring, dst_offset & 0xffffffff); radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xffffffff); src_offset += cur_size_in_bytes; dst_offset += cur_size_in_bytes; -- cgit v0.10.2 From 227ae10f17a5f2fd1307b7e582b603ef7bbb7e97 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 11 Dec 2013 11:43:58 -0500 Subject: drm/radeon: add missing display tiling setup for oland Fixes improperly set up display params for 2D tiling on oland. Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index 80a2012..b197059 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -1196,7 +1196,9 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, } else if ((rdev->family == CHIP_TAHITI) || (rdev->family == CHIP_PITCAIRN)) fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P8_32x32_8x16); - else if (rdev->family == CHIP_VERDE) + else if ((rdev->family == CHIP_VERDE) || + (rdev->family == CHIP_OLAND) || + (rdev->family == CHIP_HAINAN)) /* for completeness. HAINAN has no display hw */ fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P4_8x16); switch (radeon_crtc->crtc_id) { -- cgit v0.10.2 From 59aebe2b32466f7ed5b79f18646be6dddc926a6d Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 12 Dec 2013 09:05:06 -0500 Subject: Revert "drm/radeon: Implement radeon_pci_shutdown" This causes a race condition between drm_dev_unregister() and pci_driver.shutdown at shutdown or driver unload time. We need to revisit how to properly support kexec within the drm. This reverts commit 846ae41ae99d314bf2a02784152208a6ddf7eddc. diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 9f5ff28..1958b36a 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -508,15 +508,6 @@ static const struct file_operations radeon_driver_kms_fops = { #endif }; - -static void -radeon_pci_shutdown(struct pci_dev *pdev) -{ - struct drm_device *dev = pci_get_drvdata(pdev); - - radeon_driver_unload_kms(dev); -} - static struct drm_driver kms_driver = { .driver_features = DRIVER_USE_AGP | @@ -586,7 +577,6 @@ static struct pci_driver radeon_kms_pci_driver = { .probe = radeon_pci_probe, .remove = radeon_pci_remove, .driver.pm = &radeon_pm_ops, - .shutdown = radeon_pci_shutdown, }; static int __init radeon_init(void) -- cgit v0.10.2 From 4cc629b7a20945ce35628179180329b6bc9e552b Mon Sep 17 00:00:00 2001 From: Stephen Boyd Date: Tue, 10 Dec 2013 15:19:03 -0800 Subject: gpio: msm: Fix irq mask/unmask by writing bits instead of numbers We should be writing bits here but instead we're writing the numbers that correspond to the bits we want to write. Fix it by wrapping the numbers in the BIT() macro. This fixes gpios acting as interrupts. Cc: stable@vger.kernel.org Signed-off-by: Stephen Boyd Signed-off-by: Linus Walleij diff --git a/drivers/gpio/gpio-msm-v2.c b/drivers/gpio/gpio-msm-v2.c index 7b37300..2baf0dd 100644 --- a/drivers/gpio/gpio-msm-v2.c +++ b/drivers/gpio/gpio-msm-v2.c @@ -252,7 +252,7 @@ static void msm_gpio_irq_mask(struct irq_data *d) spin_lock_irqsave(&tlmm_lock, irq_flags); writel(TARGET_PROC_NONE, GPIO_INTR_CFG_SU(gpio)); - clear_gpio_bits(INTR_RAW_STATUS_EN | INTR_ENABLE, GPIO_INTR_CFG(gpio)); + clear_gpio_bits(BIT(INTR_RAW_STATUS_EN) | BIT(INTR_ENABLE), GPIO_INTR_CFG(gpio)); __clear_bit(gpio, msm_gpio.enabled_irqs); spin_unlock_irqrestore(&tlmm_lock, irq_flags); } @@ -264,7 +264,7 @@ static void msm_gpio_irq_unmask(struct irq_data *d) spin_lock_irqsave(&tlmm_lock, irq_flags); __set_bit(gpio, msm_gpio.enabled_irqs); - set_gpio_bits(INTR_RAW_STATUS_EN | INTR_ENABLE, GPIO_INTR_CFG(gpio)); + set_gpio_bits(BIT(INTR_RAW_STATUS_EN) | BIT(INTR_ENABLE), GPIO_INTR_CFG(gpio)); writel(TARGET_PROC_SCORPION, GPIO_INTR_CFG_SU(gpio)); spin_unlock_irqrestore(&tlmm_lock, irq_flags); } -- cgit v0.10.2 From 8808b64daac68a2c85366c767a3ef850824ede74 Mon Sep 17 00:00:00 2001 From: Valentine Barshak Date: Fri, 29 Nov 2013 22:04:09 +0400 Subject: gpio: rcar: Fix level interrupt handling According to the manual, if a port is set for level detection using the corresponding bit in the edge/level select register and an external level interrupt signal is asserted, the corresponding bit in INTDT does not use the FF to hold the input. Thus, writing 1 to the corresponding bits in INTCLR cannot clear the corresponding bits in the INTDT register. Instead, when an external input signal is stopped, the corresponding bit in INTDT is cleared automatically. Since the INTDT bit cannot be cleared for the level interrupts until the interrupt signal is stopped, we end up with the infinite loop when using deferred (threaded) IRQ handling. Since a deferred interrupt is disabled by the low-level handler and re-enabled only when the deferred handler is completed, Fix the issue by dropping disabled interrupts from the pending mask as suggested by Laurent Pinchart Changes in V2: * Drop disabled interrupts from pending mask altogether instead of dropping level interrupts one by one once they get handled. Signed-off-by: Valentine Barshak Acked-by: Laurent Pinchart Acked-by: Magnus Damm Signed-off-by: Linus Walleij diff --git a/drivers/gpio/gpio-rcar.c b/drivers/gpio/gpio-rcar.c index fe088a3..8b7e719 100644 --- a/drivers/gpio/gpio-rcar.c +++ b/drivers/gpio/gpio-rcar.c @@ -169,7 +169,8 @@ static irqreturn_t gpio_rcar_irq_handler(int irq, void *dev_id) u32 pending; unsigned int offset, irqs_handled = 0; - while ((pending = gpio_rcar_read(p, INTDT))) { + while ((pending = gpio_rcar_read(p, INTDT) & + gpio_rcar_read(p, INTMSK))) { offset = __ffs(pending); gpio_rcar_write(p, INTCLR, BIT(offset)); generic_handle_irq(irq_find_mapping(p->irq_domain, offset)); -- cgit v0.10.2 From 0f58411d4fd704d8958879fb08751eae3573271b Mon Sep 17 00:00:00 2001 From: Ilia Mirkin Date: Thu, 5 Dec 2013 09:42:49 -0500 Subject: drm: don't double-free on driver load error MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit All instances of drm_dev_register are followed by drm_dev_free on failure. Don't free dev->control/render/primary on failure, as they will be freed by drm_dev_free since commit 8f6599da8e (drm: delay minor destruction to drm_dev_free()). Instead unplug them. Signed-off-by: Ilia Mirkin Reported-and-tested-by: Bruno Prémont Reviewed-by: David Herrmann Signed-off-by: Dave Airlie diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index f53d524..66dd3a0 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -566,11 +566,11 @@ err_unload: if (dev->driver->unload) dev->driver->unload(dev); err_primary_node: - drm_put_minor(dev->primary); + drm_unplug_minor(dev->primary); err_render_node: - drm_put_minor(dev->render); + drm_unplug_minor(dev->render); err_control_node: - drm_put_minor(dev->control); + drm_unplug_minor(dev->control); err_agp: if (dev->driver->bus->agp_destroy) dev->driver->bus->agp_destroy(dev); -- cgit v0.10.2 From 36aa1b180e7398e2bd27642760bfaa4ad8c65ab6 Mon Sep 17 00:00:00 2001 From: Ulrich Weigand Date: Thu, 12 Dec 2013 15:59:34 +1100 Subject: powerpc: PTRACE_PEEKUSR always returns FPR0 There is a bug in using ptrace to access FPRs via PTRACE_PEEKUSR / PTRACE_POKEUSR. In effect, trying to access any of the FPRs always really accesses FPR0, which does seriously break debugging :-) The problem seems to have been introduced by commit 3ad26e5c4459d (Merge branch 'for-kvm' into next). [ It is indeed a merge conflict between Paul's FPU/VSX state rework and my LE patches - Anton ] Signed-off-by: Ulrich Weigand Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/ptrace.c b/arch/powerpc/kernel/ptrace.c index 75fb404..2e3d2bf 100644 --- a/arch/powerpc/kernel/ptrace.c +++ b/arch/powerpc/kernel/ptrace.c @@ -1555,7 +1555,7 @@ long arch_ptrace(struct task_struct *child, long request, flush_fp_to_thread(child); if (fpidx < (PT_FPSCR - PT_FPR0)) - memcpy(&tmp, &child->thread.fp_state.fpr, + memcpy(&tmp, &child->thread.TS_FPR(fpidx), sizeof(long)); else tmp = child->thread.fp_state.fpscr; @@ -1588,7 +1588,7 @@ long arch_ptrace(struct task_struct *child, long request, flush_fp_to_thread(child); if (fpidx < (PT_FPSCR - PT_FPR0)) - memcpy(&child->thread.fp_state.fpr, &data, + memcpy(&child->thread.TS_FPR(fpidx), &data, sizeof(long)); else child->thread.fp_state.fpscr = data; -- cgit v0.10.2 From 01666c8ee2b6afcd31de2064fbb7c097a75e5089 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Thu, 12 Dec 2013 15:59:35 +1100 Subject: powerpc: Fix endian issue in setup-common.c During on LE boot we see: Partition configured for 1073741824 cpus, operating system maximum is 2048. Clearly missing a byteswap here. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/setup-common.c b/arch/powerpc/kernel/setup-common.c index febc804..bc76cc6 100644 --- a/arch/powerpc/kernel/setup-common.c +++ b/arch/powerpc/kernel/setup-common.c @@ -479,7 +479,7 @@ void __init smp_setup_cpu_maps(void) if (machine_is(pseries) && firmware_has_feature(FW_FEATURE_LPAR) && (dn = of_find_node_by_path("/rtas"))) { int num_addr_cell, num_size_cell, maxcpus; - const unsigned int *ireg; + const __be32 *ireg; num_addr_cell = of_n_addr_cells(dn); num_size_cell = of_n_size_cells(dn); @@ -489,7 +489,7 @@ void __init smp_setup_cpu_maps(void) if (!ireg) goto out; - maxcpus = ireg[num_addr_cell + num_size_cell]; + maxcpus = be32_to_cpup(ireg + num_addr_cell + num_size_cell); /* Double maxcpus for processors which have SMT capability */ if (cpu_has_feature(CPU_FTR_SMT)) -- cgit v0.10.2 From f8a1883a833bbad8e6b5ec4f0918b7797e652d65 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Thu, 12 Dec 2013 15:59:36 +1100 Subject: powerpc: Fix topology core_id endian issue on LE builds cpu_to_core_id() is missing a byteswap: cat /sys/devices/system/cpu/cpu63/topology/core_id 201326592 Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c index a3b64f3..c1cf4a1 100644 --- a/arch/powerpc/kernel/smp.c +++ b/arch/powerpc/kernel/smp.c @@ -580,7 +580,7 @@ int __cpu_up(unsigned int cpu, struct task_struct *tidle) int cpu_to_core_id(int cpu) { struct device_node *np; - const int *reg; + const __be32 *reg; int id = -1; np = of_get_cpu_node(cpu, NULL); @@ -591,7 +591,7 @@ int cpu_to_core_id(int cpu) if (!reg) goto out; - id = *reg; + id = be32_to_cpup(reg); out: of_node_put(np); return id; -- cgit v0.10.2 From ca5de4e652ea09c0c18ebbd12dd4c2149271245a Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Thu, 12 Dec 2013 15:59:37 +1100 Subject: powerpc/pseries: Fix endian issues in /proc/ppc64/lparcfg Some obvious issues: cat /proc/ppc64/lparcfg ... partition_id=16777216 ... partition_potential_processors=268435456 Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/platforms/pseries/lparcfg.c b/arch/powerpc/platforms/pseries/lparcfg.c index e738007..c9fecf0 100644 --- a/arch/powerpc/platforms/pseries/lparcfg.c +++ b/arch/powerpc/platforms/pseries/lparcfg.c @@ -157,7 +157,7 @@ static void parse_ppp_data(struct seq_file *m) { struct hvcall_ppp_data ppp_data; struct device_node *root; - const int *perf_level; + const __be32 *perf_level; int rc; rc = h_get_ppp(&ppp_data); @@ -201,7 +201,7 @@ static void parse_ppp_data(struct seq_file *m) perf_level = of_get_property(root, "ibm,partition-performance-parameters-level", NULL); - if (perf_level && (*perf_level >= 1)) { + if (perf_level && (be32_to_cpup(perf_level) >= 1)) { seq_printf(m, "physical_procs_allocated_to_virtualization=%d\n", ppp_data.phys_platform_procs); @@ -435,7 +435,7 @@ static int pseries_lparcfg_data(struct seq_file *m, void *v) int partition_potential_processors; int partition_active_processors; struct device_node *rtas_node; - const int *lrdrp = NULL; + const __be32 *lrdrp = NULL; rtas_node = of_find_node_by_path("/rtas"); if (rtas_node) @@ -444,7 +444,7 @@ static int pseries_lparcfg_data(struct seq_file *m, void *v) if (lrdrp == NULL) { partition_potential_processors = vdso_data->processorCount; } else { - partition_potential_processors = *(lrdrp + 4); + partition_potential_processors = be32_to_cpup(lrdrp + 4); } of_node_put(rtas_node); @@ -654,7 +654,7 @@ static int lparcfg_data(struct seq_file *m, void *v) const char *model = ""; const char *system_id = ""; const char *tmp; - const unsigned int *lp_index_ptr; + const __be32 *lp_index_ptr; unsigned int lp_index = 0; seq_printf(m, "%s %s\n", MODULE_NAME, MODULE_VERS); @@ -670,7 +670,7 @@ static int lparcfg_data(struct seq_file *m, void *v) lp_index_ptr = of_get_property(rootdn, "ibm,partition-no", NULL); if (lp_index_ptr) - lp_index = *lp_index_ptr; + lp_index = be32_to_cpup(lp_index_ptr); of_node_put(rootdn); } seq_printf(m, "serial_number=%s\n", system_id); -- cgit v0.10.2 From 9fa2984d1b5d9736a88c813ae89c160a2f9d5308 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Thu, 12 Dec 2013 15:59:38 +1100 Subject: powerpc/pseries: Fix endian issues in nvram code The NVRAM code has a number of endian issues. I noticed a very confused error log count: RTAS: 100663330 -------- RTAS event begin -------- 100663330 == 0x06000022. 0x6 LE error logs and 0x22 BE error logs. The pstore code has similar issues - if we write an oops in one endian and attempt to read it in another we get junk. Make both of these formats big endian, and byteswap as required. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/platforms/pseries/nvram.c b/arch/powerpc/platforms/pseries/nvram.c index 7bfaf58..d7096f2 100644 --- a/arch/powerpc/platforms/pseries/nvram.c +++ b/arch/powerpc/platforms/pseries/nvram.c @@ -43,8 +43,8 @@ static char nvram_buf[NVRW_CNT]; /* assume this is in the first 4GB */ static DEFINE_SPINLOCK(nvram_lock); struct err_log_info { - int error_type; - unsigned int seq_num; + __be32 error_type; + __be32 seq_num; }; struct nvram_os_partition { @@ -79,9 +79,9 @@ static const char *pseries_nvram_os_partitions[] = { }; struct oops_log_info { - u16 version; - u16 report_length; - u64 timestamp; + __be16 version; + __be16 report_length; + __be64 timestamp; } __attribute__((packed)); static void oops_to_nvram(struct kmsg_dumper *dumper, @@ -291,8 +291,8 @@ int nvram_write_os_partition(struct nvram_os_partition *part, char * buff, length = part->size; } - info.error_type = err_type; - info.seq_num = error_log_cnt; + info.error_type = cpu_to_be32(err_type); + info.seq_num = cpu_to_be32(error_log_cnt); tmp_index = part->index; @@ -364,8 +364,8 @@ int nvram_read_partition(struct nvram_os_partition *part, char *buff, } if (part->os_partition) { - *error_log_cnt = info.seq_num; - *err_type = info.error_type; + *error_log_cnt = be32_to_cpu(info.seq_num); + *err_type = be32_to_cpu(info.error_type); } return 0; @@ -529,9 +529,9 @@ static int zip_oops(size_t text_len) pr_err("nvram: logging uncompressed oops/panic report\n"); return -1; } - oops_hdr->version = OOPS_HDR_VERSION; - oops_hdr->report_length = (u16) zipped_len; - oops_hdr->timestamp = get_seconds(); + oops_hdr->version = cpu_to_be16(OOPS_HDR_VERSION); + oops_hdr->report_length = cpu_to_be16(zipped_len); + oops_hdr->timestamp = cpu_to_be64(get_seconds()); return 0; } @@ -574,9 +574,9 @@ static int nvram_pstore_write(enum pstore_type_id type, clobbering_unread_rtas_event()) return -1; - oops_hdr->version = OOPS_HDR_VERSION; - oops_hdr->report_length = (u16) size; - oops_hdr->timestamp = get_seconds(); + oops_hdr->version = cpu_to_be16(OOPS_HDR_VERSION); + oops_hdr->report_length = cpu_to_be16(size); + oops_hdr->timestamp = cpu_to_be64(get_seconds()); if (compressed) err_type = ERR_TYPE_KERNEL_PANIC_GZ; @@ -670,16 +670,16 @@ static ssize_t nvram_pstore_read(u64 *id, enum pstore_type_id *type, size_t length, hdr_size; oops_hdr = (struct oops_log_info *)buff; - if (oops_hdr->version < OOPS_HDR_VERSION) { + if (be16_to_cpu(oops_hdr->version) < OOPS_HDR_VERSION) { /* Old format oops header had 2-byte record size */ hdr_size = sizeof(u16); - length = oops_hdr->version; + length = be16_to_cpu(oops_hdr->version); time->tv_sec = 0; time->tv_nsec = 0; } else { hdr_size = sizeof(*oops_hdr); - length = oops_hdr->report_length; - time->tv_sec = oops_hdr->timestamp; + length = be16_to_cpu(oops_hdr->report_length); + time->tv_sec = be64_to_cpu(oops_hdr->timestamp); time->tv_nsec = 0; } *buf = kmalloc(length, GFP_KERNEL); @@ -889,13 +889,13 @@ static void oops_to_nvram(struct kmsg_dumper *dumper, kmsg_dump_get_buffer(dumper, false, oops_data, oops_data_sz, &text_len); err_type = ERR_TYPE_KERNEL_PANIC; - oops_hdr->version = OOPS_HDR_VERSION; - oops_hdr->report_length = (u16) text_len; - oops_hdr->timestamp = get_seconds(); + oops_hdr->version = cpu_to_be16(OOPS_HDR_VERSION); + oops_hdr->report_length = cpu_to_be16(text_len); + oops_hdr->timestamp = cpu_to_be64(get_seconds()); } (void) nvram_write_os_partition(&oops_log_partition, oops_buf, - (int) (sizeof(*oops_hdr) + oops_hdr->report_length), err_type, + (int) (sizeof(*oops_hdr) + text_len), err_type, ++oops_count); spin_unlock_irqrestore(&lock, flags); -- cgit v0.10.2 From 5091f0c969d2d0639db2efddce7352e7c48069ba Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Thu, 12 Dec 2013 15:59:39 +1100 Subject: powerpc/pseries: Fix PCIE link speed endian issue We need to byteswap ibm,pcie-link-speed-stats. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/platforms/pseries/pci.c b/arch/powerpc/platforms/pseries/pci.c index 5f93856..70670a2 100644 --- a/arch/powerpc/platforms/pseries/pci.c +++ b/arch/powerpc/platforms/pseries/pci.c @@ -113,7 +113,7 @@ int pseries_root_bridge_prepare(struct pci_host_bridge *bridge) { struct device_node *dn, *pdn; struct pci_bus *bus; - const uint32_t *pcie_link_speed_stats; + const __be32 *pcie_link_speed_stats; bus = bridge->bus; @@ -122,7 +122,7 @@ int pseries_root_bridge_prepare(struct pci_host_bridge *bridge) return 0; for (pdn = dn; pdn != NULL; pdn = of_get_next_parent(pdn)) { - pcie_link_speed_stats = (const uint32_t *) of_get_property(pdn, + pcie_link_speed_stats = of_get_property(pdn, "ibm,pcie-link-speed-stats", NULL); if (pcie_link_speed_stats) break; @@ -135,7 +135,7 @@ int pseries_root_bridge_prepare(struct pci_host_bridge *bridge) return 0; } - switch (pcie_link_speed_stats[0]) { + switch (be32_to_cpup(pcie_link_speed_stats)) { case 0x01: bus->max_bus_speed = PCIE_SPEED_2_5GT; break; @@ -147,7 +147,7 @@ int pseries_root_bridge_prepare(struct pci_host_bridge *bridge) break; } - switch (pcie_link_speed_stats[1]) { + switch (be32_to_cpup(pcie_link_speed_stats)) { case 0x01: bus->cur_bus_speed = PCIE_SPEED_2_5GT; break; -- cgit v0.10.2 From 8d15315537578a9ef5d35912d06127205dd82be8 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Thu, 12 Dec 2013 15:59:40 +1100 Subject: powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/platforms/pseries/msi.c b/arch/powerpc/platforms/pseries/msi.c index 6d2f0ab..0c882e8 100644 --- a/arch/powerpc/platforms/pseries/msi.c +++ b/arch/powerpc/platforms/pseries/msi.c @@ -130,7 +130,8 @@ static int check_req(struct pci_dev *pdev, int nvec, char *prop_name) { struct device_node *dn; struct pci_dn *pdn; - const u32 *req_msi; + const __be32 *p; + u32 req_msi; pdn = pci_get_pdn(pdev); if (!pdn) @@ -138,19 +139,20 @@ static int check_req(struct pci_dev *pdev, int nvec, char *prop_name) dn = pdn->node; - req_msi = of_get_property(dn, prop_name, NULL); - if (!req_msi) { + p = of_get_property(dn, prop_name, NULL); + if (!p) { pr_debug("rtas_msi: No %s on %s\n", prop_name, dn->full_name); return -ENOENT; } - if (*req_msi < nvec) { + req_msi = be32_to_cpup(p); + if (req_msi < nvec) { pr_debug("rtas_msi: %s requests < %d MSIs\n", prop_name, nvec); - if (*req_msi == 0) /* Be paranoid */ + if (req_msi == 0) /* Be paranoid */ return -ENOSPC; - return *req_msi; + return req_msi; } return 0; @@ -171,7 +173,7 @@ static int check_req_msix(struct pci_dev *pdev, int nvec) static struct device_node *find_pe_total_msi(struct pci_dev *dev, int *total) { struct device_node *dn; - const u32 *p; + const __be32 *p; dn = of_node_get(pci_device_to_OF_node(dev)); while (dn) { @@ -179,7 +181,7 @@ static struct device_node *find_pe_total_msi(struct pci_dev *dev, int *total) if (p) { pr_debug("rtas_msi: found prop on dn %s\n", dn->full_name); - *total = *p; + *total = be32_to_cpup(p); return dn; } @@ -232,13 +234,13 @@ struct msi_counts { static void *count_non_bridge_devices(struct device_node *dn, void *data) { struct msi_counts *counts = data; - const u32 *p; + const __be32 *p; u32 class; pr_debug("rtas_msi: counting %s\n", dn->full_name); p = of_get_property(dn, "class-code", NULL); - class = p ? *p : 0; + class = p ? be32_to_cpup(p) : 0; if ((class >> 8) != PCI_CLASS_BRIDGE_PCI) counts->num_devices++; @@ -249,7 +251,7 @@ static void *count_non_bridge_devices(struct device_node *dn, void *data) static void *count_spare_msis(struct device_node *dn, void *data) { struct msi_counts *counts = data; - const u32 *p; + const __be32 *p; int req; if (dn == counts->requestor) @@ -260,11 +262,11 @@ static void *count_spare_msis(struct device_node *dn, void *data) req = 0; p = of_get_property(dn, "ibm,req#msi", NULL); if (p) - req = *p; + req = be32_to_cpup(p); p = of_get_property(dn, "ibm,req#msi-x", NULL); if (p) - req = max(req, (int)*p); + req = max(req, (int)be32_to_cpup(p)); } if (req < counts->quota) -- cgit v0.10.2 From a29e30efa3a1b253fd1252731aa774a3544c5c3b Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Thu, 12 Dec 2013 15:59:41 +1100 Subject: powerpc: Fix endian issues in crash dump code A couple more device tree properties that need byte swapping. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/crash_dump.c b/arch/powerpc/kernel/crash_dump.c index 779a78c..11c1d06 100644 --- a/arch/powerpc/kernel/crash_dump.c +++ b/arch/powerpc/kernel/crash_dump.c @@ -124,15 +124,15 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, void crash_free_reserved_phys_range(unsigned long begin, unsigned long end) { unsigned long addr; - const u32 *basep, *sizep; + const __be32 *basep, *sizep; unsigned int rtas_start = 0, rtas_end = 0; basep = of_get_property(rtas.dev, "linux,rtas-base", NULL); sizep = of_get_property(rtas.dev, "rtas-size", NULL); if (basep && sizep) { - rtas_start = *basep; - rtas_end = *basep + *sizep; + rtas_start = be32_to_cpup(basep); + rtas_end = rtas_start + be32_to_cpup(sizep); } for (addr = begin; addr < end; addr += PAGE_SIZE) { -- cgit v0.10.2 From 01a9dbccbdfc7a93950b791405e476c510f73ab0 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Fri, 13 Dec 2013 15:53:43 +1100 Subject: powerpc/powernv: Fix endian issue in opal_xscom_read opal_xscom_read uses a pointer to return the data so we need to byteswap it on LE builds. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/include/asm/opal.h b/arch/powerpc/include/asm/opal.h index 033c06b..522f50c 100644 --- a/arch/powerpc/include/asm/opal.h +++ b/arch/powerpc/include/asm/opal.h @@ -720,7 +720,7 @@ int64_t opal_pci_next_error(uint64_t phb_id, uint64_t *first_frozen_pe, int64_t opal_pci_poll(uint64_t phb_id); int64_t opal_return_cpu(void); -int64_t opal_xscom_read(uint32_t gcid, uint32_t pcb_addr, uint64_t *val); +int64_t opal_xscom_read(uint32_t gcid, uint32_t pcb_addr, __be64 *val); int64_t opal_xscom_write(uint32_t gcid, uint32_t pcb_addr, uint64_t val); int64_t opal_lpc_write(uint32_t chip_id, enum OpalLPCAddressType addr_type, diff --git a/arch/powerpc/platforms/powernv/opal-xscom.c b/arch/powerpc/platforms/powernv/opal-xscom.c index 4d99a8f..4fbf276 100644 --- a/arch/powerpc/platforms/powernv/opal-xscom.c +++ b/arch/powerpc/platforms/powernv/opal-xscom.c @@ -96,9 +96,11 @@ static int opal_scom_read(scom_map_t map, u64 reg, u64 *value) { struct opal_scom_map *m = map; int64_t rc; + __be64 v; reg = opal_scom_unmangle(reg); - rc = opal_xscom_read(m->chip, m->addr + reg, (uint64_t *)__pa(value)); + rc = opal_xscom_read(m->chip, m->addr + reg, (__be64 *)__pa(&v)); + *value = be64_to_cpu(v); return opal_xscom_err_xlate(rc); } -- cgit v0.10.2 From 803c2d2f84da9dc2619449994af34d27148ab20d Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Fri, 13 Dec 2013 15:56:06 +1100 Subject: powerpc/powernv: Fix OPAL LPC access in Little Endian We are passing pointers to the firmware for reads, we need to properly convert the result as OPAL is always BE. Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/include/asm/opal.h b/arch/powerpc/include/asm/opal.h index 522f50c..7bdcf34 100644 --- a/arch/powerpc/include/asm/opal.h +++ b/arch/powerpc/include/asm/opal.h @@ -726,7 +726,7 @@ int64_t opal_xscom_write(uint32_t gcid, uint32_t pcb_addr, uint64_t val); int64_t opal_lpc_write(uint32_t chip_id, enum OpalLPCAddressType addr_type, uint32_t addr, uint32_t data, uint32_t sz); int64_t opal_lpc_read(uint32_t chip_id, enum OpalLPCAddressType addr_type, - uint32_t addr, uint32_t *data, uint32_t sz); + uint32_t addr, __be32 *data, uint32_t sz); int64_t opal_validate_flash(uint64_t buffer, uint32_t *size, uint32_t *result); int64_t opal_manage_flash(uint8_t op); int64_t opal_update_flash(uint64_t blk_list); diff --git a/arch/powerpc/platforms/powernv/opal-lpc.c b/arch/powerpc/platforms/powernv/opal-lpc.c index e7e59e4..79d83ca 100644 --- a/arch/powerpc/platforms/powernv/opal-lpc.c +++ b/arch/powerpc/platforms/powernv/opal-lpc.c @@ -24,25 +24,25 @@ static int opal_lpc_chip_id = -1; static u8 opal_lpc_inb(unsigned long port) { int64_t rc; - uint32_t data; + __be32 data; if (opal_lpc_chip_id < 0 || port > 0xffff) return 0xff; rc = opal_lpc_read(opal_lpc_chip_id, OPAL_LPC_IO, port, &data, 1); - return rc ? 0xff : data; + return rc ? 0xff : be32_to_cpu(data); } static __le16 __opal_lpc_inw(unsigned long port) { int64_t rc; - uint32_t data; + __be32 data; if (opal_lpc_chip_id < 0 || port > 0xfffe) return 0xffff; if (port & 1) return (__le16)opal_lpc_inb(port) << 8 | opal_lpc_inb(port + 1); rc = opal_lpc_read(opal_lpc_chip_id, OPAL_LPC_IO, port, &data, 2); - return rc ? 0xffff : data; + return rc ? 0xffff : be32_to_cpu(data); } static u16 opal_lpc_inw(unsigned long port) { @@ -52,7 +52,7 @@ static u16 opal_lpc_inw(unsigned long port) static __le32 __opal_lpc_inl(unsigned long port) { int64_t rc; - uint32_t data; + __be32 data; if (opal_lpc_chip_id < 0 || port > 0xfffc) return 0xffffffff; @@ -62,7 +62,7 @@ static __le32 __opal_lpc_inl(unsigned long port) (__le32)opal_lpc_inb(port + 2) << 8 | opal_lpc_inb(port + 3); rc = opal_lpc_read(opal_lpc_chip_id, OPAL_LPC_IO, port, &data, 4); - return rc ? 0xffffffff : data; + return rc ? 0xffffffff : be32_to_cpu(data); } static u32 opal_lpc_inl(unsigned long port) -- cgit v0.10.2 From f984841bc0c4d596c4e95d1798f318291c4f78f5 Mon Sep 17 00:00:00 2001 From: Jason Cooper Date: Mon, 9 Dec 2013 11:15:56 -0800 Subject: dma: mv_xor: remove mv_desc_get_dest_addr() The following commit: 54f8d501e842 dmaengine: remove DMA unmap from drivers removed the last caller to mv_desc_get_dest_addr(), creating the warning: drivers/dma/mv_xor.c:57:12: warning: mv_desc_get_dest_addr defined but not used [-Wunused-function] Remove it. Signed-off-by: Jason Cooper Acked-by: Vinod Koul Signed-off-by: Dan Williams diff --git a/drivers/dma/mv_xor.c b/drivers/dma/mv_xor.c index 7807f0e..23bcc91 100644 --- a/drivers/dma/mv_xor.c +++ b/drivers/dma/mv_xor.c @@ -54,12 +54,6 @@ static void mv_desc_init(struct mv_xor_desc_slot *desc, unsigned long flags) hw_desc->desc_command = (1 << 31); } -static u32 mv_desc_get_dest_addr(struct mv_xor_desc_slot *desc) -{ - struct mv_xor_desc *hw_desc = desc->hw_desc; - return hw_desc->phy_dest_addr; -} - static void mv_desc_set_byte_count(struct mv_xor_desc_slot *desc, u32 byte_count) { -- cgit v0.10.2 From d7fb0300fe663a5e338e3adeb7f811a2f9727e6c Mon Sep 17 00:00:00 2001 From: Olof Johansson Date: Mon, 9 Dec 2013 11:15:57 -0800 Subject: dmaengine: at_hdmac: remove unused function commit 54f8d501e8428 ('dmaengine: remove DMA unmap from drivers') refactored some code which resulted in an unused function in the at_hdmac driver: drivers/dma/at_hdmac_regs.h:350:23: warning: 'chan2parent' defined but not used [-Wunused-function] Fixes: 54f8d501e8428 ('dmaengine: remove DMA unmap from drivers') Signed-off-by: Olof Johansson Cc: Bartlomiej Zolnierkiewicz Acked-by: Nicolas Ferre Acked-by: Bartlomiej Zolnierkiewicz Acked-by: Vinod Koul Signed-off-by: Dan Williams diff --git a/drivers/dma/at_hdmac_regs.h b/drivers/dma/at_hdmac_regs.h index f31d647..2787aba 100644 --- a/drivers/dma/at_hdmac_regs.h +++ b/drivers/dma/at_hdmac_regs.h @@ -347,10 +347,6 @@ static struct device *chan2dev(struct dma_chan *chan) { return &chan->dev->device; } -static struct device *chan2parent(struct dma_chan *chan) -{ - return chan->dev->device.parent; -} #if defined(VERBOSE_DEBUG) static void vdbg_dump_regs(struct at_dma_chan *atchan) -- cgit v0.10.2 From 6aa2731ce2c7bd1305b553b5fc14ae4856d36569 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Mon, 9 Dec 2013 11:15:59 -0800 Subject: dma: fix build warnings in ppc4xx drivers/dma/ppc4xx/adma.c:1507:6: warning: unused variable 'i' [-Wunused-variable] - due to unmap reworks drivers/dma/ppc4xx/adma.c:3900:2: warning: format '%s' expects a matching 'char *' argument [-Wformat] - due to memset removal drivers/dma/ppc4xx/adma.c:538:13: warning: 'ppc440spe_desc_init_memset' defined but not used [-Wunused-function] - due to memset removal Cc: Bartlomiej Zolnierkiewicz Signed-off-by: Dan Williams diff --git a/drivers/dma/ppc4xx/adma.c b/drivers/dma/ppc4xx/adma.c index 8da48c6..8bba298 100644 --- a/drivers/dma/ppc4xx/adma.c +++ b/drivers/dma/ppc4xx/adma.c @@ -533,29 +533,6 @@ static void ppc440spe_desc_init_memcpy(struct ppc440spe_adma_desc_slot *desc, } /** - * ppc440spe_desc_init_memset - initialize the descriptor for MEMSET operation - */ -static void ppc440spe_desc_init_memset(struct ppc440spe_adma_desc_slot *desc, - int value, unsigned long flags) -{ - struct dma_cdb *hw_desc = desc->hw_desc; - - memset(desc->hw_desc, 0, sizeof(struct dma_cdb)); - desc->hw_next = NULL; - desc->src_cnt = 1; - desc->dst_cnt = 1; - - if (flags & DMA_PREP_INTERRUPT) - set_bit(PPC440SPE_DESC_INT, &desc->flags); - else - clear_bit(PPC440SPE_DESC_INT, &desc->flags); - - hw_desc->sg1u = hw_desc->sg1l = cpu_to_le32((u32)value); - hw_desc->sg3u = hw_desc->sg3l = cpu_to_le32((u32)value); - hw_desc->opc = DMA_CDB_OPC_DFILL128; -} - -/** * ppc440spe_desc_set_src_addr - set source address into the descriptor */ static void ppc440spe_desc_set_src_addr(struct ppc440spe_adma_desc_slot *desc, @@ -1504,8 +1481,6 @@ static dma_cookie_t ppc440spe_adma_run_tx_complete_actions( struct ppc440spe_adma_chan *chan, dma_cookie_t cookie) { - int i; - BUG_ON(desc->async_tx.cookie < 0); if (desc->async_tx.cookie > 0) { cookie = desc->async_tx.cookie; @@ -3898,7 +3873,7 @@ static void ppc440spe_adma_init_capabilities(struct ppc440spe_adma_device *adev) ppc440spe_adma_prep_dma_interrupt; } pr_info("%s: AMCC(R) PPC440SP(E) ADMA Engine: " - "( %s%s%s%s%s%s%s)\n", + "( %s%s%s%s%s%s)\n", dev_name(adev->dev), dma_has_cap(DMA_PQ, adev->common.cap_mask) ? "pq " : "", dma_has_cap(DMA_PQ_VAL, adev->common.cap_mask) ? "pq_val " : "", -- cgit v0.10.2 From bbc76560d488c437dbddff72242b0a07e42a0fd0 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Mon, 9 Dec 2013 11:16:00 -0800 Subject: dma: fix fsldma build warnings drivers/dma/fsldma.c: In function 'fsldma_cleanup_descriptor': drivers/dma/fsldma.c:860:6: warning: unused variable 'len' [-Wunused-variable] drivers/dma/fsldma.c:859:13: warning: unused variable 'dst' [-Wunused-variable] drivers/dma/fsldma.c:858:13: warning: unused variable 'src' [-Wunused-variable] drivers/dma/fsldma.c:857:17: warning: unused variable 'dev' [-Wunused-variable] - due to unmap changes drivers/dma/fsldma.c: In function 'fsl_dma_tx_submit': drivers/dma/fsldma.c:428:2: warning: 'cookie' may be used uninitialized in this function [-Wuninitialized] - long standing warning Cc: Bartlomiej Zolnierkiewicz Cc: Li Yang Cc: Zhang Wei Signed-off-by: Dan Williams diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c index 7086a16..f157c6f 100644 --- a/drivers/dma/fsldma.c +++ b/drivers/dma/fsldma.c @@ -86,11 +86,6 @@ static void set_desc_cnt(struct fsldma_chan *chan, hw->count = CPU_TO_DMA(chan, count, 32); } -static u32 get_desc_cnt(struct fsldma_chan *chan, struct fsl_desc_sw *desc) -{ - return DMA_TO_CPU(chan, desc->hw.count, 32); -} - static void set_desc_src(struct fsldma_chan *chan, struct fsl_dma_ld_hw *hw, dma_addr_t src) { @@ -101,16 +96,6 @@ static void set_desc_src(struct fsldma_chan *chan, hw->src_addr = CPU_TO_DMA(chan, snoop_bits | src, 64); } -static dma_addr_t get_desc_src(struct fsldma_chan *chan, - struct fsl_desc_sw *desc) -{ - u64 snoop_bits; - - snoop_bits = ((chan->feature & FSL_DMA_IP_MASK) == FSL_DMA_IP_85XX) - ? ((u64)FSL_DMA_SATR_SREADTYPE_SNOOP_READ << 32) : 0; - return DMA_TO_CPU(chan, desc->hw.src_addr, 64) & ~snoop_bits; -} - static void set_desc_dst(struct fsldma_chan *chan, struct fsl_dma_ld_hw *hw, dma_addr_t dst) { @@ -121,16 +106,6 @@ static void set_desc_dst(struct fsldma_chan *chan, hw->dst_addr = CPU_TO_DMA(chan, snoop_bits | dst, 64); } -static dma_addr_t get_desc_dst(struct fsldma_chan *chan, - struct fsl_desc_sw *desc) -{ - u64 snoop_bits; - - snoop_bits = ((chan->feature & FSL_DMA_IP_MASK) == FSL_DMA_IP_85XX) - ? ((u64)FSL_DMA_DATR_DWRITETYPE_SNOOP_WRITE << 32) : 0; - return DMA_TO_CPU(chan, desc->hw.dst_addr, 64) & ~snoop_bits; -} - static void set_desc_next(struct fsldma_chan *chan, struct fsl_dma_ld_hw *hw, dma_addr_t next) { @@ -408,7 +383,7 @@ static dma_cookie_t fsl_dma_tx_submit(struct dma_async_tx_descriptor *tx) struct fsl_desc_sw *desc = tx_to_fsl_desc(tx); struct fsl_desc_sw *child; unsigned long flags; - dma_cookie_t cookie; + dma_cookie_t cookie = -EINVAL; spin_lock_irqsave(&chan->desc_lock, flags); @@ -854,10 +829,6 @@ static void fsldma_cleanup_descriptor(struct fsldma_chan *chan, struct fsl_desc_sw *desc) { struct dma_async_tx_descriptor *txd = &desc->async_tx; - struct device *dev = chan->common.device->dev; - dma_addr_t src = get_desc_src(chan, desc); - dma_addr_t dst = get_desc_dst(chan, desc); - u32 len = get_desc_cnt(chan, desc); /* Run the link descriptor callback function */ if (txd->callback) { -- cgit v0.10.2 From 745c00daf9a75bacb53d0fe8635a132673ab0b46 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Mon, 9 Dec 2013 11:16:01 -0800 Subject: dmatest: fix build warning on mips drivers/dma/dmatest.c:543:11: warning: passing argument 1 of 'virt_to_phys' makes pointer from integer without a cast [enabled by default] mips expects virt_to_phys() to take a pointer. Fix up the types accordingly. Signed-off-by: Dan Williams diff --git a/drivers/dma/dmatest.c b/drivers/dma/dmatest.c index 20f9a3a..9dfcaf5 100644 --- a/drivers/dma/dmatest.c +++ b/drivers/dma/dmatest.c @@ -539,9 +539,9 @@ static int dmatest_func(void *data) um->len = params->buf_size; for (i = 0; i < src_cnt; i++) { - unsigned long buf = (unsigned long) thread->srcs[i]; + void *buf = thread->srcs[i]; struct page *pg = virt_to_page(buf); - unsigned pg_off = buf & ~PAGE_MASK; + unsigned pg_off = (unsigned long) buf & ~PAGE_MASK; um->addr[i] = dma_map_page(dev->dev, pg, pg_off, um->len, DMA_TO_DEVICE); @@ -559,9 +559,9 @@ static int dmatest_func(void *data) /* map with DMA_BIDIRECTIONAL to force writeback/invalidate */ dsts = &um->addr[src_cnt]; for (i = 0; i < dst_cnt; i++) { - unsigned long buf = (unsigned long) thread->dsts[i]; + void *buf = thread->dsts[i]; struct page *pg = virt_to_page(buf); - unsigned pg_off = buf & ~PAGE_MASK; + unsigned pg_off = (unsigned long) buf & ~PAGE_MASK; dsts[i] = dma_map_page(dev->dev, pg, pg_off, um->len, DMA_BIDIRECTIONAL); -- cgit v0.10.2 From 8e5ee258d98a6643227d958361aec2a62559b804 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Mon, 9 Dec 2013 11:16:03 -0800 Subject: dma: fix build warnings in txx9 The unmap rework missed this: drivers/dma/txx9dmac.c:409:25: warning: unused variable 'ds' [-Wunused-variable] Cc: Bartlomiej Zolnierkiewicz Signed-off-by: Dan Williams diff --git a/drivers/dma/txx9dmac.c b/drivers/dma/txx9dmac.c index bae6c29..17686ca 100644 --- a/drivers/dma/txx9dmac.c +++ b/drivers/dma/txx9dmac.c @@ -406,7 +406,6 @@ txx9dmac_descriptor_complete(struct txx9dmac_chan *dc, dma_async_tx_callback callback; void *param; struct dma_async_tx_descriptor *txd = &desc->txd; - struct txx9dmac_slave *ds = dc->chan.private; dev_vdbg(chan2dev(&dc->chan), "descriptor %u %p complete\n", txd->cookie, desc); -- cgit v0.10.2 From 3cc377b9ae4bd3133bf8ba388d2b2b66b2b973c1 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Mon, 9 Dec 2013 10:33:16 -0800 Subject: dmaengine: fix enable for high order unmap pools The higher order mempools support raid operations, and we want to disable them when raid support is not enabled. Making them conditional on ASYNC_TX_DMA is not sufficient as other users (specifically dmatest) will also issue raid operations. Make raid drivers explicitly request that the core carry the higher order pools. Reported-by: Ezequiel Garcia Tested-by: Ezequiel Garcia Signed-off-by: Dan Williams diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig index 446687c..132a4fd 100644 --- a/drivers/dma/Kconfig +++ b/drivers/dma/Kconfig @@ -62,6 +62,7 @@ config INTEL_IOATDMA tristate "Intel I/OAT DMA support" depends on PCI && X86 select DMA_ENGINE + select DMA_ENGINE_RAID select DCA help Enable support for the Intel(R) I/OAT DMA engine present @@ -112,6 +113,7 @@ config MV_XOR bool "Marvell XOR engine support" depends on PLAT_ORION select DMA_ENGINE + select DMA_ENGINE_RAID select ASYNC_TX_ENABLE_CHANNEL_SWITCH ---help--- Enable support for the Marvell XOR engine. @@ -187,6 +189,7 @@ config AMCC_PPC440SPE_ADMA tristate "AMCC PPC440SPe ADMA support" depends on 440SPe || 440SP select DMA_ENGINE + select DMA_ENGINE_RAID select ARCH_HAS_ASYNC_TX_FIND_CHANNEL select ASYNC_TX_ENABLE_CHANNEL_SWITCH help @@ -377,4 +380,7 @@ config DMATEST Simple DMA test client. Say N unless you're debugging a DMA Device driver. +config DMA_ENGINE_RAID + bool + endif diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c index ea806bd..b601024 100644 --- a/drivers/dma/dmaengine.c +++ b/drivers/dma/dmaengine.c @@ -912,7 +912,7 @@ struct dmaengine_unmap_pool { #define __UNMAP_POOL(x) { .size = x, .name = "dmaengine-unmap-" __stringify(x) } static struct dmaengine_unmap_pool unmap_pool[] = { __UNMAP_POOL(2), - #if IS_ENABLED(CONFIG_ASYNC_TX_DMA) + #if IS_ENABLED(CONFIG_DMA_ENGINE_RAID) __UNMAP_POOL(16), __UNMAP_POOL(128), __UNMAP_POOL(256), -- cgit v0.10.2 From d16695a75019ac4baad7a117dc86d1d292e09115 Mon Sep 17 00:00:00 2001 From: Ezequiel Garcia Date: Tue, 10 Dec 2013 09:32:36 -0300 Subject: dma: mv_xor: Use dmaengine_unmap_data for the self-tests The driver-specific unmap code was removed in: commit 54f8d501e842879143e867e70996574a54d1e130 Author: Bartlomiej Zolnierkiewicz Date: Fri Oct 18 19:35:32 2013 +0200 dmaengine: remove DMA unmap from drivers which had the side-effect of not unmapping the self-test mappings. Fix this by using dmaengine_unmap_data in the self-test routines. In addition, since dmaengine_unmap() assumes that all mappings were created with dma_map_page, this commit changes the single mapping to a page mapping to avoid an incorrect unmapping of the memcpy self-test. The allocation could be changed to be alloc_page(), but sticking to kmalloc results in a less intrusive patch. The size of the test buffer is increased, since dma_map_page() seem to fail when the source and destination pages are the same page. Signed-off-by: Ezequiel Garcia Signed-off-by: Dan Williams diff --git a/drivers/dma/mv_xor.c b/drivers/dma/mv_xor.c index 23bcc91..7962088 100644 --- a/drivers/dma/mv_xor.c +++ b/drivers/dma/mv_xor.c @@ -781,7 +781,6 @@ static void mv_xor_issue_pending(struct dma_chan *chan) /* * Perform a transaction to verify the HW works. */ -#define MV_XOR_TEST_SIZE 2000 static int mv_xor_memcpy_self_test(struct mv_xor_chan *mv_chan) { @@ -791,20 +790,21 @@ static int mv_xor_memcpy_self_test(struct mv_xor_chan *mv_chan) struct dma_chan *dma_chan; dma_cookie_t cookie; struct dma_async_tx_descriptor *tx; + struct dmaengine_unmap_data *unmap; int err = 0; - src = kmalloc(sizeof(u8) * MV_XOR_TEST_SIZE, GFP_KERNEL); + src = kmalloc(sizeof(u8) * PAGE_SIZE, GFP_KERNEL); if (!src) return -ENOMEM; - dest = kzalloc(sizeof(u8) * MV_XOR_TEST_SIZE, GFP_KERNEL); + dest = kzalloc(sizeof(u8) * PAGE_SIZE, GFP_KERNEL); if (!dest) { kfree(src); return -ENOMEM; } /* Fill in src buffer */ - for (i = 0; i < MV_XOR_TEST_SIZE; i++) + for (i = 0; i < PAGE_SIZE; i++) ((u8 *) src)[i] = (u8)i; dma_chan = &mv_chan->dmachan; @@ -813,14 +813,26 @@ static int mv_xor_memcpy_self_test(struct mv_xor_chan *mv_chan) goto out; } - dest_dma = dma_map_single(dma_chan->device->dev, dest, - MV_XOR_TEST_SIZE, DMA_FROM_DEVICE); + unmap = dmaengine_get_unmap_data(dma_chan->device->dev, 2, GFP_KERNEL); + if (!unmap) { + err = -ENOMEM; + goto free_resources; + } + + src_dma = dma_map_page(dma_chan->device->dev, virt_to_page(src), 0, + PAGE_SIZE, DMA_TO_DEVICE); + unmap->to_cnt = 1; + unmap->addr[0] = src_dma; - src_dma = dma_map_single(dma_chan->device->dev, src, - MV_XOR_TEST_SIZE, DMA_TO_DEVICE); + dest_dma = dma_map_page(dma_chan->device->dev, virt_to_page(dest), 0, + PAGE_SIZE, DMA_FROM_DEVICE); + unmap->from_cnt = 1; + unmap->addr[1] = dest_dma; + + unmap->len = PAGE_SIZE; tx = mv_xor_prep_dma_memcpy(dma_chan, dest_dma, src_dma, - MV_XOR_TEST_SIZE, 0); + PAGE_SIZE, 0); cookie = mv_xor_tx_submit(tx); mv_xor_issue_pending(dma_chan); async_tx_ack(tx); @@ -835,8 +847,8 @@ static int mv_xor_memcpy_self_test(struct mv_xor_chan *mv_chan) } dma_sync_single_for_cpu(dma_chan->device->dev, dest_dma, - MV_XOR_TEST_SIZE, DMA_FROM_DEVICE); - if (memcmp(src, dest, MV_XOR_TEST_SIZE)) { + PAGE_SIZE, DMA_FROM_DEVICE); + if (memcmp(src, dest, PAGE_SIZE)) { dev_err(dma_chan->device->dev, "Self-test copy failed compare, disabling\n"); err = -ENODEV; @@ -844,6 +856,7 @@ static int mv_xor_memcpy_self_test(struct mv_xor_chan *mv_chan) } free_resources: + dmaengine_unmap_put(unmap); mv_xor_free_chan_resources(dma_chan); out: kfree(src); @@ -861,13 +874,15 @@ mv_xor_xor_self_test(struct mv_xor_chan *mv_chan) dma_addr_t dma_srcs[MV_XOR_NUM_SRC_TEST]; dma_addr_t dest_dma; struct dma_async_tx_descriptor *tx; + struct dmaengine_unmap_data *unmap; struct dma_chan *dma_chan; dma_cookie_t cookie; u8 cmp_byte = 0; u32 cmp_word; int err = 0; + int src_count = MV_XOR_NUM_SRC_TEST; - for (src_idx = 0; src_idx < MV_XOR_NUM_SRC_TEST; src_idx++) { + for (src_idx = 0; src_idx < src_count; src_idx++) { xor_srcs[src_idx] = alloc_page(GFP_KERNEL); if (!xor_srcs[src_idx]) { while (src_idx--) @@ -884,13 +899,13 @@ mv_xor_xor_self_test(struct mv_xor_chan *mv_chan) } /* Fill in src buffers */ - for (src_idx = 0; src_idx < MV_XOR_NUM_SRC_TEST; src_idx++) { + for (src_idx = 0; src_idx < src_count; src_idx++) { u8 *ptr = page_address(xor_srcs[src_idx]); for (i = 0; i < PAGE_SIZE; i++) ptr[i] = (1 << src_idx); } - for (src_idx = 0; src_idx < MV_XOR_NUM_SRC_TEST; src_idx++) + for (src_idx = 0; src_idx < src_count; src_idx++) cmp_byte ^= (u8) (1 << src_idx); cmp_word = (cmp_byte << 24) | (cmp_byte << 16) | @@ -904,16 +919,29 @@ mv_xor_xor_self_test(struct mv_xor_chan *mv_chan) goto out; } + unmap = dmaengine_get_unmap_data(dma_chan->device->dev, src_count + 1, + GFP_KERNEL); + if (!unmap) { + err = -ENOMEM; + goto free_resources; + } + /* test xor */ - dest_dma = dma_map_page(dma_chan->device->dev, dest, 0, PAGE_SIZE, - DMA_FROM_DEVICE); + for (i = 0; i < src_count; i++) { + unmap->addr[i] = dma_map_page(dma_chan->device->dev, xor_srcs[i], + 0, PAGE_SIZE, DMA_TO_DEVICE); + dma_srcs[i] = unmap->addr[i]; + unmap->to_cnt++; + } - for (i = 0; i < MV_XOR_NUM_SRC_TEST; i++) - dma_srcs[i] = dma_map_page(dma_chan->device->dev, xor_srcs[i], - 0, PAGE_SIZE, DMA_TO_DEVICE); + unmap->addr[src_count] = dma_map_page(dma_chan->device->dev, dest, 0, PAGE_SIZE, + DMA_FROM_DEVICE); + dest_dma = unmap->addr[src_count]; + unmap->from_cnt = 1; + unmap->len = PAGE_SIZE; tx = mv_xor_prep_dma_xor(dma_chan, dest_dma, dma_srcs, - MV_XOR_NUM_SRC_TEST, PAGE_SIZE, 0); + src_count, PAGE_SIZE, 0); cookie = mv_xor_tx_submit(tx); mv_xor_issue_pending(dma_chan); @@ -942,9 +970,10 @@ mv_xor_xor_self_test(struct mv_xor_chan *mv_chan) } free_resources: + dmaengine_unmap_put(unmap); mv_xor_free_chan_resources(dma_chan); out: - src_idx = MV_XOR_NUM_SRC_TEST; + src_idx = src_count; while (src_idx--) __free_page(xor_srcs[src_idx]); __free_page(dest); -- cgit v0.10.2 From 0be8253fa2b4385e6246387db1d6067366e987ba Mon Sep 17 00:00:00 2001 From: Russell King Date: Thu, 12 Dec 2013 23:59:08 +0000 Subject: dmaengine: mv_xor: fix oops when channels fail to initialise When a channel fails to initialise, we error out and clean up any previously unregistered channels by walking the entire xordev->channels array. Unfortunately, there are paths which end up storing an error pointer in this array, which we then try and dereference in the cleanup code, which causes an oops. Fix this by avoiding writing invalid pointers to this array in the first place. Tested-by: Aaro Koskinen Signed-off-by: Russell King Signed-off-by: Dan Williams diff --git a/drivers/dma/mv_xor.c b/drivers/dma/mv_xor.c index 7962088..53fb0c8 100644 --- a/drivers/dma/mv_xor.c +++ b/drivers/dma/mv_xor.c @@ -1199,6 +1199,7 @@ static int mv_xor_probe(struct platform_device *pdev) int i = 0; for_each_child_of_node(pdev->dev.of_node, np) { + struct mv_xor_chan *chan; dma_cap_mask_t cap_mask; int irq; @@ -1216,21 +1217,21 @@ static int mv_xor_probe(struct platform_device *pdev) goto err_channel_add; } - xordev->channels[i] = - mv_xor_channel_add(xordev, pdev, i, - cap_mask, irq); - if (IS_ERR(xordev->channels[i])) { - ret = PTR_ERR(xordev->channels[i]); - xordev->channels[i] = NULL; + chan = mv_xor_channel_add(xordev, pdev, i, + cap_mask, irq); + if (IS_ERR(chan)) { + ret = PTR_ERR(chan); irq_dispose_mapping(irq); goto err_channel_add; } + xordev->channels[i] = chan; i++; } } else if (pdata && pdata->channels) { for (i = 0; i < MV_XOR_MAX_CHANNELS; i++) { struct mv_xor_channel_data *cd; + struct mv_xor_chan *chan; int irq; cd = &pdata->channels[i]; @@ -1245,13 +1246,14 @@ static int mv_xor_probe(struct platform_device *pdev) goto err_channel_add; } - xordev->channels[i] = - mv_xor_channel_add(xordev, pdev, i, - cd->cap_mask, irq); - if (IS_ERR(xordev->channels[i])) { - ret = PTR_ERR(xordev->channels[i]); + chan = mv_xor_channel_add(xordev, pdev, i, + cd->cap_mask, irq); + if (IS_ERR(chan)) { + ret = PTR_ERR(chan); goto err_channel_add; } + + xordev->channels[i] = chan; } } -- cgit v0.10.2 From c29cb5eb8157a0049c882672a7f941261f23ea34 Mon Sep 17 00:00:00 2001 From: Hui Wang Date: Fri, 13 Dec 2013 11:57:05 +0800 Subject: ALSA: hda - Add Dell headset detection quirk for three laptop models On the Dell machines with codec whose Subsystem Id is 0x10280610, 0x10280629 or 0x1028063e, no external microphone can be detected when plugging a 3-ring headset. If we add "model=dell-headset-multi" for the snd-hda-intel.ko, the problem will disappear. The codecs on these machines belong to alc_269 family. BugLink: https://bugs.launchpad.net/bugs/1260303 Cc: David Henningsson Cc: stable@vger.kernel.org Signed-off-by: Hui Wang Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 34de5dc..5ab8e16 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -4247,11 +4247,14 @@ static const struct snd_pci_quirk alc269_fixup_tbl[] = { SND_PCI_QUIRK(0x1028, 0x0606, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x0608, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x0609, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), + SND_PCI_QUIRK(0x1028, 0x0610, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x0613, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x0614, "Dell Inspiron 3135", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x0616, "Dell Vostro 5470", ALC290_FIXUP_MONO_SPEAKERS), SND_PCI_QUIRK(0x1028, 0x061f, "Dell", ALC255_FIXUP_DELL1_MIC_NO_PRESENCE), + SND_PCI_QUIRK(0x1028, 0x0629, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x0638, "Dell Inspiron 5439", ALC290_FIXUP_MONO_SPEAKERS), + SND_PCI_QUIRK(0x1028, 0x063e, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x063f, "Dell", ALC255_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x15cc, "Dell X5 Precision", ALC269_FIXUP_DELL2_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x15cd, "Dell X5 Precision", ALC269_FIXUP_DELL2_MIC_NO_PRESENCE), -- cgit v0.10.2 From 8194ee27764b1a86fa7a6b0d411f0a225a6abd5f Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Fri, 13 Dec 2013 00:57:03 -0800 Subject: dmaengine: fix sleep in atomic BUG: sleeping function called from invalid context at mm/mempool.c:203 in_atomic(): 1, irqs_disabled(): 0, pid: 43502, name: linbug no locks held by linbug/43502. CPU: 7 PID: 43502 Comm: linbug Not tainted 3.13.0-rc1+ #15 Hardware name: 0000000000000010 ffff88005ebd1878 ffffffff8172d512 ffff8801752bc1c0 ffff8801752bc1c0 ffff88005ebd1898 ffffffff8109d1f6 ffff88005f9a3c58 ffff880177f0f080 ffff88005ebd1918 ffffffff81161f43 ffff88005ebd18f8 Call Trace: [] dump_stack+0x4e/0x68 [] __might_sleep+0xe6/0x120 [] mempool_alloc+0x93/0x170 [] ? mark_held_locks+0x74/0x140 [] ? follow_page_mask+0x556/0x600 [] dmaengine_get_unmap_data+0x2e/0x60 [] dma_async_memcpy_pg_to_pg+0x41/0x1c0 [] dma_async_memcpy_buf_to_pg+0x50/0x60 [] dma_memcpy_to_iovec+0xfc/0x190 [] dma_skb_copy_datagram_iovec+0x6f/0x2b0 Signed-off-by: Dan Williams diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c index b601024..ef63b90 100644 --- a/drivers/dma/dmaengine.c +++ b/drivers/dma/dmaengine.c @@ -1054,7 +1054,7 @@ dma_async_memcpy_pg_to_pg(struct dma_chan *chan, struct page *dest_pg, dma_cookie_t cookie; unsigned long flags; - unmap = dmaengine_get_unmap_data(dev->dev, 2, GFP_NOIO); + unmap = dmaengine_get_unmap_data(dev->dev, 2, GFP_NOWAIT); if (!unmap) return -ENOMEM; -- cgit v0.10.2 From 380108d891acf8db5cf0d477176c7ed2b62b7928 Mon Sep 17 00:00:00 2001 From: Julien Grall Date: Tue, 3 Dec 2013 15:40:37 +0000 Subject: xen/block: Correctly define structures in public headers on ARM32 and ARM64 On ARM (32 bits and 64 bits), the double-word is 8-bytes aligned. This will result on different structure from Xen and Linux repositories. As Linux is using __packed__ attribute, it must have a 4-bytes padding before each "id" field. This change breaks guest block support with older kernel. IMHO, it's acceptable because Xen on ARM is still on Tech Preview and the hypercall ABI is not yet freezed. Only one architecture (x86_32) doesn't have 64-bit ABI for the block interface. Don't add padding if Linux is compiled for this architecture. Signed-off-by: Julien Grall Acked-by: Roger Pau Monne Acked-by: David Vrabel Cc: Boris Ostrovsky Cc: Ian Campbell Acked-by: Stefano Stabellini [I had asked for confirmation that it did not break x86 and Ian went beyound the call of duty to confirm it. Also a internal regression bucket with 32/64 dom0 with 32/64 domU (PV and HVM) confirmed no regressions. ABI changes are a drag..] Signed-off-by: Konrad Rzeszutek Wilk diff --git a/include/xen/interface/io/blkif.h b/include/xen/interface/io/blkif.h index 65e1209..ae665ac 100644 --- a/include/xen/interface/io/blkif.h +++ b/include/xen/interface/io/blkif.h @@ -146,7 +146,7 @@ struct blkif_request_segment_aligned { struct blkif_request_rw { uint8_t nr_segments; /* number of segments */ blkif_vdev_t handle; /* only for read/write requests */ -#ifdef CONFIG_X86_64 +#ifndef CONFIG_X86_32 uint32_t _pad1; /* offsetof(blkif_request,u.rw.id) == 8 */ #endif uint64_t id; /* private guest value, echoed in resp */ @@ -163,7 +163,7 @@ struct blkif_request_discard { uint8_t flag; /* BLKIF_DISCARD_SECURE or zero. */ #define BLKIF_DISCARD_SECURE (1<<0) /* ignored if discard-secure=0 */ blkif_vdev_t _pad1; /* only for read/write requests */ -#ifdef CONFIG_X86_64 +#ifndef CONFIG_X86_32 uint32_t _pad2; /* offsetof(blkif_req..,u.discard.id)==8*/ #endif uint64_t id; /* private guest value, echoed in resp */ @@ -175,7 +175,7 @@ struct blkif_request_discard { struct blkif_request_other { uint8_t _pad1; blkif_vdev_t _pad2; /* only for read/write requests */ -#ifdef CONFIG_X86_64 +#ifndef CONFIG_X86_32 uint32_t _pad3; /* offsetof(blkif_req..,u.other.id)==8*/ #endif uint64_t id; /* private guest value, echoed in resp */ @@ -184,7 +184,7 @@ struct blkif_request_other { struct blkif_request_indirect { uint8_t indirect_op; uint16_t nr_segments; -#ifdef CONFIG_X86_64 +#ifndef CONFIG_X86_32 uint32_t _pad1; /* offsetof(blkif_...,u.indirect.id) == 8 */ #endif uint64_t id; @@ -192,7 +192,7 @@ struct blkif_request_indirect { blkif_vdev_t handle; uint16_t _pad2; grant_ref_t indirect_grefs[BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST]; -#ifdef CONFIG_X86_64 +#ifndef CONFIG_X86_32 uint32_t _pad3; /* make it 64 byte aligned */ #else uint64_t _pad3; /* make it 64 byte aligned */ -- cgit v0.10.2 From d7ec435fdd03cfee70dba934ee384acc87bd6d00 Mon Sep 17 00:00:00 2001 From: David Howells Date: Fri, 13 Dec 2013 15:20:19 +0000 Subject: X.509: Fix certificate gathering Fix the gathering of certificates from both the source tree and the build tree to correctly calculate the pathnames of all the certificates. The problem was that if the default generated cert, signing_key.x509, didn't exist then it would not have a path attached and if it did, it would have a path attached. This means that the contents of kernel/.x509.list would change between the first compilation in a directory and the second. After the second it would remain stable because the signing_key.x509 file exists. The consequence was that the kernel would get relinked unconditionally on the second recompilation. The second recompilation would also show something like this: X.509 certificate list changed CERTS kernel/x509_certificate_list - Including cert /home/torvalds/v2.6/linux/signing_key.x509 AS kernel/system_certificates.o LD kernel/built-in.o which is why the relink would happen. Unfortunately, it isn't a simple matter of just sticking a path on the front of the filename of the certificate in the build directory as make can't then work out how to build it. So the path has to be prepended to the name for sorting and duplicate elimination and then removed for the make rule if it is in the build tree. Reported-by: Linus Torvalds Signed-off-by: David Howells diff --git a/kernel/Makefile b/kernel/Makefile index bbaf7d5..c23bb0b 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -137,9 +137,10 @@ $(obj)/timeconst.h: $(obj)/hz.bc $(src)/timeconst.bc FORCE ############################################################################### ifeq ($(CONFIG_SYSTEM_TRUSTED_KEYRING),y) X509_CERTIFICATES-y := $(wildcard *.x509) $(wildcard $(srctree)/*.x509) -X509_CERTIFICATES-$(CONFIG_MODULE_SIG) += signing_key.x509 -X509_CERTIFICATES := $(sort $(foreach CERT,$(X509_CERTIFICATES-y), \ +X509_CERTIFICATES-$(CONFIG_MODULE_SIG) += $(objtree)/signing_key.x509 +X509_CERTIFICATES-raw := $(sort $(foreach CERT,$(X509_CERTIFICATES-y), \ $(or $(realpath $(CERT)),$(CERT)))) +X509_CERTIFICATES := $(subst $(realpath $(objtree))/,,$(X509_CERTIFICATES-raw)) ifeq ($(X509_CERTIFICATES),) $(warning *** No X.509 certificates found ***) -- cgit v0.10.2 From f46a3cbbebdaa5ca7b3ab23d7b81925dbe152bcb Mon Sep 17 00:00:00 2001 From: Kirill Tkhai Date: Tue, 10 Dec 2013 22:39:57 +0400 Subject: KEYS: Remove files generated when SYSTEM_TRUSTED_KEYRING=y Always remove generated SYSTEM_TRUSTED_KEYRING files while doing make mrproper. Signed-off-by: Kirill Tkhai Signed-off-by: David Howells diff --git a/kernel/Makefile b/kernel/Makefile index c23bb0b..bc010ee 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -165,9 +165,9 @@ $(obj)/x509_certificate_list: $(X509_CERTIFICATES) $(obj)/.x509.list targets += $(obj)/.x509.list $(obj)/.x509.list: @echo $(X509_CERTIFICATES) >$@ +endif clean-files := x509_certificate_list .x509.list -endif ifeq ($(CONFIG_MODULE_SIG),y) ############################################################################### -- cgit v0.10.2 From 6bd364d82920be726c2d678e7ba9e27112686e11 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Fri, 13 Dec 2013 15:00:32 +0800 Subject: KEYS: fix uninitialized persistent_keyring_register_sem We run into this bug: [ 2736.063245] Unable to handle kernel paging request for data at address 0x00000000 [ 2736.063293] Faulting instruction address: 0xc00000000037efb0 [ 2736.063300] Oops: Kernel access of bad area, sig: 11 [#1] [ 2736.063303] SMP NR_CPUS=2048 NUMA pSeries [ 2736.063310] Modules linked in: sg nfsv3 rpcsec_gss_krb5 nfsv4 dns_resolver nfs fscache nf_conntrack_netbios_ns nf_conntrack_broadcast ipt_MASQUERADE ip6table_mangle ip6table_security ip6table_raw ip6t_REJECT iptable_nat nf_nat_ipv4 iptable_mangle iptable_security iptable_raw ipt_REJECT nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack ebtable_filter ebtables ip6table_filter iptable_filter ip_tables ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 nf_nat nf_conntrack ip6_tables ibmveth pseries_rng nx_crypto nfsd auth_rpcgss nfs_acl lockd sunrpc binfmt_misc xfs libcrc32c dm_service_time sd_mod crc_t10dif crct10dif_common ibmvfc scsi_transport_fc scsi_tgt dm_mirror dm_region_hash dm_log dm_multipath dm_mod [ 2736.063383] CPU: 1 PID: 7128 Comm: ssh Not tainted 3.10.0-48.el7.ppc64 #1 [ 2736.063389] task: c000000131930120 ti: c0000001319a0000 task.ti: c0000001319a0000 [ 2736.063394] NIP: c00000000037efb0 LR: c0000000006c40f8 CTR: 0000000000000000 [ 2736.063399] REGS: c0000001319a3870 TRAP: 0300 Not tainted (3.10.0-48.el7.ppc64) [ 2736.063403] MSR: 8000000000009032 CR: 28824242 XER: 20000000 [ 2736.063415] SOFTE: 0 [ 2736.063418] CFAR: c00000000000908c [ 2736.063421] DAR: 0000000000000000, DSISR: 40000000 [ 2736.063425] GPR00: c0000000006c40f8 c0000001319a3af0 c000000001074788 c0000001319a3bf0 GPR04: 0000000000000000 0000000000000000 0000000000000020 000000000000000a GPR08: fffffffe00000002 00000000ffff0000 0000000080000001 c000000000924888 GPR12: 0000000028824248 c000000007e00400 00001fffffa0f998 0000000000000000 GPR16: 0000000000000022 00001fffffa0f998 0000010022e92470 0000000000000000 GPR20: 0000000000000000 0000000000000000 0000000000000000 0000000000000000 GPR24: 0000000000000000 c000000000f4a828 00003ffffe527108 0000000000000000 GPR28: c000000000f4a730 c000000000f4a828 0000000000000000 c0000001319a3bf0 [ 2736.063498] NIP [c00000000037efb0] .__list_add+0x30/0x110 [ 2736.063504] LR [c0000000006c40f8] .rwsem_down_write_failed+0x78/0x264 [ 2736.063508] PACATMSCRATCH [800000000280f032] [ 2736.063511] Call Trace: [ 2736.063516] [c0000001319a3af0] [c0000001319a3b80] 0xc0000001319a3b80 (unreliable) [ 2736.063523] [c0000001319a3b80] [c0000000006c40f8] .rwsem_down_write_failed+0x78/0x264 [ 2736.063530] [c0000001319a3c50] [c0000000006c1bb0] .down_write+0x70/0x78 [ 2736.063536] [c0000001319a3cd0] [c0000000002e5ffc] .keyctl_get_persistent+0x20c/0x320 [ 2736.063542] [c0000001319a3dc0] [c0000000002e2388] .SyS_keyctl+0x238/0x260 [ 2736.063548] [c0000001319a3e30] [c000000000009e7c] syscall_exit+0x0/0x7c [ 2736.063553] Instruction dump: [ 2736.063556] 7c0802a6 fba1ffe8 fbc1fff0 fbe1fff8 7cbd2b78 7c9e2378 7c7f1b78 f8010010 [ 2736.063566] f821ff71 e8a50008 7fa52040 40de00c0 7fbd2840 40de0094 7fbff040 [ 2736.063579] ---[ end trace 2708241785538296 ]--- It's caused by uninitialized persistent_keyring_register_sem. The bug was introduced by commit f36f8c75, two typos are in that commit: CONFIG_KEYS_KERBEROS_CACHE should be CONFIG_PERSISTENT_KEYRINGS and krb_cache_register_sem should be persistent_keyring_register_sem. Signed-off-by: Xiao Guangrong Signed-off-by: David Howells diff --git a/kernel/user.c b/kernel/user.c index a3a0dbf..c006131 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -51,9 +51,9 @@ struct user_namespace init_user_ns = { .owner = GLOBAL_ROOT_UID, .group = GLOBAL_ROOT_GID, .proc_inum = PROC_USER_INIT_INO, -#ifdef CONFIG_KEYS_KERBEROS_CACHE - .krb_cache_register_sem = - __RWSEM_INITIALIZER(init_user_ns.krb_cache_register_sem), +#ifdef CONFIG_PERSISTENT_KEYRINGS + .persistent_keyring_register_sem = + __RWSEM_INITIALIZER(init_user_ns.persistent_keyring_register_sem), #endif }; EXPORT_SYMBOL_GPL(init_user_ns); -- cgit v0.10.2 From 3cafea3076423987726023235e548af1d534ff1a Mon Sep 17 00:00:00 2001 From: James Solner Date: Wed, 6 Nov 2013 12:53:36 -0600 Subject: Add Documentation/module-signing.txt file This patch adds the Documentation/module-signing.txt file that is currently missing from the Documentation directory. The init/Kconfig file references the Documentation/module-signing.txt file to explain how kernel module signing works. This patch supplies this documentation. Signed-off-by: James Solner Signed-off-by: David Howells diff --git a/Documentation/module-signing.txt b/Documentation/module-signing.txt new file mode 100644 index 0000000..2b40e04 --- /dev/null +++ b/Documentation/module-signing.txt @@ -0,0 +1,240 @@ + ============================== + KERNEL MODULE SIGNING FACILITY + ============================== + +CONTENTS + + - Overview. + - Configuring module signing. + - Generating signing keys. + - Public keys in the kernel. + - Manually signing modules. + - Signed modules and stripping. + - Loading signed modules. + - Non-valid signatures and unsigned modules. + - Administering/protecting the private key. + + +======== +OVERVIEW +======== + +The kernel module signing facility cryptographically signs modules during +installation and then checks the signature upon loading the module. This +allows increased kernel security by disallowing the loading of unsigned modules +or modules signed with an invalid key. Module signing increases security by +making it harder to load a malicious module into the kernel. The module +signature checking is done by the kernel so that it is not necessary to have +trusted userspace bits. + +This facility uses X.509 ITU-T standard certificates to encode the public keys +involved. The signatures are not themselves encoded in any industrial standard +type. The facility currently only supports the RSA public key encryption +standard (though it is pluggable and permits others to be used). The possible +hash algorithms that can be used are SHA-1, SHA-224, SHA-256, SHA-384, and +SHA-512 (the algorithm is selected by data in the signature). + + +========================== +CONFIGURING MODULE SIGNING +========================== + +The module signing facility is enabled by going to the "Enable Loadable Module +Support" section of the kernel configuration and turning on + + CONFIG_MODULE_SIG "Module signature verification" + +This has a number of options available: + + (1) "Require modules to be validly signed" (CONFIG_MODULE_SIG_FORCE) + + This specifies how the kernel should deal with a module that has a + signature for which the key is not known or a module that is unsigned. + + If this is off (ie. "permissive"), then modules for which the key is not + available and modules that are unsigned are permitted, but the kernel will + be marked as being tainted. + + If this is on (ie. "restrictive"), only modules that have a valid + signature that can be verified by a public key in the kernel's possession + will be loaded. All other modules will generate an error. + + Irrespective of the setting here, if the module has a signature block that + cannot be parsed, it will be rejected out of hand. + + + (2) "Automatically sign all modules" (CONFIG_MODULE_SIG_ALL) + + If this is on then modules will be automatically signed during the + modules_install phase of a build. If this is off, then the modules must + be signed manually using: + + scripts/sign-file + + + (3) "Which hash algorithm should modules be signed with?" + + This presents a choice of which hash algorithm the installation phase will + sign the modules with: + + CONFIG_SIG_SHA1 "Sign modules with SHA-1" + CONFIG_SIG_SHA224 "Sign modules with SHA-224" + CONFIG_SIG_SHA256 "Sign modules with SHA-256" + CONFIG_SIG_SHA384 "Sign modules with SHA-384" + CONFIG_SIG_SHA512 "Sign modules with SHA-512" + + The algorithm selected here will also be built into the kernel (rather + than being a module) so that modules signed with that algorithm can have + their signatures checked without causing a dependency loop. + + +======================= +GENERATING SIGNING KEYS +======================= + +Cryptographic keypairs are required to generate and check signatures. A +private key is used to generate a signature and the corresponding public key is +used to check it. The private key is only needed during the build, after which +it can be deleted or stored securely. The public key gets built into the +kernel so that it can be used to check the signatures as the modules are +loaded. + +Under normal conditions, the kernel build will automatically generate a new +keypair using openssl if one does not exist in the files: + + signing_key.priv + signing_key.x509 + +during the building of vmlinux (the public part of the key needs to be built +into vmlinux) using parameters in the: + + x509.genkey + +file (which is also generated if it does not already exist). + +It is strongly recommended that you provide your own x509.genkey file. + +Most notably, in the x509.genkey file, the req_distinguished_name section +should be altered from the default: + + [ req_distinguished_name ] + O = Magrathea + CN = Glacier signing key + emailAddress = slartibartfast@magrathea.h2g2 + +The generated RSA key size can also be set with: + + [ req ] + default_bits = 4096 + + +It is also possible to manually generate the key private/public files using the +x509.genkey key generation configuration file in the root node of the Linux +kernel sources tree and the openssl command. The following is an example to +generate the public/private key files: + + openssl req -new -nodes -utf8 -sha256 -days 36500 -batch -x509 \ + -config x509.genkey -outform DER -out signing_key.x509 \ + -keyout signing_key.priv + + +========================= +PUBLIC KEYS IN THE KERNEL +========================= + +The kernel contains a ring of public keys that can be viewed by root. They're +in a keyring called ".system_keyring" that can be seen by: + + [root@deneb ~]# cat /proc/keys + ... + 223c7853 I------ 1 perm 1f030000 0 0 keyring .system_keyring: 1 + 302d2d52 I------ 1 perm 1f010000 0 0 asymmetri Fedora kernel signing key: d69a84e6bce3d216b979e9505b3e3ef9a7118079: X509.RSA a7118079 [] + ... + +Beyond the public key generated specifically for module signing, any file +placed in the kernel source root directory or the kernel build root directory +whose name is suffixed with ".x509" will be assumed to be an X.509 public key +and will be added to the keyring. + +Further, the architecture code may take public keys from a hardware store and +add those in also (e.g. from the UEFI key database). + +Finally, it is possible to add additional public keys by doing: + + keyctl padd asymmetric "" [.system_keyring-ID] <[key-file] + +e.g.: + + keyctl padd asymmetric "" 0x223c7853 Date: Wed, 11 Dec 2013 16:58:42 +0000 Subject: xen/balloon: Seperate the auto-translate logic properly (v2) The auto-xlat logic vs the non-xlat means that we don't need to for auto-xlat guests (like PVH, HVM or ARM): - use P2M - use scratch page. However the code in increase_reservation does modify the p2m for auto_translate guests, but not in decrease_reservation. Fix that by avoiding any p2m modifications in both increase_reservation and decrease_reservation for auto_translated guests. And also avoid allocating or using scratch pages for auto_translated guests. Lastly, since !auto-xlat is really another way of saying 'xen_pv' remove the redundant 'xen_pv_domain' check. Signed-off-by: Stefano Stabellini [v2: Updated the description] Signed-off-by: Konrad Rzeszutek Wilk diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c index 55ea73f..4c02e2b 100644 --- a/drivers/xen/balloon.c +++ b/drivers/xen/balloon.c @@ -350,17 +350,19 @@ static enum bp_state increase_reservation(unsigned long nr_pages) pfn = page_to_pfn(page); - set_phys_to_machine(pfn, frame_list[i]); - #ifdef CONFIG_XEN_HAVE_PVMMU - /* Link back into the page tables if not highmem. */ - if (xen_pv_domain() && !PageHighMem(page)) { - int ret; - ret = HYPERVISOR_update_va_mapping( - (unsigned long)__va(pfn << PAGE_SHIFT), - mfn_pte(frame_list[i], PAGE_KERNEL), - 0); - BUG_ON(ret); + if (!xen_feature(XENFEAT_auto_translated_physmap)) { + set_phys_to_machine(pfn, frame_list[i]); + + /* Link back into the page tables if not highmem. */ + if (!PageHighMem(page)) { + int ret; + ret = HYPERVISOR_update_va_mapping( + (unsigned long)__va(pfn << PAGE_SHIFT), + mfn_pte(frame_list[i], PAGE_KERNEL), + 0); + BUG_ON(ret); + } } #endif @@ -378,7 +380,6 @@ static enum bp_state decrease_reservation(unsigned long nr_pages, gfp_t gfp) enum bp_state state = BP_DONE; unsigned long pfn, i; struct page *page; - struct page *scratch_page; int ret; struct xen_memory_reservation reservation = { .address_bits = 0, @@ -411,27 +412,29 @@ static enum bp_state decrease_reservation(unsigned long nr_pages, gfp_t gfp) scrub_page(page); +#ifdef CONFIG_XEN_HAVE_PVMMU /* * Ballooned out frames are effectively replaced with * a scratch frame. Ensure direct mappings and the * p2m are consistent. */ - scratch_page = get_balloon_scratch_page(); -#ifdef CONFIG_XEN_HAVE_PVMMU - if (xen_pv_domain() && !PageHighMem(page)) { - ret = HYPERVISOR_update_va_mapping( - (unsigned long)__va(pfn << PAGE_SHIFT), - pfn_pte(page_to_pfn(scratch_page), - PAGE_KERNEL_RO), 0); - BUG_ON(ret); - } -#endif if (!xen_feature(XENFEAT_auto_translated_physmap)) { unsigned long p; + struct page *scratch_page = get_balloon_scratch_page(); + + if (!PageHighMem(page)) { + ret = HYPERVISOR_update_va_mapping( + (unsigned long)__va(pfn << PAGE_SHIFT), + pfn_pte(page_to_pfn(scratch_page), + PAGE_KERNEL_RO), 0); + BUG_ON(ret); + } p = page_to_pfn(scratch_page); __set_phys_to_machine(pfn, pfn_to_mfn(p)); + + put_balloon_scratch_page(); } - put_balloon_scratch_page(); +#endif balloon_append(pfn_to_page(pfn)); } @@ -627,15 +630,17 @@ static int __init balloon_init(void) if (!xen_domain()) return -ENODEV; - for_each_online_cpu(cpu) - { - per_cpu(balloon_scratch_page, cpu) = alloc_page(GFP_KERNEL); - if (per_cpu(balloon_scratch_page, cpu) == NULL) { - pr_warn("Failed to allocate balloon_scratch_page for cpu %d\n", cpu); - return -ENOMEM; + if (!xen_feature(XENFEAT_auto_translated_physmap)) { + for_each_online_cpu(cpu) + { + per_cpu(balloon_scratch_page, cpu) = alloc_page(GFP_KERNEL); + if (per_cpu(balloon_scratch_page, cpu) == NULL) { + pr_warn("Failed to allocate balloon_scratch_page for cpu %d\n", cpu); + return -ENOMEM; + } } + register_cpu_notifier(&balloon_cpu_notifier); } - register_cpu_notifier(&balloon_cpu_notifier); pr_info("Initialising balloon driver\n"); -- cgit v0.10.2 From 86b58d13134ef14f09f8c8f37797ccc37cf823a3 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 5 Dec 2013 12:38:59 +0800 Subject: ceph: initialize inode before instantiating dentry commit b18825a7c8 (Put a small type field into struct dentry::d_flags) put a type field into struct dentry::d_flags. __d_instantiate() set the field by checking inode->i_mode. So we should initialize inode before instantiating dentry when handling mds reply. Fixes: http://tracker.ceph.com/issues/6930 Signed-off-by: Yan, Zheng Reviewed-by: Sage Weil diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 9a8e396..278fd28 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -978,7 +978,6 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, struct ceph_mds_reply_inode *ininfo; struct ceph_vino vino; struct ceph_fs_client *fsc = ceph_sb_to_client(sb); - int i = 0; int err = 0; dout("fill_trace %p is_dentry %d is_target %d\n", req, @@ -1039,6 +1038,29 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, } } + if (rinfo->head->is_target) { + vino.ino = le64_to_cpu(rinfo->targeti.in->ino); + vino.snap = le64_to_cpu(rinfo->targeti.in->snapid); + + in = ceph_get_inode(sb, vino); + if (IS_ERR(in)) { + err = PTR_ERR(in); + goto done; + } + req->r_target_inode = in; + + err = fill_inode(in, &rinfo->targeti, NULL, + session, req->r_request_started, + (le32_to_cpu(rinfo->head->result) == 0) ? + req->r_fmode : -1, + &req->r_caps_reservation); + if (err < 0) { + pr_err("fill_inode badness %p %llx.%llx\n", + in, ceph_vinop(in)); + goto done; + } + } + /* * ignore null lease/binding on snapdir ENOENT, or else we * will have trouble splicing in the virtual snapdir later @@ -1108,7 +1130,6 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, ceph_dentry(req->r_old_dentry)->offset); dn = req->r_old_dentry; /* use old_dentry */ - in = dn->d_inode; } /* null dentry? */ @@ -1130,44 +1151,28 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, } /* attach proper inode */ - ininfo = rinfo->targeti.in; - vino.ino = le64_to_cpu(ininfo->ino); - vino.snap = le64_to_cpu(ininfo->snapid); - in = dn->d_inode; - if (!in) { - in = ceph_get_inode(sb, vino); - if (IS_ERR(in)) { - pr_err("fill_trace bad get_inode " - "%llx.%llx\n", vino.ino, vino.snap); - err = PTR_ERR(in); - d_drop(dn); - goto done; - } + if (!dn->d_inode) { + ihold(in); dn = splice_dentry(dn, in, &have_lease, true); if (IS_ERR(dn)) { err = PTR_ERR(dn); goto done; } req->r_dentry = dn; /* may have spliced */ - ihold(in); - } else if (ceph_ino(in) == vino.ino && - ceph_snap(in) == vino.snap) { - ihold(in); - } else { + } else if (dn->d_inode && dn->d_inode != in) { dout(" %p links to %p %llx.%llx, not %llx.%llx\n", - dn, in, ceph_ino(in), ceph_snap(in), - vino.ino, vino.snap); + dn, dn->d_inode, ceph_vinop(dn->d_inode), + ceph_vinop(in)); have_lease = false; - in = NULL; } if (have_lease) update_dentry_lease(dn, rinfo->dlease, session, req->r_request_started); dout(" final dn %p\n", dn); - i++; - } else if ((req->r_op == CEPH_MDS_OP_LOOKUPSNAP || - req->r_op == CEPH_MDS_OP_MKSNAP) && !req->r_aborted) { + } else if (!req->r_aborted && + (req->r_op == CEPH_MDS_OP_LOOKUPSNAP || + req->r_op == CEPH_MDS_OP_MKSNAP)) { struct dentry *dn = req->r_dentry; /* fill out a snapdir LOOKUPSNAP dentry */ @@ -1177,52 +1182,15 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, ininfo = rinfo->targeti.in; vino.ino = le64_to_cpu(ininfo->ino); vino.snap = le64_to_cpu(ininfo->snapid); - in = ceph_get_inode(sb, vino); - if (IS_ERR(in)) { - pr_err("fill_inode get_inode badness %llx.%llx\n", - vino.ino, vino.snap); - err = PTR_ERR(in); - d_delete(dn); - goto done; - } dout(" linking snapped dir %p to dn %p\n", in, dn); + ihold(in); dn = splice_dentry(dn, in, NULL, true); if (IS_ERR(dn)) { err = PTR_ERR(dn); goto done; } req->r_dentry = dn; /* may have spliced */ - ihold(in); - rinfo->head->is_dentry = 1; /* fool notrace handlers */ - } - - if (rinfo->head->is_target) { - vino.ino = le64_to_cpu(rinfo->targeti.in->ino); - vino.snap = le64_to_cpu(rinfo->targeti.in->snapid); - - if (in == NULL || ceph_ino(in) != vino.ino || - ceph_snap(in) != vino.snap) { - in = ceph_get_inode(sb, vino); - if (IS_ERR(in)) { - err = PTR_ERR(in); - goto done; - } - } - req->r_target_inode = in; - - err = fill_inode(in, - &rinfo->targeti, NULL, - session, req->r_request_started, - (le32_to_cpu(rinfo->head->result) == 0) ? - req->r_fmode : -1, - &req->r_caps_reservation); - if (err < 0) { - pr_err("fill_inode badness %p %llx.%llx\n", - in, ceph_vinop(in)); - goto done; - } } - done: dout("fill_trace done err=%d\n", err); return err; @@ -1272,7 +1240,7 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, struct qstr dname; struct dentry *dn; struct inode *in; - int err = 0, i; + int err = 0, ret, i; struct inode *snapdir = NULL; struct ceph_mds_request_head *rhead = req->r_request->front.iov_base; struct ceph_dentry_info *di; @@ -1305,6 +1273,7 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, ceph_fill_dirfrag(parent->d_inode, rinfo->dir_dir); } + /* FIXME: release caps/leases if error occurs */ for (i = 0; i < rinfo->dir_nr; i++) { struct ceph_vino vino; @@ -1329,9 +1298,10 @@ retry_lookup: err = -ENOMEM; goto out; } - err = ceph_init_dentry(dn); - if (err < 0) { + ret = ceph_init_dentry(dn); + if (ret < 0) { dput(dn); + err = ret; goto out; } } else if (dn->d_inode && @@ -1351,9 +1321,6 @@ retry_lookup: spin_unlock(&parent->d_lock); } - di = dn->d_fsdata; - di->offset = ceph_make_fpos(frag, i + r_readdir_offset); - /* inode */ if (dn->d_inode) { in = dn->d_inode; @@ -1366,26 +1333,39 @@ retry_lookup: err = PTR_ERR(in); goto out; } - dn = splice_dentry(dn, in, NULL, false); - if (IS_ERR(dn)) - dn = NULL; } if (fill_inode(in, &rinfo->dir_in[i], NULL, session, req->r_request_started, -1, &req->r_caps_reservation) < 0) { pr_err("fill_inode badness on %p\n", in); + if (!dn->d_inode) + iput(in); + d_drop(dn); goto next_item; } - if (dn) - update_dentry_lease(dn, rinfo->dir_dlease[i], - req->r_session, - req->r_request_started); + + if (!dn->d_inode) { + dn = splice_dentry(dn, in, NULL, false); + if (IS_ERR(dn)) { + err = PTR_ERR(dn); + dn = NULL; + goto next_item; + } + } + + di = dn->d_fsdata; + di->offset = ceph_make_fpos(frag, i + r_readdir_offset); + + update_dentry_lease(dn, rinfo->dir_dlease[i], + req->r_session, + req->r_request_started); next_item: if (dn) dput(dn); } - req->r_did_prepopulate = true; + if (err == 0) + req->r_did_prepopulate = true; out: if (snapdir) { -- cgit v0.10.2 From 56f91aad69444d650237295f68c195b74d888d95 Mon Sep 17 00:00:00 2001 From: Li Wang Date: Wed, 13 Nov 2013 15:22:14 +0800 Subject: ceph: Avoid data inconsistency due to d-cache aliasing in readpage() If the length of data to be read in readpage() is exactly PAGE_CACHE_SIZE, the original code does not flush d-cache for data consistency after finishing reading. This patches fixes this. Signed-off-by: Li Wang Signed-off-by: Sage Weil diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 1e561c0..ec3ba43 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -210,9 +210,13 @@ static int readpage_nounlock(struct file *filp, struct page *page) if (err < 0) { SetPageError(page); goto out; - } else if (err < PAGE_CACHE_SIZE) { + } else { + if (err < PAGE_CACHE_SIZE) { /* zero fill remainder of page */ - zero_user_segment(page, err, PAGE_CACHE_SIZE); + zero_user_segment(page, err, PAGE_CACHE_SIZE); + } else { + flush_dcache_page(page); + } } SetPageUptodate(page); -- cgit v0.10.2 From cb12057256ec58b6f798855b27e120579d6c9aee Mon Sep 17 00:00:00 2001 From: Tomasz Figa Date: Fri, 13 Dec 2013 20:59:39 +0100 Subject: ARM: s3c64xx: dt: Fix boot failure due to double clock initialization Commit 4178bac ARM: call of_clk_init from default time_init handler added implicit call to of_clk_init() from default time_init callback, but it did not change platforms calling it from other callbacks, despite of not having custom time_init callbacks. This caused double clock initialization on such platforms, leading to boot failures. An example of such platform is mach-s3c64xx. This patch fixes boot failure on s3c64xx by dropping custom init_irq callback, which had a call to of_clk_init() and moving system reset initialization to init_machine callback. This allows us to have clocks initialized properly without a need to have custom init_time or init_irq callbacks. Signed-off-by: Tomasz Figa Signed-off-by: Olof Johansson diff --git a/arch/arm/mach-s3c64xx/mach-s3c64xx-dt.c b/arch/arm/mach-s3c64xx/mach-s3c64xx-dt.c index 7eb9a10..2fddf38 100644 --- a/arch/arm/mach-s3c64xx/mach-s3c64xx-dt.c +++ b/arch/arm/mach-s3c64xx/mach-s3c64xx-dt.c @@ -8,8 +8,6 @@ * published by the Free Software Foundation. */ -#include -#include #include #include @@ -48,15 +46,9 @@ static void __init s3c64xx_dt_map_io(void) panic("SoC is not S3C64xx!"); } -static void __init s3c64xx_dt_init_irq(void) -{ - of_clk_init(NULL); - samsung_wdt_reset_of_init(); - irqchip_init(); -}; - static void __init s3c64xx_dt_init_machine(void) { + samsung_wdt_reset_of_init(); of_platform_populate(NULL, of_default_bus_match_table, NULL, NULL); } @@ -79,7 +71,6 @@ DT_MACHINE_START(S3C6400_DT, "Samsung S3C64xx (Flattened Device Tree)") /* Maintainer: Tomasz Figa */ .dt_compat = s3c64xx_dt_compat, .map_io = s3c64xx_dt_map_io, - .init_irq = s3c64xx_dt_init_irq, .init_machine = s3c64xx_dt_init_machine, .restart = s3c64xx_dt_restart, MACHINE_END -- cgit v0.10.2 From ce027ed98fd176710fb14be9d6015697b62436f0 Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Sun, 15 Dec 2013 16:18:01 +0100 Subject: firewire: sbp2: bring back WRITE SAME support Commit 54b2b50c20a6 "[SCSI] Disable WRITE SAME for RAID and virtual host adapter drivers" disabled WRITE SAME support for all SBP-2 attached targets. But as described in the changelog of commit b0ea5f19d3d8 "firewire: sbp2: allow WRITE SAME and REPORT SUPPORTED OPERATION CODES", it is not required to blacklist WRITE SAME. Bring the feature back by reverting the sbp2.c hunk of commit 54b2b50c20a6. Signed-off-by: Stefan Richter Cc: stable@kernel.org diff --git a/drivers/firewire/sbp2.c b/drivers/firewire/sbp2.c index b0bb056..281029d 100644 --- a/drivers/firewire/sbp2.c +++ b/drivers/firewire/sbp2.c @@ -1623,7 +1623,6 @@ static struct scsi_host_template scsi_driver_template = { .cmd_per_lun = 1, .can_queue = 1, .sdev_attrs = sbp2_scsi_sysfs_attrs, - .no_write_same = 1, }; MODULE_AUTHOR("Kristian Hoegsberg "); -- cgit v0.10.2 From 128d6637cce0747e21b8936e449b1c78a497b189 Mon Sep 17 00:00:00 2001 From: Beomho Seo Date: Mon, 9 Dec 2013 02:17:00 +0000 Subject: iio: cm36651: Changed return value of read function A return value of callback have been changed to IIO_VAL_INT. If not IIO_VAL_INT, driver will print wrong value(*_read_int_time). A follow up patch will deal with a related bug in the new event handling code. Signed-off-by: Beomho Seo Signed-off-by: Jonathan Cameron diff --git a/drivers/iio/light/cm36651.c b/drivers/iio/light/cm36651.c index 21df571..0922e39 100644 --- a/drivers/iio/light/cm36651.c +++ b/drivers/iio/light/cm36651.c @@ -387,7 +387,7 @@ static int cm36651_read_int_time(struct cm36651_data *cm36651, return -EINVAL; } - return IIO_VAL_INT_PLUS_MICRO; + return IIO_VAL_INT; } static int cm36651_write_int_time(struct cm36651_data *cm36651, -- cgit v0.10.2 From b25b4427e9dfba073cf9bc86603956ed395eb6e3 Mon Sep 17 00:00:00 2001 From: Ilia Mirkin Date: Wed, 11 Dec 2013 22:19:01 -0500 Subject: drm/nouveau: only runtime suspend by default in optimus configuration The intent was to only enable it by default for optimus, e.g. see the runtime_idle callback. The suspend callback may be called directly, e.g. as a result of nouveau_crtc_set_config. Reported-by: Stefan Lippers-Hollmann Signed-off-by: Ilia Mirkin Tested-by: Stefan Lippers-Hollmann Cc: stable@vger.kernel.org Signed-off-by: Dave Airlie diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index 7a3759f..98a22e6 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -858,6 +858,12 @@ static int nouveau_pmops_runtime_suspend(struct device *dev) if (nouveau_runtime_pm == 0) return -EINVAL; + /* are we optimus enabled? */ + if (nouveau_runtime_pm == -1 && !nouveau_is_optimus() && !nouveau_is_v1_dsm()) { + DRM_DEBUG_DRIVER("failing to power off - not optimus\n"); + return -EINVAL; + } + nv_debug_level(SILENT); drm_kms_helper_poll_disable(drm_dev); vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_OFF); -- cgit v0.10.2 From 1b1ccee1e821e59c2a45c95b007aeb2c9dafd9be Mon Sep 17 00:00:00 2001 From: Krzysztof Kozlowski Date: Wed, 11 Dec 2013 15:07:43 +0100 Subject: mfd: s2mps11: Fix build after regmap field rename in sec-core.c Fix building of s2mps11 regulator and clock drivers after renaming regmap field in struct sec_pmic_dev in commit: - "mfd/rtc: s5m: Fix register updating by adding regmap for RTC" Signed-off-by: Krzysztof Kozlowski Cc: Kyungmin Park Cc: Marek Szyprowski Signed-off-by: Mark Brown diff --git a/drivers/clk/clk-s2mps11.c b/drivers/clk/clk-s2mps11.c index 7be41e6..00a3abe 100644 --- a/drivers/clk/clk-s2mps11.c +++ b/drivers/clk/clk-s2mps11.c @@ -60,7 +60,7 @@ static int s2mps11_clk_prepare(struct clk_hw *hw) struct s2mps11_clk *s2mps11 = to_s2mps11_clk(hw); int ret; - ret = regmap_update_bits(s2mps11->iodev->regmap, + ret = regmap_update_bits(s2mps11->iodev->regmap_pmic, S2MPS11_REG_RTC_CTRL, s2mps11->mask, s2mps11->mask); if (!ret) @@ -74,7 +74,7 @@ static void s2mps11_clk_unprepare(struct clk_hw *hw) struct s2mps11_clk *s2mps11 = to_s2mps11_clk(hw); int ret; - ret = regmap_update_bits(s2mps11->iodev->regmap, S2MPS11_REG_RTC_CTRL, + ret = regmap_update_bits(s2mps11->iodev->regmap_pmic, S2MPS11_REG_RTC_CTRL, s2mps11->mask, ~s2mps11->mask); if (!ret) @@ -174,7 +174,7 @@ static int s2mps11_clk_probe(struct platform_device *pdev) s2mps11_clk->hw.init = &s2mps11_clks_init[i]; s2mps11_clk->mask = 1 << i; - ret = regmap_read(s2mps11_clk->iodev->regmap, + ret = regmap_read(s2mps11_clk->iodev->regmap_pmic, S2MPS11_REG_RTC_CTRL, &val); if (ret < 0) goto err_reg; diff --git a/drivers/regulator/s2mps11.c b/drivers/regulator/s2mps11.c index 333677d..9e61922 100644 --- a/drivers/regulator/s2mps11.c +++ b/drivers/regulator/s2mps11.c @@ -438,7 +438,7 @@ common_reg: platform_set_drvdata(pdev, s2mps11); config.dev = &pdev->dev; - config.regmap = iodev->regmap; + config.regmap = iodev->regmap_pmic; config.driver_data = s2mps11; for (i = 0; i < S2MPS11_REGULATOR_MAX; i++) { if (!reg_np) { -- cgit v0.10.2 From c4602c1c818bd6626178d6d3fcc152d9f2f48ac0 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Mon, 16 Dec 2013 15:20:01 +0800 Subject: ftrace: Initialize the ftrace profiler for each possible cpu Ftrace currently initializes only the online CPUs. This implementation has two problems: - If we online a CPU after we enable the function profile, and then run the test, we will lose the trace information on that CPU. Steps to reproduce: # echo 0 > /sys/devices/system/cpu/cpu1/online # cd /tracing/ # echo >> set_ftrace_filter # echo 1 > function_profile_enabled # echo 1 > /sys/devices/system/cpu/cpu1/online # run test - If we offline a CPU before we enable the function profile, we will not clear the trace information when we enable the function profile. It will trouble the users. Steps to reproduce: # cd /tracing/ # echo >> set_ftrace_filter # echo 1 > function_profile_enabled # run test # cat trace_stat/function* # echo 0 > /sys/devices/system/cpu/cpu1/online # echo 0 > function_profile_enabled # echo 1 > function_profile_enabled # cat trace_stat/function* # run test # cat trace_stat/function* So it is better that we initialize the ftrace profiler for each possible cpu every time we enable the function profile instead of just the online ones. Link: http://lkml.kernel.org/r/1387178401-10619-1-git-send-email-miaox@cn.fujitsu.com Cc: stable@vger.kernel.org # 2.6.31+ Signed-off-by: Miao Xie Signed-off-by: Steven Rostedt diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 0e9f9ea..72a0f81 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -775,7 +775,7 @@ static int ftrace_profile_init(void) int cpu; int ret = 0; - for_each_online_cpu(cpu) { + for_each_possible_cpu(cpu) { ret = ftrace_profile_init_cpu(cpu); if (ret) break; -- cgit v0.10.2 From d386735588c3e22129c2bc6eb64fc1d37a8f805c Mon Sep 17 00:00:00 2001 From: Thomas Hellstrom Date: Sun, 8 Dec 2013 23:23:57 -0800 Subject: drm/ttm: Fix accesses through vmas with only partial coverage VMAs covering a bo but that didn't start at the same address space offset as the bo they were mapping were incorrectly generating SEGFAULT errors in the fault handler. Reported-by: Joseph Dolinak Signed-off-by: Thomas Hellstrom Reviewed-by: Jakob Bornecrantz Cc: stable@vger.kernel.org diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c index b249ab9..6440eea 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_vm.c +++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c @@ -169,9 +169,9 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) } page_offset = ((address - vma->vm_start) >> PAGE_SHIFT) + - drm_vma_node_start(&bo->vma_node) - vma->vm_pgoff; - page_last = vma_pages(vma) + - drm_vma_node_start(&bo->vma_node) - vma->vm_pgoff; + vma->vm_pgoff - drm_vma_node_start(&bo->vma_node); + page_last = vma_pages(vma) + vma->vm_pgoff - + drm_vma_node_start(&bo->vma_node); if (unlikely(page_offset >= bo->num_pages)) { retval = VM_FAULT_SIGBUS; -- cgit v0.10.2 From 317929cd8ef32121fb6574f12d02621f78f24732 Mon Sep 17 00:00:00 2001 From: Santosh Shilimkar Date: Sat, 23 Nov 2013 17:31:27 -0500 Subject: MAINTAINERS: Add keystone git tree information Update the Keystone entry to add git tree information. Signed-off-by: Santosh Shilimkar diff --git a/MAINTAINERS b/MAINTAINERS index 8285ed4..25b945b 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1013,6 +1013,7 @@ M: Santosh Shilimkar L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) S: Maintained F: arch/arm/mach-keystone/ +T: git git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git ARM/LOGICPD PXA270 MACHINE SUPPORT M: Lennert Buytenhek -- cgit v0.10.2 From cffa8e3b5c7dcf7617268561397eb236651446f8 Mon Sep 17 00:00:00 2001 From: Santosh Shilimkar Date: Sat, 23 Nov 2013 17:35:23 -0500 Subject: MAINTAINERS: Add keystone clock drivers Cc: Mike Turquette Signed-off-by: Santosh Shilimkar diff --git a/MAINTAINERS b/MAINTAINERS index 25b945b..249651f 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1013,6 +1013,7 @@ M: Santosh Shilimkar L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) S: Maintained F: arch/arm/mach-keystone/ +F: drivers/clk/keystone/ T: git git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git ARM/LOGICPD PXA270 MACHINE SUPPORT -- cgit v0.10.2 From cf872776fc84128bb779ce2b83a37c884c3203ae Mon Sep 17 00:00:00 2001 From: Peter Hurley Date: Wed, 11 Dec 2013 21:11:58 -0500 Subject: tty: Fix hang at ldsem_down_read() When a controlling tty is being hung up and the hang up is waiting for a just-signalled tty reader or writer to exit, and a new tty reader/writer tries to acquire an ldisc reference concurrently with the ldisc reference release from the signalled reader/writer, the hangup can hang. The new reader/writer is sleeping in ldsem_down_read() and the hangup is sleeping in ldsem_down_write() [1]. The new reader/writer fails to wakeup the waiting hangup because the wrong lock count value is checked (the old lock count rather than the new lock count) to see if the lock is unowned. Change helper function to return the new lock count if the cmpxchg was successful; document this behavior. [1] edited dmesg log from reporter SysRq : Show Blocked State task PC stack pid father systemd D ffff88040c4f0000 0 1 0 0x00000000 ffff88040c49fbe0 0000000000000046 ffff88040c4a0000 ffff88040c49ffd8 00000000001d3980 00000000001d3980 ffff88040c4a0000 ffff88040593d840 ffff88040c49fb40 ffffffff810a4cc0 0000000000000006 0000000000000023 Call Trace: [] ? sched_clock_cpu+0x9f/0xe4 [] ? sched_clock_cpu+0x9f/0xe4 [] ? sched_clock_cpu+0x9f/0xe4 [] ? sched_clock_cpu+0x9f/0xe4 [] schedule+0x24/0x5e [] schedule_timeout+0x15b/0x1ec [] ? sched_clock_cpu+0x9f/0xe4 [] ? _raw_spin_unlock_irq+0x24/0x26 [] down_read_failed+0xe3/0x1b9 [] ldsem_down_read+0x8b/0xa5 [] ? tty_ldisc_ref_wait+0x1b/0x44 [] tty_ldisc_ref_wait+0x1b/0x44 [] tty_write+0x7d/0x28a [] redirected_tty_write+0x8d/0x98 [] ? tty_write+0x28a/0x28a [] do_loop_readv_writev+0x56/0x79 [] do_readv_writev+0x1b0/0x1ff [] ? do_vfs_ioctl+0x32a/0x489 [] ? final_putname+0x1d/0x3a [] vfs_writev+0x2e/0x49 [] SyS_writev+0x47/0xaa [] system_call_fastpath+0x16/0x1b bash D ffffffff81c104c0 0 5469 5302 0x00000082 ffff8800cf817ac0 0000000000000046 ffff8804086b22a0 ffff8800cf817fd8 00000000001d3980 00000000001d3980 ffff8804086b22a0 ffff8800cf817a48 000000000000b9a0 ffff8800cf817a78 ffffffff81004675 ffff8800cf817a44 Call Trace: [] ? dump_trace+0x165/0x29c [] ? sched_clock_cpu+0x9f/0xe4 [] ? save_stack_trace+0x26/0x41 [] schedule+0x24/0x5e [] schedule_timeout+0x15b/0x1ec [] ? sched_clock_cpu+0x9f/0xe4 [] ? down_write_failed+0xa3/0x1c9 [] ? _raw_spin_unlock_irq+0x24/0x26 [] down_write_failed+0xab/0x1c9 [] ldsem_down_write+0x79/0xb1 [] ? tty_ldisc_lock_pair_timeout+0xa5/0xd9 [] tty_ldisc_lock_pair_timeout+0xa5/0xd9 [] tty_ldisc_hangup+0xc4/0x218 [] __tty_hangup+0x2e2/0x3ed [] disassociate_ctty+0x63/0x226 [] do_exit+0x79f/0xa11 [] ? get_signal_to_deliver+0x206/0x62f [] ? lock_release_holdtime.part.8+0xf/0x16e [] do_group_exit+0x47/0xb5 [] get_signal_to_deliver+0x241/0x62f [] do_signal+0x43/0x59d [] ? __audit_syscall_exit+0x21a/0x2a8 [] ? lock_release_holdtime.part.8+0xf/0x16e [] do_notify_resume+0x54/0x6c [] int_signal+0x12/0x17 Reported-by: Sami Farin Cc: # 3.12.x Signed-off-by: Peter Hurley Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/tty/tty_ldsem.c b/drivers/tty/tty_ldsem.c index 22fad8a..d8a55e8 100644 --- a/drivers/tty/tty_ldsem.c +++ b/drivers/tty/tty_ldsem.c @@ -86,11 +86,21 @@ static inline long ldsem_atomic_update(long delta, struct ld_semaphore *sem) return atomic_long_add_return(delta, (atomic_long_t *)&sem->count); } +/* + * ldsem_cmpxchg() updates @*old with the last-known sem->count value. + * Returns 1 if count was successfully changed; @*old will have @new value. + * Returns 0 if count was not changed; @*old will have most recent sem->count + */ static inline int ldsem_cmpxchg(long *old, long new, struct ld_semaphore *sem) { - long tmp = *old; - *old = atomic_long_cmpxchg(&sem->count, *old, new); - return *old == tmp; + long tmp = atomic_long_cmpxchg(&sem->count, *old, new); + if (tmp == *old) { + *old = new; + return 1; + } else { + *old = tmp; + return 0; + } } /* -- cgit v0.10.2 From 9bd7d20c45157c175bfd45449e904a404506f99c Mon Sep 17 00:00:00 2001 From: wangweidong Date: Fri, 13 Dec 2013 11:00:10 +0800 Subject: sctp: loading sctp when load sctp_probe when I modprobe sctp_probe, it failed with "FATAL: ". I found that sctp should load before sctp_probe register jprobe. So I add a sctp_setup_jprobe for loading 'sctp' when first failed to register jprobe, just do this similar to dccp_probe. v2: add MODULE_SOFTDEP and check of request_module, as suggested by Neil Signed-off-by: Wang Weidong Acked-by: Neil Horman Signed-off-by: David S. Miller diff --git a/net/sctp/probe.c b/net/sctp/probe.c index 53c452e..5e68b94 100644 --- a/net/sctp/probe.c +++ b/net/sctp/probe.c @@ -38,6 +38,7 @@ #include #include +MODULE_SOFTDEP("pre: sctp"); MODULE_AUTHOR("Wei Yongjun "); MODULE_DESCRIPTION("SCTP snooper"); MODULE_LICENSE("GPL"); @@ -182,6 +183,20 @@ static struct jprobe sctp_recv_probe = { .entry = jsctp_sf_eat_sack, }; +static __init int sctp_setup_jprobe(void) +{ + int ret = register_jprobe(&sctp_recv_probe); + + if (ret) { + if (request_module("sctp")) + goto out; + ret = register_jprobe(&sctp_recv_probe); + } + +out: + return ret; +} + static __init int sctpprobe_init(void) { int ret = -ENOMEM; @@ -202,7 +217,7 @@ static __init int sctpprobe_init(void) &sctpprobe_fops)) goto free_kfifo; - ret = register_jprobe(&sctp_recv_probe); + ret = sctp_setup_jprobe(); if (ret) goto remove_proc; -- cgit v0.10.2 From 6979f8d28049879e6147767d93ba6732c8bd94f4 Mon Sep 17 00:00:00 2001 From: James Hogan Date: Tue, 10 Dec 2013 22:28:04 +0000 Subject: serial: 8250_dw: Fix LCR workaround regression Commit c49436b657d0 (serial: 8250_dw: Improve unwritable LCR workaround) caused a regression. It added a check that the LCR was written properly to detect and workaround the busy quirk, but the behaviour of bit 5 (UART_LCR_SPAR) differs between IP versions 3.00a and 3.14c per the docs. On older versions this caused the check to fail and it would repeatedly force idle and rewrite the LCR register, causing delays and preventing any input from serial being received. This is fixed by masking out UART_LCR_SPAR before making the comparison. Signed-off-by: James Hogan Cc: Greg Kroah-Hartman Cc: Jiri Slaby Cc: Tim Kryger Cc: Ezequiel Garcia Cc: Matt Porter Cc: Markus Mayer Tested-by: Tim Kryger Tested-by: Ezequiel Garcia Tested-by: Heikki Krogerus Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/tty/serial/8250/8250_dw.c b/drivers/tty/serial/8250/8250_dw.c index 4658e3e..5f096c7 100644 --- a/drivers/tty/serial/8250/8250_dw.c +++ b/drivers/tty/serial/8250/8250_dw.c @@ -96,7 +96,8 @@ static void dw8250_serial_out(struct uart_port *p, int offset, int value) if (offset == UART_LCR) { int tries = 1000; while (tries--) { - if (value == p->serial_in(p, UART_LCR)) + unsigned int lcr = p->serial_in(p, UART_LCR); + if ((value & ~UART_LCR_SPAR) == (lcr & ~UART_LCR_SPAR)) return; dw8250_force_idle(p); writeb(value, p->membase + (UART_LCR << p->regshift)); @@ -132,7 +133,8 @@ static void dw8250_serial_out32(struct uart_port *p, int offset, int value) if (offset == UART_LCR) { int tries = 1000; while (tries--) { - if (value == p->serial_in(p, UART_LCR)) + unsigned int lcr = p->serial_in(p, UART_LCR); + if ((value & ~UART_LCR_SPAR) == (lcr & ~UART_LCR_SPAR)) return; dw8250_force_idle(p); writel(value, p->membase + (UART_LCR << p->regshift)); -- cgit v0.10.2 From 49d45a31b71d7d9da74485922bdb63faf3dc9684 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= Date: Sat, 7 Dec 2013 13:22:42 +0100 Subject: drm/edid: add quirk for BPC in Samsung NP700G7A-S01PL notebook MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This bug in EDID was exposed by: commit eccea7920cfb009c2fa40e9ecdce8c36f61cab66 Author: Alex Deucher Date: Mon Mar 26 15:12:54 2012 -0400 drm/radeon/kms: improve bpc handling (v2) Which resulted in kind of regression in 3.5. This fixes https://bugs.freedesktop.org/show_bug.cgi?id=70934 Signed-off-by: Rafał Miłecki Cc: stable@vger.kernel.org Signed-off-by: Dave Airlie diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c index 0a1e4a5..8835dcd 100644 --- a/drivers/gpu/drm/drm_edid.c +++ b/drivers/gpu/drm/drm_edid.c @@ -68,6 +68,8 @@ #define EDID_QUIRK_DETAILED_SYNC_PP (1 << 6) /* Force reduced-blanking timings for detailed modes */ #define EDID_QUIRK_FORCE_REDUCED_BLANKING (1 << 7) +/* Force 8bpc */ +#define EDID_QUIRK_FORCE_8BPC (1 << 8) struct detailed_mode_closure { struct drm_connector *connector; @@ -128,6 +130,9 @@ static struct edid_quirk { /* Medion MD 30217 PG */ { "MED", 0x7b8, EDID_QUIRK_PREFER_LARGE_75 }, + + /* Panel in Samsung NP700G7A-S01PL notebook reports 6bpc */ + { "SEC", 0xd033, EDID_QUIRK_FORCE_8BPC }, }; /* @@ -3435,6 +3440,9 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid) drm_add_display_info(edid, &connector->display_info); + if (quirks & EDID_QUIRK_FORCE_8BPC) + connector->display_info.bpc = 8; + return num_modes; } EXPORT_SYMBOL(drm_add_edid_modes); -- cgit v0.10.2 From 7cd0c298f6e09476f474b7ae6c1ca876c5d7f881 Mon Sep 17 00:00:00 2001 From: Felipe Balbi Date: Tue, 10 Dec 2013 17:00:06 -0600 Subject: usb: phy: fix driver dependencies both isp1301-omap and fsl_usb2_otg drivers depend on usb_bus_start_enum() which is only defined if CONFIG_USB != n. There is a problem, however, where both those drivers could be statically linked, while CONFIG_USB=m. Fix the problem by fixing driver dependency. Signed-off-by: Felipe Balbi Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/usb/phy/Kconfig b/drivers/usb/phy/Kconfig index 08e2f39..2b41c63 100644 --- a/drivers/usb/phy/Kconfig +++ b/drivers/usb/phy/Kconfig @@ -19,8 +19,9 @@ config AB8500_USB in host mode, low speed. config FSL_USB2_OTG - bool "Freescale USB OTG Transceiver Driver" + tristate "Freescale USB OTG Transceiver Driver" depends on USB_EHCI_FSL && USB_FSL_USB2 && PM_RUNTIME + depends on USB select USB_OTG select USB_PHY help @@ -29,6 +30,7 @@ config FSL_USB2_OTG config ISP1301_OMAP tristate "Philips ISP1301 with OMAP OTG" depends on I2C && ARCH_OMAP_OTG + depends on USB select USB_PHY help If you say yes here you get support for the Philips ISP1301 -- cgit v0.10.2 From a1c31f1d057130cc63e72a09189410d169db7ecf Mon Sep 17 00:00:00 2001 From: Alexey Khoroshilov Date: Sat, 7 Dec 2013 03:10:35 +0400 Subject: can: ems_usb: fix urb leaks on failure paths There are a couple failure paths where urb leaks. Is spare code within ems_usb_start_xmit(), usb_free_urb() should be used to deallocate urb instead of usb_unanchor_urb(). In ems_usb_start() there is no usb_free_urb() if usb_submit_urb() fails. Found by Linux Driver Verification project (linuxtesting.org). Signed-off-by: Alexey Khoroshilov Acked-by: Sebastian Haas Signed-off-by: Marc Kleine-Budde diff --git a/drivers/net/can/usb/ems_usb.c b/drivers/net/can/usb/ems_usb.c index 5f9a7ad..8aeec0b 100644 --- a/drivers/net/can/usb/ems_usb.c +++ b/drivers/net/can/usb/ems_usb.c @@ -625,6 +625,7 @@ static int ems_usb_start(struct ems_usb *dev) usb_unanchor_urb(urb); usb_free_coherent(dev->udev, RX_BUFFER_SIZE, buf, urb->transfer_dma); + usb_free_urb(urb); break; } @@ -798,8 +799,8 @@ static netdev_tx_t ems_usb_start_xmit(struct sk_buff *skb, struct net_device *ne * allowed (MAX_TX_URBS). */ if (!context) { - usb_unanchor_urb(urb); usb_free_coherent(dev->udev, size, buf, urb->transfer_dma); + usb_free_urb(urb); netdev_warn(netdev, "couldn't find free context\n"); -- cgit v0.10.2 From 20fb4eb96fb0350d28fc4d7cbfd5506711079592 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Sat, 14 Dec 2013 14:36:25 +0100 Subject: can: peak_usb: fix mem leak in pcan_usb_pro_init() This patch fixes a memory leak in pcan_usb_pro_init(). In patch f14e224 net: can: peak_usb: Do not do dma on the stack the struct pcan_usb_pro_fwinfo *fi and struct pcan_usb_pro_blinfo *bi were converted from stack to dynamic allocation va kmalloc(). However the corresponding kfree() was not introduced. This patch adds the missing kfree(). Cc: linux-stable # v3.10 Reported-by: Stephane Grosjean Signed-off-by: Marc Kleine-Budde diff --git a/drivers/net/can/usb/peak_usb/pcan_usb_pro.c b/drivers/net/can/usb/peak_usb/pcan_usb_pro.c index 8ee9d15..263dd92 100644 --- a/drivers/net/can/usb/peak_usb/pcan_usb_pro.c +++ b/drivers/net/can/usb/peak_usb/pcan_usb_pro.c @@ -927,6 +927,9 @@ static int pcan_usb_pro_init(struct peak_usb_device *dev) /* set LED in default state (end of init phase) */ pcan_usb_pro_set_led(dev, 0, 1); + kfree(bi); + kfree(fi); + return 0; err_out: -- cgit v0.10.2 From 280484e708a3cc38fe6807718caa460e744c0b20 Mon Sep 17 00:00:00 2001 From: Charles Keepax Date: Tue, 17 Dec 2013 13:16:16 +0000 Subject: ASoC: wm5110: Correct HPOUT3 DAPM route typo Reported-by: Kyung-Kwee Ryu Signed-off-by: Charles Keepax Signed-off-by: Mark Brown Cc: stable@vger.kernel.org diff --git a/sound/soc/codecs/wm5110.c b/sound/soc/codecs/wm5110.c index c3c7396..b3d0b28 100644 --- a/sound/soc/codecs/wm5110.c +++ b/sound/soc/codecs/wm5110.c @@ -1037,7 +1037,7 @@ static const struct snd_soc_dapm_route wm5110_dapm_routes[] = { { "AEC Loopback", "HPOUT3L", "OUT3L" }, { "AEC Loopback", "HPOUT3R", "OUT3R" }, { "HPOUT3L", NULL, "OUT3L" }, - { "HPOUT3R", NULL, "OUT3L" }, + { "HPOUT3R", NULL, "OUT3R" }, { "AEC Loopback", "SPKOUTL", "OUT4L" }, { "SPKOUTLN", NULL, "OUT4L" }, -- cgit v0.10.2 From 443772776c69ac9293d66b4d69fd9af16299cc2a Mon Sep 17 00:00:00 2001 From: Alexander Shishkin Date: Mon, 16 Dec 2013 14:17:36 +0200 Subject: perf: Disable all pmus on unthrottling and rescheduling Currently, only one PMU in a context gets disabled during unthrottling and event_sched_{out,in}(), however, events in one context may belong to different pmus, which results in PMUs being reprogrammed while they are still enabled. This means that mixed PMU use [which is rare in itself] resulted in potentially completely unreliable results: corrupted events, bogus results, etc. This patch temporarily disables PMUs that correspond to each event in the context while these events are being modified. Signed-off-by: Alexander Shishkin Reviewed-by: Andi Kleen Signed-off-by: Peter Zijlstra Cc: Frederic Weisbecker Cc: Mike Galbraith Cc: Paul Mackerras Cc: Stephane Eranian Cc: Alexander Shishkin Link: http://lkml.kernel.org/r/1387196256-8030-1-git-send-email-alexander.shishkin@linux.intel.com Signed-off-by: Ingo Molnar diff --git a/kernel/events/core.c b/kernel/events/core.c index 72348dc..f574401 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1396,6 +1396,8 @@ event_sched_out(struct perf_event *event, if (event->state != PERF_EVENT_STATE_ACTIVE) return; + perf_pmu_disable(event->pmu); + event->state = PERF_EVENT_STATE_INACTIVE; if (event->pending_disable) { event->pending_disable = 0; @@ -1412,6 +1414,8 @@ event_sched_out(struct perf_event *event, ctx->nr_freq--; if (event->attr.exclusive || !cpuctx->active_oncpu) cpuctx->exclusive = 0; + + perf_pmu_enable(event->pmu); } static void @@ -1652,6 +1656,7 @@ event_sched_in(struct perf_event *event, struct perf_event_context *ctx) { u64 tstamp = perf_event_time(event); + int ret = 0; if (event->state <= PERF_EVENT_STATE_OFF) return 0; @@ -1674,10 +1679,13 @@ event_sched_in(struct perf_event *event, */ smp_wmb(); + perf_pmu_disable(event->pmu); + if (event->pmu->add(event, PERF_EF_START)) { event->state = PERF_EVENT_STATE_INACTIVE; event->oncpu = -1; - return -EAGAIN; + ret = -EAGAIN; + goto out; } event->tstamp_running += tstamp - event->tstamp_stopped; @@ -1693,7 +1701,10 @@ event_sched_in(struct perf_event *event, if (event->attr.exclusive) cpuctx->exclusive = 1; - return 0; +out: + perf_pmu_enable(event->pmu); + + return ret; } static int @@ -2743,6 +2754,8 @@ static void perf_adjust_freq_unthr_context(struct perf_event_context *ctx, if (!event_filter_match(event)) continue; + perf_pmu_disable(event->pmu); + hwc = &event->hw; if (hwc->interrupts == MAX_INTERRUPTS) { @@ -2752,7 +2765,7 @@ static void perf_adjust_freq_unthr_context(struct perf_event_context *ctx, } if (!event->attr.freq || !event->attr.sample_freq) - continue; + goto next; /* * stop the event and update event->count @@ -2774,6 +2787,8 @@ static void perf_adjust_freq_unthr_context(struct perf_event_context *ctx, perf_adjust_period(event, period, delta, false); event->pmu->start(event, delta > 0 ? PERF_EF_RELOAD : 0); + next: + perf_pmu_enable(event->pmu); } perf_pmu_enable(ctx->pmu); -- cgit v0.10.2 From 189b84fb54490ae24111124346a8e63f8e019385 Mon Sep 17 00:00:00 2001 From: Vince Weaver Date: Fri, 13 Dec 2013 15:52:25 -0500 Subject: perf: Document the new transaction sample type Commit fdfbbd07e91f8fe3871 ("perf: Add generic transaction flags") added support for PERF_SAMPLE_TRANSACTION but forgot to add documentation for the sample type to include/uapi/linux/perf_event.h Signed-off-by: Vince Weaver Signed-off-by: Peter Zijlstra Cc: Andi Kleen Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1312131548450.10372@pianoman.cluster.toy Signed-off-by: Ingo Molnar diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h index e1802d6..959d454 100644 --- a/include/uapi/linux/perf_event.h +++ b/include/uapi/linux/perf_event.h @@ -679,6 +679,7 @@ enum perf_event_type { * * { u64 weight; } && PERF_SAMPLE_WEIGHT * { u64 data_src; } && PERF_SAMPLE_DATA_SRC + * { u64 transaction; } && PERF_SAMPLE_TRANSACTION * }; */ PERF_RECORD_SAMPLE = 9, -- cgit v0.10.2 From 5d4cf996cf134e8ddb4f906b8197feb9267c2b77 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Tue, 17 Dec 2013 09:21:25 +0000 Subject: sched: Assign correct scheduling domain to 'sd_llc' Commit 42eb088e (sched: Avoid NULL dereference on sd_busy) corrected a NULL dereference on sd_busy but the fix also altered what scheduling domain it used for the 'sd_llc' percpu variable. One impact of this is that a task selecting a runqueue may consider idle CPUs that are not cache siblings as candidates for running. Tasks are then running on CPUs that are not cache hot. This was found through bisection where ebizzy threads were not seeing equal performance and it looked like a scheduling fairness issue. This patch mitigates but does not completely fix the problem on all machines tested implying there may be an additional bug or a common root cause. Here are the average range of performance seen by individual ebizzy threads. It was tested on top of candidate patches related to x86 TLB range flushing. 4-core machine 3.13.0-rc3 3.13.0-rc3 vanilla fixsd-v3r3 Mean 1 0.00 ( 0.00%) 0.00 ( 0.00%) Mean 2 0.34 ( 0.00%) 0.10 ( 70.59%) Mean 3 1.29 ( 0.00%) 0.93 ( 27.91%) Mean 4 7.08 ( 0.00%) 0.77 ( 89.12%) Mean 5 193.54 ( 0.00%) 2.14 ( 98.89%) Mean 6 151.12 ( 0.00%) 2.06 ( 98.64%) Mean 7 115.38 ( 0.00%) 2.04 ( 98.23%) Mean 8 108.65 ( 0.00%) 1.92 ( 98.23%) 8-core machine Mean 1 0.00 ( 0.00%) 0.00 ( 0.00%) Mean 2 0.40 ( 0.00%) 0.21 ( 47.50%) Mean 3 23.73 ( 0.00%) 0.89 ( 96.25%) Mean 4 12.79 ( 0.00%) 1.04 ( 91.87%) Mean 5 13.08 ( 0.00%) 2.42 ( 81.50%) Mean 6 23.21 ( 0.00%) 69.46 (-199.27%) Mean 7 15.85 ( 0.00%) 101.72 (-541.77%) Mean 8 109.37 ( 0.00%) 19.13 ( 82.51%) Mean 12 124.84 ( 0.00%) 28.62 ( 77.07%) Mean 16 113.50 ( 0.00%) 24.16 ( 78.71%) It's eliminated for one machine and reduced for another. Signed-off-by: Mel Gorman Signed-off-by: Peter Zijlstra Cc: Alex Shi Cc: Andrew Morton Cc: Fengguang Wu Cc: H Peter Anvin Cc: Linus Torvalds Link: http://lkml.kernel.org/r/20131217092124.GV11295@suse.de Signed-off-by: Ingo Molnar diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 19af58f..a88f4a4 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4902,6 +4902,7 @@ DEFINE_PER_CPU(struct sched_domain *, sd_asym); static void update_top_cache_domain(int cpu) { struct sched_domain *sd; + struct sched_domain *busy_sd = NULL; int id = cpu; int size = 1; @@ -4909,9 +4910,9 @@ static void update_top_cache_domain(int cpu) if (sd) { id = cpumask_first(sched_domain_span(sd)); size = cpumask_weight(sched_domain_span(sd)); - sd = sd->parent; /* sd_busy */ + busy_sd = sd->parent; /* sd_busy */ } - rcu_assign_pointer(per_cpu(sd_busy, cpu), sd); + rcu_assign_pointer(per_cpu(sd_busy, cpu), busy_sd); rcu_assign_pointer(per_cpu(sd_llc, cpu), sd); per_cpu(sd_llc_size, cpu) = size; -- cgit v0.10.2 From 757dfcaa41844595964f1220f1d33182dae49976 Mon Sep 17 00:00:00 2001 From: Kirill Tkhai Date: Wed, 27 Nov 2013 19:59:13 +0400 Subject: sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities This patch touches the RT group scheduling case. Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority, while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because changing of priority on a child level does not guarantee that the priority is the highest all over the rq. So, this leak makes RT balancing unusable. The short example: the task having the highest priority among all rq's RT tasks (no one other task has the same priority) are waking on a throttle rt_rq. The rq's cpupri is set to the task's priority equivalent, but real rq->rt.highest_prio.curr is less. The patch below fixes the problem. Signed-off-by: Kirill Tkhai Signed-off-by: Peter Zijlstra CC: Steven Rostedt CC: stable@vger.kernel.org Link: http://lkml.kernel.org/r/49231385567953@web4m.yandex.ru Signed-off-by: Ingo Molnar diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 7d57275..1c40655 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -901,6 +901,13 @@ inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) { struct rq *rq = rq_of_rt_rq(rt_rq); +#ifdef CONFIG_RT_GROUP_SCHED + /* + * Change rq's cpupri only if rt_rq is the top queue. + */ + if (&rq->rt != rt_rq) + return; +#endif if (rq->online && prio < prev_prio) cpupri_set(&rq->rd->cpupri, rq->cpu, prio); } @@ -910,6 +917,13 @@ dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) { struct rq *rq = rq_of_rt_rq(rt_rq); +#ifdef CONFIG_RT_GROUP_SCHED + /* + * Change rq's cpupri only if rt_rq is the top queue. + */ + if (&rq->rt != rt_rq) + return; +#endif if (rq->online && rt_rq->highest_prio.curr != prev_prio) cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr); } -- cgit v0.10.2 From 02fc17c10258ad70c1b9a93f8884bdaf0ac3f766 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Wed, 27 Nov 2013 21:10:24 +0100 Subject: ASoC: kirkwood: Fix the CPU DAI rates This patch fixes the rates declared in the CPU DAI parameters: - SNDRV_PCM_RATE_KNOT and the discrete rates SNDRV_PCM_RATE_xxx should not be used with SNDRV_PCM_RATE_CONTINUOUS, - SNDRV_PCM_RATE_CONTINUOUS asks for rate_min and rate_max, - the device may do streaming down to 5512Hz. Signed-off-by: Jean-Francois Moine Signed-off-by: Mark Brown diff --git a/sound/soc/kirkwood/kirkwood-i2s.c b/sound/soc/kirkwood/kirkwood-i2s.c index 0b18f65..3920a5e 100644 --- a/sound/soc/kirkwood/kirkwood-i2s.c +++ b/sound/soc/kirkwood/kirkwood-i2s.c @@ -473,17 +473,17 @@ static struct snd_soc_dai_driver kirkwood_i2s_dai_extclk[2] = { .playback = { .channels_min = 1, .channels_max = 2, - .rates = SNDRV_PCM_RATE_8000_192000 | - SNDRV_PCM_RATE_CONTINUOUS | - SNDRV_PCM_RATE_KNOT, + .rates = SNDRV_PCM_RATE_CONTINUOUS, + .rate_min = 5512, + .rate_max = 192000, .formats = KIRKWOOD_I2S_FORMATS, }, .capture = { .channels_min = 1, .channels_max = 2, - .rates = SNDRV_PCM_RATE_8000_192000 | - SNDRV_PCM_RATE_CONTINUOUS | - SNDRV_PCM_RATE_KNOT, + .rates = SNDRV_PCM_RATE_CONTINUOUS, + .rate_min = 5512, + .rate_max = 192000, .formats = KIRKWOOD_I2S_FORMATS, }, .ops = &kirkwood_i2s_dai_ops, @@ -494,17 +494,17 @@ static struct snd_soc_dai_driver kirkwood_i2s_dai_extclk[2] = { .playback = { .channels_min = 1, .channels_max = 2, - .rates = SNDRV_PCM_RATE_8000_192000 | - SNDRV_PCM_RATE_CONTINUOUS | - SNDRV_PCM_RATE_KNOT, + .rates = SNDRV_PCM_RATE_CONTINUOUS, + .rate_min = 5512, + .rate_max = 192000, .formats = KIRKWOOD_SPDIF_FORMATS, }, .capture = { .channels_min = 1, .channels_max = 2, - .rates = SNDRV_PCM_RATE_8000_192000 | - SNDRV_PCM_RATE_CONTINUOUS | - SNDRV_PCM_RATE_KNOT, + .rates = SNDRV_PCM_RATE_CONTINUOUS, + .rate_min = 5512, + .rate_max = 192000, .formats = KIRKWOOD_SPDIF_FORMATS, }, .ops = &kirkwood_i2s_dai_ops, -- cgit v0.10.2 From 3a8c92086d1c05ae1cf6a52ffd8dde2851e6b1d3 Mon Sep 17 00:00:00 2001 From: Mark Tinguely Date: Sat, 5 Oct 2013 21:48:25 -0500 Subject: xfs: fix memory leak in xfs_dir2_node_removename Fix the leak of kernel memory in xfs_dir2_node_removename() when xfs_dir2_leafn_remove() returns an error code. Signed-off-by: Mark Tinguely Reviewed-by: Ben Myers Signed-off-by: Ben Myers (cherry picked from commit ef701600fd26cace9d513ee174688a2b83832126) diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c index 56369d4..48c7d18 100644 --- a/fs/xfs/xfs_dir2_node.c +++ b/fs/xfs/xfs_dir2_node.c @@ -2067,12 +2067,12 @@ xfs_dir2_node_lookup( */ int /* error */ xfs_dir2_node_removename( - xfs_da_args_t *args) /* operation arguments */ + struct xfs_da_args *args) /* operation arguments */ { - xfs_da_state_blk_t *blk; /* leaf block */ + struct xfs_da_state_blk *blk; /* leaf block */ int error; /* error return value */ int rval; /* operation return value */ - xfs_da_state_t *state; /* btree cursor */ + struct xfs_da_state *state; /* btree cursor */ trace_xfs_dir2_node_removename(args); @@ -2084,19 +2084,18 @@ xfs_dir2_node_removename( state->mp = args->dp->i_mount; state->blocksize = state->mp->m_dirblksize; state->node_ents = state->mp->m_dir_node_ents; - /* - * Look up the entry we're deleting, set up the cursor. - */ + + /* Look up the entry we're deleting, set up the cursor. */ error = xfs_da3_node_lookup_int(state, &rval); if (error) - rval = error; - /* - * Didn't find it, upper layer screwed up. - */ + goto out_free; + + /* Didn't find it, upper layer screwed up. */ if (rval != EEXIST) { - xfs_da_state_free(state); - return rval; + error = rval; + goto out_free; } + blk = &state->path.blk[state->path.active - 1]; ASSERT(blk->magic == XFS_DIR2_LEAFN_MAGIC); ASSERT(state->extravalid); @@ -2107,7 +2106,7 @@ xfs_dir2_node_removename( error = xfs_dir2_leafn_remove(args, blk->bp, blk->index, &state->extrablk, &rval); if (error) - return error; + goto out_free; /* * Fix the hash values up the btree. */ @@ -2122,6 +2121,7 @@ xfs_dir2_node_removename( */ if (!error) error = xfs_dir2_node_to_leaf(state); +out_free: xfs_da_state_free(state); return error; } -- cgit v0.10.2 From 30d161c9aacb4b719789c30623b9eec7d1aa1d08 Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Tue, 26 Nov 2013 21:38:54 +0800 Subject: xfs: fix false assertion at xfs_qm_vop_create_dqattach After the previous fix, there still has another ASSERT failure if turning off any type of quota while fsstress is running at the same time. Backtrace in this case: [ 50.867897] XFS: Assertion failed: XFS_IS_GQUOTA_ON(mp), file: fs/xfs/xfs_qm.c, line: 2118 [ 50.867924] ------------[ cut here ]------------ ... [ 50.867957] Kernel BUG at ffffffffa0b55a32 [verbose debug info unavailable] [ 50.867999] invalid opcode: 0000 [#1] SMP [ 50.869407] Call Trace: [ 50.869446] [] xfs_qm_vop_create_dqattach+0x19a/0x2d0 [xfs] [ 50.869512] [] xfs_create+0x5c5/0x6a0 [xfs] [ 50.869564] [] xfs_vn_mknod+0xac/0x1d0 [xfs] [ 50.869615] [] xfs_vn_mkdir+0x16/0x20 [xfs] [ 50.869655] [] vfs_mkdir+0x95/0x130 [ 50.869689] [] SyS_mkdirat+0xaa/0xe0 [ 50.869723] [] SyS_mkdir+0x19/0x20 [ 50.869757] [] system_call_fastpath+0x1a/0x1f [ 50.869793] Code: 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 55 48 89 [ 50.870003] RIP [] assfail+0x22/0x30 [xfs] [ 50.870050] RSP [ 50.879251] ---[ end trace c93a2b342341c65b ]--- We're hitting the ASSERT(XFS_IS_*QUOTA_ON(mp)) in xfs_qm_vop_create_dqattach(), however the assertion itself is not right IMHO. While performing quota off, we firstly clear the XFS_*QUOTA_ACTIVE bit(s) from struct xfs_mount without taking any special locks, see xfs_qm_scall_quotaoff(). Hence there is no guarantee that the desired quota is still active. Signed-off-by: Jie Liu Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers (cherry picked from commit 37eb9706ebf5b99d14c6086cdeef2c2f73f9c9fb) diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 14a4996..588e490 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -2082,24 +2082,21 @@ xfs_qm_vop_create_dqattach( ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); ASSERT(XFS_IS_QUOTA_RUNNING(mp)); - if (udqp) { + if (udqp && XFS_IS_UQUOTA_ON(mp)) { ASSERT(ip->i_udquot == NULL); - ASSERT(XFS_IS_UQUOTA_ON(mp)); ASSERT(ip->i_d.di_uid == be32_to_cpu(udqp->q_core.d_id)); ip->i_udquot = xfs_qm_dqhold(udqp); xfs_trans_mod_dquot(tp, udqp, XFS_TRANS_DQ_ICOUNT, 1); } - if (gdqp) { + if (gdqp && XFS_IS_GQUOTA_ON(mp)) { ASSERT(ip->i_gdquot == NULL); - ASSERT(XFS_IS_GQUOTA_ON(mp)); ASSERT(ip->i_d.di_gid == be32_to_cpu(gdqp->q_core.d_id)); ip->i_gdquot = xfs_qm_dqhold(gdqp); xfs_trans_mod_dquot(tp, gdqp, XFS_TRANS_DQ_ICOUNT, 1); } - if (pdqp) { + if (pdqp && XFS_IS_PQUOTA_ON(mp)) { ASSERT(ip->i_pdquot == NULL); - ASSERT(XFS_IS_PQUOTA_ON(mp)); ASSERT(xfs_get_projid(ip) == be32_to_cpu(pdqp->q_core.d_id)); ip->i_pdquot = xfs_qm_dqhold(pdqp); -- cgit v0.10.2 From 5c22727895bf61cb851835be0d30260fb36de648 Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Tue, 26 Nov 2013 21:38:34 +0800 Subject: xfs: fix assertion failure at xfs_setattr_nonsize For CRC enabled v5 super block, change a file's ownership can simply trigger an ASSERT failure at xfs_setattr_nonsize() if both group and project quota are enabled, i.e, [ 305.337609] XFS: Assertion failed: !XFS_IS_PQUOTA_ON(mp), file: fs/xfs/xfs_iops.c, line: 621 [ 305.339250] Kernel BUG at ffffffffa0a7fa32 [verbose debug info unavailable] [ 305.383939] Call Trace: [ 305.385536] [] xfs_setattr_nonsize+0x69a/0x720 [xfs] [ 305.387142] [] xfs_vn_setattr+0x29/0x70 [xfs] [ 305.388727] [] notify_change+0x1a8/0x350 [ 305.390298] [] chown_common+0xfd/0x110 [ 305.391868] [] SyS_fchownat+0xaf/0x110 [ 305.393440] [] SyS_lchown+0x20/0x30 [ 305.394995] [] system_call_fastpath+0x1a/0x1f [ 305.399870] RIP [] assfail+0x22/0x30 [xfs] This fix adjust the assertion to check if the super block support both quota inodes or not. Signed-off-by: Jie Liu Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers (cherry picked from commit 5a01dd54f4a7fb513062070c5acef20d13cad980) diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index 27e0e54..104455b 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -618,7 +618,8 @@ xfs_setattr_nonsize( } if (!gid_eq(igid, gid)) { if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_GQUOTA_ON(mp)) { - ASSERT(!XFS_IS_PQUOTA_ON(mp)); + ASSERT(xfs_sb_version_has_pquotino(&mp->m_sb) || + !XFS_IS_PQUOTA_ON(mp)); ASSERT(mask & ATTR_GID); ASSERT(gdqp); olddquot2 = xfs_qm_vop_chown(tp, ip, -- cgit v0.10.2 From 718cc6f88cbfc4fbd39609f28c4c86883945f90d Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Tue, 26 Nov 2013 21:38:49 +0800 Subject: xfs: fix infinite loop by detaching the group/project hints from user dquot xfs_quota(8) will hang up if trying to turn group/project quota off before the user quota is off, this could be 100% reproduced by: # mount -ouquota,gquota /dev/sda7 /xfs # mkdir /xfs/test # xfs_quota -xc 'off -g' /xfs <-- hangs up # echo w > /proc/sysrq-trigger # dmesg SysRq : Show Blocked State task PC stack pid father xfs_quota D 0000000000000000 0 27574 2551 0x00000000 [snip] Call Trace: [] schedule+0xad/0xc0 [] schedule_timeout+0x35e/0x3c0 [] ? mark_held_locks+0x176/0x1c0 [] ? call_timer_fn+0x2c0/0x2c0 [] ? xfs_qm_shrink_count+0x30/0x30 [xfs] [] schedule_timeout_uninterruptible+0x26/0x30 [] xfs_qm_dquot_walk+0x235/0x260 [xfs] [] ? xfs_perag_get+0x1d8/0x2d0 [xfs] [] ? xfs_perag_get+0x5/0x2d0 [xfs] [] ? xfs_inode_ag_iterator+0xae/0xf0 [xfs] [] ? xfs_trans_free_dqinfo+0x50/0x50 [xfs] [] ? xfs_inode_ag_iterator+0xcf/0xf0 [xfs] [] xfs_qm_dqpurge_all+0x66/0xb0 [xfs] [] xfs_qm_scall_quotaoff+0x20a/0x5f0 [xfs] [] xfs_fs_set_xstate+0x136/0x180 [xfs] [] do_quotactl+0x53a/0x6b0 [] ? iput+0x5b/0x90 [] SyS_quotactl+0x167/0x1d0 [] ? trace_hardirqs_on_thunk+0x3a/0x3f [] system_call_fastpath+0x16/0x1b It's fine if we turn user quota off at first, then turn off other kind of quotas if they are enabled since the group/project dquot refcount is decreased to zero once the user quota if off. Otherwise, those dquots refcount is non-zero due to the user dquot might refer to them as hint(s). Hence, above operation cause an infinite loop at xfs_qm_dquot_walk() while trying to purge dquot cache. This problem has been around since Linux 3.4, it was introduced by: [ b84a3a9675 xfs: remove the per-filesystem list of dquots ] Originally we will release the group dquot pointers because the user dquots maybe carrying around as a hint via xfs_qm_detach_gdquots(). However, with above change, there is no such work to be done before purging group/project dquot cache. In order to solve this problem, this patch introduces a special routine xfs_qm_dqpurge_hints(), and it would release the group/project dquot pointers the user dquots maybe carrying around as a hint, and then it will proceed to purge the user dquot cache if requested. Cc: stable@vger.kernel.org Signed-off-by: Jie Liu Reviewed-by: Dave Chinner Signed-off-by: Ben Myers (cherry picked from commit df8052e7dae00bde6f21b40b6e3e1099770f3afc) diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 588e490..dd88f0e 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -134,8 +134,6 @@ xfs_qm_dqpurge( { struct xfs_mount *mp = dqp->q_mount; struct xfs_quotainfo *qi = mp->m_quotainfo; - struct xfs_dquot *gdqp = NULL; - struct xfs_dquot *pdqp = NULL; xfs_dqlock(dqp); if ((dqp->dq_flags & XFS_DQ_FREEING) || dqp->q_nrefs != 0) { @@ -143,21 +141,6 @@ xfs_qm_dqpurge( return EAGAIN; } - /* - * If this quota has a hint attached, prepare for releasing it now. - */ - gdqp = dqp->q_gdquot; - if (gdqp) { - xfs_dqlock(gdqp); - dqp->q_gdquot = NULL; - } - - pdqp = dqp->q_pdquot; - if (pdqp) { - xfs_dqlock(pdqp); - dqp->q_pdquot = NULL; - } - dqp->dq_flags |= XFS_DQ_FREEING; xfs_dqflock(dqp); @@ -206,11 +189,47 @@ xfs_qm_dqpurge( XFS_STATS_DEC(xs_qm_dquot_unused); xfs_qm_dqdestroy(dqp); + return 0; +} + +/* + * Release the group or project dquot pointers the user dquots maybe carrying + * around as a hint, and proceed to purge the user dquot cache if requested. +*/ +STATIC int +xfs_qm_dqpurge_hints( + struct xfs_dquot *dqp, + void *data) +{ + struct xfs_dquot *gdqp = NULL; + struct xfs_dquot *pdqp = NULL; + uint flags = *((uint *)data); + + xfs_dqlock(dqp); + if (dqp->dq_flags & XFS_DQ_FREEING) { + xfs_dqunlock(dqp); + return EAGAIN; + } + + /* If this quota has a hint attached, prepare for releasing it now */ + gdqp = dqp->q_gdquot; + if (gdqp) + dqp->q_gdquot = NULL; + + pdqp = dqp->q_pdquot; + if (pdqp) + dqp->q_pdquot = NULL; + + xfs_dqunlock(dqp); if (gdqp) - xfs_qm_dqput(gdqp); + xfs_qm_dqrele(gdqp); if (pdqp) - xfs_qm_dqput(pdqp); + xfs_qm_dqrele(pdqp); + + if (flags & XFS_QMOPT_UQUOTA) + return xfs_qm_dqpurge(dqp, NULL); + return 0; } @@ -222,8 +241,18 @@ xfs_qm_dqpurge_all( struct xfs_mount *mp, uint flags) { - if (flags & XFS_QMOPT_UQUOTA) - xfs_qm_dquot_walk(mp, XFS_DQ_USER, xfs_qm_dqpurge, NULL); + /* + * We have to release group/project dquot hint(s) from the user dquot + * at first if they are there, otherwise we would run into an infinite + * loop while walking through radix tree to purge other type of dquots + * since their refcount is not zero if the user dquot refers to them + * as hint. + * + * Call the special xfs_qm_dqpurge_hints() will end up go through the + * general xfs_qm_dqpurge() against user dquot cache if requested. + */ + xfs_qm_dquot_walk(mp, XFS_DQ_USER, xfs_qm_dqpurge_hints, &flags); + if (flags & XFS_QMOPT_GQUOTA) xfs_qm_dquot_walk(mp, XFS_DQ_GROUP, xfs_qm_dqpurge, NULL); if (flags & XFS_QMOPT_PQUOTA) -- cgit v0.10.2 From 809625ca544bdc9e02a5728c1939a76dc29decb1 Mon Sep 17 00:00:00 2001 From: Namjae Jeon Date: Sun, 8 Dec 2013 23:33:50 +0900 Subject: MAINTAINERS: fix incorrect mail address of XFS maintainer When I tried to send the patches to XFS Maintainers, I got returned mail included delivery fail message for Dave's mail. Maybe, Dave Chinner mail address is incorrect. I try to fix it correctly. Signed-off-by: Namjae Jeon Reviewed-by: Ben Myers Signed-off-by: Ben Myers (cherry picked from commit db10bddc7d4f412bcd8630fc479fa1eb009e325b) diff --git a/MAINTAINERS b/MAINTAINERS index f216db8..f5d0bdd 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -9530,7 +9530,7 @@ F: drivers/xen/*swiotlb* XFS FILESYSTEM P: Silicon Graphics Inc -M: Dave Chinner +M: Dave Chinner M: Ben Myers M: xfs@oss.sgi.com L: xfs@oss.sgi.com -- cgit v0.10.2 From 6e708bcf6583f663da9fe7bc5292cc62f0a8410d Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Fri, 22 Nov 2013 10:41:16 +1100 Subject: xfs: align initial file allocations correctly The function xfs_bmap_isaeof() is used to indicate that an allocation is occurring at or past the end of file, and as such should be aligned to the underlying storage geometry if possible. Commit 27a3f8f ("xfs: introduce xfs_bmap_last_extent") changed the behaviour of this function for empty files - it turned off allocation alignment for this case accidentally. Hence large initial allocations from direct IO are not getting correctly aligned to the underlying geometry, and that is cause write performance to drop in alignment sensitive configurations. Fix it by considering allocation into empty files as requiring aligned allocation again. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers (cherry picked from commit f9b395a8ef8f34d19cae2cde361e19c96e097fad) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 3ef11b2..8401f11 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -1635,7 +1635,7 @@ xfs_bmap_last_extent( * blocks at the end of the file which do not start at the previous data block, * we will try to align the new blocks at stripe unit boundaries. * - * Returns 0 in bma->aeof if the file (fork) is empty as any new write will be + * Returns 1 in bma->aeof if the file (fork) is empty as any new write will be * at, or past the EOF. */ STATIC int @@ -1650,9 +1650,14 @@ xfs_bmap_isaeof( bma->aeof = 0; error = xfs_bmap_last_extent(NULL, bma->ip, whichfork, &rec, &is_empty); - if (error || is_empty) + if (error) return error; + if (is_empty) { + bma->aeof = 1; + return 0; + } + /* * Check if we are allocation or past the last extent, or at least into * the last delayed allocated extent. -- cgit v0.10.2 From 83a0adc3f93aae4ab9c59113e3145c7bdb2b4a8c Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 17 Dec 2013 00:03:52 -0800 Subject: xfs: remove xfsbdstrat error The xfsbdstrat helper is a small but useless wrapper for xfs_buf_iorequest that handles the case of a shut down filesystem. Most of the users have private, uncached buffers that can just be freed in this case, but the complex error handling in xfs_bioerror_relse messes up the case when it's called without a locked buffer. Remove xfsbdstrat and opencode the error handling in the callers. All but one can simply return an error and don't need to deal with buffer state, and the one caller that cares about the buffer state could do with a major cleanup as well, but we'll defer that to later. Signed-off-by: Christoph Hellwig Reviewed-by: Ben Myers Signed-off-by: Ben Myers diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index 5887e41..1394106 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -1187,7 +1187,12 @@ xfs_zero_remaining_bytes( XFS_BUF_UNWRITE(bp); XFS_BUF_READ(bp); XFS_BUF_SET_ADDR(bp, xfs_fsb_to_db(ip, imap.br_startblock)); - xfsbdstrat(mp, bp); + + if (XFS_FORCED_SHUTDOWN(mp)) { + error = XFS_ERROR(EIO); + break; + } + xfs_buf_iorequest(bp); error = xfs_buf_iowait(bp); if (error) { xfs_buf_ioerror_alert(bp, @@ -1200,7 +1205,12 @@ xfs_zero_remaining_bytes( XFS_BUF_UNDONE(bp); XFS_BUF_UNREAD(bp); XFS_BUF_WRITE(bp); - xfsbdstrat(mp, bp); + + if (XFS_FORCED_SHUTDOWN(mp)) { + error = XFS_ERROR(EIO); + break; + } + xfs_buf_iorequest(bp); error = xfs_buf_iowait(bp); if (error) { xfs_buf_ioerror_alert(bp, diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index c7f0b77..9fa9c43 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -698,7 +698,11 @@ xfs_buf_read_uncached( bp->b_flags |= XBF_READ; bp->b_ops = ops; - xfsbdstrat(target->bt_mount, bp); + if (XFS_FORCED_SHUTDOWN(target->bt_mount)) { + xfs_buf_relse(bp); + return NULL; + } + xfs_buf_iorequest(bp); xfs_buf_iowait(bp); return bp; } @@ -1089,7 +1093,7 @@ xfs_bioerror( * This is meant for userdata errors; metadata bufs come with * iodone functions attached, so that we can track down errors. */ -STATIC int +int xfs_bioerror_relse( struct xfs_buf *bp) { @@ -1164,25 +1168,6 @@ xfs_bwrite( return error; } -/* - * Wrapper around bdstrat so that we can stop data from going to disk in case - * we are shutting down the filesystem. Typically user data goes thru this - * path; one of the exceptions is the superblock. - */ -void -xfsbdstrat( - struct xfs_mount *mp, - struct xfs_buf *bp) -{ - if (XFS_FORCED_SHUTDOWN(mp)) { - trace_xfs_bdstrat_shut(bp, _RET_IP_); - xfs_bioerror_relse(bp); - return; - } - - xfs_buf_iorequest(bp); -} - STATIC void _xfs_buf_ioend( xfs_buf_t *bp, diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index e656833..7e41b08 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -269,9 +269,6 @@ extern void xfs_buf_unlock(xfs_buf_t *); /* Buffer Read and Write Routines */ extern int xfs_bwrite(struct xfs_buf *bp); - -extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *); - extern void xfs_buf_ioend(xfs_buf_t *, int); extern void xfs_buf_ioerror(xfs_buf_t *, int); extern void xfs_buf_ioerror_alert(struct xfs_buf *, const char *func); @@ -282,6 +279,8 @@ extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *, #define xfs_buf_zero(bp, off, len) \ xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO) +extern int xfs_bioerror_relse(struct xfs_buf *); + static inline int xfs_buf_geterror(xfs_buf_t *bp) { return bp ? bp->b_error : ENOMEM; diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index b6b669d..eae1692 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -193,7 +193,10 @@ xlog_bread_noalign( bp->b_io_length = nbblks; bp->b_error = 0; - xfsbdstrat(log->l_mp, bp); + if (XFS_FORCED_SHUTDOWN(log->l_mp)) + return XFS_ERROR(EIO); + + xfs_buf_iorequest(bp); error = xfs_buf_iowait(bp); if (error) xfs_buf_ioerror_alert(bp, __func__); @@ -4397,7 +4400,13 @@ xlog_do_recover( XFS_BUF_READ(bp); XFS_BUF_UNASYNC(bp); bp->b_ops = &xfs_sb_buf_ops; - xfsbdstrat(log->l_mp, bp); + + if (XFS_FORCED_SHUTDOWN(log->l_mp)) { + xfs_buf_relse(bp); + return XFS_ERROR(EIO); + } + + xfs_buf_iorequest(bp); error = xfs_buf_iowait(bp); if (error) { xfs_buf_ioerror_alert(bp, __func__); diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index c035d11..647b6f1 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -314,7 +314,18 @@ xfs_trans_read_buf_map( ASSERT(bp->b_iodone == NULL); XFS_BUF_READ(bp); bp->b_ops = ops; - xfsbdstrat(tp->t_mountp, bp); + + /* + * XXX(hch): clean up the error handling here to be less + * of a mess.. + */ + if (XFS_FORCED_SHUTDOWN(mp)) { + trace_xfs_bdstrat_shut(bp, _RET_IP_); + xfs_bioerror_relse(bp); + } else { + xfs_buf_iorequest(bp); + } + error = xfs_buf_iowait(bp); if (error) { xfs_buf_ioerror_alert(bp, __func__); -- cgit v0.10.2 From 33177f05364c6cd13b06d0f3500dad07cf4647c2 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 12 Dec 2013 16:34:36 +1100 Subject: xfs: swalloc doesn't align allocations properly When swalloc is specified as a mount option, allocations are supposed to be aligned to the stripe width rather than the stripe unit of the underlying filesystem. However, it does not do this. What the implementation does is round up the allocation size to a stripe width, hence ensuring that all allocations span a full stripe width. It does not, however, ensure that that allocation is aligned to a stripe width, and hence the allocations can span multiple underlying stripes and so still see RMW cycles for things like direct IO on MD RAID. So, if the swalloc mount option is set, change the allocation alignment in xfs_bmap_btalloc() to use the stripe width rather than the stripe unit. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Reviewed-by: Ben Myers Signed-off-by: Ben Myers diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 8401f11..3b2c14b 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -3648,10 +3648,19 @@ xfs_bmap_btalloc( int isaligned; int tryagain; int error; + int stripe_align; ASSERT(ap->length); mp = ap->ip->i_mount; + + /* stripe alignment for allocation is determined by mount parameters */ + stripe_align = 0; + if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC)) + stripe_align = mp->m_swidth; + else if (mp->m_dalign) + stripe_align = mp->m_dalign; + align = ap->userdata ? xfs_get_extsz_hint(ap->ip) : 0; if (unlikely(align)) { error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev, @@ -3660,6 +3669,8 @@ xfs_bmap_btalloc( ASSERT(!error); ASSERT(ap->length); } + + nullfb = *ap->firstblock == NULLFSBLOCK; fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock); if (nullfb) { @@ -3735,7 +3746,7 @@ xfs_bmap_btalloc( */ if (!ap->flist->xbf_low && ap->aeof) { if (!ap->offset) { - args.alignment = mp->m_dalign; + args.alignment = stripe_align; atype = args.type; isaligned = 1; /* @@ -3760,13 +3771,13 @@ xfs_bmap_btalloc( * of minlen+alignment+slop doesn't go up * between the calls. */ - if (blen > mp->m_dalign && blen <= args.maxlen) - nextminlen = blen - mp->m_dalign; + if (blen > stripe_align && blen <= args.maxlen) + nextminlen = blen - stripe_align; else nextminlen = args.minlen; - if (nextminlen + mp->m_dalign > args.minlen + 1) + if (nextminlen + stripe_align > args.minlen + 1) args.minalignslop = - nextminlen + mp->m_dalign - + nextminlen + stripe_align - args.minlen - 1; else args.minalignslop = 0; @@ -3788,7 +3799,7 @@ xfs_bmap_btalloc( */ args.type = atype; args.fsbno = ap->blkno; - args.alignment = mp->m_dalign; + args.alignment = stripe_align; args.minlen = nextminlen; args.minalignslop = 0; isaligned = 1; -- cgit v0.10.2 From ac8809f9ab01a73de1a47b5a37bd8dcca8712fb3 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 12 Dec 2013 16:34:38 +1100 Subject: xfs: abort metadata writeback on permanent errors If we are doing aysnc writeback of metadata, we can get write errors but have nobody to report them to. At the moment, we simply attempt to reissue the write from io completion in the hope that it's a transient error. When it's not a transient error, the buffer is stuck forever in this loop, and we cannot break out of it. Eventually, unmount will hang because the AIL cannot be emptied and everything goes downhill from them. To solve this problem, only retry the write IO once before aborting it. We don't throw the buffer away because some transient errors can last minutes (e.g. FC path failover) or even hours (thin provisioned devices that have run out of backing space) before they go away. Hence we really want to keep trying until we can't try any more. Because the buffer was not cleaned, however, it does not get removed from the AIL and hence the next pass across the AIL will start IO on it again. As such, we still get the "retry forever" semantics that we currently have, but we allow other access to the buffer in the mean time. Meanwhile the filesystem can continue to modify the buffer and relog it, so the IO errors won't hang the log or the filesystem. Now when we are pushing the AIL, we can see all these "permanent IO error" buffers and we can issue a warning about failures before we retry the IO. We can also catch these buffers when unmounting an issue a corruption warning, too. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 9fa9c43..afe7645 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1156,7 +1156,7 @@ xfs_bwrite( ASSERT(xfs_buf_islocked(bp)); bp->b_flags |= XBF_WRITE; - bp->b_flags &= ~(XBF_ASYNC | XBF_READ | _XBF_DELWRI_Q); + bp->b_flags &= ~(XBF_ASYNC | XBF_READ | _XBF_DELWRI_Q | XBF_WRITE_FAIL); xfs_bdstrat_cb(bp); @@ -1501,6 +1501,12 @@ xfs_wait_buftarg( struct xfs_buf *bp; bp = list_first_entry(&dispose, struct xfs_buf, b_lru); list_del_init(&bp->b_lru); + if (bp->b_flags & XBF_WRITE_FAIL) { + xfs_alert(btp->bt_mount, +"Corruption Alert: Buffer at block 0x%llx had permanent write failures!\n" +"Please run xfs_repair to determine the extent of the problem.", + (long long)bp->b_bn); + } xfs_buf_rele(bp); } if (loop++ != 0) @@ -1784,7 +1790,7 @@ __xfs_buf_delwri_submit( blk_start_plug(&plug); list_for_each_entry_safe(bp, n, io_list, b_list) { - bp->b_flags &= ~(_XBF_DELWRI_Q | XBF_ASYNC); + bp->b_flags &= ~(_XBF_DELWRI_Q | XBF_ASYNC | XBF_WRITE_FAIL); bp->b_flags |= XBF_WRITE; if (!wait) { diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 7e41b08..1cf21a4 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -45,6 +45,7 @@ typedef enum { #define XBF_ASYNC (1 << 4) /* initiator will not wait for completion */ #define XBF_DONE (1 << 5) /* all pages in the buffer uptodate */ #define XBF_STALE (1 << 6) /* buffer has been staled, do not find it */ +#define XBF_WRITE_FAIL (1 << 24)/* async writes have failed on this buffer */ /* I/O hints for the BIO layer */ #define XBF_SYNCIO (1 << 10)/* treat this buffer as synchronous I/O */ @@ -70,6 +71,7 @@ typedef unsigned int xfs_buf_flags_t; { XBF_ASYNC, "ASYNC" }, \ { XBF_DONE, "DONE" }, \ { XBF_STALE, "STALE" }, \ + { XBF_WRITE_FAIL, "WRITE_FAIL" }, \ { XBF_SYNCIO, "SYNCIO" }, \ { XBF_FUA, "FUA" }, \ { XBF_FLUSH, "FLUSH" }, \ @@ -80,6 +82,7 @@ typedef unsigned int xfs_buf_flags_t; { _XBF_DELWRI_Q, "DELWRI_Q" }, \ { _XBF_COMPOUND, "COMPOUND" } + /* * Internal state flags. */ @@ -300,7 +303,8 @@ extern void xfs_buf_terminate(void); #define XFS_BUF_ZEROFLAGS(bp) \ ((bp)->b_flags &= ~(XBF_READ|XBF_WRITE|XBF_ASYNC| \ - XBF_SYNCIO|XBF_FUA|XBF_FLUSH)) + XBF_SYNCIO|XBF_FUA|XBF_FLUSH| \ + XBF_WRITE_FAIL)) void xfs_buf_stale(struct xfs_buf *bp); #define XFS_BUF_UNSTALE(bp) ((bp)->b_flags &= ~XBF_STALE) diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index a64f67b..2227b9b 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -496,6 +496,14 @@ xfs_buf_item_unpin( } } +/* + * Buffer IO error rate limiting. Limit it to no more than 10 messages per 30 + * seconds so as to not spam logs too much on repeated detection of the same + * buffer being bad.. + */ + +DEFINE_RATELIMIT_STATE(xfs_buf_write_fail_rl_state, 30 * HZ, 10); + STATIC uint xfs_buf_item_push( struct xfs_log_item *lip, @@ -524,6 +532,14 @@ xfs_buf_item_push( trace_xfs_buf_item_push(bip); + /* has a previous flush failed due to IO errors? */ + if ((bp->b_flags & XBF_WRITE_FAIL) && + ___ratelimit(&xfs_buf_write_fail_rl_state, "XFS:")) { + xfs_warn(bp->b_target->bt_mount, +"Detected failing async write on buffer block 0x%llx. Retrying async write.\n", + (long long)bp->b_bn); + } + if (!xfs_buf_delwri_queue(bp, buffer_list)) rval = XFS_ITEM_FLUSHING; xfs_buf_unlock(bp); @@ -1096,8 +1112,9 @@ xfs_buf_iodone_callbacks( xfs_buf_ioerror(bp, 0); /* errno of 0 unsets the flag */ - if (!XFS_BUF_ISSTALE(bp)) { - bp->b_flags |= XBF_WRITE | XBF_ASYNC | XBF_DONE; + if (!(bp->b_flags & (XBF_STALE|XBF_WRITE_FAIL))) { + bp->b_flags |= XBF_WRITE | XBF_ASYNC | + XBF_DONE | XBF_WRITE_FAIL; xfs_buf_iorequest(bp); } else { xfs_buf_relse(bp); -- cgit v0.10.2 From ed697e1aaf7237b1a62af39f64463b05c262808d Mon Sep 17 00:00:00 2001 From: JongHo Kim Date: Tue, 17 Dec 2013 23:02:24 +0900 Subject: ALSA: Add SNDRV_PCM_STATE_PAUSED case in wait_for_avail function When the process is sleeping at the SNDRV_PCM_STATE_PAUSED state from the wait_for_avail function, the sleep process will be woken by timeout(10 seconds). Even if the sleep process wake up by timeout, by this patch, the process will continue with sleep and wait for the other state. Signed-off-by: JongHo Kim Cc: Signed-off-by: Takashi Iwai diff --git a/sound/core/pcm_lib.c b/sound/core/pcm_lib.c index 6e03b46..a210467 100644 --- a/sound/core/pcm_lib.c +++ b/sound/core/pcm_lib.c @@ -1937,6 +1937,8 @@ static int wait_for_avail(struct snd_pcm_substream *substream, case SNDRV_PCM_STATE_DISCONNECTED: err = -EBADFD; goto _endloop; + case SNDRV_PCM_STATE_PAUSED: + continue; } if (!tout) { snd_printd("%s write error (DMA or IRQ trouble?)\n", -- cgit v0.10.2 From d24c195f90cb1adb178d26d84c722d4b9e551e05 Mon Sep 17 00:00:00 2001 From: Mika Westerberg Date: Tue, 10 Dec 2013 12:56:59 +0200 Subject: serial: 8250_dw: add new ACPI IDs Newer Intel PCHs with LPSS have the same Designware controllers than Haswell but ACPI IDs are different. Add these IDs to the driver list. Signed-off-by: Mika Westerberg Cc: stable Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/tty/serial/8250/8250_dw.c b/drivers/tty/serial/8250/8250_dw.c index 5f096c7..06525f1 100644 --- a/drivers/tty/serial/8250/8250_dw.c +++ b/drivers/tty/serial/8250/8250_dw.c @@ -457,6 +457,8 @@ MODULE_DEVICE_TABLE(of, dw8250_of_match); static const struct acpi_device_id dw8250_acpi_match[] = { { "INT33C4", 0 }, { "INT33C5", 0 }, + { "INT3434", 0 }, + { "INT3435", 0 }, { "80860F0A", 0 }, { }, }; -- cgit v0.10.2 From 1075a6e2dc7e2a96efc417b98dd98f57fdae985d Mon Sep 17 00:00:00 2001 From: Peter Hurley Date: Mon, 9 Dec 2013 18:06:07 -0500 Subject: n_tty: Fix apparent order of echoed output With block processing of echoed output, observed output order is still required. Push completed echoes and echo commands prior to output. Introduce echo_mark echo buffer index, which tracks completed echo commands; ie., those submitted via commit_echoes but which may not have been committed. Ensure that completed echoes are output prior to subsequent terminal writes in process_echoes(). Fixes newline/prompt output order in cooked mode shell. Cc: # 3.12.x : 39434ab n_tty: Fix missing newline echo Reported-by: Karl Dahlke Reported-by: Mikulas Patocka Signed-off-by: Peter Hurley Tested-by: Karl Dahlke Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c index 268b627..34aacaa 100644 --- a/drivers/tty/n_tty.c +++ b/drivers/tty/n_tty.c @@ -93,6 +93,7 @@ struct n_tty_data { size_t canon_head; size_t echo_head; size_t echo_commit; + size_t echo_mark; DECLARE_BITMAP(char_map, 256); /* private to n_tty_receive_overrun (single-threaded) */ @@ -336,6 +337,7 @@ static void reset_buffer_flags(struct n_tty_data *ldata) { ldata->read_head = ldata->canon_head = ldata->read_tail = 0; ldata->echo_head = ldata->echo_tail = ldata->echo_commit = 0; + ldata->echo_mark = 0; ldata->line_start = 0; ldata->erasing = 0; @@ -787,6 +789,7 @@ static void commit_echoes(struct tty_struct *tty) size_t head; head = ldata->echo_head; + ldata->echo_mark = head; old = ldata->echo_commit - ldata->echo_tail; /* Process committed echoes if the accumulated # of bytes @@ -811,10 +814,11 @@ static void process_echoes(struct tty_struct *tty) size_t echoed; if ((!L_ECHO(tty) && !L_ECHONL(tty)) || - ldata->echo_commit == ldata->echo_tail) + ldata->echo_mark == ldata->echo_tail) return; mutex_lock(&ldata->output_lock); + ldata->echo_commit = ldata->echo_mark; echoed = __process_echoes(tty); mutex_unlock(&ldata->output_lock); @@ -822,6 +826,7 @@ static void process_echoes(struct tty_struct *tty) tty->ops->flush_chars(tty); } +/* NB: echo_mark and echo_head should be equivalent here */ static void flush_echoes(struct tty_struct *tty) { struct n_tty_data *ldata = tty->disc_data; -- cgit v0.10.2 From f447ef4a56dee4b68a91460bcdfe06b5011085f2 Mon Sep 17 00:00:00 2001 From: Josh Boyer Date: Fri, 11 Oct 2013 08:45:51 -0400 Subject: cpupower: Fix segfault due to incorrect getopt_long arugments If a user calls 'cpupower set --perf-bias 15', the process will end with a SIGSEGV in libc because cpupower-set passes a NULL optarg to the atoi call. This is because the getopt_long structure currently has all of the options as having an optional_argument when they really have a required argument. We change the structure to use required_argument to match the short options and it resolves the issue. This fixes https://bugzilla.redhat.com/show_bug.cgi?id=1000439 Signed-off-by: Josh Boyer Cc: Dominik Brodowski Cc: Thomas Renninger Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds diff --git a/tools/power/cpupower/utils/cpupower-set.c b/tools/power/cpupower/utils/cpupower-set.c index dc4de37..bcf1d2f 100644 --- a/tools/power/cpupower/utils/cpupower-set.c +++ b/tools/power/cpupower/utils/cpupower-set.c @@ -18,9 +18,9 @@ #include "helpers/bitmask.h" static struct option set_opts[] = { - { .name = "perf-bias", .has_arg = optional_argument, .flag = NULL, .val = 'b'}, - { .name = "sched-mc", .has_arg = optional_argument, .flag = NULL, .val = 'm'}, - { .name = "sched-smt", .has_arg = optional_argument, .flag = NULL, .val = 's'}, + { .name = "perf-bias", .has_arg = required_argument, .flag = NULL, .val = 'b'}, + { .name = "sched-mc", .has_arg = required_argument, .flag = NULL, .val = 'm'}, + { .name = "sched-smt", .has_arg = required_argument, .flag = NULL, .val = 's'}, { }, }; -- cgit v0.10.2 From 50dc875f2e6e2e04aed3b3033eb0ac99192d6d02 Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Fri, 13 Dec 2013 17:21:27 +0800 Subject: netvsc: don't flush peers notifying work during setting mtu There's a possible deadlock if we flush the peers notifying work during setting mtu: [ 22.991149] ====================================================== [ 22.991173] [ INFO: possible circular locking dependency detected ] [ 22.991198] 3.10.0-54.0.1.el7.x86_64.debug #1 Not tainted [ 22.991219] ------------------------------------------------------- [ 22.991243] ip/974 is trying to acquire lock: [ 22.991261] ((&(&net_device_ctx->dwork)->work)){+.+.+.}, at: [] flush_work+0x5/0x2e0 [ 22.991307] but task is already holding lock: [ 22.991330] (rtnl_mutex){+.+.+.}, at: [] rtnetlink_rcv+0x1b/0x40 [ 22.991367] which lock already depends on the new lock. [ 22.991398] the existing dependency chain (in reverse order) is: [ 22.991426] -> #1 (rtnl_mutex){+.+.+.}: [ 22.991449] [] __lock_acquire+0xb19/0x1260 [ 22.991477] [] lock_acquire+0xa2/0x1f0 [ 22.991501] [] mutex_lock_nested+0x89/0x4f0 [ 22.991529] [] rtnl_lock+0x17/0x20 [ 22.991552] [] netdev_notify_peers+0x12/0x30 [ 22.991579] [] netvsc_send_garp+0x22/0x30 [hv_netvsc] [ 22.991610] [] process_one_work+0x211/0x6e0 [ 22.991637] [] worker_thread+0x11b/0x3a0 [ 22.991663] [] kthread+0xed/0x100 [ 22.991686] [] ret_from_fork+0x7c/0xb0 [ 22.991715] -> #0 ((&(&net_device_ctx->dwork)->work)){+.+.+.}: [ 22.991715] [] check_prevs_add+0x967/0x970 [ 22.991715] [] __lock_acquire+0xb19/0x1260 [ 22.991715] [] lock_acquire+0xa2/0x1f0 [ 22.991715] [] flush_work+0x4e/0x2e0 [ 22.991715] [] __cancel_work_timer+0x95/0x130 [ 22.991715] [] cancel_delayed_work_sync+0x13/0x20 [ 22.991715] [] netvsc_change_mtu+0x84/0x200 [hv_netvsc] [ 22.991715] [] dev_set_mtu+0x34/0x80 [ 22.991715] [] do_setlink+0x23a/0xa00 [ 22.991715] [] rtnl_newlink+0x394/0x5e0 [ 22.991715] [] rtnetlink_rcv_msg+0x9c/0x260 [ 22.991715] [] netlink_rcv_skb+0xa9/0xc0 [ 22.991715] [] rtnetlink_rcv+0x2a/0x40 [ 22.991715] [] netlink_unicast+0xdd/0x190 [ 22.991715] [] netlink_sendmsg+0x337/0x750 [ 22.991715] [] sock_sendmsg+0x99/0xd0 [ 22.991715] [] ___sys_sendmsg+0x39e/0x3b0 [ 22.991715] [] __sys_sendmsg+0x42/0x80 [ 22.991715] [] SyS_sendmsg+0x12/0x20 [ 22.991715] [] system_call_fastpath+0x16/0x1b This is because we hold the rtnl_lock() before ndo_change_mtu() and try to flush the work in netvsc_change_mtu(), in the mean time, netdev_notify_peers() may be called from worker and also trying to hold the rtnl_lock. This will lead the flush won't succeed forever. Solve this by not canceling and flushing the work, this is safe because the transmission done by NETDEV_NOTIFY_PEERS was synchronized with the netif_tx_disable() called by netvsc_change_mtu(). Reported-by: Yaju Cao Tested-by: Yaju Cao Cc: K. Y. Srinivasan Cc: Haiyang Zhang Signed-off-by: Jason Wang Acked-by: Haiyang Zhang Signed-off-by: David S. Miller diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c index 524f713..f813572 100644 --- a/drivers/net/hyperv/netvsc_drv.c +++ b/drivers/net/hyperv/netvsc_drv.c @@ -327,7 +327,6 @@ static int netvsc_change_mtu(struct net_device *ndev, int mtu) return -EINVAL; nvdev->start_remove = true; - cancel_delayed_work_sync(&ndevctx->dwork); cancel_work_sync(&ndevctx->work); netif_tx_disable(ndev); rndis_filter_device_remove(hdev); -- cgit v0.10.2 From e47eb5dfb296bf217e9ebee7b2f07486670b9c1b Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Sun, 15 Dec 2013 10:53:46 -0800 Subject: udp: ipv4: do not use sk_dst_lock from softirq context Using sk_dst_lock from softirq context is not supported right now. Instead of adding BH protection everywhere, udp_sk_rx_dst_set() can instead use xchg(), as suggested by David. Reported-by: Fengguang Wu Fixes: 975022310233 ("udp: ipv4: must add synchronization in udp_sk_rx_dst_set()") Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 62c19fd..f140048 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1600,20 +1600,15 @@ static void flush_stack(struct sock **stack, unsigned int count, } /* For TCP sockets, sk_rx_dst is protected by socket lock - * For UDP, we use sk_dst_lock to guard against concurrent changes. + * For UDP, we use xchg() to guard against concurrent changes. */ static void udp_sk_rx_dst_set(struct sock *sk, struct dst_entry *dst) { struct dst_entry *old; - spin_lock(&sk->sk_dst_lock); - old = sk->sk_rx_dst; - if (likely(old != dst)) { - dst_hold(dst); - sk->sk_rx_dst = dst; - dst_release(old); - } - spin_unlock(&sk->sk_dst_lock); + dst_hold(dst); + old = xchg(&sk->sk_rx_dst, dst); + dst_release(old); } /* -- cgit v0.10.2 From d1fc50247693bf9116184c19ca16386ff14f5aed Mon Sep 17 00:00:00 2001 From: Steffen Klassert Date: Fri, 13 Dec 2013 12:56:05 +0100 Subject: MAINTAINERS: Update the IPsec maintainer entry Add the IPsec git trees and some pure IPsec modules to the IPsec section in the MAINTAINERS file. Signed-off-by: Steffen Klassert Signed-off-by: David S. Miller diff --git a/MAINTAINERS b/MAINTAINERS index 1344816..b8af16d 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -5911,12 +5911,21 @@ M: Steffen Klassert M: Herbert Xu M: "David S. Miller" L: netdev@vger.kernel.org -T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git +T: git git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec.git +T: git git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next.git S: Maintained F: net/xfrm/ F: net/key/ F: net/ipv4/xfrm* +F: net/ipv4/esp4.c +F: net/ipv4/ah4.c +F: net/ipv4/ipcomp.c +F: net/ipv4/ip_vti.c F: net/ipv6/xfrm* +F: net/ipv6/esp6.c +F: net/ipv6/ah6.c +F: net/ipv6/ipcomp6.c +F: net/ipv6/ip6_vti.c F: include/uapi/linux/xfrm.h F: include/net/xfrm.h -- cgit v0.10.2 From 37ab4fa7844a044dc21fde45e2a0fc2f3c3b6490 Mon Sep 17 00:00:00 2001 From: Sasha Levin Date: Fri, 13 Dec 2013 10:54:22 -0500 Subject: net: unix: allow bind to fail on mutex lock This is similar to the set_peek_off patch where calling bind while the socket is stuck in unix_dgram_recvmsg() will block and cause a hung task spew after a while. This is also the last place that did a straightforward mutex_lock(), so there shouldn't be any more of these patches. Signed-off-by: Sasha Levin Signed-off-by: David S. Miller diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index a0ca162..a427623 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -718,7 +718,9 @@ static int unix_autobind(struct socket *sock) int err; unsigned int retries = 0; - mutex_lock(&u->readlock); + err = mutex_lock_interruptible(&u->readlock); + if (err) + return err; err = 0; if (u->addr) @@ -877,7 +879,9 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len) goto out; addr_len = err; - mutex_lock(&u->readlock); + err = mutex_lock_interruptible(&u->readlock); + if (err) + goto out; err = -EINVAL; if (u->addr) -- cgit v0.10.2 From 188b04d580ab7acf11eb77cb564692050c10edfe Mon Sep 17 00:00:00 2001 From: Hannes Frederic Sowa Date: Sat, 14 Dec 2013 04:42:13 +0100 Subject: ipv4: improve documentation of ip_no_pmtu_disc Signed-off-by: Hannes Frederic Sowa Signed-off-by: David S. Miller diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt index 3c12d9a..8a984e9 100644 --- a/Documentation/networking/ip-sysctl.txt +++ b/Documentation/networking/ip-sysctl.txt @@ -16,8 +16,12 @@ ip_default_ttl - INTEGER Default: 64 (as recommended by RFC1700) ip_no_pmtu_disc - BOOLEAN - Disable Path MTU Discovery. - default FALSE + Disable Path MTU Discovery. If enabled and a + fragmentation-required ICMP is received, the PMTU to this + destination will be set to min_pmtu (see below). You will need + to raise min_pmtu to the smallest interface MTU on your system + manually if you want to avoid locally generated fragments. + Default: FALSE min_pmtu - INTEGER default 552 - minimum discovered Path MTU -- cgit v0.10.2 From c5fe7a41ad51e0c2df8381e9c85d7343b36b2c1e Mon Sep 17 00:00:00 2001 From: Jonathan Cameron Date: Wed, 11 Dec 2013 18:45:00 +0000 Subject: staging:iio:mag:hmc5843 fix incorrect endianness of channel as a result of missuse of the IIO_ST macro. This driver sets the shift value equal to IIO_BE (or 1) rather than setting that to 0 and specificying the endianness. This means the channel type is missreported as [be|le]:u16/16>>1 where the be|le is dependent on the cpu native endianness, rather than be:u16/16>>0 resulting in any userspace code using this information, miss converting the channel and generating thoroughly trashed data. Signed-off-by: Jonathan Cameron Acked-by: Lars-Peter Clausen Cc: stable@vger.kernel.org diff --git a/drivers/staging/iio/magnetometer/hmc5843.c b/drivers/staging/iio/magnetometer/hmc5843.c index 99421f9..0485d7f 100644 --- a/drivers/staging/iio/magnetometer/hmc5843.c +++ b/drivers/staging/iio/magnetometer/hmc5843.c @@ -451,7 +451,12 @@ done: .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE) | \ BIT(IIO_CHAN_INFO_SAMP_FREQ), \ .scan_index = idx, \ - .scan_type = IIO_ST('s', 16, 16, IIO_BE), \ + .scan_type = { \ + .sign = 's', \ + .realbits = 16, \ + .storagebits = 16, \ + .endianness = IIO_BE, \ + }, \ } static const struct iio_chan_spec hmc5843_channels[] = { -- cgit v0.10.2 From 3425c0f7ac61f2fcfb7f2728e9b7ba7e27aec429 Mon Sep 17 00:00:00 2001 From: Jonathan Cameron Date: Wed, 11 Dec 2013 18:45:00 +0000 Subject: iio:imu:adis16400 fix pressure channel scan type A single channel in this driver was using the IIO_ST macro. This does not provide a parameter for setting the endianness of the channel. Thus this channel will have been reported as whatever is the native endianness of the cpu rather than big endian. This means it would be incorrect on little endian platforms. Signed-off-by: Jonathan Cameron Acked-by: Lars-Peter Clausen Cc: stable@vger.kernel.org diff --git a/drivers/iio/imu/adis16400_core.c b/drivers/iio/imu/adis16400_core.c index 3fb7757..368660d 100644 --- a/drivers/iio/imu/adis16400_core.c +++ b/drivers/iio/imu/adis16400_core.c @@ -651,7 +651,12 @@ static const struct iio_chan_spec adis16448_channels[] = { .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE), .address = ADIS16448_BARO_OUT, .scan_index = ADIS16400_SCAN_BARO, - .scan_type = IIO_ST('s', 16, 16, 0), + .scan_type = { + .sign = 's', + .realbits = 16, + .storagebits = 16, + .endianness = IIO_BE, + }, }, ADIS16400_TEMP_CHAN(ADIS16448_TEMP_OUT, 12), IIO_CHAN_SOFT_TIMESTAMP(11) -- cgit v0.10.2 From e39d99059ad7f75d7ae2d3c59055d3c476cdb0d9 Mon Sep 17 00:00:00 2001 From: Jonathan Cameron Date: Wed, 11 Dec 2013 18:45:00 +0000 Subject: iio:adc:ad7887 Fix channel reported endianness from cpu to big endian Note this also sets the endianness to big endian whereas it would previously have defaulted to the cpu endian. Hence technically this is a bug fix on LE platforms. Signed-off-by: Jonathan Cameron Acked-by: Lars-Peter Clausen Cc: stable@vger.kernel.org diff --git a/drivers/iio/adc/ad7887.c b/drivers/iio/adc/ad7887.c index acb7f90..749a6ca 100644 --- a/drivers/iio/adc/ad7887.c +++ b/drivers/iio/adc/ad7887.c @@ -200,7 +200,13 @@ static const struct ad7887_chip_info ad7887_chip_info_tbl[] = { .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE), .address = 1, .scan_index = 1, - .scan_type = IIO_ST('u', 12, 16, 0), + .scan_type = { + .sign = 'u', + .realbits = 12, + .storagebits = 16, + .shift = 0, + .endianness = IIO_BE, + }, }, .channel[1] = { .type = IIO_VOLTAGE, @@ -210,7 +216,13 @@ static const struct ad7887_chip_info ad7887_chip_info_tbl[] = { .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE), .address = 0, .scan_index = 0, - .scan_type = IIO_ST('u', 12, 16, 0), + .scan_type = { + .sign = 'u', + .realbits = 12, + .storagebits = 16, + .shift = 0, + .endianness = IIO_BE, + }, }, .channel[2] = IIO_CHAN_SOFT_TIMESTAMP(2), .int_vref_mv = 2500, -- cgit v0.10.2 From 0283f7a100882684ad32b768f9f1ad81658a0b92 Mon Sep 17 00:00:00 2001 From: Ian Abbott Date: Fri, 13 Dec 2013 12:00:30 +0000 Subject: staging: comedi: 8255_pci: fix for newer PCI-DIO48H At some point, Measurement Computing / ComputerBoards redesigned the PCI-DIO48H to use a PLX PCI interface chip instead of an AMCC chip. This meant they had to put their hardware registers in the PCI BAR 2 region instead of PCI BAR 1. Unfortunately, they kept the same PCI device ID for the new design. This means the driver recognizes the newer cards, but doesn't work (and is likely to screw up the local configuration registers of the PLX chip) because it's using the wrong region. Since the PCI subvendor and subdevice IDs were both zero on the old design, but are the same as the vendor and device on the new design, we can tell the old design and new design apart easily enough. Split the existing entry for the PCI-DIO48H in `pci_8255_boards[]` into two new entries, referenced by different entries in the PCI device ID table `pci_8255_pci_table[]`. Use the same board name for both entries. Signed-off-by: Ian Abbott Cc: stablle # 3.10.y # 3.11.y # 3.12.y # 3.13.y Acked-by: H Hartley Sweeten Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/comedi/drivers/8255_pci.c b/drivers/staging/comedi/drivers/8255_pci.c index 432e3f9..c55f234 100644 --- a/drivers/staging/comedi/drivers/8255_pci.c +++ b/drivers/staging/comedi/drivers/8255_pci.c @@ -63,7 +63,8 @@ enum pci_8255_boardid { BOARD_ADLINK_PCI7296, BOARD_CB_PCIDIO24, BOARD_CB_PCIDIO24H, - BOARD_CB_PCIDIO48H, + BOARD_CB_PCIDIO48H_OLD, + BOARD_CB_PCIDIO48H_NEW, BOARD_CB_PCIDIO96H, BOARD_NI_PCIDIO96, BOARD_NI_PCIDIO96B, @@ -106,11 +107,16 @@ static const struct pci_8255_boardinfo pci_8255_boards[] = { .dio_badr = 2, .n_8255 = 1, }, - [BOARD_CB_PCIDIO48H] = { + [BOARD_CB_PCIDIO48H_OLD] = { .name = "cb_pci-dio48h", .dio_badr = 1, .n_8255 = 2, }, + [BOARD_CB_PCIDIO48H_NEW] = { + .name = "cb_pci-dio48h", + .dio_badr = 2, + .n_8255 = 2, + }, [BOARD_CB_PCIDIO96H] = { .name = "cb_pci-dio96h", .dio_badr = 2, @@ -263,7 +269,10 @@ static DEFINE_PCI_DEVICE_TABLE(pci_8255_pci_table) = { { PCI_VDEVICE(ADLINK, 0x7296), BOARD_ADLINK_PCI7296 }, { PCI_VDEVICE(CB, 0x0028), BOARD_CB_PCIDIO24 }, { PCI_VDEVICE(CB, 0x0014), BOARD_CB_PCIDIO24H }, - { PCI_VDEVICE(CB, 0x000b), BOARD_CB_PCIDIO48H }, + { PCI_DEVICE_SUB(PCI_VENDOR_ID_CB, 0x000b, 0x0000, 0x0000), + .driver_data = BOARD_CB_PCIDIO48H_OLD }, + { PCI_DEVICE_SUB(PCI_VENDOR_ID_CB, 0x000b, PCI_VENDOR_ID_CB, 0x000b), + .driver_data = BOARD_CB_PCIDIO48H_NEW }, { PCI_VDEVICE(CB, 0x0017), BOARD_CB_PCIDIO96H }, { PCI_VDEVICE(NI, 0x0160), BOARD_NI_PCIDIO96 }, { PCI_VDEVICE(NI, 0x1630), BOARD_NI_PCIDIO96B }, -- cgit v0.10.2 From c6236c0ce39c809c336ca929f68cf8ad02cf94e0 Mon Sep 17 00:00:00 2001 From: H Hartley Sweeten Date: Tue, 10 Dec 2013 16:31:25 -0700 Subject: staging: comedi: drivers: fix return value of comedi_load_firmware() Some of the callback functions that upload the firmware in the comedi drivers return a positive value indicating the number of bytes sent to the device. Detect this condition and just return '0' to indicate a successful upload. Reported-by: Bernd Porr Signed-off-by: H Hartley Sweeten Acked-by: Ian Abbott Acked-by: Dan Carpenter Cc: stable Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/comedi/drivers.c b/drivers/staging/comedi/drivers.c index 8f02bf6..4964d2a 100644 --- a/drivers/staging/comedi/drivers.c +++ b/drivers/staging/comedi/drivers.c @@ -446,7 +446,7 @@ int comedi_load_firmware(struct comedi_device *dev, release_firmware(fw); } - return ret; + return ret < 0 ? ret : 0; } EXPORT_SYMBOL_GPL(comedi_load_firmware); -- cgit v0.10.2 From 53385d2d1de84f4036a0919ec46964c4e81b83f5 Mon Sep 17 00:00:00 2001 From: Bob Gilligan Date: Sun, 15 Dec 2013 13:39:56 -0800 Subject: neigh: Netlink notification for administrative NUD state change The neighbour code sends up an RTM_NEWNEIGH netlink notification if the NUD state of a neighbour cache entry is changed by a timer (e.g. from REACHABLE to STALE), even if the lladdr of the entry has not changed. But an administrative change to the the NUD state of a neighbour cache entry that does not change the lladdr (e.g. via "ip -4 neigh change ... nud ...") does not trigger a netlink notification. This means that netlink listeners will not hear about administrative NUD state changes such as from a resolved state to PERMANENT. This patch changes the neighbor code to generate an RTM_NEWNEIGH message when the NUD state of an entry is changed administratively. Signed-off-by: Bob Gilligan Acked-by: Nicolas Dichtel Signed-off-by: David S. Miller diff --git a/net/core/neighbour.c b/net/core/neighbour.c index ca15f32..36b1443 100644 --- a/net/core/neighbour.c +++ b/net/core/neighbour.c @@ -1161,6 +1161,7 @@ int neigh_update(struct neighbour *neigh, const u8 *lladdr, u8 new, neigh->parms->reachable_time : 0))); neigh->nud_state = new; + notify = 1; } if (lladdr != neigh->ha) { -- cgit v0.10.2 From 7022ef8b2a673db3d12a348331181f579afe9b22 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Mon, 16 Dec 2013 10:45:05 +0800 Subject: xen-netback: fix fragments error handling in checksum_setup_ip() Fix to return -EPROTO error if fragments detected in checksum_setup_ip(). Fixes: 1431fb31ecba ('xen-netback: fix fragment detection in checksum setup') Signed-off-by: Wei Yongjun Reviewed-by: Paul Durrant Acked-by: Ian Campbell Signed-off-by: David S. Miller diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c index e884ee1..27bbe58 100644 --- a/drivers/net/xen-netback/netback.c +++ b/drivers/net/xen-netback/netback.c @@ -1197,6 +1197,9 @@ static int checksum_setup_ip(struct xenvif *vif, struct sk_buff *skb, err = -EPROTO; + if (fragment) + goto out; + switch (ip_hdr(skb)->protocol) { case IPPROTO_TCP: err = maybe_pull_tail(skb, -- cgit v0.10.2 From fb5f1834c3221e459324c6885eaad75429f722a5 Mon Sep 17 00:00:00 2001 From: Boris BREZILLON Date: Sun, 8 Dec 2013 15:59:59 +0100 Subject: usb: ohci-at91: fix irq and iomem resource retrieval When using dt resources retrieval (interrupts and reg properties) there is no predefined order for these resources in the platform dev resources table. Retrieve resources using platform_get_resource and platform_get_irq functions instead of direct resource table entries to avoid resource type mismatch. Signed-off-by: Boris BREZILLON Acked-by: Nicolas Ferre Signed-off-by: Alan Stern Reviewed-by: Tomasz Figa Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/usb/host/ohci-at91.c b/drivers/usb/host/ohci-at91.c index 418444e..8c356af 100644 --- a/drivers/usb/host/ohci-at91.c +++ b/drivers/usb/host/ohci-at91.c @@ -136,23 +136,27 @@ static int usb_hcd_at91_probe(const struct hc_driver *driver, struct ohci_hcd *ohci; int retval; struct usb_hcd *hcd = NULL; - - if (pdev->num_resources != 2) { - pr_debug("hcd probe: invalid num_resources"); - return -ENODEV; + struct device *dev = &pdev->dev; + struct resource *res; + int irq; + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!res) { + dev_dbg(dev, "hcd probe: missing memory resource\n"); + return -ENXIO; } - if ((pdev->resource[0].flags != IORESOURCE_MEM) - || (pdev->resource[1].flags != IORESOURCE_IRQ)) { - pr_debug("hcd probe: invalid resource type\n"); - return -ENODEV; + irq = platform_get_irq(pdev, 0); + if (irq < 0) { + dev_dbg(dev, "hcd probe: missing irq resource\n"); + return irq; } hcd = usb_create_hcd(driver, &pdev->dev, "at91"); if (!hcd) return -ENOMEM; - hcd->rsrc_start = pdev->resource[0].start; - hcd->rsrc_len = resource_size(&pdev->resource[0]); + hcd->rsrc_start = res->start; + hcd->rsrc_len = resource_size(res); if (!request_mem_region(hcd->rsrc_start, hcd->rsrc_len, hcd_name)) { pr_debug("request_mem_region failed\n"); @@ -199,7 +203,7 @@ static int usb_hcd_at91_probe(const struct hc_driver *driver, ohci->num_ports = board->ports; at91_start_hc(pdev); - retval = usb_add_hcd(hcd, pdev->resource[1].start, IRQF_SHARED); + retval = usb_add_hcd(hcd, irq, IRQF_SHARED); if (retval == 0) return retval; -- cgit v0.10.2 From b84caae486135d588fb200973b0be8cb8a511edf Mon Sep 17 00:00:00 2001 From: Himanshu Madhani Date: Mon, 16 Dec 2013 15:36:56 -0500 Subject: qlcnic: Fix usage of netif_tx_{wake, stop} api during link change. o Driver was using netif_tx_{stop,wake}_all_queues() api during link change event. Remove these api calls to manage queue start/stop event, as core networking stack will manage this based on netif_carrier_{on,off} call. These API's were modified as part of commit id 012ec81223aa45d2b80aeafb77392fd1a19c7b10 ("qlcnic: Multi Tx queue support for 82xx Series adapter.") Signed-off-by: Shahed Shaikh Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c index 0149c94..eda6c69 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c @@ -687,17 +687,11 @@ void qlcnic_advert_link_change(struct qlcnic_adapter *adapter, int linkup) if (adapter->ahw->linkup && !linkup) { netdev_info(netdev, "NIC Link is down\n"); adapter->ahw->linkup = 0; - if (netif_running(netdev)) { - netif_carrier_off(netdev); - netif_tx_stop_all_queues(netdev); - } + netif_carrier_off(netdev); } else if (!adapter->ahw->linkup && linkup) { netdev_info(netdev, "NIC Link is up\n"); adapter->ahw->linkup = 1; - if (netif_running(netdev)) { - netif_carrier_on(netdev); - netif_wake_queue(netdev); - } + netif_carrier_on(netdev); } } -- cgit v0.10.2 From 3bf517df0d99f1cf5d369c73ab68e0afe6a3c2f9 Mon Sep 17 00:00:00 2001 From: Himanshu Madhani Date: Mon, 16 Dec 2013 15:36:57 -0500 Subject: qlcnic: Fix diagnostic test for all adapters. o Driver should re-allocate all Tx queues after completing diagnostic tests. This regression was added by commit id c2c5e3a0681bb1945c0cb211a5f4baa22cb2cbb3 ("qlcnic: Enable diagnostic test for multiple Tx queues.") Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c index b36c02f..6d3edf6 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c @@ -948,6 +948,7 @@ static int qlcnic_irq_test(struct net_device *netdev) struct qlcnic_hardware_context *ahw = adapter->ahw; struct qlcnic_cmd_args cmd; int ret, drv_sds_rings = adapter->drv_sds_rings; + int drv_tx_rings = adapter->drv_tx_rings; if (qlcnic_83xx_check(adapter)) return qlcnic_83xx_interrupt_test(netdev); @@ -980,6 +981,7 @@ free_diag_res: clear_diag_irq: adapter->drv_sds_rings = drv_sds_rings; + adapter->drv_tx_rings = drv_tx_rings; clear_bit(__QLCNIC_RESETTING, &adapter->state); return ret; diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c index 05c1eef..aa019c3 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c @@ -1940,7 +1940,6 @@ int qlcnic_diag_alloc_res(struct net_device *netdev, int test) qlcnic_detach(adapter); adapter->drv_sds_rings = QLCNIC_SINGLE_RING; - adapter->drv_tx_rings = QLCNIC_SINGLE_RING; adapter->ahw->diag_test = test; adapter->ahw->linkup = 0; -- cgit v0.10.2 From f9566265d7b44fea789072dcfa9a454e7e433af6 Mon Sep 17 00:00:00 2001 From: Himanshu Madhani Date: Mon, 16 Dec 2013 15:36:58 -0500 Subject: qlcnic: Fix TSS/RSS ring validation logic. o TSS/RSS ring validation does not take into account that either of these ring values can be 0. This patch fixes this validation and would fail set_channel operation if any of these ring value is 0. This regression was added as part of commit id 34e8c406fda5b5a9d2e126a92bab84cd28e3b5fa ("qlcnic: refactor Tx/SDS ring calculation and validation in driver.") Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c index 89208e5..fae1b71 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c @@ -2073,8 +2073,8 @@ int qlcnic_83xx_configure_opmode(struct qlcnic_adapter *adapter) ahw->nic_mode = QLCNIC_DEFAULT_MODE; adapter->nic_ops->init_driver = qlcnic_83xx_init_default_driver; ahw->idc.state_entry = qlcnic_83xx_idc_ready_state_entry; - adapter->max_sds_rings = ahw->max_rx_ques; - adapter->max_tx_rings = ahw->max_tx_ques; + adapter->max_sds_rings = QLCNIC_MAX_SDS_RINGS; + adapter->max_tx_rings = QLCNIC_MAX_TX_RINGS; } else { return -EIO; } diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c index 6d3edf6..78f5e81 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c @@ -667,9 +667,13 @@ qlcnic_set_ringparam(struct net_device *dev, static int qlcnic_validate_ring_count(struct qlcnic_adapter *adapter, u8 rx_ring, u8 tx_ring) { + if (rx_ring == 0 || tx_ring == 0) + return -EINVAL; + if (rx_ring != 0) { if (rx_ring > adapter->max_sds_rings) { - netdev_err(adapter->netdev, "Invalid ring count, SDS ring count %d should not be greater than max %d driver sds rings.\n", + netdev_err(adapter->netdev, + "Invalid ring count, SDS ring count %d should not be greater than max %d driver sds rings.\n", rx_ring, adapter->max_sds_rings); return -EINVAL; } diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c index aa019c3..2c8cac0 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c @@ -1178,6 +1178,7 @@ qlcnic_initialize_nic(struct qlcnic_adapter *adapter) } else { adapter->ahw->nic_mode = QLCNIC_DEFAULT_MODE; adapter->max_tx_rings = QLCNIC_MAX_HW_TX_RINGS; + adapter->max_sds_rings = QLCNIC_MAX_SDS_RINGS; adapter->flags &= ~QLCNIC_ESWITCH_ENABLED; } -- cgit v0.10.2 From b17a44d8b86f48e34011b884a934231ae2928d66 Mon Sep 17 00:00:00 2001 From: Himanshu Madhani Date: Mon, 16 Dec 2013 15:36:59 -0500 Subject: qlcnic: Fix TSS/RSS validation for 83xx/84xx series adapter. o Current code was not allowing the user to configure more than one Tx ring using ethtool for 83xx/84xx adapter. This regression was introduced by commit id 18afc102fdcb95d6c7d57f2967a06f2f8fe3ba4c ("qlcnic: Enable multiple Tx queue support for 83xx/84xx Series adapter.") Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c index 78f5e81..e3be276 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c @@ -680,21 +680,12 @@ static int qlcnic_validate_ring_count(struct qlcnic_adapter *adapter, } if (tx_ring != 0) { - if (qlcnic_82xx_check(adapter) && - (tx_ring > adapter->max_tx_rings)) { + if (tx_ring > adapter->max_tx_rings) { netdev_err(adapter->netdev, "Invalid ring count, Tx ring count %d should not be greater than max %d driver Tx rings.\n", tx_ring, adapter->max_tx_rings); return -EINVAL; } - - if (qlcnic_83xx_check(adapter) && - (tx_ring > QLCNIC_SINGLE_RING)) { - netdev_err(adapter->netdev, - "Invalid ring count, Tx ring count %d should not be greater than %d driver Tx rings.\n", - tx_ring, QLCNIC_SINGLE_RING); - return -EINVAL; - } } return 0; -- cgit v0.10.2 From 3fc38e267bfbea5e33e2222d6babc3b06c2bb642 Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Mon, 16 Dec 2013 15:37:00 -0500 Subject: qlcnic: Fix memory allocation o Use vzalloc() instead of kzalloc() for allocation of bootloader size memory. kzalloc() may fail to allocate the size of bootloader Signed-off-by: Manish Chopra Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c index fae1b71..cac0503 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c @@ -1254,24 +1254,24 @@ static int qlcnic_83xx_copy_bootloader(struct qlcnic_adapter *adapter) if (size & 0xF) size = (size + 16) & ~0xF; - p_cache = kzalloc(size, GFP_KERNEL); + p_cache = vzalloc(size); if (p_cache == NULL) return -ENOMEM; ret = qlcnic_83xx_lockless_flash_read32(adapter, src, p_cache, size / sizeof(u32)); if (ret) { - kfree(p_cache); + vfree(p_cache); return ret; } /* 16 byte write to MS memory */ ret = qlcnic_83xx_ms_mem_write128(adapter, dest, (u32 *)p_cache, size / 16); if (ret) { - kfree(p_cache); + vfree(p_cache); return ret; } - kfree(p_cache); + vfree(p_cache); return ret; } -- cgit v0.10.2 From 30fa15f64e40c4cf037e3379e55c2323b5d992e2 Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Mon, 16 Dec 2013 15:37:01 -0500 Subject: qlcnic: Allow firmware dump collection when auto firmware recovery is disabled o Allow driver to collect firmware dump, during a forced firmware dump operation, when auto firmware recovery is disabled. Also, during this operation, driver should not allow reset recovery to be performed. Signed-off-by: Manish Chopra Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h index 4cae6ca..a6a3350 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h @@ -662,4 +662,5 @@ pci_ers_result_t qlcnic_83xx_io_error_detected(struct pci_dev *, pci_channel_state_t); pci_ers_result_t qlcnic_83xx_io_slot_reset(struct pci_dev *); void qlcnic_83xx_io_resume(struct pci_dev *); +void qlcnic_83xx_stop_hw(struct qlcnic_adapter *); #endif diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c index cac0503..918e18d 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c @@ -740,6 +740,7 @@ static int qlcnic_83xx_idc_unknown_state(struct qlcnic_adapter *adapter) adapter->ahw->idc.err_code = -EIO; dev_err(&adapter->pdev->dev, "%s: Device in unknown state\n", __func__); + clear_bit(__QLCNIC_RESETTING, &adapter->state); return 0; } @@ -818,7 +819,6 @@ static int qlcnic_83xx_idc_ready_state(struct qlcnic_adapter *adapter) struct qlcnic_hardware_context *ahw = adapter->ahw; struct qlcnic_mailbox *mbx = ahw->mailbox; int ret = 0; - u32 owner; u32 val; /* Perform NIC configuration based ready state entry actions */ @@ -848,9 +848,9 @@ static int qlcnic_83xx_idc_ready_state(struct qlcnic_adapter *adapter) set_bit(__QLCNIC_RESETTING, &adapter->state); qlcnic_83xx_idc_enter_need_reset_state(adapter, 1); } else { - owner = qlcnic_83xx_idc_find_reset_owner_id(adapter); - if (ahw->pci_func == owner) - qlcnic_dump_fw(adapter); + netdev_info(adapter->netdev, "%s: Auto firmware recovery is disabled\n", + __func__); + qlcnic_83xx_idc_enter_failed_state(adapter, 1); } return -EIO; } @@ -948,13 +948,26 @@ static int qlcnic_83xx_idc_need_quiesce_state(struct qlcnic_adapter *adapter) return 0; } -static int qlcnic_83xx_idc_failed_state(struct qlcnic_adapter *adapter) +static void qlcnic_83xx_idc_failed_state(struct qlcnic_adapter *adapter) { - dev_err(&adapter->pdev->dev, "%s: please restart!!\n", __func__); + struct qlcnic_hardware_context *ahw = adapter->ahw; + u32 val, owner; + + val = QLCRDX(adapter->ahw, QLC_83XX_IDC_CTRL); + if (val & QLC_83XX_IDC_DISABLE_FW_RESET_RECOVERY) { + owner = qlcnic_83xx_idc_find_reset_owner_id(adapter); + if (ahw->pci_func == owner) { + qlcnic_83xx_stop_hw(adapter); + qlcnic_dump_fw(adapter); + } + } + + netdev_warn(adapter->netdev, "%s: Reboot will be required to recover the adapter!!\n", + __func__); clear_bit(__QLCNIC_RESETTING, &adapter->state); - adapter->ahw->idc.err_code = -EIO; + ahw->idc.err_code = -EIO; - return 0; + return; } static int qlcnic_83xx_idc_quiesce_state(struct qlcnic_adapter *adapter) @@ -1063,12 +1076,6 @@ void qlcnic_83xx_idc_poll_dev_state(struct work_struct *work) adapter->ahw->idc.prev_state = adapter->ahw->idc.curr_state; qlcnic_83xx_periodic_tasks(adapter); - /* Do not reschedule if firmaware is in hanged state and auto - * recovery is disabled - */ - if ((adapter->flags & QLCNIC_FW_HANG) && !qlcnic_auto_fw_reset) - return; - /* Re-schedule the function */ if (test_bit(QLC_83XX_MODULE_LOADED, &adapter->ahw->idc.status)) qlcnic_schedule_work(adapter, qlcnic_83xx_idc_poll_dev_state, @@ -1219,10 +1226,10 @@ void qlcnic_83xx_idc_request_reset(struct qlcnic_adapter *adapter, u32 key) } val = QLCRDX(adapter->ahw, QLC_83XX_IDC_CTRL); - if ((val & QLC_83XX_IDC_DISABLE_FW_RESET_RECOVERY) || - !qlcnic_auto_fw_reset) { - dev_err(&adapter->pdev->dev, - "%s:failed, device in non reset mode\n", __func__); + if (val & QLC_83XX_IDC_DISABLE_FW_RESET_RECOVERY) { + netdev_info(adapter->netdev, "%s: Auto firmware recovery is disabled\n", + __func__); + qlcnic_83xx_idc_enter_failed_state(adapter, 0); qlcnic_83xx_unlock_driver(adapter); return; } @@ -1939,7 +1946,7 @@ static void qlcnic_83xx_exec_template_cmd(struct qlcnic_adapter *p_dev, p_dev->ahw->reset.seq_index = index; } -static void qlcnic_83xx_stop_hw(struct qlcnic_adapter *p_dev) +void qlcnic_83xx_stop_hw(struct qlcnic_adapter *p_dev) { p_dev->ahw->reset.seq_index = 0; @@ -1994,6 +2001,14 @@ static int qlcnic_83xx_restart_hw(struct qlcnic_adapter *adapter) val = QLCRDX(adapter->ahw, QLC_83XX_IDC_CTRL); if (!(val & QLC_83XX_IDC_GRACEFULL_RESET)) qlcnic_dump_fw(adapter); + + if (val & QLC_83XX_IDC_DISABLE_FW_RESET_RECOVERY) { + netdev_info(adapter->netdev, "%s: Auto firmware recovery is disabled\n", + __func__); + qlcnic_83xx_idc_enter_failed_state(adapter, 1); + return err; + } + qlcnic_83xx_init_hw(adapter); if (qlcnic_83xx_copy_bootloader(adapter)) -- cgit v0.10.2 From e49df7947a48b04ee57ee0fa0c4110ef05024c4f Mon Sep 17 00:00:00 2001 From: Manish chopra Date: Mon, 16 Dec 2013 15:37:02 -0500 Subject: qlcnic: Fix mailbox processing during diagnostic test o Do not enable mailbox polling in case of legacy interrupt. Process mailbox AEN/response from the interrupt. Signed-off-by: Manish Chopra Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c index b1cb0ff..ab66e7f 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c @@ -447,8 +447,9 @@ irqreturn_t qlcnic_83xx_intr(int irq, void *data) qlcnic_83xx_poll_process_aen(adapter); - if (ahw->diag_test == QLCNIC_INTERRUPT_TEST) { - ahw->diag_cnt++; + if (ahw->diag_test) { + if (ahw->diag_test == QLCNIC_INTERRUPT_TEST) + ahw->diag_cnt++; qlcnic_83xx_enable_legacy_msix_mbx_intr(adapter); return IRQ_HANDLED; } @@ -1345,11 +1346,6 @@ static int qlcnic_83xx_diag_alloc_res(struct net_device *netdev, int test, } if (adapter->ahw->diag_test == QLCNIC_LOOPBACK_TEST) { - /* disable and free mailbox interrupt */ - if (!(adapter->flags & QLCNIC_MSIX_ENABLED)) { - qlcnic_83xx_enable_mbx_poll(adapter); - qlcnic_83xx_free_mbx_intr(adapter); - } adapter->ahw->loopback_state = 0; adapter->ahw->hw_ops->setup_link_event(adapter, 1); } @@ -1363,33 +1359,20 @@ static void qlcnic_83xx_diag_free_res(struct net_device *netdev, { struct qlcnic_adapter *adapter = netdev_priv(netdev); struct qlcnic_host_sds_ring *sds_ring; - int ring, err; + int ring; clear_bit(__QLCNIC_DEV_UP, &adapter->state); if (adapter->ahw->diag_test == QLCNIC_INTERRUPT_TEST) { for (ring = 0; ring < adapter->drv_sds_rings; ring++) { sds_ring = &adapter->recv_ctx->sds_rings[ring]; - qlcnic_83xx_disable_intr(adapter, sds_ring); - if (!(adapter->flags & QLCNIC_MSIX_ENABLED)) - qlcnic_83xx_enable_mbx_poll(adapter); + if (adapter->flags & QLCNIC_MSIX_ENABLED) + qlcnic_83xx_disable_intr(adapter, sds_ring); } } qlcnic_fw_destroy_ctx(adapter); qlcnic_detach(adapter); - if (adapter->ahw->diag_test == QLCNIC_LOOPBACK_TEST) { - if (!(adapter->flags & QLCNIC_MSIX_ENABLED)) { - err = qlcnic_83xx_setup_mbx_intr(adapter); - qlcnic_83xx_disable_mbx_poll(adapter); - if (err) { - dev_err(&adapter->pdev->dev, - "%s: failed to setup mbx interrupt\n", - __func__); - goto out; - } - } - } adapter->ahw->diag_test = 0; adapter->drv_sds_rings = drv_sds_rings; @@ -1399,9 +1382,6 @@ static void qlcnic_83xx_diag_free_res(struct net_device *netdev, if (netif_running(netdev)) __qlcnic_up(adapter, netdev); - if (adapter->ahw->diag_test == QLCNIC_INTERRUPT_TEST && - !(adapter->flags & QLCNIC_MSIX_ENABLED)) - qlcnic_83xx_disable_mbx_poll(adapter); out: netif_device_attach(netdev); } -- cgit v0.10.2 From 0951c5c214e525185572204b587e3b6762d121f4 Mon Sep 17 00:00:00 2001 From: Manish chopra Date: Mon, 16 Dec 2013 15:37:03 -0500 Subject: qlcnic: Dump mailbox registers when mailbox command times out. Signed-off-by: Manish Chopra Signed-off-by: Himanshu Madhani Signed-off-by: David S. Miller diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c index ab66e7f..6055d39 100644 --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c @@ -3734,6 +3734,19 @@ static void qlcnic_83xx_decode_mbx_rsp(struct qlcnic_adapter *adapter, return; } +static inline void qlcnic_dump_mailbox_registers(struct qlcnic_adapter *adapter) +{ + struct qlcnic_hardware_context *ahw = adapter->ahw; + u32 offset; + + offset = QLCRDX(ahw, QLCNIC_DEF_INT_MASK); + dev_info(&adapter->pdev->dev, "Mbx interrupt mask=0x%x, Mbx interrupt enable=0x%x, Host mbx control=0x%x, Fw mbx control=0x%x", + readl(ahw->pci_base0 + offset), + QLCRDX(ahw, QLCNIC_MBX_INTR_ENBL), + QLCRDX(ahw, QLCNIC_HOST_MBX_CTRL), + QLCRDX(ahw, QLCNIC_FW_MBX_CTRL)); +} + static void qlcnic_83xx_mailbox_worker(struct work_struct *work) { struct qlcnic_mailbox *mbx = container_of(work, struct qlcnic_mailbox, @@ -3778,6 +3791,8 @@ static void qlcnic_83xx_mailbox_worker(struct work_struct *work) __func__, cmd->cmd_op, cmd->type, ahw->pci_func, ahw->op_mode); clear_bit(QLC_83XX_MBX_READY, &mbx->status); + qlcnic_dump_mailbox_registers(adapter); + qlcnic_83xx_get_mbx_data(adapter, cmd); qlcnic_dump_mbx(adapter, cmd); qlcnic_83xx_idc_request_reset(adapter, QLCNIC_FORCE_FW_DUMP_KEY); -- cgit v0.10.2 From c2db11eca089b60148ded7467b956a547e8a2ae0 Mon Sep 17 00:00:00 2001 From: Soren Brinkmann Date: Mon, 2 Dec 2013 11:38:38 -0800 Subject: tty: xuartps: Properly guard sysrq specific code Commit 'tty: xuartps: Implement BREAK detection, add SYSRQ support' (0c0c47bc40a2e358d593b2d7fb93b50027fbfc0c) introduced sysrq support without properly guarding sysrq specific code which results in build errors when sysrq is disabled: DNAME=KBUILD_STR(xilinx_uartps)" -c -o drivers/tty/serial/.tmp_xilinx_uartps.o drivers/tty/serial/xilinx_uartps.c drivers/tty/serial/xilinx_uartps.c: In function 'xuartps_isr': drivers/tty/serial/xilinx_uartps.c:247:5: error: 'struct uart_port' has no member named 'sysrq' drivers/tty/serial/xilinx_uartps.c:247:5: error: 'struct uart_port' has no member named 'sysrq' drivers/tty/serial/xilinx_uartps.c:247:5: error: 'struct uart_port' has no member named 'sysrq' make[3]: *** [drivers/tty/serial/xilinx_uartps.o] Error 1 Reported-by: Masanari Iida Cc: Vlad Lungu Signed-off-by: Soren Brinkmann Signed-off-by: Mark Brown Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c index e46e9f3..f619ad5 100644 --- a/drivers/tty/serial/xilinx_uartps.c +++ b/drivers/tty/serial/xilinx_uartps.c @@ -240,6 +240,7 @@ static irqreturn_t xuartps_isr(int irq, void *dev_id) continue; } +#ifdef SUPPORT_SYSRQ /* * uart_handle_sysrq_char() doesn't work if * spinlocked, for some reason @@ -253,6 +254,7 @@ static irqreturn_t xuartps_isr(int irq, void *dev_id) } spin_lock(&port->lock); } +#endif port->icount.rx++; -- cgit v0.10.2 From 130f769e81fc472beb2211320777e26050e3fa15 Mon Sep 17 00:00:00 2001 From: Tomi Valkeinen Date: Mon, 16 Dec 2013 09:14:48 +0200 Subject: Revert "ARM: OMAP2+: Remove legacy mux code for display.c" Commit e30b06f4d5f000c31a7747a7e7ada78a5fd419a1 (ARM: OMAP2+: Remove legacy mux code for display.c) removed non-DT DSI and HDMI pinmuxing. However, DSI pinmuxing is still needed, and removing that caused DSI displays not to work. This reverts the DSI parts of the commit. Signed-off-by: Tomi Valkeinen Signed-off-by: Tony Lindgren diff --git a/arch/arm/mach-omap2/display.c b/arch/arm/mach-omap2/display.c index 58347bb..4cf1655 100644 --- a/arch/arm/mach-omap2/display.c +++ b/arch/arm/mach-omap2/display.c @@ -101,13 +101,51 @@ static const struct omap_dss_hwmod_data omap4_dss_hwmod_data[] __initconst = { { "dss_hdmi", "omapdss_hdmi", -1 }, }; +static int omap4_dsi_mux_pads(int dsi_id, unsigned lanes) +{ + u32 enable_mask, enable_shift; + u32 pipd_mask, pipd_shift; + u32 reg; + + if (dsi_id == 0) { + enable_mask = OMAP4_DSI1_LANEENABLE_MASK; + enable_shift = OMAP4_DSI1_LANEENABLE_SHIFT; + pipd_mask = OMAP4_DSI1_PIPD_MASK; + pipd_shift = OMAP4_DSI1_PIPD_SHIFT; + } else if (dsi_id == 1) { + enable_mask = OMAP4_DSI2_LANEENABLE_MASK; + enable_shift = OMAP4_DSI2_LANEENABLE_SHIFT; + pipd_mask = OMAP4_DSI2_PIPD_MASK; + pipd_shift = OMAP4_DSI2_PIPD_SHIFT; + } else { + return -ENODEV; + } + + reg = omap4_ctrl_pad_readl(OMAP4_CTRL_MODULE_PAD_CORE_CONTROL_DSIPHY); + + reg &= ~enable_mask; + reg &= ~pipd_mask; + + reg |= (lanes << enable_shift) & enable_mask; + reg |= (lanes << pipd_shift) & pipd_mask; + + omap4_ctrl_pad_writel(reg, OMAP4_CTRL_MODULE_PAD_CORE_CONTROL_DSIPHY); + + return 0; +} + static int omap_dsi_enable_pads(int dsi_id, unsigned lane_mask) { + if (cpu_is_omap44xx()) + return omap4_dsi_mux_pads(dsi_id, lane_mask); + return 0; } static void omap_dsi_disable_pads(int dsi_id, unsigned lane_mask) { + if (cpu_is_omap44xx()) + omap4_dsi_mux_pads(dsi_id, 0); } static int omap_dss_set_min_bus_tput(struct device *dev, unsigned long tput) -- cgit v0.10.2 From 82832046e285952f14e707647f9ef82466c883cc Mon Sep 17 00:00:00 2001 From: Russell King Date: Mon, 16 Dec 2013 11:33:24 +0000 Subject: imx-drm: imx-drm-core: fix error cleanup path for imx_drm_add_crtc() imx_drm_add_crtc() was kfree'ing the imx_drm_crtc structure while leaving it on the list of CRTCs. Delete it from the list first. Signed-off-by: Russell King Acked-by: Sascha Hauer Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/imx-drm/imx-drm-core.c b/drivers/staging/imx-drm/imx-drm-core.c index 6bd015a..b5ff45f3 100644 --- a/drivers/staging/imx-drm/imx-drm-core.c +++ b/drivers/staging/imx-drm/imx-drm-core.c @@ -528,6 +528,7 @@ int imx_drm_add_crtc(struct drm_crtc *crtc, return 0; err_register: + list_del(&imx_drm_crtc->list); kfree(imx_drm_crtc); err_alloc: err_busy: -- cgit v0.10.2 From 8007875f0619f36d885ba357b587e673d6f20704 Mon Sep 17 00:00:00 2001 From: Russell King Date: Mon, 16 Dec 2013 11:33:44 +0000 Subject: imx-drm: imx-drm-core: fix DRM cleanup paths We must call drm_vblank_cleanup() on the error cleanup and unload paths after we've had a successful call to drm_vblank_init(). Ensure that the calls are in the reverse order to the initialisation order. Signed-off-by: Russell King Acked-by: Sascha Hauer Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/imx-drm/imx-drm-core.c b/drivers/staging/imx-drm/imx-drm-core.c index b5ff45f3..aff4bd4 100644 --- a/drivers/staging/imx-drm/imx-drm-core.c +++ b/drivers/staging/imx-drm/imx-drm-core.c @@ -88,8 +88,9 @@ static int imx_drm_driver_unload(struct drm_device *drm) imx_drm_device_put(); - drm_mode_config_cleanup(imxdrm->drm); + drm_vblank_cleanup(imxdrm->drm); drm_kms_helper_poll_fini(imxdrm->drm); + drm_mode_config_cleanup(imxdrm->drm); return 0; } @@ -428,11 +429,11 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags) ret = drm_mode_group_init_legacy_group(imxdrm->drm, &imxdrm->drm->primary->mode_group); if (ret) - goto err_init; + goto err_kms; ret = drm_vblank_init(imxdrm->drm, MAX_CRTC); if (ret) - goto err_init; + goto err_kms; /* * with vblank_disable_allowed = true, vblank interrupt will be disabled @@ -441,12 +442,19 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags) */ imxdrm->drm->vblank_disable_allowed = true; - if (!imx_drm_device_get()) + if (!imx_drm_device_get()) { ret = -EINVAL; + goto err_vblank; + } - ret = 0; + mutex_unlock(&imxdrm->mutex); + return 0; -err_init: +err_vblank: + drm_vblank_cleanup(drm); +err_kms: + drm_kms_helper_poll_fini(drm); + drm_mode_config_cleanup(drm); mutex_unlock(&imxdrm->mutex); return ret; -- cgit v0.10.2 From 4ae078d58ac7fadace7eb30bade8da649ecc4ded Mon Sep 17 00:00:00 2001 From: Russell King Date: Mon, 16 Dec 2013 11:34:25 +0000 Subject: imx-drm: ipu-v3: fix potential CRTC device registration race Clean up the IPUv3 CRTC device registration; we don't need a separate function just to call platform_device_register_data(), and we don't need the return value converted at all. Update the IPU client id under a mutex, so that parallel probing doesn't race. Signed-off-by: Russell King Acked-by: Sascha Hauer Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/imx-drm/ipu-v3/ipu-common.c b/drivers/staging/imx-drm/ipu-v3/ipu-common.c index 7a22ce6..97ca692 100644 --- a/drivers/staging/imx-drm/ipu-v3/ipu-common.c +++ b/drivers/staging/imx-drm/ipu-v3/ipu-common.c @@ -996,35 +996,35 @@ static const struct ipu_platform_reg client_reg[] = { }, }; +static DEFINE_MUTEX(ipu_client_id_mutex); static int ipu_client_id; -static int ipu_add_subdevice_pdata(struct device *dev, - const struct ipu_platform_reg *reg) -{ - struct platform_device *pdev; - - pdev = platform_device_register_data(dev, reg->name, ipu_client_id++, - ®->pdata, sizeof(struct ipu_platform_reg)); - - return PTR_ERR_OR_ZERO(pdev); -} - static int ipu_add_client_devices(struct ipu_soc *ipu) { - int ret; - int i; + struct device *dev = ipu->dev; + unsigned i; + int id, ret; + + mutex_lock(&ipu_client_id_mutex); + id = ipu_client_id; + ipu_client_id += ARRAY_SIZE(client_reg); + mutex_unlock(&ipu_client_id_mutex); for (i = 0; i < ARRAY_SIZE(client_reg); i++) { const struct ipu_platform_reg *reg = &client_reg[i]; - ret = ipu_add_subdevice_pdata(ipu->dev, reg); - if (ret) + struct platform_device *pdev; + + pdev = platform_device_register_data(dev, reg->name, + id++, ®->pdata, sizeof(reg->pdata)); + + if (IS_ERR(pdev)) goto err_register; } return 0; err_register: - platform_device_unregister_children(to_platform_device(ipu->dev)); + platform_device_unregister_children(to_platform_device(dev)); return ret; } -- cgit v0.10.2 From bd5121bbb6b9a89d9d13878fa5bb1e833168ae18 Mon Sep 17 00:00:00 2001 From: Russell King Date: Mon, 16 Dec 2013 12:38:30 +0000 Subject: imx-drm: imx-tve: don't call sleeping functions beneath enable_lock spinlock Enable lock claims that it is serializing tve_enable/disable calls. However, DRM already serialises mode sets with a mutex, which prevents encoder/connector functions being called concurrently. Secondly, holding a spinlock while calling clk_prepare_enable() is wrong; it will cause a might_sleep() warning should that debugging be enabled. So, let's just get rid of the enable_lock. Signed-off-by: Russell King Acked-by: Sascha Hauer Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/imx-drm/imx-tve.c b/drivers/staging/imx-drm/imx-tve.c index 680f4c8..2c44fef 100644 --- a/drivers/staging/imx-drm/imx-tve.c +++ b/drivers/staging/imx-drm/imx-tve.c @@ -114,7 +114,6 @@ struct imx_tve { struct drm_encoder encoder; struct imx_drm_encoder *imx_drm_encoder; struct device *dev; - spinlock_t enable_lock; /* serializes tve_enable/disable */ spinlock_t lock; /* register lock */ bool enabled; int mode; @@ -146,10 +145,8 @@ __releases(&tve->lock) static void tve_enable(struct imx_tve *tve) { - unsigned long flags; int ret; - spin_lock_irqsave(&tve->enable_lock, flags); if (!tve->enabled) { tve->enabled = true; clk_prepare_enable(tve->clk); @@ -169,23 +166,18 @@ static void tve_enable(struct imx_tve *tve) TVE_CD_SM_IEN | TVE_CD_LM_IEN | TVE_CD_MON_END_IEN); - - spin_unlock_irqrestore(&tve->enable_lock, flags); } static void tve_disable(struct imx_tve *tve) { - unsigned long flags; int ret; - spin_lock_irqsave(&tve->enable_lock, flags); if (tve->enabled) { tve->enabled = false; ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG, TVE_IPU_CLK_EN | TVE_EN, 0); clk_disable_unprepare(tve->clk); } - spin_unlock_irqrestore(&tve->enable_lock, flags); } static int tve_setup_tvout(struct imx_tve *tve) @@ -601,7 +593,6 @@ static int imx_tve_probe(struct platform_device *pdev) tve->dev = &pdev->dev; spin_lock_init(&tve->lock); - spin_lock_init(&tve->enable_lock); ddc_node = of_parse_phandle(np, "ddc", 0); if (ddc_node) { -- cgit v0.10.2 From 942325c8b234fd67db976bd529d48320a6a73171 Mon Sep 17 00:00:00 2001 From: Russell King Date: Mon, 16 Dec 2013 12:38:50 +0000 Subject: imx-drm: imx-drm-core: use defined constant for number of CRTCs. We have this definition, there's no reason not to use it. Signed-off-by: Russell King Acked-by: Sascha Hauer Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/imx-drm/imx-drm-core.c b/drivers/staging/imx-drm/imx-drm-core.c index aff4bd4..b52b827 100644 --- a/drivers/staging/imx-drm/imx-drm-core.c +++ b/drivers/staging/imx-drm/imx-drm-core.c @@ -200,8 +200,8 @@ static void imx_drm_driver_preclose(struct drm_device *drm, if (!file->is_master) return; - for (i = 0; i < 4; i++) - imx_drm_disable_vblank(drm , i); + for (i = 0; i < MAX_CRTC; i++) + imx_drm_disable_vblank(drm, i); } static const struct file_operations imx_drm_driver_fops = { -- cgit v0.10.2 From 9fe73d46edd358fc154f7332c8ff312e067255a0 Mon Sep 17 00:00:00 2001 From: Russell King Date: Mon, 16 Dec 2013 12:39:11 +0000 Subject: imx-drm: imx-drm-core: make imx_drm_crtc_register() safer imx_drm_crtc_register() doesn't clean up the CRTC upon failure, which leaves the CRTC attached to the DRM device. Also, it does setup after attaching the CRTC to the DRM device. Fix this by reordering the function such that we do the setup before drm_crtc_init(): this fixes both issues. Signed-off-by: Russell King Acked-by: Sascha Hauer Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/imx-drm/imx-drm-core.c b/drivers/staging/imx-drm/imx-drm-core.c index b52b827..19e431d 100644 --- a/drivers/staging/imx-drm/imx-drm-core.c +++ b/drivers/staging/imx-drm/imx-drm-core.c @@ -377,8 +377,6 @@ static int imx_drm_crtc_register(struct imx_drm_crtc *imx_drm_crtc) struct imx_drm_device *imxdrm = __imx_drm_device(); int ret; - drm_crtc_init(imxdrm->drm, imx_drm_crtc->crtc, - imx_drm_crtc->imx_drm_helper_funcs.crtc_funcs); ret = drm_mode_crtc_set_gamma_size(imx_drm_crtc->crtc, 256); if (ret) return ret; @@ -386,6 +384,9 @@ static int imx_drm_crtc_register(struct imx_drm_crtc *imx_drm_crtc) drm_crtc_helper_add(imx_drm_crtc->crtc, imx_drm_crtc->imx_drm_helper_funcs.crtc_helper_funcs); + drm_crtc_init(imxdrm->drm, imx_drm_crtc->crtc, + imx_drm_crtc->imx_drm_helper_funcs.crtc_funcs); + drm_mode_group_reinit(imxdrm->drm); return 0; -- cgit v0.10.2 From fd6040ed57d8f200ab0cc2abf706c54995a48370 Mon Sep 17 00:00:00 2001 From: Russell King Date: Mon, 16 Dec 2013 12:39:31 +0000 Subject: imx-drm: imx-drm-core: improve safety of imx_drm_add_crtc() We must not add more CRTCs than we have declared to the vblank helpers, otherwise we overflow their arrays. Force failure if we exceed the number of CRTCs. Signed-off-by: Russell King Acked-by: Sascha Hauer Signed-off-by: Greg Kroah-Hartman diff --git a/drivers/staging/imx-drm/imx-drm-core.c b/drivers/staging/imx-drm/imx-drm-core.c index 19e431d..96e4eee 100644 --- a/drivers/staging/imx-drm/imx-drm-core.c +++ b/drivers/staging/imx-drm/imx-drm-core.c @@ -501,6 +501,15 @@ int imx_drm_add_crtc(struct drm_crtc *crtc, mutex_lock(&imxdrm->mutex); + /* + * The vblank arrays are dimensioned by MAX_CRTC - we can't + * pass IDs greater than this to those functions. + */ + if (imxdrm->pipes >= MAX_CRTC) { + ret = -EINVAL; + goto err_busy; + } + if (imxdrm->drm->open_count) { ret = -EBUSY; goto err_busy; -- cgit v0.10.2 From 36e7bb38028d3d812aa7749208249d600a30c22c Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Mon, 11 Nov 2013 19:29:47 +0530 Subject: powerpc: book3s: kvm: Don't abuse host r2 in exit path We don't use PACATOC for PR. Avoid updating HOST_R2 with PR KVM mode when both HV and PR are enabled in the kernel. Without this we get the below crash (qemu) Unable to handle kernel paging request for data at address 0xffffffffffff8310 Faulting instruction address: 0xc00000000001d5a4 cpu 0x2: Vector: 300 (Data Access) at [c0000001dc53aef0] pc: c00000000001d5a4: .vtime_delta.isra.1+0x34/0x1d0 lr: c00000000001d760: .vtime_account_system+0x20/0x60 sp: c0000001dc53b170 msr: 8000000000009032 dar: ffffffffffff8310 dsisr: 40000000 current = 0xc0000001d76c62d0 paca = 0xc00000000fef1100 softe: 0 irq_happened: 0x01 pid = 4472, comm = qemu-system-ppc enter ? for help [c0000001dc53b200] c00000000001d760 .vtime_account_system+0x20/0x60 [c0000001dc53b290] c00000000008d050 .kvmppc_handle_exit_pr+0x60/0xa50 [c0000001dc53b340] c00000000008f51c kvm_start_lightweight+0xb4/0xc4 [c0000001dc53b510] c00000000008cdf0 .kvmppc_vcpu_run_pr+0x150/0x2e0 [c0000001dc53b9e0] c00000000008341c .kvmppc_vcpu_run+0x2c/0x40 [c0000001dc53ba50] c000000000080af4 .kvm_arch_vcpu_ioctl_run+0x54/0x1b0 [c0000001dc53bae0] c00000000007b4c8 .kvm_vcpu_ioctl+0x478/0x730 [c0000001dc53bca0] c0000000002140cc .do_vfs_ioctl+0x4ac/0x770 [c0000001dc53bd80] c0000000002143e8 .SyS_ioctl+0x58/0xb0 [c0000001dc53be30] c000000000009e58 syscall_exit+0x0/0x98 Signed-off-by: Alexander Graf diff --git a/arch/powerpc/include/asm/kvm_book3s_asm.h b/arch/powerpc/include/asm/kvm_book3s_asm.h index 412b2f3..192917d 100644 --- a/arch/powerpc/include/asm/kvm_book3s_asm.h +++ b/arch/powerpc/include/asm/kvm_book3s_asm.h @@ -79,6 +79,7 @@ struct kvmppc_host_state { ulong vmhandler; ulong scratch0; ulong scratch1; + ulong scratch2; u8 in_guest; u8 restore_hid5; u8 napping; diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index 2ea5cc0..d3de010 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -576,6 +576,7 @@ int main(void) HSTATE_FIELD(HSTATE_VMHANDLER, vmhandler); HSTATE_FIELD(HSTATE_SCRATCH0, scratch0); HSTATE_FIELD(HSTATE_SCRATCH1, scratch1); + HSTATE_FIELD(HSTATE_SCRATCH2, scratch2); HSTATE_FIELD(HSTATE_IN_GUEST, in_guest); HSTATE_FIELD(HSTATE_RESTORE_HID5, restore_hid5); HSTATE_FIELD(HSTATE_NAPPING, napping); diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S index bde28da..be4fa04a 100644 --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S @@ -754,15 +754,14 @@ kvmppc_interrupt_hv: * guest CR, R12 saved in shadow VCPU SCRATCH1/0 * guest R13 saved in SPRN_SCRATCH0 */ - /* abuse host_r2 as third scratch area; we get r2 from PACATOC(r13) */ - std r9, HSTATE_HOST_R2(r13) + std r9, HSTATE_SCRATCH2(r13) lbz r9, HSTATE_IN_GUEST(r13) cmpwi r9, KVM_GUEST_MODE_HOST_HV beq kvmppc_bad_host_intr #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE cmpwi r9, KVM_GUEST_MODE_GUEST - ld r9, HSTATE_HOST_R2(r13) + ld r9, HSTATE_SCRATCH2(r13) beq kvmppc_interrupt_pr #endif /* We're now back in the host but in guest MMU context */ @@ -782,7 +781,7 @@ kvmppc_interrupt_hv: std r6, VCPU_GPR(R6)(r9) std r7, VCPU_GPR(R7)(r9) std r8, VCPU_GPR(R8)(r9) - ld r0, HSTATE_HOST_R2(r13) + ld r0, HSTATE_SCRATCH2(r13) std r0, VCPU_GPR(R9)(r9) std r10, VCPU_GPR(R10)(r9) std r11, VCPU_GPR(R11)(r9) -- cgit v0.10.2 From df9059bb64023da9f27e56a94a3e2b8f4b6336a9 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Mon, 16 Dec 2013 13:31:46 +1100 Subject: KVM: PPC: Book3S HV: Don't drop low-order page address bits Commit caaa4c804fae ("KVM: PPC: Book3S HV: Fix physical address calculations") unfortunately resulted in some low-order address bits getting dropped in the case where the guest is creating a 4k HPTE and the host page size is 64k. By getting the low-order bits from hva rather than gpa we miss out on bits 12 - 15 in this case, since hva is at page granularity. This puts the missing bits back in. Reported-by: Alexey Kardashevskiy Signed-off-by: Paul Mackerras Signed-off-by: Alexander Graf diff --git a/arch/powerpc/kvm/book3s_hv_rm_mmu.c b/arch/powerpc/kvm/book3s_hv_rm_mmu.c index 1931aa3..8689e2e 100644 --- a/arch/powerpc/kvm/book3s_hv_rm_mmu.c +++ b/arch/powerpc/kvm/book3s_hv_rm_mmu.c @@ -240,6 +240,7 @@ long kvmppc_do_h_enter(struct kvm *kvm, unsigned long flags, is_io = hpte_cache_bits(pte_val(pte)); pa = pte_pfn(pte) << PAGE_SHIFT; pa |= hva & (pte_size - 1); + pa |= gpa & ~PAGE_MASK; } } -- cgit v0.10.2 From 939fd1e8d9deff206f12bd9d4e54aa7a4bd0ffd6 Mon Sep 17 00:00:00 2001 From: Charles Keepax Date: Wed, 18 Dec 2013 09:25:49 +0000 Subject: ASoC: wm_adsp: Add small delay while polling DSP RAM start Some devices are getting very close to the limit whilst polling the RAM start, this patch adds a small delay to this loop to give a longer startup timeout. Signed-off-by: Charles Keepax Signed-off-by: Mark Brown Cc: stable@vger.kernel.org diff --git a/sound/soc/codecs/wm_adsp.c b/sound/soc/codecs/wm_adsp.c index 46ec0e9..4fbcab6 100644 --- a/sound/soc/codecs/wm_adsp.c +++ b/sound/soc/codecs/wm_adsp.c @@ -1474,13 +1474,17 @@ static int wm_adsp2_ena(struct wm_adsp *dsp) return ret; /* Wait for the RAM to start, should be near instantaneous */ - count = 0; - do { + for (count = 0; count < 10; ++count) { ret = regmap_read(dsp->regmap, dsp->base + ADSP2_STATUS1, &val); if (ret != 0) return ret; - } while (!(val & ADSP2_RAM_RDY) && ++count < 10); + + if (val & ADSP2_RAM_RDY) + break; + + msleep(1); + } if (!(val & ADSP2_RAM_RDY)) { adsp_err(dsp, "Failed to start DSP RAM\n"); -- cgit v0.10.2 From f0199bc5e3a3ec13f9bc938556517ec430b36437 Mon Sep 17 00:00:00 2001 From: Bo Shen Date: Wed, 18 Dec 2013 11:26:23 +0800 Subject: ASoC: wm8904: fix DSP mode B configuration When wm8904 work in DSP mode B, we still need to configure it to work in DSP mode. Or else, it will work in Right Justified mode. Signed-off-by: Bo Shen Acked-by: Charles Keepax Signed-off-by: Mark Brown Cc: stable@vger.kernel.org diff --git a/sound/soc/codecs/wm8904.c b/sound/soc/codecs/wm8904.c index 3938fb1..53bbfac 100644 --- a/sound/soc/codecs/wm8904.c +++ b/sound/soc/codecs/wm8904.c @@ -1444,7 +1444,7 @@ static int wm8904_set_fmt(struct snd_soc_dai *dai, unsigned int fmt) switch (fmt & SND_SOC_DAIFMT_FORMAT_MASK) { case SND_SOC_DAIFMT_DSP_B: - aif1 |= WM8904_AIF_LRCLK_INV; + aif1 |= 0x3 | WM8904_AIF_LRCLK_INV; case SND_SOC_DAIFMT_DSP_A: aif1 |= 0x3; break; -- cgit v0.10.2 From 3a6c5d8ad0a9253aafb76df3577edcb68c09b939 Mon Sep 17 00:00:00 2001 From: Hui Wang Date: Wed, 18 Dec 2013 18:09:56 +0800 Subject: ALSA: hda - Add Dell headset detection quirk for one more laptop model On the Dell machines with codec whose Subsystem Id is 0x10280640, no external microphone can be detected when plugging a 3-ring headset. Using ALC255_FIXUP_DELL1_MIC_NO_PRESENCE can fix this problem. The codec (Vendor ID: 0x10ec0255) on the machine belongs to alc_269 family. BugLink: https://bugs.launchpad.net/bugs/1260303 Cc: David Henningsson Cc: stable@vger.kernel.org Signed-off-by: Hui Wang Signed-off-by: Takashi Iwai diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 5ab8e16..c564694 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -4256,6 +4256,7 @@ static const struct snd_pci_quirk alc269_fixup_tbl[] = { SND_PCI_QUIRK(0x1028, 0x0638, "Dell Inspiron 5439", ALC290_FIXUP_MONO_SPEAKERS), SND_PCI_QUIRK(0x1028, 0x063e, "Dell", ALC269_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x063f, "Dell", ALC255_FIXUP_DELL1_MIC_NO_PRESENCE), + SND_PCI_QUIRK(0x1028, 0x0640, "Dell", ALC255_FIXUP_DELL1_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x15cc, "Dell X5 Precision", ALC269_FIXUP_DELL2_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x1028, 0x15cd, "Dell X5 Precision", ALC269_FIXUP_DELL2_MIC_NO_PRESENCE), SND_PCI_QUIRK(0x103c, 0x1586, "HP", ALC269_FIXUP_HP_MUTE_LED_MIC2), -- cgit v0.10.2 From 0baf8f6a2ac86c2c40ed0cacab8ea3d17371a1bb Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Mon, 2 Dec 2013 18:01:30 +0000 Subject: dma: pl330: ensure DMA descriptors are zero-initialised I see the following splat with 3.13-rc1 when attempting to perform DMA: [ 253.004516] Alignment trap: not handling instruction e1902f9f at [] [ 253.004583] Unhandled fault: alignment exception (0x221) at 0xdfdfdfd7 [ 253.004646] Internal error: : 221 [#1] PREEMPT SMP ARM [ 253.004691] Modules linked in: dmatest(+) [last unloaded: dmatest] [ 253.004798] CPU: 0 PID: 671 Comm: kthreadd Not tainted 3.13.0-rc1+ #2 [ 253.004864] task: df9b0900 ti: df03e000 task.ti: df03e000 [ 253.004937] PC is at dmaengine_unmap_put+0x14/0x34 [ 253.005010] LR is at pl330_tasklet+0x3c8/0x550 [ 253.005087] pc : [] lr : [] psr: a00e0193 [ 253.005087] sp : df03fe48 ip : 00000000 fp : df03bf18 [ 253.005178] r10: bf00e108 r9 : 00000001 r8 : 00000000 [ 253.005245] r7 : df837040 r6 : dfb41800 r5 : df837048 r4 : df837000 [ 253.005316] r3 : dfdfdfcf r2 : dfb41f80 r1 : df837048 r0 : dfdfdfd7 [ 253.005384] Flags: NzCv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel [ 253.005459] Control: 30c5387d Table: 9fb9ba80 DAC: fffffffd [ 253.005520] Process kthreadd (pid: 671, stack limit = 0xdf03e248) This is due to desc->txd.unmap containing garbage (uninitialised memory). Rather than add another dummy initialisation to _init_desc, instead ensure that the descriptors are zero-initialised during allocation and remove the dummy, per-field initialisation. Cc: Andriy Shevchenko Acked-by: Jassi Brar Signed-off-by: Will Deacon Acked-by: Vinod Koul Signed-off-by: Dan Williams diff --git a/drivers/dma/pl330.c b/drivers/dma/pl330.c index cdf0483..536632f 100644 --- a/drivers/dma/pl330.c +++ b/drivers/dma/pl330.c @@ -2492,12 +2492,9 @@ static dma_cookie_t pl330_tx_submit(struct dma_async_tx_descriptor *tx) static inline void _init_desc(struct dma_pl330_desc *desc) { - desc->pchan = NULL; desc->req.x = &desc->px; desc->req.token = desc; desc->rqcfg.swap = SWAP_NO; - desc->rqcfg.privileged = 0; - desc->rqcfg.insnaccess = 0; desc->rqcfg.scctl = SCCTRL0; desc->rqcfg.dcctl = DCCTRL0; desc->req.cfg = &desc->rqcfg; @@ -2517,7 +2514,7 @@ static int add_desc(struct dma_pl330_dmac *pdmac, gfp_t flg, int count) if (!pdmac) return 0; - desc = kmalloc(count * sizeof(*desc), flg); + desc = kcalloc(count, sizeof(*desc), flg); if (!desc) return 0; -- cgit v0.10.2 From 77873803363c9e831fc1d1e6895c084279090c22 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Tue, 17 Dec 2013 10:09:32 -0800 Subject: net_dma: mark broken net_dma can cause data to be copied to a stale mapping if a copy-on-write fault occurs during dma. The application sees missing data. The following trace is triggered by modifying the kernel to WARN if it ever triggers copy-on-write on a page that is undergoing dma: WARNING: CPU: 24 PID: 2529 at lib/dma-debug.c:485 debug_dma_assert_idle+0xd2/0x120() ioatdma 0000:00:04.0: DMA-API: cpu touching an active dma mapped page [pfn=0x16bcd9] Modules linked in: iTCO_wdt iTCO_vendor_support ioatdma lpc_ich pcspkr dca CPU: 24 PID: 2529 Comm: linbug Tainted: G W 3.13.0-rc1+ #353 00000000000001e5 ffff88016f45f688 ffffffff81751041 ffff88017ab0ef70 ffff88016f45f6d8 ffff88016f45f6c8 ffffffff8104ed9c ffffffff810f3646 ffff8801768f4840 0000000000000282 ffff88016f6cca10 00007fa2bb699349 Call Trace: [] dump_stack+0x46/0x58 [] warn_slowpath_common+0x8c/0xc0 [] ? ftrace_pid_func+0x26/0x30 [] warn_slowpath_fmt+0x46/0x50 [] debug_dma_assert_idle+0xd2/0x120 [] do_wp_page+0xd0/0x790 [] handle_mm_fault+0x51c/0xde0 [] ? copy_user_enhanced_fast_string+0x9/0x20 [] __do_page_fault+0x19c/0x530 [] ? _raw_spin_lock_bh+0x16/0x40 [] ? trace_clock_local+0x9/0x10 [] ? rb_reserve_next_event+0x64/0x310 [] ? ioat2_dma_prep_memcpy_lock+0x60/0x130 [ioatdma] [] do_page_fault+0xe/0x10 [] page_fault+0x22/0x30 [] ? __kfree_skb+0x51/0xd0 [] ? copy_user_enhanced_fast_string+0x9/0x20 [] ? memcpy_toiovec+0x52/0xa0 [] skb_copy_datagram_iovec+0x5f/0x2a0 [] tcp_rcv_established+0x674/0x7f0 [] tcp_v4_do_rcv+0x2e5/0x4a0 [..] ---[ end trace e30e3b01191b7617 ]--- Mapped at: [] debug_dma_map_page+0xb9/0x160 [] dma_async_memcpy_pg_to_pg+0x127/0x210 [] dma_memcpy_pg_to_iovec+0x119/0x1f0 [] dma_skb_copy_datagram_iovec+0x11c/0x2b0 [] tcp_rcv_established+0x74a/0x7f0: ...the problem is that the receive path falls back to cpu-copy in several locations and this trace is just one of the areas. A few options were considered to fix this: 1/ sync all dma whenever a cpu copy branch is taken 2/ modify the page fault handler to hold off while dma is in-flight Option 1 adds yet more cpu overhead to an "offload" that struggles to compete with cpu-copy. Option 2 adds checks for behavior that is already documented as broken when using get_user_pages(). At a minimum a debug mode is warranted to catch and flag these violations of the dma-api vs get_user_pages(). Thanks to David for his reproducer. Cc: Cc: Dave Jiang Cc: Vinod Koul Cc: Alexander Duyck Reported-by: David Whipple Acked-by: David S. Miller Signed-off-by: Dan Williams diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig index 132a4fd..c823daa 100644 --- a/drivers/dma/Kconfig +++ b/drivers/dma/Kconfig @@ -355,6 +355,7 @@ config NET_DMA bool "Network: TCP receive copy offload" depends on DMA_ENGINE && NET default (INTEL_IOATDMA || FSL_DMA) + depends on BROKEN help This enables the use of DMA engines in the network stack to offload receive copy-to-user operations, freeing CPU cycles. -- cgit v0.10.2 From c97102ba96324da330078ad8619ba4dfe840dbe3 Mon Sep 17 00:00:00 2001 From: Vivek Goyal Date: Wed, 18 Dec 2013 17:08:31 -0800 Subject: kexec: migrate to reboot cpu Commit 1b3a5d02ee07 ("reboot: move arch/x86 reboot= handling to generic kernel") moved reboot= handling to generic code. In the process it also removed the code in native_machine_shutdown() which are moving reboot process to reboot_cpu/cpu0. I guess that thought must have been that all reboot paths are calling migrate_to_reboot_cpu(), so we don't need this special handling. But kexec reboot path (kernel_kexec()) is not calling migrate_to_reboot_cpu() so above change broke kexec. Now reboot can happen on non-boot cpu and when INIT is sent in second kerneo to bring up BP, it brings down the machine. So start calling migrate_to_reboot_cpu() in kexec reboot path to avoid this problem. Bisected by WANG Chao. Reported-by: Matthew Whitehead Reported-by: Dave Young Signed-off-by: Vivek Goyal Tested-by: Baoquan He Tested-by: WANG Chao Acked-by: H. Peter Anvin Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/include/linux/reboot.h b/include/linux/reboot.h index 8e00f9f..9e7db9e 100644 --- a/include/linux/reboot.h +++ b/include/linux/reboot.h @@ -43,6 +43,7 @@ extern int unregister_reboot_notifier(struct notifier_block *); * Architecture-specific implementations of sys_reboot commands. */ +extern void migrate_to_reboot_cpu(void); extern void machine_restart(char *cmd); extern void machine_halt(void); extern void machine_power_off(void); diff --git a/kernel/kexec.c b/kernel/kexec.c index d0d8fca..9c97016 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -1680,6 +1680,7 @@ int kernel_kexec(void) { kexec_in_progress = true; kernel_restart_prepare(NULL); + migrate_to_reboot_cpu(); printk(KERN_EMERG "Starting new kernel\n"); machine_shutdown(); } diff --git a/kernel/reboot.c b/kernel/reboot.c index f813b34..662c83f 100644 --- a/kernel/reboot.c +++ b/kernel/reboot.c @@ -104,7 +104,7 @@ int unregister_reboot_notifier(struct notifier_block *nb) } EXPORT_SYMBOL(unregister_reboot_notifier); -static void migrate_to_reboot_cpu(void) +void migrate_to_reboot_cpu(void) { /* The boot cpu is always logical cpu 0 */ int cpu = reboot_cpu; -- cgit v0.10.2 From 2b4847e73004c10ae6666c2e27b5c5430aed8698 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:32 -0800 Subject: mm: numa: serialise parallel get_user_page against THP migration Base pages are unmapped and flushed from cache and TLB during normal page migration and replaced with a migration entry that causes any parallel NUMA hinting fault or gup to block until migration completes. THP does not unmap pages due to a lack of support for migration entries at a PMD level. This allows races with get_user_pages and get_user_pages_fast which commit 3f926ab945b6 ("mm: Close races between THP migration and PMD numa clearing") made worse by introducing a pmd_clear_flush(). This patch forces get_user_page (fast and normal) on a pmd_numa page to go through the slow get_user_page path where it will serialise against THP migration and properly account for the NUMA hinting fault. On the migration side the page table lock is taken for each PTE update. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/arch/x86/mm/gup.c b/arch/x86/mm/gup.c index dd74e46..0596e8e 100644 --- a/arch/x86/mm/gup.c +++ b/arch/x86/mm/gup.c @@ -83,6 +83,12 @@ static noinline int gup_pte_range(pmd_t pmd, unsigned long addr, pte_t pte = gup_get_pte(ptep); struct page *page; + /* Similar to the PMD case, NUMA hinting must take slow path */ + if (pte_numa(pte)) { + pte_unmap(ptep); + return 0; + } + if ((pte_flags(pte) & (mask | _PAGE_SPECIAL)) != mask) { pte_unmap(ptep); return 0; @@ -167,6 +173,13 @@ static int gup_pmd_range(pud_t pud, unsigned long addr, unsigned long end, if (pmd_none(pmd) || pmd_trans_splitting(pmd)) return 0; if (unlikely(pmd_large(pmd))) { + /* + * NUMA hinting faults need to be handled in the GUP + * slowpath for accounting purposes and so that they + * can be serialised against THP migration. + */ + if (pmd_numa(pmd)) + return 0; if (!gup_huge_pmd(pmd, addr, next, write, pages, nr)) return 0; } else { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 33a5dc4..51f0693 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1243,6 +1243,10 @@ struct page *follow_trans_huge_pmd(struct vm_area_struct *vma, if ((flags & FOLL_DUMP) && is_huge_zero_pmd(*pmd)) return ERR_PTR(-EFAULT); + /* Full NUMA hinting faults to serialise migration in fault paths */ + if ((flags & FOLL_NUMA) && pmd_numa(*pmd)) + goto out; + page = pmd_page(*pmd); VM_BUG_ON(!PageHead(page)); if (flags & FOLL_TOUCH) { @@ -1323,23 +1327,27 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, /* If the page was locked, there are no parallel migrations */ if (page_locked) goto clear_pmdnuma; + } - /* - * Otherwise wait for potential migrations and retry. We do - * relock and check_same as the page may no longer be mapped. - * As the fault is being retried, do not account for it. - */ + /* + * If there are potential migrations, wait for completion and retry. We + * do not relock and check_same as the page may no longer be mapped. + * Furtermore, even if the page is currently misplaced, there is no + * guarantee it is still misplaced after the migration completes. + */ + if (!page_locked) { spin_unlock(ptl); wait_on_page_locked(page); page_nid = -1; goto out; } - /* Page is misplaced, serialise migrations and parallel THP splits */ + /* + * Page is misplaced. Page lock serialises migrations. Acquire anon_vma + * to serialises splits + */ get_page(page); spin_unlock(ptl); - if (!page_locked) - lock_page(page); anon_vma = page_lock_anon_vma_read(page); /* Confirm the PMD did not change while page_table_lock was released */ diff --git a/mm/migrate.c b/mm/migrate.c index bb94004..2cabbd5 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1722,6 +1722,7 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, struct page *new_page = NULL; struct mem_cgroup *memcg = NULL; int page_lru = page_is_file_cache(page); + pmd_t orig_entry; /* * Rate-limit the amount of data that is being migrated to a node. @@ -1756,7 +1757,8 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, /* Recheck the target PMD */ ptl = pmd_lock(mm, pmd); - if (unlikely(!pmd_same(*pmd, entry))) { + if (unlikely(!pmd_same(*pmd, entry) || page_count(page) != 2)) { +fail_putback: spin_unlock(ptl); /* Reverse changes made by migrate_page_copy() */ @@ -1786,16 +1788,34 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, */ mem_cgroup_prepare_migration(page, new_page, &memcg); + orig_entry = *pmd; entry = mk_pmd(new_page, vma->vm_page_prot); - entry = pmd_mknonnuma(entry); - entry = maybe_pmd_mkwrite(pmd_mkdirty(entry), vma); entry = pmd_mkhuge(entry); + entry = maybe_pmd_mkwrite(pmd_mkdirty(entry), vma); + /* + * Clear the old entry under pagetable lock and establish the new PTE. + * Any parallel GUP will either observe the old page blocking on the + * page lock, block on the page table lock or observe the new page. + * The SetPageUptodate on the new page and page_add_new_anon_rmap + * guarantee the copy is visible before the pagetable update. + */ + flush_cache_range(vma, haddr, haddr + HPAGE_PMD_SIZE); + page_add_new_anon_rmap(new_page, vma, haddr); pmdp_clear_flush(vma, haddr, pmd); set_pmd_at(mm, haddr, pmd, entry); - page_add_new_anon_rmap(new_page, vma, haddr); update_mmu_cache_pmd(vma, address, &entry); + + if (page_count(page) != 2) { + set_pmd_at(mm, haddr, pmd, orig_entry); + flush_tlb_range(vma, haddr, haddr + HPAGE_PMD_SIZE); + update_mmu_cache_pmd(vma, address, &entry); + page_remove_rmap(new_page); + goto fail_putback; + } + page_remove_rmap(page); + /* * Finish the charge transaction under the page table lock to * prevent split_huge_page() from dividing up the charge @@ -1820,9 +1840,13 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, out_fail: count_vm_events(PGMIGRATE_FAIL, HPAGE_PMD_NR); out_dropref: - entry = pmd_mknonnuma(entry); - set_pmd_at(mm, haddr, pmd, entry); - update_mmu_cache_pmd(vma, address, &entry); + ptl = pmd_lock(mm, pmd); + if (pmd_same(*pmd, entry)) { + entry = pmd_mknonnuma(entry); + set_pmd_at(mm, haddr, pmd, entry); + update_mmu_cache_pmd(vma, address, &entry); + } + spin_unlock(ptl); unlock_page(page); put_page(page); -- cgit v0.10.2 From f714f4f20e59ea6eea264a86b9a51fd51b88fc54 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:33 -0800 Subject: mm: numa: call MMU notifiers on THP migration MMU notifiers must be called on THP page migration or secondary MMUs will get very confused. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/migrate.c b/mm/migrate.c index 2cabbd5..be787d5 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -36,6 +36,7 @@ #include #include #include +#include #include @@ -1716,12 +1717,13 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, struct page *page, int node) { spinlock_t *ptl; - unsigned long haddr = address & HPAGE_PMD_MASK; pg_data_t *pgdat = NODE_DATA(node); int isolated = 0; struct page *new_page = NULL; struct mem_cgroup *memcg = NULL; int page_lru = page_is_file_cache(page); + unsigned long mmun_start = address & HPAGE_PMD_MASK; + unsigned long mmun_end = mmun_start + HPAGE_PMD_SIZE; pmd_t orig_entry; /* @@ -1756,10 +1758,12 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, WARN_ON(PageLRU(new_page)); /* Recheck the target PMD */ + mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end); ptl = pmd_lock(mm, pmd); if (unlikely(!pmd_same(*pmd, entry) || page_count(page) != 2)) { fail_putback: spin_unlock(ptl); + mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end); /* Reverse changes made by migrate_page_copy() */ if (TestClearPageActive(new_page)) @@ -1800,15 +1804,16 @@ fail_putback: * The SetPageUptodate on the new page and page_add_new_anon_rmap * guarantee the copy is visible before the pagetable update. */ - flush_cache_range(vma, haddr, haddr + HPAGE_PMD_SIZE); - page_add_new_anon_rmap(new_page, vma, haddr); - pmdp_clear_flush(vma, haddr, pmd); - set_pmd_at(mm, haddr, pmd, entry); + flush_cache_range(vma, mmun_start, mmun_end); + page_add_new_anon_rmap(new_page, vma, mmun_start); + pmdp_clear_flush(vma, mmun_start, pmd); + set_pmd_at(mm, mmun_start, pmd, entry); + flush_tlb_range(vma, mmun_start, mmun_end); update_mmu_cache_pmd(vma, address, &entry); if (page_count(page) != 2) { - set_pmd_at(mm, haddr, pmd, orig_entry); - flush_tlb_range(vma, haddr, haddr + HPAGE_PMD_SIZE); + set_pmd_at(mm, mmun_start, pmd, orig_entry); + flush_tlb_range(vma, mmun_start, mmun_end); update_mmu_cache_pmd(vma, address, &entry); page_remove_rmap(new_page); goto fail_putback; @@ -1823,6 +1828,7 @@ fail_putback: */ mem_cgroup_end_migration(memcg, page, new_page, true); spin_unlock(ptl); + mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end); unlock_page(new_page); unlock_page(page); @@ -1843,7 +1849,7 @@ out_dropref: ptl = pmd_lock(mm, pmd); if (pmd_same(*pmd, entry)) { entry = pmd_mknonnuma(entry); - set_pmd_at(mm, haddr, pmd, entry); + set_pmd_at(mm, mmun_start, pmd, entry); update_mmu_cache_pmd(vma, address, &entry); } spin_unlock(ptl); -- cgit v0.10.2 From 67f87463d3a3362424efcbe8b40e4772fd34fc61 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:34 -0800 Subject: mm: clear pmd_numa before invalidating On x86, PMD entries are similar to _PAGE_PROTNONE protection and are handled as NUMA hinting faults. The following two page table protection bits are what defines them _PAGE_NUMA:set _PAGE_PRESENT:clear A PMD is considered present if any of the _PAGE_PRESENT, _PAGE_PROTNONE, _PAGE_PSE or _PAGE_NUMA bits are set. If pmdp_invalidate encounters a pmd_numa, it clears the present bit leaving _PAGE_NUMA which will be considered not present by the CPU but present by pmd_present. The existing caller of pmdp_invalidate should handle it but it's an inconsistent state for a PMD. This patch keeps the state consistent when calling pmdp_invalidate. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/pgtable-generic.c b/mm/pgtable-generic.c index cbb3854..e84cad2 100644 --- a/mm/pgtable-generic.c +++ b/mm/pgtable-generic.c @@ -191,6 +191,9 @@ pgtable_t pgtable_trans_huge_withdraw(struct mm_struct *mm, pmd_t *pmdp) void pmdp_invalidate(struct vm_area_struct *vma, unsigned long address, pmd_t *pmdp) { + pmd_t entry = *pmdp; + if (pmd_numa(entry)) + entry = pmd_mknonnuma(entry); set_pmd_at(vma->vm_mm, address, pmdp, pmd_mknotpresent(*pmdp)); flush_tlb_range(vma, address, address + HPAGE_PMD_SIZE); } -- cgit v0.10.2 From 5a6dac3ec5f583cc8ee7bc53b5500a207c4ca433 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:36 -0800 Subject: mm: numa: do not clear PMD during PTE update scan If the PMD is flushed then a parallel fault in handle_mm_fault() will enter the pmd_none and do_huge_pmd_anonymous_page() path where it'll attempt to insert a huge zero page. This is wasteful so the patch avoids clearing the PMD when setting pmd_numa. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 51f0693..420826e 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1539,7 +1539,7 @@ int change_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, */ if (!is_huge_zero_page(page) && !pmd_numa(*pmd)) { - entry = pmdp_get_and_clear(mm, addr, pmd); + entry = *pmd; entry = pmd_mknuma(entry); ret = HPAGE_PMD_NR; } -- cgit v0.10.2 From 0c5f83c23ca703d32f930393825487257a5cde6d Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:37 -0800 Subject: mm: numa: do not clear PTE for pte_numa update The TLB must be flushed if the PTE is updated but change_pte_range is clearing the PTE while marking PTEs pte_numa without necessarily flushing the TLB if it reinserts the same entry. Without the flush, it's conceivable that two processors have different TLBs for the same virtual address and at the very least it would generate spurious faults. This patch only unmaps the pages in change_pte_range for a full protection change. [riel@redhat.com: write pte_numa pte back to the page tables] Signed-off-by: Mel Gorman Signed-off-by: Rik van Riel Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Chegu Vinod Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/mprotect.c b/mm/mprotect.c index 2666797..1291a05 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -52,17 +52,19 @@ static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd, pte_t ptent; bool updated = false; - ptent = ptep_modify_prot_start(mm, addr, pte); if (!prot_numa) { + ptent = ptep_modify_prot_start(mm, addr, pte); ptent = pte_modify(ptent, newprot); updated = true; } else { struct page *page; + ptent = *pte; page = vm_normal_page(vma, addr, oldpte); if (page) { if (!pte_numa(oldpte)) { ptent = pte_mknuma(ptent); + set_pte_at(mm, addr, pte, ptent); updated = true; } } @@ -79,7 +81,10 @@ static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd, if (updated) pages++; - ptep_modify_prot_commit(mm, addr, pte, ptent); + + /* Only !prot_numa always clears the pte */ + if (!prot_numa) + ptep_modify_prot_commit(mm, addr, pte, ptent); } else if (IS_ENABLED(CONFIG_MIGRATION) && !pte_file(oldpte)) { swp_entry_t entry = pte_to_swp_entry(oldpte); -- cgit v0.10.2 From c3a489cac38d43ea6dc4ac240473b44b46deecf7 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:38 -0800 Subject: mm: numa: ensure anon_vma is locked to prevent parallel THP splits The anon_vma lock prevents parallel THP splits and any associated complexity that arises when handling splits during THP migration. This patch checks if the lock was successfully acquired and bails from THP migration if it failed for any reason. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 420826e..dbafffa 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1359,6 +1359,13 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, goto out_unlock; } + /* Bail if we fail to protect against THP splits for any reason */ + if (unlikely(!anon_vma)) { + put_page(page); + page_nid = -1; + goto clear_pmdnuma; + } + /* * Migrate the THP to the requested node, returns with page unlocked * and pmd_numa cleared. -- cgit v0.10.2 From eb4489f69f224356193364dc2762aa009738ca7f Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:39 -0800 Subject: mm: numa: avoid unnecessary work on the failure path If a PMD changes during a THP migration then migration aborts but the failure path is doing more work than is necessary. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/migrate.c b/mm/migrate.c index be787d5..a987525 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1780,7 +1780,8 @@ fail_putback: putback_lru_page(page); mod_zone_page_state(page_zone(page), NR_ISOLATED_ANON + page_lru, -HPAGE_PMD_NR); - goto out_fail; + + goto out_unlock; } /* @@ -1854,6 +1855,7 @@ out_dropref: } spin_unlock(ptl); +out_unlock: unlock_page(page); put_page(page); return 0; -- cgit v0.10.2 From 3c67f474558748b604e247d92b55dfe89654c81d Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:40 -0800 Subject: sched: numa: skip inaccessible VMAs Inaccessible VMA should not be trapping NUMA hint faults. Skip them. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 9030da7..c7395d9 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1738,6 +1738,13 @@ void task_numa_work(struct callback_head *work) (vma->vm_file && (vma->vm_flags & (VM_READ|VM_WRITE)) == (VM_READ))) continue; + /* + * Skip inaccessible VMAs to avoid any confusion between + * PROT_NONE and NUMA hinting ptes + */ + if (!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE))) + continue; + do { start = max(start, vma->vm_start); end = ALIGN(start + (pages << PAGE_SHIFT), HPAGE_SIZE); -- cgit v0.10.2 From 1667918b6483b12a6496bf54151b827b8235d7b1 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:41 -0800 Subject: mm: numa: clear numa hinting information on mprotect On a protection change it is no longer clear if the page should be still accessible. This patch clears the NUMA hinting fault bits on a protection change. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/huge_memory.c b/mm/huge_memory.c index dbafffa..70e7429 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1532,6 +1532,8 @@ int change_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, ret = 1; if (!prot_numa) { entry = pmdp_get_and_clear(mm, addr, pmd); + if (pmd_numa(entry)) + entry = pmd_mknonnuma(entry); entry = pmd_modify(entry, newprot); ret = HPAGE_PMD_NR; BUG_ON(pmd_write(entry)); diff --git a/mm/mprotect.c b/mm/mprotect.c index 1291a05..f842172 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -54,6 +54,8 @@ static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd, if (!prot_numa) { ptent = ptep_modify_prot_start(mm, addr, pte); + if (pte_numa(ptent)) + ptent = pte_mknonnuma(ptent); ptent = pte_modify(ptent, newprot); updated = true; } else { -- cgit v0.10.2 From de466bd628e8d663fdf3f791bc8db318ee85c714 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:42 -0800 Subject: mm: numa: avoid unnecessary disruption of NUMA hinting during migration do_huge_pmd_numa_page() handles the case where there is parallel THP migration. However, by the time it is checked the NUMA hinting information has already been disrupted. This patch adds an earlier check with some helpers. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/include/linux/migrate.h b/include/linux/migrate.h index f5096b5..b7717d7 100644 --- a/include/linux/migrate.h +++ b/include/linux/migrate.h @@ -90,10 +90,19 @@ static inline int migrate_huge_page_move_mapping(struct address_space *mapping, #endif /* CONFIG_MIGRATION */ #ifdef CONFIG_NUMA_BALANCING +extern bool pmd_trans_migrating(pmd_t pmd); +extern void wait_migrate_huge_page(struct anon_vma *anon_vma, pmd_t *pmd); extern int migrate_misplaced_page(struct page *page, struct vm_area_struct *vma, int node); extern bool migrate_ratelimited(int node); #else +static inline bool pmd_trans_migrating(pmd_t pmd) +{ + return false; +} +static inline void wait_migrate_huge_page(struct anon_vma *anon_vma, pmd_t *pmd) +{ +} static inline int migrate_misplaced_page(struct page *page, struct vm_area_struct *vma, int node) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 70e7429..7de1bf8 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -882,6 +882,10 @@ int copy_huge_pmd(struct mm_struct *dst_mm, struct mm_struct *src_mm, ret = 0; goto out_unlock; } + + /* mmap_sem prevents this happening but warn if that changes */ + WARN_ON(pmd_trans_migrating(pmd)); + if (unlikely(pmd_trans_splitting(pmd))) { /* split huge page running from under us */ spin_unlock(src_ptl); @@ -1299,6 +1303,17 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, if (unlikely(!pmd_same(pmd, *pmdp))) goto out_unlock; + /* + * If there are potential migrations, wait for completion and retry + * without disrupting NUMA hinting information. Do not relock and + * check_same as the page may no longer be mapped. + */ + if (unlikely(pmd_trans_migrating(*pmdp))) { + spin_unlock(ptl); + wait_migrate_huge_page(vma->anon_vma, pmdp); + goto out; + } + page = pmd_page(pmd); BUG_ON(is_huge_zero_page(page)); page_nid = page_to_nid(page); @@ -1329,12 +1344,7 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, goto clear_pmdnuma; } - /* - * If there are potential migrations, wait for completion and retry. We - * do not relock and check_same as the page may no longer be mapped. - * Furtermore, even if the page is currently misplaced, there is no - * guarantee it is still misplaced after the migration completes. - */ + /* Migration could have started since the pmd_trans_migrating check */ if (!page_locked) { spin_unlock(ptl); wait_on_page_locked(page); diff --git a/mm/migrate.c b/mm/migrate.c index a987525..cfb4190 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1655,6 +1655,18 @@ int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page) return 1; } +bool pmd_trans_migrating(pmd_t pmd) +{ + struct page *page = pmd_page(pmd); + return PageLocked(page); +} + +void wait_migrate_huge_page(struct anon_vma *anon_vma, pmd_t *pmd) +{ + struct page *page = pmd_page(*pmd); + wait_on_page_locked(page); +} + /* * Attempt to migrate a misplaced page to the specified destination * node. Caller is expected to have an elevated reference count on -- cgit v0.10.2 From 20841405940e7be0617612d521e206e4b6b325db Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Wed, 18 Dec 2013 17:08:44 -0800 Subject: mm: fix TLB flush race between migration, and change_protection_range There are a few subtle races, between change_protection_range (used by mprotect and change_prot_numa) on one side, and NUMA page migration and compaction on the other side. The basic race is that there is a time window between when the PTE gets made non-present (PROT_NONE or NUMA), and the TLB is flushed. During that time, a CPU may continue writing to the page. This is fine most of the time, however compaction or the NUMA migration code may come in, and migrate the page away. When that happens, the CPU may continue writing, through the cached translation, to what is no longer the current memory location of the process. This only affects x86, which has a somewhat optimistic pte_accessible. All other architectures appear to be safe, and will either always flush, or flush whenever there is a valid mapping, even with no permissions (SPARC). The basic race looks like this: CPU A CPU B CPU C load TLB entry make entry PTE/PMD_NUMA fault on entry read/write old page start migrating page change PTE/PMD to new page read/write old page [*] flush TLB reload TLB from new entry read/write new page lose data [*] the old page may belong to a new user at this point! The obvious fix is to flush remote TLB entries, by making sure that pte_accessible aware of the fact that PROT_NONE and PROT_NUMA memory may still be accessible if there is a TLB flush pending for the mm. This should fix both NUMA migration and compaction. [mgorman@suse.de: fix build] Signed-off-by: Rik van Riel Signed-off-by: Mel Gorman Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/arch/sparc/include/asm/pgtable_64.h b/arch/sparc/include/asm/pgtable_64.h index 8358dc1..0f9e945 100644 --- a/arch/sparc/include/asm/pgtable_64.h +++ b/arch/sparc/include/asm/pgtable_64.h @@ -619,7 +619,7 @@ static inline unsigned long pte_present(pte_t pte) } #define pte_accessible pte_accessible -static inline unsigned long pte_accessible(pte_t a) +static inline unsigned long pte_accessible(struct mm_struct *mm, pte_t a) { return pte_val(a) & _PAGE_VALID; } @@ -847,7 +847,7 @@ static inline void __set_pte_at(struct mm_struct *mm, unsigned long addr, * SUN4V NOTE: _PAGE_VALID is the same value in both the SUN4U * and SUN4V pte layout, so this inline test is fine. */ - if (likely(mm != &init_mm) && pte_accessible(orig)) + if (likely(mm != &init_mm) && pte_accessible(mm, orig)) tlb_batch_add(mm, addr, ptep, orig, fullmm); } diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 3d19994..bbc8b12 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -452,9 +452,16 @@ static inline int pte_present(pte_t a) } #define pte_accessible pte_accessible -static inline int pte_accessible(pte_t a) +static inline bool pte_accessible(struct mm_struct *mm, pte_t a) { - return pte_flags(a) & _PAGE_PRESENT; + if (pte_flags(a) & _PAGE_PRESENT) + return true; + + if ((pte_flags(a) & (_PAGE_PROTNONE | _PAGE_NUMA)) && + mm_tlb_flush_pending(mm)) + return true; + + return false; } static inline int pte_hidden(pte_t pte) diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h index f330d28..b12079a 100644 --- a/include/asm-generic/pgtable.h +++ b/include/asm-generic/pgtable.h @@ -217,7 +217,7 @@ static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b) #endif #ifndef pte_accessible -# define pte_accessible(pte) ((void)(pte),1) +# define pte_accessible(mm, pte) ((void)(pte), 1) #endif #ifndef flush_tlb_fix_spurious_fault diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index bd29941..e5c49c3 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -443,6 +443,14 @@ struct mm_struct { /* numa_scan_seq prevents two threads setting pte_numa */ int numa_scan_seq; #endif +#if defined(CONFIG_NUMA_BALANCING) || defined(CONFIG_COMPACTION) + /* + * An operation with batched TLB flushing is going on. Anything that + * can move process memory needs to flush the TLB when moving a + * PROT_NONE or PROT_NUMA mapped page. + */ + bool tlb_flush_pending; +#endif struct uprobes_state uprobes_state; }; @@ -459,4 +467,40 @@ static inline cpumask_t *mm_cpumask(struct mm_struct *mm) return mm->cpu_vm_mask_var; } +#if defined(CONFIG_NUMA_BALANCING) || defined(CONFIG_COMPACTION) +/* + * Memory barriers to keep this state in sync are graciously provided by + * the page table locks, outside of which no page table modifications happen. + * The barriers below prevent the compiler from re-ordering the instructions + * around the memory barriers that are already present in the code. + */ +static inline bool mm_tlb_flush_pending(struct mm_struct *mm) +{ + barrier(); + return mm->tlb_flush_pending; +} +static inline void set_tlb_flush_pending(struct mm_struct *mm) +{ + mm->tlb_flush_pending = true; + barrier(); +} +/* Clearing is done after a TLB flush, which also provides a barrier. */ +static inline void clear_tlb_flush_pending(struct mm_struct *mm) +{ + barrier(); + mm->tlb_flush_pending = false; +} +#else +static inline bool mm_tlb_flush_pending(struct mm_struct *mm) +{ + return false; +} +static inline void set_tlb_flush_pending(struct mm_struct *mm) +{ +} +static inline void clear_tlb_flush_pending(struct mm_struct *mm) +{ +} +#endif + #endif /* _LINUX_MM_TYPES_H */ diff --git a/kernel/fork.c b/kernel/fork.c index 728d5be..5721f0e 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -537,6 +537,7 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p) spin_lock_init(&mm->page_table_lock); mm_init_aio(mm); mm_init_owner(mm, p); + clear_tlb_flush_pending(mm); if (likely(!mm_alloc_pgd(mm))) { mm->def_flags = 0; diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 7de1bf8..3d2783e 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1377,6 +1377,13 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, } /* + * The page_table_lock above provides a memory barrier + * with change_protection_range. + */ + if (mm_tlb_flush_pending(mm)) + flush_tlb_range(vma, haddr, haddr + HPAGE_PMD_SIZE); + + /* * Migrate the THP to the requested node, returns with page unlocked * and pmd_numa cleared. */ diff --git a/mm/mprotect.c b/mm/mprotect.c index f842172..bb53a65 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -188,6 +188,7 @@ static unsigned long change_protection_range(struct vm_area_struct *vma, BUG_ON(addr >= end); pgd = pgd_offset(mm, addr); flush_cache_range(vma, addr, end); + set_tlb_flush_pending(mm); do { next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(pgd)) @@ -199,6 +200,7 @@ static unsigned long change_protection_range(struct vm_area_struct *vma, /* Only flush the TLB if we actually modified any entries: */ if (pages) flush_tlb_range(vma, start, end); + clear_tlb_flush_pending(mm); return pages; } diff --git a/mm/pgtable-generic.c b/mm/pgtable-generic.c index e84cad2..a8b9199 100644 --- a/mm/pgtable-generic.c +++ b/mm/pgtable-generic.c @@ -110,9 +110,10 @@ int pmdp_clear_flush_young(struct vm_area_struct *vma, pte_t ptep_clear_flush(struct vm_area_struct *vma, unsigned long address, pte_t *ptep) { + struct mm_struct *mm = (vma)->vm_mm; pte_t pte; - pte = ptep_get_and_clear((vma)->vm_mm, address, ptep); - if (pte_accessible(pte)) + pte = ptep_get_and_clear(mm, address, ptep); + if (pte_accessible(mm, pte)) flush_tlb_page(vma, address); return pte; } -- cgit v0.10.2 From af2c1401e6f9177483be4fad876d0073669df9df Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:45 -0800 Subject: mm: numa: guarantee that tlb_flush_pending updates are visible before page table updates According to documentation on barriers, stores issued before a LOCK can complete after the lock implying that it's possible tlb_flush_pending can be visible after a page table update. As per revised documentation, this patch adds a smp_mb__before_spinlock to guarantee the correct ordering. Signed-off-by: Mel Gorman Acked-by: Paul E. McKenney Reviewed-by: Rik van Riel Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index e5c49c3..ad0616f 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -482,7 +482,12 @@ static inline bool mm_tlb_flush_pending(struct mm_struct *mm) static inline void set_tlb_flush_pending(struct mm_struct *mm) { mm->tlb_flush_pending = true; - barrier(); + + /* + * Guarantee that the tlb_flush_pending store does not leak into the + * critical section updating the page tables + */ + smp_mb__before_spinlock(); } /* Clearing is done after a TLB flush, which also provides a barrier. */ static inline void clear_tlb_flush_pending(struct mm_struct *mm) -- cgit v0.10.2 From b0943d61b8fa420180f92f64ef67662b4f6cc493 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:46 -0800 Subject: mm: numa: defer TLB flush for THP migration as long as possible THP migration can fail for a variety of reasons. Avoid flushing the TLB to deal with THP migration races until the copy is ready to start. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 3d2783e..7de1bf8 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1377,13 +1377,6 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, } /* - * The page_table_lock above provides a memory barrier - * with change_protection_range. - */ - if (mm_tlb_flush_pending(mm)) - flush_tlb_range(vma, haddr, haddr + HPAGE_PMD_SIZE); - - /* * Migrate the THP to the requested node, returns with page unlocked * and pmd_numa cleared. */ diff --git a/mm/migrate.c b/mm/migrate.c index cfb4190..e9b7102 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1759,6 +1759,9 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, goto out_fail; } + if (mm_tlb_flush_pending(mm)) + flush_tlb_range(vma, mmun_start, mmun_end); + /* Prepare a page as a migration target */ __set_page_locked(new_page); SetPageSwapBacked(new_page); -- cgit v0.10.2 From 73f038b863dfe98acabc7c36c17342b84ad52e94 Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Wed, 18 Dec 2013 17:08:47 -0800 Subject: mm: page_alloc: exclude unreclaimable allocations from zone fairness policy Dave Hansen noted a regression in a microbenchmark that loops around open() and close() on an 8-node NUMA machine and bisected it down to commit 81c0a2bb515f ("mm: page_alloc: fair zone allocator policy"). That change forces the slab allocations of the file descriptor to spread out to all 8 nodes, causing remote references in the page allocator and slab. The round-robin policy is only there to provide fairness among memory allocations that are reclaimed involuntarily based on pressure in each zone. It does not make sense to apply it to unreclaimable kernel allocations that are freed manually, in this case instantly after the allocation, and incur the remote reference costs twice for no reason. Only round-robin allocations that are usually freed through page reclaim or slab shrinking. Bisected by Dave Hansen. Signed-off-by: Johannes Weiner Cc: Dave Hansen Cc: Mel Gorman Reviewed-by: Rik van Riel Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 580a5f0..f861d02 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1920,7 +1920,8 @@ zonelist_scan: * back to remote zones that do not partake in the * fairness round-robin cycle of this zonelist. */ - if (alloc_flags & ALLOC_WMARK_LOW) { + if ((alloc_flags & ALLOC_WMARK_LOW) && + (gfp_mask & GFP_MOVABLE_MASK)) { if (zone_page_state(zone, NR_ALLOC_BATCH) <= 0) continue; if (zone_reclaim_mode && -- cgit v0.10.2 From 84ed8a99058e61567f495cc43118344261641c5f Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Wed, 18 Dec 2013 17:08:48 -0800 Subject: sh: always link in helper functions extracted from libgcc E.g. landisk_defconfig, which has CONFIG_NTFS_FS=m: ERROR: "__ashrdi3" [fs/ntfs/ntfs.ko] undefined! For "lib-y", if no symbols in a compilation unit are referenced by other units, the compilation unit will not be included in vmlinux. This breaks modules that do reference those symbols. Use "obj-y" instead to fix this. http://kisskb.ellerman.id.au/kisskb/buildresult/8838077/ This doesn't fix all cases. There are others, e.g. udivsi3. This is also not limited to sh, many architectures handle this in the same way. A simple solution is to unconditionally include all helper functions. A more complex solution is to make the choice of "lib-y" or "obj-y" depend on CONFIG_MODULES: obj-$(CONFIG_MODULES) += ... lib-y($CONFIG_MODULES) += ... Signed-off-by: Geert Uytterhoeven Cc: Paul Mundt Tested-by: Nobuhiro Iwamatsu Reviewed-by: Nobuhiro Iwamatsu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/arch/sh/lib/Makefile b/arch/sh/lib/Makefile index 7b95f29..3baff31 100644 --- a/arch/sh/lib/Makefile +++ b/arch/sh/lib/Makefile @@ -6,7 +6,7 @@ lib-y = delay.o memmove.o memchr.o \ checksum.o strlen.o div64.o div64-generic.o # Extracted from libgcc -lib-y += movmem.o ashldi3.o ashrdi3.o lshrdi3.o \ +obj-y += movmem.o ashldi3.o ashrdi3.o lshrdi3.o \ ashlsi3.o ashrsi3.o ashiftrt.o lshrsi3.o \ udiv_qrnnd.o -- cgit v0.10.2 From a844f38671d45ee6d981cfe41c5c4c2334578d73 Mon Sep 17 00:00:00 2001 From: Sima Baymani Date: Wed, 18 Dec 2013 17:08:49 -0800 Subject: mm: add missing dependency in Kconfig Eliminate the following (rand)config warning by adding missing PROC_FS dependency: warning: (HWPOISON_INJECT && MEM_SOFT_DIRTY) selects PROC_PAGE_MONITOR which has unmet direct dependencies (PROC_FS && MMU) Signed-off-by: Sima Baymani Suggested-by: David Rientjes Acked-by: David Rientjes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/Kconfig b/mm/Kconfig index eb69f35..723bbe0 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -543,7 +543,7 @@ config ZSWAP config MEM_SOFT_DIRTY bool "Track memory changes" - depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY + depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS select PROC_PAGE_MONITOR help This option enables memory changes tracking by introducing a -- cgit v0.10.2 From b0e5fd7359f1ce8db4ccb862b3aa80d2f2cbf4d0 Mon Sep 17 00:00:00 2001 From: Joonsoo Kim Date: Wed, 18 Dec 2013 17:08:51 -0800 Subject: mm/mempolicy: correct putback method for isolate pages if failed queue_pages_range() isolates hugetlbfs pages and putback_lru_pages() can't handle these. We should change it to putback_movable_pages(). Naoya said that it is worth going into stable, because it can break in-use hugepage list. Signed-off-by: Joonsoo Kim Acked-by: Rafael Aquini Reviewed-by: Naoya Horiguchi Reviewed-by: Wanpeng Li Cc: Christoph Lameter Cc: Vlastimil Babka Cc: Wanpeng Li Cc: Mel Gorman Cc: Rik van Riel Cc: Vlastimil Babka Cc: Zhang Yanfei Cc: [3.12.x] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/mempolicy.c b/mm/mempolicy.c index eca4a31..6d04d37 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -1318,7 +1318,7 @@ static long do_mbind(unsigned long start, unsigned long len, if (nr_failed && (flags & MPOL_MF_STRICT)) err = -EIO; } else - putback_lru_pages(&pagelist); + putback_movable_pages(&pagelist); up_write(&mm->mmap_sem); mpol_out: -- cgit v0.10.2 From 6815bf3f233e0b10c99a758497d5d236063b010b Mon Sep 17 00:00:00 2001 From: Joonsoo Kim Date: Wed, 18 Dec 2013 17:08:52 -0800 Subject: mm/compaction: respect ignore_skip_hint in update_pageblock_skip update_pageblock_skip() only fits to compaction which tries to isolate by pageblock unit. If isolate_migratepages_range() is called by CMA, it try to isolate regardless of pageblock unit and it don't reference get_pageblock_skip() by ignore_skip_hint. We should also respect it on update_pageblock_skip() to prevent from setting the wrong information. Signed-off-by: Joonsoo Kim Acked-by: Vlastimil Babka Reviewed-by: Naoya Horiguchi Reviewed-by: Wanpeng Li Cc: Christoph Lameter Cc: Rafael Aquini Cc: Vlastimil Babka Cc: Wanpeng Li Cc: Mel Gorman Cc: Rik van Riel Cc: Zhang Yanfei Cc: [3.7+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/compaction.c b/mm/compaction.c index 805165b..f58bcd0 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -134,6 +134,10 @@ static void update_pageblock_skip(struct compact_control *cc, bool migrate_scanner) { struct zone *zone = cc->zone; + + if (cc->ignore_skip_hint) + return; + if (!page) return; -- cgit v0.10.2 From a49ecbcd7b0d5a1cda7d60e03df402dd0ef76ac8 Mon Sep 17 00:00:00 2001 From: Jianguo Wu Date: Wed, 18 Dec 2013 17:08:54 -0800 Subject: mm/memory-failure.c: recheck PageHuge() after hugetlb page migrate successfully After a successful hugetlb page migration by soft offline, the source page will either be freed into hugepage_freelists or buddy(over-commit page). If page is in buddy, page_hstate(page) will be NULL. It will hit a NULL pointer dereference in dequeue_hwpoisoned_huge_page(). BUG: unable to handle kernel NULL pointer dereference at 0000000000000058 IP: [] dequeue_hwpoisoned_huge_page+0x131/0x1d0 PGD c23762067 PUD c24be2067 PMD 0 Oops: 0000 [#1] SMP So check PageHuge(page) after call migrate_pages() successfully. Signed-off-by: Jianguo Wu Tested-by: Naoya Horiguchi Reviewed-by: Naoya Horiguchi Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/memory-failure.c b/mm/memory-failure.c index b7c1716..db08af9 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -1505,10 +1505,16 @@ static int soft_offline_huge_page(struct page *page, int flags) if (ret > 0) ret = -EIO; } else { - set_page_hwpoison_huge_page(hpage); - dequeue_hwpoisoned_huge_page(hpage); - atomic_long_add(1 << compound_order(hpage), - &num_poisoned_pages); + /* overcommit hugetlb page will be freed to buddy */ + if (PageHuge(page)) { + set_page_hwpoison_huge_page(hpage); + dequeue_hwpoisoned_huge_page(hpage); + atomic_long_add(1 << compound_order(hpage), + &num_poisoned_pages); + } else { + SetPageHWPoison(page); + atomic_long_inc(&num_poisoned_pages); + } } return ret; } -- cgit v0.10.2 From 584ec9794293ff69f597f5bec23296e62e94e857 Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Wed, 18 Dec 2013 17:08:55 -0800 Subject: MAINTAINERS: add Davidlohr as GPT maintainer Add a new entry for the GPT standard. Any future changes will now be routed through linux-efi. Signed-off-by: Davidlohr Bueso Acked-by: Matt Fleming Cc: Jens Axboe Acked-by: Matt Domsch Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/MAINTAINERS b/MAINTAINERS index 75e3db9..3c99c25 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3833,6 +3833,12 @@ T: git git://linuxtv.org/media_tree.git S: Maintained F: drivers/media/usb/gspca/ +GUID PARTITION TABLE (GPT) +M: Davidlohr Bueso +L: linux-efi@vger.kernel.org +S: Maintained +F: block/partitions/efi.* + STK1160 USB VIDEO CAPTURE DRIVER M: Ezequiel Garcia L: linux-media@vger.kernel.org -- cgit v0.10.2 From 11c731e81bb0d8d2e835447a2dd645b34bb74706 Mon Sep 17 00:00:00 2001 From: Wanpeng Li Date: Wed, 18 Dec 2013 17:08:56 -0800 Subject: mm/mempolicy: fix !vma in new_vma_page() BUG_ON(!vma) assumption is introduced by commit 0bf598d863e3 ("mbind: add BUG_ON(!vma) in new_vma_page()"), however, even if address = __vma_address(page, vma); and vma->start < address < vma->end page_address_in_vma() may still return -EFAULT because of many other conditions in it. As a result the while loop in new_vma_page() may end with vma=NULL. This patch revert the commit and also fix the potential dereference NULL pointer reported by Dan. http://marc.info/?l=linux-mm&m=137689530323257&w=2 kernel BUG at mm/mempolicy.c:1204! invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC CPU: 3 PID: 7056 Comm: trinity-child3 Not tainted 3.13.0-rc3+ #2 task: ffff8801ca5295d0 ti: ffff88005ab20000 task.ti: ffff88005ab20000 RIP: new_vma_page+0x70/0x90 RSP: 0000:ffff88005ab21db0 EFLAGS: 00010246 RAX: fffffffffffffff2 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000008040075 RSI: ffff8801c3d74600 RDI: ffffea00079a8b80 RBP: ffff88005ab21dc8 R08: 0000000000000004 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000000 R12: fffffffffffffff2 R13: ffffea00079a8b80 R14: 0000000000400000 R15: 0000000000400000 FS: 00007ff49c6f4740(0000) GS:ffff880244e00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007ff49c68f994 CR3: 000000005a205000 CR4: 00000000001407e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Stack: ffffea00079a8b80 ffffea00079a8bc0 ffffea00079a8ba0 ffff88005ab21e50 ffffffff811adc7a 0000000000000000 ffff8801ca5295d0 0000000464e224f8 0000000000000000 0000000000000002 0000000000000000 ffff88020ce75c00 Call Trace: migrate_pages+0x12a/0x850 SYSC_mbind+0x513/0x6a0 SyS_mbind+0xe/0x10 ia32_do_call+0x13/0x13 Code: 85 c0 75 2f 4c 89 e1 48 89 da 31 f6 bf da 00 02 00 65 44 8b 04 25 08 f7 1c 00 e8 ec fd ff ff 5b 41 5c 41 5d 5d c3 0f 1f 44 00 00 <0f> 0b 66 0f 1f 44 00 00 4c 89 e6 48 89 df ba 01 00 00 00 e8 48 RIP [] new_vma_page+0x70/0x90 RSP Signed-off-by: Wanpeng Li Reported-by: Dave Jones Reported-by: Sasha Levin Reviewed-by: Naoya Horiguchi Reviewed-by: Bob Liu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/mempolicy.c b/mm/mempolicy.c index 6d04d37..0cd2c4d 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -1197,14 +1197,16 @@ static struct page *new_vma_page(struct page *page, unsigned long private, int * break; vma = vma->vm_next; } + + if (PageHuge(page)) { + if (vma) + return alloc_huge_page_noerr(vma, address, 1); + else + return NULL; + } /* - * queue_pages_range() confirms that @page belongs to some vma, - * so vma shouldn't be NULL. + * if !vma, alloc_page_vma() will use task or system default policy */ - BUG_ON(!vma); - - if (PageHuge(page)) - return alloc_huge_page_noerr(vma, address, 1); return alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address); } #else -- cgit v0.10.2 From 7ac181568342ec618d4da1ab03c2babcaa7ee84f Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Wed, 18 Dec 2013 17:08:57 -0800 Subject: fix build with make 3.80 According to Documentation/Changes, make 3.80 is still being supported for building the kernel, hence make files must not make (unconditional) use of features introduced only in newer versions. Commit 1bf49dd4be0b ("./Makefile: export initial ramdisk compression config option") however introduced "else ifeq" constructs which make 3.80 doesn't understand. Replace the logic there with more conventional (in the kernel build infrastructure) list constructs (except that the list here is intentionally limited to exactly one element). Signed-off-by: Jan Beulich Cc: P J P Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/Makefile b/Makefile index 858a147..89cee86 100644 --- a/Makefile +++ b/Makefile @@ -732,19 +732,13 @@ export mod_strip_cmd # Select initial ramdisk compression format, default is gzip(1). # This shall be used by the dracut(8) tool while creating an initramfs image. # -INITRD_COMPRESS=gzip -ifeq ($(CONFIG_RD_BZIP2), y) - INITRD_COMPRESS=bzip2 -else ifeq ($(CONFIG_RD_LZMA), y) - INITRD_COMPRESS=lzma -else ifeq ($(CONFIG_RD_XZ), y) - INITRD_COMPRESS=xz -else ifeq ($(CONFIG_RD_LZO), y) - INITRD_COMPRESS=lzo -else ifeq ($(CONFIG_RD_LZ4), y) - INITRD_COMPRESS=lz4 -endif -export INITRD_COMPRESS +INITRD_COMPRESS-y := gzip +INITRD_COMPRESS-$(CONFIG_RD_BZIP2) := bzip2 +INITRD_COMPRESS-$(CONFIG_RD_LZMA) := lzma +INITRD_COMPRESS-$(CONFIG_RD_XZ) := xz +INITRD_COMPRESS-$(CONFIG_RD_LZO) := lzo +INITRD_COMPRESS-$(CONFIG_RD_LZ4) := lz4 +export INITRD_COMPRESS := $(INITRD_COMPRESS-y) ifdef CONFIG_MODULE_SIG_ALL MODSECKEY = ./signing_key.priv -- cgit v0.10.2 From 98398c32f6687ee1e1f3ae084effb4b75adb0747 Mon Sep 17 00:00:00 2001 From: Jianguo Wu Date: Wed, 18 Dec 2013 17:08:59 -0800 Subject: mm/hugetlb: check for pte NULL pointer in __page_check_address() In __page_check_address(), if address's pud is not present, huge_pte_offset() will return NULL, we should check the return value. Signed-off-by: Jianguo Wu Cc: Naoya Horiguchi Cc: Mel Gorman Cc: qiuxishi Cc: Hanjun Guo Acked-by: Kirill A. Shutemov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds diff --git a/mm/rmap.c b/mm/rmap.c index 55c8b8d..068522d 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -600,7 +600,11 @@ pte_t *__page_check_address(struct page *page, struct mm_struct *mm, spinlock_t *ptl; if (unlikely(PageHuge(page))) { + /* when pud is not present, pte will be NULL */ pte = huge_pte_offset(mm, address); + if (!pte) + return NULL; + ptl = huge_pte_lockptr(page_hstate(page), mm, pte); goto check; } -- cgit v0.10.2 From db6077fd0b7dd41dc6ff18329cec979379071f87 Mon Sep 17 00:00:00 2001 From: Nicholas Bellinger Date: Wed, 11 Dec 2013 15:45:32 -0800 Subject: iscsi-target: Fix incorrect np->np_thread NULL assignment When shutting down a target there is a race condition between iscsit_del_np() and __iscsi_target_login_thread(). The latter sets the thread pointer to NULL, and the former tries to issue kthread_stop() on that pointer without any synchronization. This patch moves the np->np_thread NULL assignment into iscsit_del_np(), after kthread_stop() has completed. It also removes the signal_pending() + np_state check, and only exits when kthread_should_stop() is true. Reported-by: Hannes Reinecke Cc: #3.12+ Signed-off-by: Nicholas Bellinger diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c index 02182ab..0086719 100644 --- a/drivers/target/iscsi/iscsi_target.c +++ b/drivers/target/iscsi/iscsi_target.c @@ -465,6 +465,7 @@ int iscsit_del_np(struct iscsi_np *np) */ send_sig(SIGINT, np->np_thread, 1); kthread_stop(np->np_thread); + np->np_thread = NULL; } np->np_transport->iscsit_free_np(np); diff --git a/drivers/target/iscsi/iscsi_target_login.c b/drivers/target/iscsi/iscsi_target_login.c index 4eb93b2..e29279e 100644 --- a/drivers/target/iscsi/iscsi_target_login.c +++ b/drivers/target/iscsi/iscsi_target_login.c @@ -1403,11 +1403,6 @@ old_sess_out: out: stop = kthread_should_stop(); - if (!stop && signal_pending(current)) { - spin_lock_bh(&np->np_thread_lock); - stop = (np->np_thread_state == ISCSI_NP_THREAD_SHUTDOWN); - spin_unlock_bh(&np->np_thread_lock); - } /* Wait for another socket.. */ if (!stop) return 1; @@ -1415,7 +1410,6 @@ exit: iscsi_stop_login_thread_timer(np); spin_lock_bh(&np->np_thread_lock); np->np_thread_state = ISCSI_NP_THREAD_EXIT; - np->np_thread = NULL; spin_unlock_bh(&np->np_thread_lock); return 0; -- cgit v0.10.2 From 2853c2b6671509591be09213954d7249ca6ff224 Mon Sep 17 00:00:00 2001 From: Nicholas Bellinger Date: Wed, 11 Dec 2013 16:20:13 -0800 Subject: iser-target: Move INIT_WORK setup into isert_create_device_ib_res This patch moves INIT_WORK setup for cq_desc->cq_[rx,tx]_work into isert_create_device_ib_res(), instead of being done each callback invocation in isert_cq_[rx,tx]_callback(). This also fixes a 'INFO: trying to register non-static key' warning when cancel_work_sync() is called before INIT_WORK has setup the struct work_struct. Reported-by: Or Gerlitz Cc: #3.12+ Signed-off-by: Nicholas Bellinger diff --git a/drivers/infiniband/ulp/isert/ib_isert.c b/drivers/infiniband/ulp/isert/ib_isert.c index 78f6e92..9804fca 100644 --- a/drivers/infiniband/ulp/isert/ib_isert.c +++ b/drivers/infiniband/ulp/isert/ib_isert.c @@ -207,7 +207,9 @@ isert_free_rx_descriptors(struct isert_conn *isert_conn) isert_conn->conn_rx_descs = NULL; } +static void isert_cq_tx_work(struct work_struct *); static void isert_cq_tx_callback(struct ib_cq *, void *); +static void isert_cq_rx_work(struct work_struct *); static void isert_cq_rx_callback(struct ib_cq *, void *); static int @@ -259,6 +261,7 @@ isert_create_device_ib_res(struct isert_device *device) cq_desc[i].device = device; cq_desc[i].cq_index = i; + INIT_WORK(&cq_desc[i].cq_rx_work, isert_cq_rx_work); device->dev_rx_cq[i] = ib_create_cq(device->ib_device, isert_cq_rx_callback, isert_cq_event_callback, @@ -270,6 +273,7 @@ isert_create_device_ib_res(struct isert_device *device) goto out_cq; } + INIT_WORK(&cq_desc[i].cq_tx_work, isert_cq_tx_work); device->dev_tx_cq[i] = ib_create_cq(device->ib_device, isert_cq_tx_callback, isert_cq_event_callback, @@ -1732,7 +1736,6 @@ isert_cq_tx_callback(struct ib_cq *cq, void *context) { struct isert_cq_desc *cq_desc = (struct isert_cq_desc *)context; - INIT_WORK(&cq_desc->cq_tx_work, isert_cq_tx_work); queue_work(isert_comp_wq, &cq_desc->cq_tx_work); } @@ -1776,7 +1779,6 @@ isert_cq_rx_callback(struct ib_cq *cq, void *context) { struct isert_cq_desc *cq_desc = (struct isert_cq_desc *)context; - INIT_WORK(&cq_desc->cq_rx_work, isert_cq_rx_work); queue_work(isert_rx_wq, &cq_desc->cq_rx_work); } -- cgit v0.10.2 From 95cadace8f3959282e76ebf8b382bd0930807d2c Mon Sep 17 00:00:00 2001 From: Nicholas Bellinger Date: Thu, 12 Dec 2013 12:24:11 -0800 Subject: target/file: Update hw_max_sectors based on current block_size This patch allows FILEIO to update hw_max_sectors based on the current max_bytes_per_io. This is required because vfs_[writev,readv]() can accept a maximum of 2048 iovecs per call, so the enforced hw_max_sectors really needs to be calculated based on block_size. This addresses a >= v3.5 bug where block_size=512 was rejecting > 1M sized I/O requests, because FD_MAX_SECTORS was hardcoded to 2048 for the block_size=4096 case. (v2: Use max_bytes_per_io instead of ->update_hw_max_sectors) Reported-by: Henrik Goldman Cc: #3.5+ Signed-off-by: Nicholas Bellinger diff --git a/drivers/target/target_core_device.c b/drivers/target/target_core_device.c index 207b340..d06de84 100644 --- a/drivers/target/target_core_device.c +++ b/drivers/target/target_core_device.c @@ -1106,6 +1106,11 @@ int se_dev_set_block_size(struct se_device *dev, u32 block_size) dev->dev_attrib.block_size = block_size; pr_debug("dev[%p]: SE Device block_size changed to %u\n", dev, block_size); + + if (dev->dev_attrib.max_bytes_per_io) + dev->dev_attrib.hw_max_sectors = + dev->dev_attrib.max_bytes_per_io / block_size; + return 0; } diff --git a/drivers/target/target_core_file.c b/drivers/target/target_core_file.c index 0e34cda..78241a5 100644 --- a/drivers/target/target_core_file.c +++ b/drivers/target/target_core_file.c @@ -66,9 +66,8 @@ static int fd_attach_hba(struct se_hba *hba, u32 host_id) pr_debug("CORE_HBA[%d] - TCM FILEIO HBA Driver %s on Generic" " Target Core Stack %s\n", hba->hba_id, FD_VERSION, TARGET_CORE_MOD_VERSION); - pr_debug("CORE_HBA[%d] - Attached FILEIO HBA: %u to Generic" - " MaxSectors: %u\n", - hba->hba_id, fd_host->fd_host_id, FD_MAX_SECTORS); + pr_debug("CORE_HBA[%d] - Attached FILEIO HBA: %u to Generic\n", + hba->hba_id, fd_host->fd_host_id); return 0; } @@ -220,7 +219,8 @@ static int fd_configure_device(struct se_device *dev) } dev->dev_attrib.hw_block_size = fd_dev->fd_block_size; - dev->dev_attrib.hw_max_sectors = FD_MAX_SECTORS; + dev->dev_attrib.max_bytes_per_io = FD_MAX_BYTES; + dev->dev_attrib.hw_max_sectors = FD_MAX_BYTES / fd_dev->fd_block_size; dev->dev_attrib.hw_queue_depth = FD_MAX_DEVICE_QUEUE_DEPTH; if (fd_dev->fbd_flags & FDBD_HAS_BUFFERED_IO_WCE) { diff --git a/drivers/target/target_core_file.h b/drivers/target/target_core_file.h index 37ffc5b..d7772c1 100644 --- a/drivers/target/target_core_file.h +++ b/drivers/target/target_core_file.h @@ -7,7 +7,10 @@ #define FD_DEVICE_QUEUE_DEPTH 32 #define FD_MAX_DEVICE_QUEUE_DEPTH 128 #define FD_BLOCKSIZE 512 -#define FD_MAX_SECTORS 2048 +/* + * Limited by the number of iovecs (2048) per vfs_[writev,readv] call + */ +#define FD_MAX_BYTES 8388608 #define RRF_EMULATE_CDB 0x01 #define RRF_GOT_LBA 0x02 diff --git a/include/target/target_core_base.h b/include/target/target_core_base.h index 9f1dda6..321301c 100644 --- a/include/target/target_core_base.h +++ b/include/target/target_core_base.h @@ -620,6 +620,7 @@ struct se_dev_attrib { u32 unmap_granularity; u32 unmap_granularity_alignment; u32 max_write_same_len; + u32 max_bytes_per_io; struct se_device *da_dev; struct config_group da_group; }; -- cgit v0.10.2 From 1e01c7eb7c431a74437d73fe54670398b4d2b222 Mon Sep 17 00:00:00 2001 From: Vineet Gupta Date: Thu, 19 Dec 2013 18:55:58 +0530 Subject: ARC: Allow conditional multiple inclusion of uapi/asm/unistd.h Commit 97bc386fc12d "ARC: Add guard macro to uapi/asm/unistd.h" inhibited multiple inclusion of ARCH unistd.h. This however hosed the system since Generic syscall table generator relies on it being included twice, and in lack-of an empty table was emitted by C preprocessor. Fix that by allowing one exception to rule for the special case (just like Xtensa) Suggested-by: Chen Gang Signed-off-by: Vineet Gupta diff --git a/arch/arc/include/uapi/asm/unistd.h b/arch/arc/include/uapi/asm/unistd.h index 68125dd..39e58d1 100644 --- a/arch/arc/include/uapi/asm/unistd.h +++ b/arch/arc/include/uapi/asm/unistd.h @@ -8,7 +8,11 @@ /******** no-legacy-syscalls-ABI *******/ -#ifndef _UAPI_ASM_ARC_UNISTD_H +/* + * Non-typical guard macro to enable inclusion twice in ARCH sys.c + * That is how the Generic syscall wrapper generator works + */ +#if !defined(_UAPI_ASM_ARC_UNISTD_H) || defined(__SYSCALL) #define _UAPI_ASM_ARC_UNISTD_H #define __ARCH_WANT_SYS_EXECVE @@ -36,4 +40,6 @@ __SYSCALL(__NR_arc_gettls, sys_arc_gettls) #define __NR_sysfs (__NR_arch_specific_syscall + 3) __SYSCALL(__NR_sysfs, sys_sysfs) +#undef __SYSCALL + #endif -- cgit v0.10.2 From cdc27c27843248ae7eb0df5fc261dd004eaa5670 Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Tue, 17 Dec 2013 17:09:08 +0000 Subject: arm64: ptrace: avoid using HW_BREAKPOINT_EMPTY for disabled events Commit 8f34a1da35ae ("arm64: ptrace: use HW_BREAKPOINT_EMPTY type for disabled breakpoints") fixed an issue with GDB trying to zero breakpoint control registers. The problem there is that the arch hw_breakpoint code will attempt to create a (disabled), execute breakpoint of length 0. This will fail validation and report unexpected failure to GDB. To avoid this, we treated disabled breakpoints as HW_BREAKPOINT_EMPTY, but that seems to have broken with recent kernels, causing watchpoints to be treated as TYPE_INST in the core code and returning ENOSPC for any further breakpoints. This patch fixes the problem by prioritising the `enable' field of the breakpoint: if it is cleared, we simply update the perf_event_attr to indicate that the thing is disabled and don't bother changing either the type or the length. This reinforces the behaviour that the breakpoint control register is essentially read-only apart from the enable bit when disabling a breakpoint. Cc: Reported-by: Aaron Liu Signed-off-by: Will Deacon Signed-off-by: Catalin Marinas diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c index 6777a21..6a8928b 100644 --- a/arch/arm64/kernel/ptrace.c +++ b/arch/arm64/kernel/ptrace.c @@ -214,31 +214,29 @@ static int ptrace_hbp_fill_attr_ctrl(unsigned int note_type, { int err, len, type, disabled = !ctrl.enabled; - if (disabled) { - len = 0; - type = HW_BREAKPOINT_EMPTY; - } else { - err = arch_bp_generic_fields(ctrl, &len, &type); - if (err) - return err; - - switch (note_type) { - case NT_ARM_HW_BREAK: - if ((type & HW_BREAKPOINT_X) != type) - return -EINVAL; - break; - case NT_ARM_HW_WATCH: - if ((type & HW_BREAKPOINT_RW) != type) - return -EINVAL; - break; - default: + attr->disabled = disabled; + if (disabled) + return 0; + + err = arch_bp_generic_fields(ctrl, &len, &type); + if (err) + return err; + + switch (note_type) { + case NT_ARM_HW_BREAK: + if ((type & HW_BREAKPOINT_X) != type) return -EINVAL; - } + break; + case NT_ARM_HW_WATCH: + if ((type & HW_BREAKPOINT_RW) != type) + return -EINVAL; + break; + default: + return -EINVAL; } attr->bp_len = len; attr->bp_type = type; - attr->disabled = disabled; return 0; } -- cgit v0.10.2 From de06875f089678f4f9f1e8d5e1421fb0ceab12d0 Mon Sep 17 00:00:00 2001 From: Andy Grover Date: Tue, 26 Nov 2013 11:49:24 -0800 Subject: target: Remove extra percpu_ref_init lun->lun_ref is also initialized in core_tpg_post_addlun, so it doesn't need to be done in core_tpg_setup_virtual_lun0. (nab: Drop left-over percpu_ref_cancel_init in failure path) Signed-off-by: Andy Grover Signed-off-by: Nicholas Bellinger diff --git a/drivers/target/target_core_tpg.c b/drivers/target/target_core_tpg.c index f755712..2a573de 100644 --- a/drivers/target/target_core_tpg.c +++ b/drivers/target/target_core_tpg.c @@ -656,15 +656,9 @@ static int core_tpg_setup_virtual_lun0(struct se_portal_group *se_tpg) spin_lock_init(&lun->lun_sep_lock); init_completion(&lun->lun_ref_comp); - ret = percpu_ref_init(&lun->lun_ref, core_tpg_lun_ref_release); - if (ret < 0) - return ret; - ret = core_tpg_post_addlun(se_tpg, lun, lun_access, dev); - if (ret < 0) { - percpu_ref_cancel_init(&lun->lun_ref); + if (ret < 0) return ret; - } return 0; } -- cgit v0.10.2 From dcd211997ddba3229e875f5990ea14f920934729 Mon Sep 17 00:00:00 2001 From: Nicholas Bellinger Date: Tue, 17 Dec 2013 01:51:22 -0800 Subject: qla2xxx: Fix scsi_host leak on qlt_lport_register callback failure This patch fixes a possible scsi_host reference leak in qlt_lport_register(), when a non zero return from the passed (*callback) does not call drop the local reference via scsi_host_put() before returning. This currently does not effect existing tcm_qla2xxx code as the passed callback will never fail, but fix this up regardless for future code. Cc: Chad Dupuis Signed-off-by: Nicholas Bellinger diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c index 3bb0a1d..38a1257 100644 --- a/drivers/scsi/qla2xxx/qla_target.c +++ b/drivers/scsi/qla2xxx/qla_target.c @@ -4291,6 +4291,7 @@ int qlt_lport_register(struct qla_tgt_func_tmpl *qla_tgt_ops, u64 wwpn, if (rc != 0) { ha->tgt.tgt_ops = NULL; ha->tgt.target_lport_ptr = NULL; + scsi_host_put(host); } mutex_unlock(&qla_tgt_mutex); return rc; -- cgit v0.10.2 From ee53664bda169f519ce3c6a22d378f0b946c8178 Mon Sep 17 00:00:00 2001 From: "Kirill A. Shutemov" Date: Fri, 20 Dec 2013 15:10:03 +0200 Subject: mm: Fix NULL pointer dereference in madvise(MADV_WILLNEED) support Sasha Levin found a NULL pointer dereference that is due to a missing page table lock, which in turn is due to the pmd entry in question being a transparent huge-table entry. The code - introduced in commit 1998cc048901 ("mm: make madvise(MADV_WILLNEED) support swap file prefetch") - correctly checks for this situation using pmd_none_or_trans_huge_or_clear_bad(), but it turns out that that function doesn't work correctly. pmd_none_or_trans_huge_or_clear_bad() expected that pmd_bad() would trigger if the transparent hugepage bit was set, but it doesn't do that if pmd_numa() is also set. Note that the NUMA bit only gets set on real NUMA machines, so people trying to reproduce this on most normal development systems would never actually trigger this. Fix it by removing the very subtle (and subtly incorrect) expectation, and instead just checking pmd_trans_huge() explicitly. Reported-by: Sasha Levin Acked-by: Andrea Arcangeli [ Additionally remove the now stale test for pmd_trans_huge() inside the pmd_bad() case - Linus ] Signed-off-by: Linus Torvalds diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h index b12079a..db09234 100644 --- a/include/asm-generic/pgtable.h +++ b/include/asm-generic/pgtable.h @@ -599,11 +599,10 @@ static inline int pmd_none_or_trans_huge_or_clear_bad(pmd_t *pmd) #ifdef CONFIG_TRANSPARENT_HUGEPAGE barrier(); #endif - if (pmd_none(pmdval)) + if (pmd_none(pmdval) || pmd_trans_huge(pmdval)) return 1; if (unlikely(pmd_bad(pmdval))) { - if (!pmd_trans_huge(pmdval)) - pmd_clear_bad(pmd); + pmd_clear_bad(pmd); return 1; } return 0; -- cgit v0.10.2 From 8798cee2f90292ca4ca84baccc757e4d06f8797f Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Fri, 20 Dec 2013 14:54:11 +0000 Subject: Revert "mm: page_alloc: exclude unreclaimable allocations from zone fairness policy" This reverts commit 73f038b863df. The NUMA behaviour of this patch is less than ideal. An alternative approch is to interleave allocations only within local zones which is implemented in the next patch. Cc: stable@vger.kernel.org Signed-off-by: Mel Gorman Signed-off-by: Linus Torvalds diff --git a/mm/page_alloc.c b/mm/page_alloc.c index f861d02..580a5f0 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1920,8 +1920,7 @@ zonelist_scan: * back to remote zones that do not partake in the * fairness round-robin cycle of this zonelist. */ - if ((alloc_flags & ALLOC_WMARK_LOW) && - (gfp_mask & GFP_MOVABLE_MASK)) { + if (alloc_flags & ALLOC_WMARK_LOW) { if (zone_page_state(zone, NR_ALLOC_BATCH) <= 0) continue; if (zone_reclaim_mode && -- cgit v0.10.2 From fff4068cba484e6b0abe334ed6b15d5a215a3b25 Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Fri, 20 Dec 2013 14:54:12 +0000 Subject: mm: page_alloc: revert NUMA aspect of fair allocation policy Commit 81c0a2bb515f ("mm: page_alloc: fair zone allocator policy") meant to bring aging fairness among zones in system, but it was overzealous and badly regressed basic workloads on NUMA systems. Due to the way kswapd and page allocator interacts, we still want to make sure that all zones in any given node are used equally for all allocations to maximize memory utilization and prevent thrashing on the highest zone in the node. While the same principle applies to NUMA nodes - memory utilization is obviously improved by spreading allocations throughout all nodes - remote references can be costly and so many workloads prefer locality over memory utilization. The original change assumed that zone_reclaim_mode would be a good enough predictor for that, but it turned out to be as indicative as a coin flip. Revert the NUMA aspect of the fairness until we can find a proper way to make it configurable and agree on a sane default. Signed-off-by: Johannes Weiner Reviewed-by: Michal Hocko Signed-off-by: Mel Gorman Cc: # 3.12 Signed-off-by: Linus Torvalds diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 580a5f0..5248fe0 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1816,7 +1816,7 @@ static void zlc_clear_zones_full(struct zonelist *zonelist) static bool zone_local(struct zone *local_zone, struct zone *zone) { - return node_distance(local_zone->node, zone->node) == LOCAL_DISTANCE; + return local_zone->node == zone->node; } static bool zone_allows_reclaim(struct zone *local_zone, struct zone *zone) @@ -1913,18 +1913,17 @@ zonelist_scan: * page was allocated in should have no effect on the * time the page has in memory before being reclaimed. * - * When zone_reclaim_mode is enabled, try to stay in - * local zones in the fastpath. If that fails, the - * slowpath is entered, which will do another pass - * starting with the local zones, but ultimately fall - * back to remote zones that do not partake in the - * fairness round-robin cycle of this zonelist. + * Try to stay in local zones in the fastpath. If + * that fails, the slowpath is entered, which will do + * another pass starting with the local zones, but + * ultimately fall back to remote zones that do not + * partake in the fairness round-robin cycle of this + * zonelist. */ if (alloc_flags & ALLOC_WMARK_LOW) { if (zone_page_state(zone, NR_ALLOC_BATCH) <= 0) continue; - if (zone_reclaim_mode && - !zone_local(preferred_zone, zone)) + if (!zone_local(preferred_zone, zone)) continue; } /* @@ -2390,7 +2389,7 @@ static void prepare_slowpath(gfp_t gfp_mask, unsigned int order, * thrash fairness information for zones that are not * actually part of this zonelist's round-robin cycle. */ - if (zone_reclaim_mode && !zone_local(preferred_zone, zone)) + if (!zone_local(preferred_zone, zone)) continue; mod_zone_page_state(zone, NR_ALLOC_BATCH, high_wmark_pages(zone) - -- cgit v0.10.2 From 597d795a2a786d22dd872332428e2b9439ede639 Mon Sep 17 00:00:00 2001 From: "Kirill A. Shutemov" Date: Fri, 20 Dec 2013 13:35:58 +0200 Subject: mm: do not allocate page->ptl dynamically, if spinlock_t fits to long In struct page we have enough space to fit long-size page->ptl there, but we use dynamically-allocated page->ptl if size(spinlock_t) is larger than sizeof(int). It hurts 64-bit architectures with CONFIG_GENERIC_LOCKBREAK, where sizeof(spinlock_t) == 8, but it easily fits into struct page. Signed-off-by: Kirill A. Shutemov Acked-by: Hugh Dickins Signed-off-by: Linus Torvalds diff --git a/include/linux/lockref.h b/include/linux/lockref.h index c8929c3..4bfde0e 100644 --- a/include/linux/lockref.h +++ b/include/linux/lockref.h @@ -19,7 +19,7 @@ #define USE_CMPXCHG_LOCKREF \ (IS_ENABLED(CONFIG_ARCH_USE_CMPXCHG_LOCKREF) && \ - IS_ENABLED(CONFIG_SMP) && !BLOATED_SPINLOCKS) + IS_ENABLED(CONFIG_SMP) && SPINLOCK_SIZE <= 4) struct lockref { union { diff --git a/include/linux/mm.h b/include/linux/mm.h index 1cedd00..3552717 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1317,7 +1317,7 @@ static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long a #endif /* CONFIG_MMU && !__ARCH_HAS_4LEVEL_HACK */ #if USE_SPLIT_PTE_PTLOCKS -#if BLOATED_SPINLOCKS +#if ALLOC_SPLIT_PTLOCKS extern bool ptlock_alloc(struct page *page); extern void ptlock_free(struct page *page); @@ -1325,7 +1325,7 @@ static inline spinlock_t *ptlock_ptr(struct page *page) { return page->ptl; } -#else /* BLOATED_SPINLOCKS */ +#else /* ALLOC_SPLIT_PTLOCKS */ static inline bool ptlock_alloc(struct page *page) { return true; @@ -1339,7 +1339,7 @@ static inline spinlock_t *ptlock_ptr(struct page *page) { return &page->ptl; } -#endif /* BLOATED_SPINLOCKS */ +#endif /* ALLOC_SPLIT_PTLOCKS */ static inline spinlock_t *pte_lockptr(struct mm_struct *mm, pmd_t *pmd) { diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index ad0616f..290901a 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -26,6 +26,7 @@ struct address_space; #define USE_SPLIT_PTE_PTLOCKS (NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS) #define USE_SPLIT_PMD_PTLOCKS (USE_SPLIT_PTE_PTLOCKS && \ IS_ENABLED(CONFIG_ARCH_ENABLE_SPLIT_PMD_PTLOCK)) +#define ALLOC_SPLIT_PTLOCKS (SPINLOCK_SIZE > BITS_PER_LONG/8) /* * Each physical page in the system has a struct page associated with @@ -155,7 +156,7 @@ struct page { * system if PG_buddy is set. */ #if USE_SPLIT_PTE_PTLOCKS -#if BLOATED_SPINLOCKS +#if ALLOC_SPLIT_PTLOCKS spinlock_t *ptl; #else spinlock_t ptl; diff --git a/kernel/bounds.c b/kernel/bounds.c index 5253204..9fd4246 100644 --- a/kernel/bounds.c +++ b/kernel/bounds.c @@ -22,6 +22,6 @@ void foo(void) #ifdef CONFIG_SMP DEFINE(NR_CPUS_BITS, ilog2(CONFIG_NR_CPUS)); #endif - DEFINE(BLOATED_SPINLOCKS, sizeof(spinlock_t) > sizeof(int)); + DEFINE(SPINLOCK_SIZE, sizeof(spinlock_t)); /* End of constants */ } diff --git a/mm/memory.c b/mm/memory.c index 5d9025f..b6e211b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -4271,7 +4271,7 @@ void copy_user_huge_page(struct page *dst, struct page *src, } #endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HUGETLBFS */ -#if USE_SPLIT_PTE_PTLOCKS && BLOATED_SPINLOCKS +#if ALLOC_SPLIT_PTLOCKS bool ptlock_alloc(struct page *page) { spinlock_t *ptl; -- cgit v0.10.2 From df36ac1bc2a166eef90785d584e4cfed6f52bd32 Mon Sep 17 00:00:00 2001 From: "Luck, Tony" Date: Wed, 18 Dec 2013 15:17:10 -0800 Subject: pstore: Don't allow high traffic options on fragile devices Some pstore backing devices use on board flash as persistent storage. These have limited numbers of write cycles so it is a poor idea to use them from high frequency operations. Signed-off-by: Tony Luck Signed-off-by: Linus Torvalds diff --git a/drivers/acpi/apei/erst.c b/drivers/acpi/apei/erst.c index 26311f2..cb1d557 100644 --- a/drivers/acpi/apei/erst.c +++ b/drivers/acpi/apei/erst.c @@ -942,6 +942,7 @@ static int erst_clearer(enum pstore_type_id type, u64 id, int count, static struct pstore_info erst_info = { .owner = THIS_MODULE, .name = "erst", + .flags = PSTORE_FLAGS_FRAGILE, .open = erst_open_pstore, .close = erst_close_pstore, .read = erst_reader, diff --git a/drivers/firmware/efi/efi-pstore.c b/drivers/firmware/efi/efi-pstore.c index 743fd42..4b9dc83 100644 --- a/drivers/firmware/efi/efi-pstore.c +++ b/drivers/firmware/efi/efi-pstore.c @@ -356,6 +356,7 @@ static int efi_pstore_erase(enum pstore_type_id type, u64 id, int count, static struct pstore_info efi_pstore_info = { .owner = THIS_MODULE, .name = "efi", + .flags = PSTORE_FLAGS_FRAGILE, .open = efi_pstore_open, .close = efi_pstore_close, .read = efi_pstore_read, diff --git a/fs/pstore/platform.c b/fs/pstore/platform.c index b8e93a4..78c3c20 100644 --- a/fs/pstore/platform.c +++ b/fs/pstore/platform.c @@ -443,8 +443,11 @@ int pstore_register(struct pstore_info *psi) pstore_get_records(0); kmsg_dump_register(&pstore_dumper); - pstore_register_console(); - pstore_register_ftrace(); + + if ((psi->flags & PSTORE_FLAGS_FRAGILE) == 0) { + pstore_register_console(); + pstore_register_ftrace(); + } if (pstore_update_ms >= 0) { pstore_timer.expires = jiffies + diff --git a/include/linux/pstore.h b/include/linux/pstore.h index abd437d..ece0c6b 100644 --- a/include/linux/pstore.h +++ b/include/linux/pstore.h @@ -51,6 +51,7 @@ struct pstore_info { char *buf; size_t bufsize; struct mutex read_mutex; /* serialize open/read/close */ + int flags; int (*open)(struct pstore_info *psi); int (*close)(struct pstore_info *psi); ssize_t (*read)(u64 *id, enum pstore_type_id *type, @@ -70,6 +71,8 @@ struct pstore_info { void *data; }; +#define PSTORE_FLAGS_FRAGILE 1 + #ifdef CONFIG_PSTORE extern int pstore_register(struct pstore_info *); extern bool pstore_cannot_block_path(enum kmsg_dump_reason reason); -- cgit v0.10.2 From 40b64acd17a2200579db265048b4a51998a84729 Mon Sep 17 00:00:00 2001 From: Olof Johansson Date: Fri, 20 Dec 2013 14:28:05 -0800 Subject: mm: fix build of split ptlock code Commit 597d795a2a78 ('mm: do not allocate page->ptl dynamically, if spinlock_t fits to long') restructures some allocators that are compiled even if USE_SPLIT_PTLOCKS arn't used. It results in compilation failure: mm/memory.c:4282:6: error: 'struct page' has no member named 'ptl' mm/memory.c:4288:12: error: 'struct page' has no member named 'ptl' Add in the missing ifdef. Fixes: 597d795a2a78 ('mm: do not allocate page->ptl dynamically, if spinlock_t fits to long') Signed-off-by: Olof Johansson Cc: Kirill A. Shutemov Cc: Hugh Dickins Signed-off-by: Linus Torvalds diff --git a/mm/memory.c b/mm/memory.c index b6e211b..6768ce9 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -4271,7 +4271,7 @@ void copy_user_huge_page(struct page *dst, struct page *src, } #endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HUGETLBFS */ -#if ALLOC_SPLIT_PTLOCKS +#if USE_SPLIT_PTE_PTLOCKS && ALLOC_SPLIT_PTLOCKS bool ptlock_alloc(struct page *page) { spinlock_t *ptl; -- cgit v0.10.2 From b7000adef17a5cce85636e40fa2c2d9851a89e28 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 20 Dec 2013 16:52:45 -0800 Subject: Don't set the INITRD_COMPRESS environment variable automatically Commit 1bf49dd4be0b ("./Makefile: export initial ramdisk compression config option") started setting the INITRD_COMPRESS environment variable depending on which decompression models the kernel had available. That is completely broken. For example, we by default have CONFIG_RD_LZ4 enabled, and are able to decompress such an initrd, but the user tools to *create* such an initrd may not be availble. So trying to tell dracut to generate an lz4-compressed image just because we can decode such an image is completely inappropriate. Cc: J P Cc: Andrew Morton Cc: Jan Beulich Signed-off-by: Linus Torvalds diff --git a/Makefile b/Makefile index 89cee86..82ea7dd 100644 --- a/Makefile +++ b/Makefile @@ -738,7 +738,9 @@ INITRD_COMPRESS-$(CONFIG_RD_LZMA) := lzma INITRD_COMPRESS-$(CONFIG_RD_XZ) := xz INITRD_COMPRESS-$(CONFIG_RD_LZO) := lzo INITRD_COMPRESS-$(CONFIG_RD_LZ4) := lz4 -export INITRD_COMPRESS := $(INITRD_COMPRESS-y) +# do not export INITRD_COMPRESS, since we didn't actually +# choose a sane default compression above. +# export INITRD_COMPRESS := $(INITRD_COMPRESS-y) ifdef CONFIG_MODULE_SIG_ALL MODSECKEY = ./signing_key.priv -- cgit v0.10.2 From 1881686f842065d2f92ec9c6424830ffc17d23b0 Mon Sep 17 00:00:00 2001 From: Benjamin LaHaise Date: Sat, 21 Dec 2013 15:49:28 -0500 Subject: aio: fix kioctx leak introduced by "aio: Fix a trinity splat" e34ecee2ae791df674dfb466ce40692ca6218e43 reworked the percpu reference counting to correct a bug trinity found. Unfortunately, the change lead to kioctxes being leaked because there was no final reference count to put. Add that reference count back in to fix things. Signed-off-by: Benjamin LaHaise Cc: stable@vger.kernel.org diff --git a/fs/aio.c b/fs/aio.c index 6efb7f6..fd1c0ba 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -652,7 +652,8 @@ static struct kioctx *ioctx_alloc(unsigned nr_events) aio_nr += ctx->max_reqs; spin_unlock(&aio_nr_lock); - percpu_ref_get(&ctx->users); /* io_setup() will drop this ref */ + percpu_ref_get(&ctx->users); /* io_setup() will drop this ref */ + percpu_ref_get(&ctx->reqs); /* free_ioctx_users() will drop this */ err = ioctx_add_table(ctx, mm); if (err) -- cgit v0.10.2 From 8e321fefb0e60bae4e2a28d20fc4fa30758d27c6 Mon Sep 17 00:00:00 2001 From: Benjamin LaHaise Date: Sat, 21 Dec 2013 17:56:08 -0500 Subject: aio/migratepages: make aio migrate pages sane The arbitrary restriction on page counts offered by the core migrate_page_move_mapping() code results in rather suspicious looking fiddling with page reference counts in the aio_migratepage() operation. To fix this, make migrate_page_move_mapping() take an extra_count parameter that allows aio to tell the code about its own reference count on the page being migrated. While cleaning up aio_migratepage(), make it validate that the old page being passed in is actually what aio_migratepage() expects to prevent misbehaviour in the case of races. Signed-off-by: Benjamin LaHaise diff --git a/fs/aio.c b/fs/aio.c index fd1c0ba..efa708b 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -244,9 +244,14 @@ static void aio_free_ring(struct kioctx *ctx) int i; for (i = 0; i < ctx->nr_pages; i++) { + struct page *page; pr_debug("pid(%d) [%d] page->count=%d\n", current->pid, i, page_count(ctx->ring_pages[i])); - put_page(ctx->ring_pages[i]); + page = ctx->ring_pages[i]; + if (!page) + continue; + ctx->ring_pages[i] = NULL; + put_page(page); } put_aio_ring_file(ctx); @@ -280,18 +285,38 @@ static int aio_migratepage(struct address_space *mapping, struct page *new, unsigned long flags; int rc; + rc = 0; + + /* Make sure the old page hasn't already been changed */ + spin_lock(&mapping->private_lock); + ctx = mapping->private_data; + if (ctx) { + pgoff_t idx; + spin_lock_irqsave(&ctx->completion_lock, flags); + idx = old->index; + if (idx < (pgoff_t)ctx->nr_pages) { + if (ctx->ring_pages[idx] != old) + rc = -EAGAIN; + } else + rc = -EINVAL; + spin_unlock_irqrestore(&ctx->completion_lock, flags); + } else + rc = -EINVAL; + spin_unlock(&mapping->private_lock); + + if (rc != 0) + return rc; + /* Writeback must be complete */ BUG_ON(PageWriteback(old)); - put_page(old); + get_page(new); - rc = migrate_page_move_mapping(mapping, new, old, NULL, mode); + rc = migrate_page_move_mapping(mapping, new, old, NULL, mode, 1); if (rc != MIGRATEPAGE_SUCCESS) { - get_page(old); + put_page(new); return rc; } - get_page(new); - /* We can potentially race against kioctx teardown here. Use the * address_space's private data lock to protect the mapping's * private_data. @@ -303,13 +328,24 @@ static int aio_migratepage(struct address_space *mapping, struct page *new, spin_lock_irqsave(&ctx->completion_lock, flags); migrate_page_copy(new, old); idx = old->index; - if (idx < (pgoff_t)ctx->nr_pages) - ctx->ring_pages[idx] = new; + if (idx < (pgoff_t)ctx->nr_pages) { + /* And only do the move if things haven't changed */ + if (ctx->ring_pages[idx] == old) + ctx->ring_pages[idx] = new; + else + rc = -EAGAIN; + } else + rc = -EINVAL; spin_unlock_irqrestore(&ctx->completion_lock, flags); } else rc = -EBUSY; spin_unlock(&mapping->private_lock); + if (rc == MIGRATEPAGE_SUCCESS) + put_page(old); + else + put_page(new); + return rc; } #endif diff --git a/include/linux/migrate.h b/include/linux/migrate.h index b7717d7..f015c05 100644 --- a/include/linux/migrate.h +++ b/include/linux/migrate.h @@ -55,7 +55,8 @@ extern int migrate_huge_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page); extern int migrate_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page, - struct buffer_head *head, enum migrate_mode mode); + struct buffer_head *head, enum migrate_mode mode, + int extra_count); #else static inline void putback_lru_pages(struct list_head *l) {} diff --git a/mm/migrate.c b/mm/migrate.c index e9b7102..9194375 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -317,14 +317,15 @@ static inline bool buffer_migrate_lock_buffers(struct buffer_head *head, */ int migrate_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page, - struct buffer_head *head, enum migrate_mode mode) + struct buffer_head *head, enum migrate_mode mode, + int extra_count) { - int expected_count = 0; + int expected_count = 1 + extra_count; void **pslot; if (!mapping) { /* Anonymous page without mapping */ - if (page_count(page) != 1) + if (page_count(page) != expected_count) return -EAGAIN; return MIGRATEPAGE_SUCCESS; } @@ -334,7 +335,7 @@ int migrate_page_move_mapping(struct address_space *mapping, pslot = radix_tree_lookup_slot(&mapping->page_tree, page_index(page)); - expected_count = 2 + page_has_private(page); + expected_count += 1 + page_has_private(page); if (page_count(page) != expected_count || radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { spin_unlock_irq(&mapping->tree_lock); @@ -584,7 +585,7 @@ int migrate_page(struct address_space *mapping, BUG_ON(PageWriteback(page)); /* Writeback must be complete */ - rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode); + rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode, 0); if (rc != MIGRATEPAGE_SUCCESS) return rc; @@ -611,7 +612,7 @@ int buffer_migrate_page(struct address_space *mapping, head = page_buffers(page); - rc = migrate_page_move_mapping(mapping, newpage, page, head, mode); + rc = migrate_page_move_mapping(mapping, newpage, page, head, mode, 0); if (rc != MIGRATEPAGE_SUCCESS) return rc; -- cgit v0.10.2 From 3dc9acb67600393249a795934ccdfc291a200e6b Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 20 Dec 2013 05:11:12 +0900 Subject: aio: clean up and fix aio_setup_ring page mapping Since commit 36bc08cc01709 ("fs/aio: Add support to aio ring pages migration") the aio ring setup code has used a special per-ring backing inode for the page allocations, rather than just using random anonymous pages. However, rather than remembering the pages as it allocated them, it would allocate the pages, insert them into the file mapping (dirty, so that they couldn't be free'd), and then forget about them. And then to look them up again, it would mmap the mapping, and then use "get_user_pages()" to get back an array of the pages we just created. Now, not only is that incredibly inefficient, it also leaked all the pages if the mmap failed (which could happen due to excessive number of mappings, for example). So clean it all up, making it much more straightforward. Also remove some left-overs of the previous (broken) mm_populate() usage that was removed in commit d6c355c7dabc ("aio: fix race in ring buffer page lookup introduced by page migration support") but left the pointless and now misleading MAP_POPULATE flag around. Tested-and-acked-by: Benjamin LaHaise Signed-off-by: Linus Torvalds diff --git a/fs/aio.c b/fs/aio.c index 6efb7f6..643db8f 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -326,7 +326,7 @@ static int aio_setup_ring(struct kioctx *ctx) struct aio_ring *ring; unsigned nr_events = ctx->max_reqs; struct mm_struct *mm = current->mm; - unsigned long size, populate; + unsigned long size, unused; int nr_pages; int i; struct file *file; @@ -347,6 +347,20 @@ static int aio_setup_ring(struct kioctx *ctx) return -EAGAIN; } + ctx->aio_ring_file = file; + nr_events = (PAGE_SIZE * nr_pages - sizeof(struct aio_ring)) + / sizeof(struct io_event); + + ctx->ring_pages = ctx->internal_pages; + if (nr_pages > AIO_RING_PAGES) { + ctx->ring_pages = kcalloc(nr_pages, sizeof(struct page *), + GFP_KERNEL); + if (!ctx->ring_pages) { + put_aio_ring_file(ctx); + return -ENOMEM; + } + } + for (i = 0; i < nr_pages; i++) { struct page *page; page = find_or_create_page(file->f_inode->i_mapping, @@ -358,19 +372,14 @@ static int aio_setup_ring(struct kioctx *ctx) SetPageUptodate(page); SetPageDirty(page); unlock_page(page); + + ctx->ring_pages[i] = page; } - ctx->aio_ring_file = file; - nr_events = (PAGE_SIZE * nr_pages - sizeof(struct aio_ring)) - / sizeof(struct io_event); + ctx->nr_pages = i; - ctx->ring_pages = ctx->internal_pages; - if (nr_pages > AIO_RING_PAGES) { - ctx->ring_pages = kcalloc(nr_pages, sizeof(struct page *), - GFP_KERNEL); - if (!ctx->ring_pages) { - put_aio_ring_file(ctx); - return -ENOMEM; - } + if (unlikely(i != nr_pages)) { + aio_free_ring(ctx); + return -EAGAIN; } ctx->mmap_size = nr_pages * PAGE_SIZE; @@ -379,9 +388,9 @@ static int aio_setup_ring(struct kioctx *ctx) down_write(&mm->mmap_sem); ctx->mmap_base = do_mmap_pgoff(ctx->aio_ring_file, 0, ctx->mmap_size, PROT_READ | PROT_WRITE, - MAP_SHARED | MAP_POPULATE, 0, &populate); + MAP_SHARED, 0, &unused); + up_write(&mm->mmap_sem); if (IS_ERR((void *)ctx->mmap_base)) { - up_write(&mm->mmap_sem); ctx->mmap_size = 0; aio_free_ring(ctx); return -EAGAIN; @@ -389,27 +398,6 @@ static int aio_setup_ring(struct kioctx *ctx) pr_debug("mmap address: 0x%08lx\n", ctx->mmap_base); - /* We must do this while still holding mmap_sem for write, as we - * need to be protected against userspace attempting to mremap() - * or munmap() the ring buffer. - */ - ctx->nr_pages = get_user_pages(current, mm, ctx->mmap_base, nr_pages, - 1, 0, ctx->ring_pages, NULL); - - /* Dropping the reference here is safe as the page cache will hold - * onto the pages for us. It is also required so that page migration - * can unmap the pages and get the right reference count. - */ - for (i = 0; i < ctx->nr_pages; i++) - put_page(ctx->ring_pages[i]); - - up_write(&mm->mmap_sem); - - if (unlikely(ctx->nr_pages != nr_pages)) { - aio_free_ring(ctx); - return -EAGAIN; - } - ctx->user_id = ctx->mmap_base; ctx->nr_events = nr_events; /* trusted copy */ -- cgit v0.10.2 From 413541dd66d51f791a0b169d9b9014e4f56be13c Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 22 Dec 2013 13:08:32 -0800 Subject: Linux 3.13-rc5 diff --git a/Makefile b/Makefile index 82ea7dd..14d592c 100644 --- a/Makefile +++ b/Makefile @@ -1,7 +1,7 @@ VERSION = 3 PATCHLEVEL = 13 SUBLEVEL = 0 -EXTRAVERSION = -rc4 +EXTRAVERSION = -rc5 NAME = One Giant Leap for Frogkind # *DOCUMENTATION* -- cgit v0.10.2