From ca81bd6d5a9d81d0aab4cfb780f2d401d8c5e85a Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Thu, 21 Sep 2017 16:26:41 -0400 Subject: [PATCH 1900/4131] drm/amdgpu: Handle GPUVM fault storms When many wavefronts cause VM faults at the same time, it can overwhelm the interrupt handler and cause IH ring overflows before the driver can notify or kill the faulting application. As a workaround I'm introducing limited per-VM fault credit. After that number of VM faults have occurred, further VM faults are filtered out at the prescreen stage of processing. This depends on the PASID in the interrupt packet, so it currently only works for KFD contexts. Signed-off-by: Felix Kuehling Reviewed-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 31 +++++++++++++++++++++++++++++++ drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 7 ++++++- drivers/gpu/drm/amd/amdgpu/cik_ih.c | 19 +++++++++++++++++-- drivers/gpu/drm/amd/amdgpu/cz_ih.c | 19 +++++++++++++++++-- drivers/gpu/drm/amd/amdgpu/iceland_ih.c | 19 +++++++++++++++++-- drivers/gpu/drm/amd/amdgpu/tonga_ih.c | 19 +++++++++++++++++-- drivers/gpu/drm/amd/amdgpu/vega10_ih.c | 11 +++++++---- 7 files changed, 112 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index c542eef..3675c2f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -2727,6 +2727,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, } INIT_KFIFO(vm->faults); + vm->fault_credit = 16; return 0; @@ -2843,6 +2844,36 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) } /** + * amdgpu_vm_pasid_fault_credit - Check fault credit for given PASID + * + * @adev: amdgpu_device pointer + * @pasid: PASID do identify the VM + * + * This function is expected to be called in interrupt context. Returns + * true if there was fault credit, false otherwise + */ +bool amdgpu_vm_pasid_fault_credit(struct amdgpu_device *adev, + unsigned int pasid) +{ + struct amdgpu_vm *vm; + + spin_lock(&adev->vm_manager.pasid_lock); + vm = idr_find(&adev->vm_manager.pasid_idr, pasid); + spin_unlock(&adev->vm_manager.pasid_lock); + if (!vm) + /* VM not found, can't track fault credit */ + return true; + + /* No lock needed. only accessed by IRQ handler */ + if (!vm->fault_credit) + /* Too many faults in this VM */ + return false; + + vm->fault_credit--; + return true; +} + +/** * amdgpu_vm_manager_init - init the VM manager * * @adev: amdgpu_device pointer diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index 2056f99..1a41459 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -168,8 +168,11 @@ struct amdgpu_vm { /* Flag to indicate ATS support from PTE for GFX9 */ bool pte_support_ats; - /* Up to 128 pending page faults */ + /* Up to 128 pending retry page faults */ DECLARE_KFIFO(faults, u64, 128); + + /* Limit non-retry fault storms */ + unsigned int fault_credit; }; struct amdgpu_vm_id { @@ -249,6 +252,8 @@ void amdgpu_vm_manager_fini(struct amdgpu_device *adev); int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, int vm_context, unsigned int pasid); void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm); +bool amdgpu_vm_pasid_fault_credit(struct amdgpu_device *adev, + unsigned int pasid); void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, struct list_head *validated, struct amdgpu_bo_list_entry *entry); diff --git a/drivers/gpu/drm/amd/amdgpu/cik_ih.c b/drivers/gpu/drm/amd/amdgpu/cik_ih.c index 07d3d89..a870b35 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_ih.c @@ -237,8 +237,23 @@ static u32 cik_ih_get_wptr(struct amdgpu_device *adev) */ static bool cik_ih_prescreen_iv(struct amdgpu_device *adev) { - /* Process all interrupts */ - return true; + u32 ring_index = adev->irq.ih.rptr >> 2; + u16 pasid; + + switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { + case 146: + case 147: + pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; + if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) + return true; + break; + default: + /* Not a VM fault */ + return true; + } + + adev->irq.ih.rptr += 16; + return false; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c index b6cdf4a..fa61d64 100644 --- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c @@ -216,8 +216,23 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev) */ static bool cz_ih_prescreen_iv(struct amdgpu_device *adev) { - /* Process all interrupts */ - return true; + u32 ring_index = adev->irq.ih.rptr >> 2; + u16 pasid; + + switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { + case 146: + case 147: + pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; + if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) + return true; + break; + default: + /* Not a VM fault */ + return true; + } + + adev->irq.ih.rptr += 16; + return false; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c index 65ed6d3..bd592cb 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c @@ -216,8 +216,23 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev) */ static bool iceland_ih_prescreen_iv(struct amdgpu_device *adev) { - /* Process all interrupts */ - return true; + u32 ring_index = adev->irq.ih.rptr >> 2; + u16 pasid; + + switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { + case 146: + case 147: + pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; + if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) + return true; + break; + default: + /* Not a VM fault */ + return true; + } + + adev->irq.ih.rptr += 16; + return false; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c index 5ed0069..aa4e320 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c @@ -227,8 +227,23 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev) */ static bool tonga_ih_prescreen_iv(struct amdgpu_device *adev) { - /* Process all interrupts */ - return true; + u32 ring_index = adev->irq.ih.rptr >> 2; + u16 pasid; + + switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { + case 146: + case 147: + pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; + if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) + return true; + break; + default: + /* Not a VM fault */ + return true; + } + + adev->irq.ih.rptr += 16; + return false; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c index a3b30d8..6973257 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c @@ -260,15 +260,18 @@ static bool vega10_ih_prescreen_iv(struct amdgpu_device *adev) return true; } - /* Not a retry fault */ - if (!(dw5 & 0x80)) - return true; - pasid = dw3 & 0xffff; /* No PASID, can't identify faulting process */ if (!pasid) return true; + /* Not a retry fault, check fault credit */ + if (!(dw5 & 0x80)) { + if (!amdgpu_vm_pasid_fault_credit(adev, pasid)) + goto ignore_iv; + return true; + } + addr = ((u64)(dw5 & 0xf) << 44) | ((u64)dw4 << 12); key = AMDGPU_VM_FAULT(pasid, addr); r = amdgpu_ih_add_fault(adev, key); -- 2.7.4