drm/amdkfd: check access permisson to restore retry fault
Subsystems (3)
PATCH 1/2 drm/amdkfd: check access permisson to restore retry fault
Check range access permission to restore GPU retry fault, if GPU retry fault on address which belongs to VMA, and VMA has no read or write permission requested by GPU, failed to restore the address. The vm fault event will pass back to user space.
--- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 5 +++-- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 2 +- drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 3 ++- drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 3 ++- drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 30 +++++++++++++++++++++++++- drivers/gpu/drm/amd/amdkfd/kfd_svm.h | 5 +++-- 6 files changed, 40 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +3 -2
@@ -3347,12 +3347,13 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)* @adev: amdgpu device pointer * @pasid: PASID of the VM * @addr: Address of the fault +* @rw_fault: 0 is read fault, 1 is write fault * * Try to gracefully handle a VM fault. Return true if the fault was handled and * shouldn't be reported any more. */ bool amdgpu_vm_handle_fault (struct amdgpu_device *adev,u32 pasid, -uint64_t addr) +uint64_t addr,uint32_t rw_fault) {bool is_compute_context = false;struct amdgpu_bo *root;
@@ -3377,7 +3378,7 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, addr /=AMDGPU_GPU_PAGE_SIZE ;if (is_compute_context && - !svm_range_restore_pages (adev, pasid, addr)) { + !svm_range_restore_pages (adev, pasid, addr, rw_fault)) {amdgpu_bo_unref (&root);return true; }
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +1 -1
@@ -448,7 +448,7 @@ void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev);void amdgpu_vm_get_task_info (struct amdgpu_device *adev,u32 pasid,struct amdgpu_task_info *task_info);bool amdgpu_vm_handle_fault (struct amdgpu_device *adev,u32 pasid, -uint64_t addr); +uint64_t addr,uint32_t rw_fault);void amdgpu_vm_set_task_info (struct amdgpu_vm *vm);
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +2 -1
@@ -93,6 +93,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) {bool retry_fault = !!(entry->src_data[1 ] &0x80 ); +bool rw_fault = !!(entry->src_data[1 ] &0x20 );struct amdgpu_vmhub *hub = &adev->vmhub[entry->vmid_src];struct amdgpu_task_info task_info;uint32_t status =0 ;
@@ -121,7 +122,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,/* Try to handle the recoverable page faults by filling page * tables */ -if (amdgpu_vm_handle_fault (adev, entry->pasid, addr)) +if (amdgpu_vm_handle_fault (adev, entry->pasid, addr, rw_fault))return 1 ; }
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +2 -1
@@ -506,6 +506,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) {bool retry_fault = !!(entry->src_data[1 ] &0x80 ); +bool rw_fault = !!(entry->src_data[1 ] &0x20 );uint32_t status =0 , cid =0 , rw =0 ;struct amdgpu_task_info task_info;struct amdgpu_vmhub *hub;
@@ -536,7 +537,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,/* Try to handle the recoverable page faults by filling page * tables */ -if (amdgpu_vm_handle_fault (adev, entry->pasid, addr)) +if (amdgpu_vm_handle_fault (adev, entry->pasid, addr, rw_fault))return 1 ; }
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +29 -1
@@ -2400,9 +2400,29 @@ svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p,WRITE_ONCE (pdd->faults, pdd->faults +1 ); } +static bool +svm_range_allow_access (struct mm_struct *mm,uint64_t addr,uint32_t rw_fault) +{ +unsigned long requested =VM_READ ; +struct vm_area_struct *vma; + +if (rw_fault) + requested |=VM_WRITE ; + + vma =find_vma (mm, addr <<PAGE_SHIFT ); +if (!vma || (addr <<PAGE_SHIFT ) < vma->vm_start) { +pr_debug ("address 0x%llx VMA is removed\n" , addr); +return true; + } + +pr_debug ("requested 0x%lx, vma permission flags 0x%lx\n" , requested, + vma->vm_flags); +return (requested & ~vma->vm_flags) ==0 ; +} +int svm_range_restore_pages (struct amdgpu_device *adev,unsigned int pasid, -uint64_t addr) +uint64_t addr,uint32_t rw_fault) {struct mm_struct *mm =NULL ;struct svm_range_list *svms;
@@ -2440,6 +2460,7 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, }mmap_read_lock (mm); + retry_write_locked:mutex_lock (&svms->lock); prange =svm_range_from_addr (svms, addr,NULL );
@@ -2484,6 +2505,13 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, goto out_unlock_range; } +if (!svm_range_allow_access (mm, addr, rw_fault)) { +pr_debug ("fault addr 0x%llx no %s permission\n" , addr, + rw_fault ?"write" :"read" ); + r = -EPERM ; + goto out_unlock_range; + } + best_loc =svm_range_best_restore_location (prange, adev, &gpuidx);if (best_loc ==-1 ) { pr_debug("svms %p failed get best restore loc [0x%lx 0x%lx]\n",
drivers/gpu/drm/amd/amdkfd/kfd_svm.h +3 -2
@@ -175,7 +175,7 @@ int svm_range_split_by_granularity(struct kfd_process *p, struct mm_struct *mm, unsigned long addr, struct svm_range *parent, struct svm_range *prange);int svm_range_restore_pages (struct amdgpu_device *adev, -unsigned int pasid,uint64_t addr); +unsigned int pasid,uint64_t addr,uint32_t rw);int svm_range_schedule_evict_svm_bo (struct amdgpu_amdkfd_fence *fence); void svm_range_add_list_work(struct svm_range_list *svms, struct svm_range *prange, struct mm_struct *mm,
@@ -210,7 +210,8 @@ static inline void svm_range_list_fini(struct kfd_process *p) }static inline int svm_range_restore_pages (struct amdgpu_device *adev, -unsigned int pasid,uint64_t addr) +unsigned int pasid,uint64_t addr, +uint32_t rw_fault) {return -EFAULT ; }
Signature · 1 line
2.17.1
PATCH 2/2 drm/amdkfd: map SVM range with correct access permission
Restore retry fault or prefetch range, or restore svm range after eviction to map range to GPU with correct read or write access permission.
Range may includes multiple VMAs, update GPU page table with offset of prange, number of pages for each VMA according VMA access permission.
--- drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 131 +++++++++++++++++---------- 1 file changed, 84 insertions(+), 47 deletions(-)
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +84 -47
@@ -120,6 +120,7 @@ static void svm_range_remove_notifier(struct svm_range *prange)static int svm_range_dma_map_dev (struct amdgpu_device *adev,struct svm_range *prange, +unsigned long offset,unsigned long npages,unsigned long *hmm_pfns,uint32_t gpuidx) {enum dma_data_direction dir =DMA_BIDIRECTIONAL ;
@@ -136,7 +137,8 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange, prange->dma_addr[gpuidx] = addr; } -for (i =0 ; i < prange->npages; i++) { + addr += offset; +for (i =0 ; i < npages; i++) {if (WARN_ONCE (addr[i] && !dma_mapping_error (dev, addr[i]),"leaking dma mapping\n" ))dma_unmap_page (dev, addr[i],PAGE_SIZE , dir);
@@ -167,6 +169,7 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange,static int svm_range_dma_map (struct svm_range *prange,unsigned long *bitmap, +unsigned long offset,unsigned long npages,unsigned long *hmm_pfns) {struct kfd_process *p;
@@ -187,7 +190,8 @@ svm_range_dma_map(struct svm_range *prange, unsigned long *bitmap, } adev = (struct amdgpu_device *)pdd->dev->kgd; - r =svm_range_dma_map_dev (adev, prange, hmm_pfns, gpuidx); + r =svm_range_dma_map_dev (adev, prange, offset, npages, + hmm_pfns, gpuidx);if (r)break ; }
@@ -1088,11 +1092,6 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange, pte_flags |= snoop ?AMDGPU_PTE_SNOOPED :0 ; pte_flags |=amdgpu_gem_va_map_flags (adev, mapping_flags); - -pr_debug ("svms 0x%p [0x%lx 0x%lx] vram %d PTE 0x%llx mapping 0x%x\n" , - prange->svms, prange->start, prange->last, - (domain ==SVM_RANGE_VRAM_DOMAIN ) ?1 :0 , pte_flags, mapping_flags); -return pte_flags; }
@@ -1156,7 +1155,8 @@ svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start,static int svm_range_map_to_gpu (struct amdgpu_device *adev,struct amdgpu_vm *vm, -struct svm_range *prange,dma_addr_t *dma_addr, +struct svm_range *prange,unsigned long offset, +unsigned long npages,bool readonly,dma_addr_t *dma_addr,struct amdgpu_device *bo_adev,struct dma_fence **fence) {struct amdgpu_bo_va bo_va;
@@ -1167,14 +1167,15 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,int r =0 ;int64_t i; -pr_debug ("svms 0x%p [0x%lx 0x%lx]\n" , prange->svms, prange->start, - prange->last); + last_start = prange->start + offset; + +pr_debug ("svms 0x%p [0x%lx 0x%lx] readonly %d\n" , prange->svms, + last_start, last_start + npages -1 , readonly);if (prange->svm_bo && prange->ttm_res) bo_va.is_xgmi =amdgpu_xgmi_same_hive (adev, bo_adev); - last_start = prange->start; -for (i =0 ; i < prange->npages; i++) { +for (i = offset; i < offset + npages; i++) { last_domain = dma_addr[i] &SVM_RANGE_VRAM_DOMAIN ; dma_addr[i] &= ~SVM_RANGE_VRAM_DOMAIN ; if ((prange->start + i) < prange->last &&
@@ -1183,13 +1184,21 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,pr_debug ("Mapping range [0x%lx 0x%llx] on domain: %s\n" , last_start, prange->start + i, last_domain ?"GPU" :"CPU" ); + pte_flags =svm_range_get_pte_flags (adev, prange, last_domain); - r =amdgpu_vm_bo_update_mapping (adev, bo_adev, vm, false, false,NULL , - last_start, +if (readonly) + pte_flags &= ~AMDGPU_PTE_WRITEABLE ; + +pr_debug ("svms 0x%p map [0x%lx 0x%llx] vram %d PTE 0x%llx\n" , + prange->svms, last_start, prange->start + i, + (last_domain ==SVM_RANGE_VRAM_DOMAIN ) ?1 :0 , + pte_flags); + + r =amdgpu_vm_bo_update_mapping (adev, bo_adev, vm, false, false, +NULL , last_start, prange->start + i, pte_flags, last_start - prange->start, -NULL , - dma_addr, +NULL , dma_addr, &vm->last_update, &table_freed);if (r) {
@@ -1220,8 +1229,10 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,return r; } -static int svm_range_map_to_gpus (struct svm_range *prange, -unsigned long *bitmap,bool wait) +static int +svm_range_map_to_gpus (struct svm_range *prange,unsigned long offset, +unsigned long npages,bool readonly, +unsigned long *bitmap,bool wait) {struct kfd_process_device *pdd;struct amdgpu_device *bo_adev;
@@ -1257,7 +1268,8 @@ static int svm_range_map_to_gpus(struct svm_range *prange, } r =svm_range_map_to_gpu (adev,drm_priv_to_vm (pdd->drm_priv), - prange, prange->dma_addr[gpuidx], + prange, offset, npages, readonly, + prange->dma_addr[gpuidx], bo_adev, wait ? &fence :NULL );if (r)break ;
@@ -1390,6 +1402,7 @@ static int svm_range_validate_and_map(struct mm_struct *mm, int32_t gpuidx, bool intr, bool wait) {struct svm_validate_context ctx; +unsigned long start, end, addr;struct hmm_range *hmm_range;struct kfd_process *p;void *owner;
@@ -1448,40 +1461,64 @@ static int svm_range_validate_and_map(struct mm_struct *mm,break ; } } - r =amdgpu_hmm_range_get_pages (&prange->notifier, mm,NULL , - prange->start <<PAGE_SHIFT , - prange->npages, &hmm_range, - false, true, owner); -if (r) { -pr_debug ("failed %d to get svm range pages\n" , r); - goto unreserve_out; - } - r =svm_range_dma_map (prange, ctx.bitmap, - hmm_range->hmm_pfns); -if (r) { -pr_debug ("failed %d to dma map range\n" , r); - goto unreserve_out; - } + start = prange->start <<PAGE_SHIFT ; + end = (prange->last +1 ) <<PAGE_SHIFT ; +for (addr = start; addr < end && !r; ) { +struct vm_area_struct *vma; +unsigned long next; +unsigned long offset; +unsigned long npages; +bool readonly; - prange->validated_once = true; + vma =find_vma (mm, addr); +if (!vma || addr < vma->vm_start) { + r = -EINVAL ; + goto unreserve_out; + } + readonly = !(vma->vm_flags &VM_WRITE ); -svm_range_lock (prange); -if (amdgpu_hmm_range_get_pages_done (hmm_range)) { -pr_debug ("hmm update the range, need validate again\n" ); - r = -EAGAIN ; - goto unlock_out; - } -if (!list_empty (&prange->child_list)) { -pr_debug ("range split by unmap in parallel, validate again\n" ); - r = -EAGAIN ; - goto unlock_out; - } + next =min (vma->vm_end, end); + npages = (next - addr) /PAGE_SIZE ; + r =amdgpu_hmm_range_get_pages (&prange->notifier, mm,NULL , + addr, npages, &hmm_range, + readonly, true, owner); +if (r) { +pr_debug ("failed %d to get svm range pages\n" , r); + goto unreserve_out; + } - r =svm_range_map_to_gpus (prange, ctx.bitmap, wait); + offset = (addr - start) /PAGE_SIZE ; + r =svm_range_dma_map (prange, ctx.bitmap, offset, npages, + hmm_range->hmm_pfns); +if (r) { +pr_debug ("failed %d to dma map range\n" , r); + goto unreserve_out; + } + +svm_range_lock (prange); +if (amdgpu_hmm_range_get_pages_done (hmm_range)) { +pr_debug ("hmm update the range, need validate again\n" ); + r = -EAGAIN ; + goto unlock_out; + } +if (!list_empty (&prange->child_list)) { +pr_debug ("range split by unmap in parallel, validate again\n" ); + r = -EAGAIN ; + goto unlock_out; + } + + r =svm_range_map_to_gpus (prange, offset, npages, readonly, + ctx.bitmap, wait); unlock_out: -svm_range_unlock (prange); +svm_range_unlock (prange); + + addr = next; + } + + prange->validated_once = true; + unreserve_out:svm_range_unreserve_bos (&ctx);
Signature · 1 line
2.17.1
Felix Kuehling
maintainerRe: [PATCH 2/2] drm/amdkfd: map SVM range with correct access permission
Am 2021-08-19 um 10:56 a.m. schrieb Philip Yang:
Quoted 8 lines
Restore retry fault or prefetch range, or restore svm range after eviction to map range to GPU with correct read or write access permission.
Range may includes multiple VMAs, update GPU page table with offset of prange, number of pages for each VMA according VMA access permission.
Signed-off-by Philip Yang <[email protected]>
Minor nitpicks, and one question. See inline. It looks good otherwise.
Quoted 165 lines
--- drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 131 +++++++++++++++++---------- 1 file changed, 84 insertions(+), 47 deletions(-)drivers/gpu/drm/amd/amdkfd/kfd_svm.c +35
-33@@ -120,6 +120,7 @@ static void svm_range_remove_notifier(struct svm_range *prange)static int svm_range_dma_map_dev (struct amdgpu_device *adev,struct svm_range *prange, +unsigned long offset,unsigned long npages,unsigned long *hmm_pfns,uint32_t gpuidx) {enum dma_data_direction dir =DMA_BIDIRECTIONAL ;@@ -136,7 +137,8 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange, prange->dma_addr[gpuidx] = addr; } -for (i =0 ; i < prange->npages; i++) { + addr += offset; +for (i =0 ; i < npages; i++) {if (WARN_ONCE (addr[i] && !dma_mapping_error (dev, addr[i]),"leaking dma mapping\n" ))dma_unmap_page (dev, addr[i],PAGE_SIZE , dir);@@ -167,6 +169,7 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange,static int svm_range_dma_map (struct svm_range *prange,unsigned long *bitmap, +unsigned long offset,unsigned long npages,unsigned long *hmm_pfns) {struct kfd_process *p;@@ -187,7 +190,8 @@ svm_range_dma_map(struct svm_range *prange, unsigned long *bitmap, } adev = (struct amdgpu_device *)pdd->dev->kgd; - r =svm_range_dma_map_dev (adev, prange, hmm_pfns, gpuidx); + r =svm_range_dma_map_dev (adev, prange, offset, npages, + hmm_pfns, gpuidx);if (r)break ; }@@ -1088,11 +1092,6 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange, pte_flags |= snoop ?AMDGPU_PTE_SNOOPED :0 ; pte_flags |=amdgpu_gem_va_map_flags (adev, mapping_flags); - -pr_debug ("svms 0x%p [0x%lx 0x%lx] vram %d PTE 0x%llx mapping 0x%x\n" , - prange->svms, prange->start, prange->last, - (domain ==SVM_RANGE_VRAM_DOMAIN ) ?1 :0 , pte_flags, mapping_flags); -return pte_flags; }@@ -1156,7 +1155,8 @@ svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start,static int svm_range_map_to_gpu (struct amdgpu_device *adev,struct amdgpu_vm *vm, -struct svm_range *prange,dma_addr_t *dma_addr, +struct svm_range *prange,unsigned long offset, +unsigned long npages,bool readonly,dma_addr_t *dma_addr,struct amdgpu_device *bo_adev,struct dma_fence **fence) {struct amdgpu_bo_va bo_va;@@ -1167,14 +1167,15 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,int r =0 ;int64_t i; -pr_debug ("svms 0x%p [0x%lx 0x%lx]\n" , prange->svms, prange->start, - prange->last); + last_start = prange->start + offset; + +pr_debug ("svms 0x%p [0x%lx 0x%lx] readonly %d\n" , prange->svms, + last_start, last_start + npages -1 , readonly);if (prange->svm_bo && prange->ttm_res) bo_va.is_xgmi =amdgpu_xgmi_same_hive (adev, bo_adev); - last_start = prange->start; -for (i =0 ; i < prange->npages; i++) { +for (i = offset; i < offset + npages; i++) { last_domain = dma_addr[i] &SVM_RANGE_VRAM_DOMAIN ; dma_addr[i] &= ~SVM_RANGE_VRAM_DOMAIN ; if ((prange->start + i) < prange->last &&@@ -1183,13 +1184,21 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,pr_debug ("Mapping range [0x%lx 0x%llx] on domain: %s\n" , last_start, prange->start + i, last_domain ?"GPU" :"CPU" ); + pte_flags =svm_range_get_pte_flags (adev, prange, last_domain); - r =amdgpu_vm_bo_update_mapping (adev, bo_adev, vm, false, false,NULL , - last_start, +if (readonly) + pte_flags &= ~AMDGPU_PTE_WRITEABLE ; + +pr_debug ("svms 0x%p map [0x%lx 0x%llx] vram %d PTE 0x%llx\n" , + prange->svms, last_start, prange->start + i, + (last_domain ==SVM_RANGE_VRAM_DOMAIN ) ?1 :0 , + pte_flags); + + r =amdgpu_vm_bo_update_mapping (adev, bo_adev, vm, false, false, +NULL , last_start, prange->start + i, pte_flags, last_start - prange->start, -NULL , - dma_addr, +NULL , dma_addr, &vm->last_update, &table_freed);if (r) {@@ -1220,8 +1229,10 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,return r; } -static int svm_range_map_to_gpus (struct svm_range *prange, -unsigned long *bitmap,bool wait) +static int +svm_range_map_to_gpus (struct svm_range *prange,unsigned long offset, +unsigned long npages,bool readonly, +unsigned long *bitmap,bool wait) {struct kfd_process_device *pdd;struct amdgpu_device *bo_adev;@@ -1257,7 +1268,8 @@ static int svm_range_map_to_gpus(struct svm_range *prange, } r =svm_range_map_to_gpu (adev,drm_priv_to_vm (pdd->drm_priv), - prange, prange->dma_addr[gpuidx], + prange, offset, npages, readonly, + prange->dma_addr[gpuidx], bo_adev, wait ? &fence :NULL );if (r)break ;@@ -1390,6 +1402,7 @@ static int svm_range_validate_and_map(struct mm_struct *mm, int32_t gpuidx, bool intr, bool wait) {struct svm_validate_context ctx; +unsigned long start, end, addr;struct hmm_range *hmm_range;struct kfd_process *p;void *owner;@@ -1448,40 +1461,64 @@ static int svm_range_validate_and_map(struct mm_struct *mm,break ; } } - r =amdgpu_hmm_range_get_pages (&prange->notifier, mm,NULL , - prange->start <<PAGE_SHIFT , - prange->npages, &hmm_range, - false, true, owner); -if (r) { -pr_debug ("failed %d to get svm range pages\n" , r); - goto unreserve_out; - } - r =svm_range_dma_map (prange, ctx.bitmap, - hmm_range->hmm_pfns); -if (r) { -pr_debug ("failed %d to dma map range\n" , r); - goto unreserve_out; - } + start = prange->start <<PAGE_SHIFT ; + end = (prange->last +1 ) <<PAGE_SHIFT ; +for (addr = start; addr < end && !r; ) {
drivers/gpu/drm/amd/amdkfd/kfd_svm.c static int svm_range_validate_and_map(struct mm_struct *mm, · line 1476
Quoted 10 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +8
-1+struct vm_area_struct *vma; +unsigned long next; +unsigned long offset; +unsigned long npages; +bool readonly; - prange->validated_once = true; + vma =find_vma (mm, addr); +if (!vma || addr < vma->vm_start) { + r = -EINVAL ;
I think -EFAULT would be the appropriate error code here.
Quoted 7 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +3
-3+ goto unreserve_out; + } + readonly = !(vma->vm_flags &VM_WRITE ); -svm_range_lock (prange); -if (amdgpu_hmm_range_get_pages_done (hmm_range)) { -pr_debug ("hmm update the range, need validate again\n" );
drivers/gpu/drm/amd/amdkfd/kfd_svm.c static int svm_range_validate_and_map(struct mm_struct *mm, · line 1482
Quoted 10 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +2
-8- r = -EAGAIN ; - goto unlock_out; - } -if (!list_empty (&prange->child_list)) { -pr_debug ("range split by unmap in parallel, validate again\n" ); - r = -EAGAIN ; - goto unlock_out; - } + next =min (vma->vm_end, end); + npages = (next - addr) /PAGE_SIZE ;
Use >> PAGE_SHIFT for consistency.
Quoted 34 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +8
-1+ r =amdgpu_hmm_range_get_pages (&prange->notifier, mm,NULL , + addr, npages, &hmm_range, + readonly, true, owner); +if (r) { +pr_debug ("failed %d to get svm range pages\n" , r); + goto unreserve_out; + } - r =svm_range_map_to_gpus (prange, ctx.bitmap, wait); + offset = (addr - start) /PAGE_SIZE ;Quoted 1 line
PAGE_SHIFTdrivers/gpu/drm/amd/amdkfd/kfd_svm.c +20
-0+ r =svm_range_dma_map (prange, ctx.bitmap, offset, npages, + hmm_range->hmm_pfns); +if (r) { +pr_debug ("failed %d to dma map range\n" , r); + goto unreserve_out; + } + +svm_range_lock (prange); +if (amdgpu_hmm_range_get_pages_done (hmm_range)) { +pr_debug ("hmm update the range, need validate again\n" ); + r = -EAGAIN ; + goto unlock_out; + } +if (!list_empty (&prange->child_list)) { +pr_debug ("range split by unmap in parallel, validate again\n" ); + r = -EAGAIN ; + goto unlock_out; + } + + r = svm_range_map_to_gpus(prange, offset, npages, readonly,
drivers/gpu/drm/amd/amdkfd/kfd_svm.c static int svm_range_validate_and_map(struct mm_struct *mm, · line 1520
Quoted 10 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +7
-1+ ctx.bitmap, wait); unlock_out: -svm_range_unlock (prange); +svm_range_unlock (prange); + + addr = next; + } + + prange->validated_once = true;
Should this be conditional on "!r"?
Regards,
Felix
Quoted 4 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +1
-0+ unreserve_out:svm_range_unreserve_bos (&ctx);
Re: [PATCH 2/2] drm/amdkfd: map SVM range with correct access permission
On 2021-08-19 5:32 p.m., Felix Kuehling wrote:
Am 2021-08-19 um 10:56 a.m. schrieb Philip Yang: Restore retry fault or prefetch range, or restore svm range after eviction to map range to GPU with correct read or write access permission. Range may includes multiple VMAs, update GPU page table with offset of prange, number of pages for each VMA according VMA access permission. Signed-off-by: Philip Yang <[email protected]> Minor nitpicks, and one question. See inline. It looks good otherwise. --- drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 131 +++++++++++++++++---------- 1 file changed, 84 insertions(+), 47 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index cf1009bb532a..94612581963f 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -120,6 +120,7 @@ static void svm_range_remove_notifier(struct svm_range *prange) static int svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange, + unsigned long offset, unsigned long npages, unsigned long *hmm_pfns, uint32_t gpuidx) { enum dma_data_direction dir = DMA_BIDIRECTIONAL; @@ -136,7 +137,8 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange, prange->dma_addr[gpuidx] = addr; } - for (i = 0; i < prange->npages; i++) { + addr += offset; + for (i = 0; i < npages; i++) { if (WARN_ONCE(addr[i] && !dma_mapping_error(dev, addr[i]), "leaking dma mapping\n")) dma_unmap_page(dev, addr[i], PAGE_SIZE, dir); @@ -167,6 +169,7 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange, static int svm_range_dma_map(struct svm_range *prange, unsigned long *bitmap, + unsigned long offset, unsigned long npages, unsigned long *hmm_pfns) { struct kfd_process *p; @@ -187,7 +190,8 @@ svm_range_dma_map(struct svm_range *prange, unsigned long *bitmap, } adev = (struct amdgpu_device *)pdd->dev->kgd; - r = svm_range_dma_map_dev(adev, prange, hmm_pfns, gpuidx); + r = svm_range_dma_map_dev(adev, prange, offset, npages, + hmm_pfns, gpuidx); if (r) break; } @@ -1088,11 +1092,6 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange, pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0; pte_flags |= amdgpu_gem_va_map_flags(adev, mapping_flags); - - pr_debug("svms 0x%p [0x%lx 0x%lx] vram %d PTE 0x%llx mapping 0x%x\n", - prange->svms, prange->start, prange->last, - (domain == SVM_RANGE_VRAM_DOMAIN) ? 1:0, pte_flags, mapping_flags); - return pte_flags; } @@ -1156,7 +1155,8 @@ svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start, static int svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm, - struct svm_range *prange, dma_addr_t *dma_addr, + struct svm_range *prange, unsigned long offset, + unsigned long npages, bool readonly, dma_addr_t *dma_addr, struct amdgpu_device *bo_adev, struct dma_fence **fence) { struct amdgpu_bo_va bo_va; @@ -1167,14 +1167,15 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm, int r = 0; int64_t i; - pr_debug("svms 0x%p [0x%lx 0x%lx]\n", prange->svms, prange->start, - prange->last); + last_start = prange->start + offset; + + pr_debug("svms 0x%p [0x%lx 0x%lx] readonly %d\n", prange->svms, + last_start, last_start + npages - 1, readonly); if (prange->svm_bo && prange->ttm_res) bo_va.is_xgmi = amdgpu_xgmi_same_hive(adev, bo_adev); - last_start = prange->start; - for (i = 0; i < prange->npages; i++) { + for (i = offset; i < offset + npages; i++) { last_domain = dma_addr[i] & SVM_RANGE_VRAM_DOMAIN; dma_addr[i] &= ~SVM_RANGE_VRAM_DOMAIN; if ((prange->start + i) < prange->last && @@ -1183,13 +1184,21 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm, pr_debug("Mapping range [0x%lx 0x%llx] on domain: %s\n", last_start, prange->start + i, last_domain ? "GPU" : "CPU"); + pte_flags = svm_range_get_pte_flags(adev, prange, last_domain); - r = amdgpu_vm_bo_update_mapping(adev, bo_adev, vm, false, false, NULL, - last_start, + if (readonly) + pte_flags &= ~AMDGPU_PTE_WRITEABLE; + + pr_debug("svms 0x%p map [0x%lx 0x%llx] vram %d PTE 0x%llx\n", + prange->svms, last_start, prange->start + i, + (last_domain == SVM_RANGE_VRAM_DOMAIN) ? 1 : 0, + pte_flags); + + r = amdgpu_vm_bo_update_mapping(adev, bo_adev, vm, false, false, + NULL, last_start, prange->start + i, pte_flags, last_start - prange->start, - NULL, - dma_addr, + NULL, dma_addr, &vm->last_update, &table_freed); if (r) { @@ -1220,8 +1229,10 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm, return r; } -static int svm_range_map_to_gpus(struct svm_range *prange, - unsigned long *bitmap, bool wait) +static int +svm_range_map_to_gpus(struct svm_range *prange, unsigned long offset, + unsigned long npages, bool readonly, + unsigned long *bitmap, bool wait) { struct kfd_process_device *pdd; struct amdgpu_device *bo_adev; @@ -1257,7 +1268,8 @@ static int svm_range_map_to_gpus(struct svm_range *prange, } r = svm_range_map_to_gpu(adev, drm_priv_to_vm(pdd->drm_priv), - prange, prange->dma_addr[gpuidx], + prange, offset, npages, readonly, + prange->dma_addr[gpuidx], bo_adev, wait ? &fence : NULL); if (r) break; @@ -1390,6 +1402,7 @@ static int svm_range_validate_and_map(struct mm_struct *mm, int32_t gpuidx, bool intr, bool wait) { struct svm_validate_context ctx; + unsigned long start, end, addr; struct hmm_range *hmm_range; struct kfd_process *p; void *owner; @@ -1448,40 +1461,64 @@ static int svm_range_validate_and_map(struct mm_struct *mm, break; } } - r = amdgpu_hmm_range_get_pages(&prange->notifier, mm, NULL, - prange->start << PAGE_SHIFT, - prange->npages, &hmm_range, - false, true, owner); - if (r) { - pr_debug("failed %d to get svm range pages\n", r); - goto unreserve_out; - } - r = svm_range_dma_map(prange, ctx.bitmap, - hmm_range->hmm_pfns); - if (r) { - pr_debug("failed %d to dma map range\n", r); - goto unreserve_out; - } + start = prange->start << PAGE_SHIFT; + end = (prange->last + 1) << PAGE_SHIFT; + for (addr = start; addr < end && !r; ) { + struct vm_area_struct *vma; + unsigned long next; + unsigned long offset; + unsigned long npages; + bool readonly; - prange->validated_once = true; + vma = find_vma(mm, addr); + if (!vma || addr < vma->vm_start) { + r = -EINVAL; I think -EFAULT would be the appropriate error code here. Yes, this error code will pass to user space, -EFAULT means bad address, is appropriate error code here.
+ goto unreserve_out; + } + readonly = !(vma->vm_flags & VM_WRITE); - svm_range_lock(prange); - if (amdgpu_hmm_range_get_pages_done(hmm_range)) { - pr_debug("hmm update the range, need validate again\n"); - r = -EAGAIN; - goto unlock_out; - } - if (!list_empty(&prange->child_list)) { - pr_debug("range split by unmap in parallel, validate again\n"); - r = -EAGAIN; - goto unlock_out; - } + next = min(vma->vm_end, end); + npages = (next - addr) / PAGE_SIZE; Use >> PAGE_SHIFT for consistency. + r = amdgpu_hmm_range_get_pages(&prange->notifier, mm, NULL, + addr, npages, &hmm_range, + readonly, true, owner); + if (r) { + pr_debug("failed %d to get svm range pages\n", r); + goto unreserve_out; + } - r = svm_range_map_to_gpus(prange, ctx.bitmap, wait); + offset = (addr - start) / PAGE_SIZE; PAGE_SHIFT done.
+ r = svm_range_dma_map(prange, ctx.bitmap, offset, npages, + hmm_range->hmm_pfns); + if (r) { + pr_debug("failed %d to dma map range\n", r); + goto unreserve_out; + } + + svm_range_lock(prange); + if (amdgpu_hmm_range_get_pages_done(hmm_range)) { + pr_debug("hmm update the range, need validate again\n"); + r = -EAGAIN; + goto unlock_out; + } + if (!list_empty(&prange->child_list)) { + pr_debug("range split by unmap in parallel, validate again\n"); + r = -EAGAIN; + goto unlock_out; + } + + r = svm_range_map_to_gpus(prange, offset, npages, readonly, + ctx.bitmap, wait); unlock_out: - svm_range_unlock(prange); + svm_range_unlock(prange); + + addr = next; + } + + prange->validated_once = true; Should this be conditional on "!r"? Add if (addr == end) condition, to ensure all pages of range are validated once.
Regards,
Philip
Regards, Felix + unreserve_out: svm_range_unreserve_bos(&ctx);
Felix Kuehling
maintainerAm 2021-08-19 um 10:56 a.m. schrieb Philip Yang:
Quoted 6 lines
Check range access permission to restore GPU retry fault, if GPU retry fault on address which belongs to VMA, and VMA has no read or write permission requested by GPU, failed to restore the address. The vm fault event will pass back to user space.
Signed-off-by Philip Yang <[email protected]>
Just some nit-picks. Other than that the patch looks good to me. See inline ...
Quoted 15 lines
--- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 5 +++-- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 2 +- drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 3 ++- drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 3 ++- drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 30 +++++++++++++++++++++++++- drivers/gpu/drm/amd/amdkfd/kfd_svm.h | 5 +++-- 6 files changed, 40 insertions(+), 8 deletions(-)drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +0
-0@@ -3347,12 +3347,13 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)* @adev: amdgpu device pointer
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c void amdgpu_vm_set_task_info(struct amdgpu_vm *vm) · line 3356
Quoted 10 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +2
-1* @pasid: PASID of the VM * @addr: Address of the fault +* @rw_fault: 0 is read fault, 1 is write fault * * Try to gracefully handle a VM fault. Return true if the fault was handled and * shouldn't be reported any more. */ bool amdgpu_vm_handle_fault (struct amdgpu_device *adev,u32 pasid, -uint64_t addr) +uint64_t addr,uint32_t rw_fault)
Should rw_fault be a bool? And maybe call it write_fault to clarify what "true" means.
Quoted 17 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +1
-1{bool is_compute_context = false;struct amdgpu_bo *root;@@ -3377,7 +3378,7 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, addr /=AMDGPU_GPU_PAGE_SIZE ;if (is_compute_context && - !svm_range_restore_pages (adev, pasid, addr)) { + !svm_range_restore_pages (adev, pasid, addr, rw_fault)) { amdgpu_bo_unref(&root)return true; }drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +0
-0
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev); · line 451
Quoted 5 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +1
-1@@ -448,7 +448,7 @@ void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev);void amdgpu_vm_get_task_info (struct amdgpu_device *adev,u32 pasid,struct amdgpu_task_info *task_info);bool amdgpu_vm_handle_fault (struct amdgpu_device *adev,u32 pasid, -uint64_t addr); +uint64_t addr,uint32_t rw_fault);
bool write_fault
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, · line 96
Quoted 12 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +0
-0void amdgpu_vm_set_task_info (struct amdgpu_vm *vm);drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +1
-0@@ -93,6 +93,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) {bool retry_fault = !!(entry->src_data[1 ] &0x80 ); +bool rw_fault = !!(entry->src_data[1 ] &0x20 );
write_fault
Quoted 17 lines
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +1
-1struct amdgpu_vmhub *hub = &adev->vmhub[entry->vmid_src];struct amdgpu_task_info task_info;uint32_t status =0 ;@@ -121,7 +122,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,/* Try to handle the recoverable page faults by filling page * tables */ -if (amdgpu_vm_handle_fault (adev, entry->pasid, addr)) +if (amdgpu_vm_handle_fault (adev, entry->pasid, addr, rw_fault))return 1 ; }drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +0
-0
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, · line 509
Quoted 4 lines
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +1
-0@@ -506,6 +506,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) {bool retry_fault = !!(entry->src_data[1 ] &0x80 ); +bool rw_fault = !!(entry->src_data[1 ] &0x20 );
write_fault
Quoted 17 lines
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +1
-1uint32_t status =0 , cid =0 , rw =0 ;struct amdgpu_task_info task_info;struct amdgpu_vmhub *hub;@@ -536,7 +537,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,/* Try to handle the recoverable page faults by filling page * tables */ -if (amdgpu_vm_handle_fault (adev, entry->pasid, addr)) +if (amdgpu_vm_handle_fault (adev, entry->pasid, addr, rw_fault))return 1 ; }drivers/gpu/drm/amd/amdkfd/kfd_svm.c +0
-0
drivers/gpu/drm/amd/amdkfd/kfd_svm.c svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p, · line 2404
Quoted 5 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +2
-0@@ -2400,9 +2400,29 @@ svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p,WRITE_ONCE (pdd->faults, pdd->faults +1 ); } +static bool +svm_range_allow_access (struct mm_struct *mm,uint64_t addr,uint32_t rw_fault)
I think the function name "svm_range_..." is a bit misleading because it doesn't do anything for an address range. It only checks one VMA at one virtual address. I'd suggest "svm_fault_allowed".
Quoted 6 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +6
-0+{ +unsigned long requested =VM_READ ; +struct vm_area_struct *vma; + +if (rw_fault) + requested |=VM_WRITE ;
drivers/gpu/drm/amd/amdkfd/kfd_svm.c svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p, · line 2420
Quoted 10 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +10
-0+ + vma =find_vma (mm, addr <<PAGE_SHIFT ); +if (!vma || (addr <<PAGE_SHIFT ) < vma->vm_start) { +pr_debug ("address 0x%llx VMA is removed\n" , addr); +return true; + } + +pr_debug ("requested 0x%lx, vma permission flags 0x%lx\n" , requested, + vma->vm_flags); +return (requested & ~vma->vm_flags) ==0 ;
I think this is the same as "(vma->vm_flags & requested) == requested".
drivers/gpu/drm/amd/amdkfd/kfd_svm.c svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p, · line 2425
Quoted 6 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +3
-1+} +int svm_range_restore_pages (struct amdgpu_device *adev,unsigned int pasid, -uint64_t addr) +uint64_t addr,uint32_t rw_fault)
bool write_fault
Regards,
Felix
Quoted 48 lines
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +8
-0{struct mm_struct *mm =NULL ;struct svm_range_list *svms;@@ -2440,6 +2460,7 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, }mmap_read_lock (mm); + retry_write_locked:mutex_lock (&svms->lock); prange =svm_range_from_addr (svms, addr,NULL );@@ -2484,6 +2505,13 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, goto out_unlock_range; } +if (!svm_range_allow_access (mm, addr, rw_fault)) { +pr_debug ("fault addr 0x%llx no %s permission\n" , addr, + rw_fault ?"write" :"read" ); + r = -EPERM ; + goto out_unlock_range; + } + best_loc =svm_range_best_restore_location (prange, adev, &gpuidx);if (best_loc ==-1 ) { pr_debug("svms %p failed get best restore loc [0x%lx 0x%lx]\n",drivers/gpu/drm/amd/amdkfd/kfd_svm.h +3
-2@@ -175,7 +175,7 @@ int svm_range_split_by_granularity(struct kfd_process *p, struct mm_struct *mm, unsigned long addr, struct svm_range *parent, struct svm_range *prange);int svm_range_restore_pages (struct amdgpu_device *adev, -unsigned int pasid,uint64_t addr); +unsigned int pasid,uint64_t addr,uint32_t rw);int svm_range_schedule_evict_svm_bo (struct amdgpu_amdkfd_fence *fence); void svm_range_add_list_work(struct svm_range_list *svms, struct svm_range *prange, struct mm_struct *mm,@@ -210,7 +210,8 @@ static inline void svm_range_list_fini(struct kfd_process *p) }static inline int svm_range_restore_pages (struct amdgpu_device *adev, -unsigned int pasid,uint64_t addr) +unsigned int pasid,uint64_t addr, +uint32_t rw_fault) {return -EFAULT ; }