diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdkfd/kfd_svm.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 38 |
1 files changed, 24 insertions, 14 deletions
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index 069b81eeea03..bd9c2921e0dc 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -1171,7 +1171,6 @@ svm_range_get_pte_flags(struct kfd_node *node, bool snoop = (domain != SVM_RANGE_VRAM_DOMAIN); bool coherent = flags & (KFD_IOCTL_SVM_FLAG_COHERENT | KFD_IOCTL_SVM_FLAG_EXT_COHERENT); bool ext_coherent = flags & KFD_IOCTL_SVM_FLAG_EXT_COHERENT; - bool uncached = false; /*flags & KFD_IOCTL_SVM_FLAG_UNCACHED;*/ unsigned int mtype_local; if (domain == SVM_RANGE_VRAM_DOMAIN) @@ -1213,15 +1212,14 @@ svm_range_get_pte_flags(struct kfd_node *node, } break; case IP_VERSION(9, 4, 3): + case IP_VERSION(9, 4, 4): if (ext_coherent) mtype_local = node->adev->rev_id ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_UC; else mtype_local = amdgpu_mtype_local == 1 ? AMDGPU_VM_MTYPE_NC : amdgpu_mtype_local == 2 ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW; snoop = true; - if (uncached) { - mapping_flags |= AMDGPU_VM_MTYPE_UC; - } else if (domain == SVM_RANGE_VRAM_DOMAIN) { + if (domain == SVM_RANGE_VRAM_DOMAIN) { /* local HBM region close to partition */ if (bo_node->adev == node->adev && (!bo_node->xcp || !node->xcp || bo_node->xcp->mem_id == node->xcp->mem_id)) @@ -1248,6 +1246,16 @@ svm_range_get_pte_flags(struct kfd_node *node, mapping_flags |= AMDGPU_VM_MTYPE_UC; } break; + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + if (domain == SVM_RANGE_VRAM_DOMAIN) { + if (bo_node != node) + mapping_flags |= AMDGPU_VM_MTYPE_NC; + } else { + mapping_flags |= coherent ? + AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC; + } + break; default: mapping_flags |= coherent ? AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC; @@ -1263,6 +1271,8 @@ svm_range_get_pte_flags(struct kfd_node *node, pte_flags = AMDGPU_PTE_VALID; pte_flags |= (domain == SVM_RANGE_VRAM_DOMAIN) ? 0 : AMDGPU_PTE_SYSTEM; pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0; + if (KFD_GC_VERSION(node) >= IP_VERSION(12, 0, 0)) + pte_flags |= AMDGPU_PTE_IS_PTE; pte_flags |= amdgpu_gem_va_map_flags(node->adev, mapping_flags); return pte_flags; @@ -1658,7 +1668,7 @@ static int svm_range_validate_and_map(struct mm_struct *mm, start = map_start << PAGE_SHIFT; end = (map_last + 1) << PAGE_SHIFT; for (addr = start; !r && addr < end; ) { - struct hmm_range *hmm_range; + struct hmm_range *hmm_range = NULL; unsigned long map_start_vma; unsigned long map_last_vma; struct vm_area_struct *vma; @@ -1678,11 +1688,8 @@ static int svm_range_validate_and_map(struct mm_struct *mm, readonly, owner, NULL, &hmm_range); WRITE_ONCE(p->svms.faulting_task, NULL); - if (r) { + if (r) pr_debug("failed %d to get svm range pages\n", r); - if (r == -EBUSY) - r = -EAGAIN; - } } else { r = -EFAULT; } @@ -1696,7 +1703,12 @@ static int svm_range_validate_and_map(struct mm_struct *mm, } svm_range_lock(prange); - if (!r && amdgpu_hmm_range_get_pages_done(hmm_range)) { + + /* Free backing memory of hmm_range if it was initialized + * Overrride return value to TRY AGAIN only if prior returns + * were successful + */ + if (hmm_range && amdgpu_hmm_range_get_pages_done(hmm_range) && !r) { pr_debug("hmm update the range, need validate again\n"); r = -EAGAIN; } @@ -2619,8 +2631,7 @@ svm_range_best_restore_location(struct svm_range *prange, return -1; } - if (node->adev->gmc.is_app_apu || - node->adev->flags & AMD_IS_APU) + if (node->adev->flags & AMD_IS_APU) return 0; if (prange->preferred_loc == gpuid || @@ -3338,8 +3349,7 @@ svm_range_best_prefetch_location(struct svm_range *prange) goto out; } - if (bo_node->adev->gmc.is_app_apu || - bo_node->adev->flags & AMD_IS_APU) { + if (bo_node->adev->flags & AMD_IS_APU) { best_loc = 0; goto out; } |