From 0e4101c11aed803e6e69a83f8cc48160a676c19d Mon Sep 17 00:00:00 2001 From: Andrew Martin Date: Mon, 14 Sep 2026 12:19:10 +0000 Subject: [PATCH] drm/amdkfd: Fix always mapped range mapping to all ACCESS GPUs CWSR ranges use the always mapped flag and only need a mapping on the GPU running the queue. However, svm_range_validate_and_map now updates the mapping on every GPU with the ACCESS attribute, which is all supported GPUs when XNACK is on. After the range is prefetched to VRAM, those GPUs may span different XGMI hives, so the page owner mismatches and is set to NULL. hmm_range_fault then tries to migrate the VRAM pages back to system memory, but svm_migrate_to_ram skips it because migrating pages back to system memory is not allowed while the GPU mapping is being updated, so the CPU page fault is never resolved. Only map to the ACCESS GPUs if the range is not mapped to any GPU and is not in VRAM. Fixes: b96f26104946 ("drm/amdkfd: Unmap svm range from GPU set to no-access") Signed-off-by: Andrew Martin Assisted-by: Claude:claude-opus-5 Acked-by: Alex Deucher Reviewed-by: Philip Yang Signed-off-by: Alex Deucher (cherry picked from commit 093cf8842ace35ce212858d5059fee007ec8e778) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index fa4054d51f60..d7f2ff5c8f02 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -1752,12 +1752,15 @@ static int svm_range_validate_and_map(struct mm_struct *mm, } /* - * If prange with always mapped flag, update mapping on GPUs with - * ACCESS attribute + * If prange with always mapped flag is not mapped or migrated + * yet, update mapping on GPUs with ACCESS attribute. Don't add + * them otherwise, ACCESS GPUs may be on different XGMI hives, + * that would force hmm_range_fault to migrate VRAM back to ram. */ - if (prange->flags & KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED) - bitmap_or(ctx->bitmap, ctx->bitmap, prange->bitmap_access, - MAX_GPU_INSTANCE); + if (bitmap_empty(ctx->bitmap, MAX_GPU_INSTANCE) && + prange->flags & KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED) + bitmap_copy(ctx->bitmap, prange->bitmap_access, + MAX_GPU_INSTANCE); } else { bitmap_or(ctx->bitmap, prange->bitmap_access, prange->bitmap_aip, MAX_GPU_INSTANCE);