From: Felix Kuehling <Felix.Kuehling@amd.com> To: amd-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org Subject: [PATCH v2 03/10] drm/amdgpu: Keep a bo-reference per-attachment Date: Wed, 21 Apr 2021 21:30:51 -0400 [thread overview] Message-ID: <20210422013058.6305-4-Felix.Kuehling@amd.com> (raw) In-Reply-To: <20210422013058.6305-1-Felix.Kuehling@amd.com> For now they all reference the same BO. For correct DMA mappings they will refer to different BOs per-GPU. Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> --- .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 22 ++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index fee4c64dd051..34c9a2d0028e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -489,11 +489,11 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, struct amdgpu_vm *vm, bool is_aql, struct kfd_mem_attachment **p_attachment) { - int ret; - struct kfd_mem_attachment *attachment; - struct amdgpu_bo *bo = mem->bo; + unsigned long bo_size = mem->bo->tbo.base.size; uint64_t va = mem->va; - unsigned long bo_size = bo->tbo.base.size; + struct kfd_mem_attachment *attachment; + struct amdgpu_bo *bo; + int ret; if (!va) { pr_err("Invalid VA when adding BO to VM\n"); @@ -510,6 +510,14 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, pr_debug("\t add VA 0x%llx - 0x%llx to vm %p\n", va, va + bo_size, vm); + /* FIXME: For now all attachments use the same BO. This is incorrect + * because one BO can only have one DMA mapping for one GPU. We need + * one BO per GPU, e.g. a DMABuf import with dynamic attachment. This + * will be addressed one BO-type at a time in subsequent patches. + */ + bo = mem->bo; + drm_gem_object_get(&bo->tbo.base); + /* Add BO to VM internal data structures*/ attachment->bo_va = amdgpu_vm_bo_add(adev, vm, bo); if (!attachment->bo_va) { @@ -529,7 +537,7 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, /* Allocate validate page tables if needed */ ret = vm_validate_pt_pd_bos(vm); - if (ret) { + if (unlikely(ret)) { pr_err("validate_pt_pd_bos() failed\n"); goto err_alloc_pts; } @@ -540,15 +548,19 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, amdgpu_vm_bo_rmv(adev, attachment->bo_va); list_del(&attachment->list); err_vmadd: + drm_gem_object_put(&bo->tbo.base); kfree(attachment); return ret; } static void kfd_mem_detach(struct kfd_mem_attachment *attachment) { + struct amdgpu_bo *bo = attachment->bo_va->base.bo; + pr_debug("\t remove VA 0x%llx in entry %p\n", attachment->va, attachment); amdgpu_vm_bo_rmv(attachment->adev, attachment->bo_va); + drm_gem_object_put(&bo->tbo.base); list_del(&attachment->list); kfree(attachment); } -- 2.31.1 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
WARNING: multiple messages have this Message-ID (diff)
From: Felix Kuehling <Felix.Kuehling@amd.com> To: amd-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org Subject: [PATCH v2 03/10] drm/amdgpu: Keep a bo-reference per-attachment Date: Wed, 21 Apr 2021 21:30:51 -0400 [thread overview] Message-ID: <20210422013058.6305-4-Felix.Kuehling@amd.com> (raw) In-Reply-To: <20210422013058.6305-1-Felix.Kuehling@amd.com> For now they all reference the same BO. For correct DMA mappings they will refer to different BOs per-GPU. Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> --- .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 22 ++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index fee4c64dd051..34c9a2d0028e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -489,11 +489,11 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, struct amdgpu_vm *vm, bool is_aql, struct kfd_mem_attachment **p_attachment) { - int ret; - struct kfd_mem_attachment *attachment; - struct amdgpu_bo *bo = mem->bo; + unsigned long bo_size = mem->bo->tbo.base.size; uint64_t va = mem->va; - unsigned long bo_size = bo->tbo.base.size; + struct kfd_mem_attachment *attachment; + struct amdgpu_bo *bo; + int ret; if (!va) { pr_err("Invalid VA when adding BO to VM\n"); @@ -510,6 +510,14 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, pr_debug("\t add VA 0x%llx - 0x%llx to vm %p\n", va, va + bo_size, vm); + /* FIXME: For now all attachments use the same BO. This is incorrect + * because one BO can only have one DMA mapping for one GPU. We need + * one BO per GPU, e.g. a DMABuf import with dynamic attachment. This + * will be addressed one BO-type at a time in subsequent patches. + */ + bo = mem->bo; + drm_gem_object_get(&bo->tbo.base); + /* Add BO to VM internal data structures*/ attachment->bo_va = amdgpu_vm_bo_add(adev, vm, bo); if (!attachment->bo_va) { @@ -529,7 +537,7 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, /* Allocate validate page tables if needed */ ret = vm_validate_pt_pd_bos(vm); - if (ret) { + if (unlikely(ret)) { pr_err("validate_pt_pd_bos() failed\n"); goto err_alloc_pts; } @@ -540,15 +548,19 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, amdgpu_vm_bo_rmv(adev, attachment->bo_va); list_del(&attachment->list); err_vmadd: + drm_gem_object_put(&bo->tbo.base); kfree(attachment); return ret; } static void kfd_mem_detach(struct kfd_mem_attachment *attachment) { + struct amdgpu_bo *bo = attachment->bo_va->base.bo; + pr_debug("\t remove VA 0x%llx in entry %p\n", attachment->va, attachment); amdgpu_vm_bo_rmv(attachment->adev, attachment->bo_va); + drm_gem_object_put(&bo->tbo.base); list_del(&attachment->list); kfree(attachment); } -- 2.31.1 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx
next prev parent reply other threads:[~2021-04-22 1:31 UTC|newest] Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-04-22 1:30 [PATCH v2 00/10] Implement multi-GPU DMA mappings for KFD Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-04-22 1:30 ` [PATCH v2 01/10] rock-dbg_defconfig: Enable Intel IOMMU Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-04-22 1:30 ` [PATCH v2 02/10] drm/amdgpu: Rename kfd_bo_va_list to kfd_mem_attachment Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-05-10 22:00 ` Errabolu, Ramesh 2021-05-10 22:00 ` Errabolu, Ramesh 2021-04-22 1:30 ` Felix Kuehling [this message] 2021-04-22 1:30 ` [PATCH v2 03/10] drm/amdgpu: Keep a bo-reference per-attachment Felix Kuehling 2021-05-10 22:00 ` Errabolu, Ramesh 2021-05-10 22:00 ` Errabolu, Ramesh 2021-04-22 1:30 ` [PATCH v2 04/10] drm/amdgpu: Simplify AQL queue mapping Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-04-23 1:33 ` Zeng, Oak 2021-04-23 1:33 ` Zeng, Oak 2021-04-23 7:23 ` Felix Kuehling 2021-04-23 7:23 ` Felix Kuehling 2021-05-10 22:03 ` Errabolu, Ramesh 2021-05-10 22:03 ` Errabolu, Ramesh 2021-04-22 1:30 ` [PATCH v2 05/10] drm/amdgpu: Add multi-GPU DMA mapping helpers Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-04-27 0:09 ` Zeng, Oak 2021-04-27 0:09 ` Zeng, Oak 2021-04-27 3:41 ` Felix Kuehling 2021-04-27 3:41 ` Felix Kuehling 2021-05-10 22:05 ` Errabolu, Ramesh 2021-05-10 22:05 ` Errabolu, Ramesh 2021-04-22 1:30 ` [PATCH v2 06/10] drm/amdgpu: DMA map/unmap when updating GPU mappings Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-04-27 0:23 ` Zeng, Oak 2021-04-27 0:23 ` Zeng, Oak 2021-04-27 3:47 ` Felix Kuehling 2021-04-27 3:47 ` Felix Kuehling 2021-05-10 22:06 ` Errabolu, Ramesh 2021-05-10 22:06 ` Errabolu, Ramesh 2021-04-22 1:30 ` [PATCH v2 07/10] drm/amdgpu: Move kfd_mem_attach outside reservation Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-05-10 22:06 ` Errabolu, Ramesh 2021-05-10 22:06 ` Errabolu, Ramesh 2021-04-22 1:30 ` [PATCH v2 08/10] drm/amdgpu: Add DMA mapping of GTT BOs Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-04-27 0:35 ` Zeng, Oak 2021-04-27 0:35 ` Zeng, Oak 2021-04-27 3:56 ` Felix Kuehling 2021-04-27 3:56 ` Felix Kuehling 2021-04-27 14:29 ` Zeng, Oak 2021-04-27 14:29 ` Zeng, Oak 2021-04-27 15:08 ` Felix Kuehling 2021-04-27 15:08 ` Felix Kuehling 2021-05-10 22:07 ` Errabolu, Ramesh 2021-05-10 22:07 ` Errabolu, Ramesh 2021-04-22 1:30 ` [PATCH v2 09/10] drm/ttm: Don't count pages in SG BOs against pages_limit Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-05-10 22:08 ` Errabolu, Ramesh 2021-05-10 22:08 ` Errabolu, Ramesh 2021-04-22 1:30 ` [PATCH v2 10/10] drm/amdgpu: Move dmabuf attach/detach to backend_(un)bind Felix Kuehling 2021-04-22 1:30 ` Felix Kuehling 2021-04-22 11:20 ` Christian König 2021-04-22 11:20 ` Christian König 2021-05-10 22:09 ` Errabolu, Ramesh 2021-05-10 22:09 ` Errabolu, Ramesh 2021-04-27 15:16 ` [PATCH v2 00/10] Implement multi-GPU DMA mappings for KFD Zeng, Oak 2021-04-27 15:16 ` Zeng, Oak
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20210422013058.6305-4-Felix.Kuehling@amd.com \ --to=felix.kuehling@amd.com \ --cc=amd-gfx@lists.freedesktop.org \ --cc=dri-devel@lists.freedesktop.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.