From 649fdce23cdf516e69aa8c18f4b44c62127f0e83 Mon Sep 17 00:00:00 2001 From: Chunming Zhou Date: Mon, 15 Oct 2018 16:55:47 +0800 Subject: drm: add flags to drm_syncobj_find_fence MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit flags can be used by driver to decide whether need to block wait submission. Signed-off-by: Chunming Zhou SIgned-off-by: Christian König Reviewed-by: Christian König Link: https://patchwork.kernel.org/patch/10641339/ --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 663043c8f0f5..8e9a65a15875 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -1104,7 +1104,7 @@ static int amdgpu_syncobj_lookup_and_add_to_sync(struct amdgpu_cs_parser *p, { int r; struct dma_fence *fence; - r = drm_syncobj_find_fence(p->filp, handle, 0, &fence); + r = drm_syncobj_find_fence(p->filp, handle, 0, 0, &fence); if (r) return r; -- cgit From ca05359f1e64cf8303ee532e50efe4ab7563d4a9 Mon Sep 17 00:00:00 2001 From: Christian König Date: Wed, 19 Sep 2018 16:12:25 +0200 Subject: dma-buf: allow reserving more than one shared fence slot MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Let's support simultaneous submissions to multiple engines. Signed-off-by: Christian König Reviewed-by: Michel Dänzer Reviewed-by: Junwei Zhang Reviewed-by: Huang Rui Link: https://patchwork.kernel.org/patch/10626149/ --- drivers/dma-buf/reservation.c | 13 ++++++++----- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 ++-- drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c | 2 +- drivers/gpu/drm/i915/i915_vma.c | 2 +- drivers/gpu/drm/msm/msm_gem_submit.c | 3 ++- drivers/gpu/drm/nouveau/nouveau_fence.c | 2 +- drivers/gpu/drm/qxl/qxl_release.c | 2 +- drivers/gpu/drm/radeon/radeon_vm.c | 2 +- drivers/gpu/drm/ttm/ttm_bo.c | 4 ++-- drivers/gpu/drm/ttm/ttm_execbuf_util.c | 4 ++-- drivers/gpu/drm/v3d/v3d_gem.c | 2 +- drivers/gpu/drm/vc4/vc4_gem.c | 2 +- drivers/gpu/drm/vgem/vgem_fence.c | 2 +- include/linux/reservation.h | 3 ++- 16 files changed, 28 insertions(+), 23 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/dma-buf/reservation.c b/drivers/dma-buf/reservation.c index 5825fc336a13..5fb4fd461908 100644 --- a/drivers/dma-buf/reservation.c +++ b/drivers/dma-buf/reservation.c @@ -56,9 +56,10 @@ const char reservation_seqcount_string[] = "reservation_seqcount"; EXPORT_SYMBOL(reservation_seqcount_string); /** - * reservation_object_reserve_shared - Reserve space to add a shared - * fence to a reservation_object. + * reservation_object_reserve_shared - Reserve space to add shared fences to + * a reservation_object. * @obj: reservation object + * @num_fences: number of fences we want to add * * Should be called before reservation_object_add_shared_fence(). Must * be called with obj->lock held. @@ -66,7 +67,8 @@ EXPORT_SYMBOL(reservation_seqcount_string); * RETURNS * Zero for success, or -errno */ -int reservation_object_reserve_shared(struct reservation_object *obj) +int reservation_object_reserve_shared(struct reservation_object *obj, + unsigned int num_fences) { struct reservation_object_list *old, *new; unsigned int i, j, k, max; @@ -74,10 +76,11 @@ int reservation_object_reserve_shared(struct reservation_object *obj) old = reservation_object_get_list(obj); if (old && old->shared_max) { - if (old->shared_count < old->shared_max) + if ((old->shared_count + num_fences) <= old->shared_max) return 0; else - max = old->shared_max * 2; + max = max(old->shared_count + num_fences, + old->shared_max * 2); } else { max = 4; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 8e9a65a15875..35bc8fc3bc70 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -955,7 +955,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p) if (r) return r; - r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv); + r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv, 1); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 904014dc5915..cf768acb51dc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -640,7 +640,7 @@ int amdgpu_bo_backup_to_shadow(struct amdgpu_device *adev, bo_addr = amdgpu_bo_gpu_offset(bo); shadow_addr = amdgpu_bo_gpu_offset(bo->shadow); - r = reservation_object_reserve_shared(bo->tbo.resv); + r = reservation_object_reserve_shared(bo->tbo.resv, 1); if (r) goto err; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 6904d794d60a..bdce05183edb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -772,7 +772,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, ring = container_of(vm->entity.rq->sched, struct amdgpu_ring, sched); - r = reservation_object_reserve_shared(bo->tbo.resv); + r = reservation_object_reserve_shared(bo->tbo.resv, 1); if (r) return r; @@ -1839,7 +1839,7 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, if (r) goto error_free; - r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv); + r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv, 1); if (r) goto error_free; diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c b/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c index 983e67f19e45..30875f8f2933 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c @@ -179,7 +179,7 @@ static int submit_fence_sync(struct etnaviv_gem_submit *submit) struct reservation_object *robj = bo->obj->resv; if (!(bo->flags & ETNA_SUBMIT_BO_WRITE)) { - ret = reservation_object_reserve_shared(robj); + ret = reservation_object_reserve_shared(robj, 1); if (ret) return ret; } diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index 31efc971a3a8..35fce4c88629 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -892,7 +892,7 @@ static void export_fence(struct i915_vma *vma, reservation_object_lock(resv, NULL); if (flags & EXEC_OBJECT_WRITE) reservation_object_add_excl_fence(resv, &rq->fence); - else if (reservation_object_reserve_shared(resv) == 0) + else if (reservation_object_reserve_shared(resv, 1) == 0) reservation_object_add_shared_fence(resv, &rq->fence); reservation_object_unlock(resv); } diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c index 7a7923e6220d..a90aedd6883a 100644 --- a/drivers/gpu/drm/msm/msm_gem_submit.c +++ b/drivers/gpu/drm/msm/msm_gem_submit.c @@ -241,7 +241,8 @@ static int submit_fence_sync(struct msm_gem_submit *submit, bool no_implicit) * strange place to call it. OTOH this is a * convenient can-fail point to hook it in. */ - ret = reservation_object_reserve_shared(msm_obj->resv); + ret = reservation_object_reserve_shared(msm_obj->resv, + 1); if (ret) return ret; } diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c index 99be61ddeb75..d4964f3397a1 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.c +++ b/drivers/gpu/drm/nouveau/nouveau_fence.c @@ -341,7 +341,7 @@ nouveau_fence_sync(struct nouveau_bo *nvbo, struct nouveau_channel *chan, bool e int ret = 0, i; if (!exclusive) { - ret = reservation_object_reserve_shared(resv); + ret = reservation_object_reserve_shared(resv, 1); if (ret) return ret; diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c index e37f0097f744..a8d5457a1af9 100644 --- a/drivers/gpu/drm/qxl/qxl_release.c +++ b/drivers/gpu/drm/qxl/qxl_release.c @@ -234,7 +234,7 @@ static int qxl_release_validate_bo(struct qxl_bo *bo) return ret; } - ret = reservation_object_reserve_shared(bo->tbo.resv); + ret = reservation_object_reserve_shared(bo->tbo.resv, 1); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index 7f1a9c787bd1..fed11ece0de6 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -831,7 +831,7 @@ static int radeon_vm_update_ptes(struct radeon_device *rdev, int r; radeon_sync_resv(rdev, &ib->sync, pt->tbo.resv, true); - r = reservation_object_reserve_shared(pt->tbo.resv); + r = reservation_object_reserve_shared(pt->tbo.resv, 1); if (r) return r; diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 26b889f86670..83b4657ffb10 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -872,7 +872,7 @@ static int ttm_bo_add_move_fence(struct ttm_buffer_object *bo, if (fence) { reservation_object_add_shared_fence(bo->resv, fence); - ret = reservation_object_reserve_shared(bo->resv); + ret = reservation_object_reserve_shared(bo->resv, 1); if (unlikely(ret)) return ret; @@ -977,7 +977,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo, bool has_erestartsys = false; int i, ret; - ret = reservation_object_reserve_shared(bo->resv); + ret = reservation_object_reserve_shared(bo->resv, 1); if (unlikely(ret)) return ret; diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c index e73ae0d22897..e493edb0d3e7 100644 --- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c +++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c @@ -129,7 +129,7 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, if (!entry->shared) continue; - ret = reservation_object_reserve_shared(bo->resv); + ret = reservation_object_reserve_shared(bo->resv, 1); if (!ret) continue; } @@ -151,7 +151,7 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, } if (!ret && entry->shared) - ret = reservation_object_reserve_shared(bo->resv); + ret = reservation_object_reserve_shared(bo->resv, 1); if (unlikely(ret != 0)) { if (ret == -EINTR) diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c index e688369ca82b..b88c96911453 100644 --- a/drivers/gpu/drm/v3d/v3d_gem.c +++ b/drivers/gpu/drm/v3d/v3d_gem.c @@ -305,7 +305,7 @@ retry: for (i = 0; i < exec->bo_count; i++) { bo = to_v3d_bo(&exec->bo[i]->base); - ret = reservation_object_reserve_shared(bo->resv); + ret = reservation_object_reserve_shared(bo->resv, 1); if (ret) { v3d_unlock_bo_reservations(dev, exec, acquire_ctx); return ret; diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c index 251198194c38..41881ce4132d 100644 --- a/drivers/gpu/drm/vc4/vc4_gem.c +++ b/drivers/gpu/drm/vc4/vc4_gem.c @@ -635,7 +635,7 @@ retry: for (i = 0; i < exec->bo_count; i++) { bo = to_vc4_bo(&exec->bo[i]->base); - ret = reservation_object_reserve_shared(bo->resv); + ret = reservation_object_reserve_shared(bo->resv, 1); if (ret) { vc4_unlock_bo_reservations(dev, exec, acquire_ctx); return ret; diff --git a/drivers/gpu/drm/vgem/vgem_fence.c b/drivers/gpu/drm/vgem/vgem_fence.c index e6ee71323a66..c1c420afe2dd 100644 --- a/drivers/gpu/drm/vgem/vgem_fence.c +++ b/drivers/gpu/drm/vgem/vgem_fence.c @@ -180,7 +180,7 @@ int vgem_fence_attach_ioctl(struct drm_device *dev, reservation_object_lock(resv, NULL); if (arg->flags & VGEM_FENCE_WRITE) reservation_object_add_excl_fence(resv, fence); - else if ((ret = reservation_object_reserve_shared(resv)) == 0) + else if ((ret = reservation_object_reserve_shared(resv, 1)) == 0) reservation_object_add_shared_fence(resv, fence); reservation_object_unlock(resv); diff --git a/include/linux/reservation.h b/include/linux/reservation.h index 54cf6773a14c..5ddb0e143721 100644 --- a/include/linux/reservation.h +++ b/include/linux/reservation.h @@ -261,7 +261,8 @@ reservation_object_get_excl_rcu(struct reservation_object *obj) return fence; } -int reservation_object_reserve_shared(struct reservation_object *obj); +int reservation_object_reserve_shared(struct reservation_object *obj, + unsigned int num_fences); void reservation_object_add_shared_fence(struct reservation_object *obj, struct dma_fence *fence); -- cgit From 26efecf9558895a89c2920d258601b4afba10fd0 Mon Sep 17 00:00:00 2001 From: Sharat Masetty Date: Mon, 29 Oct 2018 15:02:28 +0530 Subject: drm/scheduler: Add drm_sched_job_cleanup MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This patch adds a new API to clean up the scheduler job resources. This is primarliy needed in cases the job was created but was not queued to the scheduler queue. Additionally with this change, the layer which creates the scheduler job also gets to free up the job's resources and this entails moving the dma_fence_put(finished_fence) to the drivers ops free handler routines. Signed-off-by: Sharat Masetty Reviewed-by: Christian König Acked-by: Andrey Grodzovsky Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 3 +-- drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 2 ++ drivers/gpu/drm/etnaviv/etnaviv_sched.c | 3 +++ drivers/gpu/drm/scheduler/sched_entity.c | 1 - drivers/gpu/drm/scheduler/sched_main.c | 13 ++++++++++++- drivers/gpu/drm/v3d/v3d_sched.c | 2 ++ include/drm/gpu_scheduler.h | 1 + 7 files changed, 21 insertions(+), 4 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 663043c8f0f5..5d768f95de24 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -1260,8 +1260,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, return 0; error_abort: - dma_fence_put(&job->base.s_fence->finished); - job->base.s_fence = NULL; + drm_sched_job_cleanup(&job->base); amdgpu_mn_unlock(p->mn); error_unlock: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index 755f733bf0d9..e0af44fd6a0c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -112,6 +112,8 @@ static void amdgpu_job_free_cb(struct drm_sched_job *s_job) struct amdgpu_ring *ring = to_amdgpu_ring(s_job->sched); struct amdgpu_job *job = to_amdgpu_job(s_job); + drm_sched_job_cleanup(s_job); + amdgpu_ring_priority_put(ring, s_job->s_priority); dma_fence_put(job->fence); amdgpu_sync_free(&job->sync); diff --git a/drivers/gpu/drm/etnaviv/etnaviv_sched.c b/drivers/gpu/drm/etnaviv/etnaviv_sched.c index f8c5f1e6c5e2..88b2768d91c9 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_sched.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_sched.c @@ -125,6 +125,8 @@ static void etnaviv_sched_free_job(struct drm_sched_job *sched_job) { struct etnaviv_gem_submit *submit = to_etnaviv_submit(sched_job); + drm_sched_job_cleanup(sched_job); + etnaviv_submit_put(submit); } @@ -157,6 +159,7 @@ int etnaviv_sched_push_job(struct drm_sched_entity *sched_entity, submit->out_fence, 0, INT_MAX, GFP_KERNEL); if (submit->out_fence_id < 0) { + drm_sched_job_cleanup(&submit->sched_job); ret = -ENOMEM; goto out_unlock; } diff --git a/drivers/gpu/drm/scheduler/sched_entity.c b/drivers/gpu/drm/scheduler/sched_entity.c index ba54c30a466e..4463d3826ecb 100644 --- a/drivers/gpu/drm/scheduler/sched_entity.c +++ b/drivers/gpu/drm/scheduler/sched_entity.c @@ -211,7 +211,6 @@ static void drm_sched_entity_kill_jobs_cb(struct dma_fence *f, drm_sched_fence_finished(job->s_fence); WARN_ON(job->s_fence->parent); - dma_fence_put(&job->s_fence->finished); job->sched->ops->free_job(job); } diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 6b2fd49334f7..18ebbb05762e 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -233,7 +233,6 @@ static void drm_sched_job_finish(struct work_struct *work) drm_sched_start_timeout(sched); spin_unlock(&sched->job_list_lock); - dma_fence_put(&s_job->s_fence->finished); sched->ops->free_job(s_job); } @@ -440,6 +439,18 @@ int drm_sched_job_init(struct drm_sched_job *job, } EXPORT_SYMBOL(drm_sched_job_init); +/** + * drm_sched_job_cleanup - clean up scheduler job resources + * + * @job: scheduler job to clean up + */ +void drm_sched_job_cleanup(struct drm_sched_job *job) +{ + dma_fence_put(&job->s_fence->finished); + job->s_fence = NULL; +} +EXPORT_SYMBOL(drm_sched_job_cleanup); + /** * drm_sched_ready - is the scheduler ready * diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index 80b641ffc3be..445b2ef03303 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -35,6 +35,8 @@ v3d_job_free(struct drm_sched_job *sched_job) { struct v3d_job *job = to_v3d_job(sched_job); + drm_sched_job_cleanup(sched_job); + v3d_exec_put(job->exec); } diff --git a/include/drm/gpu_scheduler.h b/include/drm/gpu_scheduler.h index 4ae192a21c3f..926379d53484 100644 --- a/include/drm/gpu_scheduler.h +++ b/include/drm/gpu_scheduler.h @@ -296,6 +296,7 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched); int drm_sched_job_init(struct drm_sched_job *job, struct drm_sched_entity *entity, void *owner); +void drm_sched_job_cleanup(struct drm_sched_job *job); void drm_sched_wakeup(struct drm_gpu_scheduler *sched); void drm_sched_hw_job_reset(struct drm_gpu_scheduler *sched, struct drm_sched_job *job); -- cgit From e0519696cc31c111f11e7c67bd663dbb88d6673f Mon Sep 17 00:00:00 2001 From: Samuel Pitoiset Date: Mon, 29 Oct 2018 12:23:06 +0100 Subject: drm/amdgpu: print an error when the parser can't be initialized MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Similar to other error messages, might help for tracking down issues. Signed-off-by: Samuel Pitoiset Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 5d768f95de24..fd12e9162f3c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -1284,7 +1284,7 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) r = amdgpu_cs_parser_init(&parser, data); if (r) { - DRM_ERROR("Failed to initialize parser !\n"); + DRM_ERROR("Failed to initialize parser %d!\n", r); goto out; } -- cgit From 0b258ed1a219a9776e8f6967eb34837ae0332e64 Mon Sep 17 00:00:00 2001 From: Christian König Date: Wed, 14 Nov 2018 14:24:27 +0100 Subject: drm: revert "expand replace_fence to support timeline point v2" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This reverts commit 9a09a42369a4a37a959c051d8e1a1f948c1529a4. The whole interface isn't thought through. Since this function can't fail we actually can't allocate an object to store the sync point. Sorry, I should have taken the lead on this from the very beginning and reviewed it more thoughtfully. Going to propose a new interface as a follow up change. Signed-off-by: Christian König Reviewed-by: Chunming Zhou Link: https://patchwork.freedesktop.org/patch/265580/ --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- drivers/gpu/drm/drm_syncobj.c | 14 ++++++-------- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 2 +- drivers/gpu/drm/v3d/v3d_gem.c | 3 +-- drivers/gpu/drm/vc4/vc4_gem.c | 2 +- include/drm/drm_syncobj.h | 2 +- 6 files changed, 11 insertions(+), 14 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 024dfbd87f11..dc54e9efd910 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -1193,7 +1193,7 @@ static void amdgpu_cs_post_dependencies(struct amdgpu_cs_parser *p) int i; for (i = 0; i < p->num_post_dep_syncobjs; ++i) - drm_syncobj_replace_fence(p->post_dep_syncobjs[i], 0, p->fence); + drm_syncobj_replace_fence(p->post_dep_syncobjs[i], p->fence); } static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c index 5c5ba1f14307..db30a0e89db8 100644 --- a/drivers/gpu/drm/drm_syncobj.c +++ b/drivers/gpu/drm/drm_syncobj.c @@ -140,13 +140,11 @@ void drm_syncobj_remove_callback(struct drm_syncobj *syncobj, /** * drm_syncobj_replace_fence - replace fence in a sync object. * @syncobj: Sync object to replace fence in - * @point: timeline point * @fence: fence to install in sync file. * - * This replaces the fence on a sync object, or a timeline point fence. + * This replaces the fence on a sync object. */ void drm_syncobj_replace_fence(struct drm_syncobj *syncobj, - u64 point, struct dma_fence *fence) { struct dma_fence *old_fence; @@ -184,7 +182,7 @@ static void drm_syncobj_assign_null_handle(struct drm_syncobj *syncobj) { struct dma_fence *fence = dma_fence_get_stub(); - drm_syncobj_replace_fence(syncobj, 0, fence); + drm_syncobj_replace_fence(syncobj, fence); dma_fence_put(fence); } @@ -233,7 +231,7 @@ void drm_syncobj_free(struct kref *kref) struct drm_syncobj *syncobj = container_of(kref, struct drm_syncobj, refcount); - drm_syncobj_replace_fence(syncobj, 0, NULL); + drm_syncobj_replace_fence(syncobj, NULL); kfree(syncobj); } EXPORT_SYMBOL(drm_syncobj_free); @@ -267,7 +265,7 @@ int drm_syncobj_create(struct drm_syncobj **out_syncobj, uint32_t flags, drm_syncobj_assign_null_handle(syncobj); if (fence) - drm_syncobj_replace_fence(syncobj, 0, fence); + drm_syncobj_replace_fence(syncobj, fence); *out_syncobj = syncobj; return 0; @@ -452,7 +450,7 @@ static int drm_syncobj_import_sync_file_fence(struct drm_file *file_private, return -ENOENT; } - drm_syncobj_replace_fence(syncobj, 0, fence); + drm_syncobj_replace_fence(syncobj, fence); dma_fence_put(fence); drm_syncobj_put(syncobj); return 0; @@ -923,7 +921,7 @@ drm_syncobj_reset_ioctl(struct drm_device *dev, void *data, return ret; for (i = 0; i < args->count_handles; i++) - drm_syncobj_replace_fence(syncobjs[i], 0, NULL); + drm_syncobj_replace_fence(syncobjs[i], NULL); drm_syncobj_array_free(syncobjs, args->count_handles); diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index d4fac09095f8..10a4afb4f235 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -2191,7 +2191,7 @@ signal_fence_array(struct i915_execbuffer *eb, if (!(flags & I915_EXEC_FENCE_SIGNAL)) continue; - drm_syncobj_replace_fence(syncobj, 0, fence); + drm_syncobj_replace_fence(syncobj, fence); } } diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c index f565b197cba9..cb99e53f7607 100644 --- a/drivers/gpu/drm/v3d/v3d_gem.c +++ b/drivers/gpu/drm/v3d/v3d_gem.c @@ -609,8 +609,7 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void *data, /* Update the return sync object for the */ sync_out = drm_syncobj_find(file_priv, args->out_sync); if (sync_out) { - drm_syncobj_replace_fence(sync_out, 0, - exec->render_done_fence); + drm_syncobj_replace_fence(sync_out, exec->render_done_fence); drm_syncobj_put(sync_out); } diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c index 41881ce4132d..aea2b8dfec17 100644 --- a/drivers/gpu/drm/vc4/vc4_gem.c +++ b/drivers/gpu/drm/vc4/vc4_gem.c @@ -681,7 +681,7 @@ vc4_queue_submit(struct drm_device *dev, struct vc4_exec_info *exec, exec->fence = &fence->base; if (out_sync) - drm_syncobj_replace_fence(out_sync, 0, exec->fence); + drm_syncobj_replace_fence(out_sync, exec->fence); vc4_update_bo_seqnos(exec, seqno); diff --git a/include/drm/drm_syncobj.h b/include/drm/drm_syncobj.h index 2eda44def639..b1fe921f8e8f 100644 --- a/include/drm/drm_syncobj.h +++ b/include/drm/drm_syncobj.h @@ -131,7 +131,7 @@ drm_syncobj_fence_get(struct drm_syncobj *syncobj) struct drm_syncobj *drm_syncobj_find(struct drm_file *file_private, u32 handle); -void drm_syncobj_replace_fence(struct drm_syncobj *syncobj, u64 point, +void drm_syncobj_replace_fence(struct drm_syncobj *syncobj, struct dma_fence *fence); int drm_syncobj_find_fence(struct drm_file *file_private, u32 handle, u64 point, u64 flags, -- cgit From a9f34c70fd168b164aadffd46bb757ded52e25b9 Mon Sep 17 00:00:00 2001 From: Christian König Date: Wed, 19 Sep 2018 16:25:08 +0200 Subject: drm/ttm: allow reserving more than one shared slot v3 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Let's support simultaneous submissions to multiple engines. v2: rename the field to num_shared and fix up all users v3: rebased Signed-off-by: Christian König Reviewed-by: Michel Dänzer Reviewed-by: Junwei Zhang Reviewed-by: Huang Rui Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 10 +++++----- drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 7 +++++-- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 2 +- drivers/gpu/drm/qxl/qxl_release.c | 2 +- drivers/gpu/drm/radeon/radeon_cs.c | 4 ++-- drivers/gpu/drm/radeon/radeon_gem.c | 2 +- drivers/gpu/drm/radeon/radeon_vm.c | 4 ++-- drivers/gpu/drm/ttm/ttm_execbuf_util.c | 12 +++++++----- drivers/gpu/drm/vmwgfx/vmwgfx_resource.c | 10 +++++----- drivers/gpu/drm/vmwgfx/vmwgfx_validation.c | 2 +- include/drm/ttm/ttm_execbuf_util.h | 4 ++-- 14 files changed, 35 insertions(+), 30 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index f3129b912714..b29ef088fa14 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -535,7 +535,7 @@ static void add_kgd_mem_to_kfd_bo_list(struct kgd_mem *mem, struct amdgpu_bo *bo = mem->bo; INIT_LIST_HEAD(&entry->head); - entry->shared = true; + entry->num_shared = 1; entry->bo = &bo->tbo; mutex_lock(&process_info->lock); if (userptr) @@ -676,7 +676,7 @@ static int reserve_bo_and_vm(struct kgd_mem *mem, ctx->kfd_bo.priority = 0; ctx->kfd_bo.tv.bo = &bo->tbo; - ctx->kfd_bo.tv.shared = true; + ctx->kfd_bo.tv.num_shared = 1; ctx->kfd_bo.user_pages = NULL; list_add(&ctx->kfd_bo.tv.head, &ctx->list); @@ -740,7 +740,7 @@ static int reserve_bo_and_cond_vms(struct kgd_mem *mem, ctx->kfd_bo.priority = 0; ctx->kfd_bo.tv.bo = &bo->tbo; - ctx->kfd_bo.tv.shared = true; + ctx->kfd_bo.tv.num_shared = 1; ctx->kfd_bo.user_pages = NULL; list_add(&ctx->kfd_bo.tv.head, &ctx->list); @@ -1830,7 +1830,7 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) validate_list.head) { list_add_tail(&mem->resv_list.head, &resv_list); mem->resv_list.bo = mem->validate_list.bo; - mem->resv_list.shared = mem->validate_list.shared; + mem->resv_list.num_shared = mem->validate_list.num_shared; } /* Reserve all BOs and page tables for validation */ @@ -2049,7 +2049,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef) list_add_tail(&mem->resv_list.head, &ctx.list); mem->resv_list.bo = mem->validate_list.bo; - mem->resv_list.shared = mem->validate_list.shared; + mem->resv_list.num_shared = mem->validate_list.num_shared; } ret = ttm_eu_reserve_buffers(&ctx.ticket, &ctx.list, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c index 14d2982a47cc..b75d30ee80c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c @@ -118,7 +118,7 @@ int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp, entry->priority = min(info[i].bo_priority, AMDGPU_BO_LIST_MAX_PRIORITY); entry->tv.bo = &bo->tbo; - entry->tv.shared = !bo->prime_shared_count; + entry->tv.num_shared = !bo->prime_shared_count; if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GDS) list->gds_obj = bo; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 024dfbd87f11..a4b2ac541475 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -50,7 +50,7 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p, bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj)); p->uf_entry.priority = 0; p->uf_entry.tv.bo = &bo->tbo; - p->uf_entry.tv.shared = true; + p->uf_entry.tv.num_shared = 1; p->uf_entry.user_pages = NULL; drm_gem_object_put_unlocked(gobj); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c index 5b550706ee76..7e22be7ca68a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c @@ -74,7 +74,7 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, INIT_LIST_HEAD(&list); INIT_LIST_HEAD(&csa_tv.head); csa_tv.bo = &bo->tbo; - csa_tv.shared = true; + csa_tv.num_shared = 1; list_add(&csa_tv.head, &list); amdgpu_vm_get_pd_bo(vm, &list, &pd); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 7b3d1ebda9df..f4f00217546e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -169,7 +169,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj, INIT_LIST_HEAD(&duplicates); tv.bo = &bo->tbo; - tv.shared = true; + tv.num_shared = 1; list_add(&tv.head, &list); amdgpu_vm_get_pd_bo(vm, &list, &vm_pd); @@ -604,7 +604,10 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, return -ENOENT; abo = gem_to_amdgpu_bo(gobj); tv.bo = &abo->tbo; - tv.shared = !!(abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID); + if (abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) + tv.num_shared = 1; + else + tv.num_shared = 0; list_add(&tv.head, &list); } else { gobj = NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 58a2363040dd..39fb5d24cc51 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -617,7 +617,7 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, { entry->priority = 0; entry->tv.bo = &vm->root.base.bo->tbo; - entry->tv.shared = true; + entry->tv.num_shared = 1; entry->user_pages = NULL; list_add(&entry->tv.head, validated); } diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c index 3813ec198900..18030e2be71f 100644 --- a/drivers/gpu/drm/qxl/qxl_release.c +++ b/drivers/gpu/drm/qxl/qxl_release.c @@ -217,7 +217,7 @@ int qxl_release_list_add(struct qxl_release *release, struct qxl_bo *bo) qxl_bo_ref(bo); entry->tv.bo = &bo->tbo; - entry->tv.shared = false; + entry->tv.num_shared = 0; list_add_tail(&entry->tv.head, &release->bos); return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 1ae31dbc61c6..f43305329939 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -178,7 +178,7 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) } p->relocs[i].tv.bo = &p->relocs[i].robj->tbo; - p->relocs[i].tv.shared = !r->write_domain; + p->relocs[i].tv.num_shared = !r->write_domain; radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head, priority); @@ -253,7 +253,7 @@ static int radeon_cs_sync_rings(struct radeon_cs_parser *p) resv = reloc->robj->tbo.resv; r = radeon_sync_resv(p->rdev, &p->ib.sync, resv, - reloc->tv.shared); + reloc->tv.num_shared); if (r) return r; } diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index 27d8e7dd2d06..44617dec8183 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -552,7 +552,7 @@ static void radeon_gem_va_update_vm(struct radeon_device *rdev, INIT_LIST_HEAD(&list); tv.bo = &bo_va->bo->tbo; - tv.shared = true; + tv.num_shared = 1; list_add(&tv.head, &list); vm_bos = radeon_vm_get_bos(rdev, bo_va->vm, &list); diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index a3d2ca07a058..0d374211661c 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -142,7 +142,7 @@ struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, list[0].preferred_domains = RADEON_GEM_DOMAIN_VRAM; list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM; list[0].tv.bo = &vm->page_directory->tbo; - list[0].tv.shared = true; + list[0].tv.num_shared = 1; list[0].tiling_flags = 0; list_add(&list[0].tv.head, head); @@ -154,7 +154,7 @@ struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, list[idx].preferred_domains = RADEON_GEM_DOMAIN_VRAM; list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM; list[idx].tv.bo = &list[idx].robj->tbo; - list[idx].tv.shared = true; + list[idx].tv.num_shared = 1; list[idx].tiling_flags = 0; list_add(&list[idx++].tv.head, head); } diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c index efa005a1c1b7..93860346c426 100644 --- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c +++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c @@ -126,10 +126,11 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, } if (!ret) { - if (!entry->shared) + if (!entry->num_shared) continue; - ret = reservation_object_reserve_shared(bo->resv, 1); + ret = reservation_object_reserve_shared(bo->resv, + entry->num_shared); if (!ret) continue; } @@ -150,8 +151,9 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, } } - if (!ret && entry->shared) - ret = reservation_object_reserve_shared(bo->resv, 1); + if (!ret && entry->num_shared) + ret = reservation_object_reserve_shared(bo->resv, + entry->num_shared); if (unlikely(ret != 0)) { if (ret == -EINTR) @@ -199,7 +201,7 @@ void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket, list_for_each_entry(entry, list, head) { bo = entry->bo; - if (entry->shared) + if (entry->num_shared) reservation_object_add_shared_fence(bo->resv, fence); else reservation_object_add_excl_fence(bo->resv, fence); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index 8a029bade32a..3025bfc001a1 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -85,7 +85,7 @@ static void vmw_resource_release(struct kref *kref) struct ttm_validate_buffer val_buf; val_buf.bo = bo; - val_buf.shared = false; + val_buf.num_shared = 0; res->func->unbind(res, false, &val_buf); } res->backup_dirty = false; @@ -462,7 +462,7 @@ vmw_resource_check_buffer(struct ww_acquire_ctx *ticket, INIT_LIST_HEAD(&val_list); val_buf->bo = ttm_bo_reference(&res->backup->base); - val_buf->shared = false; + val_buf->num_shared = 0; list_add_tail(&val_buf->head, &val_list); ret = ttm_eu_reserve_buffers(ticket, &val_list, interruptible, NULL); if (unlikely(ret != 0)) @@ -565,7 +565,7 @@ static int vmw_resource_do_evict(struct ww_acquire_ctx *ticket, BUG_ON(!func->may_evict); val_buf.bo = NULL; - val_buf.shared = false; + val_buf.num_shared = 0; ret = vmw_resource_check_buffer(ticket, res, interruptible, &val_buf); if (unlikely(ret != 0)) return ret; @@ -614,7 +614,7 @@ int vmw_resource_validate(struct vmw_resource *res, bool intr) return 0; val_buf.bo = NULL; - val_buf.shared = false; + val_buf.num_shared = 0; if (res->backup) val_buf.bo = &res->backup->base; do { @@ -685,7 +685,7 @@ void vmw_resource_unbind_list(struct vmw_buffer_object *vbo) struct vmw_resource *res, *next; struct ttm_validate_buffer val_buf = { .bo = &vbo->base, - .shared = false + .num_shared = 0 }; lockdep_assert_held(&vbo->base.resv->lock.base); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c index 184025fa938e..fef22753f4de 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c @@ -266,7 +266,7 @@ int vmw_validation_add_bo(struct vmw_validation_context *ctx, val_buf->bo = ttm_bo_get_unless_zero(&vbo->base); if (!val_buf->bo) return -ESRCH; - val_buf->shared = false; + val_buf->num_shared = 0; list_add_tail(&val_buf->head, &ctx->bo_list); bo_node->as_mob = as_mob; bo_node->cpu_blit = cpu_blit; diff --git a/include/drm/ttm/ttm_execbuf_util.h b/include/drm/ttm/ttm_execbuf_util.h index b0fdd1980034..621615fa7728 100644 --- a/include/drm/ttm/ttm_execbuf_util.h +++ b/include/drm/ttm/ttm_execbuf_util.h @@ -40,13 +40,13 @@ * * @head: list head for thread-private list. * @bo: refcounted buffer object pointer. - * @shared: should the fence be added shared? + * @num_shared: How many shared fences we want to add. */ struct ttm_validate_buffer { struct list_head head; struct ttm_buffer_object *bo; - bool shared; + unsigned int num_shared; }; /** -- cgit From 049aca4363d8af87cab8d53de5401602db3b9999 Mon Sep 17 00:00:00 2001 From: Christian König Date: Wed, 19 Sep 2018 16:54:35 +0200 Subject: drm/amdgpu: fix using shared fence for exported BOs v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It is perfectly possible that the BO list is created before the BO is exported. While at it clean up setting shared to one instead of true. v2: add comment and simplify logic Signed-off-by: Christian König Reviewed-by: Michel Dänzer Reviewed-by: Huang Rui Acked-by: Junwei Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c | 1 - drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 13 +++++++++++-- 2 files changed, 11 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c index b75d30ee80c6..5c79da8e1150 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c @@ -118,7 +118,6 @@ int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp, entry->priority = min(info[i].bo_priority, AMDGPU_BO_LIST_MAX_PRIORITY); entry->tv.bo = &bo->tbo; - entry->tv.num_shared = !bo->prime_shared_count; if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GDS) list->gds_obj = bo; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index a4b2ac541475..ce58af62d523 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -598,6 +598,9 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, return r; } + amdgpu_bo_list_for_each_entry(e, p->bo_list) + e->tv.num_shared = 1; + amdgpu_bo_list_get_list(p->bo_list, &p->validated); if (p->bo_list->first_userptr != p->bo_list->num_entries) p->mn = amdgpu_mn_get(p->adev, AMDGPU_MN_TYPE_GFX); @@ -717,8 +720,14 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, gws = p->bo_list->gws_obj; oa = p->bo_list->oa_obj; - amdgpu_bo_list_for_each_entry(e, p->bo_list) - e->bo_va = amdgpu_vm_bo_find(vm, ttm_to_amdgpu_bo(e->tv.bo)); + amdgpu_bo_list_for_each_entry(e, p->bo_list) { + struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo); + + /* Make sure we use the exclusive slot for shared BOs */ + if (bo->prime_shared_count) + e->tv.num_shared = 0; + e->bo_va = amdgpu_vm_bo_find(vm, bo); + } if (gds) { p->job->gds_base = amdgpu_bo_gpu_offset(gds) >> PAGE_SHIFT; -- cgit From 0aa7aa24cc11720a05b4492345f0adba8373c226 Mon Sep 17 00:00:00 2001 From: Christian König Date: Fri, 21 Sep 2018 18:09:59 +0200 Subject: drm/amdgpu: always reserve two slots for the VM MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit And drop the now superflous extra reservations. Signed-off-by: Christian König Reviewed-by: Michel Dänzer Reviewed-by: Junwei Zhang Reviewed-by: Huang Rui Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 4 ---- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 15 ++++++--------- 2 files changed, 6 insertions(+), 13 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index ce58af62d523..8cc640abe3de 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -964,10 +964,6 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p) if (r) return r; - r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv, 1); - if (r) - return r; - p->job->vm_pd_addr = amdgpu_gmc_pd_addr(vm->root.base.bo); if (amdgpu_vm_debug) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 39fb5d24cc51..b303ac79c2f6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -617,7 +617,8 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, { entry->priority = 0; entry->tv.bo = &vm->root.base.bo->tbo; - entry->tv.num_shared = 1; + /* One for the VM updates and one for the CS job */ + entry->tv.num_shared = 2; entry->user_pages = NULL; list_add(&entry->tv.head, validated); } @@ -773,10 +774,6 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, ring = container_of(vm->entity.rq->sched, struct amdgpu_ring, sched); - r = reservation_object_reserve_shared(bo->tbo.resv, 1); - if (r) - return r; - r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (r) goto error; @@ -1842,10 +1839,6 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, if (r) goto error_free; - r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv, 1); - if (r) - goto error_free; - r = amdgpu_vm_update_ptes(¶ms, start, last + 1, addr, flags); if (r) goto error_free; @@ -3026,6 +3019,10 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, if (r) goto error_free_root; + r = reservation_object_reserve_shared(root->tbo.resv, 1); + if (r) + goto error_unreserve; + r = amdgpu_vm_clear_bo(adev, vm, root, adev->vm_manager.root_level, vm->pte_support_ats); -- cgit From 07daa8a0784e3085c4c8e282a83dd8c92bf59b53 Mon Sep 17 00:00:00 2001 From: Christian König Date: Mon, 24 Sep 2018 13:35:08 +0200 Subject: drm/amdgpu: always reserve one more shared slot for pipelined BO moves MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This allows us to drop the extra reserve in TTM. Signed-off-by: Christian König Reviewed-by: Michel Dänzer Reviewed-by: Junwei Zhang Reviewed-by: Huang Rui Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 6 ++++-- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 ++-- 2 files changed, 6 insertions(+), 4 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 8cc640abe3de..4a6a1d4a88ea 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -50,7 +50,8 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p, bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj)); p->uf_entry.priority = 0; p->uf_entry.tv.bo = &bo->tbo; - p->uf_entry.tv.num_shared = 1; + /* One for TTM and one for the CS job */ + p->uf_entry.tv.num_shared = 2; p->uf_entry.user_pages = NULL; drm_gem_object_put_unlocked(gobj); @@ -598,8 +599,9 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, return r; } + /* One for TTM and one for the CS job */ amdgpu_bo_list_for_each_entry(e, p->bo_list) - e->tv.num_shared = 1; + e->tv.num_shared = 2; amdgpu_bo_list_get_list(p->bo_list, &p->validated); if (p->bo_list->first_userptr != p->bo_list->num_entries) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index b303ac79c2f6..b095bbbbc302 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -617,8 +617,8 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, { entry->priority = 0; entry->tv.bo = &vm->root.base.bo->tbo; - /* One for the VM updates and one for the CS job */ - entry->tv.num_shared = 2; + /* One for the VM updates, one for TTM and one for the CS job */ + entry->tv.num_shared = 3; entry->user_pages = NULL; list_add(&entry->tv.head, validated); } -- cgit