Update xe_sched_job_arm to return a fence that the caller can use to install dma-resv fences, signal syncobjs, or wait for job completion. The caller is responsible for dropping the fence reference when it is finished with it.
Update xe_sched_job_push to take a reference to the job owned by the scheduler. The caller is now responsible for dropping the initial job creation reference. These semantics better align with DRM dep semantics. Signed-off-by: Matthew Brost <[email protected]> --- drivers/gpu/drm/xe/tests/xe_migrate.c | 10 ++++++---- drivers/gpu/drm/xe/xe_exec.c | 12 ++++++------ drivers/gpu/drm/xe/xe_gsc.c | 5 +++-- drivers/gpu/drm/xe/xe_gsc_submit.c | 5 +++-- drivers/gpu/drm/xe/xe_gt.c | 5 +++-- drivers/gpu/drm/xe/xe_migrate.c | 25 +++++++++++++++---------- drivers/gpu/drm/xe/xe_oa.c | 5 +++-- drivers/gpu/drm/xe/xe_pxp_submit.c | 10 ++++++---- drivers/gpu/drm/xe/xe_sched_job.c | 5 +++-- drivers/gpu/drm/xe/xe_sched_job.h | 2 +- 10 files changed, 49 insertions(+), 35 deletions(-) diff --git a/drivers/gpu/drm/xe/tests/xe_migrate.c b/drivers/gpu/drm/xe/tests/xe_migrate.c index 34e2f0f4631f..257b373fcaec 100644 --- a/drivers/gpu/drm/xe/tests/xe_migrate.c +++ b/drivers/gpu/drm/xe/tests/xe_migrate.c @@ -50,9 +50,10 @@ static int run_sanity_job(struct xe_migrate *m, struct xe_device *xe, return PTR_ERR(job); } - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); if (sanity_fence_failed(xe, fence, str, test)) return -ETIMEDOUT; @@ -463,10 +464,11 @@ static struct dma_fence *blt_copy(struct xe_tile *tile, xe_sched_job_add_migrate_flush(job, flush_flags); mutex_lock(&m->job_mutex); - xe_sched_job_arm(job); dma_fence_put(fence); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); dma_fence_put(m->fence); m->fence = dma_fence_get(fence); diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index e05dabfcd43c..0c0ddf500981 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -124,6 +124,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) struct xe_validation_ctx ctx; struct xe_sched_job *job; struct xe_vm *vm; + struct dma_fence *fence; bool write_locked; int err = 0; struct xe_hw_engine_group *group; @@ -320,19 +321,19 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) * Point of no return, if we error after this point just set an error on * the job and let the DRM scheduler / backend clean up the job. */ - xe_sched_job_arm(job); + fence = xe_sched_job_arm(job); if (!xe_vm_in_lr_mode(vm)) - drm_gpuvm_resv_add_fence(&vm->gpuvm, exec, &job->drm.s_fence->finished, + drm_gpuvm_resv_add_fence(&vm->gpuvm, exec, fence, DMA_RESV_USAGE_BOOKKEEP, DMA_RESV_USAGE_BOOKKEEP); for (i = 0; i < num_syncs; i++) { - xe_sync_entry_signal(&syncs[i], &job->drm.s_fence->finished); + xe_sync_entry_signal(&syncs[i], fence); xe_sched_job_init_user_fence(job, &syncs[i]); } if (!xe_vm_in_lr_mode(vm)) - xe_exec_queue_last_fence_set(q, vm, &job->drm.s_fence->finished); + xe_exec_queue_last_fence_set(q, vm, fence); xe_sched_job_push(job); xe_vm_reactivate_rebind(vm); @@ -349,8 +350,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) if (!xe_vm_in_lr_mode(vm)) xe_svm_notifier_unlock(vm); err_put_job: - if (err) - xe_sched_job_put(job); + xe_sched_job_put(job); err_exec: if (!xe_vm_in_lr_mode(vm)) xe_validation_ctx_fini(&ctx); diff --git a/drivers/gpu/drm/xe/xe_gsc.c b/drivers/gpu/drm/xe/xe_gsc.c index e5c234f3d795..10168c27fc22 100644 --- a/drivers/gpu/drm/xe/xe_gsc.c +++ b/drivers/gpu/drm/xe/xe_gsc.c @@ -92,9 +92,10 @@ static int emit_gsc_upload(struct xe_gsc *gsc) return PTR_ERR(job); } - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); timeout = dma_fence_wait_timeout(fence, false, HZ); dma_fence_put(fence); diff --git a/drivers/gpu/drm/xe/xe_gsc_submit.c b/drivers/gpu/drm/xe/xe_gsc_submit.c index 08082b596501..2d506fb66017 100644 --- a/drivers/gpu/drm/xe/xe_gsc_submit.c +++ b/drivers/gpu/drm/xe/xe_gsc_submit.c @@ -202,9 +202,10 @@ int xe_gsc_pkt_submit_kernel(struct xe_gsc *gsc, u64 addr_in, u32 size_in, return PTR_ERR(job); } - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); timeout = dma_fence_wait_timeout(fence, false, HZ); dma_fence_put(fence); diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c index 79f4128fe325..10885ba7c2d2 100644 --- a/drivers/gpu/drm/xe/xe_gt.c +++ b/drivers/gpu/drm/xe/xe_gt.c @@ -182,9 +182,10 @@ static int emit_job_sync(struct xe_exec_queue *q, struct xe_bb *bb, if (IS_ERR(job)) return PTR_ERR(job); - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); timeout = dma_fence_wait_timeout(fence, false, timeout_jiffies); dma_fence_put(fence); diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index fc918b4fba54..519f7c70abfb 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -1022,10 +1022,11 @@ static struct dma_fence *__xe_migrate_copy(struct xe_migrate *m, } mutex_lock(&m->job_mutex); - xe_sched_job_arm(job); dma_fence_put(fence); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); dma_fence_put(m->fence); m->fence = dma_fence_get(fence); @@ -1438,10 +1439,11 @@ struct dma_fence *xe_migrate_vram_copy_chunk(struct xe_bo *vram_bo, u64 vram_off DMA_RESV_USAGE_BOOKKEEP)); scoped_guard(mutex, &m->job_mutex) { - xe_sched_job_arm(job); dma_fence_put(fence); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); dma_fence_put(m->fence); m->fence = dma_fence_get(fence); @@ -1676,10 +1678,11 @@ struct dma_fence *xe_migrate_clear(struct xe_migrate *m, } mutex_lock(&m->job_mutex); - xe_sched_job_arm(job); dma_fence_put(fence); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); dma_fence_put(m->fence); m->fence = dma_fence_get(fence); @@ -1996,9 +1999,10 @@ __xe_migrate_update_pgtables(struct xe_migrate *m, if (is_migrate) mutex_lock(&m->job_mutex); - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); if (is_migrate) mutex_unlock(&m->job_mutex); @@ -2282,9 +2286,10 @@ static struct dma_fence *xe_migrate_vram(struct xe_migrate *m, } mutex_lock(&m->job_mutex); - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); dma_fence_put(m->fence); m->fence = dma_fence_get(fence); diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c index c176a61febb2..8de8952a2ca7 100644 --- a/drivers/gpu/drm/xe/xe_oa.c +++ b/drivers/gpu/drm/xe/xe_oa.c @@ -653,9 +653,10 @@ static struct dma_fence *xe_oa_submit_bb(struct xe_oa_stream *stream, enum xe_oa } } - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); xe_oa_unlock_vma(q); diff --git a/drivers/gpu/drm/xe/xe_pxp_submit.c b/drivers/gpu/drm/xe/xe_pxp_submit.c index e60526e30030..5af384acae82 100644 --- a/drivers/gpu/drm/xe/xe_pxp_submit.c +++ b/drivers/gpu/drm/xe/xe_pxp_submit.c @@ -315,9 +315,10 @@ int xe_pxp_submit_session_termination(struct xe_pxp *pxp, u32 id) if (IS_ERR(job)) return PTR_ERR(job); - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); timeout = dma_fence_wait_timeout(fence, false, HZ); @@ -378,9 +379,10 @@ static int pxp_pkt_submit(struct xe_exec_queue *q, u64 batch_addr) if (IS_ERR(job)) return PTR_ERR(job); - xe_sched_job_arm(job); - fence = dma_fence_get(&job->drm.s_fence->finished); + fence = xe_sched_job_arm(job); + dma_fence_get(fence); xe_sched_job_push(job); + xe_sched_job_put(job); timeout = dma_fence_wait_timeout(fence, false, HZ); dma_fence_put(fence); diff --git a/drivers/gpu/drm/xe/xe_sched_job.c b/drivers/gpu/drm/xe/xe_sched_job.c index ae5b38b2a884..99f11bb4d2b9 100644 --- a/drivers/gpu/drm/xe/xe_sched_job.c +++ b/drivers/gpu/drm/xe/xe_sched_job.c @@ -243,7 +243,7 @@ bool xe_sched_job_completed(struct xe_sched_job *job) xe_lrc_seqno(lrc)); } -void xe_sched_job_arm(struct xe_sched_job *job) +struct dma_fence *xe_sched_job_arm(struct xe_sched_job *job) { struct xe_exec_queue *q = job->q; struct dma_fence *fence, *prev; @@ -288,6 +288,8 @@ void xe_sched_job_arm(struct xe_sched_job *job) job->fence = dma_fence_get(fence); /* Pairs with put in scheduler */ drm_sched_job_arm(&job->drm); + + return &job->drm.s_fence->finished; } void xe_sched_job_push(struct xe_sched_job *job) @@ -295,7 +297,6 @@ void xe_sched_job_push(struct xe_sched_job *job) xe_sched_job_get(job); trace_xe_sched_job_exec(job); drm_sched_entity_push_job(&job->drm); - xe_sched_job_put(job); } /** diff --git a/drivers/gpu/drm/xe/xe_sched_job.h b/drivers/gpu/drm/xe/xe_sched_job.h index 1c1cb44216c3..a39cc4ab980b 100644 --- a/drivers/gpu/drm/xe/xe_sched_job.h +++ b/drivers/gpu/drm/xe/xe_sched_job.h @@ -55,7 +55,7 @@ static inline bool xe_sched_job_is_error(struct xe_sched_job *job) bool xe_sched_job_started(struct xe_sched_job *job); bool xe_sched_job_completed(struct xe_sched_job *job); -void xe_sched_job_arm(struct xe_sched_job *job); +struct dma_fence *xe_sched_job_arm(struct xe_sched_job *job); void xe_sched_job_push(struct xe_sched_job *job); void xe_sched_job_init_user_fence(struct xe_sched_job *job, -- 2.34.1
