drm/amdgpu: rework TDR in scheduler (v2)

Add two callbacks to scheduler to maintain jobs, and invoked for
job timeout calculations. Now TDR measures time gap from
job is processed by hw.

v2:
fix typo

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

3 files changed, 17 insertions, 1 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 9bf72b24495c..ccb28468ece8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -754,6 +754,7 @@ void amdgpu_job_free(struct amdgpu_job *job);
 int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring,
                       struct amd_sched_entity *entity, void *owner,
                       struct fence **f);
+void amdgpu_job_timeout_func(struct work_struct *work);
 
 struct amdgpu_ring {
         struct amdgpu_device            *adev;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 23266b454aec..9025671d21c3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -871,6 +871,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 
         r = amd_sched_job_init(&job->base, &ring->sched,
                                                 &p->ctx->rings[ring->idx].entity,
+                                                amdgpu_job_timeout_func,
                                                 p->filp, &fence);
         if (r) {
                 amdgpu_job_free(job);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index 23468088a995..961cae4a1955 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -34,6 +34,15 @@ static void amdgpu_job_free_handler(struct work_struct *ws)
         kfree(job);
 }
 
+void amdgpu_job_timeout_func(struct work_struct *work)
+{
+        struct amdgpu_job *job = container_of(work, struct amdgpu_job, base.work_tdr.work);
+        DRM_ERROR("ring %s timeout, last signaled seq=%u, last emitted seq=%u\n",
+                                job->base.sched->name,
+                                (uint32_t)atomic_read(&job->ring->fence_drv.last_seq),
+                                job->ring->fence_drv.sync_seq);
+}
+
 int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
                      struct amdgpu_job **job)
 {
@@ -103,7 +112,10 @@ int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring,
         if (!f)
                 return -EINVAL;
 
-        r = amd_sched_job_init(&job->base, &ring->sched, entity, owner, &fence);
+        r = amd_sched_job_init(&job->base, &ring->sched,
+                                                        entity, owner,
+                                                        amdgpu_job_timeout_func,
+                                                        &fence);
         if (r)
                 return r;
 
@@ -180,4 +192,6 @@ err:
 struct amd_sched_backend_ops amdgpu_sched_ops = {
         .dependency = amdgpu_job_dependency,
         .run_job = amdgpu_job_run,
+        .begin_job = amd_sched_job_begin,
+        .finish_job = amd_sched_job_finish,
 };