1 files changed, 56 insertions, 39 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 8443cea6821a..34e35423b78e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -74,9 +74,6 @@ int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			ib->gpu_addr = amdgpu_sa_bo_gpu_addr(ib->sa_bo);
 	}
 
-	ib->vm = vm;
-	ib->vm_id = 0;
-
 	return 0;
 }
 
@@ -89,7 +86,8 @@ int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
  *
  * Free an IB (all asics).
  */
-void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib, struct fence *f)
+void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
+		    struct fence *f)
 {
 	amdgpu_sa_bo_free(adev, &ib->sa_bo, f);
 }
@@ -117,28 +115,37 @@ void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib, struct fen
  */
 int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		       struct amdgpu_ib *ibs, struct fence *last_vm_update,
-		       struct fence **f)
+		       struct amdgpu_job *job, struct fence **f)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_ib *ib = &ibs[0];
-	struct amdgpu_ctx *ctx, *old_ctx;
+	bool skip_preamble, need_ctx_switch;
+	unsigned patch_offset = ~0;
 	struct amdgpu_vm *vm;
 	struct fence *hwf;
+	uint64_t ctx;
+
 	unsigned i;
 	int r = 0;
 
 	if (num_ibs == 0)
 		return -EINVAL;
 
-	ctx = ibs->ctx;
-	vm = ibs->vm;
+	/* ring tests don't use a job */
+	if (job) {
+		vm = job->vm;
+		ctx = job->ctx;
+	} else {
+		vm = NULL;
+		ctx = 0;
+	}
 
 	if (!ring->ready) {
 		dev_err(adev->dev, "couldn't schedule ib\n");
 		return -EINVAL;
 	}
 
-	if (vm && !ibs->vm_id) {
+	if (vm && !job->vm_id) {
 		dev_err(adev->dev, "VM IB without ID\n");
 		return -EINVAL;
 	}
@@ -149,58 +156,68 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		return r;
 	}
 
+	if (ring->type == AMDGPU_RING_TYPE_SDMA && ring->funcs->init_cond_exec)
+		patch_offset = amdgpu_ring_init_cond_exec(ring);
+
 	if (vm) {
-		/* do context switch */
-		amdgpu_vm_flush(ring, ib->vm_id, ib->vm_pd_addr,
-				ib->gds_base, ib->gds_size,
-				ib->gws_base, ib->gws_size,
-				ib->oa_base, ib->oa_size);
-
-		if (ring->funcs->emit_hdp_flush)
-			amdgpu_ring_emit_hdp_flush(ring);
+		r = amdgpu_vm_flush(ring, job->vm_id, job->vm_pd_addr,
+				    job->gds_base, job->gds_size,
+				    job->gws_base, job->gws_size,
+				    job->oa_base, job->oa_size);
+		if (r) {
+			amdgpu_ring_undo(ring);
+			return r;
+		}
 	}
 
-	old_ctx = ring->current_ctx;
+	if (ring->funcs->emit_hdp_flush)
+		amdgpu_ring_emit_hdp_flush(ring);
+
+	/* always set cond_exec_polling to CONTINUE */
+	*ring->cond_exe_cpu_addr = 1;
+
+	skip_preamble = ring->current_ctx == ctx;
+	need_ctx_switch = ring->current_ctx != ctx;
 	for (i = 0; i < num_ibs; ++i) {
 		ib = &ibs[i];
 
-		if (ib->ctx != ctx || ib->vm != vm) {
-			ring->current_ctx = old_ctx;
-			if (ib->vm_id)
-				amdgpu_vm_reset_id(adev, ib->vm_id);
-			amdgpu_ring_undo(ring);
-			return -EINVAL;
-		}
-		amdgpu_ring_emit_ib(ring, ib);
-		ring->current_ctx = ctx;
-	}
+		/* drop preamble IBs if we don't have a context switch */
+		if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) && skip_preamble)
+			continue;
 
-	if (vm) {
-		if (ring->funcs->emit_hdp_invalidate)
-			amdgpu_ring_emit_hdp_invalidate(ring);
+		amdgpu_ring_emit_ib(ring, ib, job ? job->vm_id : 0,
+				    need_ctx_switch);
+		need_ctx_switch = false;
 	}
 
+	if (ring->funcs->emit_hdp_invalidate)
+		amdgpu_ring_emit_hdp_invalidate(ring);
+
 	r = amdgpu_fence_emit(ring, &hwf);
 	if (r) {
 		dev_err(adev->dev, "failed to emit fence (%d)\n", r);
-		ring->current_ctx = old_ctx;
-		if (ib->vm_id)
-			amdgpu_vm_reset_id(adev, ib->vm_id);
+		if (job && job->vm_id)
+			amdgpu_vm_reset_id(adev, job->vm_id);
 		amdgpu_ring_undo(ring);
 		return r;
 	}
 
 	/* wrap the last IB with fence */
-	if (ib->user) {
-		uint64_t addr = amdgpu_bo_gpu_offset(ib->user->bo);
-		addr += ib->user->offset;
-		amdgpu_ring_emit_fence(ring, addr, ib->sequence,
+	if (job && job->uf_bo) {
+		uint64_t addr = amdgpu_bo_gpu_offset(job->uf_bo);
+
+		addr += job->uf_offset;
+		amdgpu_ring_emit_fence(ring, addr, job->uf_sequence,
 				       AMDGPU_FENCE_FLAG_64BIT);
 	}
 
 	if (f)
 		*f = fence_get(hwf);
 
+	if (patch_offset != ~0 && ring->funcs->patch_cond_exec)
+		amdgpu_ring_patch_cond_exec(ring, patch_offset);
+
+	ring->current_ctx = ctx;
 	amdgpu_ring_commit(ring);
 	return 0;
 }
@@ -315,7 +332,7 @@ static int amdgpu_debugfs_sa_info(struct seq_file *m, void *data)
 
 }
 
-static struct drm_info_list amdgpu_debugfs_sa_list[] = {
+static const struct drm_info_list amdgpu_debugfs_sa_list[] = {
 	{"amdgpu_sa_info", &amdgpu_debugfs_sa_info, 0, NULL},
 };