drm/i915/gvt: Move request alloc to dispatch_workload path only

Previously the performance is improved through the workload auditing
and shadowing ahead of vGPU scheduling, however, there is the case that
more requests are allocated in submit_context before the previous request
is added, the timeline will hold its seqno which is later.

This patch is to move the request alloc to dispatch_workload function,
where is the same place as request is added.

It will fix the issue of kernel BUG for (timeline->seqno != request->fence.seqno)
check when add_request.

Fixes: 89ea20b930cb ("drm/i915/gvt: Factor out scan and shadow from workload dispatch")
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Signed-off-by: fred gao <fred.gao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
(cherry picked from commit f2880e04f3a5419366926182fc97a3c2e4fd8f2a)

3 files changed, 29 insertions, 4 deletions

diff --git a/drivers/gpu/drm/i915/gvt/execlist.c b/drivers/gpu/drm/i915/gvt/execlist.c
index 4427be18e4a9..940cdaaa3f24 100644
--- a/drivers/gpu/drm/i915/gvt/execlist.c
+++ b/drivers/gpu/drm/i915/gvt/execlist.c
@@ -496,6 +496,12 @@ static int prepare_execlist_workload(struct intel_vgpu_workload *workload)
 		goto err_unpin_mm;
 	}
 
+	ret = intel_gvt_generate_request(workload);
+	if (ret) {
+		gvt_vgpu_err("fail to generate request\n");
+		goto err_unpin_mm;
+	}
+
 	ret = prepare_shadow_batch_buffer(workload);
 	if (ret) {
 		gvt_vgpu_err("fail to prepare_shadow_batch_buffer\n");
diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c
index f031234290f6..3ac1dc97a7a0 100644
--- a/drivers/gpu/drm/i915/gvt/scheduler.c
+++ b/drivers/gpu/drm/i915/gvt/scheduler.c
@@ -254,7 +254,6 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 	struct i915_gem_context *shadow_ctx = workload->vgpu->shadow_ctx;
 	struct drm_i915_private *dev_priv = workload->vgpu->gvt->dev_priv;
 	struct intel_engine_cs *engine = dev_priv->engine[ring_id];
-	struct drm_i915_gem_request *rq;
 	struct intel_vgpu *vgpu = workload->vgpu;
 	struct intel_ring *ring;
 	int ret;
@@ -300,6 +299,26 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 	ret = populate_shadow_context(workload);
 	if (ret)
 		goto err_unpin;
+	workload->shadowed = true;
+	return 0;
+
+err_unpin:
+	engine->context_unpin(engine, shadow_ctx);
+err_shadow:
+	release_shadow_wa_ctx(&workload->wa_ctx);
+err_scan:
+	return ret;
+}
+
+int intel_gvt_generate_request(struct intel_vgpu_workload *workload)
+{
+	int ring_id = workload->ring_id;
+	struct drm_i915_private *dev_priv = workload->vgpu->gvt->dev_priv;
+	struct intel_engine_cs *engine = dev_priv->engine[ring_id];
+	struct drm_i915_gem_request *rq;
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct i915_gem_context *shadow_ctx = vgpu->shadow_ctx;
+	int ret;
 
 	rq = i915_gem_request_alloc(dev_priv->engine[ring_id], shadow_ctx);
 	if (IS_ERR(rq)) {
@@ -314,14 +333,11 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 	ret = copy_workload_to_ring_buffer(workload);
 	if (ret)
 		goto err_unpin;
-	workload->shadowed = true;
 	return 0;
 
 err_unpin:
 	engine->context_unpin(engine, shadow_ctx);
-err_shadow:
 	release_shadow_wa_ctx(&workload->wa_ctx);
-err_scan:
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/i915/gvt/scheduler.h b/drivers/gpu/drm/i915/gvt/scheduler.h
index 2d694f6c0907..b9f872204d7e 100644
--- a/drivers/gpu/drm/i915/gvt/scheduler.h
+++ b/drivers/gpu/drm/i915/gvt/scheduler.h
@@ -142,4 +142,7 @@ int intel_vgpu_init_gvt_context(struct intel_vgpu *vgpu);
 void intel_vgpu_clean_gvt_context(struct intel_vgpu *vgpu);
 
 void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx);
+
+int intel_gvt_generate_request(struct intel_vgpu_workload *workload);
+
 #endif