drm/i915/gvt: Separate cmd scan from request allocation

Currently i915 request structure and shadow ring buffer are allocated
before command scan, so it will have to restore to previous states once
any error happens afterwards in the long dispatch_workload path.

This patch is to introduce a reserved ring buffer created at the beginning
of vGPU initialization. Workload will be coped to this reserved buffer and
be scanned first, the i915 request and shadow ring buffer are only
allocated after the result of scan is successful.

To balance the memory usage and buffer alloc time, the coming bigger ring
buffer will be reallocated and kept until more bigger buffer is coming.

v2:
- use kmalloc for the smaller ring buffer, realloc if required. (Zhenyu)

v3:
- remove the dynamically allocated ring buffer. (Zhenyu)

v4:
- code style polish.
- kfree previous allocated buffer once kmalloc failed. (Zhenyu)

Signed-off-by: fred gao <fred.gao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

1 files changed, 57 insertions, 27 deletions

diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c
index 391800d2067b..0e480f59f659 100644
--- a/drivers/gpu/drm/i915/gvt/scheduler.c
+++ b/drivers/gpu/drm/i915/gvt/scheduler.c
@@ -201,6 +201,34 @@ static void shadow_context_descriptor_update(struct i915_gem_context *ctx,
 	ce->lrc_desc = desc;
 }
 
+static int copy_workload_to_ring_buffer(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	void *shadow_ring_buffer_va;
+	u32 *cs;
+
+	/* allocate shadow ring buffer */
+	cs = intel_ring_begin(workload->req, workload->rb_len / sizeof(u32));
+	if (IS_ERR(cs)) {
+		gvt_vgpu_err("fail to alloc size =%ld shadow  ring buffer\n",
+			workload->rb_len);
+		return PTR_ERR(cs);
+	}
+
+	shadow_ring_buffer_va = workload->shadow_ring_buffer_va;
+
+	/* get shadow ring buffer va */
+	workload->shadow_ring_buffer_va = cs;
+
+	memcpy(cs, shadow_ring_buffer_va,
+			workload->rb_len);
+
+	cs += workload->rb_len / sizeof(u32);
+	intel_ring_advance(workload->req, cs);
+
+	return 0;
+}
+
 /**
  * intel_gvt_scan_and_shadow_workload - audit the workload by scanning and
  * shadow it as well, include ringbuffer,wa_ctx and ctx.
@@ -214,8 +242,10 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 	int ring_id = workload->ring_id;
 	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;
 
 	lockdep_assert_held(&dev_priv->drm.struct_mutex);
@@ -231,17 +261,6 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 		shadow_context_descriptor_update(shadow_ctx,
 					dev_priv->engine[ring_id]);
 
-	rq = i915_gem_request_alloc(dev_priv->engine[ring_id], shadow_ctx);
-	if (IS_ERR(rq)) {
-		gvt_vgpu_err("fail to allocate gem request\n");
-		ret = PTR_ERR(rq);
-		goto out;
-	}
-
-	gvt_dbg_sched("ring id %d get i915 gem request %p\n", ring_id, rq);
-
-	workload->req = i915_gem_request_get(rq);
-
 	ret = intel_gvt_scan_and_shadow_ringbuffer(workload);
 	if (ret)
 		goto out;
@@ -253,10 +272,37 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 			goto out;
 	}
 
+	/* pin shadow context by gvt even the shadow context will be pinned
+	 * when i915 alloc request. That is because gvt will update the guest
+	 * context from shadow context when workload is completed, and at that
+	 * moment, i915 may already unpined the shadow context to make the
+	 * shadow_ctx pages invalid. So gvt need to pin itself. After update
+	 * the guest context, gvt can unpin the shadow_ctx safely.
+	 */
+	ring = engine->context_pin(engine, shadow_ctx);
+	if (IS_ERR(ring)) {
+		ret = PTR_ERR(ring);
+		gvt_vgpu_err("fail to pin shadow context\n");
+		goto out;
+	}
+
 	ret = populate_shadow_context(workload);
 	if (ret)
 		goto out;
 
+	rq = i915_gem_request_alloc(dev_priv->engine[ring_id], shadow_ctx);
+	if (IS_ERR(rq)) {
+		gvt_vgpu_err("fail to allocate gem request\n");
+		ret = PTR_ERR(rq);
+		goto out;
+	}
+
+	gvt_dbg_sched("ring id %d get i915 gem request %p\n", ring_id, rq);
+
+	workload->req = i915_gem_request_get(rq);
+	ret = copy_workload_to_ring_buffer(workload);
+	if (ret)
+		goto out;
 	workload->shadowed = true;
 
 out:
@@ -269,8 +315,6 @@ static int dispatch_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 intel_vgpu *vgpu = workload->vgpu;
-	struct intel_ring *ring;
 	int ret = 0;
 
 	gvt_dbg_sched("ring id %d prepare to dispatch workload %p\n",
@@ -288,20 +332,6 @@ static int dispatch_workload(struct intel_vgpu_workload *workload)
 			goto out;
 	}
 
-	/* pin shadow context by gvt even the shadow context will be pinned
-	 * when i915 alloc request. That is because gvt will update the guest
-	 * context from shadow context when workload is completed, and at that
-	 * moment, i915 may already unpined the shadow context to make the
-	 * shadow_ctx pages invalid. So gvt need to pin itself. After update
-	 * the guest context, gvt can unpin the shadow_ctx safely.
-	 */
-	ring = engine->context_pin(engine, shadow_ctx);
-	if (IS_ERR(ring)) {
-		ret = PTR_ERR(ring);
-		gvt_vgpu_err("fail to pin shadow context\n");
-		goto out;
-	}
-
 out:
 	if (ret)
 		workload->status = ret;