[Intel-gfx] [PATCH 2/4] drm/i915/gem: Use a single chained reloc batches for a single execbuf

[Chris Wilson]

As we can now keep chaining together a relocation batch to process any
number of relocations, we can keep building that relocation batch for
all of the target vma. This avoiding emitting a new request into the
ring for each target, consuming precious ring space and a potential
stall.

Testcase: igt/gem_exec_reloc/basic-wide-active
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 .../gpu/drm/i915/gem/i915_gem_execbuffer.c    | 23 +++++++++++--------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
index 47b1192a159e..9d68d66555b0 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -268,6 +268,7 @@ struct i915_execbuffer {
 		bool has_fence : 1;
 		bool needs_unfenced : 1;
 
+		struct i915_vma *target;
 		struct i915_request *rq;
 		u32 *rq_cmd;
 		unsigned int rq_size;
@@ -1070,9 +1071,6 @@ static void reloc_cache_reset(struct reloc_cache *cache)
 {
 	void *vaddr;
 
-	if (cache->rq)
-		reloc_gpu_flush(cache);
-
 	if (!cache->vaddr)
 		return;
 
@@ -1265,7 +1263,6 @@ static int reloc_move_to_gpu(struct i915_request *rq, struct i915_vma *vma)
 }
 
 static int __reloc_gpu_alloc(struct i915_execbuffer *eb,
-			     struct i915_vma *vma,
 			     unsigned int len)
 {
 	struct reloc_cache *cache = &eb->reloc_cache;
@@ -1288,7 +1285,7 @@ static int __reloc_gpu_alloc(struct i915_execbuffer *eb,
 		goto out_pool;
 	}
 
-	batch = i915_vma_instance(pool->obj, vma->vm, NULL);
+	batch = i915_vma_instance(pool->obj, eb->context->vm, NULL);
 	if (IS_ERR(batch)) {
 		err = PTR_ERR(batch);
 		goto err_unmap;
@@ -1308,10 +1305,6 @@ static int __reloc_gpu_alloc(struct i915_execbuffer *eb,
 	if (err)
 		goto err_request;
 
-	err = reloc_move_to_gpu(rq, vma);
-	if (err)
-		goto err_request;
-
 	err = eb->engine->emit_bb_start(rq,
 					batch->node.start, PAGE_SIZE,
 					cache->gen > 5 ? 0 : I915_DISPATCH_SECURE);
@@ -1361,9 +1354,17 @@ static u32 *reloc_gpu(struct i915_execbuffer *eb,
 		if (!intel_engine_can_store_dword(eb->engine))
 			return ERR_PTR(-ENODEV);
 
-		err = __reloc_gpu_alloc(eb, vma, len);
+		err = __reloc_gpu_alloc(eb, len);
+		if (unlikely(err))
+			return ERR_PTR(err);
+	}
+
+	if (vma != cache->target) {
+		err = reloc_move_to_gpu(cache->rq, vma);
 		if (unlikely(err))
 			return ERR_PTR(err);
+
+		cache->target = vma;
 	}
 
 	if (unlikely(cache->rq_size > PAGE_SIZE / sizeof(u32) - len - 4)) {
@@ -1680,6 +1681,8 @@ static int eb_relocate(struct i915_execbuffer *eb)
 			if (err)
 				return err;
 		}
+
+		reloc_gpu_flush(&eb->reloc_cache);
 	}
 
 	return 0;
-- 
2.20.1