drm/i915: Remove obsolete engine->gpu_caches_dirty

Space for flushing the GPU cache prior to completing the request is
preallocated and so cannot fail - the GPU caches will always be flushed
along with the completed request. This means we no longer have to track
whether the GPU cache is dirty between batches like we had to with the
outstanding_lazy_seqno.

With the removal of the duplication in the per-backend entry points for
emitting the obsolete lazy flush, we can then further unify the
engine->emit_flush.

v2: Expand a bit on the legacy of gpu_caches_dirty

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/1469432687-22756-18-git-send-email-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1470174640-18242-7-git-send-email-chris@chris-wilson.co.uk

drivers/gpu/drm/i915/i915_gem_context.c

@@ -568,7 +568,7 @@ mi_set_context(struct drm_i915_gem_request *req, u32 hw_flags)
 	 * itlb_before_ctx_switch.
 	 */
 	if (IS_GEN6(dev_priv)) {
-		ret = engine->flush(req, I915_GEM_GPU_DOMAINS, 0);
+		ret = engine->emit_flush(req, I915_GEM_GPU_DOMAINS, 0);
 		if (ret)
 			return ret;
 	}

drivers/gpu/drm/i915/i915_gem_execbuffer.c

@@ -998,10 +998,8 @@ i915_gem_execbuffer_move_to_gpu(struct drm_i915_gem_request *req,
 	if (flush_domains & I915_GEM_DOMAIN_GTT)
 		wmb();
 
-	/* Unconditionally invalidate gpu caches and ensure that we do flush
-	 * any residual writes from the previous batch.
-	 */
-	return intel_engine_invalidate_all_caches(req);
+	/* Unconditionally invalidate GPU caches and TLBs. */
+	return req->engine->emit_flush(req, I915_GEM_GPU_DOMAINS, 0);
 }
 
 static bool
@@ -1163,9 +1161,6 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 static void
 i915_gem_execbuffer_retire_commands(struct i915_execbuffer_params *params)
 {
-	/* Unconditionally force add_request to emit a full flush. */
-	params->engine->gpu_caches_dirty = true;
-
 	/* Add a breadcrumb for the completion of the batch buffer */
 	__i915_add_request(params->request, params->batch_obj, true);
 }

drivers/gpu/drm/i915/i915_gem_gtt.c

@@ -1666,7 +1666,8 @@ static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
 	int ret;
 
 	/* NB: TLBs must be flushed and invalidated before a switch */
-	ret = engine->flush(req, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
+	ret = engine->emit_flush(req,
+				 I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
 	if (ret)
 		return ret;
 
@@ -1693,7 +1694,8 @@ static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
 	int ret;
 
 	/* NB: TLBs must be flushed and invalidated before a switch */
-	ret = engine->flush(req, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
+	ret = engine->emit_flush(req,
+				 I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
 	if (ret)
 		return ret;
 
@@ -1711,8 +1713,9 @@ static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
 
 	/* XXX: RCS is the only one to auto invalidate the TLBs? */
 	if (engine->id != RCS) {
-		ret = engine->flush(req,
-				    I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
+		ret = engine->emit_flush(req,
+					 I915_GEM_GPU_DOMAINS,
+					 I915_GEM_GPU_DOMAINS);
 		if (ret)
 			return ret;
 	}

drivers/gpu/drm/i915/i915_gem_request.c

@@ -451,12 +451,10 @@ void __i915_add_request(struct drm_i915_gem_request *request,
 	 * what.
 	 */
 	if (flush_caches) {
-		if (i915.enable_execlists)
-			ret = logical_ring_flush_all_caches(request);
-		else
-			ret = intel_engine_flush_all_caches(request);
+		ret = engine->emit_flush(request, 0, I915_GEM_GPU_DOMAINS);
+
 		/* Not allowed to fail! */
-		WARN(ret, "*_ring_flush_all_caches failed: %d!\n", ret);
+		WARN(ret, "engine->emit_flush() failed: %d!\n", ret);
 	}
 
 	trace_i915_gem_request_add(request);

drivers/gpu/drm/i915/intel_lrc.c

@@ -642,24 +642,6 @@ static void execlists_context_queue(struct drm_i915_gem_request *request)
 	spin_unlock_bh(&engine->execlist_lock);
 }
 
-static int logical_ring_invalidate_all_caches(struct drm_i915_gem_request *req)
-{
-	struct intel_engine_cs *engine = req->engine;
-	uint32_t flush_domains;
-	int ret;
-
-	flush_domains = 0;
-	if (engine->gpu_caches_dirty)
-		flush_domains = I915_GEM_GPU_DOMAINS;
-
-	ret = engine->emit_flush(req, I915_GEM_GPU_DOMAINS, flush_domains);
-	if (ret)
-		return ret;
-
-	engine->gpu_caches_dirty = false;
-	return 0;
-}
-
 static int execlists_move_to_gpu(struct drm_i915_gem_request *req,
 				 struct list_head *vmas)
 {
@@ -690,7 +672,7 @@ static int execlists_move_to_gpu(struct drm_i915_gem_request *req,
 	/* Unconditionally invalidate gpu caches and ensure that we do flush
 	 * any residual writes from the previous batch.
 	 */
-	return logical_ring_invalidate_all_caches(req);
+	return req->engine->emit_flush(req, I915_GEM_GPU_DOMAINS, 0);
 }
 
 int intel_logical_ring_alloc_request_extras(struct drm_i915_gem_request *request)
@@ -930,22 +912,6 @@ void intel_logical_ring_stop(struct intel_engine_cs *engine)
 	I915_WRITE_MODE(engine, _MASKED_BIT_DISABLE(STOP_RING));
 }
 
-int logical_ring_flush_all_caches(struct drm_i915_gem_request *req)
-{
-	struct intel_engine_cs *engine = req->engine;
-	int ret;
-
-	if (!engine->gpu_caches_dirty)
-		return 0;
-
-	ret = engine->emit_flush(req, 0, I915_GEM_GPU_DOMAINS);
-	if (ret)
-		return ret;
-
-	engine->gpu_caches_dirty = false;
-	return 0;
-}
-
 static int intel_lr_context_pin(struct i915_gem_context *ctx,
 				struct intel_engine_cs *engine)
 {
@@ -1026,15 +992,15 @@ void intel_lr_context_unpin(struct i915_gem_context *ctx,
 static int intel_logical_ring_workarounds_emit(struct drm_i915_gem_request *req)
 {
 	int ret, i;
-	struct intel_engine_cs *engine = req->engine;
 	struct intel_ring *ring = req->ring;
 	struct i915_workarounds *w = &req->i915->workarounds;
 
 	if (w->count == 0)
 		return 0;
 
-	engine->gpu_caches_dirty = true;
-	ret = logical_ring_flush_all_caches(req);
+	ret = req->engine->emit_flush(req,
+				      I915_GEM_GPU_DOMAINS,
+				      I915_GEM_GPU_DOMAINS);
 	if (ret)
 		return ret;
 
@@ -1051,8 +1017,9 @@ static int intel_logical_ring_workarounds_emit(struct drm_i915_gem_request *req)
 
 	intel_ring_advance(ring);
 
-	engine->gpu_caches_dirty = true;
-	ret = logical_ring_flush_all_caches(req);
+	ret = req->engine->emit_flush(req,
+				      I915_GEM_GPU_DOMAINS,
+				      I915_GEM_GPU_DOMAINS);
 	if (ret)
 		return ret;
 

drivers/gpu/drm/i915/intel_lrc.h

@@ -72,8 +72,6 @@ int logical_xcs_ring_init(struct intel_engine_cs *engine);
 
 int intel_engines_init(struct drm_device *dev);
 
-int logical_ring_flush_all_caches(struct drm_i915_gem_request *req);
-
 /* Logical Ring Contexts */
 
 /* One extra page is added before LRC for GuC as shared data */

drivers/gpu/drm/i915/intel_ringbuffer.c

@@ -688,8 +688,9 @@ static int intel_ring_workarounds_emit(struct drm_i915_gem_request *req)
 	if (w->count == 0)
 		return 0;
 
-	req->engine->gpu_caches_dirty = true;
-	ret = intel_engine_flush_all_caches(req);
+	ret = req->engine->emit_flush(req,
+				      I915_GEM_GPU_DOMAINS,
+				      I915_GEM_GPU_DOMAINS);
 	if (ret)
 		return ret;
 
@@ -706,8 +707,9 @@ static int intel_ring_workarounds_emit(struct drm_i915_gem_request *req)
 
 	intel_ring_advance(ring);
 
-	req->engine->gpu_caches_dirty = true;
-	ret = intel_engine_flush_all_caches(req);
+	ret = req->engine->emit_flush(req,
+				      I915_GEM_GPU_DOMAINS,
+				      I915_GEM_GPU_DOMAINS);
 	if (ret)
 		return ret;
 
@@ -2860,21 +2862,21 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine)
 	if (INTEL_GEN(dev_priv) >= 8) {
 		engine->init_context = intel_rcs_ctx_init;
 		engine->add_request = gen8_render_add_request;
-		engine->flush = gen8_render_ring_flush;
+		engine->emit_flush = gen8_render_ring_flush;
 		if (i915.semaphores)
 			engine->semaphore.signal = gen8_rcs_signal;
 	} else if (INTEL_GEN(dev_priv) >= 6) {
 		engine->init_context = intel_rcs_ctx_init;
-		engine->flush = gen7_render_ring_flush;
+		engine->emit_flush = gen7_render_ring_flush;
 		if (IS_GEN6(dev_priv))
-			engine->flush = gen6_render_ring_flush;
+			engine->emit_flush = gen6_render_ring_flush;
 	} else if (IS_GEN5(dev_priv)) {
-		engine->flush = gen4_render_ring_flush;
+		engine->emit_flush = gen4_render_ring_flush;
 	} else {
 		if (INTEL_GEN(dev_priv) < 4)
-			engine->flush = gen2_render_ring_flush;
+			engine->emit_flush = gen2_render_ring_flush;
 		else
-			engine->flush = gen4_render_ring_flush;
+			engine->emit_flush = gen4_render_ring_flush;
 		engine->irq_enable_mask = I915_USER_INTERRUPT;
 	}
 
@@ -2911,12 +2913,12 @@ int intel_init_bsd_ring_buffer(struct intel_engine_cs *engine)
 		/* gen6 bsd needs a special wa for tail updates */
 		if (IS_GEN6(dev_priv))
 			engine->write_tail = gen6_bsd_ring_write_tail;
-		engine->flush = gen6_bsd_ring_flush;
+		engine->emit_flush = gen6_bsd_ring_flush;
 		if (INTEL_GEN(dev_priv) < 8)
 			engine->irq_enable_mask = GT_BSD_USER_INTERRUPT;
 	} else {
 		engine->mmio_base = BSD_RING_BASE;
-		engine->flush = bsd_ring_flush;
+		engine->emit_flush = bsd_ring_flush;
 		if (IS_GEN5(dev_priv))
 			engine->irq_enable_mask = ILK_BSD_USER_INTERRUPT;
 		else
@@ -2935,7 +2937,7 @@ int intel_init_bsd2_ring_buffer(struct intel_engine_cs *engine)
 
 	intel_ring_default_vfuncs(dev_priv, engine);
 
-	engine->flush = gen6_bsd_ring_flush;
+	engine->emit_flush = gen6_bsd_ring_flush;
 
 	return intel_init_ring_buffer(engine);
 }
@@ -2946,7 +2948,7 @@ int intel_init_blt_ring_buffer(struct intel_engine_cs *engine)
 
 	intel_ring_default_vfuncs(dev_priv, engine);
 
-	engine->flush = gen6_ring_flush;
+	engine->emit_flush = gen6_ring_flush;
 	if (INTEL_GEN(dev_priv) < 8)
 		engine->irq_enable_mask = GT_BLT_USER_INTERRUPT;
 
@@ -2959,7 +2961,7 @@ int intel_init_vebox_ring_buffer(struct intel_engine_cs *engine)
 
 	intel_ring_default_vfuncs(dev_priv, engine);
 
-	engine->flush = gen6_ring_flush;
+	engine->emit_flush = gen6_ring_flush;
 
 	if (INTEL_GEN(dev_priv) < 8) {
 		engine->irq_enable_mask = PM_VEBOX_USER_INTERRUPT;
@@ -2970,46 +2972,6 @@ int intel_init_vebox_ring_buffer(struct intel_engine_cs *engine)
 	return intel_init_ring_buffer(engine);
 }
 
-int
-intel_engine_flush_all_caches(struct drm_i915_gem_request *req)
-{
-	struct intel_engine_cs *engine = req->engine;
-	int ret;
-
-	if (!engine->gpu_caches_dirty)
-		return 0;
-
-	ret = engine->flush(req, 0, I915_GEM_GPU_DOMAINS);
-	if (ret)
-		return ret;
-
-	trace_i915_gem_ring_flush(req, 0, I915_GEM_GPU_DOMAINS);
-
-	engine->gpu_caches_dirty = false;
-	return 0;
-}
-
-int
-intel_engine_invalidate_all_caches(struct drm_i915_gem_request *req)
-{
-	struct intel_engine_cs *engine = req->engine;
-	uint32_t flush_domains;
-	int ret;
-
-	flush_domains = 0;
-	if (engine->gpu_caches_dirty)
-		flush_domains = I915_GEM_GPU_DOMAINS;
-
-	ret = engine->flush(req, I915_GEM_GPU_DOMAINS, flush_domains);
-	if (ret)
-		return ret;
-
-	trace_i915_gem_ring_flush(req, I915_GEM_GPU_DOMAINS, flush_domains);
-
-	engine->gpu_caches_dirty = false;
-	return 0;
-}
-
 void intel_engine_stop(struct intel_engine_cs *engine)
 {
 	int ret;

drivers/gpu/drm/i915/intel_ringbuffer.h

@@ -206,9 +206,6 @@ struct intel_engine_cs {
 
 	void		(*write_tail)(struct intel_engine_cs *engine,
 				      u32 value);
-	int __must_check (*flush)(struct drm_i915_gem_request *req,
-				  u32	invalidate_domains,
-				  u32	flush_domains);
 	int		(*add_request)(struct drm_i915_gem_request *req);
 	/* Some chipsets are not quite as coherent as advertised and need
 	 * an expensive kick to force a true read of the up-to-date seqno.
@@ -325,8 +322,6 @@ struct intel_engine_cs {
 	 */
 	u32 last_submitted_seqno;
 
-	bool gpu_caches_dirty;
-
 	struct i915_gem_context *last_context;
 
 	struct intel_engine_hangcheck hangcheck;
@@ -474,8 +469,6 @@ void intel_ring_update_space(struct intel_ring *ring);
 
 int __must_check intel_engine_idle(struct intel_engine_cs *engine);
 void intel_engine_init_seqno(struct intel_engine_cs *engine, u32 seqno);
-int intel_engine_flush_all_caches(struct drm_i915_gem_request *req);
-int intel_engine_invalidate_all_caches(struct drm_i915_gem_request *req);
 
 int intel_init_pipe_control(struct intel_engine_cs *engine, int size);
 void intel_fini_pipe_control(struct intel_engine_cs *engine);