[Intel-gfx] [PATCH v2 07/12] drm/i915: Extract "emit write" part of emit breadcrumb functions

[Michał Winiarski]

Let's separate the "emit" part from touching any internal structures,
this way we can have a generic "emit coherent GGTT write" function.
We would like to reuse this functionality for emitting HWSP write, to
confirm that preempt-to-idle has finished.

v2: Reorder args to match emit_pipe_control, s/render/rcs (Chris)

Signed-off-by: Michał Winiarski <michal.winiarski at intel.com>
Cc: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/intel_lrc.c        | 28 +++++++-----------------
 drivers/gpu/drm/i915/intel_ringbuffer.h | 38 +++++++++++++++++++++++++++++++++
 2 files changed, 46 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index e821c1eba7a4..599c709fc5a7 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -1794,10 +1794,8 @@ static void gen8_emit_breadcrumb(struct drm_i915_gem_request *request, u32 *cs)
 	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
 	BUILD_BUG_ON(I915_GEM_HWS_INDEX_ADDR & (1 << 5));
 
-	*cs++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW;
-	*cs++ = intel_hws_seqno_address(request->engine) | MI_FLUSH_DW_USE_GTT;
-	*cs++ = 0;
-	*cs++ = request->global_seqno;
+	cs = gen8_emit_ggtt_write(cs, request->global_seqno,
+				  intel_hws_seqno_address(request->engine));
 	*cs++ = MI_USER_INTERRUPT;
 	*cs++ = MI_NOOP;
 	request->tail = intel_ring_offset(request, cs);
@@ -1807,24 +1805,14 @@ static void gen8_emit_breadcrumb(struct drm_i915_gem_request *request, u32 *cs)
 }
 static const int gen8_emit_breadcrumb_sz = 6 + WA_TAIL_DWORDS;
 
-static void gen8_emit_breadcrumb_render(struct drm_i915_gem_request *request,
+static void gen8_emit_breadcrumb_rcs(struct drm_i915_gem_request *request,
 					u32 *cs)
 {
 	/* We're using qword write, seqno should be aligned to 8 bytes. */
 	BUILD_BUG_ON(I915_GEM_HWS_INDEX & 1);
 
-	/* w/a for post sync ops following a GPGPU operation we
-	 * need a prior CS_STALL, which is emitted by the flush
-	 * following the batch.
-	 */
-	*cs++ = GFX_OP_PIPE_CONTROL(6);
-	*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_CS_STALL |
-		PIPE_CONTROL_QW_WRITE;
-	*cs++ = intel_hws_seqno_address(request->engine);
-	*cs++ = 0;
-	*cs++ = request->global_seqno;
-	/* We're thrashing one dword of HWS. */
-	*cs++ = 0;
+	cs = gen8_emit_ggtt_write_rcs(cs, request->global_seqno,
+				      intel_hws_seqno_address(request->engine));
 	*cs++ = MI_USER_INTERRUPT;
 	*cs++ = MI_NOOP;
 	request->tail = intel_ring_offset(request, cs);
@@ -1832,7 +1820,7 @@ static void gen8_emit_breadcrumb_render(struct drm_i915_gem_request *request,
 
 	gen8_emit_wa_tail(request, cs);
 }
-static const int gen8_emit_breadcrumb_render_sz = 8 + WA_TAIL_DWORDS;
+static const int gen8_emit_breadcrumb_rcs_sz = 8 + WA_TAIL_DWORDS;
 
 static int gen8_init_rcs_context(struct drm_i915_gem_request *req)
 {
@@ -1991,8 +1979,8 @@ int logical_render_ring_init(struct intel_engine_cs *engine)
 		engine->init_hw = gen8_init_render_ring;
 	engine->init_context = gen8_init_rcs_context;
 	engine->emit_flush = gen8_emit_flush_render;
-	engine->emit_breadcrumb = gen8_emit_breadcrumb_render;
-	engine->emit_breadcrumb_sz = gen8_emit_breadcrumb_render_sz;
+	engine->emit_breadcrumb = gen8_emit_breadcrumb_rcs;
+	engine->emit_breadcrumb_sz = gen8_emit_breadcrumb_rcs_sz;
 
 	ret = intel_engine_create_scratch(engine, PAGE_SIZE);
 	if (ret)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index e5a0d489a304..c15161e56964 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -869,6 +869,44 @@ static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset)
 	return batch + 6;
 }
 
+static inline u32 *
+gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset)
+{
+	/* We're using qword write, offset should be aligned to 8 bytes. */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	/* w/a for post sync ops following a GPGPU operation we
+	 * need a prior CS_STALL, which is emitted by the flush
+	 * following the batch.
+	 */
+	*cs++ = GFX_OP_PIPE_CONTROL(6);
+	*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_CS_STALL |
+		PIPE_CONTROL_QW_WRITE;
+	*cs++ = gtt_offset;
+	*cs++ = 0;
+	*cs++ = value;
+	/* We're thrashing one dword of HWS. */
+	*cs++ = 0;
+
+	return cs;
+}
+
+static inline u32 *
+gen8_emit_ggtt_write(u32 *cs, u32 value, u32 gtt_offset)
+{
+	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
+	GEM_BUG_ON(gtt_offset & (1 << 5));
+	/* Offset should be aligned to 8 bytes for both (QW/DW) write types */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	*cs++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW;
+	*cs++ = gtt_offset | MI_FLUSH_DW_USE_GTT;
+	*cs++ = 0;
+	*cs++ = value;
+
+	return cs;
+}
+
 bool intel_engine_is_idle(struct intel_engine_cs *engine);
 bool intel_engines_are_idle(struct drm_i915_private *dev_priv);
 
-- 
2.13.6