sna/gen4: Workaround render corruption with multiple SF threads

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=57410
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

2 files changed, 22 insertions, 12 deletions

diff --git a/src/sna/gen4_render.c b/src/sna/gen4_render.c
index fb1c742e..4e8ad578 100644
--- a/src/sna/gen4_render.c
+++ b/src/sna/gen4_render.c
@@ -1256,7 +1256,7 @@ gen4_emit_pipelined_pointers(struct sna *sna,
 	bp = gen4_get_blend(blend, op->has_component_alpha, op->dst.format);
 
 	DBG(("%s: sp=%d, bp=%d\n", __FUNCTION__, sp, bp));
-	key = sp | (uint32_t)bp << 16;
+	key = sp | (uint32_t)bp << 16 | op->u.gen4.sf << 31;
 	if (key == sna->render_state.gen4.last_pipelined_pointers)
 		return;
 
@@ -1264,7 +1264,7 @@ gen4_emit_pipelined_pointers(struct sna *sna,
 	OUT_BATCH(sna->render_state.gen4.vs);
 	OUT_BATCH(GEN4_GS_DISABLE); /* passthrough */
 	OUT_BATCH(GEN4_CLIP_DISABLE); /* passthrough */
-	OUT_BATCH(sna->render_state.gen4.sf[1]);
+	OUT_BATCH(sna->render_state.gen4.sf[op->u.gen4.sf]);
 	OUT_BATCH(sna->render_state.gen4.wm + sp);
 	OUT_BATCH(sna->render_state.gen4.cc + bp);
 
@@ -2339,6 +2339,7 @@ gen4_render_composite(struct sna *sna,
 	tmp->is_affine = tmp->src.is_affine;
 	tmp->has_component_alpha = false;
 	tmp->need_magic_ca_pass = false;
+	tmp->u.gen4.sf = 0;
 
 	tmp->prim_emit = gen4_emit_composite_primitive;
 	if (mask) {
@@ -2391,11 +2392,13 @@ gen4_render_composite(struct sna *sna,
 
 		tmp->floats_per_vertex = 5 + 2 * !tmp->is_affine;
 	} else {
-		if (tmp->src.is_solid)
+		if (tmp->src.is_solid) {
 			tmp->prim_emit = gen4_emit_composite_primitive_solid;
-		else if (tmp->src.transform == NULL)
+		} else if (tmp->src.transform == NULL) {
 			tmp->prim_emit = gen4_emit_composite_primitive_identity_source;
-		else if (tmp->src.is_affine)
+			/* XXX using more then one thread causes corruption? */
+			tmp->u.gen4.sf = 1;
+		} else if (tmp->src.is_affine)
 			tmp->prim_emit = gen4_emit_composite_primitive_affine_source;
 
 		tmp->floats_per_vertex = 3 + !tmp->is_affine;
@@ -2715,6 +2718,7 @@ gen4_render_composite_spans(struct sna *sna,
 
 	tmp->base.u.gen4.wm_kernel = WM_KERNEL_OPACITY | !tmp->base.is_affine;
 	tmp->base.u.gen4.ve_id = 1 << 1 | tmp->base.is_affine;
+	tmp->base.u.gen4.sf = 0;
 
 	tmp->box   = gen4_render_composite_spans_box;
 	tmp->boxes = gen4_render_composite_spans_boxes;
@@ -2925,6 +2929,7 @@ fallback_blt:
 	tmp.floats_per_rect = 9;
 	tmp.u.gen4.wm_kernel = WM_KERNEL;
 	tmp.u.gen4.ve_id = 1;
+	tmp.u.gen4.sf = 0;
 
 	if (!kgem_check_bo(&sna->kgem, dst_bo, src_bo, NULL)) {
 		kgem_submit(&sna->kgem);
@@ -3057,6 +3062,7 @@ fallback:
 	op->base.floats_per_rect = 9;
 	op->base.u.gen4.wm_kernel = WM_KERNEL;
 	op->base.u.gen4.ve_id = 1;
+	op->base.u.gen4.sf = 0;
 
 	if (!kgem_check_bo(&sna->kgem, dst_bo, src_bo, NULL)) {
 		kgem_submit(&sna->kgem);
@@ -3179,6 +3185,7 @@ gen4_render_fill_boxes(struct sna *sna,
 	tmp.floats_per_rect = 9;
 	tmp.u.gen4.wm_kernel = WM_KERNEL;
 	tmp.u.gen4.ve_id = 1;
+	tmp.u.gen4.sf = 0;
 
 	if (!kgem_check_bo(&sna->kgem, dst_bo, NULL)) {
 		kgem_submit(&sna->kgem);
@@ -3284,6 +3291,7 @@ gen4_render_fill(struct sna *sna, uint8_t alu,
 	op->base.floats_per_rect = 9;
 	op->base.u.gen4.wm_kernel = WM_KERNEL;
 	op->base.u.gen4.ve_id = 1;
+	op->base.u.gen4.sf = 0;
 
 	if (!kgem_check_bo(&sna->kgem, dst_bo, NULL)) {
 		kgem_submit(&sna->kgem);
@@ -3363,6 +3371,7 @@ gen4_render_fill_one(struct sna *sna, PixmapPtr dst, struct kgem_bo *bo,
 
 	tmp.u.gen4.wm_kernel = WM_KERNEL;
 	tmp.u.gen4.ve_id = 1;
+	tmp.u.gen4.sf = 0;
 
 	if (!kgem_check_bo(&sna->kgem, bo, NULL)) {
 		_kgem_submit(&sna->kgem);
@@ -3461,7 +3470,8 @@ static uint32_t gen4_create_vs_unit_state(struct sna_static_stream *stream)
 }
 
 static uint32_t gen4_create_sf_state(struct sna_static_stream *stream,
-				     int gen, uint32_t kernel)
+				     int gen, uint32_t kernel,
+				     bool single_thread)
 {
 	struct gen4_sf_unit_state *sf;
 
@@ -3475,7 +3485,7 @@ static uint32_t gen4_create_sf_state(struct sna_static_stream *stream,
 	/* don't smash vertex header, read start from dw8 */
 	sf->thread3.urb_entry_read_offset = 1;
 	sf->thread3.dispatch_grf_start_reg = 3;
-	sf->thread4.max_threads = GEN4_MAX_SF_THREADS - 1;
+	sf->thread4.max_threads = single_thread ? 0 : GEN4_MAX_SF_THREADS - 1;
 	sf->thread4.urb_entry_allocation_size = URB_SF_ENTRY_SIZE - 1;
 	sf->thread4.nr_urb_entries = URB_SF_ENTRIES;
 	sf->sf5.viewport_transform = false;	/* skip viewport */
@@ -3590,7 +3600,7 @@ static bool gen4_render_setup(struct sna *sna)
 	struct gen4_render_state *state = &sna->render_state.gen4;
 	struct sna_static_stream general;
 	struct gen4_wm_unit_state_padded *wm_state;
-	uint32_t sf[2], wm[KERNEL_COUNT];
+	uint32_t sf, wm[KERNEL_COUNT];
 	int i, j, k, l, m;
 
 	sna_static_stream_init(&general);
@@ -3600,8 +3610,7 @@ static bool gen4_render_setup(struct sna *sna)
 	 */
 	null_create(&general);
 
-	sf[0] = sna_static_stream_compile_sf(sna, &general, brw_sf_kernel__nomask);
-	sf[1] = sna_static_stream_compile_sf(sna, &general, brw_sf_kernel__mask);
+	sf = sna_static_stream_compile_sf(sna, &general, brw_sf_kernel__mask);
 	for (m = 0; m < KERNEL_COUNT; m++) {
 		if (wm_kernels[m].size) {
 			wm[m] = sna_static_stream_add(&general,
@@ -3616,8 +3625,8 @@ static bool gen4_render_setup(struct sna *sna)
 	}
 
 	state->vs = gen4_create_vs_unit_state(&general);
-	state->sf[0] = gen4_create_sf_state(&general, sna->kgem.gen, sf[0]);
-	state->sf[1] = gen4_create_sf_state(&general, sna->kgem.gen, sf[1]);
+	state->sf[0] = gen4_create_sf_state(&general, sna->kgem.gen, sf, false);
+	state->sf[1] = gen4_create_sf_state(&general, sna->kgem.gen, sf, true);
 
 	wm_state = sna_static_stream_map(&general,
 					  sizeof(*wm_state) * KERNEL_COUNT *
diff --git a/src/sna/sna_render.h b/src/sna/sna_render.h
index e056e3fb..a2281f47 100644
--- a/src/sna/sna_render.h
+++ b/src/sna/sna_render.h
@@ -119,6 +119,7 @@ struct sna_composite_op {
 		struct {
 			int wm_kernel;
 			int ve_id;
+			int sf;
 		} gen4;
 
 		struct {