summaryrefslogtreecommitdiffstats
path: root/src/mesa/drivers/dri/i965
diff options
context:
space:
mode:
authorKenneth Graunke <[email protected]>2012-12-03 18:28:29 -0800
committerKenneth Graunke <[email protected]>2014-01-31 17:50:07 -0800
commit5c0d7dbcb9575bd8126d4d4ef83753664b848d27 (patch)
tree7e8e2523442594d98e2541f8f7b8de476c45a221 /src/mesa/drivers/dri/i965
parent08a471495935665c55f2968e310d6e20193b02f1 (diff)
i965: Rework vertex uploads for Broadwell.
v2: Emit a dummy 3DSTATE_VF_SGVS packet when not needed. v3: Add WARN_ONCE and perf_debugs requested by Eric Anholt. v4: Program 3DSTATE_SGVS even in the no-elements case so gl_VertexID continues working. Fix 3DSTATE_VF_INSTANCING to not use an element index to access the buffers array. Some ARB_draw_indirect prep work. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
Diffstat (limited to 'src/mesa/drivers/dri/i965')
-rw-r--r--src/mesa/drivers/dri/i965/Makefile.sources1
-rw-r--r--src/mesa/drivers/dri/i965/brw_context.h1
-rw-r--r--src/mesa/drivers/dri/i965/brw_defines.h13
-rw-r--r--src/mesa/drivers/dri/i965/brw_draw_upload.c3
-rw-r--r--src/mesa/drivers/dri/i965/brw_state.h3
-rw-r--r--src/mesa/drivers/dri/i965/brw_state_upload.c6
-rw-r--r--src/mesa/drivers/dri/i965/gen8_draw_upload.c250
7 files changed, 274 insertions, 3 deletions
diff --git a/src/mesa/drivers/dri/i965/Makefile.sources b/src/mesa/drivers/dri/i965/Makefile.sources
index 57afc9a822d..87962d92fd8 100644
--- a/src/mesa/drivers/dri/i965/Makefile.sources
+++ b/src/mesa/drivers/dri/i965/Makefile.sources
@@ -143,6 +143,7 @@ i965_FILES = \
gen7_wm_surface_state.c \
gen8_blend_state.c \
gen8_disasm.c \
+ gen8_draw_upload.c \
gen8_fs_generator.cpp \
gen8_generator.cpp \
gen8_instruction.c \
diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 8d098e6c3b7..b6a1363a6a1 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -1617,6 +1617,7 @@ gl_clip_plane *brw_select_clip_planes(struct gl_context *ctx);
unsigned brw_get_vertex_surface_type(struct brw_context *brw,
const struct gl_client_array *glarray);
unsigned brw_get_index_type(GLenum type);
+void brw_prepare_vertices(struct brw_context *brw);
/* brw_wm_surface_state.c */
void brw_init_surface_formats(struct brw_context *brw);
diff --git a/src/mesa/drivers/dri/i965/brw_defines.h b/src/mesa/drivers/dri/i965/brw_defines.h
index a7f5d1a5f2f..4496e57311e 100644
--- a/src/mesa/drivers/dri/i965/brw_defines.h
+++ b/src/mesa/drivers/dri/i965/brw_defines.h
@@ -1409,6 +1409,19 @@ enum brw_message_target {
#define _3DSTATE_VF 0x780c /* GEN7.5+ */
#define HSW_CUT_INDEX_ENABLE (1 << 8)
+#define _3DSTATE_VF_INSTANCING 0x7849 /* GEN8+ */
+# define GEN8_VF_INSTANCING_ENABLE (1 << 8)
+
+#define _3DSTATE_VF_SGVS 0x784a /* GEN8+ */
+# define GEN8_SGVS_ENABLE_INSTANCE_ID (1 << 31)
+# define GEN8_SGVS_INSTANCE_ID_COMPONENT_SHIFT 29
+# define GEN8_SGVS_INSTANCE_ID_ELEMENT_OFFSET_SHIFT 16
+# define GEN8_SGVS_ENABLE_VERTEX_ID (1 << 15)
+# define GEN8_SGVS_VERTEX_ID_COMPONENT_SHIFT 13
+# define GEN8_SGVS_VERTEX_ID_ELEMENT_OFFSET_SHIFT 0
+
+#define _3DSTATE_VF_TOPOLOGY 0x784b /* GEN8+ */
+
#define _3DSTATE_URB_VS 0x7830 /* GEN7+ */
#define _3DSTATE_URB_HS 0x7831 /* GEN7+ */
#define _3DSTATE_URB_DS 0x7832 /* GEN7+ */
diff --git a/src/mesa/drivers/dri/i965/brw_draw_upload.c b/src/mesa/drivers/dri/i965/brw_draw_upload.c
index 5418398afdf..52d38478d8d 100644
--- a/src/mesa/drivers/dri/i965/brw_draw_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_draw_upload.c
@@ -400,7 +400,8 @@ copy_array_to_vbo_array(struct brw_context *brw,
buffer->stride = dst_stride;
}
-static void brw_prepare_vertices(struct brw_context *brw)
+void
+brw_prepare_vertices(struct brw_context *brw)
{
struct gl_context *ctx = &brw->ctx;
/* CACHE_NEW_VS_PROG */
diff --git a/src/mesa/drivers/dri/i965/brw_state.h b/src/mesa/drivers/dri/i965/brw_state.h
index 619f4c8d29a..dbad4661d3e 100644
--- a/src/mesa/drivers/dri/i965/brw_state.h
+++ b/src/mesa/drivers/dri/i965/brw_state.h
@@ -131,6 +131,7 @@ extern const struct brw_tracked_state gen7_urb;
extern const struct brw_tracked_state gen7_vs_state;
extern const struct brw_tracked_state gen7_wm_state;
extern const struct brw_tracked_state haswell_cut_index;
+extern const struct brw_tracked_state gen8_index_buffer;
extern const struct brw_tracked_state gen8_ps_blend;
extern const struct brw_tracked_state gen8_ps_extra;
extern const struct brw_tracked_state gen8_ps_state;
@@ -140,6 +141,8 @@ extern const struct brw_tracked_state gen8_raster_state;
extern const struct brw_tracked_state gen8_sbe_state;
extern const struct brw_tracked_state gen8_sf_state;
extern const struct brw_tracked_state gen8_state_base_address;
+extern const struct brw_tracked_state gen8_vertices;
+extern const struct brw_tracked_state gen8_vf_topology;
extern const struct brw_tracked_state gen8_vs_state;
/* brw_misc_state.c */
diff --git a/src/mesa/drivers/dri/i965/brw_state_upload.c b/src/mesa/drivers/dri/i965/brw_state_upload.c
index e4f8104bfab..76698dac3a7 100644
--- a/src/mesa/drivers/dri/i965/brw_state_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_state_upload.c
@@ -322,9 +322,11 @@ static const struct brw_tracked_state *gen8_atoms[] =
&brw_drawing_rect,
+ &gen8_vf_topology,
+
&brw_indices,
- &brw_index_buffer,
- &brw_vertices,
+ &gen8_index_buffer,
+ &gen8_vertices,
&haswell_cut_index,
};
diff --git a/src/mesa/drivers/dri/i965/gen8_draw_upload.c b/src/mesa/drivers/dri/i965/gen8_draw_upload.c
new file mode 100644
index 00000000000..f927c135dd9
--- /dev/null
+++ b/src/mesa/drivers/dri/i965/gen8_draw_upload.c
@@ -0,0 +1,250 @@
+/*
+ * Copyright © 2012 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "main/glheader.h"
+#include "main/bufferobj.h"
+#include "main/context.h"
+#include "main/enums.h"
+#include "main/macros.h"
+
+#include "brw_draw.h"
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_state.h"
+
+#include "intel_batchbuffer.h"
+#include "intel_buffer_objects.h"
+
+static void
+gen8_emit_vertices(struct brw_context *brw)
+{
+ struct gl_context *ctx = &brw->ctx;
+
+ brw_prepare_vertices(brw);
+
+ if (brw->vs.prog_data->uses_vertexid) {
+ unsigned vue = brw->vb.nr_enabled;
+
+ WARN_ONCE(brw->vs.prog_data->inputs_read & VERT_BIT_EDGEFLAG,
+ "Using VID/IID with edgeflags, need to reorder the "
+ "vertex attributes");
+ WARN_ONCE(vue >= 33,
+ "Trying to insert VID/IID past 33rd vertex element, "
+ "need to reorder the vertex attrbutes.");
+
+ BEGIN_BATCH(2);
+ OUT_BATCH(_3DSTATE_VF_SGVS << 16 | (2 - 2));
+ OUT_BATCH(GEN8_SGVS_ENABLE_VERTEX_ID |
+ (0 << GEN8_SGVS_VERTEX_ID_COMPONENT_SHIFT) | /* .x channel */
+ (vue << GEN8_SGVS_VERTEX_ID_ELEMENT_OFFSET_SHIFT) |
+ GEN8_SGVS_ENABLE_INSTANCE_ID |
+ (1 << GEN8_SGVS_INSTANCE_ID_COMPONENT_SHIFT) | /* .y channel */
+ (vue << GEN8_SGVS_INSTANCE_ID_ELEMENT_OFFSET_SHIFT));
+ ADVANCE_BATCH();
+ } else {
+ BEGIN_BATCH(2);
+ OUT_BATCH(_3DSTATE_VF_SGVS << 16 | (2 - 2));
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+ }
+
+ /* If the VS doesn't read any inputs (calculating vertex position from
+ * a state variable for some reason, for example), emit a single pad
+ * VERTEX_ELEMENT struct and bail.
+ *
+ * The stale VB state stays in place, but they don't do anything unless
+ * a VE loads from them.
+ */
+ if (brw->vb.nr_enabled == 0) {
+ BEGIN_BATCH(3);
+ OUT_BATCH((_3DSTATE_VERTEX_ELEMENTS << 16) | (3 - 2));
+ OUT_BATCH((0 << GEN6_VE0_INDEX_SHIFT) |
+ GEN6_VE0_VALID |
+ (BRW_SURFACEFORMAT_R32G32B32A32_FLOAT << BRW_VE0_FORMAT_SHIFT) |
+ (0 << BRW_VE0_SRC_OFFSET_SHIFT));
+ OUT_BATCH((BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_0_SHIFT) |
+ (BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_1_SHIFT) |
+ (BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_2_SHIFT) |
+ (BRW_VE1_COMPONENT_STORE_1_FLT << BRW_VE1_COMPONENT_3_SHIFT));
+ ADVANCE_BATCH();
+ return;
+ }
+
+ /* Now emit 3DSTATE_VERTEX_BUFFERS and 3DSTATE_VERTEX_ELEMENTS packets. */
+ if (brw->vb.nr_buffers) {
+ assert(brw->vb.nr_buffers <= 33);
+
+ perf_debug("Missing MOCS setup for 3DSTATE_VERTEX_BUFFERS.");
+
+ BEGIN_BATCH(1 + 4*brw->vb.nr_buffers);
+ OUT_BATCH((_3DSTATE_VERTEX_BUFFERS << 16) | (4*brw->vb.nr_buffers - 1));
+ for (unsigned i = 0; i < brw->vb.nr_buffers; i++) {
+ struct brw_vertex_buffer *buffer = &brw->vb.buffers[i];
+ uint32_t dw0 = 0;
+
+ dw0 |= i << GEN6_VB0_INDEX_SHIFT;
+ dw0 |= GEN7_VB0_ADDRESS_MODIFYENABLE;
+ dw0 |= buffer->stride << BRW_VB0_PITCH_SHIFT;
+
+ OUT_BATCH(dw0);
+ OUT_RELOC64(buffer->bo, I915_GEM_DOMAIN_VERTEX, 0, buffer->offset);
+ OUT_BATCH(buffer->bo->size);
+ }
+ ADVANCE_BATCH();
+ }
+
+ unsigned nr_elements = brw->vb.nr_enabled;
+
+ /* The hardware allows one more VERTEX_ELEMENTS than VERTEX_BUFFERS,
+ * presumably for VertexID/InstanceID.
+ */
+ assert(nr_elements <= 34);
+
+ struct brw_vertex_element *gen6_edgeflag_input = NULL;
+
+ BEGIN_BATCH(1 + nr_elements * 2);
+ OUT_BATCH((_3DSTATE_VERTEX_ELEMENTS << 16) | (2 * nr_elements - 1));
+ for (unsigned i = 0; i < brw->vb.nr_enabled; i++) {
+ struct brw_vertex_element *input = brw->vb.enabled[i];
+ uint32_t format = brw_get_vertex_surface_type(brw, input->glarray);
+ uint32_t comp0 = BRW_VE1_COMPONENT_STORE_SRC;
+ uint32_t comp1 = BRW_VE1_COMPONENT_STORE_SRC;
+ uint32_t comp2 = BRW_VE1_COMPONENT_STORE_SRC;
+ uint32_t comp3 = BRW_VE1_COMPONENT_STORE_SRC;
+
+ /* The gen4 driver expects edgeflag to come in as a float, and passes
+ * that float on to the tests in the clipper. Mesa's current vertex
+ * attribute value for EdgeFlag is stored as a float, which works out.
+ * glEdgeFlagPointer, on the other hand, gives us an unnormalized
+ * integer ubyte. Just rewrite that to convert to a float.
+ */
+ if (input->attrib == VERT_ATTRIB_EDGEFLAG) {
+ /* Gen6+ passes edgeflag as sideband along with the vertex, instead
+ * of in the VUE. We have to upload it sideband as the last vertex
+ * element according to the B-Spec.
+ */
+ gen6_edgeflag_input = input;
+ continue;
+ }
+
+ switch (input->glarray->Size) {
+ case 0: comp0 = BRW_VE1_COMPONENT_STORE_0;
+ case 1: comp1 = BRW_VE1_COMPONENT_STORE_0;
+ case 2: comp2 = BRW_VE1_COMPONENT_STORE_0;
+ case 3: comp3 = input->glarray->Integer ? BRW_VE1_COMPONENT_STORE_1_INT
+ : BRW_VE1_COMPONENT_STORE_1_FLT;
+ break;
+ }
+
+ OUT_BATCH((input->buffer << GEN6_VE0_INDEX_SHIFT) |
+ GEN6_VE0_VALID |
+ (format << BRW_VE0_FORMAT_SHIFT) |
+ (input->offset << BRW_VE0_SRC_OFFSET_SHIFT));
+
+ OUT_BATCH((comp0 << BRW_VE1_COMPONENT_0_SHIFT) |
+ (comp1 << BRW_VE1_COMPONENT_1_SHIFT) |
+ (comp2 << BRW_VE1_COMPONENT_2_SHIFT) |
+ (comp3 << BRW_VE1_COMPONENT_3_SHIFT));
+ }
+
+ if (gen6_edgeflag_input) {
+ uint32_t format =
+ brw_get_vertex_surface_type(brw, gen6_edgeflag_input->glarray);
+
+ OUT_BATCH((gen6_edgeflag_input->buffer << GEN6_VE0_INDEX_SHIFT) |
+ GEN6_VE0_VALID |
+ GEN6_VE0_EDGE_FLAG_ENABLE |
+ (format << BRW_VE0_FORMAT_SHIFT) |
+ (gen6_edgeflag_input->offset << BRW_VE0_SRC_OFFSET_SHIFT));
+ OUT_BATCH((BRW_VE1_COMPONENT_STORE_SRC << BRW_VE1_COMPONENT_0_SHIFT) |
+ (BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_1_SHIFT) |
+ (BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_2_SHIFT) |
+ (BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_3_SHIFT));
+ }
+ ADVANCE_BATCH();
+
+ for (unsigned i = 0; i < brw->vb.nr_enabled; i++) {
+ const struct brw_vertex_element *input = brw->vb.enabled[i];
+ const struct brw_vertex_buffer *buffer = &brw->vb.buffers[input->buffer];
+
+ BEGIN_BATCH(3);
+ OUT_BATCH(_3DSTATE_VF_INSTANCING << 16 | (3 - 2));
+ OUT_BATCH(i | (buffer->step_rate ? GEN8_VF_INSTANCING_ENABLE : 0));
+ OUT_BATCH(buffer->step_rate);
+ ADVANCE_BATCH();
+ }
+}
+
+const struct brw_tracked_state gen8_vertices = {
+ .dirty = {
+ .mesa = _NEW_POLYGON,
+ .brw = BRW_NEW_BATCH | BRW_NEW_VERTICES,
+ .cache = CACHE_NEW_VS_PROG,
+ },
+ .emit = gen8_emit_vertices,
+};
+
+static void
+gen8_emit_index_buffer(struct brw_context *brw)
+{
+ const struct _mesa_index_buffer *index_buffer = brw->ib.ib;
+
+ if (index_buffer == NULL)
+ return;
+
+ perf_debug("Missing MOCS setup for 3DSTATE_INDEX_BUFFER.");
+
+ BEGIN_BATCH(5);
+ OUT_BATCH(CMD_INDEX_BUFFER << 16 | (5 - 2));
+ OUT_BATCH(brw_get_index_type(index_buffer->type) << 8);
+ OUT_RELOC64(brw->ib.bo, I915_GEM_DOMAIN_VERTEX, 0, 0);
+ OUT_BATCH(brw->ib.bo->size);
+ ADVANCE_BATCH();
+}
+
+const struct brw_tracked_state gen8_index_buffer = {
+ .dirty = {
+ .mesa = 0,
+ .brw = BRW_NEW_BATCH | BRW_NEW_INDEX_BUFFER,
+ .cache = 0,
+ },
+ .emit = gen8_emit_index_buffer,
+};
+
+static void
+gen8_emit_vf_topology(struct brw_context *brw)
+{
+ BEGIN_BATCH(2);
+ OUT_BATCH(_3DSTATE_VF_TOPOLOGY << 16 | (2 - 2));
+ OUT_BATCH(brw->primitive);
+ ADVANCE_BATCH();
+}
+
+const struct brw_tracked_state gen8_vf_topology = {
+ .dirty = {
+ .mesa = 0,
+ .brw = BRW_NEW_PRIMITIVE,
+ .cache = 0,
+ },
+ .emit = gen8_emit_vf_topology,
+};