/*
Copyright (C) Intel Corp. 2006. All Rights Reserved.
- Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ Intel funded Tungsten Graphics to
develop this 3D driver.
-
+
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
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.
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.
-
+
**********************************************************************/
/*
* Authors:
- * Keith Whitwell <keith@tungstengraphics.com>
+ * Keith Whitwell <keithw@vmware.com>
*/
-
+
#include "main/compiler.h"
#include "brw_context.h"
#include "glsl/ralloc.h"
static inline void assign_vue_slot(struct brw_vue_map *vue_map,
- int vert_result)
+ int varying)
{
- /* Make sure this vert_result hasn't been assigned a slot already */
- assert (vue_map->vert_result_to_slot[vert_result] == -1);
+ /* Make sure this varying hasn't been assigned a slot already */
+ assert (vue_map->varying_to_slot[varying] == -1);
- vue_map->vert_result_to_slot[vert_result] = vue_map->num_slots;
- vue_map->slot_to_vert_result[vue_map->num_slots++] = vert_result;
+ vue_map->varying_to_slot[varying] = vue_map->num_slots;
+ vue_map->slot_to_varying[vue_map->num_slots++] = varying;
}
/**
* Compute the VUE map for vertex shader program.
- *
- * Note that consumers of this map using cache keys must include
- * prog_data->userclip and prog_data->outputs_written in their key
- * (generated by CACHE_NEW_VS_PROG).
*/
void
-brw_compute_vue_map(struct brw_vue_map *vue_map,
- const struct intel_context *intel,
- const struct brw_vs_prog_data *prog_data)
+brw_compute_vue_map(struct brw_context *brw, struct brw_vue_map *vue_map,
+ GLbitfield64 slots_valid)
{
- bool userclip_active = prog_data->userclip;
- GLbitfield64 outputs_written = prog_data->outputs_written;
+ vue_map->slots_valid = slots_valid;
int i;
+ /* gl_Layer doesn't get its own varying slot--it's stored in the virst VUE
+ * slot (VARYING_SLOT_PSIZ).
+ */
+ slots_valid &= ~VARYING_BIT_LAYER;
+
+ /* Make sure that the values we store in vue_map->varying_to_slot and
+ * vue_map->slot_to_varying won't overflow the signed chars that are used
+ * to store them. Note that since vue_map->slot_to_varying sometimes holds
+ * values equal to BRW_VARYING_SLOT_COUNT, we need to ensure that
+ * BRW_VARYING_SLOT_COUNT is <= 127, not 128.
+ */
+ STATIC_ASSERT(BRW_VARYING_SLOT_COUNT <= 127);
+
vue_map->num_slots = 0;
- for (i = 0; i < BRW_VERT_RESULT_MAX; ++i) {
- vue_map->vert_result_to_slot[i] = -1;
- vue_map->slot_to_vert_result[i] = BRW_VERT_RESULT_MAX;
+ for (i = 0; i < BRW_VARYING_SLOT_COUNT; ++i) {
+ vue_map->varying_to_slot[i] = -1;
+ vue_map->slot_to_varying[i] = BRW_VARYING_SLOT_COUNT;
}
/* VUE header: format depends on chip generation and whether clipping is
* enabled.
*/
- switch (intel->gen) {
+ switch (brw->gen) {
case 4:
+ case 5:
/* There are 8 dwords in VUE header pre-Ironlake:
* dword 0-3 is indices, point width, clip flags.
* dword 4-7 is ndc position
* dword 8-11 is the first vertex data.
- */
- assign_vue_slot(vue_map, VERT_RESULT_PSIZ);
- assign_vue_slot(vue_map, BRW_VERT_RESULT_NDC);
- assign_vue_slot(vue_map, VERT_RESULT_HPOS);
- break;
- case 5:
- /* There are 20 DWs (D0-D19) in VUE header on Ironlake:
- * dword 0-3 of the header is indices, point width, clip flags.
- * dword 4-7 is the ndc position
- * dword 8-11 of the vertex header is the 4D space position
- * dword 12-19 of the vertex header is the user clip distance.
- * dword 20-23 is a pad so that the vertex element data is aligned
- * dword 24-27 is the first vertex data we fill.
*
- * Note: future pipeline stages expect 4D space position to be
- * contiguous with the other vert_results, so we make dword 24-27 a
- * duplicate copy of the 4D space position.
+ * On Ironlake the VUE header is nominally 20 dwords, but the hardware
+ * will accept the same header layout as Gen4 [and should be a bit faster]
*/
- assign_vue_slot(vue_map, VERT_RESULT_PSIZ);
- assign_vue_slot(vue_map, BRW_VERT_RESULT_NDC);
- assign_vue_slot(vue_map, BRW_VERT_RESULT_HPOS_DUPLICATE);
- assign_vue_slot(vue_map, VERT_RESULT_CLIP_DIST0);
- assign_vue_slot(vue_map, VERT_RESULT_CLIP_DIST1);
- assign_vue_slot(vue_map, BRW_VERT_RESULT_PAD);
- assign_vue_slot(vue_map, VERT_RESULT_HPOS);
+ assign_vue_slot(vue_map, VARYING_SLOT_PSIZ);
+ assign_vue_slot(vue_map, BRW_VARYING_SLOT_NDC);
+ assign_vue_slot(vue_map, VARYING_SLOT_POS);
break;
case 6:
case 7:
* enabled.
* dword 8-11 or 16-19 is the first vertex element data we fill.
*/
- assign_vue_slot(vue_map, VERT_RESULT_PSIZ);
- assign_vue_slot(vue_map, VERT_RESULT_HPOS);
- if (userclip_active) {
- assign_vue_slot(vue_map, VERT_RESULT_CLIP_DIST0);
- assign_vue_slot(vue_map, VERT_RESULT_CLIP_DIST1);
- }
+ assign_vue_slot(vue_map, VARYING_SLOT_PSIZ);
+ assign_vue_slot(vue_map, VARYING_SLOT_POS);
+ if (slots_valid & BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST0))
+ assign_vue_slot(vue_map, VARYING_SLOT_CLIP_DIST0);
+ if (slots_valid & BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST1))
+ assign_vue_slot(vue_map, VARYING_SLOT_CLIP_DIST1);
+
/* front and back colors need to be consecutive so that we can use
* ATTRIBUTE_SWIZZLE_INPUTATTR_FACING to swizzle them when doing
* two-sided color.
*/
- if (outputs_written & BITFIELD64_BIT(VERT_RESULT_COL0))
- assign_vue_slot(vue_map, VERT_RESULT_COL0);
- if (outputs_written & BITFIELD64_BIT(VERT_RESULT_BFC0))
- assign_vue_slot(vue_map, VERT_RESULT_BFC0);
- if (outputs_written & BITFIELD64_BIT(VERT_RESULT_COL1))
- assign_vue_slot(vue_map, VERT_RESULT_COL1);
- if (outputs_written & BITFIELD64_BIT(VERT_RESULT_BFC1))
- assign_vue_slot(vue_map, VERT_RESULT_BFC1);
+ if (slots_valid & BITFIELD64_BIT(VARYING_SLOT_COL0))
+ assign_vue_slot(vue_map, VARYING_SLOT_COL0);
+ if (slots_valid & BITFIELD64_BIT(VARYING_SLOT_BFC0))
+ assign_vue_slot(vue_map, VARYING_SLOT_BFC0);
+ if (slots_valid & BITFIELD64_BIT(VARYING_SLOT_COL1))
+ assign_vue_slot(vue_map, VARYING_SLOT_COL1);
+ if (slots_valid & BITFIELD64_BIT(VARYING_SLOT_BFC1))
+ assign_vue_slot(vue_map, VARYING_SLOT_BFC1);
break;
default:
assert (!"VUE map not known for this chip generation");
* assign them contiguously. Don't reassign outputs that already have a
* slot.
*
- * Also, prior to Gen6, don't assign a slot for VERT_RESULT_CLIP_VERTEX,
- * since it is unsupported. In Gen6 and above, VERT_RESULT_CLIP_VERTEX may
- * be needed for transform feedback; since we don't want to have to
- * recompute the VUE map (and everything that depends on it) when transform
- * feedback is enabled or disabled, just go ahead and assign a slot for it.
+ * We generally don't need to assign a slot for VARYING_SLOT_CLIP_VERTEX,
+ * since it's encoded as the clip distances by emit_clip_distances().
+ * However, it may be output by transform feedback, and we'd rather not
+ * recompute state when TF changes, so we just always include it.
*/
- for (int i = 0; i < VERT_RESULT_MAX; ++i) {
- if (intel->gen < 6 && i == VERT_RESULT_CLIP_VERTEX)
- continue;
- if ((outputs_written & BITFIELD64_BIT(i)) &&
- vue_map->vert_result_to_slot[i] == -1) {
+ for (int i = 0; i < VARYING_SLOT_MAX; ++i) {
+ if ((slots_valid & BITFIELD64_BIT(i)) &&
+ vue_map->varying_to_slot[i] == -1) {
assign_vue_slot(vue_map, i);
}
}
}
+bool
+brw_vs_prog_data_compare(const void *in_a, const void *in_b)
+{
+ const struct brw_vs_prog_data *a = in_a;
+ const struct brw_vs_prog_data *b = in_b;
+
+ /* Compare the base vec4 structure. */
+ if (!brw_vec4_prog_data_compare(&a->base, &b->base))
+ return false;
+
+ /* Compare the rest of the struct. */
+ const unsigned offset = sizeof(struct brw_vec4_prog_data);
+ if (memcmp(((char *) a) + offset, ((char *) b) + offset,
+ sizeof(struct brw_vs_prog_data) - offset)) {
+ return false;
+ }
+
+ return true;
+}
+
static bool
do_vs_prog(struct brw_context *brw,
struct gl_shader_program *prog,
struct brw_vertex_program *vp,
struct brw_vs_prog_key *key)
{
- struct gl_context *ctx = &brw->intel.ctx;
- struct intel_context *intel = &brw->intel;
GLuint program_size;
const GLuint *program;
struct brw_vs_compile c;
+ struct brw_vs_prog_data prog_data;
void *mem_ctx;
- int aux_size;
int i;
+ struct gl_shader *vs = NULL;
+
+ if (prog)
+ vs = prog->_LinkedShaders[MESA_SHADER_VERTEX];
memset(&c, 0, sizeof(c));
memcpy(&c.key, key, sizeof(*key));
+ memset(&prog_data, 0, sizeof(prog_data));
mem_ctx = ralloc_context(NULL);
- brw_init_compile(brw, &c.func, mem_ctx);
c.vp = vp;
- c.prog_data.outputs_written = vp->program.Base.OutputsWritten;
- c.prog_data.inputs_read = vp->program.Base.InputsRead;
+ /* Allocate the references to the uniforms that will end up in the
+ * prog_data associated with the compiled program, and which will be freed
+ * by the state cache.
+ */
+ int param_count;
+ if (vs) {
+ /* We add padding around uniform values below vec4 size, with the worst
+ * case being a float value that gets blown up to a vec4, so be
+ * conservative here.
+ */
+ param_count = vs->num_uniform_components * 4;
+
+ } else {
+ param_count = vp->program.Base.Parameters->NumParameters * 4;
+ }
+ /* vec4_visitor::setup_uniform_clipplane_values() also uploads user clip
+ * planes as uniforms.
+ */
+ param_count += c.key.base.nr_userclip_plane_consts * 4;
+
+ prog_data.base.param = rzalloc_array(NULL, const float *, param_count);
+ prog_data.base.pull_param = rzalloc_array(NULL, const float *, param_count);
+
+ GLbitfield64 outputs_written = vp->program.Base.OutputsWritten;
+ prog_data.inputs_read = vp->program.Base.InputsRead;
if (c.key.copy_edgeflag) {
- c.prog_data.outputs_written |= BITFIELD64_BIT(VERT_RESULT_EDGE);
- c.prog_data.inputs_read |= VERT_BIT_EDGEFLAG;
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_EDGE);
+ prog_data.inputs_read |= VERT_BIT_EDGEFLAG;
+ }
+
+ if (brw->gen < 6) {
+ /* Put dummy slots into the VUE for the SF to put the replaced
+ * point sprite coords in. We shouldn't need these dummy slots,
+ * which take up precious URB space, but it would mean that the SF
+ * doesn't get nice aligned pairs of input coords into output
+ * coords, which would be a pain to handle.
+ */
+ for (i = 0; i < 8; i++) {
+ if (c.key.point_coord_replace & (1 << i))
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_TEX0 + i);
+ }
+
+ /* if back colors are written, allocate slots for front colors too */
+ if (outputs_written & BITFIELD64_BIT(VARYING_SLOT_BFC0))
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_COL0);
+ if (outputs_written & BITFIELD64_BIT(VARYING_SLOT_BFC1))
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_COL1);
}
- /* Put dummy slots into the VUE for the SF to put the replaced
- * point sprite coords in. We shouldn't need these dummy slots,
- * which take up precious URB space, but it would mean that the SF
- * doesn't get nice aligned pairs of input coords into output
- * coords, which would be a pain to handle.
+ /* In order for legacy clipping to work, we need to populate the clip
+ * distance varying slots whenever clipping is enabled, even if the vertex
+ * shader doesn't write to gl_ClipDistance.
*/
- for (i = 0; i < 8; i++) {
- if (c.key.point_coord_replace & (1 << i))
- c.prog_data.outputs_written |= BITFIELD64_BIT(VERT_RESULT_TEX0 + i);
+ if (c.key.base.userclip_active) {
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST0);
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST1);
}
+ brw_compute_vue_map(brw, &prog_data.base.vue_map, outputs_written);
+
if (0) {
_mesa_fprint_program_opt(stdout, &c.vp->program.Base, PROG_PRINT_DEBUG,
true);
/* Emit GEN4 code.
*/
- if (prog) {
- if (!brw_vs_emit(prog, &c)) {
- ralloc_free(mem_ctx);
- return false;
- }
- } else {
- brw_old_vs_emit(&c);
+ program = brw_vs_emit(brw, prog, &c, &prog_data, mem_ctx, &program_size);
+ if (program == NULL) {
+ ralloc_free(mem_ctx);
+ return false;
}
/* Scratch space is used for register spilling */
- if (c.last_scratch) {
- c.prog_data.total_scratch = brw_get_scratch_size(c.last_scratch);
-
- brw_get_scratch_bo(intel, &brw->vs.scratch_bo,
- c.prog_data.total_scratch * brw->max_vs_threads);
- }
-
- /* get the program
- */
- program = brw_get_program(&c.func, &program_size);
+ if (c.base.last_scratch) {
+ perf_debug("Vertex shader triggered register spilling. "
+ "Try reducing the number of live vec4 values to "
+ "improve performance.\n");
- /* We upload from &c.prog_data including the constant_map assuming
- * they're packed together. It would be nice to have a
- * compile-time assert macro here.
- */
- assert(c.constant_map == (int8_t *)&c.prog_data +
- sizeof(c.prog_data));
- assert(ctx->Const.VertexProgram.MaxNativeParameters ==
- ARRAY_SIZE(c.constant_map));
- (void) ctx;
+ prog_data.base.total_scratch
+ = brw_get_scratch_size(c.base.last_scratch*REG_SIZE);
- aux_size = sizeof(c.prog_data);
- /* constant_map */
- aux_size += c.vp->program.Base.Parameters->NumParameters;
+ brw_get_scratch_bo(brw, &brw->vs.base.scratch_bo,
+ prog_data.base.total_scratch * brw->max_vs_threads);
+ }
brw_upload_cache(&brw->cache, BRW_VS_PROG,
&c.key, sizeof(c.key),
program, program_size,
- &c.prog_data, aux_size,
- &brw->vs.prog_offset, &brw->vs.prog_data);
+ &prog_data, sizeof(prog_data),
+ &brw->vs.base.prog_offset, &brw->vs.prog_data);
ralloc_free(mem_ctx);
return true;
}
+static bool
+key_debug(struct brw_context *brw, const char *name, int a, int b)
+{
+ if (a != b) {
+ perf_debug(" %s %d->%d\n", name, a, b);
+ return true;
+ }
+ return false;
+}
+
+void
+brw_vs_debug_recompile(struct brw_context *brw,
+ struct gl_shader_program *prog,
+ const struct brw_vs_prog_key *key)
+{
+ struct brw_cache_item *c = NULL;
+ const struct brw_vs_prog_key *old_key = NULL;
+ bool found = false;
+
+ perf_debug("Recompiling vertex shader for program %d\n", prog->Name);
+
+ for (unsigned int i = 0; i < brw->cache.size; i++) {
+ for (c = brw->cache.items[i]; c; c = c->next) {
+ if (c->cache_id == BRW_VS_PROG) {
+ old_key = c->key;
+
+ if (old_key->base.program_string_id == key->base.program_string_id)
+ break;
+ }
+ }
+ if (c)
+ break;
+ }
+
+ if (!c) {
+ perf_debug(" Didn't find previous compile in the shader cache for "
+ "debug\n");
+ return;
+ }
+
+ for (unsigned int i = 0; i < VERT_ATTRIB_MAX; i++) {
+ found |= key_debug(brw, "Vertex attrib w/a flags",
+ old_key->gl_attrib_wa_flags[i],
+ key->gl_attrib_wa_flags[i]);
+ }
+
+ found |= key_debug(brw, "user clip flags",
+ old_key->base.userclip_active, key->base.userclip_active);
+
+ found |= key_debug(brw, "user clipping planes as push constants",
+ old_key->base.nr_userclip_plane_consts,
+ key->base.nr_userclip_plane_consts);
+
+ found |= key_debug(brw, "copy edgeflag",
+ old_key->copy_edgeflag, key->copy_edgeflag);
+ found |= key_debug(brw, "PointCoord replace",
+ old_key->point_coord_replace, key->point_coord_replace);
+ found |= key_debug(brw, "vertex color clamping",
+ old_key->base.clamp_vertex_color, key->base.clamp_vertex_color);
+
+ found |= brw_debug_recompile_sampler_key(brw, &old_key->base.tex,
+ &key->base.tex);
+
+ if (!found) {
+ perf_debug(" Something else\n");
+ }
+}
+
+
+void
+brw_setup_vec4_key_clip_info(struct brw_context *brw,
+ struct brw_vec4_prog_key *key,
+ bool program_uses_clip_distance)
+{
+ struct gl_context *ctx = &brw->ctx;
+
+ key->userclip_active = (ctx->Transform.ClipPlanesEnabled != 0);
+ if (key->userclip_active && !program_uses_clip_distance) {
+ key->nr_userclip_plane_consts
+ = _mesa_logbase2(ctx->Transform.ClipPlanesEnabled) + 1;
+ }
+}
+
static void brw_upload_vs_prog(struct brw_context *brw)
{
- struct intel_context *intel = &brw->intel;
- struct gl_context *ctx = &intel->ctx;
+ struct gl_context *ctx = &brw->ctx;
struct brw_vs_prog_key key;
/* BRW_NEW_VERTEX_PROGRAM */
- struct brw_vertex_program *vp =
+ struct brw_vertex_program *vp =
(struct brw_vertex_program *)brw->vertex_program;
struct gl_program *prog = (struct gl_program *) brw->vertex_program;
int i;
/* Just upload the program verbatim for now. Always send it all
* the inputs it asks for, whether they are varying or not.
*/
- key.program_string_id = vp->id;
- key.userclip_active = (ctx->Transform.ClipPlanesEnabled != 0);
- key.uses_clip_distance = vp->program.UsesClipDistance;
- if (key.userclip_active && !key.uses_clip_distance) {
- if (intel->gen < 6) {
- key.nr_userclip_plane_consts
- = _mesa_bitcount_64(ctx->Transform.ClipPlanesEnabled);
- key.userclip_planes_enabled_gen_4_5
- = ctx->Transform.ClipPlanesEnabled;
- } else {
- key.nr_userclip_plane_consts
- = _mesa_logbase2(ctx->Transform.ClipPlanesEnabled) + 1;
- }
+ key.base.program_string_id = vp->id;
+ brw_setup_vec4_key_clip_info(brw, &key.base,
+ vp->program.Base.UsesClipDistanceOut);
+
+ /* _NEW_POLYGON */
+ if (brw->gen < 6) {
+ key.copy_edgeflag = (ctx->Polygon.FrontMode != GL_FILL ||
+ ctx->Polygon.BackMode != GL_FILL);
}
- key.copy_edgeflag = (ctx->Polygon.FrontMode != GL_FILL ||
- ctx->Polygon.BackMode != GL_FILL);
/* _NEW_LIGHT | _NEW_BUFFERS */
- key.clamp_vertex_color = ctx->Light._ClampVertexColor;
+ key.base.clamp_vertex_color = ctx->Light._ClampVertexColor;
/* _NEW_POINT */
- if (ctx->Point.PointSprite) {
+ if (brw->gen < 6 && ctx->Point.PointSprite) {
for (i = 0; i < 8; i++) {
if (ctx->Point.CoordReplace[i])
key.point_coord_replace |= (1 << i);
}
/* _NEW_TEXTURE */
- for (i = 0; i < BRW_MAX_TEX_UNIT; i++) {
- if (prog->TexturesUsed[i])
- brw_populate_sampler_prog_key_data(ctx, &key.tex, i);
- }
+ brw_populate_sampler_prog_key_data(ctx, prog, brw->vs.base.sampler_count,
+ &key.base.tex);
/* BRW_NEW_VERTICES */
- for (i = 0; i < VERT_ATTRIB_MAX; i++) {
- if (vp->program.Base.InputsRead & BITFIELD64_BIT(i) &&
- brw->vb.inputs[i].glarray->Type == GL_FIXED) {
- key.gl_fixed_input_size[i] = brw->vb.inputs[i].glarray->Size;
+ if (brw->gen < 8 && !brw->is_haswell) {
+ /* Prior to Haswell, the hardware can't natively support GL_FIXED or
+ * 2_10_10_10_REV vertex formats. Set appropriate workaround flags.
+ */
+ for (i = 0; i < VERT_ATTRIB_MAX; i++) {
+ if (!(vp->program.Base.InputsRead & BITFIELD64_BIT(i)))
+ continue;
+
+ uint8_t wa_flags = 0;
+
+ switch (brw->vb.inputs[i].glarray->Type) {
+
+ case GL_FIXED:
+ wa_flags = brw->vb.inputs[i].glarray->Size;
+ break;
+
+ case GL_INT_2_10_10_10_REV:
+ wa_flags |= BRW_ATTRIB_WA_SIGN;
+ /* fallthough */
+
+ case GL_UNSIGNED_INT_2_10_10_10_REV:
+ if (brw->vb.inputs[i].glarray->Format == GL_BGRA)
+ wa_flags |= BRW_ATTRIB_WA_BGRA;
+
+ if (brw->vb.inputs[i].glarray->Normalized)
+ wa_flags |= BRW_ATTRIB_WA_NORMALIZE;
+ else if (!brw->vb.inputs[i].glarray->Integer)
+ wa_flags |= BRW_ATTRIB_WA_SCALE;
+
+ break;
+ }
+
+ key.gl_attrib_wa_flags[i] = wa_flags;
}
}
if (!brw_search_cache(&brw->cache, BRW_VS_PROG,
&key, sizeof(key),
- &brw->vs.prog_offset, &brw->vs.prog_data)) {
+ &brw->vs.base.prog_offset, &brw->vs.prog_data)) {
bool success = do_vs_prog(brw, ctx->Shader.CurrentVertexProgram,
vp, &key);
-
+ (void) success;
assert(success);
}
- brw->vs.constant_map = ((int8_t *)brw->vs.prog_data +
- sizeof(*brw->vs.prog_data));
+ brw->vs.base.prog_data = &brw->vs.prog_data->base.base;
+
+ if (memcmp(&brw->vs.prog_data->base.vue_map, &brw->vue_map_geom_out,
+ sizeof(brw->vue_map_geom_out)) != 0) {
+ brw->vue_map_vs = brw->vs.prog_data->base.vue_map;
+ brw->state.dirty.brw |= BRW_NEW_VUE_MAP_VS;
+ if (brw->gen < 7) {
+ /* No geometry shader support, so the VS VUE map is the VUE map for
+ * the output of the "geometry" portion of the pipeline.
+ */
+ brw->vue_map_geom_out = brw->vue_map_vs;
+ brw->state.dirty.brw |= BRW_NEW_VUE_MAP_GEOM_OUT;
+ }
+ }
}
/* See brw_vs.c:
{
struct brw_context *brw = brw_context(ctx);
struct brw_vs_prog_key key;
- uint32_t old_prog_offset = brw->vs.prog_offset;
+ uint32_t old_prog_offset = brw->vs.base.prog_offset;
struct brw_vs_prog_data *old_prog_data = brw->vs.prog_data;
bool success;
memset(&key, 0, sizeof(key));
- key.program_string_id = bvp->id;
- key.clamp_vertex_color = true;
+ brw_vec4_setup_prog_key_for_precompile(ctx, &key.base, bvp->id, &vp->Base);
success = do_vs_prog(brw, prog, bvp, &key);
- brw->vs.prog_offset = old_prog_offset;
+ brw->vs.base.prog_offset = old_prog_offset;
brw->vs.prog_data = old_prog_data;
return success;
}
+
+
+void
+brw_vs_prog_data_free(const void *in_prog_data)
+{
+ const struct brw_vs_prog_data *prog_data = in_prog_data;
+
+ brw_vec4_prog_data_free(&prog_data->base);
+}
projects / mesa.git / blobdiff