* modes are in use, and set the appropriate bits in
* barycentric_interp_modes.
*/
- for (attr = 0; attr < FRAG_ATTRIB_MAX; ++attr) {
+ for (attr = 0; attr < VARYING_SLOT_MAX; ++attr) {
enum glsl_interp_qualifier interp_qualifier =
fprog->InterpQualifier[attr];
bool is_centroid = fprog->IsCentroid & BITFIELD64_BIT(attr);
- bool is_gl_Color = attr == FRAG_ATTRIB_COL0 || attr == FRAG_ATTRIB_COL1;
+ bool is_gl_Color = attr == VARYING_SLOT_COL0 || attr == VARYING_SLOT_COL1;
/* Ignore unused inputs. */
if (!(fprog->Base.InputsRead & BITFIELD64_BIT(attr)))
continue;
/* Ignore WPOS and FACE, because they don't require interpolation. */
- if (attr == FRAG_ATTRIB_WPOS || attr == FRAG_ATTRIB_FACE)
+ if (attr == VARYING_SLOT_POS || attr == VARYING_SLOT_FACE)
continue;
/* Determine the set (or sets) of barycentric coordinates needed to
return barycentric_interp_modes;
}
-
-void
-brw_wm_payload_setup(struct brw_context *brw,
- struct brw_wm_compile *c)
-{
- struct intel_context *intel = &brw->intel;
- bool uses_depth = (c->fp->program.Base.InputsRead &
- (1 << FRAG_ATTRIB_WPOS)) != 0;
- unsigned barycentric_interp_modes = c->prog_data.barycentric_interp_modes;
- int i;
-
- if (intel->gen >= 6) {
- /* R0-1: masks, pixel X/Y coordinates. */
- c->nr_payload_regs = 2;
- /* R2: only for 32-pixel dispatch.*/
-
- /* R3-26: barycentric interpolation coordinates. These appear in the
- * same order that they appear in the brw_wm_barycentric_interp_mode
- * enum. Each set of coordinates occupies 2 registers if dispatch width
- * == 8 and 4 registers if dispatch width == 16. Coordinates only
- * appear if they were enabled using the "Barycentric Interpolation
- * Mode" bits in WM_STATE.
- */
- for (i = 0; i < BRW_WM_BARYCENTRIC_INTERP_MODE_COUNT; ++i) {
- if (barycentric_interp_modes & (1 << i)) {
- c->barycentric_coord_reg[i] = c->nr_payload_regs;
- c->nr_payload_regs += 2;
- if (c->dispatch_width == 16) {
- c->nr_payload_regs += 2;
- }
- }
- }
-
- /* R27: interpolated depth if uses source depth */
- if (uses_depth) {
- c->source_depth_reg = c->nr_payload_regs;
- c->nr_payload_regs++;
- if (c->dispatch_width == 16) {
- /* R28: interpolated depth if not 8-wide. */
- c->nr_payload_regs++;
- }
- }
- /* R29: interpolated W set if GEN6_WM_USES_SOURCE_W.
- */
- if (uses_depth) {
- c->source_w_reg = c->nr_payload_regs;
- c->nr_payload_regs++;
- if (c->dispatch_width == 16) {
- /* R30: interpolated W if not 8-wide. */
- c->nr_payload_regs++;
- }
- }
- /* R31: MSAA position offsets. */
- /* R32-: bary for 32-pixel. */
- /* R58-59: interp W for 32-pixel. */
-
- if (c->fp->program.Base.OutputsWritten &
- BITFIELD64_BIT(FRAG_RESULT_DEPTH)) {
- c->source_depth_to_render_target = true;
- c->computes_depth = true;
- }
- } else {
- brw_wm_lookup_iz(intel, c);
- }
-}
-
bool
brw_wm_prog_data_compare(const void *in_a, const void *in_b,
int aux_size, const void *in_key)
struct brw_fragment_program *fp,
struct brw_wm_prog_key *key)
{
- struct intel_context *intel = &brw->intel;
struct brw_wm_compile *c;
const GLuint *program;
struct gl_shader *fs = NULL;
if (prog)
fs = prog->_LinkedShaders[MESA_SHADER_FRAGMENT];
- c = brw->wm.compile_data;
- if (c == NULL) {
- brw->wm.compile_data = rzalloc(NULL, struct brw_wm_compile);
- c = brw->wm.compile_data;
- if (c == NULL) {
- /* Ouch - big out of memory problem. Can't continue
- * without triggering a segfault, no way to signal,
- * so just return.
- */
- return false;
- }
- } else {
- memset(c, 0, sizeof(*brw->wm.compile_data));
- }
+ c = rzalloc(NULL, struct brw_wm_compile);
/* 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 (fs) {
- int param_count = fs->num_uniform_components;
- /* The backend also sometimes adds params for texture size. */
- param_count += 2 * BRW_MAX_TEX_UNIT;
-
- c->prog_data.param = rzalloc_array(c, const float *, param_count);
- c->prog_data.pull_param = rzalloc_array(c, const float *, param_count);
+ param_count = fs->num_uniform_components;
} else {
- /* brw_wm_pass0.c will also add references to 0.0 and 1.0 which are
- * uploaded as push parameters.
- */
- int param_count = (fp->program.Base.Parameters->NumParameters + 2) * 4;
- c->prog_data.param = rzalloc_array(c, const float *, param_count);
- /* The old backend never does pull constants. */
- c->prog_data.pull_param = NULL;
+ param_count = fp->program.Base.Parameters->NumParameters * 4;
}
+ /* The backend also sometimes adds params for texture size. */
+ param_count += 2 * BRW_MAX_TEX_UNIT;
+ c->prog_data.param = rzalloc_array(NULL, const float *, param_count);
+ c->prog_data.pull_param = rzalloc_array(NULL, const float *, param_count);
memcpy(&c->key, key, sizeof(*key));
- c->fp = fp;
-
- brw_init_compile(brw, &c->func, c);
-
c->prog_data.barycentric_interp_modes =
brw_compute_barycentric_interp_modes(brw, c->key.flat_shade,
&fp->program);
- brw_wm_fs_emit(brw, c, prog);
+ program = brw_wm_fs_emit(brw, c, &fp->program, prog, &program_size);
+ if (program == NULL)
+ 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->wm.scratch_bo,
+ brw_get_scratch_bo(brw, &brw->wm.scratch_bo,
c->prog_data.total_scratch * brw->max_wm_threads);
}
if (unlikely(INTEL_DEBUG & DEBUG_WM))
fprintf(stderr, "\n");
- /* get the program
- */
- program = brw_get_program(&c->func, &program_size);
-
brw_upload_cache(&brw->cache, BRW_WM_PROG,
&c->key, sizeof(c->key),
program, program_size,
&c->prog_data, sizeof(c->prog_data),
&brw->wm.prog_offset, &brw->wm.prog_data);
+ ralloc_free(c);
+
return true;
}
static bool
-key_debug(const char *name, int a, int b)
+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);
}
bool
-brw_debug_recompile_sampler_key(const struct brw_sampler_prog_key_data *old_key,
+brw_debug_recompile_sampler_key(struct brw_context *brw,
+ const struct brw_sampler_prog_key_data *old_key,
const struct brw_sampler_prog_key_data *key)
{
bool found = false;
for (unsigned int i = 0; i < MAX_SAMPLERS; i++) {
- found |= key_debug("EXT_texture_swizzle or DEPTH_TEXTURE_MODE",
+ found |= key_debug(brw, "EXT_texture_swizzle or DEPTH_TEXTURE_MODE",
old_key->swizzles[i], key->swizzles[i]);
}
- found |= key_debug("GL_CLAMP enabled on any texture unit's 1st coordinate",
+ found |= key_debug(brw, "GL_CLAMP enabled on any texture unit's 1st coordinate",
old_key->gl_clamp_mask[0], key->gl_clamp_mask[0]);
- found |= key_debug("GL_CLAMP enabled on any texture unit's 2nd coordinate",
+ found |= key_debug(brw, "GL_CLAMP enabled on any texture unit's 2nd coordinate",
old_key->gl_clamp_mask[1], key->gl_clamp_mask[1]);
- found |= key_debug("GL_CLAMP enabled on any texture unit's 3rd coordinate",
+ found |= key_debug(brw, "GL_CLAMP enabled on any texture unit's 3rd coordinate",
old_key->gl_clamp_mask[2], key->gl_clamp_mask[2]);
- found |= key_debug("GL_MESA_ycbcr texturing\n",
+ found |= key_debug(brw, "GL_MESA_ycbcr texturing\n",
old_key->yuvtex_mask, key->yuvtex_mask);
- found |= key_debug("GL_MESA_ycbcr UV swapping\n",
+ found |= key_debug(brw, "GL_MESA_ycbcr UV swapping\n",
old_key->yuvtex_swap_mask, key->yuvtex_swap_mask);
return found;
}
if (!c) {
- perf_debug(" Didn't find previous compile in the shader cache for "
- "debug\n");
+ perf_debug(" Didn't find previous compile in the shader cache for debug\n");
return;
}
- found |= key_debug("alphatest, computed depth, depth test, or depth write",
+ found |= key_debug(brw, "alphatest, computed depth, depth test, or "
+ "depth write",
old_key->iz_lookup, key->iz_lookup);
- found |= key_debug("depth statistics", old_key->stats_wm, key->stats_wm);
- found |= key_debug("flat shading", old_key->flat_shade, key->flat_shade);
- found |= key_debug("number of color buffers", old_key->nr_color_regions, key->nr_color_regions);
- found |= key_debug("sample alpha to coverage", old_key->sample_alpha_to_coverage, key->sample_alpha_to_coverage);
- found |= key_debug("rendering to FBO", old_key->render_to_fbo, key->render_to_fbo);
- found |= key_debug("fragment color clamping", old_key->clamp_fragment_color, key->clamp_fragment_color);
- found |= key_debug("line smoothing", old_key->line_aa, key->line_aa);
- found |= key_debug("proj_attrib_mask", old_key->proj_attrib_mask, key->proj_attrib_mask);
- found |= key_debug("renderbuffer height", old_key->drawable_height, key->drawable_height);
- found |= key_debug("vertex shader outputs", old_key->vp_outputs_written, key->vp_outputs_written);
-
- found |= brw_debug_recompile_sampler_key(&old_key->tex, &key->tex);
+ found |= key_debug(brw, "depth statistics",
+ old_key->stats_wm, key->stats_wm);
+ found |= key_debug(brw, "flat shading",
+ old_key->flat_shade, key->flat_shade);
+ found |= key_debug(brw, "number of color buffers",
+ old_key->nr_color_regions, key->nr_color_regions);
+ found |= key_debug(brw, "MRT alpha test or alpha-to-coverage",
+ old_key->replicate_alpha, key->replicate_alpha);
+ found |= key_debug(brw, "rendering to FBO",
+ old_key->render_to_fbo, key->render_to_fbo);
+ found |= key_debug(brw, "fragment color clamping",
+ old_key->clamp_fragment_color, key->clamp_fragment_color);
+ found |= key_debug(brw, "line smoothing",
+ old_key->line_aa, key->line_aa);
+ found |= key_debug(brw, "renderbuffer height",
+ old_key->drawable_height, key->drawable_height);
+ found |= key_debug(brw, "input slots valid",
+ old_key->input_slots_valid, key->input_slots_valid);
+
+ found |= brw_debug_recompile_sampler_key(brw, &old_key->tex, &key->tex);
if (!found) {
perf_debug(" Something else\n");
void
brw_populate_sampler_prog_key_data(struct gl_context *ctx,
const struct gl_program *prog,
+ unsigned sampler_count,
struct brw_sampler_prog_key_data *key)
{
- struct intel_context *intel = intel_context(ctx);
+ struct brw_context *brw = brw_context(ctx);
- for (int s = 0; s < MAX_SAMPLERS; s++) {
+ for (int s = 0; s < sampler_count; s++) {
key->swizzles[s] = SWIZZLE_NOOP;
if (!(prog->SamplersUsed & (1 << s)))
/* Haswell handles texture swizzling as surface format overrides
* (except for GL_ALPHA); all other platforms need MOVs in the shader.
*/
- if (!intel->is_haswell || alpha_depth)
- key->swizzles[s] = brw_get_texture_swizzle(t);
+ if (!brw->is_haswell || alpha_depth)
+ key->swizzles[s] = brw_get_texture_swizzle(ctx, t);
if (img->InternalFormat == GL_YCBCR_MESA) {
key->yuvtex_mask |= 1 << s;
static void brw_wm_populate_key( struct brw_context *brw,
struct brw_wm_prog_key *key )
{
- struct gl_context *ctx = &brw->intel.ctx;
- struct intel_context *intel = &brw->intel;
+ struct gl_context *ctx = &brw->ctx;
/* BRW_NEW_FRAGMENT_PROGRAM */
const struct brw_fragment_program *fp =
(struct brw_fragment_program *)brw->fragment_program;
/* Build the index for table lookup
*/
- if (intel->gen < 6) {
+ if (brw->gen < 6) {
/* _NEW_COLOR */
if (fp->program.UsesKill || ctx->Color.AlphaEnabled)
lookup |= IZ_PS_KILL_ALPHATEST_BIT;
if (ctx->Depth.Test && ctx->Depth.Mask) /* ?? */
lookup |= IZ_DEPTH_WRITE_ENABLE_BIT;
- /* _NEW_STENCIL */
+ /* _NEW_STENCIL | _NEW_BUFFERS */
if (ctx->Stencil._Enabled) {
lookup |= IZ_STENCIL_TEST_ENABLE_BIT;
/* _NEW_LINE, _NEW_POLYGON, BRW_NEW_REDUCED_PRIMITIVE */
if (ctx->Line.SmoothFlag) {
- if (brw->intel.reduced_primitive == GL_LINES) {
+ if (brw->reduced_primitive == GL_LINES) {
line_aa = AA_ALWAYS;
}
- else if (brw->intel.reduced_primitive == GL_TRIANGLES) {
+ else if (brw->reduced_primitive == GL_TRIANGLES) {
if (ctx->Polygon.FrontMode == GL_LINE) {
line_aa = AA_SOMETIMES;
key->line_aa = line_aa;
- if (intel->gen < 6)
- key->stats_wm = brw->intel.stats_wm;
-
- /* BRW_NEW_WM_INPUT_DIMENSIONS */
- /* Only set this for fixed function. The optimization it enables isn't
- * useful for programs using shaders.
- */
- if (ctx->Shader.CurrentFragmentProgram)
- key->proj_attrib_mask = 0xffffffff;
- else
- key->proj_attrib_mask = brw->wm.input_size_masks[4-1];
+ if (brw->gen < 6)
+ key->stats_wm = brw->stats_wm;
/* _NEW_LIGHT */
key->flat_shade = (ctx->Light.ShadeModel == GL_FLAT);
key->clamp_fragment_color = ctx->Color._ClampFragmentColor;
/* _NEW_TEXTURE */
- brw_populate_sampler_prog_key_data(ctx, prog, &key->tex);
+ brw_populate_sampler_prog_key_data(ctx, prog, brw->wm.sampler_count,
+ &key->tex);
/* _NEW_BUFFERS */
/*
* For DRI2 the origin_x/y will always be (0,0) but we still need the
* drawable height in order to invert the Y axis.
*/
- if (fp->program.Base.InputsRead & FRAG_BIT_WPOS) {
+ if (fp->program.Base.InputsRead & VARYING_BIT_POS) {
key->drawable_height = ctx->DrawBuffer->Height;
}
- if ((fp->program.Base.InputsRead & FRAG_BIT_WPOS) || program_uses_dfdy) {
+ if ((fp->program.Base.InputsRead & VARYING_BIT_POS) || program_uses_dfdy) {
key->render_to_fbo = _mesa_is_user_fbo(ctx->DrawBuffer);
}
/* _NEW_BUFFERS */
key->nr_color_regions = ctx->DrawBuffer->_NumColorDrawBuffers;
- /* _NEW_MULTISAMPLE */
- key->sample_alpha_to_coverage = ctx->Multisample.SampleAlphaToCoverage;
- /* CACHE_NEW_VS_PROG */
- if (intel->gen < 6)
- key->vp_outputs_written = brw->vs.prog_data->outputs_written;
+ /* _NEW_MULTISAMPLE, _NEW_COLOR, _NEW_BUFFERS */
+ key->replicate_alpha = ctx->DrawBuffer->_NumColorDrawBuffers > 1 &&
+ (ctx->Multisample.SampleAlphaToCoverage || ctx->Color.AlphaEnabled);
+
+ /* BRW_NEW_VUE_MAP_GEOM_OUT */
+ if (brw->gen < 6)
+ key->input_slots_valid = brw->vue_map_geom_out.slots_valid;
/* The unique fragment program ID */
key->program_string_id = fp->id;
static void
brw_upload_wm_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_wm_prog_key key;
struct brw_fragment_program *fp = (struct brw_fragment_program *)
brw->fragment_program;
_NEW_TEXTURE |
_NEW_MULTISAMPLE),
.brw = (BRW_NEW_FRAGMENT_PROGRAM |
- BRW_NEW_WM_INPUT_DIMENSIONS |
- BRW_NEW_REDUCED_PRIMITIVE),
- .cache = CACHE_NEW_VS_PROG,
+ BRW_NEW_REDUCED_PRIMITIVE |
+ BRW_NEW_VUE_MAP_GEOM_OUT |
+ BRW_NEW_STATS_WM)
},
.emit = brw_upload_wm_prog
};