/*
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
**********************************************************************/
/*
* Authors:
- * Keith Whitwell <keith@tungstengraphics.com>
+ * Keith Whitwell <keithw@vmware.com>
*/
#include "brw_context.h"
static unsigned
brw_compute_barycentric_interp_modes(struct brw_context *brw,
bool shade_model_flat,
+ bool persample_shading,
const struct gl_fragment_program *fprog)
{
unsigned barycentric_interp_modes = 0;
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_centroid = (fprog->IsCentroid & BITFIELD64_BIT(attr)) &&
+ !persample_shading;
+ bool is_sample = (fprog->IsSample & BITFIELD64_BIT(attr)) ||
+ persample_shading;
bool is_gl_Color = attr == VARYING_SLOT_COL0 || attr == VARYING_SLOT_COL1;
/* Ignore unused inputs. */
if (is_centroid) {
barycentric_interp_modes |=
1 << BRW_WM_NONPERSPECTIVE_CENTROID_BARYCENTRIC;
+ } else if (is_sample) {
+ barycentric_interp_modes |=
+ 1 << BRW_WM_NONPERSPECTIVE_SAMPLE_BARYCENTRIC;
}
- if (!is_centroid || brw->needs_unlit_centroid_workaround) {
+ if ((!is_centroid && !is_sample) ||
+ brw->needs_unlit_centroid_workaround) {
barycentric_interp_modes |=
1 << BRW_WM_NONPERSPECTIVE_PIXEL_BARYCENTRIC;
}
if (is_centroid) {
barycentric_interp_modes |=
1 << BRW_WM_PERSPECTIVE_CENTROID_BARYCENTRIC;
+ } else if (is_sample) {
+ barycentric_interp_modes |=
+ 1 << BRW_WM_PERSPECTIVE_SAMPLE_BARYCENTRIC;
}
- if (!is_centroid || brw->needs_unlit_centroid_workaround) {
+ if ((!is_centroid && !is_sample) ||
+ brw->needs_unlit_centroid_workaround) {
barycentric_interp_modes |=
1 << BRW_WM_PERSPECTIVE_PIXEL_BARYCENTRIC;
}
const struct brw_wm_prog_data *a = in_a;
const struct brw_wm_prog_data *b = in_b;
- /* Compare all the struct (including the base) up to the pointers. */
- if (memcmp(a, b, offsetof(struct brw_wm_prog_data, param)))
- return false;
-
- if (memcmp(a->param, b->param, a->nr_params * sizeof(void *)))
+ /* Compare the base structure. */
+ if (!brw_stage_prog_data_compare(&a->base, &b->base))
return false;
- if (memcmp(a->pull_param, b->pull_param, a->nr_pull_params * sizeof(void *)))
+ /* Compare the rest of the structure. */
+ const unsigned offset = sizeof(struct brw_stage_prog_data);
+ if (memcmp(((char *) a) + offset, ((char *) b) + offset,
+ sizeof(struct brw_wm_prog_data) - offset))
return false;
return true;
}
-void
-brw_wm_prog_data_free(const void *in_prog_data)
-{
- const struct brw_wm_prog_data *prog_data = in_prog_data;
-
- ralloc_free((void *)prog_data->param);
- ralloc_free((void *)prog_data->pull_param);
-}
-
/**
* All Mesa program -> GPU code generation goes through this function.
* Depending on the instructions used (i.e. flow control instructions)
struct brw_fragment_program *fp,
struct brw_wm_prog_key *key)
{
+ struct gl_context *ctx = &brw->ctx;
struct brw_wm_compile *c;
const GLuint *program;
struct gl_shader *fs = 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);
+ param_count += 2 * ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxTextureImageUnits;
+ c->prog_data.base.param = rzalloc_array(NULL, const float *, param_count);
+ c->prog_data.base.pull_param =
+ rzalloc_array(NULL, const float *, param_count);
+ c->prog_data.base.nr_params = param_count;
memcpy(&c->key, key, sizeof(*key));
c->prog_data.barycentric_interp_modes =
brw_compute_barycentric_interp_modes(brw, c->key.flat_shade,
+ c->key.persample_shading,
&fp->program);
program = brw_wm_fs_emit(brw, c, &fp->program, prog, &program_size);
- if (program == NULL)
+ if (program == NULL) {
+ ralloc_free(c);
return false;
+ }
- /* Scratch space is used for register spilling */
- if (c->last_scratch) {
- perf_debug("Fragment shader triggered register spilling. "
- "Try reducing the number of live scalar values to "
- "improve performance.\n");
-
- c->prog_data.total_scratch = brw_get_scratch_size(c->last_scratch);
-
+ if (c->prog_data.total_scratch) {
brw_get_scratch_bo(brw, &brw->wm.base.scratch_bo,
c->prog_data.total_scratch * brw->max_wm_threads);
}
old_key->gl_clamp_mask[1], key->gl_clamp_mask[1]);
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(brw, "GL_MESA_ycbcr texturing\n",
- old_key->yuvtex_mask, key->yuvtex_mask);
- found |= key_debug(brw, "GL_MESA_ycbcr UV swapping\n",
- old_key->yuvtex_swap_mask, key->yuvtex_swap_mask);
found |= key_debug(brw, "gather channel quirk on any texture unit",
old_key->gather_channel_quirk_mask, key->gather_channel_quirk_mask);
}
}
+static uint8_t
+gen6_gather_workaround(GLenum internalformat)
+{
+ switch (internalformat) {
+ case GL_R8I: return WA_SIGN | WA_8BIT;
+ case GL_R8UI: return WA_8BIT;
+ case GL_R16I: return WA_SIGN | WA_16BIT;
+ case GL_R16UI: return WA_16BIT;
+ /* note that even though GL_R32I and GL_R32UI have format overrides
+ * in the surface state, there is no shader w/a required */
+ default: return 0;
+ }
+}
+
void
brw_populate_sampler_prog_key_data(struct gl_context *ctx,
const struct gl_program *prog,
int unit_id = prog->SamplerUnits[s];
const struct gl_texture_unit *unit = &ctx->Texture.Unit[unit_id];
- if (unit->_ReallyEnabled && unit->_Current->Target != GL_TEXTURE_BUFFER) {
+ if (unit->_Current && unit->_Current->Target != GL_TEXTURE_BUFFER) {
const struct gl_texture_object *t = unit->_Current;
const struct gl_texture_image *img = t->Image[0][t->BaseLevel];
struct gl_sampler_object *sampler = _mesa_get_samplerobj(ctx, unit_id);
if (alpha_depth || (brw->gen < 8 && !brw->is_haswell))
key->swizzles[s] = brw_get_texture_swizzle(ctx, t);
- if (img->InternalFormat == GL_YCBCR_MESA) {
- key->yuvtex_mask |= 1 << s;
- if (img->TexFormat == MESA_FORMAT_YCBCR)
- key->yuvtex_swap_mask |= 1 << s;
- }
-
if (sampler->MinFilter != GL_NEAREST &&
sampler->MagFilter != GL_NEAREST) {
if (sampler->WrapS == GL_CLAMP)
key->gather_channel_quirk_mask |= 1 << s;
}
+ /* Gen6's gather4 is broken for UINT/SINT; we treat them as
+ * UNORM/FLOAT instead and fix it in the shader.
+ */
+ if (brw->gen == 6 && prog->UsesGather) {
+ key->gen6_gather_wa[s] = gen6_gather_workaround(img->InternalFormat);
+ }
+
/* If this is a multisample sampler, and uses the CMS MSAA layout,
* then we need to emit slightly different code to first sample the
* MCS surface.
(ctx->Multisample.SampleAlphaToCoverage || ctx->Color.AlphaEnabled);
/* _NEW_BUFFERS _NEW_MULTISAMPLE */
+ /* Ignore sample qualifier while computing this flag. */
+ key->persample_shading =
+ _mesa_get_min_invocations_per_fragment(ctx, &fp->program, true) > 1;
+
key->compute_pos_offset =
- _mesa_get_min_invocations_per_fragment(ctx, &fp->program) > 1 &&
+ _mesa_get_min_invocations_per_fragment(ctx, &fp->program, false) > 1 &&
fp->program.Base.SystemValuesRead & SYSTEM_BIT_SAMPLE_POS;
key->compute_sample_id =
if (!brw_search_cache(&brw->cache, BRW_WM_PROG,
&key, sizeof(key),
&brw->wm.base.prog_offset, &brw->wm.prog_data)) {
- bool success = do_wm_prog(brw, ctx->Shader._CurrentFragmentProgram, fp,
+ bool success = do_wm_prog(brw, ctx->_Shader->_CurrentFragmentProgram, fp,
&key);
(void) success;
assert(success);