static void
gen6_upload_wm_push_constants(struct brw_context *brw)
{
- struct gl_context *ctx = &brw->ctx;
+ struct brw_stage_state *stage_state = &brw->wm.base;
/* BRW_NEW_FRAGMENT_PROGRAM */
const struct brw_fragment_program *fp =
brw_fragment_program_const(brw->fragment_program);
/* CACHE_NEW_WM_PROG */
const struct brw_wm_prog_data *prog_data = brw->wm.prog_data;
- /* Updates the ParameterValues[i] pointers for all parameters of the
- * basic type of PROGRAM_STATE_VAR.
- */
- /* XXX: Should this happen somewhere before to get our state flag set? */
- _mesa_load_state_parameters(ctx, fp->program.Base.Parameters);
-
- if (prog_data->base.nr_params == 0) {
- brw->wm.base.push_const_size = 0;
- } else {
- float *constants;
- unsigned int i;
-
- constants = brw_state_batch(brw, AUB_TRACE_WM_CONSTANTS,
- prog_data->base.nr_params * sizeof(float),
- 32, &brw->wm.base.push_const_offset);
-
- for (i = 0; i < prog_data->base.nr_params; i++) {
- constants[i] = *prog_data->base.param[i];
- }
-
- if (0) {
- fprintf(stderr, "WM constants:\n");
- for (i = 0; i < prog_data->base.nr_params; i++) {
- if ((i & 7) == 0)
- fprintf(stderr, "g%d: ",
- prog_data->base.dispatch_grf_start_reg + i / 8);
- fprintf(stderr, "%8f ", constants[i]);
- if ((i & 7) == 7)
- fprintf(stderr, "\n");
- }
- if ((i & 7) != 0)
- fprintf(stderr, "\n");
- fprintf(stderr, "\n");
- }
-
- brw->wm.base.push_const_size = ALIGN(prog_data->base.nr_params, 8) / 8;
- }
+ gen6_upload_push_constants(brw, &fp->program.Base, &prog_data->base,
+ stage_state, AUB_TRACE_WM_CONSTANTS);
if (brw->gen >= 7) {
gen7_upload_constant_state(brw, &brw->wm.base, true,
upload_wm_state(struct brw_context *brw)
{
struct gl_context *ctx = &brw->ctx;
+ /* BRW_NEW_FRAGMENT_PROGRAM */
const struct brw_fragment_program *fp =
brw_fragment_program_const(brw->fragment_program);
- uint32_t dw2, dw4, dw5, dw6;
+ /* CACHE_NEW_WM_PROG */
+ const struct brw_wm_prog_data *prog_data = brw->wm.prog_data;
+ uint32_t dw2, dw4, dw5, dw6, ksp0, ksp2;
/* _NEW_BUFFERS */
bool multisampled_fbo = ctx->DrawBuffer->Visual.samples > 1;
- /* CACHE_NEW_WM_PROG */
- if (brw->wm.prog_data->base.nr_params == 0) {
+ /* We can't fold this into gen6_upload_wm_push_constants(), because
+ * according to the SNB PRM, vol 2 part 1 section 7.2.2
+ * (3DSTATE_CONSTANT_PS [DevSNB]):
+ *
+ * "[DevSNB]: This packet must be followed by WM_STATE."
+ */
+ if (prog_data->base.nr_params == 0) {
/* Disable the push constant buffers. */
BEGIN_BATCH(5);
OUT_BATCH(_3DSTATE_CONSTANT_PS << 16 | (5 - 2));
ADVANCE_BATCH();
}
- dw2 = dw4 = dw5 = dw6 = 0;
+ dw2 = dw4 = dw5 = dw6 = ksp2 = 0;
dw4 |= GEN6_WM_STATISTICS_ENABLE;
dw5 |= GEN6_WM_LINE_AA_WIDTH_1_0;
dw5 |= GEN6_WM_LINE_END_CAP_AA_WIDTH_0_5;
dw2 |= (ALIGN(brw->wm.base.sampler_count, 4) / 4) <<
GEN6_WM_SAMPLER_COUNT_SHIFT;
- /* CACHE_NEW_WM_PROG */
- dw2 |= ((brw->wm.prog_data->base.binding_table.size_bytes / 4) <<
+ dw2 |= ((prog_data->base.binding_table.size_bytes / 4) <<
GEN6_WM_BINDING_TABLE_ENTRY_COUNT_SHIFT);
dw5 |= (brw->max_wm_threads - 1) << GEN6_WM_MAX_THREADS_SHIFT;
- /* CACHE_NEW_WM_PROG */
-
/* In case of non 1x per sample shading, only one of SIMD8 and SIMD16
* should be enabled. We do 'SIMD16 only' dispatch if a SIMD16 shader
* is successfully compiled. In majority of the cases that bring us
_mesa_get_min_invocations_per_fragment(ctx, brw->fragment_program, false);
assert(min_inv_per_frag >= 1);
- if (brw->wm.prog_data->prog_offset_16) {
+ if (prog_data->prog_offset_16) {
dw5 |= GEN6_WM_16_DISPATCH_ENABLE;
if (min_inv_per_frag == 1) {
dw5 |= GEN6_WM_8_DISPATCH_ENABLE;
- dw4 |= (brw->wm.prog_data->base.dispatch_grf_start_reg <<
+ dw4 |= (prog_data->base.dispatch_grf_start_reg <<
GEN6_WM_DISPATCH_START_GRF_SHIFT_0);
- dw4 |= (brw->wm.prog_data->dispatch_grf_start_reg_16 <<
+ dw4 |= (prog_data->dispatch_grf_start_reg_16 <<
GEN6_WM_DISPATCH_START_GRF_SHIFT_2);
- } else
- dw4 |= (brw->wm.prog_data->dispatch_grf_start_reg_16 <<
+ ksp0 = brw->wm.base.prog_offset;
+ ksp2 = brw->wm.base.prog_offset + prog_data->prog_offset_16;
+ } else {
+ dw4 |= (prog_data->dispatch_grf_start_reg_16 <<
GEN6_WM_DISPATCH_START_GRF_SHIFT_0);
+ ksp0 = brw->wm.base.prog_offset + prog_data->prog_offset_16;
+ }
}
else {
dw5 |= GEN6_WM_8_DISPATCH_ENABLE;
- dw4 |= (brw->wm.prog_data->base.dispatch_grf_start_reg <<
+ dw4 |= (prog_data->base.dispatch_grf_start_reg <<
GEN6_WM_DISPATCH_START_GRF_SHIFT_0);
+ ksp0 = brw->wm.base.prog_offset;
}
/* CACHE_NEW_WM_PROG | _NEW_COLOR */
- if (brw->wm.prog_data->dual_src_blend &&
+ if (prog_data->dual_src_blend &&
(ctx->Color.BlendEnabled & 1) &&
ctx->Color.Blend[0]._UsesDualSrc) {
dw5 |= GEN6_WM_DUAL_SOURCE_BLEND_ENABLE;
dw5 |= GEN6_WM_USES_SOURCE_DEPTH | GEN6_WM_USES_SOURCE_W;
if (fp->program.Base.OutputsWritten & BITFIELD64_BIT(FRAG_RESULT_DEPTH))
dw5 |= GEN6_WM_COMPUTED_DEPTH;
- /* CACHE_NEW_WM_PROG */
- dw6 |= brw->wm.prog_data->barycentric_interp_modes <<
+ dw6 |= prog_data->barycentric_interp_modes <<
GEN6_WM_BARYCENTRIC_INTERPOLATION_MODE_SHIFT;
/* _NEW_COLOR, _NEW_MULTISAMPLE */
if (fp->program.UsesKill || ctx->Color.AlphaEnabled ||
ctx->Multisample.SampleAlphaToCoverage ||
- brw->wm.prog_data->uses_omask)
+ prog_data->uses_omask)
dw5 |= GEN6_WM_KILL_ENABLE;
/* _NEW_BUFFERS | _NEW_COLOR */
* Target Write messages. If present, the oMask data is used to mask off
* samples."
*/
- if(brw->wm.prog_data->uses_omask)
+ if (prog_data->uses_omask)
dw5 |= GEN6_WM_OMASK_TO_RENDER_TARGET;
- /* CACHE_NEW_WM_PROG */
- dw6 |= brw->wm.prog_data->num_varying_inputs <<
+ dw6 |= prog_data->num_varying_inputs <<
GEN6_WM_NUM_SF_OUTPUTS_SHIFT;
if (multisampled_fbo) {
/* _NEW_MULTISAMPLE */
* We only require XY sample offsets. So, this recommendation doesn't
* look useful at the moment. We might need this in future.
*/
- if (brw->wm.prog_data->uses_pos_offset)
+ if (prog_data->uses_pos_offset)
dw6 |= GEN6_WM_POSOFFSET_SAMPLE;
else
dw6 |= GEN6_WM_POSOFFSET_NONE;
BEGIN_BATCH(9);
OUT_BATCH(_3DSTATE_WM << 16 | (9 - 2));
- if (brw->wm.prog_data->prog_offset_16 && min_inv_per_frag > 1)
- OUT_BATCH(brw->wm.base.prog_offset + brw->wm.prog_data->prog_offset_16);
- else
- OUT_BATCH(brw->wm.base.prog_offset);
+ OUT_BATCH(ksp0);
OUT_BATCH(dw2);
- if (brw->wm.prog_data->total_scratch) {
+ if (prog_data->base.total_scratch) {
OUT_RELOC(brw->wm.base.scratch_bo,
I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
- ffs(brw->wm.prog_data->total_scratch) - 11);
+ ffs(prog_data->base.total_scratch) - 11);
} else {
OUT_BATCH(0);
}
OUT_BATCH(dw5);
OUT_BATCH(dw6);
OUT_BATCH(0); /* kernel 1 pointer */
- /* kernel 2 pointer */
- OUT_BATCH(brw->wm.base.prog_offset + brw->wm.prog_data->prog_offset_16);
+ OUT_BATCH(ksp2);
ADVANCE_BATCH();
}
projects / mesa.git / blobdiff