#include "main/compiler.h"
+#include "main/context.h"
#include "brw_context.h"
#include "brw_vs.h"
#include "brw_util.h"
#include "program/prog_print.h"
#include "program/prog_parameter.h"
#include "brw_nir.h"
+#include "brw_program.h"
#include "util/ralloc.h"
+GLbitfield64
+brw_vs_outputs_written(struct brw_context *brw, struct brw_vs_prog_key *key,
+ GLbitfield64 user_varyings)
+{
+ GLbitfield64 outputs_written = user_varyings;
+
+ if (key->copy_edgeflag) {
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_EDGE);
+ }
+
+ 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 (unsigned i = 0; i < 8; i++) {
+ if (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);
+ }
+
+ /* 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.
+ */
+ if (key->nr_userclip_plane_consts > 0) {
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST0);
+ outputs_written |= BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST1);
+ }
+
+ return outputs_written;
+}
+
bool
brw_codegen_vs_prog(struct brw_context *brw,
struct gl_shader_program *prog,
struct brw_vertex_program *vp,
struct brw_vs_prog_key *key)
{
+ const struct brw_compiler *compiler = brw->intelScreen->compiler;
GLuint program_size;
const GLuint *program;
struct brw_vs_prog_data prog_data;
struct brw_stage_prog_data *stage_prog_data = &prog_data.base.base;
void *mem_ctx;
- int i;
struct brw_shader *vs = NULL;
bool start_busy = false;
double start_time = 0;
- if (!vp->program.Base.nir) {
- /* Normally we generate NIR in LinkShader() or
- * ProgramStringNotify(), but Mesa's fixed-function vertex program
- * handling doesn't notify the driver at all. Just do it here, at
- * the last minute, even though it's lame.
- */
- assert(vp->program.Base.Id == 0 && prog == NULL);
- vp->program.Base.nir =
- brw_create_nir(brw, NULL, &vp->program.Base, MESA_SHADER_VERTEX,
- brw->intelScreen->compiler->scalar_vs);
- }
-
if (prog)
vs = (struct brw_shader *) prog->_LinkedShaders[MESA_SHADER_VERTEX];
* prog_data associated with the compiled program, and which will be freed
* by the state cache.
*/
- int param_count = vp->program.Base.nir->num_uniforms;
- if (!brw->intelScreen->compiler->scalar_vs)
- param_count *= 4;
+ int param_count = vp->program.Base.nir->num_uniforms / 4;
if (vs)
prog_data.base.base.nr_image_params = vs->base.NumImages;
if (prog) {
brw_nir_setup_glsl_uniforms(vp->program.Base.nir, prog, &vp->program.Base,
&prog_data.base.base,
- brw->intelScreen->compiler->scalar_vs);
+ compiler->scalar_stage[MESA_SHADER_VERTEX]);
} else {
brw_nir_setup_arb_uniforms(vp->program.Base.nir, &vp->program.Base,
&prog_data.base.base);
}
- GLbitfield64 outputs_written = vp->program.Base.OutputsWritten;
+ GLbitfield64 outputs_written =
+ brw_vs_outputs_written(brw, key, vp->program.Base.OutputsWritten);
prog_data.inputs_read = vp->program.Base.InputsRead;
if (key->copy_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 (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);
- }
-
- /* 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.
- */
- if (key->nr_userclip_plane_consts > 0) {
- outputs_written |= BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST0);
- outputs_written |= BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST1);
- }
+ prog_data.base.cull_distance_mask =
+ ((1 << vp->program.Base.CullDistanceArraySize) - 1) <<
+ vp->program.Base.ClipDistanceArraySize;
brw_compute_vue_map(brw->intelScreen->devinfo,
&prog_data.base.vue_map, outputs_written,
- prog ? prog->SeparateShader : false);
-
- unsigned nr_attributes = _mesa_bitcount_64(prog_data.inputs_read);
-
- /* gl_VertexID and gl_InstanceID are system values, but arrive via an
- * incoming vertex attribute. So, add an extra slot.
- */
- if (vp->program.Base.SystemValuesRead &
- (BITFIELD64_BIT(SYSTEM_VALUE_VERTEX_ID_ZERO_BASE) |
- BITFIELD64_BIT(SYSTEM_VALUE_INSTANCE_ID))) {
- nr_attributes++;
- }
-
- /* The 3DSTATE_VS documentation lists the lower bound on "Vertex URB Entry
- * Read Length" as 1 in vec4 mode, and 0 in SIMD8 mode. Empirically, in
- * vec4 mode, the hardware appears to wedge unless we read something.
- */
- if (brw->intelScreen->compiler->scalar_vs)
- prog_data.base.urb_read_length = DIV_ROUND_UP(nr_attributes, 2);
- else
- prog_data.base.urb_read_length = DIV_ROUND_UP(MAX2(nr_attributes, 1), 2);
-
- prog_data.nr_attributes = nr_attributes;
-
- /* Since vertex shaders reuse the same VUE entry for inputs and outputs
- * (overwriting the original contents), we need to make sure the size is
- * the larger of the two.
- */
- const unsigned vue_entries =
- MAX2(nr_attributes, prog_data.base.vue_map.num_slots);
-
- if (brw->gen == 6)
- prog_data.base.urb_entry_size = DIV_ROUND_UP(vue_entries, 8);
- else
- prog_data.base.urb_entry_size = DIV_ROUND_UP(vue_entries, 4);
+ prog ? prog->SeparateShader ||
+ prog->_LinkedShaders[MESA_SHADER_TESS_EVAL]
+ : false);
if (0) {
_mesa_fprint_program_opt(stderr, &vp->program.Base, PROG_PRINT_DEBUG,
start_time = get_time();
}
- if (unlikely(INTEL_DEBUG & DEBUG_VS))
- brw_dump_ir("vertex", prog, &vs->base, &vp->program.Base);
+ if (unlikely(INTEL_DEBUG & DEBUG_VS)) {
+ brw_dump_ir("vertex", prog, vs ? &vs->base : NULL, &vp->program.Base);
+
+ fprintf(stderr, "VS Output ");
+ brw_print_vue_map(stderr, &prog_data.base.vue_map);
+ }
int st_index = -1;
if (INTEL_DEBUG & DEBUG_SHADER_TIME)
/* Emit GEN4 code.
*/
- program = brw_vs_emit(brw, mem_ctx, key, &prog_data,
- &vp->program, prog, st_index, &program_size);
+ char *error_str;
+ program = brw_compile_vs(compiler, brw, mem_ctx, key,
+ &prog_data, vp->program.Base.nir,
+ brw_select_clip_planes(&brw->ctx),
+ !_mesa_is_gles3(&brw->ctx),
+ st_index, &program_size, &error_str);
if (program == NULL) {
+ if (prog) {
+ prog->LinkStatus = false;
+ ralloc_strcat(&prog->InfoLog, error_str);
+ }
+
+ _mesa_problem(NULL, "Failed to compile vertex shader: %s\n", error_str);
+
ralloc_free(mem_ctx);
return false;
}
}
/* Scratch space is used for register spilling */
- if (prog_data.base.base.total_scratch) {
- brw_get_scratch_bo(brw, &brw->vs.base.scratch_bo,
- prog_data.base.base.total_scratch *
- brw->max_vs_threads);
- }
+ brw_alloc_stage_scratch(brw, &brw->vs.base,
+ prog_data.base.base.total_scratch,
+ brw->max_vs_threads);
brw_upload_cache(&brw->cache, BRW_CACHE_VS_PROG,
key, sizeof(struct brw_vs_prog_key),
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,
}
static bool
-brw_vs_state_dirty(struct brw_context *brw)
+brw_vs_state_dirty(const struct brw_context *brw)
{
return brw_state_dirty(brw,
_NEW_BUFFERS |
struct brw_vertex_program *vp =
(struct brw_vertex_program *)brw->vertex_program;
struct gl_program *prog = (struct gl_program *) brw->vertex_program;
- int i;
memset(key, 0, sizeof(*key));
key->program_string_id = vp->id;
if (ctx->Transform.ClipPlanesEnabled != 0 &&
- ctx->API == API_OPENGL_COMPAT &&
- !vp->program.Base.UsesClipDistanceOut) {
+ (ctx->API == API_OPENGL_COMPAT ||
+ ctx->API == API_OPENGLES) &&
+ vp->program.Base.ClipDistanceArraySize == 0) {
key->nr_userclip_plane_consts =
_mesa_logbase2(ctx->Transform.ClipPlanesEnabled) + 1;
}
/* _NEW_POINT */
if (brw->gen < 6 && ctx->Point.PointSprite) {
- for (i = 0; i < 8; i++) {
- if (ctx->Point.CoordReplace[i])
- key->point_coord_replace |= (1 << i);
- }
+ key->point_coord_replace = ctx->Point.CoordReplace & 0xff;
}
/* _NEW_TEXTURE */
- brw_populate_sampler_prog_key_data(ctx, prog, brw->vs.base.sampler_count,
- &key->tex);
+ brw_populate_sampler_prog_key_data(ctx, prog, &key->tex);
/* BRW_NEW_VS_ATTRIB_WORKAROUNDS */
- memcpy(key->gl_attrib_wa_flags, brw->vb.attrib_wa_flags,
- sizeof(brw->vb.attrib_wa_flags));
+ if (brw->gen < 8 && !brw->is_haswell) {
+ memcpy(key->gl_attrib_wa_flags, brw->vb.attrib_wa_flags,
+ sizeof(brw->vb.attrib_wa_flags));
+ }
}
void