#include "main/macros.h"
#include "main/shaderobj.h"
-#include "main/uniforms.h"
#include "program/prog_parameter.h"
#include "program/prog_print.h"
#include "program/prog_optimize.h"
#include "brw_wm.h"
}
#include "brw_fs.h"
+#include "main/uniforms.h"
#include "glsl/glsl_types.h"
#include "glsl/ir_optimization.h"
ir->operands[operand]->print();
printf("\n");
}
+ assert(this->result.is_valid_3src());
op[operand] = this->result;
/* Matrix expression operands should have been broken down to vector
break;
case ir_unop_neg:
op[0].negate = !op[0].negate;
- this->result = op[0];
+ emit(MOV(this->result, op[0]));
break;
case ir_unop_abs:
op[0].abs = true;
op[0].negate = false;
- this->result = op[0];
+ emit(MOV(this->result, op[0]));
break;
case ir_unop_sign:
temp = fs_reg(this, ir->type);
if (brw->gen >= 7 && dispatch_width == 16)
fail("16-wide explicit accumulator operands unsupported\n");
- struct brw_reg acc = retype(brw_acc_reg(), BRW_REGISTER_TYPE_D);
+ struct brw_reg acc = retype(brw_acc_reg(), this->result.type);
emit(MUL(acc, op[0], op[1]));
emit(MACH(reg_null_d, op[0], op[1]));
emit(MUL(this->result, op[0], op[1]));
}
break;
+ case ir_binop_imul_high: {
+ if (brw->gen >= 7 && dispatch_width == 16)
+ fail("16-wide explicit accumulator operands unsupported\n");
+
+ struct brw_reg acc = retype(brw_acc_reg(), this->result.type);
+
+ emit(MUL(acc, op[0], op[1]));
+ emit(MACH(this->result, op[0], op[1]));
+ break;
+ }
case ir_binop_div:
/* Floating point should be lowered by DIV_TO_MUL_RCP in the compiler. */
assert(ir->type->is_integer());
emit_math(SHADER_OPCODE_INT_QUOTIENT, this->result, op[0], op[1]);
break;
+ case ir_binop_carry: {
+ if (brw->gen >= 7 && dispatch_width == 16)
+ fail("16-wide explicit accumulator operands unsupported\n");
+
+ struct brw_reg acc = retype(brw_acc_reg(), BRW_REGISTER_TYPE_UD);
+
+ emit(ADDC(reg_null_ud, op[0], op[1]));
+ emit(MOV(this->result, fs_reg(acc)));
+ break;
+ }
+ case ir_binop_borrow: {
+ if (brw->gen >= 7 && dispatch_width == 16)
+ fail("16-wide explicit accumulator operands unsupported\n");
+
+ struct brw_reg acc = retype(brw_acc_reg(), BRW_REGISTER_TYPE_UD);
+
+ emit(SUBB(reg_null_ud, op[0], op[1]));
+ emit(MOV(this->result, fs_reg(acc)));
+ break;
+ }
case ir_binop_mod:
/* Floating point should be lowered by MOD_TO_FRACT in the compiler. */
assert(ir->type->is_integer());
assert(!"not reached: should be handled by lower_vector_insert()");
break;
+ case ir_binop_ldexp:
+ assert(!"not reached: should be handled by ldexp_to_arith()");
+ break;
+
case ir_unop_sqrt:
emit_math(SHADER_OPCODE_SQRT, this->result, op[0]);
break;
*/
ir_constant *uniform_block = ir->operands[0]->as_constant();
ir_constant *const_offset = ir->operands[1]->as_constant();
- fs_reg surf_index = fs_reg((unsigned)SURF_INDEX_WM_UBO(uniform_block->value.u[0]));
+ fs_reg surf_index = fs_reg(c->prog_data.base.binding_table.ubo_start +
+ uniform_block->value.u[0]);
if (const_offset) {
fs_reg packed_consts = fs_reg(this, glsl_type::float_type);
packed_consts.type = result.type;
break;
}
+ case ir_triop_fma:
+ /* Note that the instruction's argument order is reversed from GLSL
+ * and the IR.
+ */
+ emit(MAD(this->result, op[2], op[1], op[0]));
+ break;
+
case ir_triop_lrp:
emit_lrp(this->result, op[0], op[1], op[2]);
break;
+
+ case ir_triop_csel:
+ emit(CMP(reg_null_d, op[0], fs_reg(0), BRW_CONDITIONAL_NZ));
+ inst = emit(BRW_OPCODE_SEL, this->result, op[1], op[2]);
+ inst->predicate = BRW_PREDICATE_NORMAL;
+ break;
}
}
emit(MOV(fs_reg(MRF, base_mrf + mlen + i), coordinate));
coordinate.reg_offset++;
}
- /* gen4's SIMD8 sampler always has the slots for u,v,r present. */
+
+ /* gen4's SIMD8 sampler always has the slots for u,v,r present.
+ * the unused slots must be zeroed.
+ */
+ for (int i = ir->coordinate->type->vector_elements; i < 3; i++) {
+ emit(MOV(fs_reg(MRF, base_mrf + mlen + i), fs_reg(0.0f)));
+ }
mlen += 3;
if (ir->op == ir_tex) {
emit(MOV(fs_reg(MRF, base_mrf + mlen + i), coordinate));
coordinate.reg_offset++;
}
+ /* zero the others. */
+ for (int i = ir->coordinate->type->vector_elements; i<3; i++) {
+ emit(MOV(fs_reg(MRF, base_mrf + mlen + i), fs_reg(0.0f)));
+ }
/* gen4's SIMD8 sampler always has the slots for u,v,r present. */
mlen += 3;
} else if (ir->op == ir_txd) {
mlen += reg_width;
inst = emit(SHADER_OPCODE_TXS, dst);
break;
+ case ir_query_levels:
+ emit(MOV(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_UD), fs_reg(0u)));
+ mlen += reg_width;
+ inst = emit(SHADER_OPCODE_TXS, dst);
+ break;
case ir_txf:
mlen = header_present + 4 * reg_width;
emit(MOV(fs_reg(MRF, base_mrf + mlen - reg_width, BRW_REGISTER_TYPE_UD), lod));
case ir_lod:
inst = emit(SHADER_OPCODE_LOD, dst);
break;
+ case ir_tg4:
+ inst = emit(SHADER_OPCODE_TG4, dst);
+ break;
+ default:
+ fail("unrecognized texture opcode");
+ break;
}
inst->base_mrf = base_mrf;
inst->mlen = mlen;
fs_reg shadow_c, fs_reg lod, fs_reg lod2,
fs_reg sample_index)
{
- int mlen = 0;
- int base_mrf = 2;
int reg_width = dispatch_width / 8;
bool header_present = false;
int offsets[3];
- if (ir->offset && ir->op != ir_txf) {
- /* The offsets set up by the ir_texture visitor are in the
- * m1 header, so we can't go headerless.
+ fs_reg payload = fs_reg(this, glsl_type::float_type);
+ fs_reg next = payload;
+
+ if (ir->op == ir_tg4 || (ir->offset && ir->op != ir_txf)) {
+ /* For general texture offsets (no txf workaround), we need a header to
+ * put them in. Note that for 16-wide we're making space for two actual
+ * hardware registers here, so the emit will have to fix up for this.
+ *
+ * * ir4_tg4 needs to place its channel select in the header,
+ * for interaction with ARB_texture_swizzle
*/
header_present = true;
- mlen++;
- base_mrf--;
+ next.reg_offset++;
}
if (ir->shadow_comparitor) {
- emit(MOV(fs_reg(MRF, base_mrf + mlen), shadow_c));
- mlen += reg_width;
+ emit(MOV(next, shadow_c));
+ next.reg_offset++;
}
/* Set up the LOD info */
switch (ir->op) {
case ir_tex:
case ir_lod:
+ case ir_tg4:
break;
case ir_txb:
- emit(MOV(fs_reg(MRF, base_mrf + mlen), lod));
- mlen += reg_width;
+ emit(MOV(next, lod));
+ next.reg_offset++;
break;
case ir_txl:
- emit(MOV(fs_reg(MRF, base_mrf + mlen), lod));
- mlen += reg_width;
+ emit(MOV(next, lod));
+ next.reg_offset++;
break;
case ir_txd: {
if (dispatch_width == 16)
* [hdr], [ref], x, dPdx.x, dPdy.x, y, dPdx.y, dPdy.y, z, dPdx.z, dPdy.z
*/
for (int i = 0; i < ir->coordinate->type->vector_elements; i++) {
- emit(MOV(fs_reg(MRF, base_mrf + mlen), coordinate));
+ emit(MOV(next, coordinate));
coordinate.reg_offset++;
- mlen += reg_width;
+ next.reg_offset++;
/* For cube map array, the coordinate is (u,v,r,ai) but there are
* only derivatives for (u, v, r).
*/
if (i < ir->lod_info.grad.dPdx->type->vector_elements) {
- emit(MOV(fs_reg(MRF, base_mrf + mlen), lod));
+ emit(MOV(next, lod));
lod.reg_offset++;
- mlen += reg_width;
+ next.reg_offset++;
- emit(MOV(fs_reg(MRF, base_mrf + mlen), lod2));
+ emit(MOV(next, lod2));
lod2.reg_offset++;
- mlen += reg_width;
+ next.reg_offset++;
}
}
break;
}
case ir_txs:
- emit(MOV(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_UD), lod));
- mlen += reg_width;
+ emit(MOV(next.retype(BRW_REGISTER_TYPE_UD), lod));
+ next.reg_offset++;
+ break;
+ case ir_query_levels:
+ emit(MOV(next.retype(BRW_REGISTER_TYPE_UD), fs_reg(0u)));
+ next.reg_offset++;
break;
case ir_txf:
/* It appears that the ld instruction used for txf does its
}
/* Unfortunately, the parameters for LD are intermixed: u, lod, v, r. */
- emit(ADD(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_D),
- coordinate, offsets[0]));
+ emit(ADD(next.retype(BRW_REGISTER_TYPE_D), coordinate, offsets[0]));
coordinate.reg_offset++;
- mlen += reg_width;
+ next.reg_offset++;
- emit(MOV(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_D), lod));
- mlen += reg_width;
+ emit(MOV(next.retype(BRW_REGISTER_TYPE_D), lod));
+ next.reg_offset++;
for (int i = 1; i < ir->coordinate->type->vector_elements; i++) {
- emit(ADD(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_D),
- coordinate, offsets[i]));
+ emit(ADD(next.retype(BRW_REGISTER_TYPE_D), coordinate, offsets[i]));
coordinate.reg_offset++;
- mlen += reg_width;
+ next.reg_offset++;
}
break;
case ir_txf_ms:
- emit(MOV(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_UD), sample_index));
- mlen += reg_width;
+ emit(MOV(next.retype(BRW_REGISTER_TYPE_UD), sample_index));
+ next.reg_offset++;
/* constant zero MCS; we arrange to never actually have a compressed
* multisample surface here for now. TODO: issue ld_mcs to get this first,
* if we ever support texturing from compressed multisample surfaces
*/
- emit(MOV(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_UD), fs_reg(0u)));
- mlen += reg_width;
+ emit(MOV(next.retype(BRW_REGISTER_TYPE_UD), fs_reg(0u)));
+ next.reg_offset++;
/* there is no offsetting for this message; just copy in the integer
* texture coordinates
*/
for (int i = 0; i < ir->coordinate->type->vector_elements; i++) {
- emit(MOV(fs_reg(MRF, base_mrf + mlen, BRW_REGISTER_TYPE_D),
- coordinate));
+ emit(MOV(next.retype(BRW_REGISTER_TYPE_D), coordinate));
coordinate.reg_offset++;
- mlen += reg_width;
+ next.reg_offset++;
}
break;
}
/* Set up the coordinate (except for cases where it was done above) */
- if (ir->op != ir_txd && ir->op != ir_txs && ir->op != ir_txf && ir->op != ir_txf_ms) {
+ if (ir->op != ir_txd && ir->op != ir_txs && ir->op != ir_txf && ir->op != ir_txf_ms && ir->op != ir_query_levels) {
for (int i = 0; i < ir->coordinate->type->vector_elements; i++) {
- emit(MOV(fs_reg(MRF, base_mrf + mlen), coordinate));
+ emit(MOV(next, coordinate));
coordinate.reg_offset++;
- mlen += reg_width;
+ next.reg_offset++;
}
}
/* Generate the SEND */
fs_inst *inst = NULL;
switch (ir->op) {
- case ir_tex: inst = emit(SHADER_OPCODE_TEX, dst); break;
- case ir_txb: inst = emit(FS_OPCODE_TXB, dst); break;
- case ir_txl: inst = emit(SHADER_OPCODE_TXL, dst); break;
- case ir_txd: inst = emit(SHADER_OPCODE_TXD, dst); break;
- case ir_txf: inst = emit(SHADER_OPCODE_TXF, dst); break;
- case ir_txf_ms: inst = emit(SHADER_OPCODE_TXF_MS, dst); break;
- case ir_txs: inst = emit(SHADER_OPCODE_TXS, dst); break;
- case ir_lod: inst = emit(SHADER_OPCODE_LOD, dst); break;
- }
- inst->base_mrf = base_mrf;
- inst->mlen = mlen;
+ case ir_tex: inst = emit(SHADER_OPCODE_TEX, dst, payload); break;
+ case ir_txb: inst = emit(FS_OPCODE_TXB, dst, payload); break;
+ case ir_txl: inst = emit(SHADER_OPCODE_TXL, dst, payload); break;
+ case ir_txd: inst = emit(SHADER_OPCODE_TXD, dst, payload); break;
+ case ir_txf: inst = emit(SHADER_OPCODE_TXF, dst, payload); break;
+ case ir_txf_ms: inst = emit(SHADER_OPCODE_TXF_MS, dst, payload); break;
+ case ir_txs: inst = emit(SHADER_OPCODE_TXS, dst, payload); break;
+ case ir_query_levels: inst = emit(SHADER_OPCODE_TXS, dst, payload); break;
+ case ir_lod: inst = emit(SHADER_OPCODE_LOD, dst, payload); break;
+ case ir_tg4: inst = emit(SHADER_OPCODE_TG4, dst, payload); break;
+ }
+ inst->base_mrf = -1;
+ if (reg_width == 2)
+ inst->mlen = next.reg_offset * reg_width - header_present;
+ else
+ inst->mlen = next.reg_offset * reg_width;
+
inst->header_present = header_present;
inst->regs_written = 4;
- if (mlen > 11) {
+ virtual_grf_sizes[payload.reg] = next.reg_offset;
+ if (inst->mlen > 11) {
fail("Message length >11 disallowed by hardware\n");
}
(brw->gen < 6 ||
(brw->gen >= 6 && (c->key.tex.gl_clamp_mask[0] & (1 << sampler) ||
c->key.tex.gl_clamp_mask[1] & (1 << sampler))))) {
- struct gl_program_parameter_list *params = fp->Base.Parameters;
+ struct gl_program_parameter_list *params = prog->Parameters;
int tokens[STATE_LENGTH] = {
STATE_INTERNAL,
STATE_TEXRECT_SCALE,
GLuint index = _mesa_add_state_reference(params,
(gl_state_index *)tokens);
c->prog_data.param[c->prog_data.nr_params++] =
- &fp->Base.Parameters->ParameterValues[index][0].f;
+ &prog->Parameters->ParameterValues[index][0].f;
c->prog_data.param[c->prog_data.nr_params++] =
- &fp->Base.Parameters->ParameterValues[index][1].f;
+ &prog->Parameters->ParameterValues[index][1].f;
}
/* The 965 requires the EU to do the normalization of GL rectangle
fs_inst *inst = NULL;
int sampler =
- _mesa_get_sampler_uniform_value(ir->sampler, shader_prog, &fp->Base);
+ _mesa_get_sampler_uniform_value(ir->sampler, shader_prog, prog);
/* FINISHME: We're failing to recompile our programs when the sampler is
* updated. This only matters for the texture rectangle scale parameters
* (pre-gen6, or gen6+ with GL_CLAMP).
*/
- int texunit = fp->Base.SamplerUnits[sampler];
+ int texunit = prog->SamplerUnits[sampler];
+
+ if (ir->op == ir_tg4) {
+ /* When tg4 is used with the degenerate ZERO/ONE swizzles, don't bother
+ * emitting anything other than setting up the constant result.
+ */
+ ir_constant *chan = ir->lod_info.component->as_constant();
+ int swiz = GET_SWZ(c->key.tex.swizzles[sampler], chan->value.i[0]);
+ if (swiz == SWIZZLE_ZERO || swiz == SWIZZLE_ONE) {
+
+ fs_reg res = fs_reg(this, glsl_type::vec4_type);
+ this->result = res;
+
+ for (int i=0; i<4; i++) {
+ emit(MOV(res, fs_reg(swiz == SWIZZLE_ZERO ? 0.0f : 1.0f)));
+ res.reg_offset++;
+ }
+ return;
+ }
+ }
/* Should be lowered by do_lower_texture_projection */
assert(!ir->projector);
switch (ir->op) {
case ir_tex:
case ir_lod:
+ case ir_tg4:
+ case ir_query_levels:
break;
case ir_txb:
ir->lod_info.bias->accept(this);
ir->lod_info.sample_index->accept(this);
sample_index = this->result;
break;
+ default:
+ assert(!"Unrecognized texture opcode");
};
/* Writemasking doesn't eliminate channels on SIMD8 texture
lod, lod2);
}
- /* The header is set up by generate_tex() when necessary. */
- inst->src[0] = reg_undef;
-
if (ir->offset != NULL && ir->op != ir_txf)
inst->texture_offset = brw_texture_offset(ir->offset->as_constant());
+ if (ir->op == ir_tg4)
+ inst->texture_offset |= gather_channel(ir, sampler) << 16; // M0.2:16-17
+
inst->sampler = sampler;
if (ir->shadow_comparitor)
swizzle_result(ir, dst, sampler);
}
+/**
+ * Set up the gather channel based on the swizzle, for gather4.
+ */
+uint32_t
+fs_visitor::gather_channel(ir_texture *ir, int sampler)
+{
+ ir_constant *chan = ir->lod_info.component->as_constant();
+ int swiz = GET_SWZ(c->key.tex.swizzles[sampler], chan->value.i[0]);
+ switch (swiz) {
+ case SWIZZLE_X: return 0;
+ case SWIZZLE_Y:
+ /* gather4 sampler is broken for green channel on RG32F --
+ * we must ask for blue instead.
+ */
+ if (c->key.tex.gather_channel_quirk_mask & (1<<sampler))
+ return 2;
+ return 1;
+ case SWIZZLE_Z: return 2;
+ case SWIZZLE_W: return 3;
+ default:
+ assert(!"Not reached"); /* zero, one swizzles handled already */
+ return 0;
+ }
+}
+
/**
* Swizzle the result of a texture result. This is necessary for
* EXT_texture_swizzle as well as DEPTH_TEXTURE_MODE for shadow comparisons.
void
fs_visitor::swizzle_result(ir_texture *ir, fs_reg orig_val, int sampler)
{
+ if (ir->op == ir_query_levels) {
+ /* # levels is in .w */
+ orig_val.reg_offset += 3;
+ this->result = orig_val;
+ return;
+ }
+
this->result = orig_val;
- if (ir->op == ir_txs || ir->op == ir_lod)
+ /* txs,lod don't actually sample the texture, so swizzling the result
+ * makes no sense.
+ */
+ if (ir->op == ir_txs || ir->op == ir_lod || ir->op == ir_tg4)
return;
if (ir->type == glsl_type::float_type) {
{
ir_expression *expr = ir->as_expression();
- if (expr) {
+ if (expr &&
+ expr->operation != ir_binop_logic_and &&
+ expr->operation != ir_binop_logic_or &&
+ expr->operation != ir_binop_logic_xor) {
fs_reg op[2];
fs_inst *inst;
inst->conditional_mod = BRW_CONDITIONAL_Z;
break;
- case ir_binop_logic_xor:
- case ir_binop_logic_or:
- case ir_binop_logic_and:
- goto out;
-
case ir_unop_f2b:
if (brw->gen >= 6) {
emit(CMP(reg_null_d, op[0], fs_reg(0.0f), BRW_CONDITIONAL_NZ));
return;
}
-out:
ir->accept(this);
fs_inst *inst = emit(AND(reg_null_d, this->result, fs_reg(1)));
inst->predicate = BRW_PREDICATE_NORMAL;
}
+/**
+ * Try to replace IF/MOV/ELSE/MOV/ENDIF with SEL.
+ *
+ * Many GLSL shaders contain the following pattern:
+ *
+ * x = condition ? foo : bar
+ *
+ * The compiler emits an ir_if tree for this, since each subexpression might be
+ * a complex tree that could have side-effects or short-circuit logic.
+ *
+ * However, the common case is to simply select one of two constants or
+ * variable values---which is exactly what SEL is for. In this case, the
+ * assembly looks like:
+ *
+ * (+f0) IF
+ * MOV dst src0
+ * ELSE
+ * MOV dst src1
+ * ENDIF
+ *
+ * which can be easily translated into:
+ *
+ * (+f0) SEL dst src0 src1
+ *
+ * If src0 is an immediate value, we promote it to a temporary GRF.
+ */
+void
+fs_visitor::try_replace_with_sel()
+{
+ fs_inst *endif_inst = (fs_inst *) instructions.get_tail();
+ assert(endif_inst->opcode == BRW_OPCODE_ENDIF);
+
+ /* Pattern match in reverse: IF, MOV, ELSE, MOV, ENDIF. */
+ int opcodes[] = {
+ BRW_OPCODE_IF, BRW_OPCODE_MOV, BRW_OPCODE_ELSE, BRW_OPCODE_MOV,
+ };
+
+ fs_inst *match = (fs_inst *) endif_inst->prev;
+ for (int i = 0; i < 4; i++) {
+ if (match->is_head_sentinel() || match->opcode != opcodes[4-i-1])
+ return;
+ match = (fs_inst *) match->prev;
+ }
+
+ /* The opcodes match; it looks like the right sequence of instructions. */
+ fs_inst *else_mov = (fs_inst *) endif_inst->prev;
+ fs_inst *then_mov = (fs_inst *) else_mov->prev->prev;
+ fs_inst *if_inst = (fs_inst *) then_mov->prev;
+
+ /* Check that the MOVs are the right form. */
+ if (then_mov->dst.equals(else_mov->dst) &&
+ !then_mov->is_partial_write() &&
+ !else_mov->is_partial_write()) {
+
+ /* Remove the matched instructions; we'll emit a SEL to replace them. */
+ while (!if_inst->next->is_tail_sentinel())
+ if_inst->next->remove();
+ if_inst->remove();
+
+ /* Only the last source register can be a constant, so if the MOV in
+ * the "then" clause uses a constant, we need to put it in a temporary.
+ */
+ fs_reg src0(then_mov->src[0]);
+ if (src0.file == IMM) {
+ src0 = fs_reg(this, glsl_type::float_type);
+ src0.type = then_mov->src[0].type;
+ emit(MOV(src0, then_mov->src[0]));
+ }
+
+ fs_inst *sel;
+ if (if_inst->conditional_mod) {
+ /* Sandybridge-specific IF with embedded comparison */
+ emit(CMP(reg_null_d, if_inst->src[0], if_inst->src[1],
+ if_inst->conditional_mod));
+ sel = emit(BRW_OPCODE_SEL, then_mov->dst, src0, else_mov->src[0]);
+ sel->predicate = BRW_PREDICATE_NORMAL;
+ } else {
+ /* Separate CMP and IF instructions */
+ sel = emit(BRW_OPCODE_SEL, then_mov->dst, src0, else_mov->src[0]);
+ sel->predicate = if_inst->predicate;
+ sel->predicate_inverse = if_inst->predicate_inverse;
+ }
+ }
+}
+
void
fs_visitor::visit(ir_if *ir)
{
}
emit(BRW_OPCODE_ENDIF);
+
+ try_replace_with_sel();
}
void
const ir_function_signature *sig;
exec_list empty;
- sig = ir->matching_signature(&empty);
+ sig = ir->matching_signature(NULL, &empty);
assert(sig);
(void)ir;
}
+void
+fs_visitor::visit(ir_emit_vertex *)
+{
+ assert(!"not reached");
+}
+
+void
+fs_visitor::visit(ir_end_primitive *)
+{
+ assert(!"not reached");
+}
+
fs_inst *
fs_visitor::emit(fs_inst inst)
{
struct brw_reg
fs_visitor::interp_reg(int location, int channel)
{
- int regnr = urb_setup[location] * 2 + channel / 2;
+ int regnr = c->prog_data.urb_setup[location] * 2 + channel / 2;
int stride = (channel & 1) * 4;
- assert(urb_setup[location] != -1);
+ assert(c->prog_data.urb_setup[location] != -1);
return brw_vec1_grf(regnr, stride);
}
fail("Missing support for simd16 depth writes on gen6\n");
}
- if (fp->Base.OutputsWritten & BITFIELD64_BIT(FRAG_RESULT_DEPTH)) {
+ if (prog->OutputsWritten & BITFIELD64_BIT(FRAG_RESULT_DEPTH)) {
/* Hand over gl_FragDepth. */
assert(this->frag_depth.file != BAD_FILE);
emit(MOV(fs_reg(MRF, nr), this->frag_depth));
this->brw = brw;
this->fp = fp;
this->shader_prog = shader_prog;
+ this->prog = &fp->Base;
this->ctx = &brw->ctx;
this->mem_ctx = ralloc_context(NULL);
if (shader_prog)
this->virtual_grf_array_size = 0;
this->virtual_grf_start = NULL;
this->virtual_grf_end = NULL;
- this->live_intervals_valid = false;
+ this->live_intervals = NULL;
this->params_remap = NULL;
this->nr_params_remap = 0;