AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

r300/vertprog: Cleanup source conflict handling

Browse Source 
Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
This commit is contained in:
Nicolai Hähnle
2009-07-23 20:24:22 +02:00
parent 86e3334333
commit d65404225d
4 changed files with 155 additions and 188 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/mesa/drivers/dri/r300/compiler/r300_fragprog.c
-6
View File
@@ -31,12 +31,6 @@
#include "../r300_reg.h"
static void reset_srcreg(struct prog_src_register* reg)
{
_mesa_bzero(reg, sizeof(*reg));
reg->Swizzle = SWIZZLE_NOOP;
}
static struct prog_src_register shadow_ambient(struct gl_program *program, int tmu)
{
gl_state_index fail_value_tokens[STATE_LENGTH] = {
src/mesa/drivers/dri/r300/compiler/r3xx_vertprog.c
+149 -176
View File
@@ -32,27 +32,18 @@
#include "shader/prog_print.h"
/* TODO: Get rid of t_src_class call */
#define CMP_SRCS(a, b) ((a.RelAddr != b.RelAddr) || (a.Index != b.Index && \
((t_src_class(a.File) == PVS_SRC_REG_CONSTANT && \
t_src_class(b.File) == PVS_SRC_REG_CONSTANT) || \
(t_src_class(a.File) == PVS_SRC_REG_INPUT && \
t_src_class(b.File) == PVS_SRC_REG_INPUT)))) \
/*
* Take an already-setup and valid source then swizzle it appropriately to
* obtain a constant ZERO or ONE source.
*/
#define __CONST(x, y) \
(PVS_SRC_OPERAND(t_src_index(vp, &src[x]), \
(PVS_SRC_OPERAND(t_src_index(vp, &vpi->SrcReg[x]), \
t_swizzle(y), \
t_swizzle(y), \
t_swizzle(y), \
t_swizzle(y), \
t_src_class(src[x].File), \
NEGATE_NONE) | (src[x].RelAddr << 4))
t_src_class(vpi->SrcReg[x].File), \
NEGATE_NONE) | (vpi->SrcReg[x].RelAddr << 4))
static unsigned long t_dst_mask(GLuint mask)
@@ -121,6 +112,24 @@ static unsigned long t_src_class(gl_register_file file)
}
}
static GLboolean t_src_conflict(struct prog_src_register a, struct prog_src_register b)
{
unsigned long aclass = t_src_class(a.File);
unsigned long bclass = t_src_class(b.File);
if (aclass != bclass)
return GL_FALSE;
if (aclass == PVS_SRC_REG_TEMPORARY)
return GL_FALSE;
if (a.RelAddr || b.RelAddr)
return GL_TRUE;
if (a.Index != b.Index)
return GL_TRUE;
return GL_FALSE;
}
static INLINE unsigned long t_swizzle(GLubyte swizzle)
{
/* this is in fact a NOP as the Mesa SWIZZLE_* are all identical to VSF_IN_COMPONENT_* */
@@ -188,11 +197,10 @@ static GLboolean valid_dst(struct r300_vertex_program_code *vp,
return GL_TRUE;
}
static GLuint * ei_vector1(struct r300_vertex_program_code *vp,
static void ei_vector1(struct r300_vertex_program_code *vp,
GLuint hw_opcode,
struct prog_instruction *vpi,
GLuint * inst,
struct prog_src_register src[3])
GLuint * inst)
{
inst[0] = PVS_OP_DST_OPERAND(hw_opcode,
GL_FALSE,
@@ -200,18 +208,15 @@ static GLuint * ei_vector1(struct r300_vertex_program_code *vp,
t_dst_index(vp, &vpi->DstReg),
t_dst_mask(vpi->DstReg.WriteMask),
t_dst_class(vpi->DstReg.File));
inst[1] = t_src(vp, &src[0]);
inst[1] = t_src(vp, &vpi->SrcReg[0]);
inst[2] = __CONST(0, SWIZZLE_ZERO);
inst[3] = __CONST(0, SWIZZLE_ZERO);
return inst;
}
static GLuint * ei_vector2(struct r300_vertex_program_code *vp,
static void ei_vector2(struct r300_vertex_program_code *vp,
GLuint hw_opcode,
struct prog_instruction *vpi,
GLuint * inst,
struct prog_src_register src[3])
GLuint * inst)
{
inst[0] = PVS_OP_DST_OPERAND(hw_opcode,
GL_FALSE,
@@ -219,18 +224,15 @@ static GLuint * ei_vector2(struct r300_vertex_program_code *vp,
t_dst_index(vp, &vpi->DstReg),
t_dst_mask(vpi->DstReg.WriteMask),
t_dst_class(vpi->DstReg.File));
inst[1] = t_src(vp, &src[0]);
inst[2] = t_src(vp, &src[1]);
inst[1] = t_src(vp, &vpi->SrcReg[0]);
inst[2] = t_src(vp, &vpi->SrcReg[1]);
inst[3] = __CONST(1, SWIZZLE_ZERO);
return inst;
}
static GLuint *ei_math1(struct r300_vertex_program_code *vp,
static void ei_math1(struct r300_vertex_program_code *vp,
GLuint hw_opcode,
struct prog_instruction *vpi,
GLuint * inst,
struct prog_src_register src[3])
GLuint * inst)
{
inst[0] = PVS_OP_DST_OPERAND(hw_opcode,
GL_TRUE,
@@ -238,17 +240,14 @@ static GLuint *ei_math1(struct r300_vertex_program_code *vp,
t_dst_index(vp, &vpi->DstReg),
t_dst_mask(vpi->DstReg.WriteMask),
t_dst_class(vpi->DstReg.File));
inst[1] = t_src_scalar(vp, &src[0]);
inst[1] = t_src_scalar(vp, &vpi->SrcReg[0]);
inst[2] = __CONST(0, SWIZZLE_ZERO);
inst[3] = __CONST(0, SWIZZLE_ZERO);
return inst;
}
static GLuint *ei_lit(struct r300_vertex_program_code *vp,
static void ei_lit(struct r300_vertex_program_code *vp,
struct prog_instruction *vpi,
GLuint * inst,
struct prog_src_register src[3])
GLuint * inst)
{
//LIT TMP 1.Y Z TMP 1{} {X W Z Y} TMP 1{} {Y W Z X} TMP 1{} {Y X Z W}
@@ -259,35 +258,32 @@ static GLuint *ei_lit(struct r300_vertex_program_code *vp,
t_dst_mask(vpi->DstReg.WriteMask),
t_dst_class(vpi->DstReg.File));
/* NOTE: Users swizzling might not work. */
inst[1] = PVS_SRC_OPERAND(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // X
t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // W
inst[1] = PVS_SRC_OPERAND(t_src_index(vp, &vpi->SrcReg[0]), t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 0)), // X
t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 3)), // W
PVS_SRC_SELECT_FORCE_0, // Z
t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // Y
t_src_class(src[0].File),
src[0].Negate ? NEGATE_XYZW : NEGATE_NONE) |
(src[0].RelAddr << 4);
inst[2] = PVS_SRC_OPERAND(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // Y
t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // W
t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 1)), // Y
t_src_class(vpi->SrcReg[0].File),
vpi->SrcReg[0].Negate ? NEGATE_XYZW : NEGATE_NONE) |
(vpi->SrcReg[0].RelAddr << 4);
inst[2] = PVS_SRC_OPERAND(t_src_index(vp, &vpi->SrcReg[0]), t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 1)), // Y
t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 3)), // W
PVS_SRC_SELECT_FORCE_0, // Z
t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // X
t_src_class(src[0].File),
src[0].Negate ? NEGATE_XYZW : NEGATE_NONE) |
(src[0].RelAddr << 4);
inst[3] = PVS_SRC_OPERAND(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // Y
t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // X
t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 0)), // X
t_src_class(vpi->SrcReg[0].File),
vpi->SrcReg[0].Negate ? NEGATE_XYZW : NEGATE_NONE) |
(vpi->SrcReg[0].RelAddr << 4);
inst[3] = PVS_SRC_OPERAND(t_src_index(vp, &vpi->SrcReg[0]), t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 1)), // Y
t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 0)), // X
PVS_SRC_SELECT_FORCE_0, // Z
t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // W
t_src_class(src[0].File),
src[0].Negate ? NEGATE_XYZW : NEGATE_NONE) |
(src[0].RelAddr << 4);
return inst;
t_swizzle(GET_SWZ(vpi->SrcReg[0].Swizzle, 3)), // W
t_src_class(vpi->SrcReg[0].File),
vpi->SrcReg[0].Negate ? NEGATE_XYZW : NEGATE_NONE) |
(vpi->SrcReg[0].RelAddr << 4);
}
static GLuint *ei_mad(struct r300_vertex_program_code *vp,
static void ei_mad(struct r300_vertex_program_code *vp,
struct prog_instruction *vpi,
GLuint * inst,
struct prog_src_register src[3])
GLuint * inst)
{
inst[0] = PVS_OP_DST_OPERAND(PVS_MACRO_OP_2CLK_MADD,
GL_FALSE,
@@ -295,17 +291,14 @@ static GLuint *ei_mad(struct r300_vertex_program_code *vp,
t_dst_index(vp, &vpi->DstReg),
t_dst_mask(vpi->DstReg.WriteMask),
t_dst_class(vpi->DstReg.File));
inst[1] = t_src(vp, &src[0]);
inst[2] = t_src(vp, &src[1]);
inst[3] = t_src(vp, &src[2]);
return inst;
inst[1] = t_src(vp, &vpi->SrcReg[0]);
inst[2] = t_src(vp, &vpi->SrcReg[1]);
inst[3] = t_src(vp, &vpi->SrcReg[2]);
}
static GLuint *ei_pow(struct r300_vertex_program_code *vp,
static void ei_pow(struct r300_vertex_program_code *vp,
struct prog_instruction *vpi,
GLuint * inst,
struct prog_src_register src[3])
GLuint * inst)
{
inst[0] = PVS_OP_DST_OPERAND(ME_POWER_FUNC_FF,
GL_TRUE,
@@ -313,11 +306,9 @@ static GLuint *ei_pow(struct r300_vertex_program_code *vp,
t_dst_index(vp, &vpi->DstReg),
t_dst_mask(vpi->DstReg.WriteMask),
t_dst_class(vpi->DstReg.File));
inst[1] = t_src_scalar(vp, &src[0]);
inst[1] = t_src_scalar(vp, &vpi->SrcReg[0]);
inst[2] = __CONST(0, SWIZZLE_ZERO);
inst[3] = t_src_scalar(vp, &src[1]);
return inst;
inst[3] = t_src_scalar(vp, &vpi->SrcReg[1]);
}
static void t_inputs_outputs(struct r300_vertex_program_code *vp, struct gl_program * glvp)
@@ -397,15 +388,7 @@ static void t_inputs_outputs(struct r300_vertex_program_code *vp, struct gl_prog
static GLboolean translate_vertex_program(struct r300_vertex_program_compiler * compiler)
{
struct prog_instruction *vpi = compiler->program->Instructions;
int i;
GLuint *inst;
unsigned long num_operands;
/* Initial value should be last tmp reg that hw supports.
Strangely enough r300 doesnt mind even though these would be out of range.
Smart enough to realize that it doesnt need it? */
int u_temp_i = VSF_MAX_FRAGMENT_TEMPS - 1;
struct prog_src_register src[3];
struct r300_vertex_program_code * vp = compiler->code;
compiler->code->pos_end = 0; /* Not supported yet */
compiler->code->length = 0;
@@ -414,109 +397,33 @@ static GLboolean translate_vertex_program(struct r300_vertex_program_compiler *
for (inst = compiler->code->body.d; vpi->Opcode != OPCODE_END;
vpi++, inst += 4) {
{
int u_temp_used = (VSF_MAX_FRAGMENT_TEMPS - 1) - u_temp_i;
if((compiler->code->num_temporaries + u_temp_used) > VSF_MAX_FRAGMENT_TEMPS) {
fprintf(stderr, "Ran out of temps, num temps %d, us %d\n", compiler->code->num_temporaries, u_temp_used);
return GL_FALSE;
}
u_temp_i=VSF_MAX_FRAGMENT_TEMPS-1;
}
/* Skip instructions writing to non-existing destination */
if (!valid_dst(compiler->code, &vpi->DstReg)) {
/* redirect result to unused temp */
vpi->DstReg.File = PROGRAM_TEMPORARY;
vpi->DstReg.Index = u_temp_i;
}
num_operands = _mesa_num_inst_src_regs(vpi->Opcode);
/* copy the sources (src) from mesa into a local variable... is this needed? */
for (i = 0; i < num_operands; i++) {
src[i] = vpi->SrcReg[i];
}
if (num_operands == 3) { /* TODO: scalars */
if (CMP_SRCS(src[1], src[2])
|| CMP_SRCS(src[0], src[2])) {
inst[0] = PVS_OP_DST_OPERAND(VE_ADD,
GL_FALSE,
GL_FALSE,
u_temp_i,
WRITEMASK_XYZW,
PVS_DST_REG_TEMPORARY);
inst[1] =
PVS_SRC_OPERAND(t_src_index(compiler->code, &src[2]),
SWIZZLE_X,
SWIZZLE_Y,
SWIZZLE_Z,
SWIZZLE_W,
t_src_class(src[2].File),
NEGATE_NONE) | (src[2].
RelAddr <<
4);
inst[2] = __CONST(2, SWIZZLE_ZERO);
inst[3] = __CONST(2, SWIZZLE_ZERO);
inst += 4;
src[2].File = PROGRAM_TEMPORARY;
src[2].Index = u_temp_i;
src[2].RelAddr = 0;
u_temp_i--;
}
}
if (num_operands >= 2) {
if (CMP_SRCS(src[1], src[0])) {
inst[0] = PVS_OP_DST_OPERAND(VE_ADD,
GL_FALSE,
GL_FALSE,
u_temp_i,
WRITEMASK_XYZW,
PVS_DST_REG_TEMPORARY);
inst[1] =
PVS_SRC_OPERAND(t_src_index(compiler->code, &src[0]),
SWIZZLE_X,
SWIZZLE_Y,
SWIZZLE_Z,
SWIZZLE_W,
t_src_class(src[0].File),
NEGATE_NONE) | (src[0].
RelAddr <<
4);
inst[2] = __CONST(0, SWIZZLE_ZERO);
inst[3] = __CONST(0, SWIZZLE_ZERO);
inst += 4;
src[0].File = PROGRAM_TEMPORARY;
src[0].Index = u_temp_i;
src[0].RelAddr = 0;
u_temp_i--;
}
inst -= 4;
continue;
}
switch (vpi->Opcode) {
case OPCODE_ADD: inst = ei_vector2(compiler->code, VE_ADD, vpi, inst, src); break;
case OPCODE_ARL: inst = ei_vector1(compiler->code, VE_FLT2FIX_DX, vpi, inst, src); break;
case OPCODE_DP4: inst = ei_vector2(compiler->code, VE_DOT_PRODUCT, vpi, inst, src); break;
case OPCODE_DST: inst = ei_vector2(compiler->code, VE_DISTANCE_VECTOR, vpi, inst, src); break;
case OPCODE_EX2: inst = ei_math1(compiler->code, ME_EXP_BASE2_FULL_DX, vpi, inst, src); break;
case OPCODE_EXP: inst = ei_math1(compiler->code, ME_EXP_BASE2_DX, vpi, inst, src); break;
case OPCODE_FRC: inst = ei_vector1(compiler->code, VE_FRACTION, vpi, inst, src); break;
case OPCODE_LG2: inst = ei_math1(compiler->code, ME_LOG_BASE2_FULL_DX, vpi, inst, src); break;
case OPCODE_LIT: inst = ei_lit(compiler->code, vpi, inst, src); break;
case OPCODE_LOG: inst = ei_math1(compiler->code, ME_LOG_BASE2_DX, vpi, inst, src); break;
case OPCODE_MAD: inst = ei_mad(compiler->code, vpi, inst, src); break;
case OPCODE_MAX: inst = ei_vector2(compiler->code, VE_MAXIMUM, vpi, inst, src); break;
case OPCODE_MIN: inst = ei_vector2(compiler->code, VE_MINIMUM, vpi, inst, src); break;
case OPCODE_MOV: inst = ei_vector1(compiler->code, VE_ADD, vpi, inst, src); break;
case OPCODE_MUL: inst = ei_vector2(compiler->code, VE_MULTIPLY, vpi, inst, src); break;
case OPCODE_POW: inst = ei_pow(compiler->code, vpi, inst, src); break;
case OPCODE_RCP: inst = ei_math1(compiler->code, ME_RECIP_DX, vpi, inst, src); break;
case OPCODE_RSQ: inst = ei_math1(compiler->code, ME_RECIP_SQRT_DX, vpi, inst, src); break;
case OPCODE_SGE: inst = ei_vector2(compiler->code, VE_SET_GREATER_THAN_EQUAL, vpi, inst, src); break;
case OPCODE_SLT: inst = ei_vector2(compiler->code, VE_SET_LESS_THAN, vpi, inst, src); break;
case OPCODE_ADD: ei_vector2(compiler->code, VE_ADD, vpi, inst); break;
case OPCODE_ARL: ei_vector1(compiler->code, VE_FLT2FIX_DX, vpi, inst); break;
case OPCODE_DP4: ei_vector2(compiler->code, VE_DOT_PRODUCT, vpi, inst); break;
case OPCODE_DST: ei_vector2(compiler->code, VE_DISTANCE_VECTOR, vpi, inst); break;
case OPCODE_EX2: ei_math1(compiler->code, ME_EXP_BASE2_FULL_DX, vpi, inst); break;
case OPCODE_EXP: ei_math1(compiler->code, ME_EXP_BASE2_DX, vpi, inst); break;
case OPCODE_FRC: ei_vector1(compiler->code, VE_FRACTION, vpi, inst); break;
case OPCODE_LG2: ei_math1(compiler->code, ME_LOG_BASE2_FULL_DX, vpi, inst); break;
case OPCODE_LIT: ei_lit(compiler->code, vpi, inst); break;
case OPCODE_LOG: ei_math1(compiler->code, ME_LOG_BASE2_DX, vpi, inst); break;
case OPCODE_MAD: ei_mad(compiler->code, vpi, inst); break;
case OPCODE_MAX: ei_vector2(compiler->code, VE_MAXIMUM, vpi, inst); break;
case OPCODE_MIN: ei_vector2(compiler->code, VE_MINIMUM, vpi, inst); break;
case OPCODE_MOV: ei_vector1(compiler->code, VE_ADD, vpi, inst); break;
case OPCODE_MUL: ei_vector2(compiler->code, VE_MULTIPLY, vpi, inst); break;
case OPCODE_POW: ei_pow(compiler->code, vpi, inst); break;
case OPCODE_RCP: ei_math1(compiler->code, ME_RECIP_DX, vpi, inst); break;
case OPCODE_RSQ: ei_math1(compiler->code, ME_RECIP_SQRT_DX, vpi, inst); break;
case OPCODE_SGE: ei_vector2(compiler->code, VE_SET_GREATER_THAN_EQUAL, vpi, inst); break;
case OPCODE_SLT: ei_vector2(compiler->code, VE_SET_LESS_THAN, vpi, inst); break;
default:
fprintf(stderr, "Unknown opcode %i\n", vpi->Opcode);
return GL_FALSE;
@@ -531,6 +438,55 @@ static GLboolean translate_vertex_program(struct r300_vertex_program_compiler *
return GL_TRUE;
}
/**
* Vertex engine cannot read two inputs or two constants at the same time.
* Introduce intermediate MOVs to temporary registers to account for this.
*/
static GLboolean transform_source_conflicts(
struct radeon_transform_context *t,
struct prog_instruction* orig_inst,
void* unused)
{
struct prog_instruction inst = *orig_inst;
struct prog_instruction * dst;
GLuint num_operands = _mesa_num_inst_src_regs(inst.Opcode);
if (num_operands == 3) {
if (t_src_conflict(inst.SrcReg[1], inst.SrcReg[2])
|| t_src_conflict(inst.SrcReg[0], inst.SrcReg[2])) {
int tmpreg = radeonFindFreeTemporary(t);
struct prog_instruction * inst_mov = radeonAppendInstructions(t->Program, 1);
inst_mov->Opcode = OPCODE_MOV;
inst_mov->DstReg.File = PROGRAM_TEMPORARY;
inst_mov->DstReg.Index = tmpreg;
inst_mov->SrcReg[0] = inst.SrcReg[2];
reset_srcreg(&inst.SrcReg[2]);
inst.SrcReg[2].File = PROGRAM_TEMPORARY;
inst.SrcReg[2].Index = tmpreg;
}
}
if (num_operands >= 2) {
if (t_src_conflict(inst.SrcReg[1], inst.SrcReg[0])) {
int tmpreg = radeonFindFreeTemporary(t);
struct prog_instruction * inst_mov = radeonAppendInstructions(t->Program, 1);
inst_mov->Opcode = OPCODE_MOV;
inst_mov->DstReg.File = PROGRAM_TEMPORARY;
inst_mov->DstReg.Index = tmpreg;
inst_mov->SrcReg[0] = inst.SrcReg[1];
reset_srcreg(&inst.SrcReg[1]);
inst.SrcReg[1].File = PROGRAM_TEMPORARY;
inst.SrcReg[1].Index = tmpreg;
}
}
dst = radeonAppendInstructions(t->Program, 1);
*dst = inst;
return GL_TRUE;
}
static void insert_wpos(struct gl_program *prog, GLuint temp_index, int tex_id)
{
struct prog_instruction *vpi;
@@ -746,6 +702,23 @@ GLboolean r3xx_compile_vertex_program(struct r300_vertex_program_compiler* compi
fflush(stdout);
}
{
/* Note: This pass has to be done seperately from ALU rewrite,
* otherwise non-native ALU instructions with source conflits
* will not be treated properly.
*/
struct radeon_program_transformation transformations[] = {
{ &transform_source_conflicts, 0 },
};
radeonLocalTransform(compiler->program, 1, transformations);
}
if (compiler->Base.Debug) {
fprintf(stderr, "Vertex program after source conflict resolve:\n");
_mesa_print_program(compiler->program);
fflush(stdout);
}
{
struct radeon_nqssadce_descr nqssadce = {
.Init = &nqssadceInit,
src/mesa/drivers/dri/r300/compiler/r500_fragprog.c
-6
View File
@@ -29,12 +29,6 @@
#include "../r300_reg.h"
static void reset_srcreg(struct prog_src_register* reg)
{
_mesa_bzero(reg, sizeof(*reg));
reg->Swizzle = SWIZZLE_NOOP;
}
static struct prog_src_register shadow_ambient(struct gl_program *program, int tmu)
{
gl_state_index fail_value_tokens[STATE_LENGTH] = {
src/mesa/drivers/dri/r300/compiler/radeon_program.h
+6
View File
@@ -79,6 +79,12 @@ static inline GLuint combine_swizzles(GLuint src, GLuint swz)
return ret;
}
static INLINE void reset_srcreg(struct prog_src_register* reg)
{
_mesa_bzero(reg, sizeof(*reg));
reg->Swizzle = SWIZZLE_NOOP;
}
/**
* Transformation context that is passed to local transformations.