Sweden-Number/dlls/wined3d/shader_sm4.c

509 lines
17 KiB
C
Raw Normal View History

/*
* Copyright 2009 Henri Verbeet for CodeWeavers
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
*/
#include "config.h"
#include "wine/port.h"
#include "wined3d_private.h"
WINE_DEFAULT_DEBUG_CHANNEL(d3d_shader);
#define WINED3D_SM4_INSTRUCTION_MODIFIER (1 << 31)
#define WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT 24
#define WINED3D_SM4_INSTRUCTION_LENGTH_MASK (0xf << WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT)
#define WINED3D_SM4_OPCODE_MASK 0xff
#define WINED3D_SM4_REGISTER_MODIFIER (1 << 31)
#define WINED3D_SM4_REGISTER_ORDER_SHIFT 20
#define WINED3D_SM4_REGISTER_ORDER_MASK (0x3 << WINED3D_SM4_REGISTER_ORDER_SHIFT)
#define WINED3D_SM4_REGISTER_TYPE_SHIFT 12
#define WINED3D_SM4_REGISTER_TYPE_MASK (0xf << WINED3D_SM4_REGISTER_TYPE_SHIFT)
#define WINED3D_SM4_IMMCONST_TYPE_SHIFT 0
#define WINED3D_SM4_IMMCONST_TYPE_MASK (0x3 << WINED3D_SM4_IMMCONST_TYPE_SHIFT)
#define WINED3D_SM4_WRITEMASK_SHIFT 4
#define WINED3D_SM4_WRITEMASK_MASK (0xf << WINED3D_SM4_WRITEMASK_SHIFT)
#define WINED3D_SM4_SWIZZLE_SHIFT 4
#define WINED3D_SM4_SWIZZLE_MASK (0xff << WINED3D_SM4_SWIZZLE_SHIFT)
#define WINED3D_SM4_VERSION_MAJOR(version) (((version) >> 4) & 0xf)
#define WINED3D_SM4_VERSION_MINOR(version) (((version) >> 0) & 0xf)
enum wined3d_sm4_opcode
{
WINED3D_SM4_OP_ADD = 0x00,
WINED3D_SM4_OP_AND = 0x01,
WINED3D_SM4_OP_BREAK = 0x02,
WINED3D_SM4_OP_BREAKC = 0x03,
WINED3D_SM4_OP_CUT = 0x09,
2010-10-13 12:26:29 +02:00
WINED3D_SM4_OP_DIV = 0x0e,
WINED3D_SM4_OP_DP3 = 0x10,
WINED3D_SM4_OP_DP4 = 0x11,
WINED3D_SM4_OP_EMIT = 0x13,
WINED3D_SM4_OP_ENDIF = 0x15,
WINED3D_SM4_OP_ENDLOOP = 0x16,
WINED3D_SM4_OP_EXP = 0x19,
WINED3D_SM4_OP_FRC = 0x1a,
WINED3D_SM4_OP_FTOI = 0x1b,
WINED3D_SM4_OP_IADD = 0x1e,
WINED3D_SM4_OP_IF = 0x1f,
2010-10-13 12:26:31 +02:00
WINED3D_SM4_OP_IEQ = 0x20,
WINED3D_SM4_OP_IGE = 0x21,
WINED3D_SM4_OP_IMUL = 0x26,
WINED3D_SM4_OP_ITOF = 0x2b,
2010-10-13 12:26:32 +02:00
WINED3D_SM4_OP_LD = 0x2d,
WINED3D_SM4_OP_LOG = 0x2f,
WINED3D_SM4_OP_LOOP = 0x30,
WINED3D_SM4_OP_LT = 0x31,
WINED3D_SM4_OP_MAD = 0x32,
WINED3D_SM4_OP_MIN = 0x33,
WINED3D_SM4_OP_MAX = 0x34,
WINED3D_SM4_OP_MOV = 0x36,
WINED3D_SM4_OP_MOVC = 0x37,
WINED3D_SM4_OP_MUL = 0x38,
WINED3D_SM4_OP_RET = 0x3e,
WINED3D_SM4_OP_RSQ = 0x44,
WINED3D_SM4_OP_SAMPLE = 0x45,
WINED3D_SM4_OP_SAMPLE_LOD = 0x48,
WINED3D_SM4_OP_SAMPLE_GRAD = 0x49,
WINED3D_SM4_OP_SQRT = 0x4b,
WINED3D_SM4_OP_SINCOS = 0x4d,
WINED3D_SM4_OP_UTOF = 0x56,
};
enum wined3d_sm4_register_type
{
WINED3D_SM4_RT_TEMP = 0x0,
WINED3D_SM4_RT_INPUT = 0x1,
WINED3D_SM4_RT_OUTPUT = 0x2,
WINED3D_SM4_RT_IMMCONST = 0x4,
WINED3D_SM4_RT_SAMPLER = 0x6,
WINED3D_SM4_RT_CONSTBUFFER = 0x8,
WINED3D_SM4_RT_NULL = 0xd,
};
enum wined3d_sm4_immconst_type
{
WINED3D_SM4_IMMCONST_SCALAR = 0x1,
WINED3D_SM4_IMMCONST_VEC4 = 0x2,
};
struct wined3d_sm4_data
{
struct wined3d_shader_version shader_version;
const DWORD *end;
const struct wined3d_shader_signature *output_signature;
};
struct wined3d_sm4_opcode_info
{
enum wined3d_sm4_opcode opcode;
enum WINED3D_SHADER_INSTRUCTION_HANDLER handler_idx;
UINT dst_count;
UINT src_count;
};
struct sysval_map
{
enum wined3d_sysval_semantic sysval;
WINED3DSHADER_PARAM_REGISTER_TYPE register_type;
UINT register_idx;
};
static const struct wined3d_sm4_opcode_info opcode_table[] =
{
{WINED3D_SM4_OP_ADD, WINED3DSIH_ADD, 1, 2},
{WINED3D_SM4_OP_AND, WINED3DSIH_AND, 1, 2},
{WINED3D_SM4_OP_BREAK, WINED3DSIH_BREAK, 0, 0},
{WINED3D_SM4_OP_BREAKC, WINED3DSIH_BREAKP, 0, 1},
{WINED3D_SM4_OP_CUT, WINED3DSIH_CUT, 0, 0},
2010-10-13 12:26:29 +02:00
{WINED3D_SM4_OP_DIV, WINED3DSIH_DIV, 1, 2},
{WINED3D_SM4_OP_DP3, WINED3DSIH_DP3, 1, 2},
{WINED3D_SM4_OP_DP4, WINED3DSIH_DP4, 1, 2},
{WINED3D_SM4_OP_EMIT, WINED3DSIH_EMIT, 0, 0},
{WINED3D_SM4_OP_ENDIF, WINED3DSIH_ENDIF, 0, 0},
{WINED3D_SM4_OP_ENDLOOP, WINED3DSIH_ENDLOOP, 0, 0},
{WINED3D_SM4_OP_EXP, WINED3DSIH_EXP, 1, 1},
{WINED3D_SM4_OP_FRC, WINED3DSIH_FRC, 1, 1},
{WINED3D_SM4_OP_FTOI, WINED3DSIH_FTOI, 1, 1},
{WINED3D_SM4_OP_IADD, WINED3DSIH_IADD, 1, 2},
{WINED3D_SM4_OP_IF, WINED3DSIH_IF, 0, 1},
2010-10-13 12:26:31 +02:00
{WINED3D_SM4_OP_IEQ, WINED3DSIH_IEQ, 1, 2},
{WINED3D_SM4_OP_IGE, WINED3DSIH_IGE, 1, 2},
{WINED3D_SM4_OP_IMUL, WINED3DSIH_IMUL, 2, 2},
{WINED3D_SM4_OP_ITOF, WINED3DSIH_ITOF, 1, 1},
2010-10-13 12:26:32 +02:00
{WINED3D_SM4_OP_LD, WINED3DSIH_LD, 1, 2},
{WINED3D_SM4_OP_LOG, WINED3DSIH_LOG, 1, 1},
{WINED3D_SM4_OP_LOOP, WINED3DSIH_LOOP, 0, 0},
{WINED3D_SM4_OP_LT, WINED3DSIH_LT, 1, 2},
{WINED3D_SM4_OP_MAD, WINED3DSIH_MAD, 1, 3},
{WINED3D_SM4_OP_MIN, WINED3DSIH_MIN, 1, 2},
{WINED3D_SM4_OP_MAX, WINED3DSIH_MAX, 1, 2},
{WINED3D_SM4_OP_MOV, WINED3DSIH_MOV, 1, 1},
{WINED3D_SM4_OP_MOVC, WINED3DSIH_MOVC, 1, 3},
{WINED3D_SM4_OP_MUL, WINED3DSIH_MUL, 1, 2},
{WINED3D_SM4_OP_RET, WINED3DSIH_RET, 0, 0},
{WINED3D_SM4_OP_RSQ, WINED3DSIH_RSQ, 1, 1},
{WINED3D_SM4_OP_SAMPLE, WINED3DSIH_SAMPLE, 1, 3},
{WINED3D_SM4_OP_SAMPLE_LOD, WINED3DSIH_SAMPLE_LOD, 1, 4},
{WINED3D_SM4_OP_SAMPLE_GRAD,WINED3DSIH_SAMPLE_GRAD, 1, 5},
{WINED3D_SM4_OP_SQRT, WINED3DSIH_SQRT, 1, 1},
{WINED3D_SM4_OP_SINCOS, WINED3DSIH_SINCOS, 2, 1},
{WINED3D_SM4_OP_UTOF, WINED3DSIH_UTOF, 1, 1},
};
static const WINED3DSHADER_PARAM_REGISTER_TYPE register_type_table[] =
{
/* WINED3D_SM4_RT_TEMP */ WINED3DSPR_TEMP,
/* WINED3D_SM4_RT_INPUT */ WINED3DSPR_INPUT,
/* WINED3D_SM4_RT_OUTPUT */ WINED3DSPR_OUTPUT,
/* UNKNOWN */ 0,
/* WINED3D_SM4_RT_IMMCONST */ WINED3DSPR_IMMCONST,
/* UNKNOWN */ 0,
/* WINED3D_SM4_RT_SAMPLER */ WINED3DSPR_SAMPLER,
/* WINED3D_SM4_RT_RESOURCE */ WINED3DSPR_RESOURCE,
/* WINED3D_SM4_RT_CONSTBUFFER */ WINED3DSPR_CONSTBUFFER,
/* UNKNOWN */ 0,
/* UNKNOWN */ 0,
/* UNKNOWN */ 0,
/* UNKNOWN */ 0,
/* WINED3D_SM4_RT_NULL */ WINED3DSPR_NULL,
};
static const struct sysval_map sysval_map[] =
{
{WINED3D_SV_DEPTH, WINED3DSPR_DEPTHOUT, 0},
{WINED3D_SV_TARGET0, WINED3DSPR_COLOROUT, 0},
{WINED3D_SV_TARGET1, WINED3DSPR_COLOROUT, 1},
{WINED3D_SV_TARGET2, WINED3DSPR_COLOROUT, 2},
{WINED3D_SV_TARGET3, WINED3DSPR_COLOROUT, 3},
{WINED3D_SV_TARGET4, WINED3DSPR_COLOROUT, 4},
{WINED3D_SV_TARGET5, WINED3DSPR_COLOROUT, 5},
{WINED3D_SV_TARGET6, WINED3DSPR_COLOROUT, 6},
{WINED3D_SV_TARGET7, WINED3DSPR_COLOROUT, 7},
};
static const struct wined3d_sm4_opcode_info *get_opcode_info(enum wined3d_sm4_opcode opcode)
{
unsigned int i;
for (i = 0; i < sizeof(opcode_table) / sizeof(*opcode_table); ++i)
{
if (opcode == opcode_table[i].opcode) return &opcode_table[i];
}
return NULL;
}
static void map_sysval(enum wined3d_sysval_semantic sysval, struct wined3d_shader_register *reg)
{
unsigned int i;
for (i = 0; i < sizeof(sysval_map) / sizeof(*sysval_map); ++i)
{
if (sysval == sysval_map[i].sysval)
{
reg->type = sysval_map[i].register_type;
reg->idx = sysval_map[i].register_idx;
}
}
}
static void map_register(struct wined3d_sm4_data *priv, struct wined3d_shader_register *reg)
{
switch (priv->shader_version.type)
{
case WINED3D_SHADER_TYPE_PIXEL:
if (reg->type == WINED3DSPR_OUTPUT)
{
unsigned int i;
const struct wined3d_shader_signature *s = priv->output_signature;
if (!s)
{
ERR("Shader has no output signature, unable to map register.\n");
break;
}
for (i = 0; i < s->element_count; ++i)
{
if (s->elements[i].register_idx == reg->idx)
{
map_sysval(s->elements[i].sysval_semantic, reg);
break;
}
}
}
break;
default:
break;
}
}
static void *shader_sm4_init(const DWORD *byte_code, const struct wined3d_shader_signature *output_signature)
{
struct wined3d_sm4_data *priv = HeapAlloc(GetProcessHeap(), 0, sizeof(*priv));
if (!priv)
{
ERR("Failed to allocate private data\n");
return NULL;
}
priv->output_signature = output_signature;
return priv;
}
static void shader_sm4_free(void *data)
{
HeapFree(GetProcessHeap(), 0, data);
}
static void shader_sm4_read_header(void *data, const DWORD **ptr, struct wined3d_shader_version *shader_version)
{
struct wined3d_sm4_data *priv = data;
DWORD version_token;
priv->end = *ptr;
version_token = *(*ptr)++;
TRACE("version: 0x%08x\n", version_token);
TRACE("token count: %u\n", **ptr);
priv->end += *(*ptr)++;
switch (version_token >> 16)
{
case WINED3D_SM4_PS:
priv->shader_version.type = WINED3D_SHADER_TYPE_PIXEL;
break;
case WINED3D_SM4_VS:
priv->shader_version.type = WINED3D_SHADER_TYPE_VERTEX;
break;
case WINED3D_SM4_GS:
priv->shader_version.type = WINED3D_SHADER_TYPE_GEOMETRY;
break;
default:
FIXME("Unrecognized shader type %#x\n", version_token >> 16);
}
priv->shader_version.major = WINED3D_SM4_VERSION_MAJOR(version_token);
priv->shader_version.minor = WINED3D_SM4_VERSION_MINOR(version_token);
*shader_version = priv->shader_version;
}
static void shader_sm4_read_opcode(void *data, const DWORD **ptr, struct wined3d_shader_instruction *ins,
UINT *param_size)
{
const struct wined3d_sm4_opcode_info *opcode_info;
DWORD token = *(*ptr)++;
DWORD opcode = token & WINED3D_SM4_OPCODE_MASK;
*param_size = ((token & WINED3D_SM4_INSTRUCTION_LENGTH_MASK) >> WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT) - 1;
opcode_info = get_opcode_info(opcode);
if (!opcode_info)
{
FIXME("Unrecognized opcode %#x, token 0x%08x\n", opcode, token);
ins->handler_idx = WINED3DSIH_TABLE_SIZE;
return;
}
ins->handler_idx = opcode_info->handler_idx;
ins->flags = 0;
ins->coissue = 0;
ins->predicate = 0;
ins->dst_count = opcode_info->dst_count;
ins->src_count = opcode_info->src_count;
if (token & WINED3D_SM4_INSTRUCTION_MODIFIER)
{
DWORD modifier = *(*ptr)++;
FIXME("Skipping modifier 0x%08x.\n", modifier);
}
}
static void shader_sm4_read_src_param(void *data, const DWORD **ptr, struct wined3d_shader_src_param *src_param,
struct wined3d_shader_src_param *src_rel_addr)
{
struct wined3d_sm4_data *priv = data;
DWORD token = *(*ptr)++;
enum wined3d_sm4_register_type register_type;
DWORD order;
register_type = (token & WINED3D_SM4_REGISTER_TYPE_MASK) >> WINED3D_SM4_REGISTER_TYPE_SHIFT;
if (register_type >= sizeof(register_type_table) / sizeof(*register_type_table))
{
FIXME("Unhandled register type %#x\n", register_type);
src_param->reg.type = WINED3DSPR_TEMP;
}
else
{
src_param->reg.type = register_type_table[register_type];
}
if (token & WINED3D_SM4_REGISTER_MODIFIER)
{
DWORD modifier = *(*ptr)++;
/* FIXME: This will probably break down at some point. The SM4
* modifiers look like flags, while wined3d currently has an enum
* with possible combinations, e.g. WINED3DSPSM_ABSNEG. */
switch (modifier)
{
case 0x41:
src_param->modifiers = WINED3DSPSM_NEG;
break;
case 0x81:
src_param->modifiers = WINED3DSPSM_ABS;
break;
default:
FIXME("Skipping modifier 0x%08x.\n", modifier);
src_param->modifiers = WINED3DSPSM_NONE;
break;
}
}
else
{
src_param->modifiers = WINED3DSPSM_NONE;
}
order = (token & WINED3D_SM4_REGISTER_ORDER_MASK) >> WINED3D_SM4_REGISTER_ORDER_SHIFT;
if (order < 1) src_param->reg.idx = ~0U;
else src_param->reg.idx = *(*ptr)++;
if (order < 2) src_param->reg.array_idx = ~0U;
else src_param->reg.array_idx = *(*ptr)++;
if (order > 2) FIXME("Unhandled order %u.\n", order);
if (register_type == WINED3D_SM4_RT_IMMCONST)
{
enum wined3d_sm4_immconst_type immconst_type =
(token & WINED3D_SM4_IMMCONST_TYPE_MASK) >> WINED3D_SM4_IMMCONST_TYPE_SHIFT;
src_param->swizzle = WINED3DSP_NOSWIZZLE;
switch(immconst_type)
{
case WINED3D_SM4_IMMCONST_SCALAR:
src_param->reg.immconst_type = WINED3D_IMMCONST_SCALAR;
memcpy(src_param->reg.immconst_data, *ptr, 1 * sizeof(DWORD));
*ptr += 1;
break;
case WINED3D_SM4_IMMCONST_VEC4:
src_param->reg.immconst_type = WINED3D_IMMCONST_VEC4;
memcpy(src_param->reg.immconst_data, *ptr, 4 * sizeof(DWORD));
*ptr += 4;
break;
default:
FIXME("Unhandled immediate constant type %#x\n", immconst_type);
break;
}
}
else
{
src_param->swizzle = (token & WINED3D_SM4_SWIZZLE_MASK) >> WINED3D_SM4_SWIZZLE_SHIFT;
}
src_param->reg.rel_addr = NULL;
map_register(priv, &src_param->reg);
}
static void shader_sm4_read_dst_param(void *data, const DWORD **ptr, struct wined3d_shader_dst_param *dst_param,
struct wined3d_shader_src_param *dst_rel_addr)
{
struct wined3d_sm4_data *priv = data;
DWORD token = *(*ptr)++;
enum wined3d_sm4_register_type register_type;
DWORD order;
register_type = (token & WINED3D_SM4_REGISTER_TYPE_MASK) >> WINED3D_SM4_REGISTER_TYPE_SHIFT;
if (register_type >= sizeof(register_type_table) / sizeof(*register_type_table))
{
FIXME("Unhandled register type %#x\n", register_type);
dst_param->reg.type = WINED3DSPR_TEMP;
}
else
{
dst_param->reg.type = register_type_table[register_type];
}
order = (token & WINED3D_SM4_REGISTER_ORDER_MASK) >> WINED3D_SM4_REGISTER_ORDER_SHIFT;
if (order < 1) dst_param->reg.idx = ~0U;
else dst_param->reg.idx = *(*ptr)++;
if (order < 2) dst_param->reg.array_idx = ~0U;
else dst_param->reg.array_idx = *(*ptr)++;
if (order > 2) FIXME("Unhandled order %u.\n", order);
dst_param->write_mask = (token & WINED3D_SM4_WRITEMASK_MASK) >> WINED3D_SM4_WRITEMASK_SHIFT;
dst_param->modifiers = 0;
dst_param->shift = 0;
dst_param->reg.rel_addr = NULL;
map_register(priv, &dst_param->reg);
}
static void shader_sm4_read_semantic(const DWORD **ptr, struct wined3d_shader_semantic *semantic)
{
FIXME("ptr %p, semantic %p stub!\n", ptr, semantic);
}
static void shader_sm4_read_comment(const DWORD **ptr, const char **comment, UINT *comment_size)
{
FIXME("ptr %p, comment %p, comment_size %p stub!\n", ptr, comment, comment_size);
*comment = NULL;
}
static BOOL shader_sm4_is_end(void *data, const DWORD **ptr)
{
struct wined3d_sm4_data *priv = data;
return *ptr == priv->end;
}
const struct wined3d_shader_frontend sm4_shader_frontend =
{
shader_sm4_init,
shader_sm4_free,
shader_sm4_read_header,
shader_sm4_read_opcode,
shader_sm4_read_src_param,
shader_sm4_read_dst_param,
shader_sm4_read_semantic,
shader_sm4_read_comment,
shader_sm4_is_end,
};