2021-03-02 13:34:46 -08:00
|
|
|
/*
|
|
|
|
|
* HLSL optimization and code generation
|
|
|
|
|
*
|
|
|
|
|
* Copyright 2019-2020 Zebediah Figura 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 "hlsl.h"
|
2021-03-28 12:46:55 -07:00
|
|
|
#include <stdio.h>
|
2024-07-10 19:47:51 -07:00
|
|
|
#include <math.h>
|
2021-03-28 12:46:55 -07:00
|
|
|
|
2022-07-01 09:36:25 -07:00
|
|
|
/* TODO: remove when no longer needed, only used for new_offset_instr_from_deref() */
|
|
|
|
|
static struct hlsl_ir_node *new_offset_from_path_index(struct hlsl_ctx *ctx, struct hlsl_block *block,
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
struct hlsl_type *type, struct hlsl_ir_node *base_offset, struct hlsl_ir_node *idx,
|
|
|
|
|
enum hlsl_regset regset, unsigned int *offset_component, const struct vkd3d_shader_location *loc)
|
2022-07-01 09:36:25 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *idx_offset = NULL;
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *c;
|
2022-07-01 09:36:25 -07:00
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
switch (type->class)
|
2022-07-01 09:36:25 -07:00
|
|
|
{
|
|
|
|
|
case HLSL_CLASS_VECTOR:
|
2024-01-11 11:54:20 -08:00
|
|
|
if (idx->type != HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &idx->loc, "Non-constant vector addressing.");
|
|
|
|
|
break;
|
|
|
|
|
}
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
*offset_component += hlsl_ir_constant(idx)->value.u[0].u;
|
2022-07-01 09:36:25 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_MATRIX:
|
|
|
|
|
{
|
2023-10-03 15:14:54 -07:00
|
|
|
idx_offset = idx;
|
2022-07-01 09:36:25 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_ARRAY:
|
|
|
|
|
{
|
2022-10-28 08:23:05 -07:00
|
|
|
unsigned int size = hlsl_type_get_array_element_reg_size(type->e.array.type, regset);
|
2022-07-01 09:36:25 -07:00
|
|
|
|
2023-10-03 15:14:54 -07:00
|
|
|
if (regset == HLSL_REGSET_NUMERIC)
|
|
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(size % 4 == 0);
|
2023-10-03 15:14:54 -07:00
|
|
|
size /= 4;
|
|
|
|
|
}
|
|
|
|
|
|
2022-07-01 09:36:25 -07:00
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, size, loc)))
|
|
|
|
|
return NULL;
|
2022-11-10 19:06:04 -08:00
|
|
|
hlsl_block_add_instr(block, c);
|
2022-07-01 09:36:25 -07:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(idx_offset = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, c, idx)))
|
2022-07-01 09:36:25 -07:00
|
|
|
return NULL;
|
2022-11-11 19:13:50 -08:00
|
|
|
hlsl_block_add_instr(block, idx_offset);
|
2022-07-01 09:36:25 -07:00
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_STRUCT:
|
|
|
|
|
{
|
2022-11-11 16:39:55 -08:00
|
|
|
unsigned int field_idx = hlsl_ir_constant(idx)->value.u[0].u;
|
2022-07-01 09:36:25 -07:00
|
|
|
struct hlsl_struct_field *field = &type->e.record.fields[field_idx];
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
unsigned int field_offset = field->reg_offset[regset];
|
2022-07-01 09:36:25 -07:00
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
if (regset == HLSL_REGSET_NUMERIC)
|
|
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(*offset_component == 0);
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
*offset_component = field_offset % 4;
|
2023-10-03 15:14:54 -07:00
|
|
|
field_offset /= 4;
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, field_offset, loc)))
|
2022-07-01 09:36:25 -07:00
|
|
|
return NULL;
|
2022-11-10 19:06:04 -08:00
|
|
|
hlsl_block_add_instr(block, c);
|
2022-07-01 09:36:25 -07:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
idx_offset = c;
|
2022-07-01 09:36:25 -07:00
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
default:
|
2022-08-31 04:25:24 -07:00
|
|
|
vkd3d_unreachable();
|
2022-07-01 09:36:25 -07:00
|
|
|
}
|
|
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
if (idx_offset)
|
2022-07-01 09:36:25 -07:00
|
|
|
{
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
if (!(base_offset = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, base_offset, idx_offset)))
|
2022-07-01 09:36:25 -07:00
|
|
|
return NULL;
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
hlsl_block_add_instr(block, base_offset);
|
2022-07-01 09:36:25 -07:00
|
|
|
}
|
|
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
return base_offset;
|
2022-07-01 09:36:25 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* TODO: remove when no longer needed, only used for replace_deref_path_with_offset() */
|
|
|
|
|
static struct hlsl_ir_node *new_offset_instr_from_deref(struct hlsl_ctx *ctx, struct hlsl_block *block,
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
const struct hlsl_deref *deref, unsigned int *offset_component, const struct vkd3d_shader_location *loc)
|
2022-07-01 09:36:25 -07:00
|
|
|
{
|
2023-06-12 08:58:07 -07:00
|
|
|
enum hlsl_regset regset = hlsl_deref_get_regset(ctx, deref);
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
struct hlsl_ir_node *offset;
|
2022-07-01 09:36:25 -07:00
|
|
|
struct hlsl_type *type;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
*offset_component = 0;
|
|
|
|
|
|
2023-02-16 15:52:15 -08:00
|
|
|
hlsl_block_init(block);
|
2022-07-01 09:36:25 -07:00
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
if (!(offset = hlsl_new_uint_constant(ctx, 0, loc)))
|
|
|
|
|
return NULL;
|
|
|
|
|
hlsl_block_add_instr(block, offset);
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(deref->var);
|
2022-07-01 09:36:25 -07:00
|
|
|
type = deref->var->data_type;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < deref->path_len; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_block idx_block;
|
|
|
|
|
|
2023-10-11 14:13:31 -07:00
|
|
|
hlsl_block_init(&idx_block);
|
|
|
|
|
|
2022-10-28 08:23:05 -07:00
|
|
|
if (!(offset = new_offset_from_path_index(ctx, &idx_block, type, offset, deref->path[i].node,
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
regset, offset_component, loc)))
|
2023-10-11 14:13:31 -07:00
|
|
|
{
|
|
|
|
|
hlsl_block_cleanup(&idx_block);
|
2022-07-01 09:36:25 -07:00
|
|
|
return NULL;
|
2023-10-11 14:13:31 -07:00
|
|
|
}
|
2022-07-01 09:36:25 -07:00
|
|
|
|
2023-02-16 16:00:01 -08:00
|
|
|
hlsl_block_add_block(block, &idx_block);
|
2022-07-01 09:36:25 -07:00
|
|
|
|
2022-08-16 09:33:51 -07:00
|
|
|
type = hlsl_get_element_type_from_path_index(ctx, type, deref->path[i].node);
|
2022-07-01 09:36:25 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return offset;
|
|
|
|
|
}
|
|
|
|
|
|
2022-06-30 15:20:20 -07:00
|
|
|
/* TODO: remove when no longer needed, only used for transform_deref_paths_into_offsets() */
|
2023-05-04 12:06:58 -07:00
|
|
|
static bool replace_deref_path_with_offset(struct hlsl_ctx *ctx, struct hlsl_deref *deref,
|
2022-06-30 15:20:20 -07:00
|
|
|
struct hlsl_ir_node *instr)
|
|
|
|
|
{
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
unsigned int offset_component;
|
2022-06-30 15:20:20 -07:00
|
|
|
struct hlsl_ir_node *offset;
|
|
|
|
|
struct hlsl_block block;
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
struct hlsl_type *type;
|
2022-06-30 15:20:20 -07:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(deref->var);
|
|
|
|
|
VKD3D_ASSERT(!hlsl_deref_is_lowered(deref));
|
2022-07-12 13:58:40 -07:00
|
|
|
|
2023-02-09 15:29:17 -08:00
|
|
|
type = hlsl_deref_get_type(ctx, deref);
|
|
|
|
|
|
|
|
|
|
/* Instructions that directly refer to structs or arrays (instead of single-register components)
|
|
|
|
|
* are removed later by dce. So it is not a problem to just cleanup their derefs. */
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class == HLSL_CLASS_STRUCT || type->class == HLSL_CLASS_ARRAY)
|
2023-02-09 15:29:17 -08:00
|
|
|
{
|
|
|
|
|
hlsl_cleanup_deref(deref);
|
2023-05-04 12:06:58 -07:00
|
|
|
return true;
|
2023-02-09 15:29:17 -08:00
|
|
|
}
|
|
|
|
|
|
2023-05-29 14:34:03 -07:00
|
|
|
deref->data_type = type;
|
2022-10-28 08:23:05 -07:00
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
if (!(offset = new_offset_instr_from_deref(ctx, &block, deref, &offset_component, &instr->loc)))
|
2023-05-04 12:06:58 -07:00
|
|
|
return false;
|
2022-06-30 15:20:20 -07:00
|
|
|
list_move_before(&instr->entry, &block.instrs);
|
|
|
|
|
|
|
|
|
|
hlsl_cleanup_deref(deref);
|
2023-10-06 09:56:24 -07:00
|
|
|
hlsl_src_from_node(&deref->rel_offset, offset);
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
deref->const_offset = offset_component;
|
2022-06-30 15:20:20 -07:00
|
|
|
|
2023-05-04 12:06:58 -07:00
|
|
|
return true;
|
2022-06-30 15:20:20 -07:00
|
|
|
}
|
|
|
|
|
|
2023-05-11 12:18:02 -07:00
|
|
|
static bool clean_constant_deref_offset_srcs(struct hlsl_ctx *ctx, struct hlsl_deref *deref,
|
|
|
|
|
struct hlsl_ir_node *instr)
|
|
|
|
|
{
|
2023-10-06 09:56:24 -07:00
|
|
|
if (deref->rel_offset.node && deref->rel_offset.node->type == HLSL_IR_CONSTANT)
|
2023-05-11 12:18:02 -07:00
|
|
|
{
|
|
|
|
|
enum hlsl_regset regset = hlsl_deref_get_regset(ctx, deref);
|
|
|
|
|
|
|
|
|
|
if (regset == HLSL_REGSET_NUMERIC)
|
2023-10-06 09:56:24 -07:00
|
|
|
deref->const_offset += 4 * hlsl_ir_constant(deref->rel_offset.node)->value.u[0].u;
|
2023-05-11 12:18:02 -07:00
|
|
|
else
|
2023-10-06 09:56:24 -07:00
|
|
|
deref->const_offset += hlsl_ir_constant(deref->rel_offset.node)->value.u[0].u;
|
|
|
|
|
hlsl_src_remove(&deref->rel_offset);
|
2023-05-11 12:18:02 -07:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2021-03-28 12:46:55 -07:00
|
|
|
/* Split uniforms into two variables representing the constant and temp
|
|
|
|
|
* registers, and copy the former to the latter, so that writes to uniforms
|
|
|
|
|
* work. */
|
2023-03-10 13:13:23 -08:00
|
|
|
static void prepend_uniform_copy(struct hlsl_ctx *ctx, struct hlsl_block *block, struct hlsl_ir_var *temp)
|
2021-03-28 12:46:55 -07:00
|
|
|
{
|
2021-04-15 17:03:45 -07:00
|
|
|
struct hlsl_ir_var *uniform;
|
2022-11-10 18:55:03 -08:00
|
|
|
struct hlsl_ir_node *store;
|
2021-03-28 12:46:55 -07:00
|
|
|
struct hlsl_ir_load *load;
|
2023-08-07 14:45:31 -07:00
|
|
|
char *new_name;
|
2021-03-28 12:46:55 -07:00
|
|
|
|
2021-04-15 17:03:45 -07:00
|
|
|
/* Use the synthetic name for the temp, rather than the uniform, so that we
|
|
|
|
|
* can write the uniform name into the shader reflection data. */
|
|
|
|
|
|
2022-03-22 14:59:11 -07:00
|
|
|
if (!(uniform = hlsl_new_var(ctx, temp->name, temp->data_type,
|
2023-04-14 00:02:14 -07:00
|
|
|
&temp->loc, NULL, temp->storage_modifiers, &temp->reg_reservation)))
|
2021-03-28 12:46:55 -07:00
|
|
|
return;
|
2021-04-15 17:03:45 -07:00
|
|
|
list_add_before(&temp->scope_entry, &uniform->scope_entry);
|
|
|
|
|
list_add_tail(&ctx->extern_vars, &uniform->extern_entry);
|
|
|
|
|
uniform->is_uniform = 1;
|
2021-04-15 17:03:46 -07:00
|
|
|
uniform->is_param = temp->is_param;
|
2021-06-21 21:37:10 -07:00
|
|
|
uniform->buffer = temp->buffer;
|
2024-04-08 13:15:20 -07:00
|
|
|
if (temp->default_values)
|
|
|
|
|
{
|
|
|
|
|
/* Transfer default values from the temp to the uniform. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!uniform->default_values);
|
|
|
|
|
VKD3D_ASSERT(hlsl_type_component_count(temp->data_type) == hlsl_type_component_count(uniform->data_type));
|
2024-04-08 13:15:20 -07:00
|
|
|
uniform->default_values = temp->default_values;
|
|
|
|
|
temp->default_values = NULL;
|
|
|
|
|
}
|
2021-04-15 17:03:45 -07:00
|
|
|
|
2023-08-07 14:45:31 -07:00
|
|
|
if (!(new_name = hlsl_sprintf_alloc(ctx, "<temp-%s>", temp->name)))
|
2021-03-28 12:46:55 -07:00
|
|
|
return;
|
2023-08-07 14:45:31 -07:00
|
|
|
temp->name = new_name;
|
2021-03-28 12:46:55 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(load = hlsl_new_var_load(ctx, uniform, &temp->loc)))
|
2021-03-28 12:46:55 -07:00
|
|
|
return;
|
2023-03-10 13:13:23 -08:00
|
|
|
list_add_head(&block->instrs, &load->node.entry);
|
2021-03-28 12:46:55 -07:00
|
|
|
|
2021-05-20 22:32:20 -07:00
|
|
|
if (!(store = hlsl_new_simple_store(ctx, temp, &load->node)))
|
2021-03-28 12:46:55 -07:00
|
|
|
return;
|
2022-11-10 18:55:03 -08:00
|
|
|
list_add_after(&load->node.entry, &store->entry);
|
2021-03-28 12:46:55 -07:00
|
|
|
}
|
2021-03-02 13:34:46 -08:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
static void validate_field_semantic(struct hlsl_ctx *ctx, struct hlsl_struct_field *field)
|
|
|
|
|
{
|
2023-11-12 17:05:50 -08:00
|
|
|
if (!field->semantic.name && hlsl_is_numeric_type(hlsl_get_multiarray_element_type(field->type))
|
2023-04-05 09:07:37 -07:00
|
|
|
&& !field->semantic.reported_missing)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &field->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_SEMANTIC,
|
|
|
|
|
"Field '%s' is missing a semantic.", field->name);
|
|
|
|
|
field->semantic.reported_missing = true;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-12 13:27:31 -07:00
|
|
|
static enum hlsl_base_type base_type_get_semantic_equivalent(enum hlsl_base_type base)
|
|
|
|
|
{
|
|
|
|
|
if (base == HLSL_TYPE_BOOL)
|
|
|
|
|
return HLSL_TYPE_UINT;
|
|
|
|
|
if (base == HLSL_TYPE_INT)
|
|
|
|
|
return HLSL_TYPE_UINT;
|
|
|
|
|
if (base == HLSL_TYPE_HALF)
|
|
|
|
|
return HLSL_TYPE_FLOAT;
|
|
|
|
|
return base;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool types_are_semantic_equivalent(struct hlsl_ctx *ctx, const struct hlsl_type *type1,
|
|
|
|
|
const struct hlsl_type *type2)
|
|
|
|
|
{
|
2023-04-14 14:46:03 -07:00
|
|
|
if (ctx->profile->major_version < 4)
|
|
|
|
|
return true;
|
|
|
|
|
|
2023-04-12 13:27:31 -07:00
|
|
|
if (type1->dimx != type2->dimx)
|
|
|
|
|
return false;
|
|
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
return base_type_get_semantic_equivalent(type1->e.numeric.type)
|
|
|
|
|
== base_type_get_semantic_equivalent(type2->e.numeric.type);
|
2023-04-12 13:27:31 -07:00
|
|
|
}
|
|
|
|
|
|
2022-07-01 08:37:47 -07:00
|
|
|
static struct hlsl_ir_var *add_semantic_var(struct hlsl_ctx *ctx, struct hlsl_ir_var *var,
|
2023-12-06 10:36:35 -08:00
|
|
|
struct hlsl_type *type, uint32_t modifiers, struct hlsl_semantic *semantic,
|
2023-04-12 12:59:06 -07:00
|
|
|
uint32_t index, bool output, const struct vkd3d_shader_location *loc)
|
2022-07-01 08:37:47 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_semantic new_semantic;
|
|
|
|
|
struct hlsl_ir_var *ext_var;
|
2023-08-07 14:45:31 -07:00
|
|
|
char *new_name;
|
2022-07-01 08:37:47 -07:00
|
|
|
|
2023-08-07 14:45:31 -07:00
|
|
|
if (!(new_name = hlsl_sprintf_alloc(ctx, "<%s-%s%u>", output ? "output" : "input", semantic->name, index)))
|
2022-07-01 08:37:47 -07:00
|
|
|
return NULL;
|
2023-04-05 08:58:23 -07:00
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(ext_var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
2023-08-07 14:45:31 -07:00
|
|
|
if (!ascii_strcasecmp(ext_var->name, new_name))
|
2023-04-05 08:58:23 -07:00
|
|
|
{
|
2023-04-12 12:59:06 -07:00
|
|
|
if (output)
|
|
|
|
|
{
|
|
|
|
|
if (index >= semantic->reported_duplicated_output_next_index)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_INVALID_SEMANTIC,
|
|
|
|
|
"Output semantic \"%s%u\" is used multiple times.", semantic->name, index);
|
2023-10-11 06:34:22 -07:00
|
|
|
hlsl_note(ctx, &ext_var->loc, VKD3D_SHADER_LOG_ERROR,
|
2023-04-12 12:59:06 -07:00
|
|
|
"First use of \"%s%u\" is here.", semantic->name, index);
|
|
|
|
|
semantic->reported_duplicated_output_next_index = index + 1;
|
|
|
|
|
}
|
|
|
|
|
}
|
2023-04-12 13:27:31 -07:00
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
if (index >= semantic->reported_duplicated_input_incompatible_next_index
|
|
|
|
|
&& !types_are_semantic_equivalent(ctx, ext_var->data_type, type))
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_INVALID_SEMANTIC,
|
|
|
|
|
"Input semantic \"%s%u\" is used multiple times with incompatible types.",
|
|
|
|
|
semantic->name, index);
|
2023-10-11 06:34:22 -07:00
|
|
|
hlsl_note(ctx, &ext_var->loc, VKD3D_SHADER_LOG_ERROR,
|
2023-04-12 13:27:31 -07:00
|
|
|
"First declaration of \"%s%u\" is here.", semantic->name, index);
|
|
|
|
|
semantic->reported_duplicated_input_incompatible_next_index = index + 1;
|
|
|
|
|
}
|
|
|
|
|
}
|
2023-04-12 12:59:06 -07:00
|
|
|
|
2023-08-07 14:45:31 -07:00
|
|
|
vkd3d_free(new_name);
|
2023-04-05 08:58:23 -07:00
|
|
|
return ext_var;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-31 03:25:12 -07:00
|
|
|
if (!(hlsl_clone_semantic(ctx, &new_semantic, semantic)))
|
2022-07-01 08:37:47 -07:00
|
|
|
{
|
2023-08-07 14:45:31 -07:00
|
|
|
vkd3d_free(new_name);
|
2022-07-01 08:37:47 -07:00
|
|
|
return NULL;
|
|
|
|
|
}
|
2023-04-05 09:07:37 -07:00
|
|
|
new_semantic.index = index;
|
2023-08-07 14:45:31 -07:00
|
|
|
if (!(ext_var = hlsl_new_var(ctx, new_name, type, loc, &new_semantic, modifiers, NULL)))
|
2022-07-01 08:37:47 -07:00
|
|
|
{
|
2023-08-07 14:45:31 -07:00
|
|
|
vkd3d_free(new_name);
|
2023-01-31 17:44:46 -08:00
|
|
|
hlsl_cleanup_semantic(&new_semantic);
|
2022-07-01 08:37:47 -07:00
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
if (output)
|
|
|
|
|
ext_var->is_output_semantic = 1;
|
|
|
|
|
else
|
|
|
|
|
ext_var->is_input_semantic = 1;
|
|
|
|
|
ext_var->is_param = var->is_param;
|
|
|
|
|
list_add_before(&var->scope_entry, &ext_var->scope_entry);
|
|
|
|
|
list_add_tail(&ctx->extern_vars, &ext_var->extern_entry);
|
|
|
|
|
|
|
|
|
|
return ext_var;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-10 13:14:30 -08:00
|
|
|
static void prepend_input_copy(struct hlsl_ctx *ctx, struct hlsl_block *block, struct hlsl_ir_load *lhs,
|
2023-12-06 10:36:35 -08:00
|
|
|
uint32_t modifiers, struct hlsl_semantic *semantic, uint32_t semantic_index)
|
2021-03-28 12:46:57 -07:00
|
|
|
{
|
2023-04-14 14:34:22 -07:00
|
|
|
struct hlsl_type *type = lhs->node.data_type, *vector_type_src, *vector_type_dst;
|
2023-04-12 12:59:06 -07:00
|
|
|
struct vkd3d_shader_location *loc = &lhs->node.loc;
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_ir_var *var = lhs->src.var;
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *c;
|
2022-07-01 09:06:30 -07:00
|
|
|
unsigned int i;
|
2021-03-28 12:46:57 -07:00
|
|
|
|
2023-11-12 17:05:50 -08:00
|
|
|
if (!hlsl_is_numeric_type(type))
|
2023-04-05 09:45:33 -07:00
|
|
|
{
|
|
|
|
|
struct vkd3d_string_buffer *string;
|
|
|
|
|
if (!(string = hlsl_type_to_string(ctx, type)))
|
|
|
|
|
return;
|
|
|
|
|
hlsl_fixme(ctx, &var->loc, "Input semantics for type %s.", string->buffer);
|
|
|
|
|
hlsl_release_string_buffer(ctx, string);
|
|
|
|
|
}
|
2023-04-05 09:07:37 -07:00
|
|
|
if (!semantic->name)
|
|
|
|
|
return;
|
2023-04-05 09:45:33 -07:00
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
vector_type_dst = hlsl_get_vector_type(ctx, type->e.numeric.type, hlsl_type_minor_size(type));
|
2023-07-31 11:00:45 -07:00
|
|
|
vector_type_src = vector_type_dst;
|
|
|
|
|
if (ctx->profile->major_version < 4 && ctx->profile->type == VKD3D_SHADER_TYPE_VERTEX)
|
2024-02-27 15:30:51 -08:00
|
|
|
vector_type_src = hlsl_get_vector_type(ctx, type->e.numeric.type, 4);
|
2022-07-01 09:06:30 -07:00
|
|
|
|
|
|
|
|
for (i = 0; i < hlsl_type_major_size(type); ++i)
|
2022-06-07 14:29:11 -07:00
|
|
|
{
|
2022-11-10 18:57:00 -08:00
|
|
|
struct hlsl_ir_node *store, *cast;
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_ir_var *input;
|
|
|
|
|
struct hlsl_ir_load *load;
|
2022-06-07 14:29:11 -07:00
|
|
|
|
2023-04-14 14:34:22 -07:00
|
|
|
if (!(input = add_semantic_var(ctx, var, vector_type_src, modifiers, semantic,
|
|
|
|
|
semantic_index + i, false, loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
2022-06-07 14:29:11 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(load = hlsl_new_var_load(ctx, input, &var->loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
|
|
|
|
list_add_after(&lhs->node.entry, &load->node.entry);
|
2021-03-28 12:46:57 -07:00
|
|
|
|
2023-04-14 14:34:22 -07:00
|
|
|
if (!(cast = hlsl_new_cast(ctx, &load->node, vector_type_dst, &var->loc)))
|
|
|
|
|
return;
|
|
|
|
|
list_add_after(&load->node.entry, &cast->entry);
|
|
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class == HLSL_CLASS_MATRIX)
|
2022-07-01 09:06:30 -07:00
|
|
|
{
|
|
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, i, &var->loc)))
|
|
|
|
|
return;
|
2022-11-10 19:06:04 -08:00
|
|
|
list_add_after(&cast->entry, &c->entry);
|
2021-03-28 12:46:57 -07:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(store = hlsl_new_store_index(ctx, &lhs->src, c, cast, 0, &var->loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
2022-11-10 19:06:04 -08:00
|
|
|
list_add_after(&c->entry, &store->entry);
|
2022-07-01 09:06:30 -07:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(i == 0);
|
2021-03-28 12:46:57 -07:00
|
|
|
|
2023-04-14 14:34:22 -07:00
|
|
|
if (!(store = hlsl_new_store_index(ctx, &lhs->src, NULL, cast, 0, &var->loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
2022-11-10 18:57:00 -08:00
|
|
|
list_add_after(&cast->entry, &store->entry);
|
2022-07-01 09:06:30 -07:00
|
|
|
}
|
|
|
|
|
}
|
2021-03-28 12:46:57 -07:00
|
|
|
}
|
|
|
|
|
|
2023-03-10 13:14:30 -08:00
|
|
|
static void prepend_input_copy_recurse(struct hlsl_ctx *ctx, struct hlsl_block *block, struct hlsl_ir_load *lhs,
|
2023-12-06 10:36:35 -08:00
|
|
|
uint32_t modifiers, struct hlsl_semantic *semantic, uint32_t semantic_index)
|
2021-03-28 12:46:57 -07:00
|
|
|
{
|
2023-04-12 12:59:06 -07:00
|
|
|
struct vkd3d_shader_location *loc = &lhs->node.loc;
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_type *type = lhs->node.data_type;
|
|
|
|
|
struct hlsl_ir_var *var = lhs->src.var;
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *c;
|
2023-04-05 09:07:37 -07:00
|
|
|
unsigned int i;
|
2021-03-28 12:46:57 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
if (type->class == HLSL_CLASS_ARRAY || type->class == HLSL_CLASS_STRUCT)
|
2021-03-28 12:46:57 -07:00
|
|
|
{
|
2023-04-05 09:07:37 -07:00
|
|
|
struct hlsl_ir_load *element_load;
|
|
|
|
|
struct hlsl_struct_field *field;
|
|
|
|
|
uint32_t elem_semantic_index;
|
2022-07-01 09:06:30 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
for (i = 0; i < hlsl_type_element_count(type); ++i)
|
|
|
|
|
{
|
2023-12-06 10:36:35 -08:00
|
|
|
uint32_t element_modifiers = modifiers;
|
2023-09-21 06:14:55 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
if (type->class == HLSL_CLASS_ARRAY)
|
|
|
|
|
{
|
|
|
|
|
elem_semantic_index = semantic_index
|
|
|
|
|
+ i * hlsl_type_get_array_element_reg_size(type->e.array.type, HLSL_REGSET_NUMERIC) / 4;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
field = &type->e.record.fields[i];
|
2022-11-25 12:31:21 -08:00
|
|
|
if (hlsl_type_is_resource(field->type))
|
2024-03-28 13:13:11 -07:00
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &field->loc, "Prepend uniform copies for resource components within structs.");
|
2022-11-25 12:31:21 -08:00
|
|
|
continue;
|
2024-03-28 13:13:11 -07:00
|
|
|
}
|
2023-04-05 09:07:37 -07:00
|
|
|
validate_field_semantic(ctx, field);
|
|
|
|
|
semantic = &field->semantic;
|
|
|
|
|
elem_semantic_index = semantic->index;
|
2023-04-12 12:59:06 -07:00
|
|
|
loc = &field->loc;
|
2023-09-21 06:14:55 -07:00
|
|
|
element_modifiers |= field->storage_modifiers;
|
|
|
|
|
|
|
|
|
|
/* TODO: 'sample' modifier is not supported yet */
|
|
|
|
|
|
|
|
|
|
/* 'nointerpolation' always takes precedence, next the same is done for 'sample',
|
|
|
|
|
remaining modifiers are combined. */
|
|
|
|
|
if (element_modifiers & HLSL_STORAGE_NOINTERPOLATION)
|
|
|
|
|
{
|
|
|
|
|
element_modifiers &= ~HLSL_INTERPOLATION_MODIFIERS_MASK;
|
|
|
|
|
element_modifiers |= HLSL_STORAGE_NOINTERPOLATION;
|
|
|
|
|
}
|
2023-04-05 09:07:37 -07:00
|
|
|
}
|
2022-07-01 09:06:30 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, i, &var->loc)))
|
|
|
|
|
return;
|
2022-11-10 19:06:04 -08:00
|
|
|
list_add_after(&lhs->node.entry, &c->entry);
|
2022-07-14 18:23:43 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
/* This redundant load is expected to be deleted later by DCE. */
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(element_load = hlsl_new_load_index(ctx, &lhs->src, c, loc)))
|
2023-04-05 09:07:37 -07:00
|
|
|
return;
|
2022-11-10 19:06:04 -08:00
|
|
|
list_add_after(&c->entry, &element_load->node.entry);
|
2023-04-05 09:07:37 -07:00
|
|
|
|
2023-09-21 06:14:55 -07:00
|
|
|
prepend_input_copy_recurse(ctx, block, element_load, element_modifiers, semantic, elem_semantic_index);
|
2023-04-05 09:07:37 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2023-03-10 13:14:30 -08:00
|
|
|
prepend_input_copy(ctx, block, lhs, modifiers, semantic, semantic_index);
|
2021-03-28 12:46:57 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-27 10:14:20 -07:00
|
|
|
/* Split inputs into two variables representing the semantic and temp registers,
|
|
|
|
|
* and copy the former to the latter, so that writes to input variables work. */
|
2023-03-10 13:14:30 -08:00
|
|
|
static void prepend_input_var_copy(struct hlsl_ctx *ctx, struct hlsl_block *block, struct hlsl_ir_var *var)
|
2021-03-28 12:46:57 -07:00
|
|
|
{
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
|
2023-03-07 17:04:37 -08:00
|
|
|
/* This redundant load is expected to be deleted later by DCE. */
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(load = hlsl_new_var_load(ctx, var, &var->loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
2023-03-10 13:14:30 -08:00
|
|
|
list_add_head(&block->instrs, &load->node.entry);
|
2022-07-01 09:06:30 -07:00
|
|
|
|
2023-03-10 13:14:30 -08:00
|
|
|
prepend_input_copy_recurse(ctx, block, load, var->storage_modifiers, &var->semantic, var->semantic.index);
|
2021-03-28 12:46:57 -07:00
|
|
|
}
|
|
|
|
|
|
2023-03-10 13:17:26 -08:00
|
|
|
static void append_output_copy(struct hlsl_ctx *ctx, struct hlsl_block *block, struct hlsl_ir_load *rhs,
|
2023-12-06 10:36:35 -08:00
|
|
|
uint32_t modifiers, struct hlsl_semantic *semantic, uint32_t semantic_index)
|
2021-03-28 12:46:59 -07:00
|
|
|
{
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_type *type = rhs->node.data_type, *vector_type;
|
2023-04-12 12:59:06 -07:00
|
|
|
struct vkd3d_shader_location *loc = &rhs->node.loc;
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_ir_var *var = rhs->src.var;
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *c;
|
2022-07-01 09:06:30 -07:00
|
|
|
unsigned int i;
|
|
|
|
|
|
2023-11-12 17:05:50 -08:00
|
|
|
if (!hlsl_is_numeric_type(type))
|
2023-04-05 09:45:33 -07:00
|
|
|
{
|
|
|
|
|
struct vkd3d_string_buffer *string;
|
|
|
|
|
if (!(string = hlsl_type_to_string(ctx, type)))
|
|
|
|
|
return;
|
|
|
|
|
hlsl_fixme(ctx, &var->loc, "Output semantics for type %s.", string->buffer);
|
|
|
|
|
hlsl_release_string_buffer(ctx, string);
|
|
|
|
|
}
|
2023-04-05 09:07:37 -07:00
|
|
|
if (!semantic->name)
|
|
|
|
|
return;
|
2023-04-05 09:45:33 -07:00
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
vector_type = hlsl_get_vector_type(ctx, type->e.numeric.type, hlsl_type_minor_size(type));
|
2021-03-28 12:46:59 -07:00
|
|
|
|
2022-07-01 09:06:30 -07:00
|
|
|
for (i = 0; i < hlsl_type_major_size(type); ++i)
|
2022-06-07 14:29:10 -07:00
|
|
|
{
|
2022-11-10 18:55:03 -08:00
|
|
|
struct hlsl_ir_node *store;
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_ir_var *output;
|
|
|
|
|
struct hlsl_ir_load *load;
|
2022-06-07 14:29:10 -07:00
|
|
|
|
2023-04-12 12:59:06 -07:00
|
|
|
if (!(output = add_semantic_var(ctx, var, vector_type, modifiers, semantic, semantic_index + i, true, loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
2022-06-07 14:29:10 -07:00
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class == HLSL_CLASS_MATRIX)
|
2022-07-01 09:06:30 -07:00
|
|
|
{
|
|
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, i, &var->loc)))
|
|
|
|
|
return;
|
2023-03-10 13:17:26 -08:00
|
|
|
hlsl_block_add_instr(block, c);
|
2021-03-28 12:46:59 -07:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(load = hlsl_new_load_index(ctx, &rhs->src, c, &var->loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
2023-03-10 13:17:26 -08:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2022-07-01 09:06:30 -07:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(i == 0);
|
2021-03-28 12:46:59 -07:00
|
|
|
|
2022-07-01 09:06:30 -07:00
|
|
|
if (!(load = hlsl_new_load_index(ctx, &rhs->src, NULL, &var->loc)))
|
|
|
|
|
return;
|
2023-03-10 13:17:26 -08:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2022-07-01 09:06:30 -07:00
|
|
|
}
|
2021-03-28 12:46:59 -07:00
|
|
|
|
2022-07-01 09:06:30 -07:00
|
|
|
if (!(store = hlsl_new_simple_store(ctx, output, &load->node)))
|
|
|
|
|
return;
|
2023-03-10 13:17:26 -08:00
|
|
|
hlsl_block_add_instr(block, store);
|
2022-07-01 09:06:30 -07:00
|
|
|
}
|
2021-03-28 12:46:59 -07:00
|
|
|
}
|
|
|
|
|
|
2023-03-10 13:17:26 -08:00
|
|
|
static void append_output_copy_recurse(struct hlsl_ctx *ctx, struct hlsl_block *block, struct hlsl_ir_load *rhs,
|
2023-12-06 10:36:35 -08:00
|
|
|
uint32_t modifiers, struct hlsl_semantic *semantic, uint32_t semantic_index)
|
2021-03-28 12:46:59 -07:00
|
|
|
{
|
2023-04-12 12:59:06 -07:00
|
|
|
struct vkd3d_shader_location *loc = &rhs->node.loc;
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_type *type = rhs->node.data_type;
|
|
|
|
|
struct hlsl_ir_var *var = rhs->src.var;
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *c;
|
2023-04-05 09:07:37 -07:00
|
|
|
unsigned int i;
|
2021-03-28 12:46:59 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
if (type->class == HLSL_CLASS_ARRAY || type->class == HLSL_CLASS_STRUCT)
|
2021-03-28 12:46:59 -07:00
|
|
|
{
|
2023-04-05 09:07:37 -07:00
|
|
|
struct hlsl_ir_load *element_load;
|
|
|
|
|
struct hlsl_struct_field *field;
|
|
|
|
|
uint32_t elem_semantic_index;
|
2022-07-01 09:06:30 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
for (i = 0; i < hlsl_type_element_count(type); ++i)
|
|
|
|
|
{
|
|
|
|
|
if (type->class == HLSL_CLASS_ARRAY)
|
|
|
|
|
{
|
|
|
|
|
elem_semantic_index = semantic_index
|
|
|
|
|
+ i * hlsl_type_get_array_element_reg_size(type->e.array.type, HLSL_REGSET_NUMERIC) / 4;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
field = &type->e.record.fields[i];
|
2022-11-25 12:31:21 -08:00
|
|
|
if (hlsl_type_is_resource(field->type))
|
|
|
|
|
continue;
|
2023-04-05 09:07:37 -07:00
|
|
|
validate_field_semantic(ctx, field);
|
|
|
|
|
semantic = &field->semantic;
|
|
|
|
|
elem_semantic_index = semantic->index;
|
2023-04-12 12:59:06 -07:00
|
|
|
loc = &field->loc;
|
2023-04-05 09:07:37 -07:00
|
|
|
}
|
2022-07-01 09:06:30 -07:00
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, i, &var->loc)))
|
|
|
|
|
return;
|
2023-03-10 13:17:26 -08:00
|
|
|
hlsl_block_add_instr(block, c);
|
2022-07-14 18:23:43 -07:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(element_load = hlsl_new_load_index(ctx, &rhs->src, c, loc)))
|
2023-04-05 09:07:37 -07:00
|
|
|
return;
|
2023-03-10 13:17:26 -08:00
|
|
|
hlsl_block_add_instr(block, &element_load->node);
|
2023-04-05 09:07:37 -07:00
|
|
|
|
2023-03-10 13:17:26 -08:00
|
|
|
append_output_copy_recurse(ctx, block, element_load, modifiers, semantic, elem_semantic_index);
|
2023-04-05 09:07:37 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2023-03-10 13:17:26 -08:00
|
|
|
append_output_copy(ctx, block, rhs, modifiers, semantic, semantic_index);
|
2021-03-28 12:46:59 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-27 10:14:20 -07:00
|
|
|
/* Split outputs into two variables representing the temp and semantic
|
2021-03-28 12:46:59 -07:00
|
|
|
* registers, and copy the former to the latter, so that reads from output
|
2021-04-27 10:14:20 -07:00
|
|
|
* variables work. */
|
2023-03-10 13:17:26 -08:00
|
|
|
static void append_output_var_copy(struct hlsl_ctx *ctx, struct hlsl_block *block, struct hlsl_ir_var *var)
|
2021-03-28 12:46:59 -07:00
|
|
|
{
|
2022-07-01 09:06:30 -07:00
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
|
2023-03-07 17:04:37 -08:00
|
|
|
/* This redundant load is expected to be deleted later by DCE. */
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(load = hlsl_new_var_load(ctx, var, &var->loc)))
|
2022-07-01 09:06:30 -07:00
|
|
|
return;
|
2023-03-10 13:17:26 -08:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2022-07-01 09:06:30 -07:00
|
|
|
|
2023-03-10 13:17:26 -08:00
|
|
|
append_output_copy_recurse(ctx, block, load, var->storage_modifiers, &var->semantic, var->semantic.index);
|
2021-03-28 12:46:59 -07:00
|
|
|
}
|
|
|
|
|
|
2023-04-25 06:04:29 -07:00
|
|
|
bool hlsl_transform_ir(struct hlsl_ctx *ctx, bool (*func)(struct hlsl_ctx *ctx, struct hlsl_ir_node *, void *),
|
2021-10-15 14:54:10 -07:00
|
|
|
struct hlsl_block *block, void *context)
|
2021-03-16 14:31:53 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr, *next;
|
2021-11-17 00:47:25 -08:00
|
|
|
bool progress = false;
|
2021-03-16 14:31:53 -07:00
|
|
|
|
2021-10-15 14:54:10 -07:00
|
|
|
LIST_FOR_EACH_ENTRY_SAFE(instr, next, &block->instrs, struct hlsl_ir_node, entry)
|
2021-03-16 14:31:53 -07:00
|
|
|
{
|
|
|
|
|
if (instr->type == HLSL_IR_IF)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
|
|
|
|
|
2023-04-25 06:04:29 -07:00
|
|
|
progress |= hlsl_transform_ir(ctx, func, &iff->then_block, context);
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, func, &iff->else_block, context);
|
2021-03-16 14:31:53 -07:00
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_LOOP)
|
2023-10-11 04:51:51 -07:00
|
|
|
{
|
2023-04-25 06:04:29 -07:00
|
|
|
progress |= hlsl_transform_ir(ctx, func, &hlsl_ir_loop(instr)->body, context);
|
2023-10-11 04:51:51 -07:00
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_SWITCH)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, func, &c->body, context);
|
|
|
|
|
}
|
|
|
|
|
}
|
2021-03-16 14:31:53 -07:00
|
|
|
|
|
|
|
|
progress |= func(ctx, instr, context);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-06 18:34:10 -08:00
|
|
|
typedef bool (*PFN_lower_func)(struct hlsl_ctx *, struct hlsl_ir_node *, struct hlsl_block *);
|
|
|
|
|
|
|
|
|
|
static bool call_lower_func(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
PFN_lower_func func = context;
|
|
|
|
|
struct hlsl_block block;
|
|
|
|
|
|
|
|
|
|
hlsl_block_init(&block);
|
|
|
|
|
if (func(ctx, instr, &block))
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *replacement = LIST_ENTRY(list_tail(&block.instrs), struct hlsl_ir_node, entry);
|
|
|
|
|
|
|
|
|
|
list_move_before(&instr->entry, &block.instrs);
|
|
|
|
|
hlsl_replace_node(instr, replacement);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
hlsl_block_cleanup(&block);
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Specific form of transform_ir() for passes which convert a single instruction
|
|
|
|
|
* to a block of one or more instructions. This helper takes care of setting up
|
|
|
|
|
* the block and calling hlsl_replace_node_with_block(). */
|
|
|
|
|
static bool lower_ir(struct hlsl_ctx *ctx, PFN_lower_func func, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
return hlsl_transform_ir(ctx, call_lower_func, block, func);
|
|
|
|
|
}
|
|
|
|
|
|
2023-05-04 12:06:58 -07:00
|
|
|
static bool transform_instr_derefs(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
bool res;
|
|
|
|
|
bool (*func)(struct hlsl_ctx *ctx, struct hlsl_deref *, struct hlsl_ir_node *) = context;
|
|
|
|
|
|
|
|
|
|
switch(instr->type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_IR_LOAD:
|
|
|
|
|
res = func(ctx, &hlsl_ir_load(instr)->src, instr);
|
|
|
|
|
return res;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_STORE:
|
|
|
|
|
res = func(ctx, &hlsl_ir_store(instr)->lhs, instr);
|
|
|
|
|
return res;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_RESOURCE_LOAD:
|
|
|
|
|
res = func(ctx, &hlsl_ir_resource_load(instr)->resource, instr);
|
|
|
|
|
if (hlsl_ir_resource_load(instr)->sampler.var)
|
|
|
|
|
res |= func(ctx, &hlsl_ir_resource_load(instr)->sampler, instr);
|
|
|
|
|
return res;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_RESOURCE_STORE:
|
|
|
|
|
res = func(ctx, &hlsl_ir_resource_store(instr)->resource, instr);
|
|
|
|
|
return res;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool transform_derefs(struct hlsl_ctx *ctx,
|
|
|
|
|
bool (*func)(struct hlsl_ctx *ctx, struct hlsl_deref *, struct hlsl_ir_node *),
|
|
|
|
|
struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
return hlsl_transform_ir(ctx, transform_instr_derefs, block, func);
|
|
|
|
|
}
|
|
|
|
|
|
2021-09-11 09:20:32 -07:00
|
|
|
struct recursive_call_ctx
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_function_decl **backtrace;
|
|
|
|
|
size_t count, capacity;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static bool find_recursive_calls(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct recursive_call_ctx *call_ctx = context;
|
|
|
|
|
struct hlsl_ir_function_decl *decl;
|
|
|
|
|
const struct hlsl_ir_call *call;
|
|
|
|
|
size_t i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_CALL)
|
|
|
|
|
return false;
|
|
|
|
|
call = hlsl_ir_call(instr);
|
|
|
|
|
decl = call->decl;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < call_ctx->count; ++i)
|
|
|
|
|
{
|
|
|
|
|
if (call_ctx->backtrace[i] == decl)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &call->node.loc, VKD3D_SHADER_ERROR_HLSL_RECURSIVE_CALL,
|
|
|
|
|
"Recursive call to \"%s\".", decl->func->name);
|
|
|
|
|
/* Native returns E_NOTIMPL instead of E_FAIL here. */
|
|
|
|
|
ctx->result = VKD3D_ERROR_NOT_IMPLEMENTED;
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!hlsl_array_reserve(ctx, (void **)&call_ctx->backtrace, &call_ctx->capacity,
|
|
|
|
|
call_ctx->count + 1, sizeof(*call_ctx->backtrace)))
|
|
|
|
|
return false;
|
|
|
|
|
call_ctx->backtrace[call_ctx->count++] = decl;
|
|
|
|
|
|
2023-04-25 06:04:29 -07:00
|
|
|
hlsl_transform_ir(ctx, find_recursive_calls, &decl->body, call_ctx);
|
2021-09-11 09:20:32 -07:00
|
|
|
|
|
|
|
|
--call_ctx->count;
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2021-09-13 21:08:34 -07:00
|
|
|
static void insert_early_return_break(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_ir_function_decl *func, struct hlsl_ir_node *cf_instr)
|
|
|
|
|
{
|
2022-11-10 18:08:44 -08:00
|
|
|
struct hlsl_ir_node *iff, *jump;
|
2022-11-10 18:04:22 -08:00
|
|
|
struct hlsl_block then_block;
|
2021-09-13 21:08:34 -07:00
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
|
2022-11-10 18:04:22 -08:00
|
|
|
hlsl_block_init(&then_block);
|
|
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(load = hlsl_new_var_load(ctx, func->early_return_var, &cf_instr->loc)))
|
2021-09-13 21:08:34 -07:00
|
|
|
return;
|
|
|
|
|
list_add_after(&cf_instr->entry, &load->node.entry);
|
|
|
|
|
|
2023-06-08 00:42:58 -07:00
|
|
|
if (!(jump = hlsl_new_jump(ctx, HLSL_IR_JUMP_BREAK, NULL, &cf_instr->loc)))
|
2021-09-13 21:08:34 -07:00
|
|
|
return;
|
2022-11-10 18:08:44 -08:00
|
|
|
hlsl_block_add_instr(&then_block, jump);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2022-11-10 18:04:22 -08:00
|
|
|
if (!(iff = hlsl_new_if(ctx, &load->node, &then_block, NULL, &cf_instr->loc)))
|
2021-09-13 21:08:34 -07:00
|
|
|
return;
|
2022-11-10 18:05:53 -08:00
|
|
|
list_add_after(&load->node.entry, &iff->entry);
|
2021-09-13 21:08:34 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Remove HLSL_IR_JUMP_RETURN calls by altering subsequent control flow. */
|
2021-09-13 21:48:38 -07:00
|
|
|
static bool lower_return(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *func,
|
2021-09-13 21:08:34 -07:00
|
|
|
struct hlsl_block *block, bool in_loop)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *return_instr = NULL, *cf_instr = NULL;
|
|
|
|
|
struct hlsl_ir_node *instr, *next;
|
2021-09-13 21:48:38 -07:00
|
|
|
bool has_early_return = false;
|
2021-09-13 21:08:34 -07:00
|
|
|
|
|
|
|
|
/* SM1 has no function calls. SM4 does, but native d3dcompiler inlines
|
|
|
|
|
* everything anyway. We are safest following suit.
|
|
|
|
|
*
|
|
|
|
|
* The basic idea is to keep track of whether the function has executed an
|
|
|
|
|
* early return in a synthesized boolean variable (func->early_return_var)
|
|
|
|
|
* and guard all code after the return on that variable being false. In the
|
|
|
|
|
* case of loops we also replace the return with a break.
|
|
|
|
|
*
|
|
|
|
|
* The following algorithm loops over instructions in a block, recursing
|
|
|
|
|
* into inferior CF blocks, until it hits one of the following two things:
|
|
|
|
|
*
|
|
|
|
|
* - A return statement. In this case, we remove everything after the return
|
|
|
|
|
* statement in this block. We have to stop and do this in a separate
|
|
|
|
|
* loop, because instructions must be deleted in reverse order (due to
|
|
|
|
|
* def-use chains.)
|
|
|
|
|
*
|
|
|
|
|
* If we're inside of a loop CF block, we can instead just turn the
|
|
|
|
|
* return into a break, which offers the right semantics—except that it
|
|
|
|
|
* won't break out of nested loops.
|
|
|
|
|
*
|
2021-09-13 21:48:38 -07:00
|
|
|
* - A CF block which contains a return statement. After calling
|
2021-09-13 21:08:34 -07:00
|
|
|
* lower_return() on the CF block body, we stop, pull out everything after
|
|
|
|
|
* the CF instruction, shove it into an if block, and then lower that if
|
|
|
|
|
* block.
|
|
|
|
|
*
|
2023-04-25 06:04:29 -07:00
|
|
|
* (We could return a "did we make progress" boolean like hlsl_transform_ir()
|
2021-09-13 21:08:34 -07:00
|
|
|
* and run this pass multiple times, but we already know the only block
|
|
|
|
|
* that still needs to be addressed, so there's not much point.)
|
|
|
|
|
*
|
|
|
|
|
* If we're inside of a loop CF block, we again do things differently. We
|
|
|
|
|
* already turned any returns into breaks. If the block we just processed
|
|
|
|
|
* was conditional, then "break" did our work for us. If it was a loop,
|
|
|
|
|
* we need to propagate that break to the outer loop.
|
2021-09-13 21:48:38 -07:00
|
|
|
*
|
|
|
|
|
* We return true if there was an early return anywhere in the block we just
|
|
|
|
|
* processed (including CF contained inside that block).
|
2021-09-13 21:08:34 -07:00
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY_SAFE(instr, next, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
if (instr->type == HLSL_IR_CALL)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_call *call = hlsl_ir_call(instr);
|
|
|
|
|
|
|
|
|
|
lower_return(ctx, call->decl, &call->decl->body, false);
|
|
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_IF)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
|
|
|
|
|
2022-11-10 18:04:22 -08:00
|
|
|
has_early_return |= lower_return(ctx, func, &iff->then_block, in_loop);
|
|
|
|
|
has_early_return |= lower_return(ctx, func, &iff->else_block, in_loop);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2021-09-13 21:48:38 -07:00
|
|
|
if (has_early_return)
|
2021-09-13 21:08:34 -07:00
|
|
|
{
|
2021-09-13 21:48:38 -07:00
|
|
|
/* If we're in a loop, we don't need to do anything here. We
|
|
|
|
|
* turned the return into a break, and that will already skip
|
|
|
|
|
* anything that comes after this "if" block. */
|
|
|
|
|
if (!in_loop)
|
|
|
|
|
{
|
|
|
|
|
cf_instr = instr;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2021-09-13 21:08:34 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_LOOP)
|
|
|
|
|
{
|
2021-09-13 21:48:38 -07:00
|
|
|
has_early_return |= lower_return(ctx, func, &hlsl_ir_loop(instr)->body, true);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2021-09-13 21:48:38 -07:00
|
|
|
if (has_early_return)
|
2021-09-13 21:08:34 -07:00
|
|
|
{
|
2021-09-13 21:48:38 -07:00
|
|
|
if (in_loop)
|
|
|
|
|
{
|
|
|
|
|
/* "instr" is a nested loop. "return" breaks out of all
|
|
|
|
|
* loops, so break out of this one too now. */
|
|
|
|
|
insert_early_return_break(ctx, func, instr);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
cf_instr = instr;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2021-09-13 21:08:34 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_JUMP)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_jump *jump = hlsl_ir_jump(instr);
|
2022-11-10 18:55:03 -08:00
|
|
|
struct hlsl_ir_node *constant, *store;
|
2021-09-13 21:08:34 -07:00
|
|
|
|
|
|
|
|
if (jump->type == HLSL_IR_JUMP_RETURN)
|
|
|
|
|
{
|
|
|
|
|
if (!(constant = hlsl_new_bool_constant(ctx, true, &jump->node.loc)))
|
2021-09-13 21:48:38 -07:00
|
|
|
return false;
|
2022-11-10 17:37:41 -08:00
|
|
|
list_add_before(&jump->node.entry, &constant->entry);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2022-11-10 17:37:41 -08:00
|
|
|
if (!(store = hlsl_new_simple_store(ctx, func->early_return_var, constant)))
|
2021-09-13 21:48:38 -07:00
|
|
|
return false;
|
2022-11-10 18:55:03 -08:00
|
|
|
list_add_after(&constant->entry, &store->entry);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2021-09-13 21:48:38 -07:00
|
|
|
has_early_return = true;
|
2021-09-13 21:08:34 -07:00
|
|
|
if (in_loop)
|
|
|
|
|
{
|
|
|
|
|
jump->type = HLSL_IR_JUMP_BREAK;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
return_instr = instr;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
2023-10-11 04:51:51 -07:00
|
|
|
else if (instr->type == HLSL_IR_SWITCH)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
has_early_return |= lower_return(ctx, func, &c->body, true);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (has_early_return)
|
|
|
|
|
{
|
|
|
|
|
if (in_loop)
|
|
|
|
|
{
|
|
|
|
|
/* For a 'switch' nested in a loop append a break after the 'switch'. */
|
|
|
|
|
insert_early_return_break(ctx, func, instr);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
cf_instr = instr;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
2021-09-13 21:08:34 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (return_instr)
|
|
|
|
|
{
|
|
|
|
|
/* If we're in a loop, we should have used "break" instead. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!in_loop);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
|
|
|
|
/* Iterate in reverse, to avoid use-after-free when unlinking sources from
|
|
|
|
|
* the "uses" list. */
|
|
|
|
|
LIST_FOR_EACH_ENTRY_SAFE_REV(instr, next, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
list_remove(&instr->entry);
|
|
|
|
|
hlsl_free_instr(instr);
|
|
|
|
|
|
|
|
|
|
/* Yes, we just freed it, but we're comparing pointers. */
|
|
|
|
|
if (instr == return_instr)
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else if (cf_instr)
|
|
|
|
|
{
|
|
|
|
|
struct list *tail = list_tail(&block->instrs);
|
2022-11-10 18:05:53 -08:00
|
|
|
struct hlsl_ir_node *not, *iff;
|
2022-11-10 18:04:22 -08:00
|
|
|
struct hlsl_block then_block;
|
2021-09-13 21:08:34 -07:00
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
|
|
|
|
|
/* If we're in a loop, we should have used "break" instead. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!in_loop);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
|
|
|
|
if (tail == &cf_instr->entry)
|
2021-09-13 21:48:38 -07:00
|
|
|
return has_early_return;
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2022-11-10 18:04:22 -08:00
|
|
|
hlsl_block_init(&then_block);
|
|
|
|
|
list_move_slice_tail(&then_block.instrs, list_next(&block->instrs, &cf_instr->entry), tail);
|
|
|
|
|
lower_return(ctx, func, &then_block, in_loop);
|
|
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(load = hlsl_new_var_load(ctx, func->early_return_var, &cf_instr->loc)))
|
2021-09-13 21:48:38 -07:00
|
|
|
return false;
|
2022-11-11 19:13:50 -08:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(not = hlsl_new_unary_expr(ctx, HLSL_OP1_LOGIC_NOT, &load->node, &cf_instr->loc)))
|
2021-09-13 21:48:38 -07:00
|
|
|
return false;
|
2022-11-11 19:13:50 -08:00
|
|
|
hlsl_block_add_instr(block, not);
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2022-11-10 18:04:22 -08:00
|
|
|
if (!(iff = hlsl_new_if(ctx, not, &then_block, NULL, &cf_instr->loc)))
|
2021-09-13 21:48:38 -07:00
|
|
|
return false;
|
2022-11-10 18:05:53 -08:00
|
|
|
list_add_tail(&block->instrs, &iff->entry);
|
2021-09-13 21:08:34 -07:00
|
|
|
}
|
2021-09-13 21:48:38 -07:00
|
|
|
|
|
|
|
|
return has_early_return;
|
2021-09-13 21:08:34 -07:00
|
|
|
}
|
|
|
|
|
|
2021-09-11 14:56:04 -07:00
|
|
|
/* Remove HLSL_IR_CALL instructions by inlining them. */
|
|
|
|
|
static bool lower_calls(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_function_decl *decl;
|
|
|
|
|
struct hlsl_ir_call *call;
|
|
|
|
|
struct hlsl_block block;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_CALL)
|
|
|
|
|
return false;
|
|
|
|
|
call = hlsl_ir_call(instr);
|
|
|
|
|
decl = call->decl;
|
|
|
|
|
|
|
|
|
|
if (!decl->has_body)
|
|
|
|
|
hlsl_error(ctx, &call->node.loc, VKD3D_SHADER_ERROR_HLSL_NOT_DEFINED,
|
|
|
|
|
"Function \"%s\" is not defined.", decl->func->name);
|
|
|
|
|
|
|
|
|
|
if (!hlsl_clone_block(ctx, &block, &decl->body))
|
|
|
|
|
return false;
|
|
|
|
|
list_move_before(&call->node.entry, &block.instrs);
|
|
|
|
|
|
|
|
|
|
list_remove(&call->node.entry);
|
|
|
|
|
hlsl_free_instr(&call->node);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-28 10:03:36 -07:00
|
|
|
static struct hlsl_ir_node *add_zero_mipmap_level(struct hlsl_ctx *ctx, struct hlsl_ir_node *index,
|
|
|
|
|
const struct vkd3d_shader_location *loc)
|
|
|
|
|
{
|
|
|
|
|
unsigned int dim_count = index->data_type->dimx;
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *store, *zero;
|
2023-04-28 10:03:36 -07:00
|
|
|
struct hlsl_ir_load *coords_load;
|
|
|
|
|
struct hlsl_deref coords_deref;
|
|
|
|
|
struct hlsl_ir_var *coords;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(dim_count < 4);
|
2023-04-28 10:03:36 -07:00
|
|
|
|
|
|
|
|
if (!(coords = hlsl_new_synthetic_var(ctx, "coords",
|
|
|
|
|
hlsl_get_vector_type(ctx, HLSL_TYPE_UINT, dim_count + 1), loc)))
|
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
|
|
hlsl_init_simple_deref_from_var(&coords_deref, coords);
|
|
|
|
|
if (!(store = hlsl_new_store_index(ctx, &coords_deref, NULL, index, (1u << dim_count) - 1, loc)))
|
|
|
|
|
return NULL;
|
2022-11-10 18:57:00 -08:00
|
|
|
list_add_after(&index->entry, &store->entry);
|
2023-04-28 10:03:36 -07:00
|
|
|
|
|
|
|
|
if (!(zero = hlsl_new_uint_constant(ctx, 0, loc)))
|
|
|
|
|
return NULL;
|
2022-11-10 19:06:04 -08:00
|
|
|
list_add_after(&store->entry, &zero->entry);
|
2023-04-28 10:03:36 -07:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(store = hlsl_new_store_index(ctx, &coords_deref, NULL, zero, 1u << dim_count, loc)))
|
2023-04-28 10:03:36 -07:00
|
|
|
return NULL;
|
2022-11-10 19:06:04 -08:00
|
|
|
list_add_after(&zero->entry, &store->entry);
|
2023-04-28 10:03:36 -07:00
|
|
|
|
|
|
|
|
if (!(coords_load = hlsl_new_var_load(ctx, coords, loc)))
|
|
|
|
|
return NULL;
|
2022-11-10 18:57:00 -08:00
|
|
|
list_add_after(&store->entry, &coords_load->node.entry);
|
2023-04-28 10:03:36 -07:00
|
|
|
|
|
|
|
|
return &coords_load->node;
|
|
|
|
|
}
|
|
|
|
|
|
2023-09-06 17:01:49 -07:00
|
|
|
/* hlsl_ir_swizzle nodes that directly point to a matrix value are only a parse-time construct that
|
|
|
|
|
* represents matrix swizzles (e.g. mat._m01_m23) before we know if they will be used in the lhs of
|
|
|
|
|
* an assignment or as a value made from different components of the matrix. The former cases should
|
|
|
|
|
* have already been split into several separate assignments, but the latter are lowered by this
|
|
|
|
|
* pass. */
|
|
|
|
|
static bool lower_matrix_swizzles(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_swizzle *swizzle;
|
|
|
|
|
struct hlsl_ir_load *var_load;
|
|
|
|
|
struct hlsl_deref var_deref;
|
|
|
|
|
struct hlsl_type *matrix_type;
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
unsigned int x, y, k, i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_SWIZZLE)
|
|
|
|
|
return false;
|
|
|
|
|
swizzle = hlsl_ir_swizzle(instr);
|
|
|
|
|
matrix_type = swizzle->val.node->data_type;
|
|
|
|
|
if (matrix_type->class != HLSL_CLASS_MATRIX)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!(var = hlsl_new_synthetic_var(ctx, "matrix-swizzle", instr->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_init_simple_deref_from_var(&var_deref, var);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < instr->data_type->dimx; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_block store_block;
|
|
|
|
|
struct hlsl_ir_node *load;
|
|
|
|
|
|
|
|
|
|
y = (swizzle->swizzle >> (8 * i + 4)) & 0xf;
|
|
|
|
|
x = (swizzle->swizzle >> 8 * i) & 0xf;
|
|
|
|
|
k = y * matrix_type->dimx + x;
|
|
|
|
|
|
|
|
|
|
if (!(load = hlsl_add_load_component(ctx, block, swizzle->val.node, k, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!hlsl_new_store_component(ctx, &store_block, &var_deref, i, load))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_block(block, &store_block);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(var_load = hlsl_new_var_load(ctx, var, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, &var_load->node);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-10 17:09:58 -08:00
|
|
|
/* hlsl_ir_index nodes are a parse-time construct used to represent array indexing and struct
|
|
|
|
|
* record access before knowing if they will be used in the lhs of an assignment --in which case
|
|
|
|
|
* they are lowered into a deref-- or as the load of an element within a larger value.
|
|
|
|
|
* For the latter case, this pass takes care of lowering hlsl_ir_indexes into individual
|
2023-03-10 12:23:49 -08:00
|
|
|
* hlsl_ir_loads, or individual hlsl_ir_resource_loads, in case the indexing is a
|
|
|
|
|
* resource access. */
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_index_loads(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-03-10 17:09:58 -08:00
|
|
|
{
|
2022-11-10 18:55:03 -08:00
|
|
|
struct hlsl_ir_node *val, *store;
|
2023-03-10 17:09:58 -08:00
|
|
|
struct hlsl_deref var_deref;
|
|
|
|
|
struct hlsl_ir_index *index;
|
|
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_INDEX)
|
|
|
|
|
return false;
|
|
|
|
|
index = hlsl_ir_index(instr);
|
|
|
|
|
val = index->val.node;
|
|
|
|
|
|
2023-03-10 12:23:49 -08:00
|
|
|
if (hlsl_index_is_resource_access(index))
|
|
|
|
|
{
|
2023-04-28 10:03:36 -07:00
|
|
|
unsigned int dim_count = hlsl_sampler_dim_count(val->data_type->sampler_dim);
|
|
|
|
|
struct hlsl_ir_node *coords = index->idx.node;
|
2023-03-10 12:23:49 -08:00
|
|
|
struct hlsl_resource_load_params params = {0};
|
2023-04-30 15:40:04 -07:00
|
|
|
struct hlsl_ir_node *resource_load;
|
2023-03-10 12:23:49 -08:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(coords->data_type->class == HLSL_CLASS_VECTOR);
|
|
|
|
|
VKD3D_ASSERT(coords->data_type->e.numeric.type == HLSL_TYPE_UINT);
|
|
|
|
|
VKD3D_ASSERT(coords->data_type->dimx == dim_count);
|
2023-04-28 10:03:36 -07:00
|
|
|
|
|
|
|
|
if (!(coords = add_zero_mipmap_level(ctx, coords, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
|
2023-03-10 12:23:49 -08:00
|
|
|
params.type = HLSL_RESOURCE_LOAD;
|
|
|
|
|
params.resource = val;
|
2023-04-28 10:03:36 -07:00
|
|
|
params.coords = coords;
|
2024-02-19 04:57:54 -08:00
|
|
|
params.format = val->data_type->e.resource.format;
|
2023-03-10 12:23:49 -08:00
|
|
|
|
2023-04-30 15:40:04 -07:00
|
|
|
if (!(resource_load = hlsl_new_resource_load(ctx, ¶ms, &instr->loc)))
|
2023-03-10 12:23:49 -08:00
|
|
|
return false;
|
2023-04-30 15:40:04 -07:00
|
|
|
hlsl_block_add_instr(block, resource_load);
|
2023-03-10 12:23:49 -08:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-10 17:09:58 -08:00
|
|
|
if (!(var = hlsl_new_synthetic_var(ctx, "index-val", val->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_init_simple_deref_from_var(&var_deref, var);
|
|
|
|
|
|
|
|
|
|
if (!(store = hlsl_new_simple_store(ctx, var, val)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, store);
|
2023-03-10 17:09:58 -08:00
|
|
|
|
|
|
|
|
if (hlsl_index_is_noncontiguous(index))
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *mat = index->val.node;
|
|
|
|
|
struct hlsl_deref row_deref;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!hlsl_type_is_row_major(mat->data_type));
|
2023-03-10 17:09:58 -08:00
|
|
|
|
|
|
|
|
if (!(var = hlsl_new_synthetic_var(ctx, "row", instr->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_init_simple_deref_from_var(&row_deref, var);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < mat->data_type->dimx; ++i)
|
|
|
|
|
{
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *c;
|
2023-03-10 17:09:58 -08:00
|
|
|
|
|
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, i, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, c);
|
2023-03-10 17:09:58 -08:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(load = hlsl_new_load_index(ctx, &var_deref, c, &instr->loc)))
|
2023-03-10 17:09:58 -08:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2023-03-10 17:09:58 -08:00
|
|
|
|
|
|
|
|
if (!(load = hlsl_new_load_index(ctx, &load->src, index->idx.node, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2023-03-10 17:09:58 -08:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(store = hlsl_new_store_index(ctx, &row_deref, c, &load->node, 0, &instr->loc)))
|
2023-03-10 17:09:58 -08:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, store);
|
2023-03-10 17:09:58 -08:00
|
|
|
}
|
|
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(load = hlsl_new_var_load(ctx, var, &instr->loc)))
|
2023-03-10 17:09:58 -08:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2023-03-10 17:09:58 -08:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
if (!(load = hlsl_new_load_index(ctx, &var_deref, index->idx.node, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, &load->node);
|
2023-03-10 17:09:58 -08:00
|
|
|
}
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2021-12-01 08:14:55 -08:00
|
|
|
/* Lower casts from vec1 to vecN to swizzles. */
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_broadcasts(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2021-12-01 08:14:55 -08:00
|
|
|
{
|
|
|
|
|
const struct hlsl_type *src_type, *dst_type;
|
|
|
|
|
struct hlsl_type *dst_scalar_type;
|
|
|
|
|
struct hlsl_ir_expr *cast;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
cast = hlsl_ir_expr(instr);
|
2021-09-10 14:35:54 -07:00
|
|
|
if (cast->op != HLSL_OP1_CAST)
|
|
|
|
|
return false;
|
2021-12-01 08:14:55 -08:00
|
|
|
src_type = cast->operands[0].node->data_type;
|
|
|
|
|
dst_type = cast->node.data_type;
|
|
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
if (src_type->class <= HLSL_CLASS_VECTOR && dst_type->class <= HLSL_CLASS_VECTOR && src_type->dimx == 1)
|
2021-12-01 08:14:55 -08:00
|
|
|
{
|
2023-06-25 16:46:10 -07:00
|
|
|
struct hlsl_ir_node *new_cast, *swizzle;
|
2021-12-01 08:14:55 -08:00
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
dst_scalar_type = hlsl_get_scalar_type(ctx, dst_type->e.numeric.type);
|
2021-12-01 08:14:55 -08:00
|
|
|
/* We need to preserve the cast since it might be doing more than just
|
|
|
|
|
* turning the scalar into a vector. */
|
|
|
|
|
if (!(new_cast = hlsl_new_cast(ctx, cast->operands[0].node, dst_scalar_type, &cast->node.loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, new_cast);
|
2022-04-27 01:56:22 -07:00
|
|
|
|
|
|
|
|
if (dst_type->dimx != 1)
|
|
|
|
|
{
|
2023-06-25 16:46:10 -07:00
|
|
|
if (!(swizzle = hlsl_new_swizzle(ctx, HLSL_SWIZZLE(X, X, X, X), dst_type->dimx, new_cast, &cast->node.loc)))
|
2022-04-27 01:56:22 -07:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, swizzle);
|
2022-04-27 01:56:22 -07:00
|
|
|
}
|
2021-12-01 08:14:55 -08:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2023-11-17 10:29:54 -08:00
|
|
|
/* Allocate a unique, ordered index to each instruction, which will be used for
|
2023-11-21 14:09:32 -08:00
|
|
|
* copy propagation and computing liveness ranges.
|
2023-11-17 10:29:54 -08:00
|
|
|
* Index 0 means unused; index 1 means function entry, so start at 2. */
|
|
|
|
|
static unsigned int index_instructions(struct hlsl_block *block, unsigned int index)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(instr, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
instr->index = index++;
|
|
|
|
|
|
|
|
|
|
if (instr->type == HLSL_IR_IF)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
|
|
|
|
index = index_instructions(&iff->then_block, index);
|
|
|
|
|
index = index_instructions(&iff->else_block, index);
|
|
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_LOOP)
|
|
|
|
|
{
|
|
|
|
|
index = index_instructions(&hlsl_ir_loop(instr)->body, index);
|
|
|
|
|
hlsl_ir_loop(instr)->next_index = index;
|
|
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_SWITCH)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
index = index_instructions(&c->body, index);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return index;
|
|
|
|
|
}
|
|
|
|
|
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
/*
|
|
|
|
|
* Copy propagation. The basic idea is to recognize instruction sequences of the
|
|
|
|
|
* form:
|
|
|
|
|
*
|
|
|
|
|
* 2: <any instruction>
|
|
|
|
|
* 3: v = @2
|
|
|
|
|
* 4: load(v)
|
|
|
|
|
*
|
|
|
|
|
* and replace the load (@4) with the original instruction (@2).
|
|
|
|
|
* This works for multiple components, even if they're written using separate
|
|
|
|
|
* store instructions, as long as the rhs is the same in every case. This basic
|
|
|
|
|
* detection is implemented by copy_propagation_replace_with_single_instr().
|
|
|
|
|
*
|
2023-01-12 15:26:03 -08:00
|
|
|
* In some cases, the load itself might not have a single source, but a
|
|
|
|
|
* subsequent swizzle might; hence we also try to replace swizzles of loads.
|
|
|
|
|
*
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
* We use the same infrastructure to implement a more specialized
|
|
|
|
|
* transformation. We recognize sequences of the form:
|
|
|
|
|
*
|
|
|
|
|
* 2: 123
|
|
|
|
|
* 3: var.x = @2
|
|
|
|
|
* 4: 345
|
|
|
|
|
* 5: var.y = @4
|
|
|
|
|
* 6: load(var.xy)
|
|
|
|
|
*
|
|
|
|
|
* where the load (@6) originates from different sources but that are constant,
|
|
|
|
|
* and transform it into a single constant vector. This latter pass is done
|
|
|
|
|
* by copy_propagation_replace_with_constant_vector().
|
|
|
|
|
*
|
|
|
|
|
* This is a specialized form of vectorization, and begs the question: why does
|
|
|
|
|
* the load need to be involved? Can we just vectorize the stores into a single
|
|
|
|
|
* instruction, and then use "normal" copy-prop to convert that into a single
|
|
|
|
|
* vector?
|
|
|
|
|
*
|
|
|
|
|
* In general, the answer is yes, but there is a special case which necessitates
|
|
|
|
|
* the use of this transformation: non-uniform control flow. Copy-prop can act
|
|
|
|
|
* across some control flow, and in cases like the following:
|
|
|
|
|
*
|
|
|
|
|
* 2: 123
|
|
|
|
|
* 3: var.x = @2
|
|
|
|
|
* 4: if (...)
|
|
|
|
|
* 5: 456
|
|
|
|
|
* 6: var.y = @5
|
|
|
|
|
* 7: load(var.xy)
|
|
|
|
|
*
|
|
|
|
|
* we can copy-prop the load (@7) into a constant vector {123, 456}, but we
|
|
|
|
|
* cannot easily vectorize the stores @3 and @6.
|
|
|
|
|
*/
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
struct copy_propagation_value
|
|
|
|
|
{
|
2023-11-21 14:09:32 -08:00
|
|
|
unsigned int timestamp;
|
|
|
|
|
/* If node is NULL, the value was dynamically written and thus, it is unknown.*/
|
2021-12-01 08:14:50 -08:00
|
|
|
struct hlsl_ir_node *node;
|
|
|
|
|
unsigned int component;
|
|
|
|
|
};
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
struct copy_propagation_component_trace
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_value *records;
|
|
|
|
|
size_t record_count, record_capacity;
|
|
|
|
|
};
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
struct copy_propagation_var_def
|
|
|
|
|
{
|
|
|
|
|
struct rb_entry entry;
|
|
|
|
|
struct hlsl_ir_var *var;
|
2023-11-21 14:09:32 -08:00
|
|
|
struct copy_propagation_component_trace traces[];
|
2021-12-01 08:14:50 -08:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
struct copy_propagation_state
|
|
|
|
|
{
|
|
|
|
|
struct rb_tree var_defs;
|
2022-04-28 06:32:05 -07:00
|
|
|
struct copy_propagation_state *parent;
|
2021-12-01 08:14:50 -08:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static int copy_propagation_var_def_compare(const void *key, const struct rb_entry *entry)
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_var_def *var_def = RB_ENTRY_VALUE(entry, struct copy_propagation_var_def, entry);
|
|
|
|
|
uintptr_t key_int = (uintptr_t)key, entry_int = (uintptr_t)var_def->var;
|
|
|
|
|
|
|
|
|
|
return (key_int > entry_int) - (key_int < entry_int);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void copy_propagation_var_def_destroy(struct rb_entry *entry, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_var_def *var_def = RB_ENTRY_VALUE(entry, struct copy_propagation_var_def, entry);
|
2023-11-21 14:09:32 -08:00
|
|
|
unsigned int component_count = hlsl_type_component_count(var_def->var->data_type);
|
|
|
|
|
unsigned int i;
|
2021-12-01 08:14:50 -08:00
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
for (i = 0; i < component_count; ++i)
|
|
|
|
|
vkd3d_free(var_def->traces[i].records);
|
2021-12-01 08:14:50 -08:00
|
|
|
vkd3d_free(var_def);
|
|
|
|
|
}
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
static struct copy_propagation_value *copy_propagation_get_value_at_time(
|
|
|
|
|
struct copy_propagation_component_trace *trace, unsigned int time)
|
|
|
|
|
{
|
|
|
|
|
int r;
|
|
|
|
|
|
|
|
|
|
for (r = trace->record_count - 1; r >= 0; --r)
|
|
|
|
|
{
|
|
|
|
|
if (trace->records[r].timestamp < time)
|
|
|
|
|
return &trace->records[r];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2022-04-28 06:32:05 -07:00
|
|
|
static struct copy_propagation_value *copy_propagation_get_value(const struct copy_propagation_state *state,
|
2023-11-21 14:09:32 -08:00
|
|
|
const struct hlsl_ir_var *var, unsigned int component, unsigned int time)
|
2021-12-01 08:14:50 -08:00
|
|
|
{
|
2022-04-28 06:32:05 -07:00
|
|
|
for (; state; state = state->parent)
|
|
|
|
|
{
|
|
|
|
|
struct rb_entry *entry = rb_get(&state->var_defs, var);
|
|
|
|
|
if (entry)
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_var_def *var_def = RB_ENTRY_VALUE(entry, struct copy_propagation_var_def, entry);
|
2022-07-20 12:37:07 -07:00
|
|
|
unsigned int component_count = hlsl_type_component_count(var->data_type);
|
2023-11-21 14:09:32 -08:00
|
|
|
struct copy_propagation_value *value;
|
2022-04-28 06:32:05 -07:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(component < component_count);
|
2023-11-21 14:09:32 -08:00
|
|
|
value = copy_propagation_get_value_at_time(&var_def->traces[component], time);
|
2022-04-28 06:32:05 -07:00
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
if (!value)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (value->node)
|
|
|
|
|
return value;
|
|
|
|
|
else
|
|
|
|
|
return NULL;
|
2022-04-28 06:32:05 -07:00
|
|
|
}
|
|
|
|
|
}
|
2021-12-01 08:14:50 -08:00
|
|
|
|
2022-04-28 06:32:05 -07:00
|
|
|
return NULL;
|
2021-12-01 08:14:50 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct copy_propagation_var_def *copy_propagation_create_var_def(struct hlsl_ctx *ctx,
|
|
|
|
|
struct copy_propagation_state *state, struct hlsl_ir_var *var)
|
|
|
|
|
{
|
|
|
|
|
struct rb_entry *entry = rb_get(&state->var_defs, var);
|
|
|
|
|
struct copy_propagation_var_def *var_def;
|
2022-07-20 12:37:07 -07:00
|
|
|
unsigned int component_count = hlsl_type_component_count(var->data_type);
|
2021-12-01 08:14:50 -08:00
|
|
|
int res;
|
|
|
|
|
|
|
|
|
|
if (entry)
|
|
|
|
|
return RB_ENTRY_VALUE(entry, struct copy_propagation_var_def, entry);
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
if (!(var_def = hlsl_alloc(ctx, offsetof(struct copy_propagation_var_def, traces[component_count]))))
|
2021-12-01 08:14:50 -08:00
|
|
|
return NULL;
|
|
|
|
|
|
|
|
|
|
var_def->var = var;
|
|
|
|
|
|
|
|
|
|
res = rb_put(&state->var_defs, var, &var_def->entry);
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!res);
|
2021-12-01 08:14:50 -08:00
|
|
|
|
|
|
|
|
return var_def;
|
|
|
|
|
}
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
static void copy_propagation_trace_record_value(struct hlsl_ctx *ctx,
|
|
|
|
|
struct copy_propagation_component_trace *trace, struct hlsl_ir_node *node,
|
|
|
|
|
unsigned int component, unsigned int time)
|
|
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!trace->record_count || trace->records[trace->record_count - 1].timestamp < time);
|
2023-11-21 14:09:32 -08:00
|
|
|
|
|
|
|
|
if (!hlsl_array_reserve(ctx, (void **)&trace->records, &trace->record_capacity,
|
|
|
|
|
trace->record_count + 1, sizeof(trace->records[0])))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
trace->records[trace->record_count].timestamp = time;
|
|
|
|
|
trace->records[trace->record_count].node = node;
|
|
|
|
|
trace->records[trace->record_count].component = component;
|
|
|
|
|
|
|
|
|
|
++trace->record_count;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void copy_propagation_invalidate_variable(struct hlsl_ctx *ctx, struct copy_propagation_var_def *var_def,
|
|
|
|
|
unsigned int comp, unsigned char writemask, unsigned int time)
|
2022-04-28 06:32:05 -07:00
|
|
|
{
|
|
|
|
|
unsigned i;
|
|
|
|
|
|
2022-07-20 12:37:07 -07:00
|
|
|
TRACE("Invalidate variable %s[%u]%s.\n", var_def->var->name, comp, debug_hlsl_writemask(writemask));
|
2022-04-28 06:32:05 -07:00
|
|
|
|
|
|
|
|
for (i = 0; i < 4; ++i)
|
|
|
|
|
{
|
|
|
|
|
if (writemask & (1u << i))
|
2023-11-21 14:09:32 -08:00
|
|
|
{
|
|
|
|
|
struct copy_propagation_component_trace *trace = &var_def->traces[comp + i];
|
|
|
|
|
|
|
|
|
|
/* Don't add an invalidate record if it is already present. */
|
|
|
|
|
if (trace->record_count && trace->records[trace->record_count - 1].timestamp == time)
|
|
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!trace->records[trace->record_count - 1].node);
|
2023-11-21 14:09:32 -08:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
copy_propagation_trace_record_value(ctx, trace, NULL, 0, time);
|
|
|
|
|
}
|
2022-04-28 06:32:05 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2022-07-20 14:42:13 -07:00
|
|
|
static void copy_propagation_invalidate_variable_from_deref_recurse(struct hlsl_ctx *ctx,
|
|
|
|
|
struct copy_propagation_var_def *var_def, const struct hlsl_deref *deref,
|
2023-11-21 14:09:32 -08:00
|
|
|
struct hlsl_type *type, unsigned int depth, unsigned int comp_start, unsigned char writemask,
|
|
|
|
|
unsigned int time)
|
2021-12-01 08:14:50 -08:00
|
|
|
{
|
2022-07-20 14:42:13 -07:00
|
|
|
unsigned int i, subtype_comp_count;
|
|
|
|
|
struct hlsl_ir_node *path_node;
|
|
|
|
|
struct hlsl_type *subtype;
|
|
|
|
|
|
|
|
|
|
if (depth == deref->path_len)
|
|
|
|
|
{
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_variable(ctx, var_def, comp_start, writemask, time);
|
2022-07-20 14:42:13 -07:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
path_node = deref->path[depth].node;
|
|
|
|
|
subtype = hlsl_get_element_type_from_path_index(ctx, type, path_node);
|
2022-04-28 06:32:05 -07:00
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class == HLSL_CLASS_STRUCT)
|
2022-07-20 14:42:13 -07:00
|
|
|
{
|
2022-11-11 16:39:55 -08:00
|
|
|
unsigned int idx = hlsl_ir_constant(path_node)->value.u[0].u;
|
2022-07-20 14:42:13 -07:00
|
|
|
|
|
|
|
|
for (i = 0; i < idx; ++i)
|
|
|
|
|
comp_start += hlsl_type_component_count(type->e.record.fields[i].type);
|
|
|
|
|
|
|
|
|
|
copy_propagation_invalidate_variable_from_deref_recurse(ctx, var_def, deref, subtype,
|
2023-11-21 14:09:32 -08:00
|
|
|
depth + 1, comp_start, writemask, time);
|
2022-07-20 14:42:13 -07:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
subtype_comp_count = hlsl_type_component_count(subtype);
|
2022-04-28 06:32:05 -07:00
|
|
|
|
2022-07-20 14:42:13 -07:00
|
|
|
if (path_node->type == HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
|
|
|
|
copy_propagation_invalidate_variable_from_deref_recurse(ctx, var_def, deref, subtype,
|
2023-11-21 14:09:32 -08:00
|
|
|
depth + 1, hlsl_ir_constant(path_node)->value.u[0].u * subtype_comp_count,
|
|
|
|
|
writemask, time);
|
2022-07-20 14:42:13 -07:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
for (i = 0; i < hlsl_type_element_count(type); ++i)
|
|
|
|
|
{
|
|
|
|
|
copy_propagation_invalidate_variable_from_deref_recurse(ctx, var_def, deref, subtype,
|
2023-11-21 14:09:32 -08:00
|
|
|
depth + 1, i * subtype_comp_count, writemask, time);
|
2022-07-20 14:42:13 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void copy_propagation_invalidate_variable_from_deref(struct hlsl_ctx *ctx,
|
2023-11-21 14:09:32 -08:00
|
|
|
struct copy_propagation_var_def *var_def, const struct hlsl_deref *deref,
|
|
|
|
|
unsigned char writemask, unsigned int time)
|
2022-07-20 14:42:13 -07:00
|
|
|
{
|
|
|
|
|
copy_propagation_invalidate_variable_from_deref_recurse(ctx, var_def, deref, deref->var->data_type,
|
2023-11-21 14:09:32 -08:00
|
|
|
0, 0, writemask, time);
|
2021-12-01 08:14:50 -08:00
|
|
|
}
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
static void copy_propagation_set_value(struct hlsl_ctx *ctx, struct copy_propagation_var_def *var_def,
|
|
|
|
|
unsigned int comp, unsigned char writemask, struct hlsl_ir_node *instr, unsigned int time)
|
2021-12-01 08:14:50 -08:00
|
|
|
{
|
|
|
|
|
unsigned int i, j = 0;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < 4; ++i)
|
|
|
|
|
{
|
|
|
|
|
if (writemask & (1u << i))
|
|
|
|
|
{
|
2023-11-21 14:09:32 -08:00
|
|
|
struct copy_propagation_component_trace *trace = &var_def->traces[comp + i];
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
TRACE("Variable %s[%u] is written by instruction %p%s.\n",
|
2022-07-20 12:37:07 -07:00
|
|
|
var_def->var->name, comp + i, instr, debug_hlsl_writemask(1u << i));
|
2023-11-21 14:09:32 -08:00
|
|
|
|
|
|
|
|
copy_propagation_trace_record_value(ctx, trace, instr, j++, time);
|
2021-12-01 08:14:50 -08:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-12 14:28:44 -08:00
|
|
|
static bool copy_propagation_replace_with_single_instr(struct hlsl_ctx *ctx,
|
2023-11-21 14:48:50 -08:00
|
|
|
const struct copy_propagation_state *state, const struct hlsl_ir_load *load,
|
2023-12-06 09:20:25 -08:00
|
|
|
uint32_t swizzle, struct hlsl_ir_node *instr)
|
2021-12-01 08:14:50 -08:00
|
|
|
{
|
2023-01-12 15:26:03 -08:00
|
|
|
const unsigned int instr_component_count = hlsl_type_component_count(instr->data_type);
|
2023-11-21 14:48:50 -08:00
|
|
|
const struct hlsl_deref *deref = &load->src;
|
2022-01-21 13:22:27 -08:00
|
|
|
const struct hlsl_ir_var *var = deref->var;
|
2023-01-12 14:28:44 -08:00
|
|
|
struct hlsl_ir_node *new_instr = NULL;
|
2023-11-21 14:48:50 -08:00
|
|
|
unsigned int time = load->node.index;
|
2022-07-20 12:37:07 -07:00
|
|
|
unsigned int start, count, i;
|
2023-12-06 09:20:25 -08:00
|
|
|
uint32_t ret_swizzle = 0;
|
2022-01-21 13:22:27 -08:00
|
|
|
|
2022-07-20 12:37:07 -07:00
|
|
|
if (!hlsl_component_index_range_from_deref(ctx, deref, &start, &count))
|
2023-01-12 14:28:44 -08:00
|
|
|
return false;
|
2021-12-01 08:14:50 -08:00
|
|
|
|
2023-01-12 15:26:03 -08:00
|
|
|
for (i = 0; i < instr_component_count; ++i)
|
2021-12-01 08:14:50 -08:00
|
|
|
{
|
2023-01-12 15:26:03 -08:00
|
|
|
struct copy_propagation_value *value;
|
2022-04-28 06:32:05 -07:00
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
if (!(value = copy_propagation_get_value(state, var, start + hlsl_swizzle_get_component(swizzle, i),
|
2023-11-21 14:48:50 -08:00
|
|
|
time)))
|
2023-01-12 14:28:44 -08:00
|
|
|
return false;
|
2022-04-28 06:32:05 -07:00
|
|
|
|
2023-01-12 14:28:44 -08:00
|
|
|
if (!new_instr)
|
2022-01-21 13:22:27 -08:00
|
|
|
{
|
2023-01-12 14:28:44 -08:00
|
|
|
new_instr = value->node;
|
2022-01-21 13:22:27 -08:00
|
|
|
}
|
2023-01-12 14:28:44 -08:00
|
|
|
else if (new_instr != value->node)
|
2022-01-21 13:22:27 -08:00
|
|
|
{
|
2023-01-12 15:26:03 -08:00
|
|
|
TRACE("No single source for propagating load from %s[%u-%u]%s\n",
|
|
|
|
|
var->name, start, start + count, debug_hlsl_swizzle(swizzle, instr_component_count));
|
2023-01-12 14:28:44 -08:00
|
|
|
return false;
|
2022-01-21 13:22:27 -08:00
|
|
|
}
|
2023-01-12 14:28:44 -08:00
|
|
|
ret_swizzle |= value->component << HLSL_SWIZZLE_SHIFT(i);
|
2021-12-01 08:14:50 -08:00
|
|
|
}
|
|
|
|
|
|
2023-01-12 15:26:03 -08:00
|
|
|
TRACE("Load from %s[%u-%u]%s propagated as instruction %p%s.\n",
|
|
|
|
|
var->name, start, start + count, debug_hlsl_swizzle(swizzle, instr_component_count),
|
|
|
|
|
new_instr, debug_hlsl_swizzle(ret_swizzle, instr_component_count));
|
2023-01-12 14:28:44 -08:00
|
|
|
|
2024-01-26 15:29:54 -08:00
|
|
|
if (new_instr->data_type->class == HLSL_CLASS_SCALAR || new_instr->data_type->class == HLSL_CLASS_VECTOR)
|
2023-01-12 14:28:44 -08:00
|
|
|
{
|
2022-11-10 19:01:18 -08:00
|
|
|
struct hlsl_ir_node *swizzle_node;
|
2023-01-12 14:28:44 -08:00
|
|
|
|
2023-02-14 14:01:18 -08:00
|
|
|
if (!(swizzle_node = hlsl_new_swizzle(ctx, ret_swizzle, instr_component_count, new_instr, &instr->loc)))
|
2023-01-12 14:28:44 -08:00
|
|
|
return false;
|
2022-11-10 19:01:18 -08:00
|
|
|
list_add_before(&instr->entry, &swizzle_node->entry);
|
|
|
|
|
new_instr = swizzle_node;
|
2023-01-12 14:28:44 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hlsl_replace_node(instr, new_instr);
|
|
|
|
|
return true;
|
2021-12-01 08:14:50 -08:00
|
|
|
}
|
|
|
|
|
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
static bool copy_propagation_replace_with_constant_vector(struct hlsl_ctx *ctx,
|
2023-11-21 14:48:50 -08:00
|
|
|
const struct copy_propagation_state *state, const struct hlsl_ir_load *load,
|
2023-12-06 09:20:25 -08:00
|
|
|
uint32_t swizzle, struct hlsl_ir_node *instr)
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
{
|
2023-01-12 15:26:03 -08:00
|
|
|
const unsigned int instr_component_count = hlsl_type_component_count(instr->data_type);
|
2023-11-21 14:48:50 -08:00
|
|
|
const struct hlsl_deref *deref = &load->src;
|
2023-01-12 15:26:03 -08:00
|
|
|
const struct hlsl_ir_var *var = deref->var;
|
2022-11-11 16:39:55 -08:00
|
|
|
struct hlsl_constant_value values = {0};
|
2023-11-21 14:48:50 -08:00
|
|
|
unsigned int time = load->node.index;
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
unsigned int start, count, i;
|
2022-11-11 17:13:26 -08:00
|
|
|
struct hlsl_ir_node *cons;
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
|
2023-01-12 15:26:03 -08:00
|
|
|
if (!hlsl_component_index_range_from_deref(ctx, deref, &start, &count))
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
return false;
|
|
|
|
|
|
2023-01-12 15:26:03 -08:00
|
|
|
for (i = 0; i < instr_component_count; ++i)
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
{
|
2023-01-12 15:26:03 -08:00
|
|
|
struct copy_propagation_value *value;
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
if (!(value = copy_propagation_get_value(state, var, start + hlsl_swizzle_get_component(swizzle, i),
|
2023-11-21 14:48:50 -08:00
|
|
|
time)) || value->node->type != HLSL_IR_CONSTANT)
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
return false;
|
|
|
|
|
|
2022-11-11 16:39:55 -08:00
|
|
|
values.u[i] = hlsl_ir_constant(value->node)->value.u[value->component];
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
}
|
|
|
|
|
|
2022-11-11 17:10:14 -08:00
|
|
|
if (!(cons = hlsl_new_constant(ctx, instr->data_type, &values, &instr->loc)))
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
return false;
|
2022-11-11 17:13:26 -08:00
|
|
|
list_add_before(&instr->entry, &cons->entry);
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
|
2023-01-12 15:26:03 -08:00
|
|
|
TRACE("Load from %s[%u-%u]%s turned into a constant %p.\n",
|
|
|
|
|
var->name, start, start + count, debug_hlsl_swizzle(swizzle, instr_component_count), cons);
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
|
2022-11-11 17:13:26 -08:00
|
|
|
hlsl_replace_node(instr, cons);
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2022-01-21 13:22:29 -08:00
|
|
|
static bool copy_propagation_transform_load(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_ir_load *load, struct copy_propagation_state *state)
|
2021-12-01 08:14:50 -08:00
|
|
|
{
|
2023-01-12 14:28:44 -08:00
|
|
|
struct hlsl_type *type = load->node.data_type;
|
2021-12-01 08:14:50 -08:00
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
switch (type->class)
|
2022-01-21 13:22:26 -08:00
|
|
|
{
|
2024-06-10 04:05:13 -07:00
|
|
|
case HLSL_CLASS_DEPTH_STENCIL_STATE:
|
2022-01-21 13:22:26 -08:00
|
|
|
case HLSL_CLASS_SCALAR:
|
|
|
|
|
case HLSL_CLASS_VECTOR:
|
2024-02-06 17:41:15 -08:00
|
|
|
case HLSL_CLASS_PIXEL_SHADER:
|
2024-04-24 02:12:08 -07:00
|
|
|
case HLSL_CLASS_RASTERIZER_STATE:
|
2024-02-05 18:13:17 -08:00
|
|
|
case HLSL_CLASS_SAMPLER:
|
2024-07-17 12:51:08 -07:00
|
|
|
case HLSL_CLASS_STRING:
|
2024-02-05 18:25:57 -08:00
|
|
|
case HLSL_CLASS_TEXTURE:
|
2024-02-05 18:32:37 -08:00
|
|
|
case HLSL_CLASS_UAV:
|
2024-02-06 15:33:26 -08:00
|
|
|
case HLSL_CLASS_VERTEX_SHADER:
|
2024-08-05 09:41:23 -07:00
|
|
|
case HLSL_CLASS_COMPUTE_SHADER:
|
|
|
|
|
case HLSL_CLASS_DOMAIN_SHADER:
|
|
|
|
|
case HLSL_CLASS_HULL_SHADER:
|
2024-08-07 09:35:51 -07:00
|
|
|
case HLSL_CLASS_RENDER_TARGET_VIEW:
|
|
|
|
|
case HLSL_CLASS_DEPTH_STENCIL_VIEW:
|
2024-08-06 07:41:46 -07:00
|
|
|
case HLSL_CLASS_GEOMETRY_SHADER:
|
2024-08-07 03:49:04 -07:00
|
|
|
case HLSL_CLASS_BLEND_STATE:
|
2024-07-23 06:30:27 -07:00
|
|
|
case HLSL_CLASS_NULL:
|
2022-01-21 13:22:26 -08:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_MATRIX:
|
|
|
|
|
case HLSL_CLASS_ARRAY:
|
|
|
|
|
case HLSL_CLASS_STRUCT:
|
2024-09-20 11:57:53 -07:00
|
|
|
/* We can't handle complex types here.
|
|
|
|
|
* They should have been already split anyway by earlier passes,
|
|
|
|
|
* but they may not have been deleted yet. We can't rely on DCE to
|
|
|
|
|
* solve that problem for us, since we may be called on a partial
|
|
|
|
|
* block, but DCE deletes dead stores, so it needs to be able to
|
|
|
|
|
* see the whole program. */
|
vkd3d-shader/hlsl: Introduce the "error" type.
Currently, if an expression successfully parses according to the bison grammar,
but for one reason or another cannot generate a meaningful IR instruction, we
abort parsing with YYABORT. This includes, for example, an undefined variable or
function, invalid swizzle or field reference, or a constructor with a complex or
non-numeric data type.
Aborting parsing is unfortunate, however, because it means that any further
errors in the program cannot be caught by the programmer, increasing the number
of times they will need to fix errors and recompile.
The idea of this patch is that any such expression will instead generate an IR
node whose data type is of HLSL_CLASS_ERROR. Any further expression which would
consume an "error" typed instruction will instead immediately return an
expression of type "error" (probably the same one) instead of aborting or doing
any other type-checking.
Currently these "error" instructions should not pass the parsing stage, since
hlsl_compile_shader() will immediately notice that compilation has failed and
skip any optimization, lowering, or bytecode-writing.
A further direction to take this is to pre-allocate one "error" expression
immediately when creating the HLSL parser, and return that expression when we
fail to allocate an hlsl_ir_node of any type. This means we do not need to
handle allocation errors when constructing nodes, saving us quite a lot of error
handling (which is not only tedious but currently often broken, if nothing else
by virtue of neglecting cleanup of local variables).
2024-08-29 10:48:23 -07:00
|
|
|
case HLSL_CLASS_ERROR:
|
2022-01-21 13:22:26 -08:00
|
|
|
return false;
|
2024-04-09 14:42:00 -07:00
|
|
|
|
2024-09-20 10:44:56 -07:00
|
|
|
case HLSL_CLASS_CONSTANT_BUFFER:
|
2024-02-05 18:35:22 -08:00
|
|
|
case HLSL_CLASS_EFFECT_GROUP:
|
2024-02-06 15:08:01 -08:00
|
|
|
case HLSL_CLASS_PASS:
|
2024-02-06 15:15:48 -08:00
|
|
|
case HLSL_CLASS_TECHNIQUE:
|
2024-04-09 14:42:00 -07:00
|
|
|
case HLSL_CLASS_VOID:
|
|
|
|
|
vkd3d_unreachable();
|
2022-01-21 13:22:26 -08:00
|
|
|
}
|
2021-12-01 08:14:50 -08:00
|
|
|
|
2023-11-21 14:48:50 -08:00
|
|
|
if (copy_propagation_replace_with_constant_vector(ctx, state, load, HLSL_SWIZZLE(X, Y, Z, W), &load->node))
|
vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.
But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.
For instance, this shader:
```
sampler s;
Texture2D t;
float4 main() : sv_target
{
return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```
results in the following IR before applying the patch:
```
float | 6.00000024e-01
float | 6.00000024e-01
uint | 0
| = (<constructor-2>[@4].x @2)
uint | 1
| = (<constructor-2>[@6].x @3)
float2 | <constructor-2>
int | 0
int | 0
uint | 0
| = (<constructor-5>[@11].x @9)
uint | 1
| = (<constructor-5>[@13].x @10)
int2 | <constructor-5>
float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
| return
| = (<output-sv_target0> @16)
```
and this IR afterwards:
```
float2 | {6.00000024e-01 6.00000024e-01 }
int2 | {0 0 }
float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
| return
| = (<output-sv_target0> @4)
```
2022-11-17 12:49:28 -08:00
|
|
|
return true;
|
|
|
|
|
|
2023-11-21 14:48:50 -08:00
|
|
|
if (copy_propagation_replace_with_single_instr(ctx, state, load, HLSL_SWIZZLE(X, Y, Z, W), &load->node))
|
2023-01-12 15:26:03 -08:00
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool copy_propagation_transform_swizzle(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_ir_swizzle *swizzle, struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
|
|
|
|
|
if (swizzle->val.node->type != HLSL_IR_LOAD)
|
|
|
|
|
return false;
|
|
|
|
|
load = hlsl_ir_load(swizzle->val.node);
|
|
|
|
|
|
2023-11-21 14:48:50 -08:00
|
|
|
if (copy_propagation_replace_with_constant_vector(ctx, state, load, swizzle->swizzle, &swizzle->node))
|
2023-01-12 15:26:03 -08:00
|
|
|
return true;
|
|
|
|
|
|
2023-11-21 14:48:50 -08:00
|
|
|
if (copy_propagation_replace_with_single_instr(ctx, state, load, swizzle->swizzle, &swizzle->node))
|
2023-01-12 14:28:44 -08:00
|
|
|
return true;
|
2021-12-01 08:14:50 -08:00
|
|
|
|
2023-01-12 14:28:44 -08:00
|
|
|
return false;
|
2021-12-01 08:14:50 -08:00
|
|
|
}
|
|
|
|
|
|
2022-01-21 13:22:29 -08:00
|
|
|
static bool copy_propagation_transform_object_load(struct hlsl_ctx *ctx,
|
2023-11-21 14:09:32 -08:00
|
|
|
struct hlsl_deref *deref, struct copy_propagation_state *state, unsigned int time)
|
2022-01-21 13:22:29 -08:00
|
|
|
{
|
2023-01-12 15:28:44 -08:00
|
|
|
struct copy_propagation_value *value;
|
2022-01-21 13:22:29 -08:00
|
|
|
struct hlsl_ir_load *load;
|
2023-01-12 15:28:44 -08:00
|
|
|
unsigned int start, count;
|
|
|
|
|
|
|
|
|
|
if (!hlsl_component_index_range_from_deref(ctx, deref, &start, &count))
|
|
|
|
|
return false;
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(count == 1);
|
2022-01-21 13:22:29 -08:00
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
if (!(value = copy_propagation_get_value(state, deref->var, start, time)))
|
2022-01-21 13:22:29 -08:00
|
|
|
return false;
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(value->component == 0);
|
2022-01-21 13:22:29 -08:00
|
|
|
|
|
|
|
|
/* Only HLSL_IR_LOAD can produce an object. */
|
2023-01-12 15:28:44 -08:00
|
|
|
load = hlsl_ir_load(value->node);
|
2022-07-20 12:37:07 -07:00
|
|
|
|
2023-01-05 10:32:11 -08:00
|
|
|
/* As we are replacing the instruction's deref (with the one in the hlsl_ir_load) and not the
|
|
|
|
|
* instruction itself, we won't be able to rely on the value retrieved by
|
|
|
|
|
* copy_propagation_get_value() for the new deref in subsequent iterations of copy propagation.
|
|
|
|
|
* This is because another value may be written to that deref between the hlsl_ir_load and
|
|
|
|
|
* this instruction.
|
|
|
|
|
*
|
|
|
|
|
* For this reason, we only replace the new deref when it corresponds to a uniform variable,
|
|
|
|
|
* which cannot be written to.
|
|
|
|
|
*
|
|
|
|
|
* In a valid shader, all object references must resolve statically to a single uniform object.
|
|
|
|
|
* If this is the case, we can expect copy propagation on regular store/loads and the other
|
|
|
|
|
* compilation passes to replace all hlsl_ir_loads with loads to uniform objects, so this
|
|
|
|
|
* implementation is complete, even with this restriction.
|
|
|
|
|
*/
|
|
|
|
|
if (!load->src.var->is_uniform)
|
|
|
|
|
{
|
|
|
|
|
TRACE("Ignoring load from non-uniform object variable %s\n", load->src.var->name);
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2022-07-20 12:37:07 -07:00
|
|
|
hlsl_cleanup_deref(deref);
|
|
|
|
|
hlsl_copy_deref(ctx, deref, &load->src);
|
|
|
|
|
|
2022-01-21 13:22:29 -08:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool copy_propagation_transform_resource_load(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_ir_resource_load *load, struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
bool progress = false;
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
progress |= copy_propagation_transform_object_load(ctx, &load->resource, state, load->node.index);
|
2022-01-21 13:22:29 -08:00
|
|
|
if (load->sampler.var)
|
2023-11-21 14:09:32 -08:00
|
|
|
progress |= copy_propagation_transform_object_load(ctx, &load->sampler, state, load->node.index);
|
2022-01-21 13:22:29 -08:00
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2022-04-05 16:26:22 -07:00
|
|
|
static bool copy_propagation_transform_resource_store(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_ir_resource_store *store, struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
bool progress = false;
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
progress |= copy_propagation_transform_object_load(ctx, &store->resource, state, store->node.index);
|
2022-04-05 16:26:22 -07:00
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
static void copy_propagation_record_store(struct hlsl_ctx *ctx, struct hlsl_ir_store *store,
|
|
|
|
|
struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_var_def *var_def;
|
|
|
|
|
struct hlsl_deref *lhs = &store->lhs;
|
|
|
|
|
struct hlsl_ir_var *var = lhs->var;
|
2022-07-20 12:37:07 -07:00
|
|
|
unsigned int start, count;
|
2021-12-01 08:14:50 -08:00
|
|
|
|
|
|
|
|
if (!(var_def = copy_propagation_create_var_def(ctx, state, var)))
|
|
|
|
|
return;
|
|
|
|
|
|
2022-07-20 12:37:07 -07:00
|
|
|
if (hlsl_component_index_range_from_deref(ctx, lhs, &start, &count))
|
2022-01-21 13:22:26 -08:00
|
|
|
{
|
|
|
|
|
unsigned int writemask = store->writemask;
|
|
|
|
|
|
2024-01-26 15:29:54 -08:00
|
|
|
if (!hlsl_is_numeric_type(store->rhs.node->data_type))
|
2022-01-21 13:22:26 -08:00
|
|
|
writemask = VKD3DSP_WRITEMASK_0;
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_set_value(ctx, var_def, start, writemask, store->rhs.node, store->node.index);
|
2022-01-21 13:22:26 -08:00
|
|
|
}
|
2021-12-01 08:14:50 -08:00
|
|
|
else
|
2022-01-21 13:22:26 -08:00
|
|
|
{
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_variable_from_deref(ctx, var_def, lhs, store->writemask,
|
|
|
|
|
store->node.index);
|
2022-01-21 13:22:26 -08:00
|
|
|
}
|
2021-12-01 08:14:50 -08:00
|
|
|
}
|
|
|
|
|
|
2022-04-28 06:32:05 -07:00
|
|
|
static void copy_propagation_state_init(struct hlsl_ctx *ctx, struct copy_propagation_state *state,
|
|
|
|
|
struct copy_propagation_state *parent)
|
|
|
|
|
{
|
|
|
|
|
rb_init(&state->var_defs, copy_propagation_var_def_compare);
|
|
|
|
|
state->parent = parent;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void copy_propagation_state_destroy(struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
rb_destroy(&state->var_defs, copy_propagation_var_def_destroy, NULL);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void copy_propagation_invalidate_from_block(struct hlsl_ctx *ctx, struct copy_propagation_state *state,
|
2023-11-21 14:09:32 -08:00
|
|
|
struct hlsl_block *block, unsigned int time)
|
2022-04-28 06:32:05 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(instr, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_IR_STORE:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_store *store = hlsl_ir_store(instr);
|
|
|
|
|
struct copy_propagation_var_def *var_def;
|
|
|
|
|
struct hlsl_deref *lhs = &store->lhs;
|
|
|
|
|
struct hlsl_ir_var *var = lhs->var;
|
|
|
|
|
|
|
|
|
|
if (!(var_def = copy_propagation_create_var_def(ctx, state, var)))
|
|
|
|
|
continue;
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_variable_from_deref(ctx, var_def, lhs, store->writemask, time);
|
2022-04-28 06:32:05 -07:00
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_IF:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &iff->then_block, time);
|
|
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &iff->else_block, time);
|
2022-04-28 06:32:05 -07:00
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_LOOP:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_loop *loop = hlsl_ir_loop(instr);
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &loop->body, time);
|
2022-04-28 06:32:05 -07:00
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-11 05:08:59 -07:00
|
|
|
case HLSL_IR_SWITCH:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &c->body, time);
|
2023-10-11 05:08:59 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2022-04-28 06:32:05 -07:00
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool copy_propagation_transform_block(struct hlsl_ctx *ctx, struct hlsl_block *block,
|
|
|
|
|
struct copy_propagation_state *state);
|
|
|
|
|
|
|
|
|
|
static bool copy_propagation_process_if(struct hlsl_ctx *ctx, struct hlsl_ir_if *iff,
|
|
|
|
|
struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_state inner_state;
|
|
|
|
|
bool progress = false;
|
|
|
|
|
|
|
|
|
|
copy_propagation_state_init(ctx, &inner_state, state);
|
2022-11-10 18:04:22 -08:00
|
|
|
progress |= copy_propagation_transform_block(ctx, &iff->then_block, &inner_state);
|
2022-04-28 06:32:05 -07:00
|
|
|
copy_propagation_state_destroy(&inner_state);
|
|
|
|
|
|
|
|
|
|
copy_propagation_state_init(ctx, &inner_state, state);
|
2022-11-10 18:04:22 -08:00
|
|
|
progress |= copy_propagation_transform_block(ctx, &iff->else_block, &inner_state);
|
2022-04-28 06:32:05 -07:00
|
|
|
copy_propagation_state_destroy(&inner_state);
|
|
|
|
|
|
|
|
|
|
/* Ideally we'd invalidate the outer state looking at what was
|
|
|
|
|
* touched in the two inner states, but this doesn't work for
|
|
|
|
|
* loops (because we need to know what is invalidated in advance),
|
|
|
|
|
* so we need copy_propagation_invalidate_from_block() anyway. */
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &iff->then_block, iff->node.index);
|
|
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &iff->else_block, iff->node.index);
|
2022-04-28 06:32:05 -07:00
|
|
|
|
|
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2022-05-03 02:57:21 -07:00
|
|
|
static bool copy_propagation_process_loop(struct hlsl_ctx *ctx, struct hlsl_ir_loop *loop,
|
|
|
|
|
struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_state inner_state;
|
|
|
|
|
bool progress = false;
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &loop->body, loop->node.index);
|
2022-05-03 02:57:21 -07:00
|
|
|
|
|
|
|
|
copy_propagation_state_init(ctx, &inner_state, state);
|
|
|
|
|
progress |= copy_propagation_transform_block(ctx, &loop->body, &inner_state);
|
|
|
|
|
copy_propagation_state_destroy(&inner_state);
|
|
|
|
|
|
|
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-11 05:08:59 -07:00
|
|
|
static bool copy_propagation_process_switch(struct hlsl_ctx *ctx, struct hlsl_ir_switch *s,
|
|
|
|
|
struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
struct copy_propagation_state inner_state;
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
bool progress = false;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
copy_propagation_state_init(ctx, &inner_state, state);
|
|
|
|
|
progress |= copy_propagation_transform_block(ctx, &c->body, &inner_state);
|
|
|
|
|
copy_propagation_state_destroy(&inner_state);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
2023-11-21 14:09:32 -08:00
|
|
|
copy_propagation_invalidate_from_block(ctx, state, &c->body, s->node.index);
|
2023-10-11 05:08:59 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
static bool copy_propagation_transform_block(struct hlsl_ctx *ctx, struct hlsl_block *block,
|
|
|
|
|
struct copy_propagation_state *state)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr, *next;
|
|
|
|
|
bool progress = false;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY_SAFE(instr, next, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_IR_LOAD:
|
2022-01-21 13:22:29 -08:00
|
|
|
progress |= copy_propagation_transform_load(ctx, hlsl_ir_load(instr), state);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_RESOURCE_LOAD:
|
|
|
|
|
progress |= copy_propagation_transform_resource_load(ctx, hlsl_ir_resource_load(instr), state);
|
2021-12-01 08:14:50 -08:00
|
|
|
break;
|
|
|
|
|
|
2022-04-05 16:26:22 -07:00
|
|
|
case HLSL_IR_RESOURCE_STORE:
|
|
|
|
|
progress |= copy_propagation_transform_resource_store(ctx, hlsl_ir_resource_store(instr), state);
|
|
|
|
|
break;
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
case HLSL_IR_STORE:
|
|
|
|
|
copy_propagation_record_store(ctx, hlsl_ir_store(instr), state);
|
|
|
|
|
break;
|
|
|
|
|
|
2023-01-12 15:26:03 -08:00
|
|
|
case HLSL_IR_SWIZZLE:
|
|
|
|
|
progress |= copy_propagation_transform_swizzle(ctx, hlsl_ir_swizzle(instr), state);
|
|
|
|
|
break;
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
case HLSL_IR_IF:
|
2022-04-28 06:32:05 -07:00
|
|
|
progress |= copy_propagation_process_if(ctx, hlsl_ir_if(instr), state);
|
|
|
|
|
break;
|
2021-12-01 08:14:50 -08:00
|
|
|
|
|
|
|
|
case HLSL_IR_LOOP:
|
2022-05-03 02:57:21 -07:00
|
|
|
progress |= copy_propagation_process_loop(ctx, hlsl_ir_loop(instr), state);
|
|
|
|
|
break;
|
2021-12-01 08:14:50 -08:00
|
|
|
|
2023-10-11 05:08:59 -07:00
|
|
|
case HLSL_IR_SWITCH:
|
|
|
|
|
progress |= copy_propagation_process_switch(ctx, hlsl_ir_switch(instr), state);
|
|
|
|
|
break;
|
|
|
|
|
|
2021-12-01 08:14:50 -08:00
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-09 06:39:38 -07:00
|
|
|
bool hlsl_copy_propagation_execute(struct hlsl_ctx *ctx, struct hlsl_block *block)
|
2021-12-01 08:14:50 -08:00
|
|
|
{
|
|
|
|
|
struct copy_propagation_state state;
|
|
|
|
|
bool progress;
|
|
|
|
|
|
2023-11-21 14:09:32 -08:00
|
|
|
index_instructions(block, 2);
|
|
|
|
|
|
2022-04-28 06:32:05 -07:00
|
|
|
copy_propagation_state_init(ctx, &state, NULL);
|
2021-12-01 08:14:50 -08:00
|
|
|
|
|
|
|
|
progress = copy_propagation_transform_block(ctx, block, &state);
|
|
|
|
|
|
2022-04-28 06:32:05 -07:00
|
|
|
copy_propagation_state_destroy(&state);
|
2021-12-01 08:14:50 -08:00
|
|
|
|
|
|
|
|
return progress;
|
|
|
|
|
}
|
|
|
|
|
|
2024-03-12 05:40:30 -07:00
|
|
|
enum validation_result
|
|
|
|
|
{
|
|
|
|
|
DEREF_VALIDATION_OK,
|
|
|
|
|
DEREF_VALIDATION_OUT_OF_BOUNDS,
|
|
|
|
|
DEREF_VALIDATION_NOT_CONSTANT,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static enum validation_result validate_component_index_range_from_deref(struct hlsl_ctx *ctx,
|
|
|
|
|
const struct hlsl_deref *deref)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_type *type = deref->var->data_type;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < deref->path_len; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *path_node = deref->path[i].node;
|
|
|
|
|
unsigned int idx = 0;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(path_node);
|
2024-03-12 05:40:30 -07:00
|
|
|
if (path_node->type != HLSL_IR_CONSTANT)
|
|
|
|
|
return DEREF_VALIDATION_NOT_CONSTANT;
|
|
|
|
|
|
|
|
|
|
/* We should always have generated a cast to UINT. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(path_node->data_type->class == HLSL_CLASS_SCALAR
|
2024-03-12 05:40:30 -07:00
|
|
|
&& path_node->data_type->e.numeric.type == HLSL_TYPE_UINT);
|
|
|
|
|
|
|
|
|
|
idx = hlsl_ir_constant(path_node)->value.u[0].u;
|
|
|
|
|
|
|
|
|
|
switch (type->class)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_CLASS_VECTOR:
|
|
|
|
|
if (idx >= type->dimx)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &path_node->loc, VKD3D_SHADER_ERROR_HLSL_OFFSET_OUT_OF_BOUNDS,
|
|
|
|
|
"Vector index is out of bounds. %u/%u", idx, type->dimx);
|
|
|
|
|
return DEREF_VALIDATION_OUT_OF_BOUNDS;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_MATRIX:
|
|
|
|
|
if (idx >= hlsl_type_major_size(type))
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &path_node->loc, VKD3D_SHADER_ERROR_HLSL_OFFSET_OUT_OF_BOUNDS,
|
|
|
|
|
"Matrix index is out of bounds. %u/%u", idx, hlsl_type_major_size(type));
|
|
|
|
|
return DEREF_VALIDATION_OUT_OF_BOUNDS;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_ARRAY:
|
|
|
|
|
if (idx >= type->e.array.elements_count)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &path_node->loc, VKD3D_SHADER_ERROR_HLSL_OFFSET_OUT_OF_BOUNDS,
|
|
|
|
|
"Array index is out of bounds. %u/%u", idx, type->e.array.elements_count);
|
|
|
|
|
return DEREF_VALIDATION_OUT_OF_BOUNDS;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_STRUCT:
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
type = hlsl_get_element_type_from_path_index(ctx, type, path_node);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return DEREF_VALIDATION_OK;
|
|
|
|
|
}
|
|
|
|
|
|
2022-07-22 08:40:24 -07:00
|
|
|
static void note_non_static_deref_expressions(struct hlsl_ctx *ctx, const struct hlsl_deref *deref,
|
|
|
|
|
const char *usage)
|
|
|
|
|
{
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < deref->path_len; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *path_node = deref->path[i].node;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(path_node);
|
2022-07-22 08:40:24 -07:00
|
|
|
if (path_node->type != HLSL_IR_CONSTANT)
|
|
|
|
|
hlsl_note(ctx, &path_node->loc, VKD3D_SHADER_LOG_ERROR,
|
|
|
|
|
"Expression for %s within \"%s\" cannot be resolved statically.",
|
|
|
|
|
usage, deref->var->name);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-03-12 05:40:30 -07:00
|
|
|
static bool validate_dereferences(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr,
|
2022-07-22 08:40:24 -07:00
|
|
|
void *context)
|
|
|
|
|
{
|
2024-05-17 15:03:34 -07:00
|
|
|
switch (instr->type)
|
2022-07-22 08:40:24 -07:00
|
|
|
{
|
2024-05-17 15:03:34 -07:00
|
|
|
case HLSL_IR_RESOURCE_LOAD:
|
2022-07-22 08:40:24 -07:00
|
|
|
{
|
2024-05-17 15:03:34 -07:00
|
|
|
struct hlsl_ir_resource_load *load = hlsl_ir_resource_load(instr);
|
2022-12-05 16:48:28 -08:00
|
|
|
|
2024-05-17 15:03:34 -07:00
|
|
|
if (!load->resource.var->is_uniform)
|
2022-12-05 16:48:28 -08:00
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_NON_STATIC_OBJECT_REF,
|
2024-05-17 15:03:34 -07:00
|
|
|
"Loaded resource must have a single uniform source.");
|
2022-12-05 16:48:28 -08:00
|
|
|
}
|
2024-03-12 05:40:30 -07:00
|
|
|
else if (validate_component_index_range_from_deref(ctx, &load->resource) == DEREF_VALIDATION_NOT_CONSTANT)
|
2022-12-05 16:48:28 -08:00
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_NON_STATIC_OBJECT_REF,
|
2024-05-17 15:03:34 -07:00
|
|
|
"Loaded resource from \"%s\" must be determinable at compile time.",
|
|
|
|
|
load->resource.var->name);
|
|
|
|
|
note_non_static_deref_expressions(ctx, &load->resource, "loaded resource");
|
2022-12-05 16:48:28 -08:00
|
|
|
}
|
2022-11-09 11:41:50 -08:00
|
|
|
|
2024-05-17 15:03:34 -07:00
|
|
|
if (load->sampler.var)
|
|
|
|
|
{
|
|
|
|
|
if (!load->sampler.var->is_uniform)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_NON_STATIC_OBJECT_REF,
|
|
|
|
|
"Resource load sampler must have a single uniform source.");
|
|
|
|
|
}
|
2024-03-12 05:40:30 -07:00
|
|
|
else if (validate_component_index_range_from_deref(ctx, &load->sampler) == DEREF_VALIDATION_NOT_CONSTANT)
|
2024-05-17 15:03:34 -07:00
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_NON_STATIC_OBJECT_REF,
|
|
|
|
|
"Resource load sampler from \"%s\" must be determinable at compile time.",
|
|
|
|
|
load->sampler.var->name);
|
|
|
|
|
note_non_static_deref_expressions(ctx, &load->sampler, "resource load sampler");
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
break;
|
2022-12-05 16:48:28 -08:00
|
|
|
}
|
2024-05-17 15:03:34 -07:00
|
|
|
case HLSL_IR_RESOURCE_STORE:
|
2022-11-09 11:41:50 -08:00
|
|
|
{
|
2024-05-17 15:03:34 -07:00
|
|
|
struct hlsl_ir_resource_store *store = hlsl_ir_resource_store(instr);
|
|
|
|
|
|
|
|
|
|
if (!store->resource.var->is_uniform)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_NON_STATIC_OBJECT_REF,
|
|
|
|
|
"Accessed resource must have a single uniform source.");
|
|
|
|
|
}
|
2024-03-12 05:40:30 -07:00
|
|
|
else if (validate_component_index_range_from_deref(ctx, &store->resource) == DEREF_VALIDATION_NOT_CONSTANT)
|
2024-05-17 15:03:34 -07:00
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_NON_STATIC_OBJECT_REF,
|
|
|
|
|
"Accessed resource from \"%s\" must be determinable at compile time.",
|
|
|
|
|
store->resource.var->name);
|
|
|
|
|
note_non_static_deref_expressions(ctx, &store->resource, "accessed resource");
|
|
|
|
|
}
|
|
|
|
|
break;
|
2022-11-09 11:41:50 -08:00
|
|
|
}
|
2024-03-12 05:40:30 -07:00
|
|
|
case HLSL_IR_LOAD:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_load *load = hlsl_ir_load(instr);
|
|
|
|
|
validate_component_index_range_from_deref(ctx, &load->src);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case HLSL_IR_STORE:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_store *store = hlsl_ir_store(instr);
|
|
|
|
|
validate_component_index_range_from_deref(ctx, &store->lhs);
|
|
|
|
|
break;
|
|
|
|
|
}
|
2024-05-17 15:03:34 -07:00
|
|
|
default:
|
|
|
|
|
break;
|
2022-11-09 11:41:50 -08:00
|
|
|
}
|
2022-07-22 08:40:24 -07:00
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2021-03-16 14:31:56 -07:00
|
|
|
static bool is_vec1(const struct hlsl_type *type)
|
|
|
|
|
{
|
2022-11-11 17:31:55 -08:00
|
|
|
return (type->class == HLSL_CLASS_SCALAR) || (type->class == HLSL_CLASS_VECTOR && type->dimx == 1);
|
2021-03-16 14:31:56 -07:00
|
|
|
}
|
|
|
|
|
|
2021-03-16 14:31:55 -07:00
|
|
|
static bool fold_redundant_casts(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
if (instr->type == HLSL_IR_EXPR)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_expr *expr = hlsl_ir_expr(instr);
|
2021-03-16 14:31:56 -07:00
|
|
|
const struct hlsl_type *dst_type = expr->node.data_type;
|
2021-09-10 14:35:54 -07:00
|
|
|
const struct hlsl_type *src_type;
|
2021-03-16 14:31:56 -07:00
|
|
|
|
2021-08-12 17:36:13 -07:00
|
|
|
if (expr->op != HLSL_OP1_CAST)
|
2021-03-16 14:31:56 -07:00
|
|
|
return false;
|
2021-03-16 14:31:55 -07:00
|
|
|
|
2021-09-10 14:35:54 -07:00
|
|
|
src_type = expr->operands[0].node->data_type;
|
|
|
|
|
|
2021-03-17 22:22:19 -07:00
|
|
|
if (hlsl_types_are_equal(src_type, dst_type)
|
2024-02-27 15:30:51 -08:00
|
|
|
|| (src_type->e.numeric.type == dst_type->e.numeric.type && is_vec1(src_type) && is_vec1(dst_type)))
|
2021-03-16 14:31:55 -07:00
|
|
|
{
|
2022-02-10 19:48:18 -08:00
|
|
|
hlsl_replace_node(&expr->node, expr->operands[0].node);
|
2021-03-16 14:31:55 -07:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2022-04-22 03:25:01 -07:00
|
|
|
/* Copy an element of a complex variable. Helper for
|
|
|
|
|
* split_array_copies(), split_struct_copies() and
|
|
|
|
|
* split_matrix_copies(). Inserts new instructions right before
|
|
|
|
|
* "store". */
|
2021-08-12 17:36:14 -07:00
|
|
|
static bool split_copy(struct hlsl_ctx *ctx, struct hlsl_ir_store *store,
|
2022-06-28 14:20:24 -07:00
|
|
|
const struct hlsl_ir_load *load, const unsigned int idx, struct hlsl_type *type)
|
2021-08-12 17:36:14 -07:00
|
|
|
{
|
2022-11-10 19:06:04 -08:00
|
|
|
struct hlsl_ir_node *split_store, *c;
|
2021-08-12 17:36:14 -07:00
|
|
|
struct hlsl_ir_load *split_load;
|
|
|
|
|
|
2022-06-28 14:20:24 -07:00
|
|
|
if (!(c = hlsl_new_uint_constant(ctx, idx, &store->node.loc)))
|
2021-08-12 17:36:14 -07:00
|
|
|
return false;
|
2022-11-10 19:06:04 -08:00
|
|
|
list_add_before(&store->node.entry, &c->entry);
|
2021-08-12 17:36:14 -07:00
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(split_load = hlsl_new_load_index(ctx, &load->src, c, &store->node.loc)))
|
2021-08-12 17:36:14 -07:00
|
|
|
return false;
|
|
|
|
|
list_add_before(&store->node.entry, &split_load->node.entry);
|
|
|
|
|
|
2022-11-10 19:06:04 -08:00
|
|
|
if (!(split_store = hlsl_new_store_index(ctx, &store->lhs, c, &split_load->node, 0, &store->node.loc)))
|
2021-08-12 17:36:14 -07:00
|
|
|
return false;
|
2022-11-10 18:57:00 -08:00
|
|
|
list_add_before(&store->node.entry, &split_store->entry);
|
2021-08-12 17:36:14 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool split_array_copies(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
2021-03-17 22:22:21 -07:00
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_node *rhs;
|
2021-08-12 17:36:14 -07:00
|
|
|
struct hlsl_type *element_type;
|
2021-03-17 22:22:21 -07:00
|
|
|
const struct hlsl_type *type;
|
2021-04-08 21:38:22 -07:00
|
|
|
struct hlsl_ir_store *store;
|
2022-06-28 14:20:24 -07:00
|
|
|
unsigned int i;
|
2021-03-17 22:22:21 -07:00
|
|
|
|
2021-04-08 21:38:22 -07:00
|
|
|
if (instr->type != HLSL_IR_STORE)
|
2021-03-17 22:22:21 -07:00
|
|
|
return false;
|
|
|
|
|
|
2021-04-08 21:38:22 -07:00
|
|
|
store = hlsl_ir_store(instr);
|
|
|
|
|
rhs = store->rhs.node;
|
2021-03-17 22:22:21 -07:00
|
|
|
type = rhs->data_type;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_ARRAY)
|
2021-03-17 22:22:21 -07:00
|
|
|
return false;
|
2021-08-12 17:36:14 -07:00
|
|
|
element_type = type->e.array.type;
|
2021-03-17 22:22:21 -07:00
|
|
|
|
2022-04-28 06:31:58 -07:00
|
|
|
if (rhs->type != HLSL_IR_LOAD)
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Array store rhs is not HLSL_IR_LOAD. Broadcast may be missing.");
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2021-08-12 17:36:14 -07:00
|
|
|
for (i = 0; i < type->e.array.elements_count; ++i)
|
2021-03-17 22:22:21 -07:00
|
|
|
{
|
2022-06-28 14:20:24 -07:00
|
|
|
if (!split_copy(ctx, store, hlsl_ir_load(rhs), i, element_type))
|
2021-03-17 22:22:21 -07:00
|
|
|
return false;
|
2021-08-12 17:36:14 -07:00
|
|
|
}
|
2021-03-17 22:22:21 -07:00
|
|
|
|
2021-08-12 17:36:14 -07:00
|
|
|
/* Remove the store instruction, so that we can split structs which contain
|
|
|
|
|
* other structs. Although assignments produce a value, we don't allow
|
|
|
|
|
* HLSL_IR_STORE to be used as a source. */
|
|
|
|
|
list_remove(&store->node.entry);
|
|
|
|
|
hlsl_free_instr(&store->node);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
2021-03-17 22:22:21 -07:00
|
|
|
|
2021-08-12 17:36:14 -07:00
|
|
|
static bool split_struct_copies(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_node *rhs;
|
|
|
|
|
const struct hlsl_type *type;
|
|
|
|
|
struct hlsl_ir_store *store;
|
2022-07-14 18:23:43 -07:00
|
|
|
size_t i;
|
2021-03-17 22:22:21 -07:00
|
|
|
|
2021-08-12 17:36:14 -07:00
|
|
|
if (instr->type != HLSL_IR_STORE)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
store = hlsl_ir_store(instr);
|
|
|
|
|
rhs = store->rhs.node;
|
|
|
|
|
type = rhs->data_type;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_STRUCT)
|
2021-08-12 17:36:14 -07:00
|
|
|
return false;
|
|
|
|
|
|
2022-04-28 06:31:58 -07:00
|
|
|
if (rhs->type != HLSL_IR_LOAD)
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Struct store rhs is not HLSL_IR_LOAD. Broadcast may be missing.");
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2022-07-14 18:23:43 -07:00
|
|
|
for (i = 0; i < type->e.record.field_count; ++i)
|
2021-08-12 17:36:14 -07:00
|
|
|
{
|
2022-07-14 18:23:43 -07:00
|
|
|
const struct hlsl_struct_field *field = &type->e.record.fields[i];
|
|
|
|
|
|
2022-06-28 14:20:24 -07:00
|
|
|
if (!split_copy(ctx, store, hlsl_ir_load(rhs), i, field->type))
|
2021-03-17 22:22:21 -07:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-08 21:38:22 -07:00
|
|
|
/* Remove the store instruction, so that we can split structs which contain
|
|
|
|
|
* other structs. Although assignments produce a value, we don't allow
|
|
|
|
|
* HLSL_IR_STORE to be used as a source. */
|
|
|
|
|
list_remove(&store->node.entry);
|
|
|
|
|
hlsl_free_instr(&store->node);
|
2021-03-17 22:22:21 -07:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2022-04-17 23:34:02 -07:00
|
|
|
static bool split_matrix_copies(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_node *rhs;
|
|
|
|
|
struct hlsl_type *element_type;
|
|
|
|
|
const struct hlsl_type *type;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
struct hlsl_ir_store *store;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_STORE)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
store = hlsl_ir_store(instr);
|
|
|
|
|
rhs = store->rhs.node;
|
|
|
|
|
type = rhs->data_type;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_MATRIX)
|
2022-04-17 23:34:02 -07:00
|
|
|
return false;
|
2024-02-27 15:30:51 -08:00
|
|
|
element_type = hlsl_get_vector_type(ctx, type->e.numeric.type, hlsl_type_minor_size(type));
|
2022-04-17 23:34:02 -07:00
|
|
|
|
|
|
|
|
if (rhs->type != HLSL_IR_LOAD)
|
|
|
|
|
{
|
2023-06-23 13:22:19 -07:00
|
|
|
hlsl_fixme(ctx, &instr->loc, "Copying from unsupported node type.");
|
2022-04-17 23:34:02 -07:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2022-06-30 13:25:12 -07:00
|
|
|
for (i = 0; i < hlsl_type_major_size(type); ++i)
|
2022-04-17 23:34:02 -07:00
|
|
|
{
|
2022-06-28 14:20:24 -07:00
|
|
|
if (!split_copy(ctx, store, hlsl_ir_load(rhs), i, element_type))
|
2022-04-17 23:34:02 -07:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
list_remove(&store->node.entry);
|
|
|
|
|
hlsl_free_instr(&store->node);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_narrowing_casts(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2021-12-01 08:14:56 -08:00
|
|
|
{
|
|
|
|
|
const struct hlsl_type *src_type, *dst_type;
|
|
|
|
|
struct hlsl_type *dst_vector_type;
|
|
|
|
|
struct hlsl_ir_expr *cast;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
cast = hlsl_ir_expr(instr);
|
2021-09-10 14:35:54 -07:00
|
|
|
if (cast->op != HLSL_OP1_CAST)
|
|
|
|
|
return false;
|
2021-12-01 08:14:56 -08:00
|
|
|
src_type = cast->operands[0].node->data_type;
|
|
|
|
|
dst_type = cast->node.data_type;
|
|
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
if (src_type->class <= HLSL_CLASS_VECTOR && dst_type->class <= HLSL_CLASS_VECTOR && dst_type->dimx < src_type->dimx)
|
2021-12-01 08:14:56 -08:00
|
|
|
{
|
2022-11-10 19:01:18 -08:00
|
|
|
struct hlsl_ir_node *new_cast, *swizzle;
|
2021-12-01 08:14:56 -08:00
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
dst_vector_type = hlsl_get_vector_type(ctx, dst_type->e.numeric.type, src_type->dimx);
|
2021-12-01 08:14:56 -08:00
|
|
|
/* We need to preserve the cast since it might be doing more than just
|
|
|
|
|
* narrowing the vector. */
|
|
|
|
|
if (!(new_cast = hlsl_new_cast(ctx, cast->operands[0].node, dst_vector_type, &cast->node.loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, new_cast);
|
|
|
|
|
|
2022-11-10 17:39:42 -08:00
|
|
|
if (!(swizzle = hlsl_new_swizzle(ctx, HLSL_SWIZZLE(X, Y, Z, W), dst_type->dimx, new_cast, &cast->node.loc)))
|
2021-12-01 08:14:56 -08:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, swizzle);
|
2021-12-01 08:14:56 -08:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2022-12-09 14:23:41 -08:00
|
|
|
static bool fold_swizzle_chains(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_swizzle *swizzle;
|
|
|
|
|
struct hlsl_ir_node *next_instr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_SWIZZLE)
|
|
|
|
|
return false;
|
|
|
|
|
swizzle = hlsl_ir_swizzle(instr);
|
|
|
|
|
|
|
|
|
|
next_instr = swizzle->val.node;
|
|
|
|
|
|
|
|
|
|
if (next_instr->type == HLSL_IR_SWIZZLE)
|
|
|
|
|
{
|
2022-11-10 19:01:18 -08:00
|
|
|
struct hlsl_ir_node *new_swizzle;
|
2023-12-06 09:20:25 -08:00
|
|
|
uint32_t combined_swizzle;
|
2022-12-09 14:23:41 -08:00
|
|
|
|
|
|
|
|
combined_swizzle = hlsl_combine_swizzles(hlsl_ir_swizzle(next_instr)->swizzle,
|
|
|
|
|
swizzle->swizzle, instr->data_type->dimx);
|
|
|
|
|
next_instr = hlsl_ir_swizzle(next_instr)->val.node;
|
|
|
|
|
|
|
|
|
|
if (!(new_swizzle = hlsl_new_swizzle(ctx, combined_swizzle, instr->data_type->dimx, next_instr, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
|
2022-11-10 19:01:18 -08:00
|
|
|
list_add_before(&instr->entry, &new_swizzle->entry);
|
|
|
|
|
hlsl_replace_node(instr, new_swizzle);
|
2022-12-09 14:23:41 -08:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2021-11-17 00:47:24 -08:00
|
|
|
static bool remove_trivial_swizzles(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_swizzle *swizzle;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_SWIZZLE)
|
|
|
|
|
return false;
|
|
|
|
|
swizzle = hlsl_ir_swizzle(instr);
|
|
|
|
|
|
|
|
|
|
if (instr->data_type->dimx != swizzle->val.node->data_type->dimx)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < instr->data_type->dimx; ++i)
|
2023-01-12 13:52:49 -08:00
|
|
|
if (hlsl_swizzle_get_component(swizzle->swizzle, i) != i)
|
2021-11-17 00:47:24 -08:00
|
|
|
return false;
|
|
|
|
|
|
2022-02-10 19:48:18 -08:00
|
|
|
hlsl_replace_node(instr, swizzle->val.node);
|
2021-11-17 00:47:24 -08:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-09-27 10:17:40 -07:00
|
|
|
static bool remove_trivial_conditional_branches(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_constant *condition;
|
|
|
|
|
struct hlsl_ir_if *iff;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_IF)
|
|
|
|
|
return false;
|
|
|
|
|
iff = hlsl_ir_if(instr);
|
|
|
|
|
if (iff->condition.node->type != HLSL_IR_CONSTANT)
|
|
|
|
|
return false;
|
|
|
|
|
condition = hlsl_ir_constant(iff->condition.node);
|
|
|
|
|
|
|
|
|
|
list_move_before(&instr->entry, condition->value.u[0].u ? &iff->then_block.instrs : &iff->else_block.instrs);
|
|
|
|
|
list_remove(&instr->entry);
|
|
|
|
|
hlsl_free_instr(instr);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-11 05:39:01 -07:00
|
|
|
static bool normalize_switch_cases(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch_case *c, *def = NULL;
|
|
|
|
|
bool missing_terminal_break = false;
|
|
|
|
|
struct hlsl_ir_node *node;
|
|
|
|
|
struct hlsl_ir_switch *s;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_SWITCH)
|
|
|
|
|
return false;
|
|
|
|
|
s = hlsl_ir_switch(instr);
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
bool terminal_break = false;
|
|
|
|
|
|
|
|
|
|
if (list_empty(&c->body.instrs))
|
|
|
|
|
{
|
|
|
|
|
terminal_break = !!list_next(&s->cases, &c->entry);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
node = LIST_ENTRY(list_tail(&c->body.instrs), struct hlsl_ir_node, entry);
|
|
|
|
|
if (node->type == HLSL_IR_JUMP)
|
2023-11-24 15:25:21 -08:00
|
|
|
terminal_break = (hlsl_ir_jump(node)->type == HLSL_IR_JUMP_BREAK);
|
2023-10-11 05:39:01 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
missing_terminal_break |= !terminal_break;
|
|
|
|
|
|
|
|
|
|
if (!terminal_break)
|
|
|
|
|
{
|
|
|
|
|
if (c->is_default)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &c->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_SYNTAX,
|
|
|
|
|
"The 'default' case block is not terminated with 'break' or 'return'.");
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &c->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_SYNTAX,
|
|
|
|
|
"Switch case block '%u' is not terminated with 'break' or 'return'.", c->value);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (missing_terminal_break)
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
if (c->is_default)
|
|
|
|
|
{
|
|
|
|
|
def = c;
|
|
|
|
|
|
|
|
|
|
/* Remove preceding empty cases. */
|
|
|
|
|
while (list_prev(&s->cases, &def->entry))
|
|
|
|
|
{
|
|
|
|
|
c = LIST_ENTRY(list_prev(&s->cases, &def->entry), struct hlsl_ir_switch_case, entry);
|
|
|
|
|
if (!list_empty(&c->body.instrs))
|
|
|
|
|
break;
|
|
|
|
|
hlsl_free_ir_switch_case(c);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (list_empty(&def->body.instrs))
|
|
|
|
|
{
|
|
|
|
|
/* Remove following empty cases. */
|
|
|
|
|
while (list_next(&s->cases, &def->entry))
|
|
|
|
|
{
|
|
|
|
|
c = LIST_ENTRY(list_next(&s->cases, &def->entry), struct hlsl_ir_switch_case, entry);
|
|
|
|
|
if (!list_empty(&c->body.instrs))
|
|
|
|
|
break;
|
|
|
|
|
hlsl_free_ir_switch_case(c);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Merge with the next case. */
|
|
|
|
|
if (list_next(&s->cases, &def->entry))
|
|
|
|
|
{
|
|
|
|
|
c = LIST_ENTRY(list_next(&s->cases, &def->entry), struct hlsl_ir_switch_case, entry);
|
|
|
|
|
c->is_default = true;
|
|
|
|
|
hlsl_free_ir_switch_case(def);
|
|
|
|
|
def = c;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (def)
|
|
|
|
|
{
|
|
|
|
|
list_remove(&def->entry);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *jump;
|
|
|
|
|
|
|
|
|
|
if (!(def = hlsl_new_switch_case(ctx, 0, true, NULL, &s->node.loc)))
|
|
|
|
|
return true;
|
|
|
|
|
if (!(jump = hlsl_new_jump(ctx, HLSL_IR_JUMP_BREAK, NULL, &s->node.loc)))
|
|
|
|
|
{
|
|
|
|
|
hlsl_free_ir_switch_case(def);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
hlsl_block_add_instr(&def->body, jump);
|
|
|
|
|
}
|
|
|
|
|
list_add_tail(&s->cases, &def->entry);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_nonconstant_vector_derefs(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-05-08 15:25:18 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *idx;
|
|
|
|
|
struct hlsl_deref *deref;
|
|
|
|
|
struct hlsl_type *type;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_LOAD)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
deref = &hlsl_ir_load(instr)->src;
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(deref->var);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
|
|
|
|
if (deref->path_len == 0)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
type = deref->var->data_type;
|
|
|
|
|
for (i = 0; i < deref->path_len - 1; ++i)
|
|
|
|
|
type = hlsl_get_element_type_from_path_index(ctx, type, deref->path[i].node);
|
|
|
|
|
|
|
|
|
|
idx = deref->path[deref->path_len - 1].node;
|
|
|
|
|
|
|
|
|
|
if (type->class == HLSL_CLASS_VECTOR && idx->type != HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
2022-11-11 17:13:26 -08:00
|
|
|
struct hlsl_ir_node *eq, *swizzle, *dot, *c, *operands[HLSL_MAX_OPERANDS] = {0};
|
2022-11-11 17:10:14 -08:00
|
|
|
struct hlsl_constant_value value;
|
2023-05-08 15:25:18 -07:00
|
|
|
struct hlsl_ir_load *vector_load;
|
|
|
|
|
enum hlsl_ir_expr_op op;
|
|
|
|
|
|
|
|
|
|
if (!(vector_load = hlsl_new_load_parent(ctx, deref, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, &vector_load->node);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
|
|
|
|
if (!(swizzle = hlsl_new_swizzle(ctx, HLSL_SWIZZLE(X, X, X, X), type->dimx, idx, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, swizzle);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
2022-11-11 17:10:14 -08:00
|
|
|
value.u[0].u = 0;
|
|
|
|
|
value.u[1].u = 1;
|
|
|
|
|
value.u[2].u = 2;
|
|
|
|
|
value.u[3].u = 3;
|
|
|
|
|
if (!(c = hlsl_new_constant(ctx, hlsl_get_vector_type(ctx, HLSL_TYPE_UINT, type->dimx), &value, &instr->loc)))
|
2023-05-08 15:25:18 -07:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, c);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
|
|
|
|
operands[0] = swizzle;
|
2022-11-11 17:13:26 -08:00
|
|
|
operands[1] = c;
|
2023-05-08 15:25:18 -07:00
|
|
|
if (!(eq = hlsl_new_expr(ctx, HLSL_OP2_EQUAL, operands,
|
|
|
|
|
hlsl_get_vector_type(ctx, HLSL_TYPE_BOOL, type->dimx), &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, eq);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
|
|
|
|
if (!(eq = hlsl_new_cast(ctx, eq, type, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, eq);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
|
|
|
|
op = HLSL_OP2_DOT;
|
|
|
|
|
if (type->dimx == 1)
|
2024-02-27 15:30:51 -08:00
|
|
|
op = type->e.numeric.type == HLSL_TYPE_BOOL ? HLSL_OP2_LOGIC_AND : HLSL_OP2_MUL;
|
2023-05-08 15:25:18 -07:00
|
|
|
|
|
|
|
|
/* Note: We may be creating a DOT for bool vectors here, which we need to lower to
|
|
|
|
|
* LOGIC_OR + LOGIC_AND. */
|
|
|
|
|
operands[0] = &vector_load->node;
|
|
|
|
|
operands[1] = eq;
|
|
|
|
|
if (!(dot = hlsl_new_expr(ctx, op, operands, instr->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, dot);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2024-01-11 11:54:20 -08:00
|
|
|
static bool validate_nonconstant_vector_store_derefs(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *idx;
|
|
|
|
|
struct hlsl_deref *deref;
|
|
|
|
|
struct hlsl_type *type;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_STORE)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
deref = &hlsl_ir_store(instr)->lhs;
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(deref->var);
|
2024-01-11 11:54:20 -08:00
|
|
|
|
|
|
|
|
if (deref->path_len == 0)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
type = deref->var->data_type;
|
|
|
|
|
for (i = 0; i < deref->path_len - 1; ++i)
|
|
|
|
|
type = hlsl_get_element_type_from_path_index(ctx, type, deref->path[i].node);
|
|
|
|
|
|
|
|
|
|
idx = deref->path[deref->path_len - 1].node;
|
|
|
|
|
|
|
|
|
|
if (type->class == HLSL_CLASS_VECTOR && idx->type != HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
|
|
|
|
/* We should turn this into an hlsl_error after we implement unrolling, because if we get
|
|
|
|
|
* here after that, it means that the HLSL is invalid. */
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Non-constant vector addressing on store. Unrolling may be missing.");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2024-07-24 14:46:04 -07:00
|
|
|
/* This pass flattens array (and row_major matrix) loads that include the indexing of a non-constant
|
|
|
|
|
* index into multiple constant loads, where the value of only one of them ends up in the resulting
|
|
|
|
|
* node.
|
2024-07-08 12:13:07 -07:00
|
|
|
* This is achieved through a synthetic variable. The non-constant index is compared for equality
|
|
|
|
|
* with every possible value it can have within the array bounds, and the ternary operator is used
|
|
|
|
|
* to update the value of the synthetic var when the equality check passes. */
|
|
|
|
|
static bool lower_nonconstant_array_loads(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr,
|
|
|
|
|
struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_constant_value zero_value = {0};
|
|
|
|
|
struct hlsl_ir_node *cut_index, *zero, *store;
|
2024-07-24 14:46:04 -07:00
|
|
|
unsigned int i, i_cut, element_count;
|
2024-07-08 12:13:07 -07:00
|
|
|
const struct hlsl_deref *deref;
|
|
|
|
|
struct hlsl_type *cut_type;
|
|
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
struct hlsl_ir_var *var;
|
2024-07-24 14:46:04 -07:00
|
|
|
bool row_major;
|
2024-07-08 12:13:07 -07:00
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_LOAD)
|
|
|
|
|
return false;
|
|
|
|
|
load = hlsl_ir_load(instr);
|
|
|
|
|
deref = &load->src;
|
|
|
|
|
|
|
|
|
|
if (deref->path_len == 0)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
for (i = deref->path_len - 1; ; --i)
|
|
|
|
|
{
|
|
|
|
|
if (deref->path[i].node->type != HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
|
|
|
|
i_cut = i;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (i == 0)
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
cut_index = deref->path[i_cut].node;
|
|
|
|
|
cut_type = deref->var->data_type;
|
|
|
|
|
for (i = 0; i < i_cut; ++i)
|
|
|
|
|
cut_type = hlsl_get_element_type_from_path_index(ctx, cut_type, deref->path[i].node);
|
|
|
|
|
|
2024-07-24 14:46:04 -07:00
|
|
|
row_major = hlsl_type_is_row_major(cut_type);
|
|
|
|
|
VKD3D_ASSERT(cut_type->class == HLSL_CLASS_ARRAY || row_major);
|
2024-07-08 12:13:07 -07:00
|
|
|
|
2024-07-24 14:46:04 -07:00
|
|
|
if (!(var = hlsl_new_synthetic_var(ctx, row_major ? "row_major-load" : "array-load", instr->data_type, &instr->loc)))
|
2024-07-08 12:13:07 -07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!(zero = hlsl_new_constant(ctx, instr->data_type, &zero_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, zero);
|
|
|
|
|
|
|
|
|
|
if (!(store = hlsl_new_simple_store(ctx, var, zero)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, store);
|
|
|
|
|
|
2024-07-24 14:46:04 -07:00
|
|
|
TRACE("Lowering non-constant %s load on variable '%s'.\n", row_major ? "row_major" : "array", deref->var->name);
|
|
|
|
|
|
|
|
|
|
element_count = hlsl_type_element_count(cut_type);
|
|
|
|
|
for (i = 0; i < element_count; ++i)
|
2024-07-08 12:13:07 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_type *btype = hlsl_get_scalar_type(ctx, HLSL_TYPE_BOOL);
|
|
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS] = {0};
|
|
|
|
|
struct hlsl_ir_node *const_i, *equals, *ternary, *var_store;
|
|
|
|
|
struct hlsl_ir_load *var_load, *specific_load;
|
|
|
|
|
struct hlsl_deref deref_copy = {0};
|
|
|
|
|
|
|
|
|
|
if (!(const_i = hlsl_new_uint_constant(ctx, i, &cut_index->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, const_i);
|
|
|
|
|
|
|
|
|
|
operands[0] = cut_index;
|
|
|
|
|
operands[1] = const_i;
|
|
|
|
|
if (!(equals = hlsl_new_expr(ctx, HLSL_OP2_EQUAL, operands, btype, &cut_index->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, equals);
|
|
|
|
|
|
|
|
|
|
if (!(equals = hlsl_new_swizzle(ctx, HLSL_SWIZZLE(X, X, X, X), var->data_type->dimx, equals, &cut_index->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, equals);
|
|
|
|
|
|
|
|
|
|
if (!(var_load = hlsl_new_var_load(ctx, var, &cut_index->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, &var_load->node);
|
|
|
|
|
|
|
|
|
|
if (!hlsl_copy_deref(ctx, &deref_copy, deref))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_src_remove(&deref_copy.path[i_cut]);
|
|
|
|
|
hlsl_src_from_node(&deref_copy.path[i_cut], const_i);
|
|
|
|
|
|
|
|
|
|
if (!(specific_load = hlsl_new_load_index(ctx, &deref_copy, NULL, &cut_index->loc)))
|
|
|
|
|
{
|
|
|
|
|
hlsl_cleanup_deref(&deref_copy);
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
hlsl_block_add_instr(block, &specific_load->node);
|
|
|
|
|
|
|
|
|
|
hlsl_cleanup_deref(&deref_copy);
|
|
|
|
|
|
|
|
|
|
operands[0] = equals;
|
|
|
|
|
operands[1] = &specific_load->node;
|
|
|
|
|
operands[2] = &var_load->node;
|
|
|
|
|
if (!(ternary = hlsl_new_expr(ctx, HLSL_OP3_TERNARY, operands, instr->data_type, &cut_index->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, ternary);
|
|
|
|
|
|
|
|
|
|
if (!(var_store = hlsl_new_simple_store(ctx, var, ternary)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, var_store);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(load = hlsl_new_var_load(ctx, var, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, &load->node);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
2023-05-29 18:59:17 -07:00
|
|
|
/* Lower combined samples and sampler variables to synthesized separated textures and samplers.
|
|
|
|
|
* That is, translate SM1-style samples in the source to SM4-style samples in the bytecode. */
|
|
|
|
|
static bool lower_combined_samples(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_resource_load *load;
|
|
|
|
|
struct vkd3d_string_buffer *name;
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_RESOURCE_LOAD)
|
|
|
|
|
return false;
|
|
|
|
|
load = hlsl_ir_resource_load(instr);
|
|
|
|
|
|
|
|
|
|
switch (load->load_type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_RESOURCE_LOAD:
|
|
|
|
|
case HLSL_RESOURCE_GATHER_RED:
|
|
|
|
|
case HLSL_RESOURCE_GATHER_GREEN:
|
|
|
|
|
case HLSL_RESOURCE_GATHER_BLUE:
|
|
|
|
|
case HLSL_RESOURCE_GATHER_ALPHA:
|
2023-06-07 10:56:02 -07:00
|
|
|
case HLSL_RESOURCE_RESINFO:
|
2023-05-29 18:59:17 -07:00
|
|
|
case HLSL_RESOURCE_SAMPLE_CMP:
|
|
|
|
|
case HLSL_RESOURCE_SAMPLE_CMP_LZ:
|
2023-06-07 10:56:02 -07:00
|
|
|
case HLSL_RESOURCE_SAMPLE_INFO:
|
2023-05-29 18:59:17 -07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
case HLSL_RESOURCE_SAMPLE:
|
2024-05-25 11:56:54 -07:00
|
|
|
case HLSL_RESOURCE_SAMPLE_GRAD:
|
2023-05-29 18:59:17 -07:00
|
|
|
case HLSL_RESOURCE_SAMPLE_LOD:
|
|
|
|
|
case HLSL_RESOURCE_SAMPLE_LOD_BIAS:
|
2023-11-09 09:43:41 -08:00
|
|
|
case HLSL_RESOURCE_SAMPLE_PROJ:
|
2023-05-29 18:59:17 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
if (load->sampler.var)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!hlsl_type_is_resource(load->resource.var->data_type))
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Lower combined samplers within structs.");
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(hlsl_deref_get_regset(ctx, &load->resource) == HLSL_REGSET_SAMPLERS);
|
2023-05-29 18:59:17 -07:00
|
|
|
|
|
|
|
|
if (!(name = hlsl_get_string_buffer(ctx)))
|
|
|
|
|
return false;
|
|
|
|
|
vkd3d_string_buffer_printf(name, "<resource>%s", load->resource.var->name);
|
|
|
|
|
|
|
|
|
|
TRACE("Lowering to separate resource %s.\n", debugstr_a(name->buffer));
|
|
|
|
|
|
|
|
|
|
if (!(var = hlsl_get_var(ctx->globals, name->buffer)))
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_type *texture_array_type = hlsl_new_texture_type(ctx, load->sampling_dim,
|
|
|
|
|
hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, 4), 0);
|
|
|
|
|
|
|
|
|
|
/* Create (possibly multi-dimensional) texture array type with the same dims as the sampler array. */
|
|
|
|
|
struct hlsl_type *arr_type = load->resource.var->data_type;
|
|
|
|
|
for (i = 0; i < load->resource.path_len; ++i)
|
|
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(arr_type->class == HLSL_CLASS_ARRAY);
|
2023-05-29 18:59:17 -07:00
|
|
|
texture_array_type = hlsl_new_array_type(ctx, texture_array_type, arr_type->e.array.elements_count);
|
|
|
|
|
arr_type = arr_type->e.array.type;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(var = hlsl_new_synthetic_var_named(ctx, name->buffer, texture_array_type, &instr->loc, false)))
|
|
|
|
|
{
|
|
|
|
|
hlsl_release_string_buffer(ctx, name);
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
var->is_uniform = 1;
|
|
|
|
|
var->is_separated_resource = true;
|
|
|
|
|
|
|
|
|
|
list_add_tail(&ctx->extern_vars, &var->extern_entry);
|
|
|
|
|
}
|
|
|
|
|
hlsl_release_string_buffer(ctx, name);
|
|
|
|
|
|
|
|
|
|
if (load->sampling_dim != var->data_type->sampler_dim)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &load->node.loc, VKD3D_SHADER_ERROR_HLSL_INCONSISTENT_SAMPLER,
|
|
|
|
|
"Cannot split combined samplers from \"%s\" if they have different usage dimensions.",
|
|
|
|
|
load->resource.var->name);
|
|
|
|
|
hlsl_note(ctx, &var->loc, VKD3D_SHADER_LOG_ERROR, "First use as combined sampler is here.");
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hlsl_copy_deref(ctx, &load->sampler, &load->resource);
|
|
|
|
|
load->resource.var = var;
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(hlsl_deref_get_type(ctx, &load->resource)->class == HLSL_CLASS_TEXTURE);
|
|
|
|
|
VKD3D_ASSERT(hlsl_deref_get_type(ctx, &load->sampler)->class == HLSL_CLASS_SAMPLER);
|
2023-05-29 18:59:17 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-08-04 12:02:39 -07:00
|
|
|
static void insert_ensuring_decreasing_bind_count(struct list *list, struct hlsl_ir_var *to_add,
|
|
|
|
|
enum hlsl_regset regset)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, list, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (var->bind_count[regset] < to_add->bind_count[regset])
|
|
|
|
|
{
|
|
|
|
|
list_add_before(&var->extern_entry, &to_add->extern_entry);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
list_add_tail(list, &to_add->extern_entry);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool sort_synthetic_separated_samplers_first(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
struct list separated_resources;
|
|
|
|
|
struct hlsl_ir_var *var, *next;
|
|
|
|
|
|
|
|
|
|
list_init(&separated_resources);
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY_SAFE(var, next, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (var->is_separated_resource)
|
|
|
|
|
{
|
|
|
|
|
list_remove(&var->extern_entry);
|
|
|
|
|
insert_ensuring_decreasing_bind_count(&separated_resources, var, HLSL_REGSET_TEXTURES);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
list_move_head(&ctx->extern_vars, &separated_resources);
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2024-02-21 11:36:43 -08:00
|
|
|
/* Turn CAST to int or uint into FLOOR + REINTERPRET (which is written as a mere MOV). */
|
2023-10-20 19:23:46 -07:00
|
|
|
static bool lower_casts_to_int(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
2024-02-21 11:36:43 -08:00
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS] = { 0 };
|
|
|
|
|
struct hlsl_ir_node *arg, *floor, *res;
|
2023-10-20 19:23:46 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP1_CAST)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
arg = expr->operands[0].node;
|
2024-02-27 15:30:51 -08:00
|
|
|
if (instr->data_type->e.numeric.type != HLSL_TYPE_INT && instr->data_type->e.numeric.type != HLSL_TYPE_UINT)
|
2023-10-20 19:23:46 -07:00
|
|
|
return false;
|
2024-02-27 15:30:51 -08:00
|
|
|
if (arg->data_type->e.numeric.type != HLSL_TYPE_FLOAT && arg->data_type->e.numeric.type != HLSL_TYPE_HALF)
|
2023-10-20 19:23:46 -07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!(floor = hlsl_new_unary_expr(ctx, HLSL_OP1_FLOOR, arg, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, floor);
|
|
|
|
|
|
2024-02-21 11:36:43 -08:00
|
|
|
memset(operands, 0, sizeof(operands));
|
|
|
|
|
operands[0] = floor;
|
|
|
|
|
if (!(res = hlsl_new_expr(ctx, HLSL_OP1_REINTERPRET, operands, instr->data_type, &instr->loc)))
|
2023-10-20 19:23:46 -07:00
|
|
|
return false;
|
2024-02-21 11:36:43 -08:00
|
|
|
hlsl_block_add_instr(block, res);
|
2023-10-20 19:23:46 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-20 22:32:24 -07:00
|
|
|
/* Lower DIV to RCP + MUL. */
|
2023-06-25 17:10:34 -07:00
|
|
|
static bool lower_division(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2021-05-20 22:32:24 -07:00
|
|
|
{
|
2023-06-25 17:10:34 -07:00
|
|
|
struct hlsl_ir_node *rcp, *mul;
|
2021-05-20 22:32:24 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
2021-08-12 17:36:13 -07:00
|
|
|
if (expr->op != HLSL_OP2_DIV)
|
2021-05-20 22:32:24 -07:00
|
|
|
return false;
|
|
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(rcp = hlsl_new_unary_expr(ctx, HLSL_OP1_RCP, expr->operands[1].node, &instr->loc)))
|
2021-05-20 22:32:24 -07:00
|
|
|
return false;
|
2023-06-25 17:10:34 -07:00
|
|
|
hlsl_block_add_instr(block, rcp);
|
|
|
|
|
|
|
|
|
|
if (!(mul = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, expr->operands[0].node, rcp)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, mul);
|
|
|
|
|
|
2021-05-20 22:32:24 -07:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-24 04:44:39 -08:00
|
|
|
/* Lower SQRT to RSQ + RCP. */
|
2023-06-25 17:11:37 -07:00
|
|
|
static bool lower_sqrt(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-01-24 04:44:39 -08:00
|
|
|
{
|
2023-06-25 17:11:37 -07:00
|
|
|
struct hlsl_ir_node *rsq, *rcp;
|
2023-01-24 04:44:39 -08:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP1_SQRT)
|
|
|
|
|
return false;
|
|
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(rsq = hlsl_new_unary_expr(ctx, HLSL_OP1_RSQ, expr->operands[0].node, &instr->loc)))
|
2023-01-24 04:44:39 -08:00
|
|
|
return false;
|
2023-06-25 17:11:37 -07:00
|
|
|
hlsl_block_add_instr(block, rsq);
|
|
|
|
|
|
|
|
|
|
if (!(rcp = hlsl_new_unary_expr(ctx, HLSL_OP1_RCP, rsq, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, rcp);
|
2023-01-24 04:44:39 -08:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-26 12:56:00 -08:00
|
|
|
/* Lower DP2 to MUL + ADD */
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_dot(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-01-26 12:56:00 -08:00
|
|
|
{
|
2022-11-10 19:01:18 -08:00
|
|
|
struct hlsl_ir_node *arg1, *arg2, *mul, *replacement, *zero, *add_x, *add_y;
|
2023-01-26 12:56:00 -08:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
arg2 = expr->operands[1].node;
|
|
|
|
|
if (expr->op != HLSL_OP2_DOT)
|
|
|
|
|
return false;
|
|
|
|
|
if (arg1->data_type->dimx != 2)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (ctx->profile->type == VKD3D_SHADER_TYPE_PIXEL)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS] = { 0 };
|
|
|
|
|
|
|
|
|
|
if (!(zero = hlsl_new_float_constant(ctx, 0.0f, &expr->node.loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, zero);
|
2023-01-26 12:56:00 -08:00
|
|
|
|
|
|
|
|
operands[0] = arg1;
|
|
|
|
|
operands[1] = arg2;
|
2022-11-10 17:45:51 -08:00
|
|
|
operands[2] = zero;
|
2023-01-26 12:56:00 -08:00
|
|
|
|
|
|
|
|
if (!(replacement = hlsl_new_expr(ctx, HLSL_OP3_DP2ADD, operands, instr->data_type, &expr->node.loc)))
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
if (!(mul = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, expr->operands[0].node, expr->operands[1].node)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, mul);
|
2023-01-26 12:56:00 -08:00
|
|
|
|
|
|
|
|
if (!(add_x = hlsl_new_swizzle(ctx, HLSL_SWIZZLE(X, X, X, X), instr->data_type->dimx, mul, &expr->node.loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, add_x);
|
2023-01-26 12:56:00 -08:00
|
|
|
|
|
|
|
|
if (!(add_y = hlsl_new_swizzle(ctx, HLSL_SWIZZLE(Y, Y, Y, Y), instr->data_type->dimx, mul, &expr->node.loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, add_y);
|
2023-01-26 12:56:00 -08:00
|
|
|
|
2022-11-10 19:01:18 -08:00
|
|
|
if (!(replacement = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, add_x, add_y)))
|
2023-01-26 12:56:00 -08:00
|
|
|
return false;
|
|
|
|
|
}
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, replacement);
|
2023-01-26 12:56:00 -08:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-02-04 15:30:36 -08:00
|
|
|
/* Lower ABS to MAX */
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_abs(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-02-04 15:30:36 -08:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *arg, *neg, *replacement;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg = expr->operands[0].node;
|
|
|
|
|
if (expr->op != HLSL_OP1_ABS)
|
|
|
|
|
return false;
|
|
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, arg, &instr->loc)))
|
2023-02-04 15:30:36 -08:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, neg);
|
2023-02-04 15:30:36 -08:00
|
|
|
|
|
|
|
|
if (!(replacement = hlsl_new_binary_expr(ctx, HLSL_OP2_MAX, neg, arg)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, replacement);
|
2023-02-04 15:30:36 -08:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-16 13:14:12 -07:00
|
|
|
/* Lower ROUND using FRC, ROUND(x) -> ((x + 0.5) - FRC(x + 0.5)). */
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_round(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-03-16 13:14:12 -07:00
|
|
|
{
|
2022-11-11 17:13:26 -08:00
|
|
|
struct hlsl_ir_node *arg, *neg, *sum, *frc, *half, *replacement;
|
2023-03-16 13:14:12 -07:00
|
|
|
struct hlsl_type *type = instr->data_type;
|
2022-11-11 17:10:14 -08:00
|
|
|
struct hlsl_constant_value half_value;
|
2023-03-16 13:14:12 -07:00
|
|
|
unsigned int i, component_count;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg = expr->operands[0].node;
|
|
|
|
|
if (expr->op != HLSL_OP1_ROUND)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
component_count = hlsl_type_component_count(type);
|
|
|
|
|
for (i = 0; i < component_count; ++i)
|
2022-11-11 17:10:14 -08:00
|
|
|
half_value.u[i].f = 0.5f;
|
|
|
|
|
if (!(half = hlsl_new_constant(ctx, type, &half_value, &expr->node.loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, half);
|
2023-03-16 13:14:12 -07:00
|
|
|
|
2022-11-11 17:13:26 -08:00
|
|
|
if (!(sum = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, arg, half)))
|
2023-03-16 13:14:12 -07:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, sum);
|
2023-03-16 13:14:12 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(frc = hlsl_new_unary_expr(ctx, HLSL_OP1_FRACT, sum, &instr->loc)))
|
2023-03-16 13:14:12 -07:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, frc);
|
2023-03-16 13:14:12 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, frc, &instr->loc)))
|
2023-03-16 13:14:12 -07:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, neg);
|
2023-03-16 13:14:12 -07:00
|
|
|
|
|
|
|
|
if (!(replacement = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, sum, neg)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, replacement);
|
2023-03-16 13:14:12 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-17 14:22:52 -07:00
|
|
|
/* Lower CEIL to FRC */
|
|
|
|
|
static bool lower_ceil(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *arg, *neg, *sum, *frc;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg = expr->operands[0].node;
|
|
|
|
|
if (expr->op != HLSL_OP1_CEIL)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, arg, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, neg);
|
|
|
|
|
|
|
|
|
|
if (!(frc = hlsl_new_unary_expr(ctx, HLSL_OP1_FRACT, neg, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, frc);
|
|
|
|
|
|
|
|
|
|
if (!(sum = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, frc, arg)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sum);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-17 14:26:04 -07:00
|
|
|
/* Lower FLOOR to FRC */
|
|
|
|
|
static bool lower_floor(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *arg, *neg, *sum, *frc;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg = expr->operands[0].node;
|
|
|
|
|
if (expr->op != HLSL_OP1_FLOOR)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!(frc = hlsl_new_unary_expr(ctx, HLSL_OP1_FRACT, arg, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, frc);
|
|
|
|
|
|
|
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, frc, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, neg);
|
|
|
|
|
|
|
|
|
|
if (!(sum = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, neg, arg)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sum);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-07-10 19:47:51 -07:00
|
|
|
/* Lower SIN/COS to SINCOS for SM1. */
|
|
|
|
|
static bool lower_trig(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *arg, *half, *two_pi, *reciprocal_two_pi, *neg_pi;
|
|
|
|
|
struct hlsl_constant_value half_value, two_pi_value, reciprocal_two_pi_value, neg_pi_value;
|
|
|
|
|
struct hlsl_ir_node *mad, *frc, *reduced;
|
|
|
|
|
struct hlsl_type *type;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
enum hlsl_ir_expr_op op;
|
|
|
|
|
struct hlsl_ir_node *sincos;
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
|
|
|
|
|
if (expr->op == HLSL_OP1_SIN)
|
|
|
|
|
op = HLSL_OP1_SIN_REDUCED;
|
|
|
|
|
else if (expr->op == HLSL_OP1_COS)
|
|
|
|
|
op = HLSL_OP1_COS_REDUCED;
|
|
|
|
|
else
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
arg = expr->operands[0].node;
|
|
|
|
|
type = arg->data_type;
|
|
|
|
|
|
|
|
|
|
/* Reduce the range of the input angles to [-pi, pi]. */
|
|
|
|
|
for (i = 0; i < type->dimx; ++i)
|
|
|
|
|
{
|
|
|
|
|
half_value.u[i].f = 0.5;
|
|
|
|
|
two_pi_value.u[i].f = 2.0 * M_PI;
|
|
|
|
|
reciprocal_two_pi_value.u[i].f = 1.0 / (2.0 * M_PI);
|
|
|
|
|
neg_pi_value.u[i].f = -M_PI;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(half = hlsl_new_constant(ctx, type, &half_value, &instr->loc))
|
|
|
|
|
|| !(two_pi = hlsl_new_constant(ctx, type, &two_pi_value, &instr->loc))
|
|
|
|
|
|| !(reciprocal_two_pi = hlsl_new_constant(ctx, type, &reciprocal_two_pi_value, &instr->loc))
|
|
|
|
|
|| !(neg_pi = hlsl_new_constant(ctx, type, &neg_pi_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, half);
|
|
|
|
|
hlsl_block_add_instr(block, two_pi);
|
|
|
|
|
hlsl_block_add_instr(block, reciprocal_two_pi);
|
|
|
|
|
hlsl_block_add_instr(block, neg_pi);
|
|
|
|
|
|
|
|
|
|
if (!(mad = hlsl_new_ternary_expr(ctx, HLSL_OP3_MAD, arg, reciprocal_two_pi, half)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, mad);
|
|
|
|
|
if (!(frc = hlsl_new_unary_expr(ctx, HLSL_OP1_FRACT, mad, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, frc);
|
|
|
|
|
if (!(reduced = hlsl_new_ternary_expr(ctx, HLSL_OP3_MAD, frc, two_pi, neg_pi)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, reduced);
|
|
|
|
|
|
|
|
|
|
if (type->dimx == 1)
|
|
|
|
|
{
|
|
|
|
|
if (!(sincos = hlsl_new_unary_expr(ctx, op, reduced, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sincos);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *comps[4] = {0};
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
struct hlsl_deref var_deref;
|
|
|
|
|
struct hlsl_ir_load *var_load;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < type->dimx; ++i)
|
|
|
|
|
{
|
|
|
|
|
uint32_t s = hlsl_swizzle_from_writemask(1 << i);
|
|
|
|
|
|
|
|
|
|
if (!(comps[i] = hlsl_new_swizzle(ctx, s, 1, reduced, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, comps[i]);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(var = hlsl_new_synthetic_var(ctx, "sincos", type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_init_simple_deref_from_var(&var_deref, var);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < type->dimx; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_block store_block;
|
|
|
|
|
|
|
|
|
|
if (!(sincos = hlsl_new_unary_expr(ctx, op, comps[i], &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sincos);
|
|
|
|
|
|
|
|
|
|
if (!hlsl_new_store_component(ctx, &store_block, &var_deref, i, sincos))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_block(block, &store_block);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(var_load = hlsl_new_load_index(ctx, &var_deref, NULL, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, &var_load->node);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-02-20 17:03:43 -08:00
|
|
|
static bool lower_logic_not(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS];
|
|
|
|
|
struct hlsl_ir_node *arg, *arg_cast, *neg, *one, *sub, *res;
|
|
|
|
|
struct hlsl_constant_value one_value;
|
|
|
|
|
struct hlsl_type *float_type;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP1_LOGIC_NOT)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
arg = expr->operands[0].node;
|
|
|
|
|
float_type = hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, arg->data_type->dimx);
|
|
|
|
|
|
2024-03-14 20:48:01 -07:00
|
|
|
/* If this is happens, it means we failed to cast the argument to boolean somewhere. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(arg->data_type->e.numeric.type == HLSL_TYPE_BOOL);
|
2024-03-14 20:48:01 -07:00
|
|
|
|
2024-02-20 17:03:43 -08:00
|
|
|
if (!(arg_cast = hlsl_new_cast(ctx, arg, float_type, &arg->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, arg_cast);
|
|
|
|
|
|
|
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, arg_cast, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, neg);
|
|
|
|
|
|
|
|
|
|
one_value.u[0].f = 1.0;
|
|
|
|
|
one_value.u[1].f = 1.0;
|
|
|
|
|
one_value.u[2].f = 1.0;
|
|
|
|
|
one_value.u[3].f = 1.0;
|
|
|
|
|
if (!(one = hlsl_new_constant(ctx, float_type, &one_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, one);
|
|
|
|
|
|
|
|
|
|
if (!(sub = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, one, neg)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sub);
|
|
|
|
|
|
|
|
|
|
memset(operands, 0, sizeof(operands));
|
|
|
|
|
operands[0] = sub;
|
|
|
|
|
if (!(res = hlsl_new_expr(ctx, HLSL_OP1_REINTERPRET, operands, instr->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, res);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-04-05 12:34:32 -07:00
|
|
|
/* Lower TERNARY to CMP for SM1. */
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_ternary(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-07-24 23:46:28 -07:00
|
|
|
{
|
2023-09-26 11:48:58 -07:00
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS] = { 0 }, *replacement;
|
2024-03-01 11:01:03 -08:00
|
|
|
struct hlsl_ir_node *cond, *first, *second, *float_cond, *neg;
|
2023-07-24 23:46:28 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
struct hlsl_type *type;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP3_TERNARY)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
cond = expr->operands[0].node;
|
|
|
|
|
first = expr->operands[1].node;
|
|
|
|
|
second = expr->operands[2].node;
|
|
|
|
|
|
2023-11-09 17:41:00 -08:00
|
|
|
if (cond->data_type->class > HLSL_CLASS_VECTOR || instr->data_type->class > HLSL_CLASS_VECTOR)
|
|
|
|
|
{
|
2024-05-22 15:57:32 -07:00
|
|
|
hlsl_fixme(ctx, &instr->loc, "Lower ternary of type other than scalar or vector.");
|
2023-11-09 17:41:00 -08:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(cond->data_type->e.numeric.type == HLSL_TYPE_BOOL);
|
2024-03-01 11:01:03 -08:00
|
|
|
|
2024-04-05 12:34:32 -07:00
|
|
|
type = hlsl_get_numeric_type(ctx, instr->data_type->class, HLSL_TYPE_FLOAT,
|
|
|
|
|
instr->data_type->dimx, instr->data_type->dimy);
|
2023-09-26 11:48:58 -07:00
|
|
|
|
2024-04-05 12:34:32 -07:00
|
|
|
if (!(float_cond = hlsl_new_cast(ctx, cond, type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, float_cond);
|
2023-09-26 11:48:58 -07:00
|
|
|
|
2024-04-05 12:34:32 -07:00
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, float_cond, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, neg);
|
2023-09-26 11:48:58 -07:00
|
|
|
|
2024-04-05 12:34:32 -07:00
|
|
|
memset(operands, 0, sizeof(operands));
|
|
|
|
|
operands[0] = neg;
|
|
|
|
|
operands[1] = second;
|
|
|
|
|
operands[2] = first;
|
|
|
|
|
if (!(replacement = hlsl_new_expr(ctx, HLSL_OP3_CMP, operands, first->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
2023-07-24 23:46:28 -07:00
|
|
|
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, replacement);
|
2023-07-24 23:46:28 -07:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-02-26 10:41:12 -08:00
|
|
|
static bool lower_comparison_operators(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr,
|
|
|
|
|
struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *arg1, *arg1_cast, *arg2, *arg2_cast, *slt, *res, *ret;
|
|
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS];
|
|
|
|
|
struct hlsl_type *float_type;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
bool negate = false;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP2_EQUAL && expr->op != HLSL_OP2_NEQUAL && expr->op != HLSL_OP2_LESS
|
|
|
|
|
&& expr->op != HLSL_OP2_GEQUAL)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
arg2 = expr->operands[1].node;
|
|
|
|
|
float_type = hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, instr->data_type->dimx);
|
|
|
|
|
|
|
|
|
|
if (!(arg1_cast = hlsl_new_cast(ctx, arg1, float_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, arg1_cast);
|
|
|
|
|
|
|
|
|
|
if (!(arg2_cast = hlsl_new_cast(ctx, arg2, float_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, arg2_cast);
|
|
|
|
|
|
|
|
|
|
switch (expr->op)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_OP2_EQUAL:
|
|
|
|
|
case HLSL_OP2_NEQUAL:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *neg, *sub, *abs, *abs_neg;
|
|
|
|
|
|
|
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, arg2_cast, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, neg);
|
|
|
|
|
|
|
|
|
|
if (!(sub = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, arg1_cast, neg)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sub);
|
|
|
|
|
|
|
|
|
|
if (ctx->profile->major_version >= 3)
|
|
|
|
|
{
|
|
|
|
|
if (!(abs = hlsl_new_unary_expr(ctx, HLSL_OP1_ABS, sub, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, abs);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
/* Use MUL as a precarious ABS. */
|
|
|
|
|
if (!(abs = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, sub, sub)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, abs);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(abs_neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, abs, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, abs_neg);
|
|
|
|
|
|
|
|
|
|
if (!(slt = hlsl_new_binary_expr(ctx, HLSL_OP2_SLT, abs_neg, abs)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, slt);
|
|
|
|
|
|
|
|
|
|
negate = (expr->op == HLSL_OP2_EQUAL);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_GEQUAL:
|
|
|
|
|
case HLSL_OP2_LESS:
|
|
|
|
|
{
|
|
|
|
|
if (!(slt = hlsl_new_binary_expr(ctx, HLSL_OP2_SLT, arg1_cast, arg2_cast)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, slt);
|
|
|
|
|
|
|
|
|
|
negate = (expr->op == HLSL_OP2_GEQUAL);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (negate)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_constant_value one_value;
|
|
|
|
|
struct hlsl_ir_node *one, *slt_neg;
|
|
|
|
|
|
|
|
|
|
one_value.u[0].f = 1.0;
|
|
|
|
|
one_value.u[1].f = 1.0;
|
|
|
|
|
one_value.u[2].f = 1.0;
|
|
|
|
|
one_value.u[3].f = 1.0;
|
|
|
|
|
if (!(one = hlsl_new_constant(ctx, float_type, &one_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, one);
|
|
|
|
|
|
|
|
|
|
if (!(slt_neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, slt, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, slt_neg);
|
|
|
|
|
|
|
|
|
|
if (!(res = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, one, slt_neg)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, res);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
res = slt;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* We need a REINTERPRET so that the HLSL IR code is valid. SLT and its arguments must be FLOAT,
|
|
|
|
|
* and casts to BOOL have already been lowered to "!= 0". */
|
|
|
|
|
memset(operands, 0, sizeof(operands));
|
|
|
|
|
operands[0] = res;
|
|
|
|
|
if (!(ret = hlsl_new_expr(ctx, HLSL_OP1_REINTERPRET, operands, instr->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, ret);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-11-01 13:07:46 -07:00
|
|
|
/* Intended to be used for SM1-SM3, lowers SLT instructions (only available in vertex shaders) to
|
|
|
|
|
* CMP instructions (only available in pixel shaders).
|
|
|
|
|
* Based on the following equivalence:
|
|
|
|
|
* SLT(x, y)
|
|
|
|
|
* = (x < y) ? 1.0 : 0.0
|
|
|
|
|
* = ((x - y) >= 0) ? 0.0 : 1.0
|
|
|
|
|
* = CMP(x - y, 0.0, 1.0)
|
|
|
|
|
*/
|
|
|
|
|
static bool lower_slt(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *arg1, *arg2, *arg1_cast, *arg2_cast, *neg, *sub, *zero, *one, *cmp;
|
|
|
|
|
struct hlsl_constant_value zero_value, one_value;
|
|
|
|
|
struct hlsl_type *float_type;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP2_SLT)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
arg2 = expr->operands[1].node;
|
|
|
|
|
float_type = hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, instr->data_type->dimx);
|
|
|
|
|
|
|
|
|
|
if (!(arg1_cast = hlsl_new_cast(ctx, arg1, float_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, arg1_cast);
|
|
|
|
|
|
|
|
|
|
if (!(arg2_cast = hlsl_new_cast(ctx, arg2, float_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, arg2_cast);
|
|
|
|
|
|
|
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, arg2_cast, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, neg);
|
|
|
|
|
|
|
|
|
|
if (!(sub = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, arg1_cast, neg)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sub);
|
|
|
|
|
|
|
|
|
|
memset(&zero_value, 0, sizeof(zero_value));
|
|
|
|
|
if (!(zero = hlsl_new_constant(ctx, float_type, &zero_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, zero);
|
|
|
|
|
|
|
|
|
|
one_value.u[0].f = 1.0;
|
|
|
|
|
one_value.u[1].f = 1.0;
|
|
|
|
|
one_value.u[2].f = 1.0;
|
|
|
|
|
one_value.u[3].f = 1.0;
|
|
|
|
|
if (!(one = hlsl_new_constant(ctx, float_type, &one_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, one);
|
|
|
|
|
|
|
|
|
|
if (!(cmp = hlsl_new_ternary_expr(ctx, HLSL_OP3_CMP, sub, zero, one)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, cmp);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-02-26 09:01:45 -08:00
|
|
|
/* Intended to be used for SM1-SM3, lowers CMP instructions (only available in pixel shaders) to
|
|
|
|
|
* SLT instructions (only available in vertex shaders).
|
|
|
|
|
* Based on the following equivalence:
|
|
|
|
|
* CMP(x, y, z)
|
|
|
|
|
* = (x >= 0) ? y : z
|
|
|
|
|
* = z * ((x < 0) ? 1.0 : 0.0) + y * ((x < 0) ? 0.0 : 1.0)
|
|
|
|
|
* = z * SLT(x, 0.0) + y * (1 - SLT(x, 0.0))
|
|
|
|
|
*/
|
|
|
|
|
static bool lower_cmp(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *args[3], *args_cast[3], *slt, *neg_slt, *sub, *zero, *one, *mul1, *mul2, *add;
|
|
|
|
|
struct hlsl_constant_value zero_value, one_value;
|
|
|
|
|
struct hlsl_type *float_type;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP3_CMP)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
float_type = hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, instr->data_type->dimx);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < 3; ++i)
|
|
|
|
|
{
|
|
|
|
|
args[i] = expr->operands[i].node;
|
|
|
|
|
|
|
|
|
|
if (!(args_cast[i] = hlsl_new_cast(ctx, args[i], float_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, args_cast[i]);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
memset(&zero_value, 0, sizeof(zero_value));
|
|
|
|
|
if (!(zero = hlsl_new_constant(ctx, float_type, &zero_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, zero);
|
|
|
|
|
|
|
|
|
|
one_value.u[0].f = 1.0;
|
|
|
|
|
one_value.u[1].f = 1.0;
|
|
|
|
|
one_value.u[2].f = 1.0;
|
|
|
|
|
one_value.u[3].f = 1.0;
|
|
|
|
|
if (!(one = hlsl_new_constant(ctx, float_type, &one_value, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, one);
|
|
|
|
|
|
|
|
|
|
if (!(slt = hlsl_new_binary_expr(ctx, HLSL_OP2_SLT, args_cast[0], zero)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, slt);
|
|
|
|
|
|
|
|
|
|
if (!(mul1 = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, args_cast[2], slt)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, mul1);
|
|
|
|
|
|
|
|
|
|
if (!(neg_slt = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, slt, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, neg_slt);
|
|
|
|
|
|
|
|
|
|
if (!(sub = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, one, neg_slt)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, sub);
|
|
|
|
|
|
|
|
|
|
if (!(mul2 = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, args_cast[1], sub)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, mul2);
|
|
|
|
|
|
|
|
|
|
if (!(add = hlsl_new_binary_expr(ctx, HLSL_OP2_ADD, mul1, mul2)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, add);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-25 17:03:26 -07:00
|
|
|
static bool lower_casts_to_bool(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2022-04-06 12:24:42 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_type *type = instr->data_type, *arg_type;
|
2022-11-11 17:10:14 -08:00
|
|
|
static const struct hlsl_constant_value zero_value;
|
2023-06-25 17:03:26 -07:00
|
|
|
struct hlsl_ir_node *zero, *neq;
|
2022-04-06 12:24:42 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
if (expr->op != HLSL_OP1_CAST)
|
|
|
|
|
return false;
|
|
|
|
|
arg_type = expr->operands[0].node->data_type;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class > HLSL_CLASS_VECTOR || arg_type->class > HLSL_CLASS_VECTOR)
|
2022-04-06 12:24:42 -07:00
|
|
|
return false;
|
2024-02-27 15:30:51 -08:00
|
|
|
if (type->e.numeric.type != HLSL_TYPE_BOOL)
|
2022-04-06 12:24:42 -07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
/* Narrowing casts should have already been lowered. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(type->dimx == arg_type->dimx);
|
2022-04-06 12:24:42 -07:00
|
|
|
|
2022-11-11 17:10:14 -08:00
|
|
|
zero = hlsl_new_constant(ctx, arg_type, &zero_value, &instr->loc);
|
2022-04-06 12:24:42 -07:00
|
|
|
if (!zero)
|
|
|
|
|
return false;
|
2023-06-25 17:03:26 -07:00
|
|
|
hlsl_block_add_instr(block, zero);
|
2022-04-06 12:24:42 -07:00
|
|
|
|
2023-06-25 17:03:26 -07:00
|
|
|
if (!(neq = hlsl_new_binary_expr(ctx, HLSL_OP2_NEQUAL, expr->operands[0].node, zero)))
|
|
|
|
|
return false;
|
|
|
|
|
neq->data_type = expr->node.data_type;
|
|
|
|
|
hlsl_block_add_instr(block, neq);
|
2022-04-06 12:24:42 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-06 18:37:41 -08:00
|
|
|
struct hlsl_ir_node *hlsl_add_conditional(struct hlsl_ctx *ctx, struct hlsl_block *instrs,
|
2021-09-21 08:12:31 -07:00
|
|
|
struct hlsl_ir_node *condition, struct hlsl_ir_node *if_true, struct hlsl_ir_node *if_false)
|
|
|
|
|
{
|
2024-03-01 11:01:03 -08:00
|
|
|
struct hlsl_type *cond_type = condition->data_type;
|
2023-09-09 11:08:52 -07:00
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS];
|
|
|
|
|
struct hlsl_ir_node *cond;
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(hlsl_types_are_equal(if_true->data_type, if_false->data_type));
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
if (cond_type->e.numeric.type != HLSL_TYPE_BOOL)
|
2024-03-01 11:01:03 -08:00
|
|
|
{
|
|
|
|
|
cond_type = hlsl_get_numeric_type(ctx, cond_type->class, HLSL_TYPE_BOOL, cond_type->dimx, cond_type->dimy);
|
|
|
|
|
|
|
|
|
|
if (!(condition = hlsl_new_cast(ctx, condition, cond_type, &condition->loc)))
|
|
|
|
|
return NULL;
|
|
|
|
|
hlsl_block_add_instr(instrs, condition);
|
|
|
|
|
}
|
|
|
|
|
|
2023-09-09 11:08:52 -07:00
|
|
|
operands[0] = condition;
|
|
|
|
|
operands[1] = if_true;
|
|
|
|
|
operands[2] = if_false;
|
|
|
|
|
if (!(cond = hlsl_new_expr(ctx, HLSL_OP3_TERNARY, operands, if_true->data_type, &condition->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(instrs, cond);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2023-09-09 11:08:52 -07:00
|
|
|
return cond;
|
2021-09-21 08:12:31 -07:00
|
|
|
}
|
|
|
|
|
|
2023-03-06 18:34:10 -08:00
|
|
|
static bool lower_int_division(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2021-09-21 08:12:31 -07:00
|
|
|
{
|
2023-03-06 18:34:10 -08:00
|
|
|
struct hlsl_ir_node *arg1, *arg2, *xor, *and, *abs1, *abs2, *div, *neg, *cast1, *cast2, *cast3, *high_bit;
|
2021-09-21 08:12:31 -07:00
|
|
|
struct hlsl_type *type = instr->data_type, *utype;
|
2022-11-11 17:10:14 -08:00
|
|
|
struct hlsl_constant_value high_bit_value;
|
2021-09-21 08:12:31 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
arg2 = expr->operands[1].node;
|
|
|
|
|
if (expr->op != HLSL_OP2_DIV)
|
|
|
|
|
return false;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_SCALAR && type->class != HLSL_CLASS_VECTOR)
|
2021-09-21 08:12:31 -07:00
|
|
|
return false;
|
2024-02-27 15:30:51 -08:00
|
|
|
if (type->e.numeric.type != HLSL_TYPE_INT)
|
2021-09-21 08:12:31 -07:00
|
|
|
return false;
|
2022-11-11 17:31:55 -08:00
|
|
|
utype = hlsl_get_numeric_type(ctx, type->class, HLSL_TYPE_UINT, type->dimx, type->dimy);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
|
|
|
|
if (!(xor = hlsl_new_binary_expr(ctx, HLSL_OP2_BIT_XOR, arg1, arg2)))
|
|
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, xor);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
|
|
|
|
for (i = 0; i < type->dimx; ++i)
|
2022-11-11 17:10:14 -08:00
|
|
|
high_bit_value.u[i].u = 0x80000000;
|
|
|
|
|
if (!(high_bit = hlsl_new_constant(ctx, type, &high_bit_value, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, high_bit);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2022-11-11 17:13:26 -08:00
|
|
|
if (!(and = hlsl_new_binary_expr(ctx, HLSL_OP2_BIT_AND, xor, high_bit)))
|
2021-09-21 08:12:31 -07:00
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, and);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(abs1 = hlsl_new_unary_expr(ctx, HLSL_OP1_ABS, arg1, &instr->loc)))
|
2021-09-21 08:12:31 -07:00
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, abs1);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
|
|
|
|
if (!(cast1 = hlsl_new_cast(ctx, abs1, utype, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, cast1);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(abs2 = hlsl_new_unary_expr(ctx, HLSL_OP1_ABS, arg2, &instr->loc)))
|
2021-09-21 08:12:31 -07:00
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, abs2);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
|
|
|
|
if (!(cast2 = hlsl_new_cast(ctx, abs2, utype, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, cast2);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2022-11-10 17:39:42 -08:00
|
|
|
if (!(div = hlsl_new_binary_expr(ctx, HLSL_OP2_DIV, cast1, cast2)))
|
2021-09-21 08:12:31 -07:00
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, div);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
|
|
|
|
if (!(cast3 = hlsl_new_cast(ctx, div, type, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, cast3);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, cast3, &instr->loc)))
|
2021-09-21 08:12:31 -07:00
|
|
|
return false;
|
2023-03-06 18:34:10 -08:00
|
|
|
hlsl_block_add_instr(block, neg);
|
2021-09-21 08:12:31 -07:00
|
|
|
|
2023-03-06 18:37:41 -08:00
|
|
|
return hlsl_add_conditional(ctx, block, and, neg, cast3);
|
2021-09-21 08:12:31 -07:00
|
|
|
}
|
|
|
|
|
|
2023-03-06 18:35:59 -08:00
|
|
|
static bool lower_int_modulus(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2021-09-22 01:27:07 -07:00
|
|
|
{
|
2023-03-06 18:35:59 -08:00
|
|
|
struct hlsl_ir_node *arg1, *arg2, *and, *abs1, *abs2, *div, *neg, *cast1, *cast2, *cast3, *high_bit;
|
2021-09-22 01:27:07 -07:00
|
|
|
struct hlsl_type *type = instr->data_type, *utype;
|
2022-11-11 17:10:14 -08:00
|
|
|
struct hlsl_constant_value high_bit_value;
|
2021-09-22 01:27:07 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
arg2 = expr->operands[1].node;
|
|
|
|
|
if (expr->op != HLSL_OP2_MOD)
|
|
|
|
|
return false;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_SCALAR && type->class != HLSL_CLASS_VECTOR)
|
2021-09-22 01:27:07 -07:00
|
|
|
return false;
|
2024-02-27 15:30:51 -08:00
|
|
|
if (type->e.numeric.type != HLSL_TYPE_INT)
|
2021-09-22 01:27:07 -07:00
|
|
|
return false;
|
2022-11-11 17:31:55 -08:00
|
|
|
utype = hlsl_get_numeric_type(ctx, type->class, HLSL_TYPE_UINT, type->dimx, type->dimy);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
|
|
|
|
for (i = 0; i < type->dimx; ++i)
|
2022-11-11 17:10:14 -08:00
|
|
|
high_bit_value.u[i].u = 0x80000000;
|
|
|
|
|
if (!(high_bit = hlsl_new_constant(ctx, type, &high_bit_value, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, high_bit);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
2022-11-11 17:13:26 -08:00
|
|
|
if (!(and = hlsl_new_binary_expr(ctx, HLSL_OP2_BIT_AND, arg1, high_bit)))
|
2021-09-22 01:27:07 -07:00
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, and);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(abs1 = hlsl_new_unary_expr(ctx, HLSL_OP1_ABS, arg1, &instr->loc)))
|
2021-09-22 01:27:07 -07:00
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, abs1);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
|
|
|
|
if (!(cast1 = hlsl_new_cast(ctx, abs1, utype, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, cast1);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(abs2 = hlsl_new_unary_expr(ctx, HLSL_OP1_ABS, arg2, &instr->loc)))
|
2021-09-22 01:27:07 -07:00
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, abs2);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
|
|
|
|
if (!(cast2 = hlsl_new_cast(ctx, abs2, utype, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, cast2);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
2022-11-10 17:39:42 -08:00
|
|
|
if (!(div = hlsl_new_binary_expr(ctx, HLSL_OP2_MOD, cast1, cast2)))
|
2021-09-22 01:27:07 -07:00
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, div);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
|
|
|
|
if (!(cast3 = hlsl_new_cast(ctx, div, type, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, cast3);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, cast3, &instr->loc)))
|
2021-09-22 01:27:07 -07:00
|
|
|
return false;
|
2023-03-06 18:35:59 -08:00
|
|
|
hlsl_block_add_instr(block, neg);
|
2021-09-22 01:27:07 -07:00
|
|
|
|
2023-03-06 18:37:41 -08:00
|
|
|
return hlsl_add_conditional(ctx, block, and, neg, cast3);
|
2021-09-22 01:27:07 -07:00
|
|
|
}
|
|
|
|
|
|
2023-06-25 17:06:45 -07:00
|
|
|
static bool lower_int_abs(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2021-09-21 08:21:44 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_type *type = instr->data_type;
|
2023-06-25 17:06:45 -07:00
|
|
|
struct hlsl_ir_node *arg, *neg, *max;
|
2021-09-21 08:21:44 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
|
|
|
|
|
if (expr->op != HLSL_OP1_ABS)
|
|
|
|
|
return false;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_SCALAR && type->class != HLSL_CLASS_VECTOR)
|
2021-09-21 08:21:44 -07:00
|
|
|
return false;
|
2024-02-27 15:30:51 -08:00
|
|
|
if (type->e.numeric.type != HLSL_TYPE_INT)
|
2021-09-21 08:21:44 -07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
arg = expr->operands[0].node;
|
|
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(neg = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, arg, &instr->loc)))
|
2021-09-21 08:21:44 -07:00
|
|
|
return false;
|
2023-06-25 17:06:45 -07:00
|
|
|
hlsl_block_add_instr(block, neg);
|
2021-09-21 08:21:44 -07:00
|
|
|
|
2023-06-25 17:06:45 -07:00
|
|
|
if (!(max = hlsl_new_binary_expr(ctx, HLSL_OP2_MAX, arg, neg)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, max);
|
2021-09-21 08:21:44 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-25 16:46:10 -07:00
|
|
|
static bool lower_int_dot(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2023-05-23 22:01:38 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *arg1, *arg2, *mult, *comps[4] = {0}, *res;
|
|
|
|
|
struct hlsl_type *type = instr->data_type;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
unsigned int i, dimx;
|
2023-05-08 15:25:18 -07:00
|
|
|
bool is_bool;
|
2023-05-23 22:01:38 -07:00
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
|
|
|
|
|
if (expr->op != HLSL_OP2_DOT)
|
|
|
|
|
return false;
|
|
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
if (type->e.numeric.type == HLSL_TYPE_INT || type->e.numeric.type == HLSL_TYPE_UINT
|
|
|
|
|
|| type->e.numeric.type == HLSL_TYPE_BOOL)
|
2023-05-23 22:01:38 -07:00
|
|
|
{
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
arg2 = expr->operands[1].node;
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(arg1->data_type->dimx == arg2->data_type->dimx);
|
2023-05-23 22:01:38 -07:00
|
|
|
dimx = arg1->data_type->dimx;
|
2024-02-27 15:30:51 -08:00
|
|
|
is_bool = type->e.numeric.type == HLSL_TYPE_BOOL;
|
2023-05-23 22:01:38 -07:00
|
|
|
|
2023-05-08 15:25:18 -07:00
|
|
|
if (!(mult = hlsl_new_binary_expr(ctx, is_bool ? HLSL_OP2_LOGIC_AND : HLSL_OP2_MUL, arg1, arg2)))
|
2023-05-23 22:01:38 -07:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, mult);
|
2023-05-23 22:01:38 -07:00
|
|
|
|
|
|
|
|
for (i = 0; i < dimx; ++i)
|
|
|
|
|
{
|
2023-12-06 09:20:25 -08:00
|
|
|
uint32_t s = hlsl_swizzle_from_writemask(1 << i);
|
2023-05-23 22:01:38 -07:00
|
|
|
|
|
|
|
|
if (!(comps[i] = hlsl_new_swizzle(ctx, s, 1, mult, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, comps[i]);
|
2023-05-23 22:01:38 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
res = comps[0];
|
|
|
|
|
for (i = 1; i < dimx; ++i)
|
|
|
|
|
{
|
2023-05-08 15:25:18 -07:00
|
|
|
if (!(res = hlsl_new_binary_expr(ctx, is_bool ? HLSL_OP2_LOGIC_OR : HLSL_OP2_ADD, res, comps[i])))
|
2023-05-23 22:01:38 -07:00
|
|
|
return false;
|
2023-06-25 16:46:10 -07:00
|
|
|
hlsl_block_add_instr(block, res);
|
2023-05-23 22:01:38 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-06 18:31:16 -08:00
|
|
|
static bool lower_float_modulus(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
2022-06-14 05:32:22 -07:00
|
|
|
{
|
2023-03-06 18:31:16 -08:00
|
|
|
struct hlsl_ir_node *arg1, *arg2, *mul1, *neg1, *ge, *neg2, *div, *mul2, *frc, *cond, *one, *mul3;
|
2022-06-14 05:32:22 -07:00
|
|
|
struct hlsl_type *type = instr->data_type, *btype;
|
2022-11-11 17:10:14 -08:00
|
|
|
struct hlsl_constant_value one_value;
|
2022-06-14 05:32:22 -07:00
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
arg2 = expr->operands[1].node;
|
|
|
|
|
if (expr->op != HLSL_OP2_MOD)
|
|
|
|
|
return false;
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_SCALAR && type->class != HLSL_CLASS_VECTOR)
|
2022-06-14 05:32:22 -07:00
|
|
|
return false;
|
2024-02-27 15:30:51 -08:00
|
|
|
if (type->e.numeric.type != HLSL_TYPE_FLOAT)
|
2022-06-14 05:32:22 -07:00
|
|
|
return false;
|
2022-11-11 17:31:55 -08:00
|
|
|
btype = hlsl_get_numeric_type(ctx, type->class, HLSL_TYPE_BOOL, type->dimx, type->dimy);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
|
|
|
|
if (!(mul1 = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, arg2, arg1)))
|
|
|
|
|
return false;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, mul1);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(neg1 = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, mul1, &instr->loc)))
|
2022-06-14 05:32:22 -07:00
|
|
|
return false;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, neg1);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
|
|
|
|
if (!(ge = hlsl_new_binary_expr(ctx, HLSL_OP2_GEQUAL, mul1, neg1)))
|
|
|
|
|
return false;
|
|
|
|
|
ge->data_type = btype;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, ge);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(neg2 = hlsl_new_unary_expr(ctx, HLSL_OP1_NEG, arg2, &instr->loc)))
|
2022-06-14 05:32:22 -07:00
|
|
|
return false;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, neg2);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
2023-03-06 18:37:41 -08:00
|
|
|
if (!(cond = hlsl_add_conditional(ctx, block, ge, arg2, neg2)))
|
2022-06-14 05:32:22 -07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < type->dimx; ++i)
|
2022-11-11 17:10:14 -08:00
|
|
|
one_value.u[i].f = 1.0f;
|
|
|
|
|
if (!(one = hlsl_new_constant(ctx, type, &one_value, &instr->loc)))
|
|
|
|
|
return false;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, one);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
2022-11-11 17:13:26 -08:00
|
|
|
if (!(div = hlsl_new_binary_expr(ctx, HLSL_OP2_DIV, one, cond)))
|
2022-06-14 05:32:22 -07:00
|
|
|
return false;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, div);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
|
|
|
|
if (!(mul2 = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, div, arg1)))
|
|
|
|
|
return false;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, mul2);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
2023-04-14 00:02:14 -07:00
|
|
|
if (!(frc = hlsl_new_unary_expr(ctx, HLSL_OP1_FRACT, mul2, &instr->loc)))
|
2022-06-14 05:32:22 -07:00
|
|
|
return false;
|
2023-03-06 18:31:16 -08:00
|
|
|
hlsl_block_add_instr(block, frc);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
2023-03-06 18:31:16 -08:00
|
|
|
if (!(mul3 = hlsl_new_binary_expr(ctx, HLSL_OP2_MUL, frc, cond)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, mul3);
|
2022-06-14 05:32:22 -07:00
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-02-21 11:17:45 -08:00
|
|
|
static bool lower_nonfloat_exprs(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
expr = hlsl_ir_expr(instr);
|
2024-02-27 15:30:51 -08:00
|
|
|
if (expr->op == HLSL_OP1_CAST || instr->data_type->e.numeric.type == HLSL_TYPE_FLOAT)
|
2024-02-21 11:17:45 -08:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
switch (expr->op)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_OP1_ABS:
|
|
|
|
|
case HLSL_OP1_NEG:
|
|
|
|
|
case HLSL_OP2_ADD:
|
|
|
|
|
case HLSL_OP2_DIV:
|
2024-03-04 11:34:29 -08:00
|
|
|
case HLSL_OP2_LOGIC_AND:
|
2024-03-04 11:31:57 -08:00
|
|
|
case HLSL_OP2_LOGIC_OR:
|
2024-02-21 11:17:45 -08:00
|
|
|
case HLSL_OP2_MAX:
|
|
|
|
|
case HLSL_OP2_MIN:
|
|
|
|
|
case HLSL_OP2_MUL:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS] = {0};
|
|
|
|
|
struct hlsl_ir_node *arg, *arg_cast, *float_expr, *ret;
|
|
|
|
|
struct hlsl_type *float_type;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < HLSL_MAX_OPERANDS; ++i)
|
|
|
|
|
{
|
|
|
|
|
arg = expr->operands[i].node;
|
|
|
|
|
if (!arg)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
float_type = hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, arg->data_type->dimx);
|
|
|
|
|
if (!(arg_cast = hlsl_new_cast(ctx, arg, float_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, arg_cast);
|
|
|
|
|
|
|
|
|
|
operands[i] = arg_cast;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
float_type = hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, instr->data_type->dimx);
|
|
|
|
|
if (!(float_expr = hlsl_new_expr(ctx, expr->op, operands, float_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, float_expr);
|
|
|
|
|
|
|
|
|
|
if (!(ret = hlsl_new_cast(ctx, float_expr, instr->data_type, &instr->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, ret);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
default:
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-08 03:47:40 -07:00
|
|
|
static bool lower_discard_neg(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *zero, *bool_false, *or, *cmp, *load;
|
|
|
|
|
static const struct hlsl_constant_value zero_value;
|
|
|
|
|
struct hlsl_type *arg_type, *cmp_type;
|
|
|
|
|
struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS] = { 0 };
|
|
|
|
|
struct hlsl_ir_jump *jump;
|
2023-06-29 21:24:53 -07:00
|
|
|
struct hlsl_block block;
|
2023-06-08 03:47:40 -07:00
|
|
|
unsigned int i, count;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_JUMP)
|
|
|
|
|
return false;
|
|
|
|
|
jump = hlsl_ir_jump(instr);
|
|
|
|
|
if (jump->type != HLSL_IR_JUMP_DISCARD_NEG)
|
|
|
|
|
return false;
|
|
|
|
|
|
2023-06-29 21:24:53 -07:00
|
|
|
hlsl_block_init(&block);
|
2023-06-08 03:47:40 -07:00
|
|
|
|
|
|
|
|
arg_type = jump->condition.node->data_type;
|
|
|
|
|
if (!(zero = hlsl_new_constant(ctx, arg_type, &zero_value, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-29 21:24:53 -07:00
|
|
|
hlsl_block_add_instr(&block, zero);
|
2023-06-08 03:47:40 -07:00
|
|
|
|
|
|
|
|
operands[0] = jump->condition.node;
|
|
|
|
|
operands[1] = zero;
|
|
|
|
|
cmp_type = hlsl_get_numeric_type(ctx, arg_type->class, HLSL_TYPE_BOOL, arg_type->dimx, arg_type->dimy);
|
|
|
|
|
if (!(cmp = hlsl_new_expr(ctx, HLSL_OP2_LESS, operands, cmp_type, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-29 21:24:53 -07:00
|
|
|
hlsl_block_add_instr(&block, cmp);
|
2023-06-08 03:47:40 -07:00
|
|
|
|
|
|
|
|
if (!(bool_false = hlsl_new_constant(ctx, hlsl_get_scalar_type(ctx, HLSL_TYPE_BOOL), &zero_value, &instr->loc)))
|
|
|
|
|
return false;
|
2023-06-29 21:24:53 -07:00
|
|
|
hlsl_block_add_instr(&block, bool_false);
|
2023-06-08 03:47:40 -07:00
|
|
|
|
|
|
|
|
or = bool_false;
|
|
|
|
|
|
|
|
|
|
count = hlsl_type_component_count(cmp_type);
|
|
|
|
|
for (i = 0; i < count; ++i)
|
|
|
|
|
{
|
2022-11-14 18:44:44 -08:00
|
|
|
if (!(load = hlsl_add_load_component(ctx, &block, cmp, i, &instr->loc)))
|
2023-06-08 03:47:40 -07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!(or = hlsl_new_binary_expr(ctx, HLSL_OP2_LOGIC_OR, or, load)))
|
|
|
|
|
return NULL;
|
2023-06-29 21:24:53 -07:00
|
|
|
hlsl_block_add_instr(&block, or);
|
2023-06-08 03:47:40 -07:00
|
|
|
}
|
|
|
|
|
|
2023-06-29 21:24:53 -07:00
|
|
|
list_move_tail(&instr->entry, &block.instrs);
|
2023-06-08 03:47:40 -07:00
|
|
|
hlsl_src_remove(&jump->condition);
|
|
|
|
|
hlsl_src_from_node(&jump->condition, or);
|
|
|
|
|
jump->type = HLSL_IR_JUMP_DISCARD_NZ;
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2021-03-16 14:31:54 -07:00
|
|
|
static bool dce(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
|
|
|
|
{
|
|
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_IR_CONSTANT:
|
2024-06-24 14:30:46 -07:00
|
|
|
case HLSL_IR_COMPILE:
|
2021-03-16 14:31:54 -07:00
|
|
|
case HLSL_IR_EXPR:
|
2023-02-24 11:39:56 -08:00
|
|
|
case HLSL_IR_INDEX:
|
2021-03-16 14:31:54 -07:00
|
|
|
case HLSL_IR_LOAD:
|
2021-10-07 19:58:57 -07:00
|
|
|
case HLSL_IR_RESOURCE_LOAD:
|
2024-06-14 16:59:21 -07:00
|
|
|
case HLSL_IR_STRING_CONSTANT:
|
2021-03-16 14:31:54 -07:00
|
|
|
case HLSL_IR_SWIZZLE:
|
2024-08-26 13:43:20 -07:00
|
|
|
case HLSL_IR_SAMPLER_STATE:
|
2021-03-16 14:31:54 -07:00
|
|
|
if (list_empty(&instr->uses))
|
|
|
|
|
{
|
|
|
|
|
list_remove(&instr->entry);
|
|
|
|
|
hlsl_free_instr(instr);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
2021-04-08 21:38:22 -07:00
|
|
|
case HLSL_IR_STORE:
|
2021-03-17 22:22:22 -07:00
|
|
|
{
|
2021-04-08 21:38:22 -07:00
|
|
|
struct hlsl_ir_store *store = hlsl_ir_store(instr);
|
|
|
|
|
struct hlsl_ir_var *var = store->lhs.var;
|
2021-03-17 22:22:22 -07:00
|
|
|
|
|
|
|
|
if (var->last_read < instr->index)
|
|
|
|
|
{
|
|
|
|
|
list_remove(&instr->entry);
|
|
|
|
|
hlsl_free_instr(instr);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2021-09-09 19:06:38 -07:00
|
|
|
case HLSL_IR_CALL:
|
2021-03-16 14:31:54 -07:00
|
|
|
case HLSL_IR_IF:
|
|
|
|
|
case HLSL_IR_JUMP:
|
|
|
|
|
case HLSL_IR_LOOP:
|
2021-08-15 10:08:32 -07:00
|
|
|
case HLSL_IR_RESOURCE_STORE:
|
2023-10-11 04:51:51 -07:00
|
|
|
case HLSL_IR_SWITCH:
|
2021-03-16 14:31:54 -07:00
|
|
|
break;
|
2024-03-18 11:31:04 -07:00
|
|
|
case HLSL_IR_STATEBLOCK_CONSTANT:
|
|
|
|
|
/* Stateblock constants should not appear in the shader program. */
|
|
|
|
|
vkd3d_unreachable();
|
2021-03-16 14:31:54 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2021-03-02 13:34:46 -08:00
|
|
|
static void dump_function(struct rb_entry *entry, void *context)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_function *func = RB_ENTRY_VALUE(entry, struct hlsl_ir_function, entry);
|
vkd3d-shader/hlsl: Store function overloads in a list.
The choice to store them in an rbtree was made early on. It does not seem likely
that HLSL programs would define many overloads for any of their functions, but I
suspect the idea was rather that intrinsics would be defined as plain
hlsl_ir_function_decl structures [cf. 447463e5900ca6a636998a65429b8a08a5441657]
and that some intrinsics that could operate on any type would therefore need
many overrides.
This is not how we deal with intrinsics, however. When the first intrinsics were
implemented I made the choice disregard this intended design, and instead match
and convert their types manually, in C. Nothing that has happened in the time
since has led me to question that choice, and in fact, the flexibility with
which we must accommodate functions has led me to believe that matching in this
way was definitely the right choice. The main other designs I see would have
been:
* define each intrinsic variant separately using existing HLSL types. Besides
efficiency concerns (i.e. this would take more space in memory, and would take
longer to generate each variant), the normal type-matching rules don't really
apply to intrinsics.
[For example: elementwise intrinsics like abs() return the same type as the
input, including preserving the distinction between float and float1. It is
legal to define separate HLSL overloads taking float and float1, but trying to
invoke these functions yields an "ambiguous function call" error.]
* introduce new (semi-)generic types. This is far more code and ends up acting
like our current scheme (with helpers) in a slightly more complex form.
So I think we can go ahead and rip out this vestige of the original design for
intrinsics.
As for why to change it: rbtrees are simply more complex to deal with, and it
seems unlikely to me that the difference is going to matter. I do not expect any
program to define large quantities of intrinsics; linked list search should be
good enough.
2023-09-08 14:27:10 -07:00
|
|
|
struct hlsl_ir_function_decl *decl;
|
2021-05-20 22:32:22 -07:00
|
|
|
struct hlsl_ctx *ctx = context;
|
|
|
|
|
|
vkd3d-shader/hlsl: Store function overloads in a list.
The choice to store them in an rbtree was made early on. It does not seem likely
that HLSL programs would define many overloads for any of their functions, but I
suspect the idea was rather that intrinsics would be defined as plain
hlsl_ir_function_decl structures [cf. 447463e5900ca6a636998a65429b8a08a5441657]
and that some intrinsics that could operate on any type would therefore need
many overrides.
This is not how we deal with intrinsics, however. When the first intrinsics were
implemented I made the choice disregard this intended design, and instead match
and convert their types manually, in C. Nothing that has happened in the time
since has led me to question that choice, and in fact, the flexibility with
which we must accommodate functions has led me to believe that matching in this
way was definitely the right choice. The main other designs I see would have
been:
* define each intrinsic variant separately using existing HLSL types. Besides
efficiency concerns (i.e. this would take more space in memory, and would take
longer to generate each variant), the normal type-matching rules don't really
apply to intrinsics.
[For example: elementwise intrinsics like abs() return the same type as the
input, including preserving the distinction between float and float1. It is
legal to define separate HLSL overloads taking float and float1, but trying to
invoke these functions yields an "ambiguous function call" error.]
* introduce new (semi-)generic types. This is far more code and ends up acting
like our current scheme (with helpers) in a slightly more complex form.
So I think we can go ahead and rip out this vestige of the original design for
intrinsics.
As for why to change it: rbtrees are simply more complex to deal with, and it
seems unlikely to me that the difference is going to matter. I do not expect any
program to define large quantities of intrinsics; linked list search should be
good enough.
2023-09-08 14:27:10 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(decl, &func->overloads, struct hlsl_ir_function_decl, entry)
|
|
|
|
|
{
|
|
|
|
|
if (decl->has_body)
|
|
|
|
|
hlsl_dump_function(ctx, decl);
|
|
|
|
|
}
|
2021-03-02 13:34:46 -08:00
|
|
|
}
|
|
|
|
|
|
2023-10-04 13:28:02 -07:00
|
|
|
static bool mark_indexable_vars(struct hlsl_ctx *ctx, struct hlsl_deref *deref,
|
|
|
|
|
struct hlsl_ir_node *instr)
|
|
|
|
|
{
|
|
|
|
|
if (!deref->rel_offset.node)
|
|
|
|
|
return false;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(deref->var);
|
|
|
|
|
VKD3D_ASSERT(deref->rel_offset.node->type != HLSL_IR_CONSTANT);
|
2023-10-04 13:28:02 -07:00
|
|
|
deref->var->indexable = true;
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-30 13:27:32 -08:00
|
|
|
static char get_regset_name(enum hlsl_regset regset)
|
|
|
|
|
{
|
|
|
|
|
switch (regset)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_REGSET_SAMPLERS:
|
|
|
|
|
return 's';
|
|
|
|
|
case HLSL_REGSET_TEXTURES:
|
|
|
|
|
return 't';
|
|
|
|
|
case HLSL_REGSET_UAVS:
|
|
|
|
|
return 'u';
|
|
|
|
|
case HLSL_REGSET_NUMERIC:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
|
|
|
|
|
2022-11-24 14:36:20 -08:00
|
|
|
static void allocate_register_reservations(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
2023-08-29 10:40:38 -07:00
|
|
|
const struct hlsl_reg_reservation *reservation = &var->reg_reservation;
|
2023-06-09 22:59:54 -07:00
|
|
|
unsigned int r;
|
2022-11-24 14:36:20 -08:00
|
|
|
|
2023-08-29 10:40:38 -07:00
|
|
|
if (reservation->reg_type)
|
2022-11-24 14:36:20 -08:00
|
|
|
{
|
2023-06-09 22:59:54 -07:00
|
|
|
for (r = 0; r <= HLSL_REGSET_LAST_OBJECT; ++r)
|
2022-11-24 14:36:20 -08:00
|
|
|
{
|
2023-06-09 22:59:54 -07:00
|
|
|
if (var->regs[r].allocation_size > 0)
|
|
|
|
|
{
|
2023-08-29 10:40:38 -07:00
|
|
|
if (reservation->reg_type != get_regset_name(r))
|
2023-06-09 22:59:54 -07:00
|
|
|
{
|
|
|
|
|
struct vkd3d_string_buffer *type_string;
|
|
|
|
|
|
|
|
|
|
/* We can throw this error because resources can only span across a single
|
|
|
|
|
* regset, but we have to check for multiple regsets if we support register
|
|
|
|
|
* reservations for structs for SM5. */
|
|
|
|
|
type_string = hlsl_type_to_string(ctx, var->data_type);
|
|
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"Object of type '%s' must be bound to register type '%c'.",
|
|
|
|
|
type_string->buffer, get_regset_name(r));
|
|
|
|
|
hlsl_release_string_buffer(ctx, type_string);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
var->regs[r].allocated = true;
|
2023-08-29 10:40:38 -07:00
|
|
|
var->regs[r].space = reservation->reg_space;
|
|
|
|
|
var->regs[r].index = reservation->reg_index;
|
2023-06-09 22:59:54 -07:00
|
|
|
}
|
|
|
|
|
}
|
2022-11-24 14:36:20 -08:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-03-25 12:33:38 -07:00
|
|
|
static void deref_mark_last_read(struct hlsl_deref *deref, unsigned int last_read)
|
|
|
|
|
{
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
if (hlsl_deref_is_lowered(deref))
|
|
|
|
|
{
|
|
|
|
|
if (deref->rel_offset.node)
|
|
|
|
|
deref->rel_offset.node->last_read = last_read;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
for (i = 0; i < deref->path_len; ++i)
|
|
|
|
|
deref->path[i].node->last_read = last_read;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2021-03-02 13:34:46 -08:00
|
|
|
/* Compute the earliest and latest liveness for each variable. In the case that
|
|
|
|
|
* a variable is accessed inside of a loop, we promote its liveness to extend
|
2023-05-10 09:30:39 -07:00
|
|
|
* to at least the range of the entire loop. We also do this for nodes, so that
|
|
|
|
|
* nodes produced before the loop have their temp register protected from being
|
|
|
|
|
* overridden after the last read within an iteration. */
|
2021-10-15 14:54:10 -07:00
|
|
|
static void compute_liveness_recurse(struct hlsl_block *block, unsigned int loop_first, unsigned int loop_last)
|
2021-03-02 13:34:46 -08:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr;
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
2021-10-15 14:54:10 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(instr, &block->instrs, struct hlsl_ir_node, entry)
|
2021-03-02 13:34:46 -08:00
|
|
|
{
|
2023-05-10 09:30:39 -07:00
|
|
|
const unsigned int last_read = loop_last ? max(instr->index, loop_last) : instr->index;
|
2021-10-15 14:54:07 -07:00
|
|
|
|
2021-03-02 13:34:46 -08:00
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
2021-09-09 19:06:38 -07:00
|
|
|
case HLSL_IR_CALL:
|
2021-09-11 14:56:04 -07:00
|
|
|
/* We should have inlined all calls before computing liveness. */
|
|
|
|
|
vkd3d_unreachable();
|
2024-03-18 11:31:04 -07:00
|
|
|
case HLSL_IR_STATEBLOCK_CONSTANT:
|
|
|
|
|
/* Stateblock constants should not appear in the shader program. */
|
|
|
|
|
vkd3d_unreachable();
|
2021-09-09 19:06:38 -07:00
|
|
|
|
2021-04-08 21:38:22 -07:00
|
|
|
case HLSL_IR_STORE:
|
2021-03-02 13:34:46 -08:00
|
|
|
{
|
2021-04-08 21:38:22 -07:00
|
|
|
struct hlsl_ir_store *store = hlsl_ir_store(instr);
|
2021-03-02 13:34:46 -08:00
|
|
|
|
2021-04-08 21:38:22 -07:00
|
|
|
var = store->lhs.var;
|
2021-03-02 13:34:46 -08:00
|
|
|
if (!var->first_write)
|
|
|
|
|
var->first_write = loop_first ? min(instr->index, loop_first) : instr->index;
|
2023-05-10 09:30:39 -07:00
|
|
|
store->rhs.node->last_read = last_read;
|
2024-03-25 12:33:38 -07:00
|
|
|
deref_mark_last_read(&store->lhs, last_read);
|
2021-03-02 13:34:46 -08:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case HLSL_IR_EXPR:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_expr *expr = hlsl_ir_expr(instr);
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(expr->operands) && expr->operands[i].node; ++i)
|
2023-05-10 09:30:39 -07:00
|
|
|
expr->operands[i].node->last_read = last_read;
|
2021-03-02 13:34:46 -08:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case HLSL_IR_IF:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
|
|
|
|
|
2022-11-10 18:04:22 -08:00
|
|
|
compute_liveness_recurse(&iff->then_block, loop_first, loop_last);
|
|
|
|
|
compute_liveness_recurse(&iff->else_block, loop_first, loop_last);
|
2023-05-10 09:30:39 -07:00
|
|
|
iff->condition.node->last_read = last_read;
|
2021-03-02 13:34:46 -08:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case HLSL_IR_LOAD:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_load *load = hlsl_ir_load(instr);
|
|
|
|
|
|
|
|
|
|
var = load->src.var;
|
2023-05-10 09:30:39 -07:00
|
|
|
var->last_read = max(var->last_read, last_read);
|
2024-03-25 12:33:38 -07:00
|
|
|
deref_mark_last_read(&load->src, last_read);
|
2021-03-02 13:34:46 -08:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case HLSL_IR_LOOP:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_loop *loop = hlsl_ir_loop(instr);
|
|
|
|
|
|
|
|
|
|
compute_liveness_recurse(&loop->body, loop_first ? loop_first : instr->index,
|
|
|
|
|
loop_last ? loop_last : loop->next_index);
|
|
|
|
|
break;
|
|
|
|
|
}
|
2021-10-07 19:58:57 -07:00
|
|
|
case HLSL_IR_RESOURCE_LOAD:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_resource_load *load = hlsl_ir_resource_load(instr);
|
|
|
|
|
|
2021-10-15 14:54:07 -07:00
|
|
|
var = load->resource.var;
|
2023-05-10 09:30:39 -07:00
|
|
|
var->last_read = max(var->last_read, last_read);
|
2024-03-25 12:33:38 -07:00
|
|
|
deref_mark_last_read(&load->resource, last_read);
|
2021-11-05 11:35:52 -07:00
|
|
|
|
|
|
|
|
if ((var = load->sampler.var))
|
|
|
|
|
{
|
2023-05-10 09:30:39 -07:00
|
|
|
var->last_read = max(var->last_read, last_read);
|
2024-03-25 12:33:38 -07:00
|
|
|
deref_mark_last_read(&load->sampler, last_read);
|
2021-11-05 11:35:52 -07:00
|
|
|
}
|
|
|
|
|
|
2023-06-07 10:56:02 -07:00
|
|
|
if (load->coords.node)
|
|
|
|
|
load->coords.node->last_read = last_read;
|
2022-01-26 06:35:29 -08:00
|
|
|
if (load->texel_offset.node)
|
2023-05-10 09:30:39 -07:00
|
|
|
load->texel_offset.node->last_read = last_read;
|
2021-08-16 18:28:47 -07:00
|
|
|
if (load->lod.node)
|
2023-05-10 09:30:39 -07:00
|
|
|
load->lod.node->last_read = last_read;
|
2023-05-05 08:13:18 -07:00
|
|
|
if (load->ddx.node)
|
2023-05-10 09:30:39 -07:00
|
|
|
load->ddx.node->last_read = last_read;
|
2023-05-05 08:13:18 -07:00
|
|
|
if (load->ddy.node)
|
2023-05-10 09:30:39 -07:00
|
|
|
load->ddy.node->last_read = last_read;
|
2023-04-27 01:15:36 -07:00
|
|
|
if (load->sample_index.node)
|
2023-05-10 09:30:39 -07:00
|
|
|
load->sample_index.node->last_read = last_read;
|
2023-05-16 11:54:22 -07:00
|
|
|
if (load->cmp.node)
|
|
|
|
|
load->cmp.node->last_read = last_read;
|
2021-10-07 19:58:57 -07:00
|
|
|
break;
|
|
|
|
|
}
|
2021-08-15 10:08:32 -07:00
|
|
|
case HLSL_IR_RESOURCE_STORE:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_resource_store *store = hlsl_ir_resource_store(instr);
|
|
|
|
|
|
|
|
|
|
var = store->resource.var;
|
2023-05-10 09:30:39 -07:00
|
|
|
var->last_read = max(var->last_read, last_read);
|
2024-03-25 12:33:38 -07:00
|
|
|
deref_mark_last_read(&store->resource, last_read);
|
2023-05-10 09:30:39 -07:00
|
|
|
store->coords.node->last_read = last_read;
|
|
|
|
|
store->value.node->last_read = last_read;
|
2021-08-15 10:08:32 -07:00
|
|
|
break;
|
|
|
|
|
}
|
2021-03-02 13:34:46 -08:00
|
|
|
case HLSL_IR_SWIZZLE:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_swizzle *swizzle = hlsl_ir_swizzle(instr);
|
|
|
|
|
|
2023-05-10 09:30:39 -07:00
|
|
|
swizzle->val.node->last_read = last_read;
|
2021-03-02 13:34:46 -08:00
|
|
|
break;
|
|
|
|
|
}
|
2023-02-24 11:39:56 -08:00
|
|
|
case HLSL_IR_INDEX:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_index *index = hlsl_ir_index(instr);
|
|
|
|
|
|
2023-05-10 09:30:39 -07:00
|
|
|
index->val.node->last_read = last_read;
|
|
|
|
|
index->idx.node->last_read = last_read;
|
2023-02-24 11:39:56 -08:00
|
|
|
break;
|
|
|
|
|
}
|
2021-03-02 13:34:46 -08:00
|
|
|
case HLSL_IR_JUMP:
|
2023-06-08 00:42:58 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_jump *jump = hlsl_ir_jump(instr);
|
|
|
|
|
|
|
|
|
|
if (jump->condition.node)
|
|
|
|
|
jump->condition.node->last_read = last_read;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2023-10-11 04:51:51 -07:00
|
|
|
case HLSL_IR_SWITCH:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
compute_liveness_recurse(&c->body, loop_first, loop_last);
|
|
|
|
|
s->selector.node->last_read = last_read;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2023-06-08 00:42:58 -07:00
|
|
|
case HLSL_IR_CONSTANT:
|
2024-06-14 16:59:21 -07:00
|
|
|
case HLSL_IR_STRING_CONSTANT:
|
2021-03-02 13:34:46 -08:00
|
|
|
break;
|
2024-06-24 14:30:46 -07:00
|
|
|
case HLSL_IR_COMPILE:
|
2024-08-26 13:43:20 -07:00
|
|
|
case HLSL_IR_SAMPLER_STATE:
|
|
|
|
|
/* These types are skipped as they are only relevant to effects. */
|
2024-06-24 14:30:46 -07:00
|
|
|
break;
|
2021-03-02 13:34:46 -08:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void compute_liveness(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry_func)
|
|
|
|
|
{
|
2021-03-17 22:22:22 -07:00
|
|
|
struct hlsl_scope *scope;
|
2021-03-02 13:34:46 -08:00
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
2021-10-15 14:54:09 -07:00
|
|
|
index_instructions(&entry_func->body, 2);
|
2021-03-17 22:22:22 -07:00
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(scope, &ctx->scopes, struct hlsl_scope, entry)
|
|
|
|
|
{
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &scope->vars, struct hlsl_ir_var, scope_entry)
|
|
|
|
|
var->first_write = var->last_read = 0;
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-15 17:03:44 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
2021-03-02 13:34:46 -08:00
|
|
|
{
|
2021-04-27 10:14:20 -07:00
|
|
|
if (var->is_uniform || var->is_input_semantic)
|
2021-03-22 15:02:40 -07:00
|
|
|
var->first_write = 1;
|
2021-04-27 10:14:20 -07:00
|
|
|
else if (var->is_output_semantic)
|
2021-03-28 12:46:59 -07:00
|
|
|
var->last_read = UINT_MAX;
|
2021-03-02 13:34:46 -08:00
|
|
|
}
|
|
|
|
|
|
2021-10-15 14:54:09 -07:00
|
|
|
compute_liveness_recurse(&entry_func->body, 0, 0);
|
2021-03-02 13:34:46 -08:00
|
|
|
}
|
|
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
struct register_allocator
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 12:09:16 -07:00
|
|
|
struct allocation
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 12:09:16 -07:00
|
|
|
uint32_t reg;
|
|
|
|
|
unsigned int writemask;
|
|
|
|
|
unsigned int first_write, last_read;
|
|
|
|
|
} *allocations;
|
2023-10-27 15:06:51 -07:00
|
|
|
size_t count, capacity;
|
2023-10-04 14:03:14 -07:00
|
|
|
|
|
|
|
|
/* Indexable temps are allocated separately and always keep their index regardless of their
|
|
|
|
|
* lifetime. */
|
|
|
|
|
size_t indexable_count;
|
2023-10-27 15:06:51 -07:00
|
|
|
|
|
|
|
|
/* Total number of registers allocated so far. Used to declare sm4 temp count. */
|
|
|
|
|
uint32_t reg_count;
|
2021-04-08 21:38:23 -07:00
|
|
|
};
|
|
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
static unsigned int get_available_writemask(const struct register_allocator *allocator,
|
|
|
|
|
unsigned int first_write, unsigned int last_read, uint32_t reg_idx)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 12:09:16 -07:00
|
|
|
unsigned int writemask = VKD3DSP_WRITEMASK_ALL;
|
|
|
|
|
size_t i;
|
2021-04-08 21:38:23 -07:00
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
for (i = 0; i < allocator->count; ++i)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 12:09:16 -07:00
|
|
|
const struct allocation *allocation = &allocator->allocations[i];
|
|
|
|
|
|
|
|
|
|
/* We do not overlap if first write == last read:
|
|
|
|
|
* this is the case where we are allocating the result of that
|
|
|
|
|
* expression, e.g. "add r0, r0, r1". */
|
|
|
|
|
|
|
|
|
|
if (allocation->reg == reg_idx
|
|
|
|
|
&& first_write < allocation->last_read && last_read > allocation->first_write)
|
|
|
|
|
writemask &= ~allocation->writemask;
|
|
|
|
|
|
|
|
|
|
if (!writemask)
|
|
|
|
|
break;
|
2021-04-08 21:38:23 -07:00
|
|
|
}
|
|
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
return writemask;
|
2021-04-08 21:38:23 -07:00
|
|
|
}
|
|
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
static void record_allocation(struct hlsl_ctx *ctx, struct register_allocator *allocator,
|
|
|
|
|
uint32_t reg_idx, unsigned int writemask, unsigned int first_write, unsigned int last_read)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 12:09:16 -07:00
|
|
|
struct allocation *allocation;
|
2021-04-08 21:38:23 -07:00
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
if (!hlsl_array_reserve(ctx, (void **)&allocator->allocations, &allocator->capacity,
|
|
|
|
|
allocator->count + 1, sizeof(*allocator->allocations)))
|
|
|
|
|
return;
|
2021-04-08 21:38:23 -07:00
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
allocation = &allocator->allocations[allocator->count++];
|
|
|
|
|
allocation->reg = reg_idx;
|
|
|
|
|
allocation->writemask = writemask;
|
|
|
|
|
allocation->first_write = first_write;
|
|
|
|
|
allocation->last_read = last_read;
|
|
|
|
|
|
2023-10-27 15:06:51 -07:00
|
|
|
allocator->reg_count = max(allocator->reg_count, reg_idx + 1);
|
2021-04-08 21:38:23 -07:00
|
|
|
}
|
|
|
|
|
|
2023-02-06 08:17:32 -08:00
|
|
|
/* reg_size is the number of register components to be reserved, while component_count is the number
|
|
|
|
|
* of components for the register's writemask. In SM1, floats and vectors allocate the whole
|
|
|
|
|
* register, even if they don't use it completely. */
|
2023-04-05 13:15:37 -07:00
|
|
|
static struct hlsl_reg allocate_register(struct hlsl_ctx *ctx, struct register_allocator *allocator,
|
2023-02-06 08:17:32 -08:00
|
|
|
unsigned int first_write, unsigned int last_read, unsigned int reg_size,
|
|
|
|
|
unsigned int component_count)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_reg ret = {0};
|
2023-04-05 12:09:16 -07:00
|
|
|
unsigned int writemask;
|
|
|
|
|
uint32_t reg_idx;
|
2021-04-08 21:38:23 -07:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(component_count <= reg_size);
|
2023-02-06 08:17:32 -08:00
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
for (reg_idx = 0;; ++reg_idx)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 12:09:16 -07:00
|
|
|
writemask = get_available_writemask(allocator, first_write, last_read, reg_idx);
|
|
|
|
|
|
|
|
|
|
if (vkd3d_popcount(writemask) >= reg_size)
|
|
|
|
|
{
|
|
|
|
|
writemask = hlsl_combine_writemasks(writemask, (1u << reg_size) - 1);
|
2021-04-08 21:38:23 -07:00
|
|
|
break;
|
2023-04-05 12:09:16 -07:00
|
|
|
}
|
2021-04-08 21:38:23 -07:00
|
|
|
}
|
2023-04-05 12:09:16 -07:00
|
|
|
|
|
|
|
|
record_allocation(ctx, allocator, reg_idx, writemask, first_write, last_read);
|
|
|
|
|
|
|
|
|
|
ret.id = reg_idx;
|
vkd3d-shader/hlsl: Rename hlsl_reg.bind_count to hlsl_reg.allocation_size.
We have to distinguish between the "bind count" and the "allocation size"
of variables.
The "allocation size" affects the starting register id for the resource to
be allocated next, while the "bind count" is determined by the last field
actually used. The former may be larger than the latter.
What we are currently calling hlsl_reg.bind_count is actually the
"allocation size", so a rename is in order.
The real "bind count", which will be introduced in following patches,
is important because it is what should be shown in the RDEF table and
some resource allocation rules depend on it.
For instance, for this shader:
texture2D texs[3];
texture2D tex;
float4 main() : sv_target
{
return texs[0].Load(int3(0, 0, 0)) + tex.Load(int3(0, 0, 0));
}
the variable "texs" has a "bind count" of 1, but an "allocation size" of
3:
// Resource Bindings:
//
// Name Type Format Dim HLSL Bind Count
// ------------------------------ ---------- ------- ----------- -------------- ------
// texs texture float4 2d t0 1
// tex texture float4 2d t3 1
2023-08-04 10:21:27 -07:00
|
|
|
ret.allocation_size = 1;
|
2023-02-06 08:17:32 -08:00
|
|
|
ret.writemask = hlsl_combine_writemasks(writemask, (1u << component_count) - 1);
|
2021-04-08 21:38:23 -07:00
|
|
|
ret.allocated = true;
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-20 16:11:44 -07:00
|
|
|
/* Allocate a register with writemask, while reserving reg_writemask. */
|
|
|
|
|
static struct hlsl_reg allocate_register_with_masks(struct hlsl_ctx *ctx, struct register_allocator *allocator,
|
|
|
|
|
unsigned int first_write, unsigned int last_read, uint32_t reg_writemask, uint32_t writemask)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_reg ret = {0};
|
|
|
|
|
uint32_t reg_idx;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT((reg_writemask & writemask) == writemask);
|
2024-06-20 16:11:44 -07:00
|
|
|
|
|
|
|
|
for (reg_idx = 0;; ++reg_idx)
|
|
|
|
|
{
|
|
|
|
|
if ((get_available_writemask(allocator, first_write, last_read, reg_idx) & reg_writemask) == reg_writemask)
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
record_allocation(ctx, allocator, reg_idx, reg_writemask, first_write, last_read);
|
|
|
|
|
|
|
|
|
|
ret.id = reg_idx;
|
|
|
|
|
ret.allocation_size = 1;
|
|
|
|
|
ret.writemask = writemask;
|
|
|
|
|
ret.allocated = true;
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
static bool is_range_available(const struct register_allocator *allocator,
|
|
|
|
|
unsigned int first_write, unsigned int last_read, uint32_t reg_idx, unsigned int reg_size)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-10-27 08:27:37 -07:00
|
|
|
unsigned int last_reg_mask = (1u << (reg_size % 4)) - 1;
|
|
|
|
|
unsigned int writemask;
|
2023-04-05 12:09:16 -07:00
|
|
|
uint32_t i;
|
2021-04-08 21:38:23 -07:00
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
for (i = 0; i < (reg_size / 4); ++i)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-10-27 08:27:37 -07:00
|
|
|
writemask = get_available_writemask(allocator, first_write, last_read, reg_idx + i);
|
|
|
|
|
if (writemask != VKD3DSP_WRITEMASK_ALL)
|
2021-04-08 21:38:23 -07:00
|
|
|
return false;
|
|
|
|
|
}
|
2023-10-27 08:27:37 -07:00
|
|
|
writemask = get_available_writemask(allocator, first_write, last_read, reg_idx + (reg_size / 4));
|
|
|
|
|
if ((writemask & last_reg_mask) != last_reg_mask)
|
|
|
|
|
return false;
|
2021-04-08 21:38:23 -07:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
static struct hlsl_reg allocate_range(struct hlsl_ctx *ctx, struct register_allocator *allocator,
|
2023-02-06 07:05:32 -08:00
|
|
|
unsigned int first_write, unsigned int last_read, unsigned int reg_size)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_reg ret = {0};
|
2023-04-05 12:09:16 -07:00
|
|
|
uint32_t reg_idx;
|
|
|
|
|
unsigned int i;
|
2021-04-08 21:38:23 -07:00
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
for (reg_idx = 0;; ++reg_idx)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 12:09:16 -07:00
|
|
|
if (is_range_available(allocator, first_write, last_read, reg_idx, reg_size))
|
2021-04-08 21:38:23 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
for (i = 0; i < reg_size / 4; ++i)
|
|
|
|
|
record_allocation(ctx, allocator, reg_idx + i, VKD3DSP_WRITEMASK_ALL, first_write, last_read);
|
2023-10-27 08:28:53 -07:00
|
|
|
if (reg_size % 4)
|
|
|
|
|
record_allocation(ctx, allocator, reg_idx + (reg_size / 4), (1u << (reg_size % 4)) - 1, first_write, last_read);
|
2023-04-05 12:09:16 -07:00
|
|
|
|
|
|
|
|
ret.id = reg_idx;
|
vkd3d-shader/hlsl: Rename hlsl_reg.bind_count to hlsl_reg.allocation_size.
We have to distinguish between the "bind count" and the "allocation size"
of variables.
The "allocation size" affects the starting register id for the resource to
be allocated next, while the "bind count" is determined by the last field
actually used. The former may be larger than the latter.
What we are currently calling hlsl_reg.bind_count is actually the
"allocation size", so a rename is in order.
The real "bind count", which will be introduced in following patches,
is important because it is what should be shown in the RDEF table and
some resource allocation rules depend on it.
For instance, for this shader:
texture2D texs[3];
texture2D tex;
float4 main() : sv_target
{
return texs[0].Load(int3(0, 0, 0)) + tex.Load(int3(0, 0, 0));
}
the variable "texs" has a "bind count" of 1, but an "allocation size" of
3:
// Resource Bindings:
//
// Name Type Format Dim HLSL Bind Count
// ------------------------------ ---------- ------- ----------- -------------- ------
// texs texture float4 2d t0 1
// tex texture float4 2d t3 1
2023-08-04 10:21:27 -07:00
|
|
|
ret.allocation_size = align(reg_size, 4) / 4;
|
2021-04-08 21:38:23 -07:00
|
|
|
ret.allocated = true;
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
static struct hlsl_reg allocate_numeric_registers_for_type(struct hlsl_ctx *ctx, struct register_allocator *allocator,
|
2023-02-02 09:41:13 -08:00
|
|
|
unsigned int first_write, unsigned int last_read, const struct hlsl_type *type)
|
|
|
|
|
{
|
2022-10-28 08:23:05 -07:00
|
|
|
unsigned int reg_size = type->reg_size[HLSL_REGSET_NUMERIC];
|
|
|
|
|
|
2024-06-17 14:42:11 -07:00
|
|
|
/* FIXME: We could potentially pack structs or arrays more efficiently... */
|
|
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class <= HLSL_CLASS_VECTOR)
|
2024-06-17 14:42:11 -07:00
|
|
|
return allocate_register(ctx, allocator, first_write, last_read, type->dimx, type->dimx);
|
2023-02-02 09:41:13 -08:00
|
|
|
else
|
2023-04-05 13:15:37 -07:00
|
|
|
return allocate_range(ctx, allocator, first_write, last_read, reg_size);
|
2023-02-02 09:41:13 -08:00
|
|
|
}
|
|
|
|
|
|
2021-04-08 21:38:23 -07:00
|
|
|
static const char *debug_register(char class, struct hlsl_reg reg, const struct hlsl_type *type)
|
|
|
|
|
{
|
2021-06-23 21:57:34 -07:00
|
|
|
static const char writemask_offset[] = {'w','x','y','z'};
|
2022-10-28 08:23:05 -07:00
|
|
|
unsigned int reg_size = type->reg_size[HLSL_REGSET_NUMERIC];
|
2021-06-23 21:57:34 -07:00
|
|
|
|
2022-10-28 08:23:05 -07:00
|
|
|
if (reg_size > 4)
|
2021-06-23 21:57:34 -07:00
|
|
|
{
|
2022-10-28 08:23:05 -07:00
|
|
|
if (reg_size & 3)
|
|
|
|
|
return vkd3d_dbg_sprintf("%c%u-%c%u.%c", class, reg.id, class, reg.id + (reg_size / 4),
|
|
|
|
|
writemask_offset[reg_size & 3]);
|
2021-06-23 21:57:34 -07:00
|
|
|
|
2022-10-28 08:23:05 -07:00
|
|
|
return vkd3d_dbg_sprintf("%c%u-%c%u", class, reg.id, class, reg.id + (reg_size / 4) - 1);
|
2021-06-23 21:57:34 -07:00
|
|
|
}
|
2021-04-08 21:38:23 -07:00
|
|
|
return vkd3d_dbg_sprintf("%c%u%s", class, reg.id, debug_hlsl_writemask(reg.writemask));
|
|
|
|
|
}
|
|
|
|
|
|
2023-05-25 14:38:31 -07:00
|
|
|
static bool track_object_components_sampler_dim(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
2022-11-25 14:47:56 -08:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_resource_load *load;
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
enum hlsl_regset regset;
|
|
|
|
|
unsigned int index;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_RESOURCE_LOAD)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
load = hlsl_ir_resource_load(instr);
|
|
|
|
|
var = load->resource.var;
|
2023-05-25 14:38:31 -07:00
|
|
|
|
2023-06-12 08:58:07 -07:00
|
|
|
regset = hlsl_deref_get_regset(ctx, &load->resource);
|
2023-05-25 14:38:31 -07:00
|
|
|
if (!hlsl_regset_index_from_deref(ctx, &load->resource, regset, &index))
|
|
|
|
|
return false;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
|
|
|
|
if (regset == HLSL_REGSET_SAMPLERS)
|
|
|
|
|
{
|
2022-11-25 15:38:33 -08:00
|
|
|
enum hlsl_sampler_dim dim;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!load->sampler.var);
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2022-11-25 15:38:33 -08:00
|
|
|
dim = var->objects_usage[regset][index].sampler_dim;
|
|
|
|
|
if (dim != load->sampling_dim)
|
|
|
|
|
{
|
|
|
|
|
if (dim == HLSL_SAMPLER_DIM_GENERIC)
|
|
|
|
|
{
|
|
|
|
|
var->objects_usage[regset][index].first_sampler_dim_loc = instr->loc;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INCONSISTENT_SAMPLER,
|
|
|
|
|
"Inconsistent generic sampler usage dimension.");
|
|
|
|
|
hlsl_note(ctx, &var->objects_usage[regset][index].first_sampler_dim_loc,
|
|
|
|
|
VKD3D_SHADER_LOG_ERROR, "First use is here.");
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
2022-11-25 14:47:56 -08:00
|
|
|
}
|
2023-05-25 14:38:31 -07:00
|
|
|
var->objects_usage[regset][index].sampler_dim = load->sampling_dim;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2023-05-25 14:38:31 -07:00
|
|
|
return false;
|
|
|
|
|
}
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 20:37:31 -07:00
|
|
|
static void register_deref_usage(struct hlsl_ctx *ctx, struct hlsl_deref *deref)
|
2023-05-25 14:38:31 -07:00
|
|
|
{
|
2024-05-06 20:37:31 -07:00
|
|
|
struct hlsl_ir_var *var = deref->var;
|
|
|
|
|
enum hlsl_regset regset = hlsl_deref_get_regset(ctx, deref);
|
2024-05-06 23:45:14 -07:00
|
|
|
uint32_t required_bind_count;
|
|
|
|
|
struct hlsl_type *type;
|
2023-05-25 14:38:31 -07:00
|
|
|
unsigned int index;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 20:37:31 -07:00
|
|
|
if (!hlsl_regset_index_from_deref(ctx, deref, regset, &index))
|
|
|
|
|
return;
|
2023-06-12 08:58:07 -07:00
|
|
|
|
2024-05-06 20:37:31 -07:00
|
|
|
if (regset <= HLSL_REGSET_LAST_OBJECT)
|
|
|
|
|
{
|
|
|
|
|
var->objects_usage[regset][index].used = true;
|
|
|
|
|
var->bind_count[regset] = max(var->bind_count[regset], index + 1);
|
|
|
|
|
}
|
2024-05-06 23:45:14 -07:00
|
|
|
else if (regset == HLSL_REGSET_NUMERIC)
|
|
|
|
|
{
|
|
|
|
|
type = hlsl_deref_get_type(ctx, deref);
|
|
|
|
|
|
|
|
|
|
hlsl_regset_index_from_deref(ctx, deref, regset, &index);
|
|
|
|
|
required_bind_count = align(index + type->reg_size[regset], 4) / 4;
|
|
|
|
|
var->bind_count[regset] = max(var->bind_count[regset], required_bind_count);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
2024-05-06 20:37:31 -07:00
|
|
|
}
|
2023-05-25 14:38:31 -07:00
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
static bool track_components_usage(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
|
2024-05-06 20:37:31 -07:00
|
|
|
{
|
|
|
|
|
switch (instr->type)
|
2023-05-25 14:38:31 -07:00
|
|
|
{
|
2024-05-06 23:45:14 -07:00
|
|
|
case HLSL_IR_LOAD:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_load *load = hlsl_ir_load(instr);
|
|
|
|
|
|
|
|
|
|
if (!load->src.var->is_uniform)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
/* These will are handled by validate_static_object_references(). */
|
|
|
|
|
if (hlsl_deref_get_regset(ctx, &load->src) != HLSL_REGSET_NUMERIC)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
register_deref_usage(ctx, &load->src);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-06 20:37:31 -07:00
|
|
|
case HLSL_IR_RESOURCE_LOAD:
|
|
|
|
|
register_deref_usage(ctx, &hlsl_ir_resource_load(instr)->resource);
|
|
|
|
|
if (hlsl_ir_resource_load(instr)->sampler.var)
|
|
|
|
|
register_deref_usage(ctx, &hlsl_ir_resource_load(instr)->sampler);
|
|
|
|
|
break;
|
2023-05-25 14:38:31 -07:00
|
|
|
|
2024-05-06 20:37:31 -07:00
|
|
|
case HLSL_IR_RESOURCE_STORE:
|
|
|
|
|
register_deref_usage(ctx, &hlsl_ir_resource_store(instr)->resource);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
break;
|
2022-11-25 14:47:56 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void calculate_resource_register_counts(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
struct hlsl_type *type;
|
2023-08-07 15:22:10 -07:00
|
|
|
unsigned int k;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
type = var->data_type;
|
|
|
|
|
|
|
|
|
|
for (k = 0; k <= HLSL_REGSET_LAST_OBJECT; ++k)
|
|
|
|
|
{
|
2023-08-07 15:22:10 -07:00
|
|
|
bool is_separated = var->is_separated_resource;
|
2022-12-01 15:17:08 -08:00
|
|
|
|
2023-08-07 15:22:10 -07:00
|
|
|
if (var->bind_count[k] > 0)
|
|
|
|
|
var->regs[k].allocation_size = (k == HLSL_REGSET_SAMPLERS || is_separated) ? var->bind_count[k] : type->reg_size[k];
|
2022-11-25 14:47:56 -08:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-20 16:11:44 -07:00
|
|
|
static void allocate_instr_temp_register(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_ir_node *instr, struct register_allocator *allocator)
|
|
|
|
|
{
|
|
|
|
|
unsigned int reg_writemask = 0, dst_writemask = 0;
|
|
|
|
|
|
|
|
|
|
if (instr->reg.allocated || !instr->last_read)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if (instr->type == HLSL_IR_EXPR)
|
|
|
|
|
{
|
|
|
|
|
switch (hlsl_ir_expr(instr)->op)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_OP1_COS_REDUCED:
|
|
|
|
|
dst_writemask = VKD3DSP_WRITEMASK_0;
|
|
|
|
|
reg_writemask = ctx->profile->major_version < 3 ? (1 << 3) - 1 : VKD3DSP_WRITEMASK_0;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP1_SIN_REDUCED:
|
|
|
|
|
dst_writemask = VKD3DSP_WRITEMASK_1;
|
|
|
|
|
reg_writemask = ctx->profile->major_version < 3 ? (1 << 3) - 1 : VKD3DSP_WRITEMASK_1;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (reg_writemask)
|
|
|
|
|
instr->reg = allocate_register_with_masks(ctx, allocator,
|
|
|
|
|
instr->index, instr->last_read, reg_writemask, dst_writemask);
|
|
|
|
|
else
|
|
|
|
|
instr->reg = allocate_numeric_registers_for_type(ctx, allocator,
|
|
|
|
|
instr->index, instr->last_read, instr->data_type);
|
|
|
|
|
|
|
|
|
|
TRACE("Allocated anonymous expression @%u to %s (liveness %u-%u).\n", instr->index,
|
|
|
|
|
debug_register('r', instr->reg, instr->data_type), instr->index, instr->last_read);
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
static void allocate_variable_temp_register(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_ir_var *var, struct register_allocator *allocator)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2021-04-27 10:14:20 -07:00
|
|
|
if (var->is_input_semantic || var->is_output_semantic || var->is_uniform)
|
2021-04-08 21:38:23 -07:00
|
|
|
return;
|
|
|
|
|
|
2022-11-24 12:03:54 -08:00
|
|
|
if (!var->regs[HLSL_REGSET_NUMERIC].allocated && var->last_read)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-10-04 14:03:14 -07:00
|
|
|
if (var->indexable)
|
|
|
|
|
{
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].id = allocator->indexable_count++;
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].allocation_size = 1;
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].writemask = 0;
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].allocated = true;
|
2023-02-02 09:41:13 -08:00
|
|
|
|
2023-10-04 14:03:14 -07:00
|
|
|
TRACE("Allocated %s to x%u[].\n", var->name, var->regs[HLSL_REGSET_NUMERIC].id);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC] = allocate_numeric_registers_for_type(ctx, allocator,
|
|
|
|
|
var->first_write, var->last_read, var->data_type);
|
|
|
|
|
|
|
|
|
|
TRACE("Allocated %s to %s (liveness %u-%u).\n", var->name, debug_register('r',
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC], var->data_type), var->first_write, var->last_read);
|
|
|
|
|
}
|
2021-04-08 21:38:23 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
static void allocate_temp_registers_recurse(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_block *block, struct register_allocator *allocator)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr;
|
|
|
|
|
|
2021-10-15 14:54:10 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(instr, &block->instrs, struct hlsl_ir_node, entry)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-05-26 05:40:30 -07:00
|
|
|
/* In SM4 all constants are inlined. */
|
|
|
|
|
if (ctx->profile->major_version >= 4 && instr->type == HLSL_IR_CONSTANT)
|
|
|
|
|
continue;
|
|
|
|
|
|
2024-06-20 16:11:44 -07:00
|
|
|
allocate_instr_temp_register(ctx, instr, allocator);
|
2021-04-08 21:38:24 -07:00
|
|
|
|
2021-04-08 21:38:23 -07:00
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_IR_IF:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
2023-04-05 13:15:37 -07:00
|
|
|
allocate_temp_registers_recurse(ctx, &iff->then_block, allocator);
|
|
|
|
|
allocate_temp_registers_recurse(ctx, &iff->else_block, allocator);
|
2021-04-08 21:38:23 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_LOAD:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_load *load = hlsl_ir_load(instr);
|
|
|
|
|
/* We need to at least allocate a variable for undefs.
|
|
|
|
|
* FIXME: We should probably find a way to remove them instead. */
|
2023-04-05 13:15:37 -07:00
|
|
|
allocate_variable_temp_register(ctx, load->src.var, allocator);
|
2021-04-08 21:38:23 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_LOOP:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_loop *loop = hlsl_ir_loop(instr);
|
2023-04-05 13:15:37 -07:00
|
|
|
allocate_temp_registers_recurse(ctx, &loop->body, allocator);
|
2021-04-08 21:38:23 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_STORE:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_store *store = hlsl_ir_store(instr);
|
2023-04-05 13:15:37 -07:00
|
|
|
allocate_variable_temp_register(ctx, store->lhs.var, allocator);
|
2021-04-08 21:38:23 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-11 04:51:51 -07:00
|
|
|
case HLSL_IR_SWITCH:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
allocate_temp_registers_recurse(ctx, &c->body, allocator);
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-08 21:38:23 -07:00
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-23 09:57:44 -07:00
|
|
|
static void record_constant(struct hlsl_ctx *ctx, unsigned int component_index, float f,
|
|
|
|
|
const struct vkd3d_shader_location *loc)
|
2023-02-24 14:42:26 -08:00
|
|
|
{
|
|
|
|
|
struct hlsl_constant_defs *defs = &ctx->constant_defs;
|
|
|
|
|
struct hlsl_constant_register *reg;
|
|
|
|
|
size_t i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < defs->count; ++i)
|
|
|
|
|
{
|
|
|
|
|
reg = &defs->regs[i];
|
|
|
|
|
if (reg->index == (component_index / 4))
|
|
|
|
|
{
|
|
|
|
|
reg->value.f[component_index % 4] = f;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!hlsl_array_reserve(ctx, (void **)&defs->regs, &defs->size, defs->count + 1, sizeof(*defs->regs)))
|
|
|
|
|
return;
|
|
|
|
|
reg = &defs->regs[defs->count++];
|
|
|
|
|
memset(reg, 0, sizeof(*reg));
|
|
|
|
|
reg->index = component_index / 4;
|
|
|
|
|
reg->value.f[component_index % 4] = f;
|
2024-05-23 09:57:44 -07:00
|
|
|
reg->loc = *loc;
|
2023-02-24 14:42:26 -08:00
|
|
|
}
|
|
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
static void allocate_const_registers_recurse(struct hlsl_ctx *ctx,
|
|
|
|
|
struct hlsl_block *block, struct register_allocator *allocator)
|
2021-04-08 21:38:26 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr;
|
|
|
|
|
|
2021-10-15 14:54:10 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(instr, &block->instrs, struct hlsl_ir_node, entry)
|
2021-04-08 21:38:26 -07:00
|
|
|
{
|
|
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_IR_CONSTANT:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_constant *constant = hlsl_ir_constant(instr);
|
2021-04-27 10:14:17 -07:00
|
|
|
const struct hlsl_type *type = instr->data_type;
|
2023-02-24 14:42:26 -08:00
|
|
|
unsigned int x, i;
|
2021-04-08 21:38:26 -07:00
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
constant->reg = allocate_numeric_registers_for_type(ctx, allocator, 1, UINT_MAX, type);
|
2021-04-27 10:14:17 -07:00
|
|
|
TRACE("Allocated constant @%u to %s.\n", instr->index, debug_register('c', constant->reg, type));
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(hlsl_is_numeric_type(type));
|
|
|
|
|
VKD3D_ASSERT(type->dimy == 1);
|
|
|
|
|
VKD3D_ASSERT(constant->reg.writemask);
|
2021-04-27 10:14:17 -07:00
|
|
|
|
2023-02-24 14:24:47 -08:00
|
|
|
for (x = 0, i = 0; x < 4; ++x)
|
2021-04-27 10:14:17 -07:00
|
|
|
{
|
2023-02-24 14:24:47 -08:00
|
|
|
const union hlsl_constant_value_component *value;
|
|
|
|
|
float f;
|
|
|
|
|
|
|
|
|
|
if (!(constant->reg.writemask & (1u << x)))
|
|
|
|
|
continue;
|
|
|
|
|
value = &constant->value.u[i++];
|
|
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
switch (type->e.numeric.type)
|
2021-04-27 10:14:17 -07:00
|
|
|
{
|
2023-02-24 14:24:47 -08:00
|
|
|
case HLSL_TYPE_BOOL:
|
|
|
|
|
f = !!value->u;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_FLOAT:
|
|
|
|
|
case HLSL_TYPE_HALF:
|
|
|
|
|
f = value->f;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_INT:
|
|
|
|
|
f = value->i;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_UINT:
|
|
|
|
|
f = value->u;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_DOUBLE:
|
|
|
|
|
FIXME("Double constant.\n");
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
2021-04-27 10:14:17 -07:00
|
|
|
}
|
2023-02-24 14:42:26 -08:00
|
|
|
|
2024-05-23 09:57:44 -07:00
|
|
|
record_constant(ctx, constant->reg.id * 4 + x, f, &constant->node.loc);
|
2021-04-27 10:14:17 -07:00
|
|
|
}
|
|
|
|
|
|
2021-04-08 21:38:26 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_IF:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
2023-04-05 13:15:37 -07:00
|
|
|
allocate_const_registers_recurse(ctx, &iff->then_block, allocator);
|
|
|
|
|
allocate_const_registers_recurse(ctx, &iff->else_block, allocator);
|
2021-04-08 21:38:26 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_LOOP:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_loop *loop = hlsl_ir_loop(instr);
|
2023-04-05 13:15:37 -07:00
|
|
|
allocate_const_registers_recurse(ctx, &loop->body, allocator);
|
2021-04-08 21:38:26 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-11 04:51:51 -07:00
|
|
|
case HLSL_IR_SWITCH:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
allocate_const_registers_recurse(ctx, &c->body, allocator);
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-08 21:38:26 -07:00
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-02 23:51:44 -07:00
|
|
|
static void sort_uniform_by_numeric_bind_count(struct list *sorted, struct hlsl_ir_var *to_sort)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
|
|
|
|
list_remove(&to_sort->extern_entry);
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, sorted, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
uint32_t to_sort_size = to_sort->bind_count[HLSL_REGSET_NUMERIC];
|
|
|
|
|
uint32_t var_size = var->bind_count[HLSL_REGSET_NUMERIC];
|
|
|
|
|
|
|
|
|
|
if (to_sort_size > var_size)
|
|
|
|
|
{
|
|
|
|
|
list_add_before(&var->extern_entry, &to_sort->extern_entry);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
list_add_tail(sorted, &to_sort->extern_entry);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void sort_uniforms_by_numeric_bind_count(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
struct list sorted = LIST_INIT(sorted);
|
|
|
|
|
struct hlsl_ir_var *var, *next;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY_SAFE(var, next, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (var->is_uniform)
|
|
|
|
|
sort_uniform_by_numeric_bind_count(&sorted, var);
|
|
|
|
|
}
|
|
|
|
|
list_move_tail(&ctx->extern_vars, &sorted);
|
|
|
|
|
}
|
|
|
|
|
|
vkd3d-shader/hlsl: Add missing src1 and src2 constants to sincos on SM2.
The sincos instruction expects two specific constants on 2.0 and 2.1 profiles.
Consider the following shader:
uniform float u;
float4 main() : sv_target
{
return sin(u);
}
On native, with ps_2_0, this compiles as:
ps_2_0
def c3, 0.159154937, 0.5, 6.28318548, -3.14159274
def c1, -1.55009923e-006, -2.17013894e-005, 0.00260416674, 0.00026041668
def c2, -0.020833334, -0.125, 1, 0.5
mov r0.xy, c3
mad r0.x, c0.x, r0.x, r0.y
frc r0.x, r0.x
mad r0.x, r0.x, c3.z, c3.w
sincos r1.y, r0.x, c1, c2
mov r0, r1.y
mov oC0, r0
We are not emitting the src1 and src2 constant arguments before this
patch.
2024-07-23 14:12:53 -07:00
|
|
|
/* In SM2, 'sincos' expects specific constants as src1 and src2 arguments.
|
|
|
|
|
* These have to be referenced directly, i.e. as 'c' not 'r'. */
|
|
|
|
|
static void allocate_sincos_const_registers(struct hlsl_ctx *ctx, struct hlsl_block *block,
|
|
|
|
|
struct register_allocator *allocator)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_node *instr;
|
|
|
|
|
struct hlsl_type *type;
|
|
|
|
|
|
|
|
|
|
if (ctx->profile->major_version >= 3)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(instr, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
if (instr->type == HLSL_IR_EXPR && (hlsl_ir_expr(instr)->op == HLSL_OP1_SIN_REDUCED
|
|
|
|
|
|| hlsl_ir_expr(instr)->op == HLSL_OP1_COS_REDUCED))
|
|
|
|
|
{
|
|
|
|
|
type = hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, 4);
|
|
|
|
|
|
|
|
|
|
ctx->d3dsincosconst1 = allocate_numeric_registers_for_type(ctx, allocator, 1, UINT_MAX, type);
|
|
|
|
|
TRACE("Allocated D3DSINCOSCONST1 to %s.\n", debug_register('c', ctx->d3dsincosconst1, type));
|
2024-05-23 09:57:44 -07:00
|
|
|
record_constant(ctx, ctx->d3dsincosconst1.id * 4 + 0, -1.55009923e-06f, &instr->loc);
|
|
|
|
|
record_constant(ctx, ctx->d3dsincosconst1.id * 4 + 1, -2.17013894e-05f, &instr->loc);
|
|
|
|
|
record_constant(ctx, ctx->d3dsincosconst1.id * 4 + 2, 2.60416674e-03f, &instr->loc);
|
|
|
|
|
record_constant(ctx, ctx->d3dsincosconst1.id * 4 + 3, 2.60416680e-04f, &instr->loc);
|
vkd3d-shader/hlsl: Add missing src1 and src2 constants to sincos on SM2.
The sincos instruction expects two specific constants on 2.0 and 2.1 profiles.
Consider the following shader:
uniform float u;
float4 main() : sv_target
{
return sin(u);
}
On native, with ps_2_0, this compiles as:
ps_2_0
def c3, 0.159154937, 0.5, 6.28318548, -3.14159274
def c1, -1.55009923e-006, -2.17013894e-005, 0.00260416674, 0.00026041668
def c2, -0.020833334, -0.125, 1, 0.5
mov r0.xy, c3
mad r0.x, c0.x, r0.x, r0.y
frc r0.x, r0.x
mad r0.x, r0.x, c3.z, c3.w
sincos r1.y, r0.x, c1, c2
mov r0, r1.y
mov oC0, r0
We are not emitting the src1 and src2 constant arguments before this
patch.
2024-07-23 14:12:53 -07:00
|
|
|
|
|
|
|
|
ctx->d3dsincosconst2 = allocate_numeric_registers_for_type(ctx, allocator, 1, UINT_MAX, type);
|
|
|
|
|
TRACE("Allocated D3DSINCOSCONST2 to %s.\n", debug_register('c', ctx->d3dsincosconst2, type));
|
2024-05-23 09:57:44 -07:00
|
|
|
record_constant(ctx, ctx->d3dsincosconst2.id * 4 + 0, -2.08333340e-02f, &instr->loc);
|
|
|
|
|
record_constant(ctx, ctx->d3dsincosconst2.id * 4 + 1, -1.25000000e-01f, &instr->loc);
|
|
|
|
|
record_constant(ctx, ctx->d3dsincosconst2.id * 4 + 2, 1.00000000e+00f, &instr->loc);
|
|
|
|
|
record_constant(ctx, ctx->d3dsincosconst2.id * 4 + 3, 5.00000000e-01f, &instr->loc);
|
vkd3d-shader/hlsl: Add missing src1 and src2 constants to sincos on SM2.
The sincos instruction expects two specific constants on 2.0 and 2.1 profiles.
Consider the following shader:
uniform float u;
float4 main() : sv_target
{
return sin(u);
}
On native, with ps_2_0, this compiles as:
ps_2_0
def c3, 0.159154937, 0.5, 6.28318548, -3.14159274
def c1, -1.55009923e-006, -2.17013894e-005, 0.00260416674, 0.00026041668
def c2, -0.020833334, -0.125, 1, 0.5
mov r0.xy, c3
mad r0.x, c0.x, r0.x, r0.y
frc r0.x, r0.x
mad r0.x, r0.x, c3.z, c3.w
sincos r1.y, r0.x, c1, c2
mov r0, r1.y
mov oC0, r0
We are not emitting the src1 and src2 constant arguments before this
patch.
2024-07-23 14:12:53 -07:00
|
|
|
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-08 21:38:27 -07:00
|
|
|
static void allocate_const_registers(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry_func)
|
2021-04-08 21:38:26 -07:00
|
|
|
{
|
2024-06-05 16:47:43 -07:00
|
|
|
struct register_allocator allocator_used = {0};
|
2023-04-05 13:15:37 -07:00
|
|
|
struct register_allocator allocator = {0};
|
2021-04-08 21:38:27 -07:00
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
2024-05-02 23:51:44 -07:00
|
|
|
sort_uniforms_by_numeric_bind_count(ctx);
|
|
|
|
|
|
2021-04-15 17:03:44 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
2021-04-08 21:38:27 -07:00
|
|
|
{
|
2023-11-07 16:13:52 -08:00
|
|
|
unsigned int reg_size = var->data_type->reg_size[HLSL_REGSET_NUMERIC];
|
2024-06-05 16:47:43 -07:00
|
|
|
unsigned int bind_count = var->bind_count[HLSL_REGSET_NUMERIC];
|
2023-11-07 16:13:52 -08:00
|
|
|
|
2024-05-03 16:32:52 -07:00
|
|
|
if (!var->is_uniform || reg_size == 0)
|
2023-11-07 16:13:52 -08:00
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (var->reg_reservation.reg_type == 'c')
|
2021-04-08 21:38:27 -07:00
|
|
|
{
|
2023-11-07 16:13:52 -08:00
|
|
|
unsigned int reg_idx = var->reg_reservation.reg_index;
|
|
|
|
|
unsigned int i;
|
2022-11-24 12:03:54 -08:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(reg_size % 4 == 0);
|
2023-11-07 16:13:52 -08:00
|
|
|
for (i = 0; i < reg_size / 4; ++i)
|
|
|
|
|
{
|
2024-06-05 16:47:43 -07:00
|
|
|
if (i < bind_count)
|
2023-11-07 16:13:52 -08:00
|
|
|
{
|
2024-06-05 16:47:43 -07:00
|
|
|
if (get_available_writemask(&allocator_used, 1, UINT_MAX, reg_idx + i) != VKD3DSP_WRITEMASK_ALL)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"Overlapping register() reservations on 'c%u'.", reg_idx + i);
|
|
|
|
|
}
|
|
|
|
|
record_allocation(ctx, &allocator_used, reg_idx + i, VKD3DSP_WRITEMASK_ALL, 1, UINT_MAX);
|
2023-11-07 16:13:52 -08:00
|
|
|
}
|
|
|
|
|
record_allocation(ctx, &allocator, reg_idx + i, VKD3DSP_WRITEMASK_ALL, 1, UINT_MAX);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].id = reg_idx;
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].allocation_size = reg_size / 4;
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].writemask = VKD3DSP_WRITEMASK_ALL;
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].allocated = true;
|
|
|
|
|
TRACE("Allocated reserved %s to %s.\n", var->name,
|
|
|
|
|
debug_register('c', var->regs[HLSL_REGSET_NUMERIC], var->data_type));
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-05 16:47:43 -07:00
|
|
|
vkd3d_free(allocator_used.allocations);
|
|
|
|
|
|
2023-11-07 16:13:52 -08:00
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
2024-05-02 23:06:37 -07:00
|
|
|
unsigned int alloc_size = 4 * var->bind_count[HLSL_REGSET_NUMERIC];
|
2023-11-07 16:13:52 -08:00
|
|
|
|
2024-05-02 23:06:37 -07:00
|
|
|
if (!var->is_uniform || alloc_size == 0)
|
2023-11-07 16:13:52 -08:00
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (!var->regs[HLSL_REGSET_NUMERIC].allocated)
|
|
|
|
|
{
|
2024-05-02 23:06:37 -07:00
|
|
|
var->regs[HLSL_REGSET_NUMERIC] = allocate_range(ctx, &allocator, 1, UINT_MAX, alloc_size);
|
2022-11-24 12:03:54 -08:00
|
|
|
TRACE("Allocated %s to %s.\n", var->name,
|
|
|
|
|
debug_register('c', var->regs[HLSL_REGSET_NUMERIC], var->data_type));
|
2021-04-08 21:38:27 -07:00
|
|
|
}
|
|
|
|
|
}
|
2023-04-05 13:16:00 -07:00
|
|
|
|
2022-11-07 15:45:35 -08:00
|
|
|
allocate_const_registers_recurse(ctx, &entry_func->body, &allocator);
|
|
|
|
|
|
vkd3d-shader/hlsl: Add missing src1 and src2 constants to sincos on SM2.
The sincos instruction expects two specific constants on 2.0 and 2.1 profiles.
Consider the following shader:
uniform float u;
float4 main() : sv_target
{
return sin(u);
}
On native, with ps_2_0, this compiles as:
ps_2_0
def c3, 0.159154937, 0.5, 6.28318548, -3.14159274
def c1, -1.55009923e-006, -2.17013894e-005, 0.00260416674, 0.00026041668
def c2, -0.020833334, -0.125, 1, 0.5
mov r0.xy, c3
mad r0.x, c0.x, r0.x, r0.y
frc r0.x, r0.x
mad r0.x, r0.x, c3.z, c3.w
sincos r1.y, r0.x, c1, c2
mov r0, r1.y
mov oC0, r0
We are not emitting the src1 and src2 constant arguments before this
patch.
2024-07-23 14:12:53 -07:00
|
|
|
allocate_sincos_const_registers(ctx, &entry_func->body, &allocator);
|
|
|
|
|
|
2023-04-05 12:09:16 -07:00
|
|
|
vkd3d_free(allocator.allocations);
|
2021-04-08 21:38:26 -07:00
|
|
|
}
|
|
|
|
|
|
2021-04-08 21:38:23 -07:00
|
|
|
/* Simple greedy temporary register allocation pass that just assigns a unique
|
|
|
|
|
* index to all (simultaneously live) variables or intermediate values. Agnostic
|
|
|
|
|
* as to how many registers are actually available for the current backend, and
|
|
|
|
|
* does not handle constants. */
|
2021-05-20 22:32:20 -07:00
|
|
|
static void allocate_temp_registers(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry_func)
|
2021-04-08 21:38:23 -07:00
|
|
|
{
|
2023-04-05 13:15:37 -07:00
|
|
|
struct register_allocator allocator = {0};
|
2023-04-05 14:18:25 -07:00
|
|
|
|
|
|
|
|
/* ps_1_* outputs are special and go in temp register 0. */
|
|
|
|
|
if (ctx->profile->major_version == 1 && ctx->profile->type == VKD3D_SHADER_TYPE_PIXEL)
|
|
|
|
|
{
|
|
|
|
|
size_t i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < entry_func->parameters.count; ++i)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_var *var = entry_func->parameters.vars[i];
|
|
|
|
|
|
|
|
|
|
if (var->is_output_semantic)
|
|
|
|
|
{
|
|
|
|
|
record_allocation(ctx, &allocator, 0, VKD3DSP_WRITEMASK_ALL, var->first_write, var->last_read);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2023-04-05 13:15:37 -07:00
|
|
|
allocate_temp_registers_recurse(ctx, &entry_func->body, &allocator);
|
2023-10-27 15:06:51 -07:00
|
|
|
ctx->temp_count = allocator.reg_count;
|
2023-04-05 12:09:16 -07:00
|
|
|
vkd3d_free(allocator.allocations);
|
2021-04-08 21:38:23 -07:00
|
|
|
}
|
|
|
|
|
|
2021-04-27 10:14:21 -07:00
|
|
|
static void allocate_semantic_register(struct hlsl_ctx *ctx, struct hlsl_ir_var *var, unsigned int *counter, bool output)
|
|
|
|
|
{
|
2023-04-05 13:22:03 -07:00
|
|
|
static const char *const shader_names[] =
|
2021-08-19 16:44:27 -07:00
|
|
|
{
|
|
|
|
|
[VKD3D_SHADER_TYPE_PIXEL] = "Pixel",
|
|
|
|
|
[VKD3D_SHADER_TYPE_VERTEX] = "Vertex",
|
|
|
|
|
[VKD3D_SHADER_TYPE_GEOMETRY] = "Geometry",
|
|
|
|
|
[VKD3D_SHADER_TYPE_HULL] = "Hull",
|
|
|
|
|
[VKD3D_SHADER_TYPE_DOMAIN] = "Domain",
|
|
|
|
|
[VKD3D_SHADER_TYPE_COMPUTE] = "Compute",
|
|
|
|
|
};
|
|
|
|
|
|
2023-11-24 05:03:25 -08:00
|
|
|
enum vkd3d_shader_register_type type;
|
2021-08-19 16:44:27 -07:00
|
|
|
uint32_t reg;
|
|
|
|
|
bool builtin;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(var->semantic.name);
|
2021-04-27 10:14:21 -07:00
|
|
|
|
|
|
|
|
if (ctx->profile->major_version < 4)
|
|
|
|
|
{
|
2024-05-17 19:06:58 -07:00
|
|
|
struct vkd3d_shader_version version;
|
2024-08-08 05:33:26 -07:00
|
|
|
enum vkd3d_decl_usage usage;
|
2021-08-19 16:44:27 -07:00
|
|
|
uint32_t usage_idx;
|
2021-05-10 21:36:07 -07:00
|
|
|
|
2023-04-05 14:18:25 -07:00
|
|
|
/* ps_1_* outputs are special and go in temp register 0. */
|
|
|
|
|
if (ctx->profile->major_version == 1 && output && ctx->profile->type == VKD3D_SHADER_TYPE_PIXEL)
|
|
|
|
|
return;
|
|
|
|
|
|
2024-05-17 19:06:58 -07:00
|
|
|
version.major = ctx->profile->major_version;
|
|
|
|
|
version.minor = ctx->profile->minor_version;
|
|
|
|
|
version.type = ctx->profile->type;
|
2024-05-21 15:19:23 -07:00
|
|
|
builtin = hlsl_sm1_register_from_semantic(&version,
|
|
|
|
|
var->semantic.name, var->semantic.index, output, &type, ®);
|
|
|
|
|
if (!builtin && !hlsl_sm1_usage_from_semantic(var->semantic.name, var->semantic.index, &usage, &usage_idx))
|
2021-05-10 21:36:07 -07:00
|
|
|
{
|
2021-12-01 08:14:57 -08:00
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_SEMANTIC,
|
2021-05-10 21:36:07 -07:00
|
|
|
"Invalid semantic '%s'.", var->semantic.name);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2021-04-27 10:14:21 -07:00
|
|
|
|
2021-08-19 16:44:27 -07:00
|
|
|
if ((!output && !var->last_read) || (output && !var->first_write))
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2024-08-07 08:23:04 -07:00
|
|
|
enum vkd3d_shader_sysval_semantic semantic;
|
2021-08-19 16:44:29 -07:00
|
|
|
bool has_idx;
|
2021-08-19 16:44:27 -07:00
|
|
|
|
2024-08-07 08:23:04 -07:00
|
|
|
if (!sysval_semantic_from_hlsl(&semantic, ctx, &var->semantic, output))
|
2021-04-27 10:14:21 -07:00
|
|
|
{
|
2021-12-01 08:14:57 -08:00
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_SEMANTIC,
|
2021-08-19 16:44:27 -07:00
|
|
|
"Invalid semantic '%s'.", var->semantic.name);
|
|
|
|
|
return;
|
2021-04-27 10:14:21 -07:00
|
|
|
}
|
2023-07-18 15:59:43 -07:00
|
|
|
if ((builtin = hlsl_sm4_register_from_semantic(ctx, &var->semantic, output, &type, &has_idx)))
|
2021-08-19 16:44:29 -07:00
|
|
|
reg = has_idx ? var->semantic.index : 0;
|
2021-08-19 16:44:27 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (builtin)
|
|
|
|
|
{
|
|
|
|
|
TRACE("%s %s semantic %s[%u] matches predefined register %#x[%u].\n", shader_names[ctx->profile->type],
|
|
|
|
|
output ? "output" : "input", var->semantic.name, var->semantic.index, type, reg);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2022-11-24 12:03:54 -08:00
|
|
|
var->regs[HLSL_REGSET_NUMERIC].allocated = true;
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC].id = (*counter)++;
|
vkd3d-shader/hlsl: Rename hlsl_reg.bind_count to hlsl_reg.allocation_size.
We have to distinguish between the "bind count" and the "allocation size"
of variables.
The "allocation size" affects the starting register id for the resource to
be allocated next, while the "bind count" is determined by the last field
actually used. The former may be larger than the latter.
What we are currently calling hlsl_reg.bind_count is actually the
"allocation size", so a rename is in order.
The real "bind count", which will be introduced in following patches,
is important because it is what should be shown in the RDEF table and
some resource allocation rules depend on it.
For instance, for this shader:
texture2D texs[3];
texture2D tex;
float4 main() : sv_target
{
return texs[0].Load(int3(0, 0, 0)) + tex.Load(int3(0, 0, 0));
}
the variable "texs" has a "bind count" of 1, but an "allocation size" of
3:
// Resource Bindings:
//
// Name Type Format Dim HLSL Bind Count
// ------------------------------ ---------- ------- ----------- -------------- ------
// texs texture float4 2d t0 1
// tex texture float4 2d t3 1
2023-08-04 10:21:27 -07:00
|
|
|
var->regs[HLSL_REGSET_NUMERIC].allocation_size = 1;
|
2022-11-24 12:03:54 -08:00
|
|
|
var->regs[HLSL_REGSET_NUMERIC].writemask = (1 << var->data_type->dimx) - 1;
|
|
|
|
|
TRACE("Allocated %s to %s.\n", var->name, debug_register(output ? 'o' : 'v',
|
|
|
|
|
var->regs[HLSL_REGSET_NUMERIC], var->data_type));
|
2021-04-27 10:14:21 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void allocate_semantic_registers(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
unsigned int input_counter = 0, output_counter = 0;
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
2021-08-19 16:44:27 -07:00
|
|
|
if (var->is_input_semantic)
|
2021-04-27 10:14:21 -07:00
|
|
|
allocate_semantic_register(ctx, var, &input_counter, false);
|
2021-08-19 16:44:27 -07:00
|
|
|
if (var->is_output_semantic)
|
2021-04-27 10:14:21 -07:00
|
|
|
allocate_semantic_register(ctx, var, &output_counter, true);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-09-26 13:35:12 -07:00
|
|
|
static const struct hlsl_buffer *get_reserved_buffer(struct hlsl_ctx *ctx,
|
|
|
|
|
uint32_t space, uint32_t index, bool allocated_only)
|
2021-06-23 21:57:35 -07:00
|
|
|
{
|
|
|
|
|
const struct hlsl_buffer *buffer;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(buffer, &ctx->buffers, const struct hlsl_buffer, entry)
|
|
|
|
|
{
|
2024-07-03 10:05:58 -07:00
|
|
|
if (buffer->reservation.reg_type == 'b'
|
2023-08-29 10:30:00 -07:00
|
|
|
&& buffer->reservation.reg_space == space && buffer->reservation.reg_index == index)
|
2024-09-26 13:35:12 -07:00
|
|
|
{
|
|
|
|
|
if (allocated_only && !buffer->reg.allocated)
|
|
|
|
|
continue;
|
|
|
|
|
|
2021-06-23 21:57:35 -07:00
|
|
|
return buffer;
|
2024-09-26 13:35:12 -07:00
|
|
|
}
|
2021-06-23 21:57:35 -07:00
|
|
|
}
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2024-02-06 06:11:47 -08:00
|
|
|
static void hlsl_calculate_buffer_offset(struct hlsl_ctx *ctx, struct hlsl_ir_var *var, bool register_reservation)
|
2021-06-23 21:57:35 -07:00
|
|
|
{
|
2023-02-21 12:08:39 -08:00
|
|
|
unsigned int var_reg_size = var->data_type->reg_size[HLSL_REGSET_NUMERIC];
|
|
|
|
|
enum hlsl_type_class var_class = var->data_type->class;
|
2021-06-23 21:57:35 -07:00
|
|
|
struct hlsl_buffer *buffer = var->buffer;
|
|
|
|
|
|
2023-11-08 13:45:26 -08:00
|
|
|
if (register_reservation)
|
2023-02-21 12:08:39 -08:00
|
|
|
{
|
2023-11-08 13:45:26 -08:00
|
|
|
var->buffer_offset = 4 * var->reg_reservation.reg_index;
|
2024-06-01 15:15:48 -07:00
|
|
|
var->has_explicit_bind_point = 1;
|
2023-11-08 13:45:26 -08:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
if (var->reg_reservation.offset_type == 'c')
|
2023-02-21 12:08:39 -08:00
|
|
|
{
|
2023-11-08 13:45:26 -08:00
|
|
|
if (var->reg_reservation.offset_index % 4)
|
2023-02-21 12:08:39 -08:00
|
|
|
{
|
2023-11-08 13:45:26 -08:00
|
|
|
if (var_class == HLSL_CLASS_MATRIX)
|
|
|
|
|
{
|
2023-02-21 12:08:39 -08:00
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
2023-11-08 13:45:26 -08:00
|
|
|
"packoffset() reservations with matrix types must be aligned with the beginning of a register.");
|
|
|
|
|
}
|
|
|
|
|
else if (var_class == HLSL_CLASS_ARRAY)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"packoffset() reservations with array types must be aligned with the beginning of a register.");
|
|
|
|
|
}
|
|
|
|
|
else if (var_class == HLSL_CLASS_STRUCT)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"packoffset() reservations with struct types must be aligned with the beginning of a register.");
|
|
|
|
|
}
|
|
|
|
|
else if (var_class == HLSL_CLASS_VECTOR)
|
|
|
|
|
{
|
|
|
|
|
unsigned int aligned_offset = hlsl_type_get_sm4_offset(var->data_type, var->reg_reservation.offset_index);
|
|
|
|
|
|
|
|
|
|
if (var->reg_reservation.offset_index != aligned_offset)
|
|
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"packoffset() reservations with vector types cannot span multiple registers.");
|
|
|
|
|
}
|
2023-02-21 12:08:39 -08:00
|
|
|
}
|
2023-11-08 13:45:26 -08:00
|
|
|
var->buffer_offset = var->reg_reservation.offset_index;
|
2024-06-01 15:15:48 -07:00
|
|
|
var->has_explicit_bind_point = 1;
|
2023-11-08 13:45:26 -08:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
var->buffer_offset = hlsl_type_get_sm4_offset(var->data_type, buffer->size);
|
2023-02-21 12:08:39 -08:00
|
|
|
}
|
|
|
|
|
}
|
2021-06-23 21:57:35 -07:00
|
|
|
|
|
|
|
|
TRACE("Allocated buffer offset %u to %s.\n", var->buffer_offset, var->name);
|
2023-02-21 12:08:39 -08:00
|
|
|
buffer->size = max(buffer->size, var->buffer_offset + var_reg_size);
|
2021-06-23 21:57:35 -07:00
|
|
|
if (var->last_read)
|
2023-02-21 12:08:39 -08:00
|
|
|
buffer->used_size = max(buffer->used_size, var->buffer_offset + var_reg_size);
|
2021-06-23 21:57:35 -07:00
|
|
|
}
|
|
|
|
|
|
2023-02-21 16:53:32 -08:00
|
|
|
static void validate_buffer_offsets(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_var *var1, *var2;
|
|
|
|
|
struct hlsl_buffer *buffer;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var1, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
2023-05-29 18:59:17 -07:00
|
|
|
if (!var1->is_uniform || hlsl_type_is_resource(var1->data_type))
|
2023-02-21 16:53:32 -08:00
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
buffer = var1->buffer;
|
|
|
|
|
if (!buffer->used_size)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var2, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
unsigned int var1_reg_size, var2_reg_size;
|
|
|
|
|
|
2023-05-29 18:59:17 -07:00
|
|
|
if (!var2->is_uniform || hlsl_type_is_resource(var2->data_type))
|
2023-02-21 16:53:32 -08:00
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (var1 == var2 || var1->buffer != var2->buffer)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
/* This is to avoid reporting the error twice for the same pair of overlapping variables. */
|
|
|
|
|
if (strcmp(var1->name, var2->name) >= 0)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
var1_reg_size = var1->data_type->reg_size[HLSL_REGSET_NUMERIC];
|
|
|
|
|
var2_reg_size = var2->data_type->reg_size[HLSL_REGSET_NUMERIC];
|
|
|
|
|
|
|
|
|
|
if (var1->buffer_offset < var2->buffer_offset + var2_reg_size
|
|
|
|
|
&& var2->buffer_offset < var1->buffer_offset + var1_reg_size)
|
|
|
|
|
hlsl_error(ctx, &buffer->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"Invalid packoffset() reservation: Variables %s and %s overlap.",
|
|
|
|
|
var1->name, var2->name);
|
|
|
|
|
}
|
|
|
|
|
}
|
2023-02-21 17:39:24 -08:00
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var1, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
buffer = var1->buffer;
|
2023-03-01 10:57:31 -08:00
|
|
|
if (!buffer || buffer == ctx->globals_buffer)
|
2023-02-21 17:39:24 -08:00
|
|
|
continue;
|
|
|
|
|
|
2023-03-01 10:57:31 -08:00
|
|
|
if (var1->reg_reservation.offset_type
|
2024-01-26 14:59:34 -08:00
|
|
|
|| var1->reg_reservation.reg_type == 's'
|
|
|
|
|
|| var1->reg_reservation.reg_type == 't'
|
|
|
|
|
|| var1->reg_reservation.reg_type == 'u')
|
2023-02-21 17:39:24 -08:00
|
|
|
buffer->manually_packed_elements = true;
|
|
|
|
|
else
|
|
|
|
|
buffer->automatically_packed_elements = true;
|
|
|
|
|
|
|
|
|
|
if (buffer->manually_packed_elements && buffer->automatically_packed_elements)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &buffer->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"packoffset() must be specified for all the buffer elements, or none of them.");
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
2023-02-21 16:53:32 -08:00
|
|
|
}
|
|
|
|
|
|
2024-02-06 06:11:47 -08:00
|
|
|
void hlsl_calculate_buffer_offsets(struct hlsl_ctx *ctx)
|
2023-11-08 13:45:26 -08:00
|
|
|
{
|
2021-06-23 21:57:35 -07:00
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
2023-11-08 13:45:26 -08:00
|
|
|
if (!var->is_uniform || hlsl_type_is_resource(var->data_type))
|
|
|
|
|
continue;
|
2021-06-23 21:57:35 -07:00
|
|
|
|
2024-02-06 06:11:47 -08:00
|
|
|
if (hlsl_var_has_buffer_offset_register_reservation(ctx, var))
|
|
|
|
|
hlsl_calculate_buffer_offset(ctx, var, true);
|
2023-11-08 13:45:26 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (!var->is_uniform || hlsl_type_is_resource(var->data_type))
|
|
|
|
|
continue;
|
|
|
|
|
|
2024-02-06 06:11:47 -08:00
|
|
|
if (!hlsl_var_has_buffer_offset_register_reservation(ctx, var))
|
|
|
|
|
hlsl_calculate_buffer_offset(ctx, var, false);
|
2023-11-08 13:45:26 -08:00
|
|
|
}
|
2024-02-06 06:11:47 -08:00
|
|
|
}
|
|
|
|
|
|
2024-07-03 10:05:58 -07:00
|
|
|
static unsigned int get_max_cbuffer_reg_index(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
if (hlsl_version_ge(ctx, 5, 1))
|
|
|
|
|
return UINT_MAX;
|
|
|
|
|
|
|
|
|
|
return 13;
|
|
|
|
|
}
|
|
|
|
|
|
2024-02-06 06:11:47 -08:00
|
|
|
static void allocate_buffers(struct hlsl_ctx *ctx)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_buffer *buffer;
|
2023-08-29 10:13:35 -07:00
|
|
|
uint32_t index = 0, id = 0;
|
2024-02-06 06:11:47 -08:00
|
|
|
struct hlsl_ir_var *var;
|
2023-11-08 13:45:26 -08:00
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (!var->is_uniform || hlsl_type_is_resource(var->data_type))
|
|
|
|
|
continue;
|
|
|
|
|
|
2024-02-06 06:11:47 -08:00
|
|
|
if (var->is_param)
|
|
|
|
|
var->buffer = ctx->params_buffer;
|
2021-06-23 21:57:35 -07:00
|
|
|
}
|
|
|
|
|
|
2024-02-06 06:11:47 -08:00
|
|
|
hlsl_calculate_buffer_offsets(ctx);
|
2023-02-21 16:53:32 -08:00
|
|
|
validate_buffer_offsets(ctx);
|
|
|
|
|
|
2021-06-23 21:57:35 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(buffer, &ctx->buffers, struct hlsl_buffer, entry)
|
|
|
|
|
{
|
|
|
|
|
if (!buffer->used_size)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (buffer->type == HLSL_BUFFER_CONSTANT)
|
|
|
|
|
{
|
2023-08-29 10:30:00 -07:00
|
|
|
const struct hlsl_reg_reservation *reservation = &buffer->reservation;
|
|
|
|
|
|
|
|
|
|
if (reservation->reg_type == 'b')
|
2021-06-23 21:57:35 -07:00
|
|
|
{
|
2024-09-26 13:35:12 -07:00
|
|
|
const struct hlsl_buffer *allocated_buffer = get_reserved_buffer(ctx,
|
|
|
|
|
reservation->reg_space, reservation->reg_index, true);
|
2024-07-03 10:05:58 -07:00
|
|
|
unsigned int max_index = get_max_cbuffer_reg_index(ctx);
|
|
|
|
|
|
|
|
|
|
if (buffer->reservation.reg_index > max_index)
|
|
|
|
|
hlsl_error(ctx, &buffer->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
|
|
|
|
"Buffer reservation cb%u exceeds target's maximum (cb%u).",
|
|
|
|
|
buffer->reservation.reg_index, max_index);
|
2021-06-23 21:57:35 -07:00
|
|
|
|
2024-09-26 13:35:12 -07:00
|
|
|
if (allocated_buffer && allocated_buffer != buffer)
|
2021-06-23 21:57:35 -07:00
|
|
|
{
|
2021-12-01 08:14:57 -08:00
|
|
|
hlsl_error(ctx, &buffer->loc, VKD3D_SHADER_ERROR_HLSL_OVERLAPPING_RESERVATIONS,
|
2023-08-29 10:30:00 -07:00
|
|
|
"Multiple buffers bound to space %u, index %u.",
|
|
|
|
|
reservation->reg_space, reservation->reg_index);
|
2024-09-26 13:35:12 -07:00
|
|
|
hlsl_note(ctx, &allocated_buffer->loc, VKD3D_SHADER_LOG_ERROR,
|
2023-08-29 10:30:00 -07:00
|
|
|
"Buffer %s is already bound to space %u, index %u.",
|
2024-09-26 13:35:12 -07:00
|
|
|
allocated_buffer->name, reservation->reg_space, reservation->reg_index);
|
2021-06-23 21:57:35 -07:00
|
|
|
}
|
|
|
|
|
|
2023-08-29 10:30:00 -07:00
|
|
|
buffer->reg.space = reservation->reg_space;
|
|
|
|
|
buffer->reg.index = reservation->reg_index;
|
2023-08-29 10:13:35 -07:00
|
|
|
if (hlsl_version_ge(ctx, 5, 1))
|
|
|
|
|
buffer->reg.id = id++;
|
|
|
|
|
else
|
|
|
|
|
buffer->reg.id = buffer->reg.index;
|
vkd3d-shader/hlsl: Rename hlsl_reg.bind_count to hlsl_reg.allocation_size.
We have to distinguish between the "bind count" and the "allocation size"
of variables.
The "allocation size" affects the starting register id for the resource to
be allocated next, while the "bind count" is determined by the last field
actually used. The former may be larger than the latter.
What we are currently calling hlsl_reg.bind_count is actually the
"allocation size", so a rename is in order.
The real "bind count", which will be introduced in following patches,
is important because it is what should be shown in the RDEF table and
some resource allocation rules depend on it.
For instance, for this shader:
texture2D texs[3];
texture2D tex;
float4 main() : sv_target
{
return texs[0].Load(int3(0, 0, 0)) + tex.Load(int3(0, 0, 0));
}
the variable "texs" has a "bind count" of 1, but an "allocation size" of
3:
// Resource Bindings:
//
// Name Type Format Dim HLSL Bind Count
// ------------------------------ ---------- ------- ----------- -------------- ------
// texs texture float4 2d t0 1
// tex texture float4 2d t3 1
2023-08-04 10:21:27 -07:00
|
|
|
buffer->reg.allocation_size = 1;
|
2021-06-23 21:57:35 -07:00
|
|
|
buffer->reg.allocated = true;
|
2023-08-29 10:30:00 -07:00
|
|
|
TRACE("Allocated reserved %s to space %u, index %u, id %u.\n",
|
|
|
|
|
buffer->name, buffer->reg.space, buffer->reg.index, buffer->reg.id);
|
2021-06-23 21:57:35 -07:00
|
|
|
}
|
2023-08-29 10:30:00 -07:00
|
|
|
else if (!reservation->reg_type)
|
2021-06-23 21:57:35 -07:00
|
|
|
{
|
2024-07-03 10:05:58 -07:00
|
|
|
unsigned int max_index = get_max_cbuffer_reg_index(ctx);
|
2024-09-26 13:35:12 -07:00
|
|
|
while (get_reserved_buffer(ctx, 0, index, false))
|
2021-06-23 21:57:35 -07:00
|
|
|
++index;
|
|
|
|
|
|
2024-07-03 10:05:58 -07:00
|
|
|
if (index > max_index)
|
|
|
|
|
hlsl_error(ctx, &buffer->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
2024-09-26 13:35:12 -07:00
|
|
|
"Too many buffers reserved, target's maximum is %u.", max_index);
|
2024-07-03 10:05:58 -07:00
|
|
|
|
2023-08-29 10:30:00 -07:00
|
|
|
buffer->reg.space = 0;
|
2023-08-29 10:13:35 -07:00
|
|
|
buffer->reg.index = index;
|
|
|
|
|
if (hlsl_version_ge(ctx, 5, 1))
|
|
|
|
|
buffer->reg.id = id++;
|
|
|
|
|
else
|
|
|
|
|
buffer->reg.id = buffer->reg.index;
|
vkd3d-shader/hlsl: Rename hlsl_reg.bind_count to hlsl_reg.allocation_size.
We have to distinguish between the "bind count" and the "allocation size"
of variables.
The "allocation size" affects the starting register id for the resource to
be allocated next, while the "bind count" is determined by the last field
actually used. The former may be larger than the latter.
What we are currently calling hlsl_reg.bind_count is actually the
"allocation size", so a rename is in order.
The real "bind count", which will be introduced in following patches,
is important because it is what should be shown in the RDEF table and
some resource allocation rules depend on it.
For instance, for this shader:
texture2D texs[3];
texture2D tex;
float4 main() : sv_target
{
return texs[0].Load(int3(0, 0, 0)) + tex.Load(int3(0, 0, 0));
}
the variable "texs" has a "bind count" of 1, but an "allocation size" of
3:
// Resource Bindings:
//
// Name Type Format Dim HLSL Bind Count
// ------------------------------ ---------- ------- ----------- -------------- ------
// texs texture float4 2d t0 1
// tex texture float4 2d t3 1
2023-08-04 10:21:27 -07:00
|
|
|
buffer->reg.allocation_size = 1;
|
2021-06-23 21:57:35 -07:00
|
|
|
buffer->reg.allocated = true;
|
2023-08-29 10:30:00 -07:00
|
|
|
TRACE("Allocated %s to space 0, index %u, id %u.\n", buffer->name, buffer->reg.index, buffer->reg.id);
|
2021-06-23 21:57:35 -07:00
|
|
|
++index;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2021-12-01 08:14:57 -08:00
|
|
|
hlsl_error(ctx, &buffer->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_RESERVATION,
|
2021-06-23 21:57:35 -07:00
|
|
|
"Constant buffers must be allocated to register type 'b'.");
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
FIXME("Allocate registers for texture buffers.\n");
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2022-11-24 14:36:20 -08:00
|
|
|
static const struct hlsl_ir_var *get_allocated_object(struct hlsl_ctx *ctx, enum hlsl_regset regset,
|
2023-08-29 10:40:38 -07:00
|
|
|
uint32_t space, uint32_t index, bool allocated_only)
|
2021-10-11 19:58:46 -07:00
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_var *var;
|
2023-04-25 09:41:38 -07:00
|
|
|
unsigned int start, count;
|
2021-10-11 19:58:46 -07:00
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, const struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
2023-06-21 10:00:39 -07:00
|
|
|
if (var->reg_reservation.reg_type == get_regset_name(regset)
|
|
|
|
|
&& var->data_type->reg_size[regset])
|
|
|
|
|
{
|
|
|
|
|
/* Vars with a reservation prevent non-reserved vars from being
|
|
|
|
|
* bound there even if the reserved vars aren't used. */
|
|
|
|
|
start = var->reg_reservation.reg_index;
|
|
|
|
|
count = var->data_type->reg_size[regset];
|
2023-06-27 11:14:37 -07:00
|
|
|
|
2023-08-29 10:40:38 -07:00
|
|
|
if (var->reg_reservation.reg_space != space)
|
|
|
|
|
continue;
|
|
|
|
|
|
2023-06-27 11:14:37 -07:00
|
|
|
if (!var->regs[regset].allocated && allocated_only)
|
|
|
|
|
continue;
|
2023-06-21 10:00:39 -07:00
|
|
|
}
|
|
|
|
|
else if (var->regs[regset].allocated)
|
|
|
|
|
{
|
2023-08-29 10:40:38 -07:00
|
|
|
if (var->regs[regset].space != space)
|
|
|
|
|
continue;
|
|
|
|
|
|
2023-08-29 10:13:35 -07:00
|
|
|
start = var->regs[regset].index;
|
vkd3d-shader/hlsl: Rename hlsl_reg.bind_count to hlsl_reg.allocation_size.
We have to distinguish between the "bind count" and the "allocation size"
of variables.
The "allocation size" affects the starting register id for the resource to
be allocated next, while the "bind count" is determined by the last field
actually used. The former may be larger than the latter.
What we are currently calling hlsl_reg.bind_count is actually the
"allocation size", so a rename is in order.
The real "bind count", which will be introduced in following patches,
is important because it is what should be shown in the RDEF table and
some resource allocation rules depend on it.
For instance, for this shader:
texture2D texs[3];
texture2D tex;
float4 main() : sv_target
{
return texs[0].Load(int3(0, 0, 0)) + tex.Load(int3(0, 0, 0));
}
the variable "texs" has a "bind count" of 1, but an "allocation size" of
3:
// Resource Bindings:
//
// Name Type Format Dim HLSL Bind Count
// ------------------------------ ---------- ------- ----------- -------------- ------
// texs texture float4 2d t0 1
// tex texture float4 2d t3 1
2023-08-04 10:21:27 -07:00
|
|
|
count = var->regs[regset].allocation_size;
|
2023-06-21 10:00:39 -07:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2022-11-24 14:36:20 -08:00
|
|
|
continue;
|
2023-06-21 10:00:39 -07:00
|
|
|
}
|
2023-04-25 09:41:38 -07:00
|
|
|
|
|
|
|
|
if (start <= index && index < start + count)
|
2021-10-11 19:58:46 -07:00
|
|
|
return var;
|
|
|
|
|
}
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-30 13:27:32 -08:00
|
|
|
static void allocate_objects(struct hlsl_ctx *ctx, enum hlsl_regset regset)
|
2021-11-19 07:44:48 -08:00
|
|
|
{
|
2023-01-30 13:27:32 -08:00
|
|
|
char regset_name = get_regset_name(regset);
|
2023-08-29 10:13:35 -07:00
|
|
|
uint32_t min_index = 0, id = 0;
|
2021-10-11 19:58:46 -07:00
|
|
|
struct hlsl_ir_var *var;
|
2021-08-12 19:26:15 -07:00
|
|
|
|
2023-01-30 13:27:32 -08:00
|
|
|
if (regset == HLSL_REGSET_UAVS)
|
2021-08-12 19:26:15 -07:00
|
|
|
{
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (var->semantic.name && (!ascii_strcasecmp(var->semantic.name, "color")
|
|
|
|
|
|| !ascii_strcasecmp(var->semantic.name, "sv_target")))
|
|
|
|
|
min_index = max(min_index, var->semantic.index + 1);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2021-10-11 19:58:46 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
vkd3d-shader/hlsl: Rename hlsl_reg.bind_count to hlsl_reg.allocation_size.
We have to distinguish between the "bind count" and the "allocation size"
of variables.
The "allocation size" affects the starting register id for the resource to
be allocated next, while the "bind count" is determined by the last field
actually used. The former may be larger than the latter.
What we are currently calling hlsl_reg.bind_count is actually the
"allocation size", so a rename is in order.
The real "bind count", which will be introduced in following patches,
is important because it is what should be shown in the RDEF table and
some resource allocation rules depend on it.
For instance, for this shader:
texture2D texs[3];
texture2D tex;
float4 main() : sv_target
{
return texs[0].Load(int3(0, 0, 0)) + tex.Load(int3(0, 0, 0));
}
the variable "texs" has a "bind count" of 1, but an "allocation size" of
3:
// Resource Bindings:
//
// Name Type Format Dim HLSL Bind Count
// ------------------------------ ---------- ------- ----------- -------------- ------
// texs texture float4 2d t0 1
// tex texture float4 2d t3 1
2023-08-04 10:21:27 -07:00
|
|
|
unsigned int count = var->regs[regset].allocation_size;
|
2023-04-25 09:41:38 -07:00
|
|
|
|
|
|
|
|
if (count == 0)
|
2021-10-11 19:58:46 -07:00
|
|
|
continue;
|
|
|
|
|
|
2023-06-27 11:14:37 -07:00
|
|
|
/* The variable was already allocated if it has a reservation. */
|
2022-11-24 14:36:20 -08:00
|
|
|
if (var->regs[regset].allocated)
|
2021-10-11 19:58:46 -07:00
|
|
|
{
|
2023-04-25 09:41:38 -07:00
|
|
|
const struct hlsl_ir_var *reserved_object, *last_reported = NULL;
|
2023-08-29 10:40:38 -07:00
|
|
|
unsigned int i;
|
2022-11-24 14:36:20 -08:00
|
|
|
|
2023-08-29 10:13:35 -07:00
|
|
|
if (var->regs[regset].index < min_index)
|
2021-08-12 19:26:15 -07:00
|
|
|
{
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(regset == HLSL_REGSET_UAVS);
|
2021-08-12 19:26:15 -07:00
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_OVERLAPPING_RESERVATIONS,
|
|
|
|
|
"UAV index (%u) must be higher than the maximum render target index (%u).",
|
2023-08-29 10:13:35 -07:00
|
|
|
var->regs[regset].index, min_index - 1);
|
2023-04-25 09:41:38 -07:00
|
|
|
continue;
|
2021-08-12 19:26:15 -07:00
|
|
|
}
|
2023-04-25 09:41:38 -07:00
|
|
|
|
|
|
|
|
for (i = 0; i < count; ++i)
|
2021-10-11 19:58:46 -07:00
|
|
|
{
|
2023-08-29 10:40:38 -07:00
|
|
|
unsigned int space = var->regs[regset].space;
|
|
|
|
|
unsigned int index = var->regs[regset].index + i;
|
2021-10-11 19:58:46 -07:00
|
|
|
|
2023-06-27 11:14:37 -07:00
|
|
|
/* get_allocated_object() may return "var" itself, but we
|
|
|
|
|
* actually want that, otherwise we'll end up reporting the
|
|
|
|
|
* same conflict between the same two variables twice. */
|
2023-08-29 10:40:38 -07:00
|
|
|
reserved_object = get_allocated_object(ctx, regset, space, index, true);
|
2023-04-25 09:41:38 -07:00
|
|
|
if (reserved_object && reserved_object != var && reserved_object != last_reported)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_OVERLAPPING_RESERVATIONS,
|
2023-08-29 10:40:38 -07:00
|
|
|
"Multiple variables bound to space %u, %c%u.", regset_name, space, index);
|
2023-04-25 09:41:38 -07:00
|
|
|
hlsl_note(ctx, &reserved_object->loc, VKD3D_SHADER_LOG_ERROR,
|
2023-08-29 10:40:38 -07:00
|
|
|
"Variable '%s' is already bound to space %u, %c%u.",
|
|
|
|
|
reserved_object->name, regset_name, space, index);
|
2023-04-25 09:41:38 -07:00
|
|
|
last_reported = reserved_object;
|
|
|
|
|
}
|
|
|
|
|
}
|
2023-08-29 10:13:35 -07:00
|
|
|
|
|
|
|
|
if (hlsl_version_ge(ctx, 5, 1))
|
|
|
|
|
var->regs[regset].id = id++;
|
|
|
|
|
else
|
|
|
|
|
var->regs[regset].id = var->regs[regset].index;
|
2023-08-29 10:40:38 -07:00
|
|
|
TRACE("Allocated reserved variable %s to space %u, indices %c%u-%c%u, id %u.\n",
|
|
|
|
|
var->name, var->regs[regset].space, regset_name, var->regs[regset].index,
|
2023-08-29 10:13:35 -07:00
|
|
|
regset_name, var->regs[regset].index + count, var->regs[regset].id);
|
2021-10-11 19:58:46 -07:00
|
|
|
}
|
2022-11-24 14:36:20 -08:00
|
|
|
else
|
2021-10-11 19:58:46 -07:00
|
|
|
{
|
2023-04-25 09:41:38 -07:00
|
|
|
unsigned int index = min_index;
|
|
|
|
|
unsigned int available = 0;
|
|
|
|
|
|
|
|
|
|
while (available < count)
|
|
|
|
|
{
|
2023-08-29 10:40:38 -07:00
|
|
|
if (get_allocated_object(ctx, regset, 0, index, false))
|
2023-04-25 09:41:38 -07:00
|
|
|
available = 0;
|
|
|
|
|
else
|
|
|
|
|
++available;
|
2021-10-11 19:58:46 -07:00
|
|
|
++index;
|
2023-04-25 09:41:38 -07:00
|
|
|
}
|
|
|
|
|
index -= count;
|
2021-10-11 19:58:46 -07:00
|
|
|
|
2023-08-29 10:40:38 -07:00
|
|
|
var->regs[regset].space = 0;
|
2023-08-29 10:13:35 -07:00
|
|
|
var->regs[regset].index = index;
|
|
|
|
|
if (hlsl_version_ge(ctx, 5, 1))
|
|
|
|
|
var->regs[regset].id = id++;
|
|
|
|
|
else
|
|
|
|
|
var->regs[regset].id = var->regs[regset].index;
|
2022-11-24 12:03:54 -08:00
|
|
|
var->regs[regset].allocated = true;
|
2023-08-29 10:40:38 -07:00
|
|
|
TRACE("Allocated variable %s to space 0, indices %c%u-%c%u, id %u.\n", var->name,
|
2023-08-29 10:13:35 -07:00
|
|
|
regset_name, index, regset_name, index + count, var->regs[regset].id);
|
2021-10-11 19:58:46 -07:00
|
|
|
++index;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2022-07-20 12:37:07 -07:00
|
|
|
bool hlsl_component_index_range_from_deref(struct hlsl_ctx *ctx, const struct hlsl_deref *deref,
|
|
|
|
|
unsigned int *start, unsigned int *count)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_type *type = deref->var->data_type;
|
|
|
|
|
unsigned int i, k;
|
|
|
|
|
|
|
|
|
|
*start = 0;
|
|
|
|
|
*count = 0;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < deref->path_len; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *path_node = deref->path[i].node;
|
|
|
|
|
unsigned int idx = 0;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(path_node);
|
2022-07-20 12:37:07 -07:00
|
|
|
if (path_node->type != HLSL_IR_CONSTANT)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
/* We should always have generated a cast to UINT. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(path_node->data_type->class == HLSL_CLASS_SCALAR
|
2024-02-27 15:30:51 -08:00
|
|
|
&& path_node->data_type->e.numeric.type == HLSL_TYPE_UINT);
|
2022-07-20 12:37:07 -07:00
|
|
|
|
2022-11-11 16:39:55 -08:00
|
|
|
idx = hlsl_ir_constant(path_node)->value.u[0].u;
|
2022-07-20 12:37:07 -07:00
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
switch (type->class)
|
2022-07-20 12:37:07 -07:00
|
|
|
{
|
|
|
|
|
case HLSL_CLASS_VECTOR:
|
|
|
|
|
if (idx >= type->dimx)
|
|
|
|
|
return false;
|
|
|
|
|
*start += idx;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_MATRIX:
|
|
|
|
|
if (idx >= hlsl_type_major_size(type))
|
|
|
|
|
return false;
|
|
|
|
|
if (hlsl_type_is_row_major(type))
|
|
|
|
|
*start += idx * type->dimx;
|
|
|
|
|
else
|
|
|
|
|
*start += idx * type->dimy;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_ARRAY:
|
|
|
|
|
if (idx >= type->e.array.elements_count)
|
|
|
|
|
return false;
|
|
|
|
|
*start += idx * hlsl_type_component_count(type->e.array.type);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_STRUCT:
|
|
|
|
|
for (k = 0; k < idx; ++k)
|
|
|
|
|
*start += hlsl_type_component_count(type->e.record.fields[k].type);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
2022-08-31 04:25:24 -07:00
|
|
|
vkd3d_unreachable();
|
2022-07-20 12:37:07 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
type = hlsl_get_element_type_from_path_index(ctx, type, path_node);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
*count = hlsl_type_component_count(type);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
/* Retrieves true if the index is constant, and false otherwise. In the latter case, the maximum
|
|
|
|
|
* possible index is retrieved, assuming there is not out-of-bounds access. */
|
2022-11-25 14:47:56 -08:00
|
|
|
bool hlsl_regset_index_from_deref(struct hlsl_ctx *ctx, const struct hlsl_deref *deref,
|
|
|
|
|
enum hlsl_regset regset, unsigned int *index)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_type *type = deref->var->data_type;
|
2024-05-06 23:45:14 -07:00
|
|
|
bool index_is_constant = true;
|
2022-11-25 14:47:56 -08:00
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
*index = 0;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < deref->path_len; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *path_node = deref->path[i].node;
|
|
|
|
|
unsigned int idx = 0;
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(path_node);
|
2024-05-06 23:45:14 -07:00
|
|
|
if (path_node->type == HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
|
|
|
|
/* We should always have generated a cast to UINT. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(path_node->data_type->class == HLSL_CLASS_SCALAR
|
2024-05-06 23:45:14 -07:00
|
|
|
&& path_node->data_type->e.numeric.type == HLSL_TYPE_UINT);
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
idx = hlsl_ir_constant(path_node)->value.u[0].u;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
switch (type->class)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_CLASS_ARRAY:
|
|
|
|
|
if (idx >= type->e.array.elements_count)
|
|
|
|
|
return false;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
*index += idx * type->e.array.type->reg_size[regset];
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_STRUCT:
|
|
|
|
|
*index += type->e.record.fields[idx].reg_offset[regset];
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_CLASS_MATRIX:
|
|
|
|
|
*index += 4 * idx;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else
|
2022-11-25 14:47:56 -08:00
|
|
|
{
|
2024-05-06 23:45:14 -07:00
|
|
|
index_is_constant = false;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
switch (type->class)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_CLASS_ARRAY:
|
|
|
|
|
idx = type->e.array.elements_count - 1;
|
|
|
|
|
*index += idx * type->e.array.type->reg_size[regset];
|
|
|
|
|
break;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
case HLSL_CLASS_MATRIX:
|
|
|
|
|
idx = hlsl_type_major_size(type) - 1;
|
|
|
|
|
*index += idx * 4;
|
|
|
|
|
break;
|
2022-11-25 14:47:56 -08:00
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
2022-11-25 14:47:56 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
type = hlsl_get_element_type_from_path_index(ctx, type, path_node);
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(!(regset <= HLSL_REGSET_LAST_OBJECT) || (type->reg_size[regset] == 1));
|
|
|
|
|
VKD3D_ASSERT(!(regset == HLSL_REGSET_NUMERIC) || type->reg_size[regset] <= 4);
|
2024-05-06 23:45:14 -07:00
|
|
|
return index_is_constant;
|
2022-11-25 14:47:56 -08:00
|
|
|
}
|
|
|
|
|
|
2022-02-24 06:06:15 -08:00
|
|
|
bool hlsl_offset_from_deref(struct hlsl_ctx *ctx, const struct hlsl_deref *deref, unsigned int *offset)
|
2021-05-10 21:36:08 -07:00
|
|
|
{
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
enum hlsl_regset regset = hlsl_deref_get_regset(ctx, deref);
|
2023-10-06 09:56:24 -07:00
|
|
|
struct hlsl_ir_node *offset_node = deref->rel_offset.node;
|
2022-10-28 08:23:05 -07:00
|
|
|
unsigned int size;
|
2021-09-23 06:42:50 -07:00
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
*offset = deref->const_offset;
|
2021-05-10 21:36:08 -07:00
|
|
|
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
if (offset_node)
|
|
|
|
|
{
|
|
|
|
|
/* We should always have generated a cast to UINT. */
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(offset_node->data_type->class == HLSL_CLASS_SCALAR
|
2024-02-27 15:30:51 -08:00
|
|
|
&& offset_node->data_type->e.numeric.type == HLSL_TYPE_UINT);
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(offset_node->type != HLSL_IR_CONSTANT);
|
2023-05-11 12:18:02 -07:00
|
|
|
return false;
|
vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:
- Since SM4's relative addressing (the capability of passing a register
as an index to another register) only has whole-register granularity,
we will need to make the offset node express the offset in
whole-registers and specify the register component in this uint,
otherwise we would have to add additional / and % operations in the
output binary.
- If, after we apply constant folding and copy propagation, we determine
that the offset is a single constant node, we can store all the offset
in this uint constant, and remove the offset src.
This allows DCE to remove a good bunch of the nodes previously required
only for the offset constants, which makes the output more liteweight
and readable, and simplifies the implementation of relative addressing
when writing tpf in the following patches.
In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:
var[@42c + 2c]
But, once we express the offset node in whole registers we will remove
the "c" from the node part:
var[@22 + 3c]
2023-10-03 12:47:13 -07:00
|
|
|
}
|
2022-02-24 06:06:15 -08:00
|
|
|
|
2023-05-29 14:34:03 -07:00
|
|
|
size = deref->var->data_type->reg_size[regset];
|
2022-10-28 08:23:05 -07:00
|
|
|
if (*offset >= size)
|
2022-02-24 06:06:15 -08:00
|
|
|
{
|
2024-07-08 12:24:53 -07:00
|
|
|
/* FIXME: Report a more specific location for the constant deref. */
|
|
|
|
|
hlsl_error(ctx, &deref->var->loc, VKD3D_SHADER_ERROR_HLSL_OFFSET_OUT_OF_BOUNDS,
|
2022-10-28 08:23:05 -07:00
|
|
|
"Dereference is out of bounds. %u/%u", *offset, size);
|
2022-02-24 06:06:15 -08:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2021-11-17 00:47:26 -08:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
unsigned int hlsl_offset_from_deref_safe(struct hlsl_ctx *ctx, const struct hlsl_deref *deref)
|
|
|
|
|
{
|
|
|
|
|
unsigned int offset;
|
|
|
|
|
|
2022-02-24 06:06:15 -08:00
|
|
|
if (hlsl_offset_from_deref(ctx, deref, &offset))
|
2021-11-17 00:47:26 -08:00
|
|
|
return offset;
|
|
|
|
|
|
2024-07-08 12:24:53 -07:00
|
|
|
if (deref->rel_offset.node)
|
|
|
|
|
hlsl_fixme(ctx, &deref->rel_offset.node->loc, "Dereference with non-constant offset of type %s.",
|
|
|
|
|
hlsl_node_type_to_string(deref->rel_offset.node->type));
|
2021-11-17 00:47:26 -08:00
|
|
|
|
|
|
|
|
return 0;
|
2021-09-23 06:42:50 -07:00
|
|
|
}
|
|
|
|
|
|
2022-03-10 07:14:05 -08:00
|
|
|
struct hlsl_reg hlsl_reg_from_deref(struct hlsl_ctx *ctx, const struct hlsl_deref *deref)
|
2021-09-23 06:42:50 -07:00
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_var *var = deref->var;
|
2022-11-24 12:03:54 -08:00
|
|
|
struct hlsl_reg ret = var->regs[HLSL_REGSET_NUMERIC];
|
2021-11-17 00:47:26 -08:00
|
|
|
unsigned int offset = hlsl_offset_from_deref_safe(ctx, deref);
|
2021-05-10 21:36:08 -07:00
|
|
|
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(deref->data_type);
|
|
|
|
|
VKD3D_ASSERT(hlsl_is_numeric_type(deref->data_type));
|
2022-10-28 08:23:05 -07:00
|
|
|
|
2023-08-29 10:13:35 -07:00
|
|
|
ret.index += offset / 4;
|
2021-05-10 21:36:08 -07:00
|
|
|
ret.id += offset / 4;
|
|
|
|
|
|
2022-03-10 07:14:05 -08:00
|
|
|
ret.writemask = 0xf & (0xf << (offset % 4));
|
2022-11-24 12:03:54 -08:00
|
|
|
if (var->regs[HLSL_REGSET_NUMERIC].writemask)
|
|
|
|
|
ret.writemask = hlsl_combine_writemasks(var->regs[HLSL_REGSET_NUMERIC].writemask, ret.writemask);
|
2022-03-10 07:14:05 -08:00
|
|
|
|
2021-05-10 21:36:08 -07:00
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-24 07:08:53 -07:00
|
|
|
static const char *get_string_argument_value(struct hlsl_ctx *ctx, const struct hlsl_attribute *attr, unsigned int i)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_node *instr = attr->args[i].node;
|
|
|
|
|
const struct hlsl_type *type = instr->data_type;
|
|
|
|
|
|
|
|
|
|
if (type->class != HLSL_CLASS_STRING)
|
|
|
|
|
{
|
|
|
|
|
struct vkd3d_string_buffer *string;
|
|
|
|
|
|
|
|
|
|
if ((string = hlsl_type_to_string(ctx, type)))
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE,
|
|
|
|
|
"Wrong type for the argument %u of [%s]: expected string, but got %s.",
|
|
|
|
|
i, attr->name, string->buffer);
|
|
|
|
|
hlsl_release_string_buffer(ctx, string);
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return hlsl_ir_string_constant(instr)->string;
|
|
|
|
|
}
|
|
|
|
|
|
2021-08-16 15:29:34 -07:00
|
|
|
static void parse_numthreads_attribute(struct hlsl_ctx *ctx, const struct hlsl_attribute *attr)
|
|
|
|
|
{
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
ctx->found_numthreads = 1;
|
|
|
|
|
|
|
|
|
|
if (attr->args_count != 3)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &attr->loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT,
|
|
|
|
|
"Expected 3 parameters for [numthreads] attribute, but got %u.", attr->args_count);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < attr->args_count; ++i)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_node *instr = attr->args[i].node;
|
|
|
|
|
const struct hlsl_type *type = instr->data_type;
|
|
|
|
|
const struct hlsl_ir_constant *constant;
|
|
|
|
|
|
2022-11-11 17:31:55 -08:00
|
|
|
if (type->class != HLSL_CLASS_SCALAR
|
2024-02-27 15:30:51 -08:00
|
|
|
|| (type->e.numeric.type != HLSL_TYPE_INT && type->e.numeric.type != HLSL_TYPE_UINT))
|
2021-08-16 15:29:34 -07:00
|
|
|
{
|
|
|
|
|
struct vkd3d_string_buffer *string;
|
|
|
|
|
|
|
|
|
|
if ((string = hlsl_type_to_string(ctx, type)))
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE,
|
|
|
|
|
"Wrong type for argument %u of [numthreads]: expected int or uint, but got %s.",
|
|
|
|
|
i, string->buffer);
|
|
|
|
|
hlsl_release_string_buffer(ctx, string);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Non-constant expression in [numthreads] initializer.");
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
constant = hlsl_ir_constant(instr);
|
|
|
|
|
|
2024-02-27 15:30:51 -08:00
|
|
|
if ((type->e.numeric.type == HLSL_TYPE_INT && constant->value.u[0].i <= 0)
|
|
|
|
|
|| (type->e.numeric.type == HLSL_TYPE_UINT && !constant->value.u[0].u))
|
2021-08-16 15:29:34 -07:00
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_THREAD_COUNT,
|
|
|
|
|
"Thread count must be a positive integer.");
|
|
|
|
|
|
2022-11-11 16:39:55 -08:00
|
|
|
ctx->thread_count[i] = constant->value.u[0].u;
|
2021-08-16 15:29:34 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-24 07:08:53 -07:00
|
|
|
static void parse_domain_attribute(struct hlsl_ctx *ctx, const struct hlsl_attribute *attr)
|
|
|
|
|
{
|
|
|
|
|
const char *value;
|
|
|
|
|
|
|
|
|
|
if (attr->args_count != 1)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &attr->loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT,
|
|
|
|
|
"Expected 1 parameter for [domain] attribute, but got %u.", attr->args_count);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(value = get_string_argument_value(ctx, attr, 0)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if (!strcmp(value, "isoline"))
|
|
|
|
|
ctx->domain = VKD3D_TESSELLATOR_DOMAIN_LINE;
|
|
|
|
|
else if (!strcmp(value, "tri"))
|
|
|
|
|
ctx->domain = VKD3D_TESSELLATOR_DOMAIN_TRIANGLE;
|
|
|
|
|
else if (!strcmp(value, "quad"))
|
|
|
|
|
ctx->domain = VKD3D_TESSELLATOR_DOMAIN_QUAD;
|
|
|
|
|
else
|
|
|
|
|
hlsl_error(ctx, &attr->args[0].node->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_DOMAIN,
|
|
|
|
|
"Invalid tessellator domain \"%s\": expected \"isoline\", \"tri\", or \"quad\".",
|
|
|
|
|
value);
|
|
|
|
|
}
|
|
|
|
|
|
2024-09-03 12:30:40 -07:00
|
|
|
static void parse_outputcontrolpoints_attribute(struct hlsl_ctx *ctx, const struct hlsl_attribute *attr)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_ir_node *instr;
|
|
|
|
|
const struct hlsl_type *type;
|
|
|
|
|
const struct hlsl_ir_constant *constant;
|
|
|
|
|
|
|
|
|
|
if (attr->args_count != 1)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &attr->loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT,
|
|
|
|
|
"Expected 1 parameter for [outputcontrolpoints] attribute, but got %u.", attr->args_count);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
instr = attr->args[0].node;
|
|
|
|
|
type = instr->data_type;
|
|
|
|
|
|
|
|
|
|
if (type->class != HLSL_CLASS_SCALAR
|
|
|
|
|
|| (type->e.numeric.type != HLSL_TYPE_INT && type->e.numeric.type != HLSL_TYPE_UINT))
|
|
|
|
|
{
|
|
|
|
|
struct vkd3d_string_buffer *string;
|
|
|
|
|
|
|
|
|
|
if ((string = hlsl_type_to_string(ctx, type)))
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE,
|
|
|
|
|
"Wrong type for argument 0 of [outputcontrolpoints]: expected int or uint, but got %s.",
|
|
|
|
|
string->buffer);
|
|
|
|
|
hlsl_release_string_buffer(ctx, string);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_CONSTANT)
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Non-constant expression in [outputcontrolpoints] initializer.");
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
constant = hlsl_ir_constant(instr);
|
|
|
|
|
|
|
|
|
|
if ((type->e.numeric.type == HLSL_TYPE_INT && constant->value.u[0].i < 0)
|
|
|
|
|
|| constant->value.u[0].u > 32)
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_CONTROL_POINT_COUNT,
|
|
|
|
|
"Output control point count must be between 0 and 32.");
|
|
|
|
|
|
|
|
|
|
ctx->output_control_point_count = constant->value.u[0].u;
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-24 07:12:34 -07:00
|
|
|
static void parse_outputtopology_attribute(struct hlsl_ctx *ctx, const struct hlsl_attribute *attr)
|
|
|
|
|
{
|
|
|
|
|
const char *value;
|
|
|
|
|
|
|
|
|
|
if (attr->args_count != 1)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &attr->loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT,
|
|
|
|
|
"Expected 1 parameter for [outputtopology] attribute, but got %u.", attr->args_count);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(value = get_string_argument_value(ctx, attr, 0)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if (!strcmp(value, "point"))
|
|
|
|
|
ctx->output_primitive = VKD3D_SHADER_TESSELLATOR_OUTPUT_POINT;
|
|
|
|
|
else if (!strcmp(value, "line"))
|
|
|
|
|
ctx->output_primitive = VKD3D_SHADER_TESSELLATOR_OUTPUT_LINE;
|
|
|
|
|
else if (!strcmp(value, "triangle_cw"))
|
|
|
|
|
ctx->output_primitive = VKD3D_SHADER_TESSELLATOR_OUTPUT_TRIANGLE_CW;
|
|
|
|
|
else if (!strcmp(value, "triangle_ccw"))
|
|
|
|
|
ctx->output_primitive = VKD3D_SHADER_TESSELLATOR_OUTPUT_TRIANGLE_CCW;
|
|
|
|
|
else
|
|
|
|
|
hlsl_error(ctx, &attr->args[0].node->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_OUTPUT_PRIMITIVE,
|
|
|
|
|
"Invalid tessellator output topology \"%s\": "
|
|
|
|
|
"expected \"point\", \"line\", \"triangle_cw\", or \"triangle_ccw\".", value);
|
|
|
|
|
}
|
|
|
|
|
|
2024-09-03 12:09:24 -07:00
|
|
|
static void parse_partitioning_attribute(struct hlsl_ctx *ctx, const struct hlsl_attribute *attr)
|
|
|
|
|
{
|
|
|
|
|
const char *value;
|
|
|
|
|
|
|
|
|
|
if (attr->args_count != 1)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &attr->loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT,
|
|
|
|
|
"Expected 1 parameter for [partitioning] attribute, but got %u.", attr->args_count);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(value = get_string_argument_value(ctx, attr, 0)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if (!strcmp(value, "integer"))
|
|
|
|
|
ctx->partitioning = VKD3D_SHADER_TESSELLATOR_PARTITIONING_INTEGER;
|
|
|
|
|
else if (!strcmp(value, "pow2"))
|
|
|
|
|
ctx->partitioning = VKD3D_SHADER_TESSELLATOR_PARTITIONING_POW2;
|
|
|
|
|
else if (!strcmp(value, "fractional_even"))
|
|
|
|
|
ctx->partitioning = VKD3D_SHADER_TESSELLATOR_PARTITIONING_FRACTIONAL_EVEN;
|
|
|
|
|
else if (!strcmp(value, "fractional_odd"))
|
|
|
|
|
ctx->partitioning = VKD3D_SHADER_TESSELLATOR_PARTITIONING_FRACTIONAL_ODD;
|
|
|
|
|
else
|
|
|
|
|
hlsl_error(ctx, &attr->args[0].node->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_PARTITIONING,
|
|
|
|
|
"Invalid tessellator partitioning \"%s\": "
|
|
|
|
|
"expected \"integer\", \"pow2\", \"fractional_even\", or \"fractional_odd\".", value);
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-24 08:54:34 -07:00
|
|
|
static void parse_patchconstantfunc_attribute(struct hlsl_ctx *ctx, const struct hlsl_attribute *attr)
|
|
|
|
|
{
|
|
|
|
|
const char *name;
|
|
|
|
|
struct hlsl_ir_function *func;
|
|
|
|
|
struct hlsl_ir_function_decl *decl;
|
|
|
|
|
|
|
|
|
|
if (attr->args_count != 1)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &attr->loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT,
|
|
|
|
|
"Expected 1 parameter for [patchconstantfunc] attribute, but got %u.", attr->args_count);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(name = get_string_argument_value(ctx, attr, 0)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
ctx->patch_constant_func = NULL;
|
|
|
|
|
if ((func = hlsl_get_function(ctx, name)))
|
|
|
|
|
{
|
|
|
|
|
/* Pick the last overload with a body. */
|
|
|
|
|
LIST_FOR_EACH_ENTRY_REV(decl, &func->overloads, struct hlsl_ir_function_decl, entry)
|
|
|
|
|
{
|
|
|
|
|
if (decl->has_body)
|
|
|
|
|
{
|
|
|
|
|
ctx->patch_constant_func = decl;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!ctx->patch_constant_func)
|
|
|
|
|
hlsl_error(ctx, &attr->loc, VKD3D_SHADER_ERROR_HLSL_NOT_DEFINED,
|
|
|
|
|
"Patch constant function \"%s\" is not defined.", name);
|
|
|
|
|
}
|
|
|
|
|
|
2024-09-06 13:03:05 -07:00
|
|
|
static void parse_entry_function_attributes(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry_func)
|
2024-08-24 07:21:06 -07:00
|
|
|
{
|
|
|
|
|
const struct hlsl_profile_info *profile = ctx->profile;
|
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < entry_func->attr_count; ++i)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_attribute *attr = entry_func->attrs[i];
|
|
|
|
|
|
|
|
|
|
if (!strcmp(attr->name, "numthreads") && profile->type == VKD3D_SHADER_TYPE_COMPUTE)
|
|
|
|
|
parse_numthreads_attribute(ctx, attr);
|
2024-08-24 07:08:53 -07:00
|
|
|
else if (!strcmp(attr->name, "domain")
|
|
|
|
|
&& (profile->type == VKD3D_SHADER_TYPE_HULL || profile->type == VKD3D_SHADER_TYPE_DOMAIN))
|
|
|
|
|
parse_domain_attribute(ctx, attr);
|
2024-09-03 12:30:40 -07:00
|
|
|
else if (!strcmp(attr->name, "outputcontrolpoints") && profile->type == VKD3D_SHADER_TYPE_HULL)
|
|
|
|
|
parse_outputcontrolpoints_attribute(ctx, attr);
|
2024-08-24 07:12:34 -07:00
|
|
|
else if (!strcmp(attr->name, "outputtopology") && profile->type == VKD3D_SHADER_TYPE_HULL)
|
|
|
|
|
parse_outputtopology_attribute(ctx, attr);
|
2024-09-03 12:09:24 -07:00
|
|
|
else if (!strcmp(attr->name, "partitioning") && profile->type == VKD3D_SHADER_TYPE_HULL)
|
|
|
|
|
parse_partitioning_attribute(ctx, attr);
|
2024-08-24 08:54:34 -07:00
|
|
|
else if (!strcmp(attr->name, "patchconstantfunc") && profile->type == VKD3D_SHADER_TYPE_HULL)
|
|
|
|
|
parse_patchconstantfunc_attribute(ctx, attr);
|
2024-09-06 13:03:05 -07:00
|
|
|
else if (!strcmp(attr->name, "earlydepthstencil") && profile->type == VKD3D_SHADER_TYPE_PIXEL)
|
|
|
|
|
entry_func->early_depth_test = true;
|
2024-08-24 07:21:06 -07:00
|
|
|
else
|
|
|
|
|
hlsl_warning(ctx, &entry_func->attrs[i]->loc, VKD3D_SHADER_WARNING_HLSL_UNKNOWN_ATTRIBUTE,
|
|
|
|
|
"Ignoring unknown attribute \"%s\".", entry_func->attrs[i]->name);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-27 07:58:02 -07:00
|
|
|
static void validate_hull_shader_attributes(struct hlsl_ctx *ctx, const struct hlsl_ir_function_decl *entry_func)
|
|
|
|
|
{
|
|
|
|
|
if (ctx->domain == VKD3D_TESSELLATOR_DOMAIN_INVALID)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_ATTRIBUTE,
|
|
|
|
|
"Entry point \"%s\" is missing a [domain] attribute.", entry_func->func->name);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (ctx->output_control_point_count == UINT_MAX)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_ATTRIBUTE,
|
|
|
|
|
"Entry point \"%s\" is missing a [outputcontrolpoints] attribute.", entry_func->func->name);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!ctx->output_primitive)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_ATTRIBUTE,
|
|
|
|
|
"Entry point \"%s\" is missing a [outputtopology] attribute.", entry_func->func->name);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!ctx->partitioning)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_ATTRIBUTE,
|
|
|
|
|
"Entry point \"%s\" is missing a [partitioning] attribute.", entry_func->func->name);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!ctx->patch_constant_func)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_ATTRIBUTE,
|
|
|
|
|
"Entry point \"%s\" is missing a [patchconstantfunc] attribute.", entry_func->func->name);
|
|
|
|
|
}
|
|
|
|
|
else if (ctx->patch_constant_func == entry_func)
|
|
|
|
|
{
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_RECURSIVE_CALL,
|
|
|
|
|
"Patch constant function cannot be the entry point function.");
|
|
|
|
|
/* Native returns E_NOTIMPL instead of E_FAIL here. */
|
|
|
|
|
ctx->result = VKD3D_ERROR_NOT_IMPLEMENTED;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
switch (ctx->domain)
|
|
|
|
|
{
|
|
|
|
|
case VKD3D_TESSELLATOR_DOMAIN_LINE:
|
|
|
|
|
if (ctx->output_primitive == VKD3D_SHADER_TESSELLATOR_OUTPUT_TRIANGLE_CW
|
|
|
|
|
|| ctx->output_primitive == VKD3D_SHADER_TESSELLATOR_OUTPUT_TRIANGLE_CCW)
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_OUTPUT_PRIMITIVE,
|
|
|
|
|
"Triangle output topologies are not available for isoline domains.");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case VKD3D_TESSELLATOR_DOMAIN_TRIANGLE:
|
|
|
|
|
if (ctx->output_primitive == VKD3D_SHADER_TESSELLATOR_OUTPUT_LINE)
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_OUTPUT_PRIMITIVE,
|
|
|
|
|
"Line output topologies are not available for triangle domains.");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case VKD3D_TESSELLATOR_DOMAIN_QUAD:
|
|
|
|
|
if (ctx->output_primitive == VKD3D_SHADER_TESSELLATOR_OUTPUT_LINE)
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_OUTPUT_PRIMITIVE,
|
|
|
|
|
"Line output topologies are not available for quad domains.");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-11 05:29:25 -07:00
|
|
|
static void remove_unreachable_code(struct hlsl_ctx *ctx, struct hlsl_block *body)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr, *next;
|
|
|
|
|
struct hlsl_block block;
|
|
|
|
|
struct list *start;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY_SAFE(instr, next, &body->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
if (instr->type == HLSL_IR_IF)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
|
|
|
|
|
|
|
|
|
remove_unreachable_code(ctx, &iff->then_block);
|
|
|
|
|
remove_unreachable_code(ctx, &iff->else_block);
|
|
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_LOOP)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_loop *loop = hlsl_ir_loop(instr);
|
|
|
|
|
|
|
|
|
|
remove_unreachable_code(ctx, &loop->body);
|
|
|
|
|
}
|
|
|
|
|
else if (instr->type == HLSL_IR_SWITCH)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
remove_unreachable_code(ctx, &c->body);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Remove instructions past unconditional jumps. */
|
|
|
|
|
LIST_FOR_EACH_ENTRY(instr, &body->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_jump *jump;
|
|
|
|
|
|
|
|
|
|
if (instr->type != HLSL_IR_JUMP)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
jump = hlsl_ir_jump(instr);
|
|
|
|
|
if (jump->type != HLSL_IR_JUMP_BREAK && jump->type != HLSL_IR_JUMP_CONTINUE)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (!(start = list_next(&body->instrs, &instr->entry)))
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
hlsl_block_init(&block);
|
|
|
|
|
list_move_slice_tail(&block.instrs, start, list_tail(&body->instrs));
|
|
|
|
|
hlsl_block_cleanup(&block);
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-09-22 01:35:16 -07:00
|
|
|
void hlsl_lower_index_loads(struct hlsl_ctx *ctx, struct hlsl_block *body)
|
|
|
|
|
{
|
|
|
|
|
lower_ir(ctx, lower_index_loads, body);
|
|
|
|
|
}
|
|
|
|
|
|
2024-04-26 09:15:52 -07:00
|
|
|
void hlsl_run_const_passes(struct hlsl_ctx *ctx, struct hlsl_block *body)
|
|
|
|
|
{
|
|
|
|
|
bool progress;
|
|
|
|
|
|
2024-04-26 11:46:56 -07:00
|
|
|
lower_ir(ctx, lower_matrix_swizzles, body);
|
|
|
|
|
|
2024-04-26 09:15:52 -07:00
|
|
|
lower_ir(ctx, lower_broadcasts, body);
|
|
|
|
|
while (hlsl_transform_ir(ctx, fold_redundant_casts, body, NULL));
|
|
|
|
|
do
|
|
|
|
|
{
|
|
|
|
|
progress = hlsl_transform_ir(ctx, split_array_copies, body, NULL);
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, split_struct_copies, body, NULL);
|
|
|
|
|
}
|
|
|
|
|
while (progress);
|
|
|
|
|
hlsl_transform_ir(ctx, split_matrix_copies, body, NULL);
|
|
|
|
|
|
|
|
|
|
lower_ir(ctx, lower_narrowing_casts, body);
|
|
|
|
|
lower_ir(ctx, lower_int_dot, body);
|
|
|
|
|
lower_ir(ctx, lower_int_division, body);
|
|
|
|
|
lower_ir(ctx, lower_int_modulus, body);
|
|
|
|
|
lower_ir(ctx, lower_int_abs, body);
|
|
|
|
|
lower_ir(ctx, lower_casts_to_bool, body);
|
|
|
|
|
lower_ir(ctx, lower_float_modulus, body);
|
|
|
|
|
hlsl_transform_ir(ctx, fold_redundant_casts, body, NULL);
|
|
|
|
|
|
|
|
|
|
do
|
|
|
|
|
{
|
|
|
|
|
progress = hlsl_transform_ir(ctx, hlsl_fold_constant_exprs, body, NULL);
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, hlsl_fold_constant_identities, body, NULL);
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, hlsl_fold_constant_swizzles, body, NULL);
|
|
|
|
|
progress |= hlsl_copy_propagation_execute(ctx, body);
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, fold_swizzle_chains, body, NULL);
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, remove_trivial_swizzles, body, NULL);
|
|
|
|
|
progress |= hlsl_transform_ir(ctx, remove_trivial_conditional_branches, body, NULL);
|
|
|
|
|
} while (progress);
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-21 11:51:17 -07:00
|
|
|
static void sm1_generate_vsir_signature_entry(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vsir_program *program, bool output, struct hlsl_ir_var *var)
|
|
|
|
|
{
|
|
|
|
|
enum vkd3d_shader_sysval_semantic sysval = VKD3D_SHADER_SV_NONE;
|
|
|
|
|
enum vkd3d_shader_register_type type;
|
|
|
|
|
struct shader_signature *signature;
|
|
|
|
|
struct signature_element *element;
|
|
|
|
|
unsigned int register_index, mask;
|
|
|
|
|
|
|
|
|
|
if ((!output && !var->last_read) || (output && !var->first_write))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if (output)
|
|
|
|
|
signature = &program->output_signature;
|
|
|
|
|
else
|
|
|
|
|
signature = &program->input_signature;
|
|
|
|
|
|
|
|
|
|
if (!vkd3d_array_reserve((void **)&signature->elements, &signature->elements_capacity,
|
|
|
|
|
signature->element_count + 1, sizeof(*signature->elements)))
|
|
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
element = &signature->elements[signature->element_count++];
|
|
|
|
|
|
|
|
|
|
if (!hlsl_sm1_register_from_semantic(&program->shader_version,
|
|
|
|
|
var->semantic.name, var->semantic.index, output, &type, ®ister_index))
|
|
|
|
|
{
|
2024-08-08 05:33:26 -07:00
|
|
|
enum vkd3d_decl_usage usage;
|
|
|
|
|
unsigned int usage_idx;
|
2024-05-21 11:51:17 -07:00
|
|
|
bool ret;
|
|
|
|
|
|
|
|
|
|
register_index = var->regs[HLSL_REGSET_NUMERIC].id;
|
|
|
|
|
|
|
|
|
|
ret = hlsl_sm1_usage_from_semantic(var->semantic.name, var->semantic.index, &usage, &usage_idx);
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(ret);
|
2024-05-21 11:51:17 -07:00
|
|
|
/* With the exception of vertex POSITION output, none of these are
|
|
|
|
|
* system values. Pixel POSITION input is not equivalent to
|
|
|
|
|
* SV_Position; the closer equivalent is VPOS, which is not declared
|
|
|
|
|
* as a semantic. */
|
|
|
|
|
if (program->shader_version.type == VKD3D_SHADER_TYPE_VERTEX
|
|
|
|
|
&& output && usage == VKD3D_DECL_USAGE_POSITION)
|
|
|
|
|
sysval = VKD3D_SHADER_SV_POSITION;
|
|
|
|
|
}
|
|
|
|
|
mask = (1 << var->data_type->dimx) - 1;
|
|
|
|
|
|
|
|
|
|
memset(element, 0, sizeof(*element));
|
|
|
|
|
if (!(element->semantic_name = vkd3d_strdup(var->semantic.name)))
|
|
|
|
|
{
|
|
|
|
|
--signature->element_count;
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
element->semantic_index = var->semantic.index;
|
|
|
|
|
element->sysval_semantic = sysval;
|
|
|
|
|
element->component_type = VKD3D_SHADER_COMPONENT_FLOAT;
|
|
|
|
|
element->register_index = register_index;
|
|
|
|
|
element->target_location = register_index;
|
|
|
|
|
element->register_count = 1;
|
|
|
|
|
element->mask = mask;
|
|
|
|
|
element->used_mask = mask;
|
|
|
|
|
if (program->shader_version.type == VKD3D_SHADER_TYPE_PIXEL && !output)
|
|
|
|
|
element->interpolation_mode = VKD3DSIM_LINEAR;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void sm1_generate_vsir_signature(struct hlsl_ctx *ctx, struct vsir_program *program)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (var->is_input_semantic)
|
|
|
|
|
sm1_generate_vsir_signature_entry(ctx, program, false, var);
|
|
|
|
|
if (var->is_output_semantic)
|
|
|
|
|
sm1_generate_vsir_signature_entry(ctx, program, true, var);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-27 16:58:05 -07:00
|
|
|
static uint32_t sm1_generate_vsir_get_src_swizzle(uint32_t src_writemask, uint32_t dst_writemask)
|
|
|
|
|
{
|
|
|
|
|
uint32_t swizzle;
|
|
|
|
|
|
|
|
|
|
swizzle = hlsl_swizzle_from_writemask(src_writemask);
|
|
|
|
|
swizzle = hlsl_map_swizzle(swizzle, dst_writemask);
|
|
|
|
|
swizzle = vsir_swizzle_from_hlsl(swizzle);
|
|
|
|
|
return swizzle;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-23 09:57:44 -07:00
|
|
|
static void sm1_generate_vsir_constant_defs(struct hlsl_ctx *ctx, struct vsir_program *program,
|
|
|
|
|
struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct vkd3d_shader_instruction_array *instructions = &program->instructions;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
unsigned int i, x;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < ctx->constant_defs.count; ++i)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_constant_register *constant_reg = &ctx->constant_defs.regs[i];
|
|
|
|
|
|
|
|
|
|
if (!shader_instruction_array_reserve(instructions, instructions->count + 1))
|
|
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ins = &instructions->elements[instructions->count];
|
|
|
|
|
if (!vsir_instruction_init_with_params(program, ins, &constant_reg->loc, VKD3DSIH_DEF, 1, 1))
|
|
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
++instructions->count;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_CONST, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
ins->dst[0].reg.dimension = VSIR_DIMENSION_VEC4;
|
|
|
|
|
ins->dst[0].reg.idx[0].offset = constant_reg->index;
|
|
|
|
|
ins->dst[0].write_mask = VKD3DSP_WRITEMASK_ALL;
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_IMMCONST, VKD3D_DATA_FLOAT, 0);
|
|
|
|
|
src_param->reg.type = VKD3DSPR_IMMCONST;
|
|
|
|
|
src_param->reg.precision = VKD3D_SHADER_REGISTER_PRECISION_DEFAULT;
|
|
|
|
|
src_param->reg.non_uniform = false;
|
|
|
|
|
src_param->reg.data_type = VKD3D_DATA_FLOAT;
|
|
|
|
|
src_param->reg.dimension = VSIR_DIMENSION_VEC4;
|
|
|
|
|
for (x = 0; x < 4; ++x)
|
|
|
|
|
src_param->reg.u.immconst_f32[x] = constant_reg->value.f[x];
|
|
|
|
|
src_param->swizzle = VKD3D_SHADER_NO_SWIZZLE;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-23 18:22:04 -07:00
|
|
|
static void sm1_generate_vsir_sampler_dcls(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vsir_program *program, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct vkd3d_shader_instruction_array *instructions = &program->instructions;
|
|
|
|
|
enum vkd3d_shader_resource_type resource_type;
|
|
|
|
|
struct vkd3d_shader_register_range *range;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_semantic *semantic;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
enum hlsl_sampler_dim sampler_dim;
|
|
|
|
|
struct hlsl_ir_var *var;
|
|
|
|
|
unsigned int i, count;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->extern_vars, struct hlsl_ir_var, extern_entry)
|
|
|
|
|
{
|
|
|
|
|
if (!var->regs[HLSL_REGSET_SAMPLERS].allocated)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
count = var->bind_count[HLSL_REGSET_SAMPLERS];
|
|
|
|
|
for (i = 0; i < count; ++i)
|
|
|
|
|
{
|
|
|
|
|
if (var->objects_usage[HLSL_REGSET_SAMPLERS][i].used)
|
|
|
|
|
{
|
|
|
|
|
sampler_dim = var->objects_usage[HLSL_REGSET_SAMPLERS][i].sampler_dim;
|
|
|
|
|
|
|
|
|
|
switch (sampler_dim)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_SAMPLER_DIM_2D:
|
|
|
|
|
resource_type = VKD3D_SHADER_RESOURCE_TEXTURE_2D;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_SAMPLER_DIM_CUBE:
|
|
|
|
|
resource_type = VKD3D_SHADER_RESOURCE_TEXTURE_CUBE;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_SAMPLER_DIM_3D:
|
|
|
|
|
resource_type = VKD3D_SHADER_RESOURCE_TEXTURE_3D;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_SAMPLER_DIM_GENERIC:
|
|
|
|
|
/* These can appear in sm4-style combined sample instructions. */
|
|
|
|
|
hlsl_fixme(ctx, &var->loc, "Generic samplers need to be lowered.");
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!shader_instruction_array_reserve(instructions, instructions->count + 1))
|
|
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ins = &instructions->elements[instructions->count];
|
|
|
|
|
if (!vsir_instruction_init_with_params(program, ins, &var->loc, VKD3DSIH_DCL, 0, 0))
|
|
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
++instructions->count;
|
|
|
|
|
|
|
|
|
|
semantic = &ins->declaration.semantic;
|
|
|
|
|
semantic->resource_type = resource_type;
|
|
|
|
|
|
|
|
|
|
dst_param = &semantic->resource.reg;
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_SAMPLER, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.dimension = VSIR_DIMENSION_NONE;
|
|
|
|
|
dst_param->reg.idx[0].offset = var->regs[HLSL_REGSET_SAMPLERS].index + i;
|
|
|
|
|
dst_param->write_mask = 0;
|
|
|
|
|
range = &semantic->resource.range;
|
|
|
|
|
range->space = 0;
|
|
|
|
|
range->first = range->last = dst_param->reg.idx[0].offset;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-28 19:11:29 -07:00
|
|
|
static struct vkd3d_shader_instruction *generate_vsir_add_program_instruction(
|
|
|
|
|
struct hlsl_ctx *ctx, struct vsir_program *program,
|
|
|
|
|
const struct vkd3d_shader_location *loc, enum vkd3d_shader_opcode opcode,
|
|
|
|
|
unsigned int dst_count, unsigned int src_count)
|
|
|
|
|
{
|
|
|
|
|
struct vkd3d_shader_instruction_array *instructions = &program->instructions;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
|
|
|
|
|
if (!shader_instruction_array_reserve(instructions, instructions->count + 1))
|
|
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
ins = &instructions->elements[instructions->count];
|
|
|
|
|
if (!vsir_instruction_init_with_params(program, ins, loc, opcode, dst_count, src_count))
|
|
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
++instructions->count;
|
|
|
|
|
return ins;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-27 16:58:05 -07:00
|
|
|
static void sm1_generate_vsir_instr_constant(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vsir_program *program, struct hlsl_ir_constant *constant)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr = &constant->node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
|
|
|
|
|
VKD3D_ASSERT(instr->reg.allocated);
|
|
|
|
|
VKD3D_ASSERT(constant->reg.allocated);
|
|
|
|
|
|
2024-05-28 19:11:29 -07:00
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_MOV, 1, 1)))
|
2024-05-27 16:58:05 -07:00
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_CONST, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = constant->reg.id;
|
|
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(constant->reg.writemask, instr->reg.writemask);
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = instr->reg.id;
|
|
|
|
|
dst_param->write_mask = instr->reg.writemask;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
/* Translate ops that can be mapped to a single vsir instruction with only one dst register. */
|
|
|
|
|
static void sm1_generate_vsir_instr_expr_single_instr_op(struct hlsl_ctx *ctx, struct vsir_program *program,
|
2024-05-31 13:06:11 -07:00
|
|
|
struct hlsl_ir_expr *expr, enum vkd3d_shader_opcode opcode, uint32_t src_mod, uint32_t dst_mod,
|
|
|
|
|
bool map_src_swizzles)
|
2024-05-30 15:44:18 -07:00
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr = &expr->node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
unsigned int i, src_count = 0;
|
|
|
|
|
|
|
|
|
|
VKD3D_ASSERT(instr->reg.allocated);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < HLSL_MAX_OPERANDS; ++i)
|
|
|
|
|
{
|
|
|
|
|
if (expr->operands[i].node)
|
|
|
|
|
src_count = i + 1;
|
|
|
|
|
}
|
|
|
|
|
VKD3D_ASSERT(!src_mod || src_count == 1);
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, opcode, 1, src_count)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = instr->reg.id;
|
|
|
|
|
dst_param->write_mask = instr->reg.writemask;
|
|
|
|
|
dst_param->modifiers = dst_mod;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < src_count; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *operand = expr->operands[i].node;
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[i];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = operand->reg.id;
|
2024-05-31 13:06:11 -07:00
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(operand->reg.writemask,
|
|
|
|
|
map_src_swizzles ? dst_param->write_mask : VKD3DSP_WRITEMASK_ALL);
|
2024-05-30 15:44:18 -07:00
|
|
|
src_param->modifiers = src_mod;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-30 23:20:36 -07:00
|
|
|
/* Translate ops that have 1 src and need one instruction for each component in
|
|
|
|
|
* the d3dbc backend. */
|
|
|
|
|
static void sm1_generate_vsir_instr_expr_per_component_instr_op(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vsir_program *program, struct hlsl_ir_expr *expr, enum vkd3d_shader_opcode opcode)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *operand = expr->operands[0].node;
|
|
|
|
|
struct hlsl_ir_node *instr = &expr->node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
uint32_t src_swizzle;
|
|
|
|
|
unsigned int i, c;
|
|
|
|
|
|
|
|
|
|
VKD3D_ASSERT(instr->reg.allocated);
|
|
|
|
|
VKD3D_ASSERT(operand);
|
|
|
|
|
|
|
|
|
|
src_swizzle = sm1_generate_vsir_get_src_swizzle(operand->reg.writemask, instr->reg.writemask);
|
|
|
|
|
for (i = 0; i < 4; ++i)
|
|
|
|
|
{
|
|
|
|
|
if (instr->reg.writemask & (1u << i))
|
|
|
|
|
{
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, opcode, 1, 1)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = instr->reg.id;
|
|
|
|
|
dst_param->write_mask = 1u << i;
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = operand->reg.id;
|
|
|
|
|
c = vsir_swizzle_get_component(src_swizzle, i);
|
|
|
|
|
src_param->swizzle = vsir_swizzle_from_writemask(1u << c);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-07-22 18:15:25 -07:00
|
|
|
static void sm1_generate_vsir_instr_expr_sincos(struct hlsl_ctx *ctx, struct vsir_program *program,
|
|
|
|
|
struct hlsl_ir_expr *expr)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *operand = expr->operands[0].node;
|
|
|
|
|
struct hlsl_ir_node *instr = &expr->node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
unsigned int src_count = 0;
|
|
|
|
|
|
|
|
|
|
VKD3D_ASSERT(instr->reg.allocated);
|
|
|
|
|
src_count = (ctx->profile->major_version < 3) ? 3 : 1;
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_SINCOS, 1, src_count)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = instr->reg.id;
|
|
|
|
|
dst_param->write_mask = instr->reg.writemask;
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = operand->reg.id;
|
|
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(operand->reg.writemask, VKD3DSP_WRITEMASK_ALL);
|
|
|
|
|
|
|
|
|
|
if (ctx->profile->major_version < 3)
|
|
|
|
|
{
|
|
|
|
|
src_param = &ins->src[1];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_CONST, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = ctx->d3dsincosconst1.id;
|
|
|
|
|
src_param->swizzle = VKD3D_SHADER_NO_SWIZZLE;
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[1];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_CONST, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = ctx->d3dsincosconst2.id;
|
|
|
|
|
src_param->swizzle = VKD3D_SHADER_NO_SWIZZLE;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-31 19:33:47 -07:00
|
|
|
static bool sm1_generate_vsir_instr_expr_cast(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vsir_program *program, struct hlsl_ir_expr *expr)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_type *src_type, *dst_type;
|
|
|
|
|
const struct hlsl_ir_node *arg1, *instr;
|
|
|
|
|
|
|
|
|
|
arg1 = expr->operands[0].node;
|
|
|
|
|
src_type = arg1->data_type;
|
|
|
|
|
instr = &expr->node;
|
|
|
|
|
dst_type = instr->data_type;
|
|
|
|
|
|
|
|
|
|
/* Narrowing casts were already lowered. */
|
|
|
|
|
VKD3D_ASSERT(src_type->dimx == dst_type->dimx);
|
|
|
|
|
|
|
|
|
|
switch (dst_type->e.numeric.type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_TYPE_HALF:
|
|
|
|
|
case HLSL_TYPE_FLOAT:
|
|
|
|
|
switch (src_type->e.numeric.type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_TYPE_INT:
|
|
|
|
|
case HLSL_TYPE_UINT:
|
|
|
|
|
case HLSL_TYPE_BOOL:
|
|
|
|
|
/* Integrals are internally represented as floats, so no change is necessary.*/
|
|
|
|
|
case HLSL_TYPE_HALF:
|
|
|
|
|
case HLSL_TYPE_FLOAT:
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MOV, 0, 0, true);
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_DOUBLE:
|
2024-08-29 07:12:31 -07:00
|
|
|
if (ctx->double_as_float_alias)
|
|
|
|
|
{
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MOV, 0, 0, true);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE,
|
|
|
|
|
"The 'double' type is not supported for the %s profile.", ctx->profile->name);
|
2024-05-31 19:33:47 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_INT:
|
|
|
|
|
case HLSL_TYPE_UINT:
|
|
|
|
|
switch(src_type->e.numeric.type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_TYPE_HALF:
|
|
|
|
|
case HLSL_TYPE_FLOAT:
|
|
|
|
|
/* A compilation pass turns these into FLOOR+REINTERPRET, so we should not
|
|
|
|
|
* reach this case unless we are missing something. */
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Unlowered SM1 cast from float to integer.");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_INT:
|
|
|
|
|
case HLSL_TYPE_UINT:
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MOV, 0, 0, true);
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_BOOL:
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "SM1 cast from bool to integer.");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_DOUBLE:
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "SM1 cast from double to integer.");
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_DOUBLE:
|
2024-08-29 07:12:31 -07:00
|
|
|
switch (src_type->e.numeric.type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_TYPE_FLOAT:
|
|
|
|
|
if (ctx->double_as_float_alias)
|
|
|
|
|
{
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MOV, 0, 0, true);
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
hlsl_error(ctx, &instr->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE,
|
|
|
|
|
"The 'double' type is not supported for the %s profile.", ctx->profile->name);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "SM1 cast to double.");
|
|
|
|
|
break;
|
|
|
|
|
}
|
2024-05-31 19:33:47 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_TYPE_BOOL:
|
|
|
|
|
/* Casts to bool should have already been lowered. */
|
|
|
|
|
default:
|
|
|
|
|
hlsl_fixme(ctx, &expr->node.loc, "SM1 cast from %s to %s.",
|
|
|
|
|
debug_hlsl_type(ctx, src_type), debug_hlsl_type(ctx, dst_type));
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
static bool sm1_generate_vsir_instr_expr(struct hlsl_ctx *ctx, struct vsir_program *program,
|
|
|
|
|
struct hlsl_ir_expr *expr)
|
|
|
|
|
{
|
2024-05-31 19:33:47 -07:00
|
|
|
struct hlsl_ir_node *instr = &expr->node;
|
|
|
|
|
|
|
|
|
|
if (expr->op != HLSL_OP1_REINTERPRET && expr->op != HLSL_OP1_CAST
|
|
|
|
|
&& instr->data_type->e.numeric.type != HLSL_TYPE_FLOAT)
|
|
|
|
|
{
|
|
|
|
|
/* These need to be lowered. */
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "SM1 non-float expression.");
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
switch (expr->op)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_OP1_ABS:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_ABS, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
2024-05-31 19:33:47 -07:00
|
|
|
case HLSL_OP1_CAST:
|
|
|
|
|
return sm1_generate_vsir_instr_expr_cast(ctx, program, expr);
|
|
|
|
|
|
2024-07-22 18:15:25 -07:00
|
|
|
case HLSL_OP1_COS_REDUCED:
|
|
|
|
|
VKD3D_ASSERT(expr->node.reg.writemask == VKD3DSP_WRITEMASK_0);
|
|
|
|
|
sm1_generate_vsir_instr_expr_sincos(ctx, program, expr);
|
|
|
|
|
break;
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
case HLSL_OP1_DSX:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_DSX, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP1_DSY:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_DSY, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
2024-05-30 23:20:36 -07:00
|
|
|
case HLSL_OP1_EXP2:
|
|
|
|
|
sm1_generate_vsir_instr_expr_per_component_instr_op(ctx, program, expr, VKD3DSIH_EXP);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP1_LOG2:
|
|
|
|
|
sm1_generate_vsir_instr_expr_per_component_instr_op(ctx, program, expr, VKD3DSIH_LOG);
|
|
|
|
|
break;
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
case HLSL_OP1_NEG:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MOV, VKD3DSPSM_NEG, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
2024-05-30 23:20:36 -07:00
|
|
|
case HLSL_OP1_RCP:
|
|
|
|
|
sm1_generate_vsir_instr_expr_per_component_instr_op(ctx, program, expr, VKD3DSIH_RCP);
|
|
|
|
|
break;
|
|
|
|
|
|
2024-05-31 17:21:57 -07:00
|
|
|
case HLSL_OP1_REINTERPRET:
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MOV, 0, 0, true);
|
|
|
|
|
break;
|
|
|
|
|
|
2024-05-30 23:20:36 -07:00
|
|
|
case HLSL_OP1_RSQ:
|
|
|
|
|
sm1_generate_vsir_instr_expr_per_component_instr_op(ctx, program, expr, VKD3DSIH_RSQ);
|
|
|
|
|
break;
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
case HLSL_OP1_SAT:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MOV, 0, VKD3DSPDM_SATURATE, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
2024-07-22 18:15:25 -07:00
|
|
|
case HLSL_OP1_SIN_REDUCED:
|
|
|
|
|
VKD3D_ASSERT(expr->node.reg.writemask == VKD3DSP_WRITEMASK_1);
|
|
|
|
|
sm1_generate_vsir_instr_expr_sincos(ctx, program, expr);
|
|
|
|
|
break;
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
case HLSL_OP2_ADD:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_ADD, 0, 0, true);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_DOT:
|
|
|
|
|
switch (expr->operands[0].node->data_type->dimx)
|
|
|
|
|
{
|
|
|
|
|
case 3:
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_DP3, 0, 0, false);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 4:
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_DP4, 0, 0, false);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
return false;
|
|
|
|
|
}
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_MAX:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MAX, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_MIN:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MIN, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_MUL:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MUL, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP1_FRACT:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_FRC, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_LOGIC_AND:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MIN, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_LOGIC_OR:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MAX, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP2_SLT:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_SLT, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_OP3_CMP:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_CMP, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
2024-05-31 16:13:25 -07:00
|
|
|
case HLSL_OP3_DP2ADD:
|
|
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_DP2ADD, 0, 0, false);
|
|
|
|
|
break;
|
|
|
|
|
|
2024-05-30 15:44:18 -07:00
|
|
|
case HLSL_OP3_MAD:
|
2024-05-31 13:06:11 -07:00
|
|
|
sm1_generate_vsir_instr_expr_single_instr_op(ctx, program, expr, VKD3DSIH_MAD, 0, 0, true);
|
2024-05-30 15:44:18 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
2024-06-04 17:25:58 -07:00
|
|
|
hlsl_fixme(ctx, &instr->loc, "SM1 \"%s\" expression.", debug_hlsl_expr_op(expr->op));
|
2024-05-30 15:44:18 -07:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-29 18:40:26 -07:00
|
|
|
static void sm1_generate_vsir_init_dst_param_from_deref(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param, struct hlsl_deref *deref,
|
|
|
|
|
const struct vkd3d_shader_location *loc, unsigned int writemask)
|
|
|
|
|
{
|
|
|
|
|
enum vkd3d_shader_register_type type = VKD3DSPR_TEMP;
|
|
|
|
|
struct vkd3d_shader_version version;
|
|
|
|
|
uint32_t register_index;
|
|
|
|
|
struct hlsl_reg reg;
|
|
|
|
|
|
|
|
|
|
reg = hlsl_reg_from_deref(ctx, deref);
|
|
|
|
|
register_index = reg.id;
|
|
|
|
|
writemask = hlsl_combine_writemasks(reg.writemask, writemask);
|
|
|
|
|
|
|
|
|
|
if (deref->var->is_output_semantic)
|
|
|
|
|
{
|
|
|
|
|
version.major = ctx->profile->major_version;
|
|
|
|
|
version.minor = ctx->profile->minor_version;
|
|
|
|
|
version.type = ctx->profile->type;
|
|
|
|
|
|
|
|
|
|
if (version.type == VKD3D_SHADER_TYPE_PIXEL && version.major == 1)
|
|
|
|
|
{
|
|
|
|
|
type = VKD3DSPR_TEMP;
|
|
|
|
|
register_index = 0;
|
|
|
|
|
}
|
|
|
|
|
else if (!hlsl_sm1_register_from_semantic(&version, deref->var->semantic.name,
|
|
|
|
|
deref->var->semantic.index, true, &type, ®ister_index))
|
|
|
|
|
{
|
|
|
|
|
VKD3D_ASSERT(reg.allocated);
|
|
|
|
|
type = VKD3DSPR_OUTPUT;
|
|
|
|
|
register_index = reg.id;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
writemask = (1u << deref->var->data_type->dimx) - 1;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
VKD3D_ASSERT(reg.allocated);
|
|
|
|
|
|
|
|
|
|
vsir_register_init(&dst_param->reg, type, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->write_mask = writemask;
|
|
|
|
|
dst_param->reg.idx[0].offset = register_index;
|
|
|
|
|
|
|
|
|
|
if (deref->rel_offset.node)
|
|
|
|
|
hlsl_fixme(ctx, loc, "Translate relative addressing on dst register for vsir.");
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-28 19:11:29 -07:00
|
|
|
static void sm1_generate_vsir_init_src_param_from_deref(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vkd3d_shader_src_param *src_param, struct hlsl_deref *deref,
|
|
|
|
|
unsigned int dst_writemask, const struct vkd3d_shader_location *loc)
|
|
|
|
|
{
|
|
|
|
|
enum vkd3d_shader_register_type type = VKD3DSPR_TEMP;
|
|
|
|
|
struct vkd3d_shader_version version;
|
|
|
|
|
uint32_t register_index;
|
|
|
|
|
unsigned int writemask;
|
|
|
|
|
struct hlsl_reg reg;
|
|
|
|
|
|
2024-06-03 19:23:53 -07:00
|
|
|
if (hlsl_type_is_resource(deref->var->data_type))
|
|
|
|
|
{
|
|
|
|
|
unsigned int sampler_offset;
|
2024-05-28 19:11:29 -07:00
|
|
|
|
2024-06-03 19:23:53 -07:00
|
|
|
type = VKD3DSPR_COMBINED_SAMPLER;
|
|
|
|
|
|
|
|
|
|
sampler_offset = hlsl_offset_from_deref_safe(ctx, deref);
|
|
|
|
|
register_index = deref->var->regs[HLSL_REGSET_SAMPLERS].index + sampler_offset;
|
|
|
|
|
writemask = VKD3DSP_WRITEMASK_ALL;
|
|
|
|
|
}
|
|
|
|
|
else if (deref->var->is_uniform)
|
2024-05-28 19:11:29 -07:00
|
|
|
{
|
|
|
|
|
type = VKD3DSPR_CONST;
|
2024-06-03 19:23:53 -07:00
|
|
|
|
|
|
|
|
reg = hlsl_reg_from_deref(ctx, deref);
|
|
|
|
|
register_index = reg.id;
|
|
|
|
|
writemask = reg.writemask;
|
|
|
|
|
VKD3D_ASSERT(reg.allocated);
|
2024-05-28 19:11:29 -07:00
|
|
|
}
|
|
|
|
|
else if (deref->var->is_input_semantic)
|
|
|
|
|
{
|
|
|
|
|
version.major = ctx->profile->major_version;
|
|
|
|
|
version.minor = ctx->profile->minor_version;
|
|
|
|
|
version.type = ctx->profile->type;
|
2024-06-03 19:23:53 -07:00
|
|
|
if (hlsl_sm1_register_from_semantic(&version, deref->var->semantic.name,
|
2024-05-28 19:11:29 -07:00
|
|
|
deref->var->semantic.index, false, &type, ®ister_index))
|
|
|
|
|
{
|
2024-06-03 19:23:53 -07:00
|
|
|
writemask = (1 << deref->var->data_type->dimx) - 1;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2024-05-28 19:11:29 -07:00
|
|
|
type = VKD3DSPR_INPUT;
|
2024-06-03 19:23:53 -07:00
|
|
|
|
|
|
|
|
reg = hlsl_reg_from_deref(ctx, deref);
|
2024-05-28 19:11:29 -07:00
|
|
|
register_index = reg.id;
|
2024-06-03 19:23:53 -07:00
|
|
|
writemask = reg.writemask;
|
|
|
|
|
VKD3D_ASSERT(reg.allocated);
|
2024-05-28 19:11:29 -07:00
|
|
|
}
|
2024-06-03 19:23:53 -07:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
type = VKD3DSPR_TEMP;
|
|
|
|
|
|
|
|
|
|
reg = hlsl_reg_from_deref(ctx, deref);
|
|
|
|
|
register_index = reg.id;
|
|
|
|
|
writemask = reg.writemask;
|
2024-05-28 19:11:29 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
vsir_register_init(&src_param->reg, type, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = register_index;
|
|
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(writemask, dst_writemask);
|
|
|
|
|
|
|
|
|
|
if (deref->rel_offset.node)
|
|
|
|
|
hlsl_fixme(ctx, loc, "Translate relative addressing on src register for vsir.");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void sm1_generate_vsir_instr_load(struct hlsl_ctx *ctx, struct vsir_program *program,
|
|
|
|
|
struct hlsl_ir_load *load)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr = &load->node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
|
|
|
|
|
VKD3D_ASSERT(instr->reg.allocated);
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_MOV, 1, 1)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = instr->reg.id;
|
|
|
|
|
dst_param->write_mask = instr->reg.writemask;
|
|
|
|
|
|
|
|
|
|
sm1_generate_vsir_init_src_param_from_deref(ctx, &ins->src[0], &load->src, dst_param->write_mask,
|
|
|
|
|
&ins->location);
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-03 19:23:53 -07:00
|
|
|
static void sm1_generate_vsir_instr_resource_load(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vsir_program *program, struct hlsl_ir_resource_load *load)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *coords = load->coords.node;
|
|
|
|
|
struct hlsl_ir_node *ddx = load->ddx.node;
|
|
|
|
|
struct hlsl_ir_node *ddy = load->ddy.node;
|
|
|
|
|
struct hlsl_ir_node *instr = &load->node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
enum vkd3d_shader_opcode opcode;
|
|
|
|
|
unsigned int src_count = 2;
|
|
|
|
|
uint32_t flags = 0;
|
|
|
|
|
|
|
|
|
|
VKD3D_ASSERT(instr->reg.allocated);
|
|
|
|
|
|
|
|
|
|
switch (load->load_type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_RESOURCE_SAMPLE:
|
|
|
|
|
opcode = VKD3DSIH_TEX;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_RESOURCE_SAMPLE_PROJ:
|
|
|
|
|
opcode = VKD3DSIH_TEX;
|
|
|
|
|
flags |= VKD3DSI_TEXLD_PROJECT;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_RESOURCE_SAMPLE_LOD_BIAS:
|
|
|
|
|
opcode = VKD3DSIH_TEX;
|
|
|
|
|
flags |= VKD3DSI_TEXLD_BIAS;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_RESOURCE_SAMPLE_GRAD:
|
|
|
|
|
opcode = VKD3DSIH_TEXLDD;
|
|
|
|
|
src_count += 2;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Resource load type %u.", load->load_type);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, opcode, 1, src_count)))
|
|
|
|
|
return;
|
|
|
|
|
ins->flags = flags;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = instr->reg.id;
|
|
|
|
|
dst_param->write_mask = instr->reg.writemask;
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = coords->reg.id;
|
|
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(coords->reg.writemask, VKD3DSP_WRITEMASK_ALL);
|
|
|
|
|
|
|
|
|
|
sm1_generate_vsir_init_src_param_from_deref(ctx, &ins->src[1], &load->resource,
|
|
|
|
|
VKD3DSP_WRITEMASK_ALL, &ins->location);
|
|
|
|
|
|
|
|
|
|
if (load->load_type == HLSL_RESOURCE_SAMPLE_GRAD)
|
|
|
|
|
{
|
|
|
|
|
src_param = &ins->src[2];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = ddx->reg.id;
|
|
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(ddx->reg.writemask, VKD3DSP_WRITEMASK_ALL);
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[3];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = ddy->reg.id;
|
|
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(ddy->reg.writemask, VKD3DSP_WRITEMASK_ALL);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-30 12:06:13 -07:00
|
|
|
static void sm1_generate_vsir_instr_swizzle(struct hlsl_ctx *ctx, struct vsir_program *program,
|
|
|
|
|
struct hlsl_ir_swizzle *swizzle_instr)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr = &swizzle_instr->node, *val = swizzle_instr->val.node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
uint32_t swizzle;
|
|
|
|
|
|
|
|
|
|
VKD3D_ASSERT(instr->reg.allocated);
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_MOV, 1, 1)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = instr->reg.id;
|
|
|
|
|
dst_param->write_mask = instr->reg.writemask;
|
|
|
|
|
|
|
|
|
|
swizzle = hlsl_swizzle_from_writemask(val->reg.writemask);
|
|
|
|
|
swizzle = hlsl_combine_swizzles(swizzle, swizzle_instr->swizzle, instr->data_type->dimx);
|
|
|
|
|
swizzle = hlsl_map_swizzle(swizzle, ins->dst[0].write_mask);
|
|
|
|
|
swizzle = vsir_swizzle_from_hlsl(swizzle);
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = val->reg.id;
|
|
|
|
|
src_param->swizzle = swizzle;
|
|
|
|
|
}
|
|
|
|
|
|
2024-05-29 18:40:26 -07:00
|
|
|
static void sm1_generate_vsir_instr_store(struct hlsl_ctx *ctx, struct vsir_program *program,
|
|
|
|
|
struct hlsl_ir_store *store)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *rhs = store->rhs.node;
|
|
|
|
|
struct hlsl_ir_node *instr = &store->node;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_MOV, 1, 1)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
sm1_generate_vsir_init_dst_param_from_deref(ctx, &ins->dst[0], &store->lhs, &ins->location, store->writemask);
|
|
|
|
|
|
|
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = rhs->reg.id;
|
|
|
|
|
src_param->swizzle = sm1_generate_vsir_get_src_swizzle(rhs->reg.writemask, ins->dst[0].write_mask);
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-04 09:30:13 -07:00
|
|
|
static void sm1_generate_vsir_instr_jump(struct hlsl_ctx *ctx,
|
|
|
|
|
struct vsir_program *program, struct hlsl_ir_jump *jump)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *condition = jump->condition.node;
|
|
|
|
|
struct hlsl_ir_node *instr = &jump->node;
|
|
|
|
|
struct vkd3d_shader_dst_param *dst_param;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
|
|
|
|
|
if (jump->type == HLSL_IR_JUMP_DISCARD_NEG)
|
|
|
|
|
{
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_TEXKILL, 1, 0)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
dst_param = &ins->dst[0];
|
|
|
|
|
vsir_register_init(&dst_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
dst_param->reg.idx[0].offset = condition->reg.id;
|
|
|
|
|
dst_param->write_mask = condition->reg.writemask;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Jump type %s.", hlsl_jump_type_to_string(jump->type));
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
static void sm1_generate_vsir_block(struct hlsl_ctx *ctx, struct hlsl_block *block, struct vsir_program *program);
|
|
|
|
|
|
|
|
|
|
static void sm1_generate_vsir_instr_if(struct hlsl_ctx *ctx, struct vsir_program *program, struct hlsl_ir_if *iff)
|
2024-05-27 16:58:05 -07:00
|
|
|
{
|
2024-06-04 13:31:26 -07:00
|
|
|
struct hlsl_ir_node *condition = iff->condition.node;
|
|
|
|
|
struct vkd3d_shader_src_param *src_param;
|
|
|
|
|
struct hlsl_ir_node *instr = &iff->node;
|
|
|
|
|
struct vkd3d_shader_instruction *ins;
|
|
|
|
|
uint32_t swizzle;
|
2024-05-27 16:58:05 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
if (hlsl_version_lt(ctx, 2, 1))
|
2024-05-27 16:58:05 -07:00
|
|
|
{
|
2024-06-04 13:31:26 -07:00
|
|
|
hlsl_fixme(ctx, &instr->loc, "Flatten \"if\" conditionals branches.");
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
VKD3D_ASSERT(condition->data_type->dimx == 1 && condition->data_type->dimy == 1);
|
2024-05-27 16:58:05 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_IFC, 0, 2)))
|
|
|
|
|
return;
|
|
|
|
|
ins->flags = VKD3D_SHADER_REL_OP_NE;
|
2024-06-04 09:30:13 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
swizzle = hlsl_swizzle_from_writemask(condition->reg.writemask);
|
|
|
|
|
swizzle = vsir_swizzle_from_hlsl(swizzle);
|
2024-05-28 19:11:29 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
src_param = &ins->src[0];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = condition->reg.id;
|
|
|
|
|
src_param->swizzle = swizzle;
|
|
|
|
|
src_param->modifiers = 0;
|
2024-06-03 19:23:53 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
src_param = &ins->src[1];
|
|
|
|
|
vsir_register_init(&src_param->reg, VKD3DSPR_TEMP, VKD3D_DATA_FLOAT, 1);
|
|
|
|
|
src_param->reg.idx[0].offset = condition->reg.id;
|
|
|
|
|
src_param->swizzle = swizzle;
|
|
|
|
|
src_param->modifiers = VKD3DSPSM_NEG;
|
2024-05-29 18:40:26 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
sm1_generate_vsir_block(ctx, &iff->then_block, program);
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_ELSE, 0, 0)))
|
|
|
|
|
return;
|
2024-05-27 16:58:05 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
sm1_generate_vsir_block(ctx, &iff->else_block, program);
|
|
|
|
|
|
|
|
|
|
if (!(ins = generate_vsir_add_program_instruction(ctx, program, &instr->loc, VKD3DSIH_ENDIF, 0, 0)))
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void sm1_generate_vsir_block(struct hlsl_ctx *ctx, struct hlsl_block *block, struct vsir_program *program)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr, *next;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY_SAFE(instr, next, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
2024-06-04 17:25:58 -07:00
|
|
|
if (instr->data_type)
|
|
|
|
|
{
|
|
|
|
|
if (instr->data_type->class != HLSL_CLASS_SCALAR && instr->data_type->class != HLSL_CLASS_VECTOR)
|
|
|
|
|
{
|
|
|
|
|
hlsl_fixme(ctx, &instr->loc, "Class %#x should have been lowered or removed.", instr->data_type->class);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
2024-06-04 17:25:58 -07:00
|
|
|
case HLSL_IR_CALL:
|
|
|
|
|
vkd3d_unreachable();
|
|
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
case HLSL_IR_CONSTANT:
|
|
|
|
|
sm1_generate_vsir_instr_constant(ctx, program, hlsl_ir_constant(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_EXPR:
|
|
|
|
|
sm1_generate_vsir_instr_expr(ctx, program, hlsl_ir_expr(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_IF:
|
|
|
|
|
sm1_generate_vsir_instr_if(ctx, program, hlsl_ir_if(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_JUMP:
|
|
|
|
|
sm1_generate_vsir_instr_jump(ctx, program, hlsl_ir_jump(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_LOAD:
|
|
|
|
|
sm1_generate_vsir_instr_load(ctx, program, hlsl_ir_load(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_RESOURCE_LOAD:
|
|
|
|
|
sm1_generate_vsir_instr_resource_load(ctx, program, hlsl_ir_resource_load(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_STORE:
|
|
|
|
|
sm1_generate_vsir_instr_store(ctx, program, hlsl_ir_store(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case HLSL_IR_SWIZZLE:
|
|
|
|
|
sm1_generate_vsir_instr_swizzle(ctx, program, hlsl_ir_swizzle(instr));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
2024-06-04 17:25:58 -07:00
|
|
|
hlsl_fixme(ctx, &instr->loc, "Instruction type %s.", hlsl_node_type_to_string(instr->type));
|
2024-06-04 13:31:26 -07:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
2024-05-27 16:58:05 -07:00
|
|
|
}
|
|
|
|
|
|
2024-05-17 15:30:59 -07:00
|
|
|
/* OBJECTIVE: Translate all the information from ctx and entry_func to the
|
2024-05-20 13:29:09 -07:00
|
|
|
* vsir_program and ctab blob, so they can be used as input to d3dbc_compile()
|
|
|
|
|
* without relying on ctx and entry_func. */
|
2024-05-17 15:30:59 -07:00
|
|
|
static void sm1_generate_vsir(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry_func,
|
2024-05-20 13:29:09 -07:00
|
|
|
uint64_t config_flags, struct vsir_program *program, struct vkd3d_shader_code *ctab)
|
2024-05-17 15:30:59 -07:00
|
|
|
{
|
|
|
|
|
struct vkd3d_shader_version version = {0};
|
2024-05-20 13:29:09 -07:00
|
|
|
struct vkd3d_bytecode_buffer buffer = {0};
|
2024-05-23 09:57:44 -07:00
|
|
|
struct hlsl_block block;
|
2024-05-17 15:30:59 -07:00
|
|
|
|
|
|
|
|
version.major = ctx->profile->major_version;
|
|
|
|
|
version.minor = ctx->profile->minor_version;
|
|
|
|
|
version.type = ctx->profile->type;
|
2024-09-10 13:22:24 -07:00
|
|
|
if (!vsir_program_init(program, NULL, &version, 0, VSIR_CF_STRUCTURED))
|
2024-05-17 15:30:59 -07:00
|
|
|
{
|
|
|
|
|
ctx->result = VKD3D_ERROR_OUT_OF_MEMORY;
|
|
|
|
|
return;
|
|
|
|
|
}
|
2024-05-20 13:29:09 -07:00
|
|
|
|
|
|
|
|
write_sm1_uniforms(ctx, &buffer);
|
|
|
|
|
if (buffer.status)
|
|
|
|
|
{
|
|
|
|
|
vkd3d_free(buffer.data);
|
|
|
|
|
ctx->result = buffer.status;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
ctab->code = buffer.data;
|
|
|
|
|
ctab->size = buffer.size;
|
2024-05-21 11:51:17 -07:00
|
|
|
|
|
|
|
|
sm1_generate_vsir_signature(ctx, program);
|
2024-05-23 09:57:44 -07:00
|
|
|
|
|
|
|
|
hlsl_block_init(&block);
|
|
|
|
|
sm1_generate_vsir_constant_defs(ctx, program, &block);
|
2024-05-23 18:22:04 -07:00
|
|
|
sm1_generate_vsir_sampler_dcls(ctx, program, &block);
|
2024-05-23 09:57:44 -07:00
|
|
|
list_move_head(&entry_func->body.instrs, &block.instrs);
|
2024-05-27 16:58:05 -07:00
|
|
|
|
2024-06-04 13:31:26 -07:00
|
|
|
sm1_generate_vsir_block(ctx, &entry_func->body, program);
|
2024-05-17 15:30:59 -07:00
|
|
|
}
|
|
|
|
|
|
2024-04-11 07:11:14 -07:00
|
|
|
static struct hlsl_ir_jump *loop_unrolling_find_jump(struct hlsl_block *block, struct hlsl_ir_node *stop_point,
|
|
|
|
|
struct hlsl_block **found_block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *node;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(node, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
if (node == stop_point)
|
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
|
|
if (node->type == HLSL_IR_IF)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(node);
|
|
|
|
|
struct hlsl_ir_jump *jump = NULL;
|
|
|
|
|
|
|
|
|
|
if ((jump = loop_unrolling_find_jump(&iff->then_block, stop_point, found_block)))
|
|
|
|
|
return jump;
|
|
|
|
|
if ((jump = loop_unrolling_find_jump(&iff->else_block, stop_point, found_block)))
|
|
|
|
|
return jump;
|
|
|
|
|
}
|
|
|
|
|
else if (node->type == HLSL_IR_JUMP)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_jump *jump = hlsl_ir_jump(node);
|
|
|
|
|
|
|
|
|
|
if (jump->type == HLSL_IR_JUMP_BREAK || jump->type == HLSL_IR_JUMP_CONTINUE)
|
|
|
|
|
{
|
|
|
|
|
*found_block = block;
|
|
|
|
|
return jump;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static unsigned int loop_unrolling_get_max_iterations(struct hlsl_ctx *ctx, struct hlsl_ir_loop *loop)
|
|
|
|
|
{
|
|
|
|
|
/* Always use the explicit limit if it has been passed. */
|
|
|
|
|
if (loop->unroll_limit)
|
|
|
|
|
return loop->unroll_limit;
|
|
|
|
|
|
|
|
|
|
/* All SMs will default to 1024 if [unroll] has been specified without an explicit limit. */
|
|
|
|
|
if (loop->unroll_type == HLSL_IR_LOOP_FORCE_UNROLL)
|
|
|
|
|
return 1024;
|
|
|
|
|
|
|
|
|
|
/* SM4 limits implicit unrolling to 254 iterations. */
|
|
|
|
|
if (hlsl_version_ge(ctx, 4, 0))
|
|
|
|
|
return 254;
|
|
|
|
|
|
|
|
|
|
/* SM<3 implicitly unrolls up to 1024 iterations. */
|
|
|
|
|
return 1024;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool loop_unrolling_unroll_loop(struct hlsl_ctx *ctx, struct hlsl_block *block,
|
|
|
|
|
struct hlsl_block *loop_parent, struct hlsl_ir_loop *loop)
|
|
|
|
|
{
|
|
|
|
|
unsigned int max_iterations, i;
|
|
|
|
|
|
|
|
|
|
max_iterations = loop_unrolling_get_max_iterations(ctx, loop);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < max_iterations; ++i)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_block tmp_dst, *jump_block;
|
|
|
|
|
struct hlsl_ir_jump *jump = NULL;
|
|
|
|
|
|
|
|
|
|
if (!hlsl_clone_block(ctx, &tmp_dst, &loop->body))
|
|
|
|
|
return false;
|
|
|
|
|
list_move_before(&loop->node.entry, &tmp_dst.instrs);
|
|
|
|
|
hlsl_block_cleanup(&tmp_dst);
|
|
|
|
|
|
|
|
|
|
hlsl_run_const_passes(ctx, block);
|
|
|
|
|
|
|
|
|
|
if ((jump = loop_unrolling_find_jump(loop_parent, &loop->node, &jump_block)))
|
|
|
|
|
{
|
|
|
|
|
enum hlsl_ir_jump_type type = jump->type;
|
|
|
|
|
|
|
|
|
|
if (jump_block != loop_parent)
|
|
|
|
|
{
|
|
|
|
|
if (loop->unroll_type == HLSL_IR_LOOP_FORCE_UNROLL)
|
|
|
|
|
hlsl_error(ctx, &jump->node.loc, VKD3D_SHADER_ERROR_HLSL_FAILED_FORCED_UNROLL,
|
|
|
|
|
"Unable to unroll loop, unrolling loops with conditional jumps is currently not supported.");
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
list_move_slice_tail(&tmp_dst.instrs, &jump->node.entry, list_prev(&loop_parent->instrs, &loop->node.entry));
|
|
|
|
|
hlsl_block_cleanup(&tmp_dst);
|
|
|
|
|
|
|
|
|
|
if (type == HLSL_IR_JUMP_BREAK)
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Native will not emit an error if max_iterations has been reached with an
|
|
|
|
|
* explicit limit. It also will not insert a loop if there are iterations left
|
|
|
|
|
* i.e [unroll(4)] for (i = 0; i < 8; ++i)) */
|
|
|
|
|
if (!loop->unroll_limit && i == max_iterations)
|
|
|
|
|
{
|
|
|
|
|
if (loop->unroll_type == HLSL_IR_LOOP_FORCE_UNROLL)
|
|
|
|
|
hlsl_error(ctx, &loop->node.loc, VKD3D_SHADER_ERROR_HLSL_FAILED_FORCED_UNROLL,
|
|
|
|
|
"Unable to unroll loop, maximum iterations reached (%u).", max_iterations);
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
list_remove(&loop->node.entry);
|
|
|
|
|
hlsl_free_instr(&loop->node);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* loop_unrolling_find_unrollable_loop() is not the normal way to do things;
|
|
|
|
|
* normal passes simply iterate over the whole block and apply a transformation
|
|
|
|
|
* to every relevant instruction. However, loop unrolling can fail, and we want
|
|
|
|
|
* to leave the loop in its previous state in that case. That isn't a problem by
|
|
|
|
|
* itself, except that loop unrolling needs copy-prop in order to work properly,
|
|
|
|
|
* and copy-prop state at the time of the loop depends on the rest of the program
|
|
|
|
|
* up to that point. This means we need to clone the whole program, and at that
|
|
|
|
|
* point we have to search it again anyway to find the clone of the loop we were
|
|
|
|
|
* going to unroll.
|
|
|
|
|
*
|
|
|
|
|
* FIXME: Ideally we wouldn't clone the whole program; instead we would run copyprop
|
|
|
|
|
* up until the loop instruction, clone just that loop, then use copyprop again
|
|
|
|
|
* with the saved state after unrolling. However, copyprop currently isn't built
|
|
|
|
|
* for that yet [notably, it still relies on indices]. Note also this still doesn't
|
|
|
|
|
* really let us use transform_ir() anyway [since we don't have a good way to say
|
|
|
|
|
* "copyprop from the beginning of the program up to the instruction we're
|
|
|
|
|
* currently processing" from the callback]; we'd have to use a dedicated
|
|
|
|
|
* recursive function instead. */
|
|
|
|
|
static struct hlsl_ir_loop *loop_unrolling_find_unrollable_loop(struct hlsl_ctx *ctx, struct hlsl_block *block,
|
|
|
|
|
struct hlsl_block **containing_block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *instr;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(instr, &block->instrs, struct hlsl_ir_node, entry)
|
|
|
|
|
{
|
|
|
|
|
switch (instr->type)
|
|
|
|
|
{
|
|
|
|
|
case HLSL_IR_LOOP:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_loop *nested_loop;
|
|
|
|
|
struct hlsl_ir_loop *loop = hlsl_ir_loop(instr);
|
|
|
|
|
|
|
|
|
|
if ((nested_loop = loop_unrolling_find_unrollable_loop(ctx, &loop->body, containing_block)))
|
|
|
|
|
return nested_loop;
|
|
|
|
|
|
|
|
|
|
if (loop->unroll_type == HLSL_IR_LOOP_UNROLL || loop->unroll_type == HLSL_IR_LOOP_FORCE_UNROLL)
|
|
|
|
|
{
|
|
|
|
|
*containing_block = block;
|
|
|
|
|
return loop;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case HLSL_IR_IF:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_loop *loop;
|
|
|
|
|
struct hlsl_ir_if *iff = hlsl_ir_if(instr);
|
|
|
|
|
|
|
|
|
|
if ((loop = loop_unrolling_find_unrollable_loop(ctx, &iff->then_block, containing_block)))
|
|
|
|
|
return loop;
|
|
|
|
|
if ((loop = loop_unrolling_find_unrollable_loop(ctx, &iff->else_block, containing_block)))
|
|
|
|
|
return loop;
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case HLSL_IR_SWITCH:
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_switch *s = hlsl_ir_switch(instr);
|
|
|
|
|
struct hlsl_ir_switch_case *c;
|
|
|
|
|
struct hlsl_ir_loop *loop;
|
|
|
|
|
|
|
|
|
|
LIST_FOR_EACH_ENTRY(c, &s->cases, struct hlsl_ir_switch_case, entry)
|
|
|
|
|
{
|
|
|
|
|
if ((loop = loop_unrolling_find_unrollable_loop(ctx, &c->body, containing_block)))
|
|
|
|
|
return loop;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void transform_unroll_loops(struct hlsl_ctx *ctx, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
while (true)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_block clone, *containing_block;
|
|
|
|
|
struct hlsl_ir_loop *loop, *cloned_loop;
|
|
|
|
|
|
|
|
|
|
if (!(loop = loop_unrolling_find_unrollable_loop(ctx, block, &containing_block)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if (!hlsl_clone_block(ctx, &clone, block))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
cloned_loop = loop_unrolling_find_unrollable_loop(ctx, &clone, &containing_block);
|
2024-08-01 01:48:48 -07:00
|
|
|
VKD3D_ASSERT(cloned_loop);
|
2024-04-11 07:11:14 -07:00
|
|
|
|
|
|
|
|
if (!loop_unrolling_unroll_loop(ctx, &clone, containing_block, cloned_loop))
|
|
|
|
|
{
|
|
|
|
|
hlsl_block_cleanup(&clone);
|
|
|
|
|
loop->unroll_type = HLSL_IR_LOOP_FORCE_LOOP;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hlsl_block_cleanup(block);
|
|
|
|
|
hlsl_block_init(block);
|
|
|
|
|
hlsl_block_add_block(block, &clone);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2024-07-06 08:24:22 -07:00
|
|
|
static bool lower_f16tof32(struct hlsl_ctx *ctx, struct hlsl_ir_node *node, struct hlsl_block *block)
|
|
|
|
|
{
|
|
|
|
|
struct hlsl_ir_node *call, *rhs, *store;
|
|
|
|
|
struct hlsl_ir_function_decl *func;
|
|
|
|
|
unsigned int component_count;
|
|
|
|
|
struct hlsl_ir_load *load;
|
|
|
|
|
struct hlsl_ir_expr *expr;
|
|
|
|
|
struct hlsl_ir_var *lhs;
|
|
|
|
|
char *body;
|
|
|
|
|
|
|
|
|
|
static const char template[] =
|
|
|
|
|
"typedef uint%u uintX;\n"
|
|
|
|
|
"float%u soft_f16tof32(uintX x)\n"
|
|
|
|
|
"{\n"
|
|
|
|
|
" uintX mantissa = x & 0x3ff;\n"
|
|
|
|
|
" uintX high2 = mantissa >> 8;\n"
|
|
|
|
|
" uintX high2_check = high2 ? high2 : mantissa;\n"
|
|
|
|
|
" uintX high6 = high2_check >> 4;\n"
|
|
|
|
|
" uintX high6_check = high6 ? high6 : high2_check;\n"
|
|
|
|
|
"\n"
|
|
|
|
|
" uintX high8 = high6_check >> 2;\n"
|
|
|
|
|
" uintX high8_check = (high8 ? high8 : high6_check) >> 1;\n"
|
|
|
|
|
" uintX shift = high6 ? (high2 ? 12 : 4) : (high2 ? 8 : 0);\n"
|
|
|
|
|
" shift = high8 ? shift + 2 : shift;\n"
|
|
|
|
|
" shift = high8_check ? shift + 1 : shift;\n"
|
|
|
|
|
" shift = -shift + 10;\n"
|
|
|
|
|
" shift = mantissa ? shift : 11;\n"
|
|
|
|
|
" uintX subnormal_mantissa = ((mantissa << shift) << 23) & 0x7fe000;\n"
|
|
|
|
|
" uintX subnormal_exp = -(shift << 23) + 0x38800000;\n"
|
|
|
|
|
" uintX subnormal_val = subnormal_exp + subnormal_mantissa;\n"
|
|
|
|
|
" uintX subnormal_or_zero = mantissa ? subnormal_val : 0;\n"
|
|
|
|
|
"\n"
|
|
|
|
|
" uintX exponent = (((x >> 10) << 23) & 0xf800000) + 0x38000000;\n"
|
|
|
|
|
"\n"
|
|
|
|
|
" uintX low_3 = (x << 13) & 0x7fe000;\n"
|
|
|
|
|
" uintX normalized_val = exponent + low_3;\n"
|
|
|
|
|
" uintX inf_nan_val = low_3 + 0x7f800000;\n"
|
|
|
|
|
"\n"
|
|
|
|
|
" uintX exp_mask = 0x7c00;\n"
|
|
|
|
|
" uintX is_inf_nan = (x & exp_mask) == exp_mask;\n"
|
|
|
|
|
" uintX is_normalized = x & exp_mask;\n"
|
|
|
|
|
"\n"
|
|
|
|
|
" uintX check = is_inf_nan ? inf_nan_val : normalized_val;\n"
|
|
|
|
|
" uintX exp_mantissa = (is_normalized ? check : subnormal_or_zero) & 0x7fffe000;\n"
|
|
|
|
|
" uintX sign_bit = (x << 16) & 0x80000000;\n"
|
|
|
|
|
"\n"
|
|
|
|
|
" return asfloat(exp_mantissa + sign_bit);\n"
|
|
|
|
|
"}\n";
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if (node->type != HLSL_IR_EXPR)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
expr = hlsl_ir_expr(node);
|
|
|
|
|
|
|
|
|
|
if (expr->op != HLSL_OP1_F16TOF32)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
rhs = expr->operands[0].node;
|
|
|
|
|
component_count = hlsl_type_component_count(rhs->data_type);
|
|
|
|
|
|
|
|
|
|
if (!(body = hlsl_sprintf_alloc(ctx, template, component_count, component_count)))
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (!(func = hlsl_compile_internal_function(ctx, "soft_f16tof32", body)))
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
lhs = func->parameters.vars[0];
|
|
|
|
|
|
|
|
|
|
if (!(store = hlsl_new_simple_store(ctx, lhs, rhs)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, store);
|
|
|
|
|
|
|
|
|
|
if (!(call = hlsl_new_call(ctx, func, &node->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, call);
|
|
|
|
|
|
|
|
|
|
if (!(load = hlsl_new_var_load(ctx, func->return_var, &node->loc)))
|
|
|
|
|
return false;
|
|
|
|
|
hlsl_block_add_instr(block, &load->node);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2024-08-24 07:25:52 -07:00
|
|
|
static void process_entry_function(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry_func)
|
2021-03-02 13:34:46 -08:00
|
|
|
{
|
2021-08-16 15:29:34 -07:00
|
|
|
const struct hlsl_profile_info *profile = ctx->profile;
|
2021-10-15 14:54:10 -07:00
|
|
|
struct hlsl_block *const body = &entry_func->body;
|
2021-09-11 09:20:32 -07:00
|
|
|
struct recursive_call_ctx recursive_call_ctx;
|
2021-03-28 12:46:55 -07:00
|
|
|
struct hlsl_ir_var *var;
|
2021-08-16 12:52:10 -07:00
|
|
|
unsigned int i;
|
2021-03-28 12:46:55 -07:00
|
|
|
|
2023-06-09 06:28:06 -07:00
|
|
|
list_move_head(&body->instrs, &ctx->static_initializers.instrs);
|
2021-03-02 13:34:46 -08:00
|
|
|
|
2021-09-11 09:20:32 -07:00
|
|
|
memset(&recursive_call_ctx, 0, sizeof(recursive_call_ctx));
|
2023-04-25 06:04:29 -07:00
|
|
|
hlsl_transform_ir(ctx, find_recursive_calls, body, &recursive_call_ctx);
|
2021-09-11 09:20:32 -07:00
|
|
|
vkd3d_free(recursive_call_ctx.backtrace);
|
|
|
|
|
|
2021-09-13 21:08:34 -07:00
|
|
|
/* Avoid going into an infinite loop when processing call instructions.
|
|
|
|
|
* lower_return() recurses into inferior calls. */
|
|
|
|
|
if (ctx->result)
|
2024-08-24 07:25:52 -07:00
|
|
|
return;
|
2021-09-13 21:08:34 -07:00
|
|
|
|
2024-07-06 08:24:22 -07:00
|
|
|
if (hlsl_version_ge(ctx, 4, 0) && hlsl_version_lt(ctx, 5, 0))
|
|
|
|
|
lower_ir(ctx, lower_f16tof32, body);
|
|
|
|
|
|
2021-09-13 21:08:34 -07:00
|
|
|
lower_return(ctx, entry_func, body, false);
|
|
|
|
|
|
2023-04-25 06:04:29 -07:00
|
|
|
while (hlsl_transform_ir(ctx, lower_calls, body, NULL));
|
2021-09-11 14:56:04 -07:00
|
|
|
|
2023-09-06 17:01:49 -07:00
|
|
|
lower_ir(ctx, lower_matrix_swizzles, body);
|
2023-06-25 16:46:10 -07:00
|
|
|
lower_ir(ctx, lower_index_loads, body);
|
2023-03-10 17:09:58 -08:00
|
|
|
|
2024-04-23 12:38:06 -07:00
|
|
|
LIST_FOR_EACH_ENTRY(var, &ctx->globals->vars, struct hlsl_ir_var, scope_entry)
|
|
|
|
|
{
|
|
|
|
|
if (var->storage_modifiers & HLSL_STORAGE_UNIFORM)
|
|
|
|
|
prepend_uniform_copy(ctx, body, var);
|
|
|
|
|
}
|
2021-03-28 12:46:55 -07:00
|
|
|
|
2023-01-31 17:27:01 -08:00
|
|
|
for (i = 0; i < entry_func->parameters.count; ++i)
|
2021-03-28 12:46:55 -07:00
|
|
|
{
|
2023-01-31 17:27:01 -08:00
|
|
|
var = entry_func->parameters.vars[i];
|
|
|
|
|
|
2023-04-05 09:07:37 -07:00
|
|
|
if (hlsl_type_is_resource(var->data_type) || (var->storage_modifiers & HLSL_STORAGE_UNIFORM))
|
2021-04-27 10:14:18 -07:00
|
|
|
{
|
2023-03-10 13:13:23 -08:00
|
|
|
prepend_uniform_copy(ctx, body, var);
|
2021-04-27 10:14:18 -07:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2023-04-05 09:07:37 -07:00
|
|
|
if (hlsl_get_multiarray_element_type(var->data_type)->class != HLSL_CLASS_STRUCT
|
|
|
|
|
&& !var->semantic.name)
|
|
|
|
|
{
|
2021-12-01 08:14:57 -08:00
|
|
|
hlsl_error(ctx, &var->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_SEMANTIC,
|
2021-10-15 14:54:08 -07:00
|
|
|
"Parameter \"%s\" is missing a semantic.", var->name);
|
2023-04-05 09:07:37 -07:00
|
|
|
var->semantic.reported_missing = true;
|
|
|
|
|
}
|
2021-10-15 14:54:08 -07:00
|
|
|
|
2022-11-29 11:10:40 -08:00
|
|
|
if (var->storage_modifiers & HLSL_STORAGE_IN)
|
2023-03-10 13:14:30 -08:00
|
|
|
prepend_input_var_copy(ctx, body, var);
|
2022-11-29 11:10:40 -08:00
|
|
|
if (var->storage_modifiers & HLSL_STORAGE_OUT)
|
2023-03-10 13:17:26 -08:00
|
|
|
append_output_var_copy(ctx, body, var);
|
2021-04-27 10:14:18 -07:00
|
|
|
}
|
2021-03-28 12:46:55 -07:00
|
|
|
}
|
2021-03-28 12:46:59 -07:00
|
|
|
if (entry_func->return_var)
|
2021-04-27 10:14:18 -07:00
|
|
|
{
|
2022-11-11 17:31:55 -08:00
|
|
|
if (entry_func->return_var->data_type->class != HLSL_CLASS_STRUCT && !entry_func->return_var->semantic.name)
|
2021-12-01 08:14:57 -08:00
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_SEMANTIC,
|
2021-04-27 10:14:18 -07:00
|
|
|
"Entry point \"%s\" is missing a return value semantic.", entry_func->func->name);
|
|
|
|
|
|
2023-03-10 13:17:26 -08:00
|
|
|
append_output_var_copy(ctx, body, entry_func->return_var);
|
2021-04-27 10:14:18 -07:00
|
|
|
}
|
2021-03-28 12:46:55 -07:00
|
|
|
|
2023-06-08 03:47:40 -07:00
|
|
|
if (profile->major_version >= 4)
|
|
|
|
|
{
|
|
|
|
|
hlsl_transform_ir(ctx, lower_discard_neg, body, NULL);
|
|
|
|
|
}
|
2023-04-25 06:04:29 -07:00
|
|
|
|
2024-04-11 07:11:14 -07:00
|
|
|
transform_unroll_loops(ctx, body);
|
2024-04-26 09:15:52 -07:00
|
|
|
hlsl_run_const_passes(ctx, body);
|
|
|
|
|
|
2023-10-11 05:29:25 -07:00
|
|
|
remove_unreachable_code(ctx, body);
|
2023-10-11 05:39:01 -07:00
|
|
|
hlsl_transform_ir(ctx, normalize_switch_cases, body, NULL);
|
2021-03-17 22:22:22 -07:00
|
|
|
|
2023-06-25 16:46:10 -07:00
|
|
|
lower_ir(ctx, lower_nonconstant_vector_derefs, body);
|
2023-06-25 17:03:26 -07:00
|
|
|
lower_ir(ctx, lower_casts_to_bool, body);
|
2023-06-25 16:46:10 -07:00
|
|
|
lower_ir(ctx, lower_int_dot, body);
|
2023-05-08 15:25:18 -07:00
|
|
|
|
2024-03-12 05:40:30 -07:00
|
|
|
hlsl_transform_ir(ctx, validate_dereferences, body, NULL);
|
2023-05-29 18:59:17 -07:00
|
|
|
hlsl_transform_ir(ctx, track_object_components_sampler_dim, body, NULL);
|
|
|
|
|
if (profile->major_version >= 4)
|
|
|
|
|
hlsl_transform_ir(ctx, lower_combined_samples, body, NULL);
|
2024-05-06 20:37:31 -07:00
|
|
|
|
|
|
|
|
do
|
|
|
|
|
compute_liveness(ctx, entry_func);
|
|
|
|
|
while (hlsl_transform_ir(ctx, dce, body, NULL));
|
|
|
|
|
|
2024-05-06 23:45:14 -07:00
|
|
|
hlsl_transform_ir(ctx, track_components_usage, body, NULL);
|
2023-08-04 12:02:39 -07:00
|
|
|
sort_synthetic_separated_samplers_first(ctx);
|
2023-05-29 18:59:17 -07:00
|
|
|
|
2021-08-16 15:29:34 -07:00
|
|
|
if (profile->major_version < 4)
|
2023-01-24 04:44:39 -08:00
|
|
|
{
|
2024-07-08 12:13:07 -07:00
|
|
|
while (lower_ir(ctx, lower_nonconstant_array_loads, body));
|
|
|
|
|
|
2024-04-05 12:34:32 -07:00
|
|
|
lower_ir(ctx, lower_ternary, body);
|
|
|
|
|
|
2024-04-02 15:47:51 -07:00
|
|
|
lower_ir(ctx, lower_nonfloat_exprs, body);
|
|
|
|
|
/* Constants casted to float must be folded, and new casts to bool also need to be lowered. */
|
|
|
|
|
hlsl_transform_ir(ctx, hlsl_fold_constant_exprs, body, NULL);
|
|
|
|
|
lower_ir(ctx, lower_casts_to_bool, body);
|
|
|
|
|
|
2023-10-20 19:23:46 -07:00
|
|
|
lower_ir(ctx, lower_casts_to_int, body);
|
2023-06-25 17:10:34 -07:00
|
|
|
lower_ir(ctx, lower_division, body);
|
2023-06-25 17:11:37 -07:00
|
|
|
lower_ir(ctx, lower_sqrt, body);
|
2023-06-25 16:46:10 -07:00
|
|
|
lower_ir(ctx, lower_dot, body);
|
|
|
|
|
lower_ir(ctx, lower_round, body);
|
2023-10-17 14:22:52 -07:00
|
|
|
lower_ir(ctx, lower_ceil, body);
|
2023-10-17 14:26:04 -07:00
|
|
|
lower_ir(ctx, lower_floor, body);
|
2024-07-10 19:47:51 -07:00
|
|
|
lower_ir(ctx, lower_trig, body);
|
2024-02-26 10:41:12 -08:00
|
|
|
lower_ir(ctx, lower_comparison_operators, body);
|
2024-02-20 17:03:43 -08:00
|
|
|
lower_ir(ctx, lower_logic_not, body);
|
2023-11-01 13:07:46 -07:00
|
|
|
if (ctx->profile->type == VKD3D_SHADER_TYPE_PIXEL)
|
|
|
|
|
lower_ir(ctx, lower_slt, body);
|
2024-02-26 09:01:45 -08:00
|
|
|
else
|
|
|
|
|
lower_ir(ctx, lower_cmp, body);
|
2023-01-24 04:44:39 -08:00
|
|
|
}
|
2021-05-20 22:32:24 -07:00
|
|
|
|
2023-02-04 15:30:36 -08:00
|
|
|
if (profile->major_version < 2)
|
|
|
|
|
{
|
2023-06-25 16:46:10 -07:00
|
|
|
lower_ir(ctx, lower_abs, body);
|
2023-02-04 15:30:36 -08:00
|
|
|
}
|
|
|
|
|
|
2024-01-11 11:54:20 -08:00
|
|
|
lower_ir(ctx, validate_nonconstant_vector_store_derefs, body);
|
|
|
|
|
|
2024-03-25 12:33:38 -07:00
|
|
|
do
|
|
|
|
|
compute_liveness(ctx, entry_func);
|
|
|
|
|
while (hlsl_transform_ir(ctx, dce, body, NULL));
|
|
|
|
|
|
2022-07-20 12:37:07 -07:00
|
|
|
/* TODO: move forward, remove when no longer needed */
|
2023-05-04 12:06:58 -07:00
|
|
|
transform_derefs(ctx, replace_deref_path_with_offset, body);
|
2023-04-25 06:04:29 -07:00
|
|
|
while (hlsl_transform_ir(ctx, hlsl_fold_constant_exprs, body, NULL));
|
2023-05-11 12:18:02 -07:00
|
|
|
transform_derefs(ctx, clean_constant_deref_offset_srcs, body);
|
2022-07-20 12:37:07 -07:00
|
|
|
|
2021-03-17 22:22:22 -07:00
|
|
|
do
|
|
|
|
|
compute_liveness(ctx, entry_func);
|
2023-04-25 06:04:29 -07:00
|
|
|
while (hlsl_transform_ir(ctx, dce, body, NULL));
|
2021-03-16 14:31:53 -07:00
|
|
|
|
2021-03-17 22:22:22 -07:00
|
|
|
compute_liveness(ctx, entry_func);
|
2021-03-02 13:34:46 -08:00
|
|
|
|
2023-10-04 13:28:02 -07:00
|
|
|
transform_derefs(ctx, mark_indexable_vars, body);
|
|
|
|
|
|
2022-11-25 14:47:56 -08:00
|
|
|
calculate_resource_register_counts(ctx);
|
2023-04-25 09:41:38 -07:00
|
|
|
|
2023-06-21 10:21:19 -07:00
|
|
|
allocate_register_reservations(ctx);
|
|
|
|
|
|
2021-05-20 22:32:20 -07:00
|
|
|
allocate_temp_registers(ctx, entry_func);
|
2024-08-24 07:25:52 -07:00
|
|
|
allocate_semantic_registers(ctx);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int hlsl_emit_bytecode(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry_func,
|
|
|
|
|
enum vkd3d_shader_target_type target_type, struct vkd3d_shader_code *out)
|
|
|
|
|
{
|
|
|
|
|
const struct hlsl_profile_info *profile = ctx->profile;
|
|
|
|
|
|
|
|
|
|
parse_entry_function_attributes(ctx, entry_func);
|
|
|
|
|
if (ctx->result)
|
|
|
|
|
return ctx->result;
|
|
|
|
|
|
|
|
|
|
if (profile->type == VKD3D_SHADER_TYPE_HULL)
|
|
|
|
|
validate_hull_shader_attributes(ctx, entry_func);
|
|
|
|
|
else if (profile->type == VKD3D_SHADER_TYPE_COMPUTE && !ctx->found_numthreads)
|
|
|
|
|
hlsl_error(ctx, &entry_func->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_ATTRIBUTE,
|
|
|
|
|
"Entry point \"%s\" is missing a [numthreads] attribute.", entry_func->func->name);
|
|
|
|
|
|
|
|
|
|
process_entry_function(ctx, entry_func);
|
|
|
|
|
if (ctx->result)
|
|
|
|
|
return ctx->result;
|
|
|
|
|
|
2021-08-16 15:29:34 -07:00
|
|
|
if (profile->major_version < 4)
|
2021-10-11 19:58:46 -07:00
|
|
|
{
|
2021-04-08 21:38:27 -07:00
|
|
|
allocate_const_registers(ctx, entry_func);
|
2021-10-11 19:58:46 -07:00
|
|
|
}
|
2021-06-23 21:57:35 -07:00
|
|
|
else
|
2021-10-11 19:58:46 -07:00
|
|
|
{
|
2021-06-23 21:57:35 -07:00
|
|
|
allocate_buffers(ctx);
|
2023-01-30 13:27:32 -08:00
|
|
|
allocate_objects(ctx, HLSL_REGSET_TEXTURES);
|
|
|
|
|
allocate_objects(ctx, HLSL_REGSET_UAVS);
|
2021-10-11 19:58:46 -07:00
|
|
|
}
|
2023-01-30 13:27:32 -08:00
|
|
|
allocate_objects(ctx, HLSL_REGSET_SAMPLERS);
|
2021-04-08 21:38:23 -07:00
|
|
|
|
2024-08-24 07:25:52 -07:00
|
|
|
if (TRACE_ON())
|
|
|
|
|
rb_for_each_entry(&ctx->functions, dump_function, ctx);
|
|
|
|
|
|
2021-05-20 22:32:23 -07:00
|
|
|
if (ctx->result)
|
|
|
|
|
return ctx->result;
|
2021-04-15 17:03:43 -07:00
|
|
|
|
2022-02-28 03:23:43 -08:00
|
|
|
switch (target_type)
|
|
|
|
|
{
|
|
|
|
|
case VKD3D_SHADER_TARGET_D3D_BYTECODE:
|
2024-05-17 15:30:59 -07:00
|
|
|
{
|
|
|
|
|
uint32_t config_flags = vkd3d_shader_init_config_flags();
|
2024-05-20 13:29:09 -07:00
|
|
|
struct vkd3d_shader_code ctab = {0};
|
2024-05-17 15:30:59 -07:00
|
|
|
struct vsir_program program;
|
|
|
|
|
int result;
|
|
|
|
|
|
2024-05-20 13:29:09 -07:00
|
|
|
sm1_generate_vsir(ctx, entry_func, config_flags, &program, &ctab);
|
2024-05-17 15:30:59 -07:00
|
|
|
if (ctx->result)
|
|
|
|
|
{
|
|
|
|
|
vsir_program_cleanup(&program);
|
2024-05-20 13:29:09 -07:00
|
|
|
vkd3d_shader_free_shader_code(&ctab);
|
2024-05-17 15:30:59 -07:00
|
|
|
return ctx->result;
|
|
|
|
|
}
|
|
|
|
|
|
2024-06-04 17:45:37 -07:00
|
|
|
result = d3dbc_compile(&program, config_flags, NULL, &ctab, out, ctx->message_context);
|
2024-05-17 15:30:59 -07:00
|
|
|
vsir_program_cleanup(&program);
|
2024-05-20 13:29:09 -07:00
|
|
|
vkd3d_shader_free_shader_code(&ctab);
|
2024-05-17 15:30:59 -07:00
|
|
|
return result;
|
|
|
|
|
}
|
2022-02-28 03:23:43 -08:00
|
|
|
|
|
|
|
|
case VKD3D_SHADER_TARGET_DXBC_TPF:
|
|
|
|
|
return hlsl_sm4_write(ctx, entry_func, out);
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
ERR("Unsupported shader target type %#x.\n", target_type);
|
|
|
|
|
return VKD3D_ERROR_INVALID_ARGUMENT;
|
|
|
|
|
}
|
2021-03-02 13:34:46 -08:00
|
|
|
}
|
