mesa/src/intel/compiler/elk/elk_predicated_break.cpp
Caio Oliveira 6648e0ebd3 intel/elk: Rename symbols
Either replace the BRW prefix with ELK or add an extra ELK prefix.  Used
the following sed script to perform the renames in this patch:

```
    # Simple prefix changes.
    s/\<BRW_/ELK_/g
    s/\<brw_/elk_/g
    s/nir_to_brw/nir_to_elk/g
    s/\<as_brw_reg\>/as_elk_reg/g
    s/\<_brw_/_elk_/g

    # Add prefix to various symbols.
    #
    # Initially I've considered using C++ namespaces here, but in various
    # cases the structs or functions had to be also visible from C code.
    # So added explicit prefix instead.
    s/\<backend_instruction/elk_\0/g
    s/\<backend_reg/elk_\0/g
    s/\<backend_shader/elk_\0/g
    s/\<bblock_t\>/elk_\0/g
    s/\<bblock_link\>/elk_\0/g
    s/\<cfg_t\>/elk_\0/g
    s/\<fs_visitor\>/elk_\0/g
    s/\<fs_reg\>/elk_\0/g
    s/\<fs_instruction_scheduler\>/elk_\0/g
    s/\<vec4_instruction_scheduler\>/elk_\0/g
    s/\<instruction_scheduler\>/elk_\0/g
    s/\<schedule_node\>/elk_\0/g
    s/\<schedule_node_child\>/elk_\0/g
    s/\<\([a-z]*_\)\?thread_payload\>/elk_\1thread_payload/g
    s/\<fs_generator\>/elk_\0/g
    s/\<fs_inst\>/elk_\0/g
    s/\<fs_reg_alloc\>/elk_\0/g
    s/\<disasm_info\>/elk_\0/g
    s/\<gfx._math\>/elk_\0/g
    s/\<gfx7_block_read_scratch\>/elk_\0/g
    s/\<gfx6_IF\>/elk_\0/g
    s/\<gfx9_fb_READ\>/elk_\0/g
    s/\<gfx6_resolve_implied_move\>/elk_\0/g

    # Opcodes.
    s/\<opcode op\>/elk_\0/g
    s/\<opcode mov_op\>/elk_\0/g
    s/\<opcode opcode\>/elk_\0/g
    s/enum opcode\>/enum elk_opcode/g
    s/static opcode\>/static elk_opcode/g
    s/\<opcode elk_op/elk_opcode elk_op/g
    s/struct opcode_desc/struct elk_opcode_desc/g
    s/NUM_BRW_OPCODES/NUM_ELK_OPCODES/g
    s/\<.._OPCODE_/ELK_\0/g
    s/\<T.._OPCODE_/ELK_\0/g
    s/\<VEC4_OPCODE_/ELK_\0/g
    s/\<VEC4_...\?_OPCODE_/ELK_\0/g
    s/\<SHADER_OPCODE_/ELK_\0/g

    # Remaining specific cases.
    s/\<wm_prog_data_barycentric_modes\>/elk_\0/g
    s/\<encode_slm_size\>/elk_\0/g
    s/\<intel_calculate_slm_size\>/elk_\0/g
    s/\<gfx6_gather_sampler_wa\>/elk_\0/g
    s/\<is_3src\>/elk_\0/g
    s/\<WA_/ELK_\0/g
    s/\<conditional_modifier\>/elk_\0/g
    s/\<pred_ctrl_align16\>/elk_\0/g
    s/\<shuffle_from_32bit_read\>/elk_\0/g
    s/\<shuffle_src_to_dst\>/elk_\0/g
    s/\<setup_imm_..\?\>/elk_\0/g

    s/\<opt_predicated_break\>/elk_\0/g
    s/\<has_bank_conflict\>/elk_\0/g
    s/\<dead_control_flow_eliminate\>/elk_\0/g

    s/\<disasm_new_inst_group\>/elk_\0/g
    s/\<disasm_initialize\>/elk_\0/g
    s/\<dump_assembly\>/elk_\0/g
    s/\<disasm_insert_error\>/elk_\0/g
    s/\<disasm_annotate\>/elk_\0/g

    s/\<enum lsc_opcode\>/enum elk_lsc_opcode/g
    s/\<lsc_opcode_/elk_lsc_opcode_/g
    s/\<lsc_aop_[a-z_]\+\>/elk_\0/g

    s/\<type_size_vec4\>/elk_\0/g
    s/\<type_size_dvec4\>/elk_\0/g
    s/\<type_size_xvec4\>/elk_\0/g
    s/\<type_size_[a-z4]\+_bytes\>/elk_\0/g

    s/\<gfx12_systolic_depth\>/elk_\0/g
```

Acked-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27563>
2024-02-24 00:24:31 +00:00

243 lines
7.6 KiB
C++

/*
* Copyright © 2013 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#include "elk_shader.h"
using namespace elk;
/** @file elk_predicated_break.cpp
*
* Loops are often structured as
*
* loop:
* CMP.f0
* (+f0) IF
* BREAK
* ENDIF
* ...
* WHILE loop
*
* This peephole pass removes the IF and ENDIF instructions and predicates the
* BREAK, dropping two instructions from the loop body.
*
* If the loop was a DO { ... } WHILE loop, it looks like
*
* loop:
* ...
* CMP.f0
* (+f0) IF
* BREAK
* ENDIF
* WHILE loop
*
* and we can remove the BREAK instruction and predicate the WHILE.
*/
#define MAX_NESTING 128
struct loop_continue_tracking {
BITSET_WORD has_continue[BITSET_WORDS(MAX_NESTING)];
unsigned depth;
};
static void
enter_loop(struct loop_continue_tracking *s)
{
s->depth++;
/* Any loops deeper than that maximum nesting will just re-use the last
* flag. This simplifies most of the code. MAX_NESTING is chosen to be
* large enough that it is unlikely to occur. Even if it does, the
* optimization that uses this tracking is unlikely to make much
* difference.
*/
if (s->depth < MAX_NESTING)
BITSET_CLEAR(s->has_continue, s->depth);
}
static void
exit_loop(struct loop_continue_tracking *s)
{
assert(s->depth > 0);
s->depth--;
}
static void
set_continue(struct loop_continue_tracking *s)
{
const unsigned i = MIN2(s->depth, MAX_NESTING - 1);
BITSET_SET(s->has_continue, i);
}
static bool
has_continue(const struct loop_continue_tracking *s)
{
const unsigned i = MIN2(s->depth, MAX_NESTING - 1);
return BITSET_TEST(s->has_continue, i);
}
bool
elk_opt_predicated_break(elk_backend_shader *s)
{
bool progress = false;
struct loop_continue_tracking state = { {0, }, 0 };
foreach_block (block, s->cfg) {
/* DO instructions, by definition, can only be found at the beginning of
* basic blocks.
*/
elk_backend_instruction *const do_inst = block->start();
/* BREAK, CONTINUE, and WHILE instructions, by definition, can only be
* found at the ends of basic blocks.
*/
elk_backend_instruction *jump_inst = block->end();
if (do_inst->opcode == ELK_OPCODE_DO)
enter_loop(&state);
if (jump_inst->opcode == ELK_OPCODE_CONTINUE)
set_continue(&state);
else if (jump_inst->opcode == ELK_OPCODE_WHILE)
exit_loop(&state);
if (block->start_ip != block->end_ip)
continue;
if (jump_inst->opcode != ELK_OPCODE_BREAK &&
jump_inst->opcode != ELK_OPCODE_CONTINUE)
continue;
elk_backend_instruction *if_inst = block->prev()->end();
if (if_inst->opcode != ELK_OPCODE_IF)
continue;
elk_backend_instruction *endif_inst = block->next()->start();
if (endif_inst->opcode != ELK_OPCODE_ENDIF)
continue;
elk_bblock_t *jump_block = block;
elk_bblock_t *if_block = jump_block->prev();
elk_bblock_t *endif_block = jump_block->next();
jump_inst->predicate = if_inst->predicate;
jump_inst->predicate_inverse = if_inst->predicate_inverse;
elk_bblock_t *earlier_block = if_block;
if (if_block->start_ip == if_block->end_ip) {
earlier_block = if_block->prev();
}
if_inst->remove(if_block);
elk_bblock_t *later_block = endif_block;
if (endif_block->start_ip == endif_block->end_ip) {
later_block = endif_block->next();
}
endif_inst->remove(endif_block);
if (!earlier_block->ends_with_control_flow()) {
/* FIXME: There is a potential problem here. If earlier_block starts
* with a DO instruction, this will delete the physical link to the
* WHILE block. It is unclear whether ENDIF has the same potential
* problem.
*/
assert(earlier_block->start() == NULL ||
earlier_block->start()->opcode != ELK_OPCODE_DO);
earlier_block->unlink_children();
earlier_block->add_successor(s->cfg->mem_ctx, jump_block,
bblock_link_logical);
}
if (!later_block->starts_with_control_flow()) {
later_block->unlink_parents();
}
/* If jump_block already has a link to later_block, don't create another
* one. Instead, promote the link to logical.
*/
bool need_to_link = true;
foreach_list_typed(elk_bblock_link, link, link, &jump_block->children) {
if (link->block == later_block) {
assert(later_block->starts_with_control_flow());
/* Update the link from later_block back to jump_block. */
foreach_list_typed(elk_bblock_link, parent_link, link, &later_block->parents) {
if (parent_link->block == jump_block) {
parent_link->kind = bblock_link_logical;
}
}
/* Update the link from jump_block to later_block. */
link->kind = bblock_link_logical;
need_to_link = false;
}
}
if (need_to_link) {
jump_block->add_successor(s->cfg->mem_ctx, later_block,
bblock_link_logical);
}
if (earlier_block->can_combine_with(jump_block)) {
earlier_block->combine_with(jump_block);
block = earlier_block;
}
/* Now look at the first instruction of the block following the BREAK. If
* it's a WHILE, we can delete the break, predicate the WHILE, and join
* the two basic blocks.
*
* This optimization can only be applied if the only instruction that
* can transfer control to the WHILE is the BREAK. If other paths can
* lead to the while, the flags may be in an unknown state, and the loop
* could terminate prematurely. This can occur if the loop contains a
* CONT instruction.
*/
elk_bblock_t *while_block = earlier_block->next();
elk_backend_instruction *while_inst = while_block->start();
if (jump_inst->opcode == ELK_OPCODE_BREAK &&
while_inst->opcode == ELK_OPCODE_WHILE &&
while_inst->predicate == ELK_PREDICATE_NONE &&
!has_continue(&state)) {
jump_inst->remove(earlier_block);
while_inst->predicate = jump_inst->predicate;
while_inst->predicate_inverse = !jump_inst->predicate_inverse;
assert(earlier_block->can_combine_with(while_block));
earlier_block->combine_with(while_block);
}
progress = true;
}
if (progress)
s->invalidate_analysis(DEPENDENCY_BLOCKS | DEPENDENCY_INSTRUCTIONS);
return progress;
}