ed/perfaware
1
0
Fork 0
mirror of https://github.com/Ed94/perfaware.git synced 2026-09-14 03:39:23 +00:00
Code Issues Packages Projects Releases Wiki Activity

Better str8_fmt_ktl_buf (using simd)

Browse Source
This commit is contained in:
ed
2026-09-08 20:45:57 -04:00
parent 749a712ceb
commit b9b1ed460a
8 changed files with 387 additions and 296 deletions
Download Patch File Download Diff File Expand all files Collapse all files
code/8086/decoder_table_generator.meta.c
+46 -55
View File
@@ -1,4 +1,5 @@
#include "duffle/dsl.h"
#include "duffle/asm.h"
#include "duffle/analysis.h"
#include "duffle/math.h"
#include "duffle/encoding.h"
@@ -274,9 +275,7 @@ x8616_decode_table_generate(X8616_DecodeGen* gen, FArena_R info_scratch) {
}
#ifndef X8616_DECODE_TABLE_OUTPUT
# define X8616_DECODE_TABLE_OUTPUT "./code/8086/gen/decoder_table.h"
#endif
#define X8616_DECODE_TABLE_OUTPUT "./code/8086/gen/decoder_table.h"
enum {
INFO_MEMORY_SIZE = kilo(64),
@@ -318,54 +317,51 @@ I_ Str8 str8_from_u4_opt(U4 num, Opt_str8_from_u4 o) { if (o.radix == 0) {o.radi
#define code_str8(...) slit8(stringify(__VA_ARGS__))
internal void
x8616_decode_gen_emit_plan(Str8Gen_R out, X8616_DecodePlan const* plan) {
defer_rewind(smem.scratch.top)
{
Str8 template = code_str8(
\t{
<flags>, <op>, <encoding_flags>, <width>,
{ <operands[0]> , <operands[1]> }, <operand_count>,
<payload>, <prefix_kind>,
<d_shift>, <w_shift>, <s_shift>, <v_shift>, <z_shift>, <reg_shift>, <sr_shift>,
{ <mod_rm.bits>, <mod_rm.mask> },
{ <post_opcode.bits>, <post_opcode.mask> },
},\n
);
KTL_Slot_Str8 tbl[] = {
#define dec(value) str8_from_u4(value, .radix = 10, .min_digits = 1)
#define hex_u1(value) str8_from_u4(value, .radix = 16, .min_digits = 2)
#define hex_u2(value) str8_from_u4(value, .radix = 16, .min_digits = 4)
#define entry(key,value) { ktl_str8_key(key), value }
entry("flags", hex_u2(plan->flags)),
entry("op", hex_u1(plan->op)),
entry("encoding_flags", hex_u1(plan->encoding_flags)),
entry("width", hex_u1(plan->width)),
entry("operands[0]", hex_u1(plan->operands[0])),
entry("operands[1]", hex_u1(plan->operands[1])),
entry("operand_count", dec(plan->operand_count)),
entry("payload", dec(plan->payload)),
entry("prefix_kind", dec(plan->prefix_kind)),
entry("d_shift", dec(plan->d_shift)),
entry("w_shift", dec(plan->w_shift)),
entry("s_shift", dec(plan->s_shift)),
entry("v_shift", dec(plan->v_shift)),
entry("z_shift", dec(plan->z_shift)),
entry("reg_shift", dec(plan->reg_shift)),
entry("sr_shift", dec(plan->sr_shift)),
entry("mod_rm.bits", hex_u1(plan->mod_rm.bits)),
entry("mod_rm.mask", hex_u1(plan->mod_rm.mask)),
entry("post_opcode.bits", hex_u1(plan->post_opcode.bits)),
entry("post_opcode.mask", hex_u1(plan->post_opcode.mask)),
#undef dec
#undef hex
#undef entry
};
str8gen_append_fmt(out, template, ktl_str8_from_arr(tbl));
}
}
x8616_decode_gen_emit_plan(Str8Gen_R out, X8616_DecodePlan_R plan) { defer_rewind(smem.scratch.top) {
Str8 template = code_str8(
\t{
<flags>, <op>, <encoding_flags>, <width>,
{<operands[0]>, <operands[1]>}, <operand_count>,
<payload>, <prefix_kind>,
<d_shift>, <w_shift>, <s_shift>, <v_shift>, <z_shift>, <reg_shift>, <sr_shift>,
{<mod_rm.bits>, <mod_rm.mask>},
{<post_opcode.bits>, <post_opcode.mask>},
},\n
);
KTL_Slot_Str8 tbl[] = {
#define dec(value) str8_from_u4(value, .radix = 10, .min_digits = 1)
#define hex_u1(value) str8_from_u4(value, .radix = 16, .min_digits = 2)
#define hex_u2(value) str8_from_u4(value, .radix = 16, .min_digits = 4)
#define entry(key,value) { ktl_str8_key(key), value }
entry("flags", hex_u2(plan->flags)),
entry("op", hex_u1(plan->op)),
entry("encoding_flags", hex_u1(plan->encoding_flags)),
entry("width", hex_u1(plan->width)),
entry("operands[0]", hex_u1(plan->operands[0])),
entry("operands[1]", hex_u1(plan->operands[1])),
entry("operand_count", dec(plan->operand_count)),
entry("payload", dec(plan->payload)),
entry("prefix_kind", dec(plan->prefix_kind)),
entry("d_shift", dec(plan->d_shift)),
entry("w_shift", dec(plan->w_shift)),
entry("s_shift", dec(plan->s_shift)),
entry("v_shift", dec(plan->v_shift)),
entry("z_shift", dec(plan->z_shift)),
entry("reg_shift", dec(plan->reg_shift)),
entry("sr_shift", dec(plan->sr_shift)),
entry("mod_rm.bits", hex_u1(plan->mod_rm.bits)),
entry("mod_rm.mask", hex_u1(plan->mod_rm.mask)),
entry("post_opcode.bits", hex_u1(plan->post_opcode.bits)),
entry("post_opcode.mask", hex_u1(plan->post_opcode.mask)),
#undef dec
#undef hex
#undef entry
};
str8gen_append_fmt(out, template, ktl_str8_from_arr(tbl));
}}
internal Str8
x8616_decode_gen_emit(Str8Gen_R out, X8616_DecodeGen const* gen) {
x8616_decode_gen_emit(Str8Gen_R out, X8616_DecodeGen_R gen) {
str8gen_append_str8(out, slit8(
"// Generated from encoder_table.h. Do not hand-edit.\n"
"// Plan 0 is the all-zero nil/invalid plan.\n\n"
@@ -397,17 +393,13 @@ x8616_decode_gen_emit(Str8Gen_R out, X8616_DecodeGen const* gen) {
return str8(out->ptr, out->len);
}
CLANG_OPTIMIZE_DISABLE
int
main(void) {
FArena info_scratch = farena_make(slice_ut_arr(smem.info));
X8616_DecodeGenInfo gen_info = x8616_decode_table_generate(& smem.gen, & info_scratch);
if (gen_info.msgs.error_count) { ms_exit_process(1); return 1; }
Str8Gen output = {
.ptr = C_(UTF8*, smem.text),
.cap = S_(smem.text),
};
Str8Gen output = str8gen_make(slice_ut_arr(smem.text));
Str8 generated = x8616_decode_gen_emit(& output, & smem.gen);
FArena file_scratch = farena_make(slice_ut_arr(smem.file));
@@ -418,4 +410,3 @@ main(void) {
ms_exit_process(0);
return 0;
}
CLANG_OPTIMIZE_ENABLE