1643 lines
48 KiB
C
1643 lines
48 KiB
C
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
|
|
│vi: set net ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi│
|
|
╞══════════════════════════════════════════════════════════════════════════════╡
|
|
│ Copyright 2020 Justine Alexandra Roberts Tunney │
|
|
│ │
|
|
│ Permission to use, copy, modify, and/or distribute this software for │
|
|
│ any purpose with or without fee is hereby granted, provided that the │
|
|
│ above copyright notice and this permission notice appear in all copies. │
|
|
│ │
|
|
│ THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL │
|
|
│ WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED │
|
|
│ WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE │
|
|
│ AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL │
|
|
│ DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR │
|
|
│ PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER │
|
|
│ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR │
|
|
│ PERFORMANCE OF THIS SOFTWARE. │
|
|
╚─────────────────────────────────────────────────────────────────────────────*/
|
|
#include "libc/bits/bits.h"
|
|
#include "libc/bits/popcnt.h"
|
|
#include "libc/calls/calls.h"
|
|
#include "libc/calls/struct/stat.h"
|
|
#include "libc/elf/def.h"
|
|
#include "libc/fmt/conv.h"
|
|
#include "libc/log/log.h"
|
|
#include "libc/macros.h"
|
|
#include "libc/mem/mem.h"
|
|
#include "libc/nexgen32e/bsr.h"
|
|
#include "libc/runtime/runtime.h"
|
|
#include "libc/stdio/stdio.h"
|
|
#include "libc/str/str.h"
|
|
#include "libc/sysv/consts/o.h"
|
|
#include "libc/x/x.h"
|
|
#include "third_party/gdtoa/gdtoa.h"
|
|
#include "tool/build/lib/elfwriter.h"
|
|
|
|
/* WORK IN PROGRESS */
|
|
|
|
#define OSZ 0x66
|
|
#define ASZ 0x67
|
|
#define REX 0x40 // byte
|
|
#define REXB 0x41 // src
|
|
#define REXX 0x42 // index
|
|
#define REXR 0x44 // dest
|
|
#define REXW 0x48 // quad
|
|
#define HASASZ 0x00010000
|
|
#define HASBASE 0x00020000
|
|
#define HASINDEX 0x00040000
|
|
#define ISRIP 0x00080000
|
|
#define ISREG 0x00100000
|
|
|
|
#define APPEND(L) L.p = realloc(L.p, ++L.n * sizeof(*L.p))
|
|
#define IS(P, N, S) (N == strlen(S) && !strncasecmp(P, S, strlen(S)))
|
|
|
|
struct Strings {
|
|
size_t n;
|
|
char **p;
|
|
};
|
|
|
|
struct Assembler {
|
|
int i;
|
|
int section;
|
|
int previous;
|
|
struct Ints {
|
|
size_t n;
|
|
long *p;
|
|
} ints;
|
|
struct Floats {
|
|
size_t n;
|
|
long double *p;
|
|
} floats;
|
|
struct Slices {
|
|
size_t n;
|
|
struct Slice {
|
|
size_t n;
|
|
char *p;
|
|
} * p;
|
|
} slices;
|
|
struct Sauces {
|
|
size_t n;
|
|
struct Sauce {
|
|
const char *path;
|
|
int line;
|
|
} * p;
|
|
} sauces;
|
|
struct Things {
|
|
size_t n;
|
|
struct Thing {
|
|
enum ThingType {
|
|
TT_INT,
|
|
TT_FLOAT,
|
|
TT_SLICE,
|
|
TT_PUNCT,
|
|
TT_FORWARD,
|
|
TT_BACKWARD,
|
|
} t;
|
|
int i;
|
|
int s;
|
|
} * p;
|
|
} things;
|
|
struct Sections {
|
|
size_t n;
|
|
struct Section {
|
|
const char *name;
|
|
int flags;
|
|
int type;
|
|
int align;
|
|
struct Slice binary;
|
|
} * p;
|
|
} sections;
|
|
struct Symbols {
|
|
size_t n;
|
|
struct Symbol {
|
|
int name;
|
|
int section;
|
|
int stb;
|
|
int stv;
|
|
int type;
|
|
long location;
|
|
long size;
|
|
} * p;
|
|
} symbols;
|
|
struct Labels {
|
|
size_t n;
|
|
struct Label {
|
|
int s;
|
|
int id;
|
|
int section;
|
|
long location;
|
|
} * p;
|
|
} labels;
|
|
};
|
|
|
|
static const char kPrefixByte[30] = {
|
|
0xf3, 0xf3, 0xf3, 0xf2, 0xf2, 0xf0, 0x26, 0x2e, 0x36, 0x3e,
|
|
0x64, 0x65, 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47,
|
|
0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, 0x66, 0x67,
|
|
};
|
|
|
|
static const char kPrefix[30][8] = {
|
|
"rep", "repz", "repe", "repne", "repnz", "lock",
|
|
"es", "cs", "ss", "ds", "fs", "gs",
|
|
"rex", "rex.b", "rex.x", "rex.xb", "rex.r", "rex.rb",
|
|
"rex.rx", "rex.rxb", "rex.w", "rex.wb", "rex.wx", "rex.wxb",
|
|
"rex.wr", "rex.wrb", "rex.wrx", "rex.wrxb", "data16", "addr32",
|
|
};
|
|
|
|
/**
|
|
* Context-sensitive register encoding information.
|
|
*
|
|
* ┌rex
|
|
* │ ┌log₂size
|
|
* │ │ ┌reg
|
|
* ├──────┐ ├─┐├─┐
|
|
* 0b0000000000000000
|
|
*/
|
|
static const struct Reg {
|
|
char s[8];
|
|
short reg;
|
|
short rm;
|
|
short base;
|
|
short index;
|
|
} kRegs[] = /* clang-format off */ {
|
|
{"ah", 4, 4, -1, -1 },
|
|
{"al", 0, 0, -1, -1 },
|
|
{"ax", 0 | 1<<3, 0 | 1<<3, -1, -1 },
|
|
{"bh", 7, 7, -1, -1 },
|
|
{"bl", 3, 3, -1, -1 },
|
|
{"bp", 5 | 1<<3, 5 | 1<<3, -1, -1 },
|
|
{"bpl", 5 | REX<<8, 5 | REX<<8, -1, -1 },
|
|
{"bx", 3 | 1<<3, 3 | 1<<3, -1, -1 },
|
|
{"ch", 5, 5, -1, -1 },
|
|
{"cl", 1, 1, -1, -1 },
|
|
{"cx", 1 | 1<<3, 1 | 1<<3, -1, -1 },
|
|
{"dh", 6, 6, -1, -1 },
|
|
{"di", 7 | 1<<3, 7 | 1<<3, -1, -1 },
|
|
{"dil", 7 | REX<<8, 7 | REX<<8, -1, -1 },
|
|
{"dl", 2, 2, -1, -1 },
|
|
{"dx", 2 | 1<<3, 2 | 1<<3, -1, -1 },
|
|
{"eax", 0 | 2<<3, 0 | 2<<3, 0 | 2<<3, 0 | 2<<3 },
|
|
{"ebp", 5 | 2<<3, 5 | 2<<3, 5 | 2<<3, 5 | 2<<3 },
|
|
{"ebx", 3 | 2<<3, 3 | 2<<3, 3 | 2<<3, 3 | 2<<3 },
|
|
{"ecx", 1 | 2<<3, 1 | 2<<3, 1 | 2<<3, 1 | 2<<3 },
|
|
{"edi", 7 | 2<<3, 7 | 2<<3, 7 | 2<<3, 7 | 2<<3 },
|
|
{"edx", 2 | 2<<3, 2 | 2<<3, 2 | 2<<3, 2 | 2<<3 },
|
|
{"esi", 6 | 2<<3, 6 | 2<<3, 6 | 2<<3, 6 | 2<<3 },
|
|
{"esp", 4 | 2<<3, 4 | 2<<3, 4 | 2<<3, 4 | 2<<3 },
|
|
{"r10", 2 | 3<<3 | REXR<<8 | REXW<<8, 2 | 3<<3 | REXB<<8 | REXW<<8, 2 | 3<<3 | REXB<<8, 2 | 3<<3 | REXX<<8 },
|
|
{"r10b", 2 | REXR<<8, 2 | REXB<<8, -1, -1 },
|
|
{"r10d", 2 | 2<<3 | REXR<<8, 2 | 2<<3 | REXB<<8, 2 | 2<<3 | REXB<<8, 2 | 2<<3 | REXX<<8 },
|
|
{"r10w", 2 | 1<<3 | REXR<<8, 2 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"r11", 3 | 3<<3 | REXR<<8 | REXW<<8, 3 | 3<<3 | REXB<<8 | REXW<<8, 3 | 3<<3 | REXB<<8, 3 | 3<<3 | REXX<<8 },
|
|
{"r11b", 3 | REXR<<8, 3 | REXB<<8, -1, -1 },
|
|
{"r11d", 3 | 2<<3 | REXR<<8, 3 | 2<<3 | REXB<<8, 3 | 2<<3 | REXB<<8, 3 | 2<<3 | REXX<<8 },
|
|
{"r11w", 3 | 1<<3 | REXR<<8, 3 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"r12", 4 | 3<<3 | REXR<<8 | REXW<<8, 4 | 3<<3 | REXB<<8 | REXW<<8, 4 | 3<<3 | REXB<<8, 4 | 3<<3 | REXX<<8 },
|
|
{"r12b", 4 | REXR<<8, 4 | REXB<<8, -1, -1 },
|
|
{"r12d", 4 | 2<<3 | REXR<<8, 4 | 2<<3 | REXB<<8, 4 | 2<<3 | REXB<<8, 4 | 2<<3 | REXX<<8 },
|
|
{"r12w", 4 | 1<<3 | REXR<<8, 4 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"r13", 5 | 3<<3 | REXR<<8 | REXW<<8, 5 | 3<<3 | REXB<<8 | REXW<<8, 5 | 3<<3 | REXB<<8, 5 | 3<<3 | REXX<<8 },
|
|
{"r13b", 5 | REXR<<8, 5 | REXB<<8, -1, -1 },
|
|
{"r13d", 5 | 2<<3 | REXR<<8, 5 | 2<<3 | REXB<<8, 5 | 2<<3 | REXB<<8, 5 | 2<<3 | REXX<<8 },
|
|
{"r13w", 5 | 1<<3 | REXR<<8, 5 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"r14", 6 | 3<<3 | REXR<<8 | REXW<<8, 6 | 3<<3 | REXB<<8 | REXW<<8, 6 | 3<<3 | REXB<<8, 6 | 3<<3 | REXX<<8 },
|
|
{"r14b", 6 | REXR<<8, 6 | REXB<<8, -1, -1 },
|
|
{"r14d", 6 | 2<<3 | REXR<<8, 6 | 2<<3 | REXB<<8, 6 | 2<<3 | REXB<<8, 6 | 2<<3 | REXX<<8 },
|
|
{"r14w", 6 | 1<<3 | REXR<<8, 6 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"r15", 7 | 3<<3 | REXR<<8 | REXW<<8, 7 | 3<<3 | REXB<<8 | REXW<<8, 7 | 3<<3 | REXB<<8, 7 | 3<<3 | REXX<<8 },
|
|
{"r15b", 7 | REXR<<8, 7 | REXB<<8, -1, -1 },
|
|
{"r15d", 7 | 2<<3 | REXR<<8, 7 | 2<<3 | REXB<<8, 7 | 2<<3 | REXB<<8, 7 | 2<<3 | REXX<<8 },
|
|
{"r15w", 7 | 1<<3 | REXR<<8, 7 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"r8", 0 | 3<<3 | REXR<<8 | REXW<<8, 0 | 3<<3 | REXB<<8 | REXW<<8, 0 | 3<<3 | REXB<<8, 0 | 3<<3 | REXX<<8 },
|
|
{"r8b", 0 | REXR<<8, 0 | REXB<<8, -1, -1 },
|
|
{"r8d", 0 | 2<<3 | REXR<<8, 0 | 2<<3 | REXB<<8, 0 | 2<<3 | REXB<<8, 0 | 2<<3 | REXX<<8 },
|
|
{"r8w", 0 | 1<<3 | REXR<<8, 0 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"r9", 1 | 3<<3 | REXR<<8 | REXW<<8, 1 | 3<<3 | REXB<<8 | REXW<<8, 1 | 3<<3 | REXB<<8, 1 | 3<<3 | REXX<<8 },
|
|
{"r9b", 1 | REXR<<8, 1 | REXB<<8, -1, -1 },
|
|
{"r9d", 1 | 2<<3 | REXR<<8, 1 | 2<<3 | REXB<<8, 1 | 2<<3 | REXB<<8, 1 | 2<<3 | REXX<<8 },
|
|
{"r9w", 1 | 1<<3 | REXR<<8, 1 | 1<<3 | REXB<<8, -1, -1 },
|
|
{"rax", 0 | 3<<3 | REXW<<8, 0 | 3<<3 | REXW<<8, 0 | 3<<3, 0 | 3<<3 },
|
|
{"rbp", 5 | 3<<3 | REXW<<8, 5 | 3<<3 | REXW<<8, 5 | 3<<3, 5 | 3<<3 },
|
|
{"rbx", 3 | 3<<3 | REXW<<8, 3 | 3<<3 | REXW<<8, 3 | 3<<3, 3 | 3<<3 },
|
|
{"rcx", 1 | 3<<3 | REXW<<8, 1 | 3<<3 | REXW<<8, 1 | 3<<3, 1 | 3<<3 },
|
|
{"rdi", 7 | 3<<3 | REXW<<8, 7 | 3<<3 | REXW<<8, 7 | 3<<3, 7 | 3<<3 },
|
|
{"rdx", 2 | 3<<3 | REXW<<8, 2 | 3<<3 | REXW<<8, 2 | 3<<3, 2 | 3<<3 },
|
|
{"rsi", 6 | 3<<3 | REXW<<8, 6 | 3<<3 | REXW<<8, 6 | 3<<3, 6 | 3<<3 },
|
|
{"rsp", 4 | 3<<3 | REXW<<8, 4 | 3<<3 | REXW<<8, 4 | 3<<3, 4 | 3<<3 },
|
|
{"si", 6 | 1<<3, 6 | 1<<3, 6 | 1<<3, 6 | 1<<3 },
|
|
{"sil", 6 | REX<<8, 6 | REX<<8, 6 | REX<<8, 6 | REX<<8 },
|
|
{"sp", 4 | 1<<3, 4 | 1<<3, 4 | 1<<3, 4 | 1<<3 },
|
|
{"spl", 4 | REX<<8, 4 | REX<<8, 4 | REX<<8, 4 | REX<<8 },
|
|
{"xmm0", 0 | 4<<3, 0 | 4<<3, -1, -1 },
|
|
{"xmm1", 1 | 4<<3, 1 | 4<<3, -1, -1 },
|
|
{"xmm10", 2 | 4<<3 | REXR<<8, 2 | 4<<3 | REXB<<8, -1, -1 },
|
|
{"xmm11", 3 | 4<<3 | REXR<<8, 3 | 4<<3 | REXB<<8, -1, -1 },
|
|
{"xmm12", 4 | 4<<3 | REXR<<8, 4 | 4<<3 | REXB<<8, -1, -1 },
|
|
{"xmm13", 5 | 4<<3 | REXR<<8, 5 | 4<<3 | REXB<<8, -1, -1 },
|
|
{"xmm14", 6 | 4<<3 | REXR<<8, 6 | 4<<3 | REXB<<8, -1, -1 },
|
|
{"xmm15", 7 | 4<<3 | REXR<<8, 7 | 4<<3 | REXB<<8, -1, -1 },
|
|
{"xmm2", 2 | 4<<3, 2 | 4<<3, -1, -1 },
|
|
{"xmm3", 3 | 4<<3, 3 | 4<<3, -1, -1 },
|
|
{"xmm4", 4 | 4<<3, 4 | 4<<3, -1, -1 },
|
|
{"xmm5", 5 | 4<<3, 5 | 4<<3, -1, -1 },
|
|
{"xmm6", 6 | 4<<3, 6 | 4<<3, -1, -1 },
|
|
{"xmm7", 7 | 4<<3, 7 | 4<<3, -1, -1 },
|
|
{"xmm8", 0 | 4<<3 | REXR<<8, 0 | 4<<3 | REXB<<8, -1, -1 },
|
|
{"xmm9", 1 | 4<<3 | REXR<<8, 1 | 4<<3 | REXB<<8, -1, -1 },
|
|
} /* clang-format on */;
|
|
|
|
static bool g_ignore_err;
|
|
static const char *g_input_path;
|
|
static const char *g_output_path;
|
|
static struct Strings g_include_paths;
|
|
|
|
static void AppendString(struct Strings *l, const char *p) {
|
|
l->p = realloc(l->p, ++l->n * sizeof(*l->p));
|
|
l->p[l->n - 1] = p;
|
|
}
|
|
|
|
static void ReadFlags(int argc, char *argv[]) {
|
|
int i;
|
|
g_input_path = "-";
|
|
g_output_path = "a.out";
|
|
for (i = 1; i < argc; ++i) {
|
|
if (!strcmp(argv[i], "-o")) {
|
|
g_output_path = argv[++i];
|
|
} else if (startswith(argv[i], "-o")) {
|
|
g_output_path = argv[i] + 2;
|
|
} else if (!strcmp(argv[i], "-I")) {
|
|
AppendString(&g_include_paths, argv[++i]);
|
|
} else if (startswith(argv[i], "-I")) {
|
|
AppendString(&g_include_paths, argv[i] + 2);
|
|
} else if (!strcmp(argv[i], "-Z")) {
|
|
g_ignore_err = true;
|
|
} else if (argv[i][0] != '-') {
|
|
g_input_path = argv[i];
|
|
}
|
|
}
|
|
}
|
|
|
|
static int AppendSection(struct Assembler *a, const char *name, int flags,
|
|
int type) {
|
|
int i;
|
|
APPEND(a->sections);
|
|
i = a->sections.n - 1;
|
|
a->sections.p[i].name = name;
|
|
a->sections.p[i].flags = flags;
|
|
a->sections.p[i].type = type;
|
|
a->sections.p[i].align = 1;
|
|
a->sections.p[i].binary.p = NULL;
|
|
a->sections.p[i].binary.n = 0;
|
|
return i;
|
|
}
|
|
|
|
static struct Assembler *NewAssembler(void) {
|
|
struct Assembler *a = calloc(1, sizeof(struct Assembler));
|
|
APPEND(a->slices);
|
|
a->slices.p[0].p = NULL;
|
|
a->slices.p[0].n = 0;
|
|
AppendSection(a, "", 0, SHT_NULL);
|
|
AppendSection(a, ".text", SHF_ALLOC | SHF_EXECINSTR, SHT_PROGBITS);
|
|
AppendSection(a, ".data", SHF_ALLOC | SHF_WRITE, SHT_PROGBITS);
|
|
AppendSection(a, ".bss", SHF_ALLOC | SHF_WRITE, SHT_NOBITS);
|
|
a->section = 1;
|
|
return a;
|
|
}
|
|
|
|
static int AppendLine(struct Assembler *a, char *path, int line) {
|
|
if (!a->sauces.n || (line != a->sauces.p[a->sauces.n - 1].line ||
|
|
path != a->sauces.p[a->sauces.n - 1].path)) {
|
|
APPEND(a->sauces);
|
|
a->sauces.p[a->sauces.n - 1].path = path;
|
|
a->sauces.p[a->sauces.n - 1].line = line;
|
|
}
|
|
return a->sauces.n - 1;
|
|
}
|
|
|
|
static int ReadCharLiteral(struct Slice *buf, int c, char *p, int *i) {
|
|
if (c != '\\') return c;
|
|
switch ((c = p[(*i)++])) {
|
|
case 'a':
|
|
return '\a';
|
|
case 'b':
|
|
return '\b';
|
|
case 't':
|
|
return '\t';
|
|
case 'n':
|
|
return '\n';
|
|
case 'v':
|
|
return '\v';
|
|
case 'f':
|
|
return '\f';
|
|
case 'r':
|
|
return '\r';
|
|
case 'e':
|
|
return 033;
|
|
case 'x':
|
|
if (isxdigit(p[*i])) {
|
|
c = hextoint(p[(*i)++]);
|
|
if (isxdigit(p[*i])) {
|
|
c = c * 16 + hextoint(p[(*i)++]);
|
|
}
|
|
}
|
|
return c;
|
|
case '0' ... '7':
|
|
c -= '0';
|
|
if ('0' <= p[*i] && p[*i] <= '7') {
|
|
c = c * 8 + (p[(*i)++] - '0');
|
|
if ('0' <= p[*i] && p[*i] <= '7') {
|
|
c = c * 8 + (p[(*i)++] - '0');
|
|
}
|
|
}
|
|
return c;
|
|
default:
|
|
return c;
|
|
}
|
|
}
|
|
|
|
static void CanonicalizeNewline(char *p) {
|
|
int i = 0, j = 0;
|
|
while (p[i]) {
|
|
if (p[i] == '\r' && p[i + 1] == '\n') {
|
|
i += 2;
|
|
p[j++] = '\n';
|
|
} else if (p[i] == '\r') {
|
|
i++;
|
|
p[j++] = '\n';
|
|
} else {
|
|
p[j++] = p[i++];
|
|
}
|
|
}
|
|
p[j] = '\0';
|
|
}
|
|
|
|
static void RemoveBackslashNewline(char *p) {
|
|
int i, j, n;
|
|
for (i = j = n = 0; p[i];) {
|
|
if (p[i] == '\\' && p[i + 1] == '\n') {
|
|
i += 2;
|
|
n++;
|
|
} else if (p[i] == '\n') {
|
|
p[j++] = p[i++];
|
|
for (; n > 0; n--) p[j++] = '\n';
|
|
} else {
|
|
p[j++] = p[i++];
|
|
}
|
|
}
|
|
for (; n > 0; n--) p[j++] = '\n';
|
|
p[j] = '\0';
|
|
}
|
|
|
|
static char *ReadFile(const char *path) {
|
|
char *p;
|
|
FILE *fp;
|
|
int buflen, nread, end, n;
|
|
if (strcmp(path, "-") == 0) {
|
|
fp = stdin;
|
|
} else {
|
|
fp = fopen(path, "r");
|
|
if (!fp) return NULL;
|
|
}
|
|
buflen = 4096;
|
|
nread = 0;
|
|
p = calloc(1, buflen);
|
|
for (;;) {
|
|
end = buflen - 2;
|
|
n = fread(p + nread, 1, end - nread, fp);
|
|
if (n == 0) break;
|
|
nread += n;
|
|
if (nread == end) {
|
|
buflen *= 2;
|
|
p = realloc(p, buflen);
|
|
}
|
|
}
|
|
if (fp != stdin) fclose(fp);
|
|
if (nread > 0 && p[nread - 1] == '\\') {
|
|
p[nread - 1] = '\n';
|
|
} else if (nread == 0 || p[nread - 1] != '\n') {
|
|
p[nread++] = '\n';
|
|
}
|
|
p[nread] = '\0';
|
|
return p;
|
|
}
|
|
|
|
static void Tokenize(struct Assembler *a, char *path) {
|
|
bool bol;
|
|
int c, i, line;
|
|
char *p, *path2;
|
|
struct Slice buf;
|
|
if (!(p = ReadFile(path))) return;
|
|
if (!memcmp(p, "\357\273\277", 3)) p += 3;
|
|
CanonicalizeNewline(p);
|
|
RemoveBackslashNewline(p);
|
|
line = 1;
|
|
bol = true;
|
|
while ((c = *p)) {
|
|
if (c == '#' || (c == '/' && bol)) {
|
|
p = strchr(p, '\n') + 1;
|
|
bol = true;
|
|
++line;
|
|
continue;
|
|
}
|
|
if (c == '\n') {
|
|
c = ';';
|
|
bol = true;
|
|
++line;
|
|
} else {
|
|
bol = false;
|
|
if (c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' ||
|
|
c == '\v' || c == ',') {
|
|
++p;
|
|
continue;
|
|
}
|
|
if (('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z') || c == '_' ||
|
|
c == '%' || c == '@' || (c == '.' && !('0' <= p[1] && p[1] <= '9'))) {
|
|
for (i = 1;; ++i) {
|
|
if (!(('a' <= p[i] && p[i] <= 'z') || ('A' <= p[i] && p[i] <= 'Z') ||
|
|
('0' <= p[i] && p[i] <= '9') || p[i] == '.' || p[i] == '_' ||
|
|
p[i] == '$' || p[i] == '(' || p[i] == ')')) {
|
|
break;
|
|
}
|
|
}
|
|
APPEND(a->things);
|
|
a->things.p[a->things.n - 1].t = TT_SLICE;
|
|
a->things.p[a->things.n - 1].s = AppendLine(a, path, line);
|
|
a->things.p[a->things.n - 1].i = a->slices.n;
|
|
APPEND(a->slices);
|
|
a->slices.p[a->slices.n - 1].p = p;
|
|
a->slices.p[a->slices.n - 1].n = i;
|
|
p += i;
|
|
continue;
|
|
}
|
|
if (('0' <= c && c <= '9') || (c == '.' && '0' <= p[1] && p[1] <= '9')) {
|
|
bool isfloat = c == '.';
|
|
if (c == '0' && p[1] != '.') {
|
|
if (p[1] == 'x' || p[1] == 'X') {
|
|
for (i = 2;; ++i) {
|
|
if (!(('0' <= p[i] && p[i] <= '9') ||
|
|
('a' <= p[i] && p[i] <= 'f') ||
|
|
('A' <= p[i] && p[i] <= 'F'))) {
|
|
break;
|
|
}
|
|
}
|
|
} else if (p[1] == 'b' || p[1] == 'B') {
|
|
for (i = 2;; ++i) {
|
|
if (!(p[i] == '0' || p[i] == '1')) break;
|
|
}
|
|
} else {
|
|
for (i = 1;; ++i) {
|
|
if (!('0' <= p[i] && p[i] <= '7')) break;
|
|
}
|
|
}
|
|
} else {
|
|
for (i = 1;; ++i) {
|
|
if (('0' <= p[i] && p[i] <= '9') || p[i] == '-' || p[i] == '+') {
|
|
continue;
|
|
} else if (p[i] == '.' || p[i] == 'e' || p[i] == 'E' ||
|
|
p[i] == 'e') {
|
|
isfloat = true;
|
|
continue;
|
|
}
|
|
break;
|
|
}
|
|
}
|
|
APPEND(a->things);
|
|
if (isfloat) {
|
|
APPEND(a->floats);
|
|
a->floats.p[a->floats.n - 1] = strtold(p, NULL);
|
|
a->things.p[a->things.n - 1].i = a->floats.n - 1;
|
|
a->things.p[a->things.n - 1].t = TT_FLOAT;
|
|
} else {
|
|
APPEND(a->ints);
|
|
a->ints.p[a->ints.n - 1] = strtol(p, NULL, 0);
|
|
a->things.p[a->things.n - 1].i = a->ints.n - 1;
|
|
if (p[i] == 'f' || p[i] == 'F') {
|
|
a->things.p[a->things.n - 1].t = TT_FORWARD;
|
|
} else if (p[i] == 'b' || p[i] == 'B') {
|
|
a->things.p[a->things.n - 1].t = TT_BACKWARD;
|
|
} else {
|
|
a->things.p[a->things.n - 1].t = TT_INT;
|
|
}
|
|
}
|
|
a->things.p[a->things.n - 1].s = AppendLine(a, path, line);
|
|
p += i;
|
|
continue;
|
|
}
|
|
if (c == '\'') {
|
|
i = 1;
|
|
c = p[i++];
|
|
c = ReadCharLiteral(&buf, c, p, &i);
|
|
if (p[i] == '\'') ++i;
|
|
p += i;
|
|
APPEND(a->things);
|
|
a->things.p[a->things.n - 1].t = TT_INT;
|
|
a->things.p[a->things.n - 1].s = AppendLine(a, path, line);
|
|
a->things.p[a->things.n - 1].i = a->ints.n;
|
|
APPEND(a->ints);
|
|
a->ints.p[a->ints.n - 1] = c;
|
|
continue;
|
|
}
|
|
if (c == '"') {
|
|
buf.n = 0;
|
|
buf.p = NULL;
|
|
for (i = 1; (c = p[i++]);) {
|
|
if (c == '"') break;
|
|
c = ReadCharLiteral(&buf, c, p, &i);
|
|
APPEND(buf);
|
|
buf.p[buf.n - 1] = c;
|
|
}
|
|
p += i;
|
|
if (a->things.n && a->things.p[a->things.n - 1].t == TT_SLICE &&
|
|
IS(a->slices.p[a->things.p[a->things.n - 1].i].p,
|
|
a->slices.p[a->things.p[a->things.n - 1].i].n, ".include")) {
|
|
APPEND(buf);
|
|
buf.p[buf.n - 1] = '\0';
|
|
--a->things.n;
|
|
if (fileexists(buf.p)) {
|
|
Tokenize(a, buf.p);
|
|
} else {
|
|
for (i = 0; i < g_include_paths.n; ++i) {
|
|
path2 = xstrcat(g_include_paths.p[i], '/', buf.p);
|
|
if (fileexists(path2)) {
|
|
Tokenize(a, path2);
|
|
free(path2);
|
|
break;
|
|
} else {
|
|
free(path2);
|
|
}
|
|
}
|
|
}
|
|
free(buf.p);
|
|
} else {
|
|
APPEND(a->things);
|
|
a->things.p[a->things.n - 1].t = TT_SLICE;
|
|
a->things.p[a->things.n - 1].s = AppendLine(a, path, line);
|
|
a->things.p[a->things.n - 1].i = a->slices.n;
|
|
APPEND(a->slices);
|
|
a->slices.p[a->slices.n - 1] = buf;
|
|
}
|
|
continue;
|
|
}
|
|
}
|
|
APPEND(a->things);
|
|
a->things.p[a->things.n - 1].t = TT_PUNCT;
|
|
a->things.p[a->things.n - 1].s = AppendLine(a, path, line);
|
|
a->things.p[a->things.n - 1].i = c;
|
|
++p;
|
|
}
|
|
}
|
|
|
|
static bool IsLocal(struct Assembler *a, int name) {
|
|
if (name < 0) return true;
|
|
return a->slices.p[name].n >= 2 && !memcmp(a->slices.p[name].p, ".L", 2);
|
|
}
|
|
|
|
static int GetSymbol(struct Assembler *a, int name) {
|
|
int i;
|
|
for (i = 0; i < a->symbols.n; ++i) {
|
|
if (a->slices.p[a->symbols.p[i].name].n == a->slices.p[name].n &&
|
|
!memcmp(a->slices.p[a->symbols.p[i].name].p, a->slices.p[name].p,
|
|
a->slices.p[name].n)) {
|
|
return i;
|
|
}
|
|
}
|
|
APPEND(a->symbols);
|
|
i = a->symbols.n - 1;
|
|
memset(&a->symbols.p[i], 0, sizeof(a->symbols.p[i]));
|
|
a->symbols.p[i].name = name;
|
|
return i;
|
|
}
|
|
|
|
static wontreturn void Fail(struct Assembler *a, const char *fmt, ...) {
|
|
va_list va;
|
|
fprintf(stderr, "%s:%d:: ", a->sauces.p[a->things.p[a->i].s].path,
|
|
a->sauces.p[a->things.p[a->i].s].line);
|
|
va_start(va, fmt);
|
|
vfprintf(stderr, fmt, va);
|
|
va_end(va);
|
|
fputc('\n', stderr);
|
|
exit(1);
|
|
}
|
|
|
|
static void Label(struct Assembler *a, int name) {
|
|
int i = GetSymbol(a, name);
|
|
if (a->symbols.p[i].section) {
|
|
Fail(a, "already defined: %.*s", a->slices.p[name].n, a->slices.p[name].p);
|
|
}
|
|
a->symbols.p[i].section = a->section;
|
|
a->symbols.p[i].location = a->sections.p[a->section].binary.n;
|
|
a->i += 2;
|
|
}
|
|
|
|
static void LocalLabel(struct Assembler *a, int id) {
|
|
APPEND(a->labels);
|
|
a->labels.p[a->labels.n - 1].s = a->things.p[a->i].s;
|
|
a->labels.p[a->labels.n - 1].id = id;
|
|
a->labels.p[a->labels.n - 1].section = a->section;
|
|
a->labels.p[a->labels.n - 1].location = a->sections.p[a->section].binary.n;
|
|
a->i += 2;
|
|
}
|
|
|
|
static void SetSection(struct Assembler *a, int section) {
|
|
a->previous = a->section;
|
|
a->section = section;
|
|
}
|
|
|
|
static bool IsInt(struct Assembler *a, int i) {
|
|
return a->things.p[i].t == TT_INT;
|
|
}
|
|
|
|
static bool IsFloat(struct Assembler *a, int i) {
|
|
return a->things.p[i].t == TT_FLOAT;
|
|
}
|
|
|
|
static bool IsSlice(struct Assembler *a, int i) {
|
|
return a->things.p[i].t == TT_SLICE;
|
|
}
|
|
|
|
static bool IsPunct(struct Assembler *a, int i, int c) {
|
|
return a->things.p[i].t == TT_PUNCT && a->things.p[i].i == c;
|
|
}
|
|
|
|
static void ConsumePunct(struct Assembler *a, int c) {
|
|
if (IsPunct(a, a->i, c)) {
|
|
++a->i;
|
|
} else {
|
|
Fail(a, "expected %`'c", c);
|
|
}
|
|
}
|
|
|
|
static long GetInt(struct Assembler *a) {
|
|
long res;
|
|
if (IsInt(a, a->i)) {
|
|
res = a->ints.p[a->things.p[a->i].i];
|
|
++a->i;
|
|
return res;
|
|
} else {
|
|
Fail(a, "expected int");
|
|
}
|
|
}
|
|
|
|
static long double GetFloat(struct Assembler *a) {
|
|
long double res;
|
|
if (IsFloat(a, a->i)) {
|
|
res = a->floats.p[a->things.p[a->i].i];
|
|
++a->i;
|
|
return res;
|
|
} else {
|
|
Fail(a, "expected float");
|
|
}
|
|
}
|
|
|
|
static struct Slice GetString(struct Assembler *a) {
|
|
struct Slice res;
|
|
if (IsSlice(a, a->i)) {
|
|
res = a->slices.p[a->things.p[a->i].i];
|
|
++a->i;
|
|
return res;
|
|
} else {
|
|
Fail(a, "expected string");
|
|
}
|
|
}
|
|
|
|
static void Emit(struct Assembler *a, int c) {
|
|
APPEND(a->sections.p[a->section].binary);
|
|
a->sections.p[a->section].binary.p[a->sections.p[a->section].binary.n - 1] =
|
|
c;
|
|
}
|
|
|
|
static void EmitWord(struct Assembler *a, long x) {
|
|
Emit(a, x >> 000);
|
|
Emit(a, x >> 010);
|
|
}
|
|
|
|
static void EmitLong(struct Assembler *a, long x) {
|
|
Emit(a, x >> 000);
|
|
Emit(a, x >> 010);
|
|
Emit(a, x >> 020);
|
|
Emit(a, x >> 030);
|
|
}
|
|
|
|
static void EmitQuad(struct Assembler *a, long x) {
|
|
EmitLong(a, x >> 000);
|
|
EmitLong(a, x >> 040);
|
|
}
|
|
|
|
static void OnZero(struct Assembler *a) {
|
|
long i, n;
|
|
while (IsInt(a, a->i)) {
|
|
n = GetInt(a);
|
|
for (i = 0; i < n; ++i) {
|
|
Emit(a, 0);
|
|
}
|
|
}
|
|
}
|
|
|
|
static void OnFill(struct Assembler *a) {
|
|
int x;
|
|
long i, n;
|
|
n = GetInt(a);
|
|
if (IsInt(a, a->i)) {
|
|
x = GetInt(a);
|
|
} else {
|
|
x = 0;
|
|
}
|
|
for (i = 0; i < n; ++i) {
|
|
Emit(a, x);
|
|
}
|
|
}
|
|
|
|
static void OnByte(struct Assembler *a) {
|
|
while (IsInt(a, a->i)) {
|
|
Emit(a, GetInt(a));
|
|
}
|
|
}
|
|
|
|
static void OnWord(struct Assembler *a) {
|
|
long x;
|
|
while (IsInt(a, a->i)) {
|
|
x = GetInt(a);
|
|
Emit(a, x >> 000);
|
|
Emit(a, x >> 010);
|
|
}
|
|
}
|
|
|
|
static void OnLong(struct Assembler *a) {
|
|
while (IsInt(a, a->i)) {
|
|
EmitLong(a, GetInt(a));
|
|
}
|
|
}
|
|
|
|
static void OnQuad(struct Assembler *a) {
|
|
while (IsInt(a, a->i)) {
|
|
EmitQuad(a, GetInt(a));
|
|
}
|
|
}
|
|
|
|
static void OnFloat(struct Assembler *a) {
|
|
float f;
|
|
char b[4];
|
|
while (IsFloat(a, a->i)) {
|
|
f = GetFloat(a);
|
|
memcpy(b, &f, 4);
|
|
Emit(a, b[0]);
|
|
Emit(a, b[1]);
|
|
Emit(a, b[2]);
|
|
Emit(a, b[3]);
|
|
}
|
|
}
|
|
|
|
static void OnDouble(struct Assembler *a) {
|
|
double f;
|
|
char b[8];
|
|
while (IsFloat(a, a->i)) {
|
|
f = GetFloat(a);
|
|
memcpy(b, &f, 8);
|
|
Emit(a, b[0]);
|
|
Emit(a, b[1]);
|
|
Emit(a, b[2]);
|
|
Emit(a, b[3]);
|
|
Emit(a, b[4]);
|
|
Emit(a, b[5]);
|
|
Emit(a, b[6]);
|
|
Emit(a, b[7]);
|
|
}
|
|
}
|
|
|
|
static void OnAscii(struct Assembler *a) {
|
|
size_t i;
|
|
struct Slice arg;
|
|
while (IsSlice(a, a->i)) {
|
|
arg = GetString(a);
|
|
for (i = 0; i < arg.n; ++i) {
|
|
Emit(a, arg.p[i]);
|
|
}
|
|
}
|
|
}
|
|
|
|
static void OnAsciz(struct Assembler *a) {
|
|
size_t i;
|
|
struct Slice arg;
|
|
while (IsSlice(a, a->i)) {
|
|
arg = GetString(a);
|
|
for (i = 0; i < arg.n; ++i) {
|
|
Emit(a, arg.p[i]);
|
|
}
|
|
Emit(a, 0);
|
|
}
|
|
}
|
|
|
|
static void OnAbort(struct Assembler *a) {
|
|
Fail(a, "aborted");
|
|
}
|
|
|
|
static void OnErr(struct Assembler *a) {
|
|
if (g_ignore_err) return;
|
|
Fail(a, "error");
|
|
}
|
|
|
|
static void OnError(struct Assembler *a) {
|
|
struct Slice msg = GetString(a);
|
|
if (g_ignore_err) return;
|
|
Fail(a, "%.*s", msg.n, msg.p);
|
|
}
|
|
|
|
static void OnText(struct Assembler *a) {
|
|
SetSection(a, 0);
|
|
}
|
|
|
|
static void OnData(struct Assembler *a) {
|
|
SetSection(a, 1);
|
|
}
|
|
|
|
static void OnBss(struct Assembler *a) {
|
|
SetSection(a, 2);
|
|
}
|
|
|
|
static void OnPrevious(struct Assembler *a) {
|
|
SetSection(a, a->previous);
|
|
}
|
|
|
|
static void OnAlign(struct Assembler *a) {
|
|
long i, n, align, fill, maxskip;
|
|
align = GetInt(a);
|
|
if (popcnt(align) != 1) Fail(a, "alignment not power of 2");
|
|
fill = (a->sections.p[a->section].flags & SHF_EXECINSTR) ? 0x90 : 0;
|
|
maxskip = 268435456;
|
|
if (IsInt(a, a->i)) {
|
|
fill = GetInt(a);
|
|
if (IsInt(a, a->i)) {
|
|
maxskip = GetInt(a);
|
|
}
|
|
}
|
|
i = a->sections.p[a->section].binary.n;
|
|
n = ROUNDUP(i, align) - i;
|
|
if (n > maxskip) return;
|
|
a->sections.p[a->section].align = MAX(a->sections.p[a->section].align, align);
|
|
for (i = 0; i < n; ++i) {
|
|
Emit(a, fill);
|
|
}
|
|
}
|
|
|
|
static int SectionFlag(struct Assembler *a, int c) {
|
|
switch (c) {
|
|
case 'a':
|
|
return SHF_ALLOC;
|
|
case 'w':
|
|
return SHF_WRITE;
|
|
case 'x':
|
|
return SHF_EXECINSTR;
|
|
case 'g':
|
|
return SHF_GROUP;
|
|
case 'M':
|
|
return SHF_MERGE;
|
|
case 'S':
|
|
return SHF_STRINGS;
|
|
case 'T':
|
|
return SHF_TLS;
|
|
default:
|
|
Fail(a, "unknown section flag: %`'c", c);
|
|
}
|
|
}
|
|
|
|
static int SectionFlags(struct Assembler *a, struct Slice s) {
|
|
int i, flags;
|
|
for (flags = i = 0; i < s.n; ++i) {
|
|
flags |= SectionFlag(a, s.p[i]);
|
|
}
|
|
return flags;
|
|
}
|
|
|
|
static int SectionType(struct Assembler *a, struct Slice s) {
|
|
if (IS(s.p, s.n, "@progbits")) {
|
|
return SHT_PROGBITS;
|
|
} else if (IS(s.p, s.n, "@note")) {
|
|
return SHT_NOTE;
|
|
} else if (IS(s.p, s.n, "@nobits")) {
|
|
return SHT_NOBITS;
|
|
} else if (IS(s.p, s.n, "@preinit_array")) {
|
|
return SHT_PREINIT_ARRAY;
|
|
} else if (IS(s.p, s.n, "@init_array")) {
|
|
return SHT_INIT_ARRAY;
|
|
} else if (IS(s.p, s.n, "@fini_array")) {
|
|
return SHT_FINI_ARRAY;
|
|
} else {
|
|
Fail(a, "unknown section type: %.*s", s.n, s.p);
|
|
}
|
|
}
|
|
|
|
static void OnSection(struct Assembler *a) {
|
|
char *name;
|
|
int flags, type;
|
|
struct Slice arg;
|
|
arg = GetString(a);
|
|
name = strndup(arg.p, arg.n);
|
|
if (startswith(name, ".text")) {
|
|
flags = SHF_ALLOC | SHF_EXECINSTR;
|
|
type = SHT_PROGBITS;
|
|
} else if (startswith(name, ".data")) {
|
|
flags = SHF_ALLOC | SHF_WRITE;
|
|
type = SHT_PROGBITS;
|
|
} else if (startswith(name, ".bss")) {
|
|
flags = SHF_ALLOC | SHF_WRITE;
|
|
type = SHT_NOBITS;
|
|
} else {
|
|
flags = SHF_ALLOC | SHF_EXECINSTR | SHF_WRITE;
|
|
type = SHT_PROGBITS;
|
|
}
|
|
if (IsSlice(a, a->i)) {
|
|
flags = SectionFlags(a, GetString(a));
|
|
if (IsSlice(a, a->i)) {
|
|
type = SectionType(a, GetString(a));
|
|
}
|
|
}
|
|
SetSection(a, AppendSection(a, name, flags, type));
|
|
}
|
|
|
|
static void OnIncbin(struct Assembler *a) {
|
|
int fd;
|
|
char *path;
|
|
struct stat st;
|
|
struct Slice arg;
|
|
arg = GetString(a);
|
|
path = strndup(arg.p, arg.n);
|
|
if ((fd = open(path, O_RDONLY)) == -1 || fstat(fd, &st) == -1) {
|
|
Fail(a, "open failed: %s", path);
|
|
}
|
|
a->sections.p[a->section].binary.p =
|
|
realloc(a->sections.p[a->section].binary.p,
|
|
a->sections.p[a->section].binary.n + st.st_size);
|
|
if (read(fd, a->sections.p[a->section].binary.p, st.st_size) != st.st_size) {
|
|
Fail(a, "read failed: %s", path);
|
|
}
|
|
a->sections.p[a->section].binary.n += st.st_size;
|
|
close(fd);
|
|
free(path);
|
|
}
|
|
|
|
static int SymbolType(struct Assembler *a, struct Slice s) {
|
|
if (IS(s.p, s.n, "@object") || IS(s.p, s.n, "STT_OBJECT")) {
|
|
return STT_OBJECT;
|
|
} else if (IS(s.p, s.n, "@function") || IS(s.p, s.n, "STT_FUNC")) {
|
|
return STT_FUNC;
|
|
} else if (IS(s.p, s.n, "@common") || IS(s.p, s.n, "STT_COMMON")) {
|
|
return STT_COMMON;
|
|
} else if (IS(s.p, s.n, "@notype") || IS(s.p, s.n, "STT_NOTYPE")) {
|
|
return STT_NOTYPE;
|
|
} else if (IS(s.p, s.n, "@tls_object") || IS(s.p, s.n, "STT_TLS")) {
|
|
return STT_TLS;
|
|
} else {
|
|
Fail(a, "unknown symbol type: %.*s", s.n, s.p);
|
|
}
|
|
}
|
|
|
|
static void OnType(struct Assembler *a) {
|
|
int i;
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].type = SymbolType(a, GetString(a));
|
|
}
|
|
|
|
static void OnSize(struct Assembler *a) {
|
|
int i;
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].size = GetInt(a);
|
|
}
|
|
|
|
static void OnInternal(struct Assembler *a) {
|
|
int i;
|
|
while (IsSlice(a, a->i)) {
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].stv = STV_INTERNAL;
|
|
}
|
|
}
|
|
|
|
static void OnHidden(struct Assembler *a) {
|
|
int i;
|
|
while (IsSlice(a, a->i)) {
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].stv = STV_HIDDEN;
|
|
}
|
|
}
|
|
|
|
static void OnProtected(struct Assembler *a) {
|
|
int i;
|
|
while (IsSlice(a, a->i)) {
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].stv = STV_PROTECTED;
|
|
}
|
|
}
|
|
|
|
static void OnLocal(struct Assembler *a) {
|
|
int i;
|
|
while (IsSlice(a, a->i)) {
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].stb = STB_LOCAL;
|
|
}
|
|
}
|
|
|
|
static void OnWeak(struct Assembler *a) {
|
|
int i;
|
|
while (IsSlice(a, a->i)) {
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].stb = STB_WEAK;
|
|
}
|
|
}
|
|
|
|
static void OnGlobal(struct Assembler *a) {
|
|
int i;
|
|
while (IsSlice(a, a->i)) {
|
|
i = GetSymbol(a, a->things.p[a->i++].i);
|
|
a->symbols.p[i].stb = STB_GLOBAL;
|
|
}
|
|
}
|
|
|
|
static bool IsSizableOp(const char *op, struct Slice s) {
|
|
size_t n = strlen(op);
|
|
if (n == s.n) return !memcmp(op, s.p, n);
|
|
if (n + 1 == s.n && !memcmp(op, s.p, n)) {
|
|
switch (s.p[n]) {
|
|
case 'b':
|
|
case 'B':
|
|
case 'w':
|
|
case 'W':
|
|
case 'l':
|
|
case 'L':
|
|
case 'q':
|
|
case 'Q':
|
|
return true;
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
static int GetOpSize(struct Assembler *a, struct Slice s) {
|
|
switch (s.p[s.n - 1]) {
|
|
case 'b':
|
|
case 'B':
|
|
return 0;
|
|
case 'w':
|
|
case 'W':
|
|
return 1;
|
|
case 'l':
|
|
case 'L':
|
|
return 2;
|
|
case 'q':
|
|
case 'Q':
|
|
return 3;
|
|
default:
|
|
Fail(a, "could not size instruction");
|
|
}
|
|
}
|
|
|
|
static int CompareString8(const char a[8], const char b[8]) {
|
|
uint64_t x, y;
|
|
x = READ64BE(a);
|
|
y = READ64BE(b);
|
|
return x > y ? 1 : x < y ? -1 : 0;
|
|
}
|
|
|
|
static bool FindReg(const char *p, size_t n, struct Reg *out_reg) {
|
|
char key[8];
|
|
int i, c, m, l, r;
|
|
if (n && n <= 8) {
|
|
if (*p == '%') ++p, --n;
|
|
memset(key, 0, sizeof(key));
|
|
for (i = 0; i < n; ++i) {
|
|
key[i] = tolower(p[i]);
|
|
}
|
|
l = 0;
|
|
r = ARRAYLEN(kRegs) - 1;
|
|
while (l <= r) {
|
|
m = (l + r) >> 1;
|
|
c = CompareString8(kRegs[m].s, key);
|
|
if (c < 0) {
|
|
l = m + 1;
|
|
} else if (c > 0) {
|
|
r = m - 1;
|
|
} else {
|
|
*out_reg = kRegs[m];
|
|
return true;
|
|
}
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
static int RemoveRexw(int x) {
|
|
if (x == -1) return x;
|
|
x &= ~0x0800;
|
|
if (((x & 0xff00) >> 8) == REX) x &= ~0xff00;
|
|
return x;
|
|
}
|
|
|
|
static int FindRegReg(struct Slice s) {
|
|
struct Reg reg;
|
|
if (!FindReg(s.p, s.n, ®)) return -1;
|
|
return reg.reg;
|
|
}
|
|
|
|
static int FindRegRm(struct Slice s) {
|
|
struct Reg reg;
|
|
if (!FindReg(s.p, s.n, ®)) return -1;
|
|
return reg.rm;
|
|
}
|
|
|
|
static int FindRegBase(struct Slice s) {
|
|
struct Reg reg;
|
|
if (!FindReg(s.p, s.n, ®)) return -1;
|
|
return reg.base;
|
|
}
|
|
|
|
static int FindRegIndex(struct Slice s) {
|
|
struct Reg reg;
|
|
if (!FindReg(s.p, s.n, ®)) return -1;
|
|
return reg.index;
|
|
}
|
|
|
|
static wontreturn void InvalidRegister(struct Assembler *a) {
|
|
Fail(a, "invalid register");
|
|
}
|
|
|
|
static int GetRegisterRm(struct Assembler *a) {
|
|
int reg;
|
|
struct Slice wut;
|
|
if ((reg = FindRegRm(GetString(a))) == -1) InvalidRegister(a);
|
|
return reg;
|
|
}
|
|
|
|
static int ParseMemory(struct Assembler *a, long *disp) {
|
|
/* ┌isrip
|
|
│┌hasindex
|
|
││┌hasbase
|
|
│││┌hasasz
|
|
││││┌rex
|
|
│││││ ┌scale
|
|
│││││ │ ┌index
|
|
│││││ │ │ ┌base
|
|
││││├──────┐├┐├─┐├─┐
|
|
0b00000000000000000000000000000000*/
|
|
struct Slice str;
|
|
int reg, scale, mem = 0;
|
|
*disp = IsInt(a, a->i) ? GetInt(a) : 0;
|
|
if (IsPunct(a, a->i, '(')) {
|
|
++a->i;
|
|
if ((str = GetString(a)).n) {
|
|
mem |= HASBASE;
|
|
if (!strncasecmp(str.p, "%rip", str.n)) {
|
|
mem |= ISRIP;
|
|
} else {
|
|
reg = FindRegBase(str);
|
|
if (reg == -1) InvalidRegister(a);
|
|
mem |= reg & 007; // reg
|
|
mem |= reg & 0xff00; // rex
|
|
if (((reg & 070) >> 3) == 2) mem |= HASASZ; // asz
|
|
}
|
|
}
|
|
if (!IsPunct(a, a->i, ')')) {
|
|
mem |= HASINDEX;
|
|
reg = FindRegIndex(GetString(a));
|
|
if (reg == -1) InvalidRegister(a);
|
|
mem |= (reg & 007) << 3; // index
|
|
mem |= reg & 0xff00; // rex
|
|
if (((reg & 070) >> 3) == 2) mem |= HASASZ; // asz
|
|
if (!IsPunct(a, a->i, ')')) {
|
|
mem |= (bsr(GetInt(a)) & 3) << 6;
|
|
}
|
|
} else {
|
|
mem |= 4 << 3; // rsp index (hint: there is none)
|
|
}
|
|
ConsumePunct(a, ')');
|
|
}
|
|
return mem;
|
|
}
|
|
|
|
static void EncodeModrm(struct Assembler *a, int reg, int mem, long disp) {
|
|
reg <<= 3;
|
|
if (mem & (HASBASE | HASINDEX)) {
|
|
if (mem & ISRIP) {
|
|
Emit(a, 005 | reg);
|
|
} else {
|
|
Emit(a, 0204 | reg); // suboptimal
|
|
Emit(a, mem);
|
|
}
|
|
} else {
|
|
Emit(a, 004 | reg);
|
|
Emit(a, 045);
|
|
}
|
|
EmitLong(a, disp);
|
|
}
|
|
|
|
static void OnMov(struct Assembler *a, struct Slice op) {
|
|
int reg, mem;
|
|
long imm, disp;
|
|
if (IsPunct(a, a->i, '$')) {
|
|
++a->i;
|
|
imm = GetInt(a);
|
|
if (IsSlice(a, a->i)) {
|
|
reg = GetRegisterRm(a);
|
|
if (reg & 0xff00) {
|
|
Emit(a, reg >> 8);
|
|
}
|
|
switch ((reg & 070) >> 3) {
|
|
case 0:
|
|
Emit(a, 0xb0 + (reg & 7));
|
|
Emit(a, imm);
|
|
break;
|
|
case 1:
|
|
Emit(a, OSZ);
|
|
Emit(a, 0xb8 + (reg & 7));
|
|
EmitWord(a, imm);
|
|
break;
|
|
case 2:
|
|
Emit(a, 0xb8 + (reg & 7));
|
|
EmitLong(a, imm);
|
|
break;
|
|
case 3:
|
|
Emit(a, 0xb8 + (reg & 7)); // suboptimal
|
|
EmitQuad(a, imm);
|
|
break;
|
|
default:
|
|
Fail(a, "todo movd/movq");
|
|
}
|
|
} else {
|
|
mem = ParseMemory(a, &disp);
|
|
if (mem & 0xff00) {
|
|
Emit(a, mem >> 8);
|
|
}
|
|
switch (GetOpSize(a, op)) {
|
|
case 0:
|
|
Emit(a, 0xc6);
|
|
EncodeModrm(a, 0, mem, disp);
|
|
Emit(a, imm);
|
|
break;
|
|
case 1:
|
|
Emit(a, OSZ);
|
|
Emit(a, 0xc7);
|
|
EncodeModrm(a, 0, mem, disp);
|
|
EmitWord(a, imm);
|
|
break;
|
|
case 2:
|
|
Emit(a, 0xc7);
|
|
EncodeModrm(a, 0, mem, disp);
|
|
EmitLong(a, imm);
|
|
break;
|
|
case 3:
|
|
Emit(a, 0xc7); // suboptimal
|
|
EncodeModrm(a, 0, mem, disp);
|
|
EmitQuad(a, imm);
|
|
break;
|
|
default:
|
|
unreachable;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
static void OnPush(struct Assembler *a) {
|
|
long x;
|
|
int reg;
|
|
if (IsPunct(a, a->i, '$')) {
|
|
++a->i;
|
|
x = GetInt(a);
|
|
if (-128 <= x && x <= 127) {
|
|
Emit(a, 0x6A);
|
|
Emit(a, x);
|
|
} else {
|
|
Emit(a, 0x68);
|
|
EmitLong(a, x);
|
|
}
|
|
} else {
|
|
reg = RemoveRexw(GetRegisterRm(a));
|
|
if (reg & 0xff00) Emit(a, reg >> 8);
|
|
if (((reg & 070) >> 3) == 1) Emit(a, OSZ);
|
|
Emit(a, 0x50 + (reg & 7));
|
|
}
|
|
}
|
|
|
|
static void OnPop(struct Assembler *a) {
|
|
int reg;
|
|
reg = RemoveRexw(GetRegisterRm(a));
|
|
if (reg & 0xff00) Emit(a, reg >> 8);
|
|
if (((reg & 070) >> 3) == 1) Emit(a, OSZ);
|
|
Emit(a, 0x58 + (reg & 7));
|
|
}
|
|
|
|
static void OnRet(struct Assembler *a) {
|
|
if (IsPunct(a, a->i, '$')) {
|
|
++a->i;
|
|
Emit(a, 0xC2);
|
|
EmitWord(a, GetInt(a));
|
|
} else {
|
|
Emit(a, 0xC3);
|
|
}
|
|
}
|
|
|
|
static void OnLeave(struct Assembler *a) {
|
|
Emit(a, 0xC9);
|
|
}
|
|
static void OnHlt(struct Assembler *a) {
|
|
Emit(a, 0xF4);
|
|
}
|
|
static void OnCmc(struct Assembler *a) {
|
|
Emit(a, 0xF5);
|
|
}
|
|
static void OnClc(struct Assembler *a) {
|
|
Emit(a, 0xF8);
|
|
}
|
|
static void OnStc(struct Assembler *a) {
|
|
Emit(a, 0xF9);
|
|
}
|
|
static void OnCli(struct Assembler *a) {
|
|
Emit(a, 0xFA);
|
|
}
|
|
static void OnSti(struct Assembler *a) {
|
|
Emit(a, 0xFB);
|
|
}
|
|
static void OnCld(struct Assembler *a) {
|
|
Emit(a, 0xFC);
|
|
}
|
|
static void OnStd(struct Assembler *a) {
|
|
Emit(a, 0xFD);
|
|
}
|
|
static void OnLodsb(struct Assembler *a) {
|
|
Emit(a, 0xAC);
|
|
}
|
|
static void OnLodsw(struct Assembler *a) {
|
|
Emit(a, OSZ);
|
|
Emit(a, 0xAD);
|
|
}
|
|
static void OnLodsl(struct Assembler *a) {
|
|
Emit(a, 0xAD);
|
|
}
|
|
static void OnLodsq(struct Assembler *a) {
|
|
Emit(a, REXW);
|
|
Emit(a, 0xAD);
|
|
}
|
|
static void OnStosb(struct Assembler *a) {
|
|
Emit(a, 0xAA);
|
|
}
|
|
static void OnStosw(struct Assembler *a) {
|
|
Emit(a, OSZ);
|
|
Emit(a, 0xAB);
|
|
}
|
|
static void OnStosl(struct Assembler *a) {
|
|
Emit(a, 0xAB);
|
|
}
|
|
static void OnStosq(struct Assembler *a) {
|
|
Emit(a, REXW);
|
|
Emit(a, 0xAB);
|
|
}
|
|
static void OnMovsb(struct Assembler *a) {
|
|
Emit(a, 0xA4);
|
|
}
|
|
static void OnMovsw(struct Assembler *a) {
|
|
Emit(a, OSZ);
|
|
Emit(a, 0xA5);
|
|
}
|
|
static void OnMovsl(struct Assembler *a) {
|
|
Emit(a, 0xA5);
|
|
}
|
|
static void OnMovsq(struct Assembler *a) {
|
|
Emit(a, REXW);
|
|
Emit(a, 0xA5);
|
|
}
|
|
|
|
static bool Prefix(struct Assembler *a, const char *s, size_t n) {
|
|
int i;
|
|
char key[8];
|
|
if (n <= 8) {
|
|
memset(key, 0, 8);
|
|
for (i = 0; i < n; ++i) key[i] = tolower(s[i]);
|
|
for (i = 0; i < sizeof(kPrefix) / sizeof(*kPrefix); i++) {
|
|
if (!memcmp(key, kPrefix[i], 8)) {
|
|
Emit(a, kPrefixByte[i]);
|
|
return true;
|
|
}
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
static void Directive(struct Assembler *a) {
|
|
struct Slice s;
|
|
for (;;) {
|
|
s = GetString(a);
|
|
if (!Prefix(a, s.p, s.n)) break;
|
|
}
|
|
if (s.n >= 1 && s.p[0] == '.') {
|
|
if (IS(s.p, s.n, ".zero")) {
|
|
OnZero(a);
|
|
} else if (IS(s.p, s.n, ".align") || IS(s.p, s.n, ".balign")) {
|
|
OnAlign(a);
|
|
} else if (IS(s.p, s.n, ".byte")) {
|
|
OnByte(a);
|
|
} else if (IS(s.p, s.n, ".word") || IS(s.p, s.n, ".short")) {
|
|
OnWord(a);
|
|
} else if (IS(s.p, s.n, ".long")) {
|
|
OnLong(a);
|
|
} else if (IS(s.p, s.n, ".quad")) {
|
|
OnQuad(a);
|
|
} else if (IS(s.p, s.n, ".float")) {
|
|
OnFloat(a);
|
|
} else if (IS(s.p, s.n, ".double")) {
|
|
OnDouble(a);
|
|
} else if (IS(s.p, s.n, ".ascii")) {
|
|
OnAscii(a);
|
|
} else if (IS(s.p, s.n, ".asciz")) {
|
|
OnAsciz(a);
|
|
} else if (IS(s.p, s.n, ".text")) {
|
|
OnText(a);
|
|
} else if (IS(s.p, s.n, ".data")) {
|
|
OnData(a);
|
|
} else if (IS(s.p, s.n, ".bss")) {
|
|
OnBss(a);
|
|
} else if (IS(s.p, s.n, ".previous")) {
|
|
OnPrevious(a);
|
|
} else if (IS(s.p, s.n, ".section")) {
|
|
OnSection(a);
|
|
} else if (IS(s.p, s.n, ".abort")) {
|
|
OnAbort(a);
|
|
} else if (IS(s.p, s.n, ".err")) {
|
|
OnErr(a);
|
|
} else if (IS(s.p, s.n, ".error")) {
|
|
OnError(a);
|
|
} else if (IS(s.p, s.n, ".fill") || IS(s.p, s.n, ".space")) {
|
|
OnFill(a);
|
|
} else if (IS(s.p, s.n, ".type")) {
|
|
OnType(a);
|
|
} else if (IS(s.p, s.n, ".size")) {
|
|
OnSize(a);
|
|
} else if (IS(s.p, s.n, ".local")) {
|
|
OnLocal(a);
|
|
} else if (IS(s.p, s.n, ".internal")) {
|
|
OnInternal(a);
|
|
} else if (IS(s.p, s.n, ".weak")) {
|
|
OnWeak(a);
|
|
} else if (IS(s.p, s.n, ".hidden")) {
|
|
OnHidden(a);
|
|
} else if (IS(s.p, s.n, ".globl") || IS(s.p, s.n, ".global")) {
|
|
OnGlobal(a);
|
|
} else if (IS(s.p, s.n, ".protected")) {
|
|
OnProtected(a);
|
|
} else if (IS(s.p, s.n, ".incbin")) {
|
|
OnIncbin(a);
|
|
} else {
|
|
Fail(a, "unexpected directive: %.*s", s.n, s.p);
|
|
}
|
|
} else if (IS(s.p, s.n, "ret")) {
|
|
OnRet(a);
|
|
} else if (IS(s.p, s.n, "leave")) {
|
|
OnLeave(a);
|
|
} else if (IS(s.p, s.n, "push")) {
|
|
OnPush(a);
|
|
} else if (IS(s.p, s.n, "pop")) {
|
|
OnPop(a);
|
|
} else if (IS(s.p, s.n, "hlt")) {
|
|
OnHlt(a);
|
|
} else if (IS(s.p, s.n, "cmc")) {
|
|
OnCmc(a);
|
|
} else if (IS(s.p, s.n, "clc")) {
|
|
OnClc(a);
|
|
} else if (IS(s.p, s.n, "stc")) {
|
|
OnStc(a);
|
|
} else if (IS(s.p, s.n, "cli")) {
|
|
OnCli(a);
|
|
} else if (IS(s.p, s.n, "sti")) {
|
|
OnSti(a);
|
|
} else if (IS(s.p, s.n, "cld")) {
|
|
OnCld(a);
|
|
} else if (IS(s.p, s.n, "std")) {
|
|
OnStd(a);
|
|
} else if (IS(s.p, s.n, "stosb")) {
|
|
OnStosb(a);
|
|
} else if (IS(s.p, s.n, "stosw")) {
|
|
OnStosw(a);
|
|
} else if (IS(s.p, s.n, "stosl")) {
|
|
OnStosl(a);
|
|
} else if (IS(s.p, s.n, "stosq")) {
|
|
OnStosq(a);
|
|
} else if (IS(s.p, s.n, "lodsb")) {
|
|
OnLodsb(a);
|
|
} else if (IS(s.p, s.n, "lodsw")) {
|
|
OnLodsw(a);
|
|
} else if (IS(s.p, s.n, "lodsl")) {
|
|
OnLodsl(a);
|
|
} else if (IS(s.p, s.n, "lodsq")) {
|
|
OnLodsq(a);
|
|
} else if (IS(s.p, s.n, "movsb")) {
|
|
OnMovsb(a);
|
|
} else if (IS(s.p, s.n, "movsw")) {
|
|
OnMovsw(a);
|
|
} else if (IS(s.p, s.n, "movsl")) {
|
|
OnMovsl(a);
|
|
} else if (IS(s.p, s.n, "movsq")) {
|
|
OnMovsq(a);
|
|
} else if (IsSizableOp("mov", s)) {
|
|
OnMov(a, s);
|
|
} else {
|
|
Fail(a, "unexpected op: %.*s", s.n, s.p);
|
|
}
|
|
ConsumePunct(a, ';');
|
|
}
|
|
|
|
static void Assemble(struct Assembler *a) {
|
|
while (a->i < a->things.n) {
|
|
if (IsPunct(a, a->i, ';')) {
|
|
++a->i;
|
|
continue;
|
|
}
|
|
switch (a->things.p[a->i].t) {
|
|
case TT_SLICE:
|
|
if (IsPunct(a, a->i + 1, ':')) {
|
|
Label(a, a->things.p[a->i].i);
|
|
} else {
|
|
Directive(a);
|
|
}
|
|
break;
|
|
case TT_INT:
|
|
if (IsPunct(a, a->i + 1, ':')) {
|
|
LocalLabel(a, a->ints.p[a->things.p[a->i].i]);
|
|
}
|
|
// fallthrough
|
|
default:
|
|
Fail(a, "unexpected token");
|
|
}
|
|
}
|
|
}
|
|
|
|
static void Objectify(struct Assembler *a, const char *path) {
|
|
size_t i, j;
|
|
struct ElfWriter *elf;
|
|
elf = elfwriter_open(path, 0644);
|
|
for (i = 0; i < a->sections.n; ++i) {
|
|
elfwriter_align(elf, a->sections.p[i].align, 0);
|
|
elfwriter_startsection(elf, a->sections.p[i].name, a->sections.p[i].type,
|
|
a->sections.p[i].flags);
|
|
for (j = 0; j < a->symbols.n; ++j) {
|
|
if (a->symbols.p[j].section != i) continue;
|
|
elfwriter_appendsym(
|
|
elf,
|
|
strndup(a->slices.p[a->symbols.p[j].name].p,
|
|
a->slices.p[a->symbols.p[j].name].n),
|
|
ELF64_ST_INFO(a->symbols.p[j].stb, a->symbols.p[j].type),
|
|
a->symbols.p[j].stv, a->symbols.p[j].location, a->symbols.p[j].size);
|
|
}
|
|
memcpy(elfwriter_reserve(elf, a->sections.p[i].binary.n),
|
|
a->sections.p[i].binary.p, a->sections.p[i].binary.n);
|
|
elfwriter_commit(elf, a->sections.p[i].binary.n);
|
|
elfwriter_finishsection(elf);
|
|
}
|
|
elfwriter_close(elf);
|
|
}
|
|
|
|
static void PrintThings(struct Assembler *a) {
|
|
int i;
|
|
char fbuf[32];
|
|
for (i = 0; i < a->things.n; ++i) {
|
|
printf("%s:%d:: ", a->sauces.p[a->things.p[i].s].path,
|
|
a->sauces.p[a->things.p[i].s].line);
|
|
switch (a->things.p[i].t) {
|
|
case TT_INT:
|
|
printf("TT_INT %ld\n", a->ints.p[a->things.p[i].i]);
|
|
break;
|
|
case TT_FLOAT:
|
|
g_xfmt_p(fbuf, &a->floats.p[a->things.p[i].i], 19, sizeof(fbuf), 0);
|
|
printf("TT_FLOAT %s\n", fbuf);
|
|
break;
|
|
case TT_SLICE:
|
|
printf("TT_SLICE %`'.*s\n", a->slices.p[a->things.p[i].i].n,
|
|
a->slices.p[a->things.p[i].i].p);
|
|
break;
|
|
case TT_PUNCT:
|
|
printf("TT_PUNCT %`'c\n", a->things.p[i].i);
|
|
break;
|
|
default:
|
|
unreachable;
|
|
}
|
|
}
|
|
}
|
|
|
|
void Assembler(int argc, char *argv[]) {
|
|
struct Assembler *a;
|
|
showcrashreports();
|
|
if (argc == 1) {
|
|
system("o//third_party/chibicc/as.com -o /tmp/o third_party/chibicc/hog.s");
|
|
system("objdump -wxd /tmp/o");
|
|
exit(0);
|
|
}
|
|
ReadFlags(argc, argv);
|
|
a = NewAssembler();
|
|
Tokenize(a, g_input_path);
|
|
/* PrintThings(a); */
|
|
Assemble(a);
|
|
Objectify(a, g_output_path);
|
|
}
|
|
|
|
int main(int argc, char *argv[]) {
|
|
Assembler(argc, argv);
|
|
return 0;
|
|
}
|