T0b1/test-repo
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Output more available information in formatter

Browse Source 
The formatter now includes the following information:

- Segment overrides
- Address-size overrides
- REP/REPNZ prefixes
- LOCK prefix
- High-byte registers (determined using presence of REX prefix)
This commit is contained in:
Alexis Engelke
2019-01-13 20:38:15 +01:00
parent 2d1a9582b8
commit 8fdfe53822
9 changed files with 121 additions and 180 deletions
Download Patch File Download Diff File Expand all files Collapse all files

195
format.c
View File

@@ -1,6 +1,7 @@
#include <stddef.h>
#include <stdint.h>
#include <stdio.h>
#include <decode.h>
@@ -17,152 +18,90 @@ static const uint16_t _mnemonic_offs[] = {
};
#undef DECODE_TABLE_STRTAB2
static
void
instr_format_decimal(char** cur, uint32_t value)
{
char buffer[32];
size_t buf_idx = sizeof(buffer) - 1;
if (value == 0)
{
buffer[buf_idx] = '0';
}
else
{
while (value > 0)
{
uint32_t digit = value % 10;
buffer[buf_idx--] = '0' + digit;
value /= 10;
}
buf_idx++;
}
size_t length = sizeof(buffer) - buf_idx;
__builtin_memcpy(*cur, buffer + buf_idx, length);
*cur += length;
}
static
void
instr_format_hex(char** cur, size_t value)
{
char buffer[32];
size_t buf_idx = sizeof(buffer) - 1;
if (value == 0)
{
buffer[buf_idx] = '0';
}
else
{
while (value > 0)
{
uint32_t nibble = value & 0xf;
buffer[buf_idx--] = "0123456789abcdef"[nibble];
value >>= 4;
}
buf_idx++;
}
buffer[--buf_idx] = 'x';
buffer[--buf_idx] = '0';
size_t length = sizeof(buffer) - buf_idx;
__builtin_memcpy(*cur, buffer + buf_idx, length);
*cur += length;
}
#define FMT_CONCAT(buf, end, ...) do { \
buf += snprintf(buf, end - buf, __VA_ARGS__); \
if (buf > end) \
buf = end; \
} while (0)
void
instr_format(const Instr* instr, char buffer[128])
{
char* cur = buffer;
*(cur++) = '[';
char* buf = buffer;
char* end = buffer + 128;
const char* mnemonic = &_mnemonic_str[_mnemonic_offs[instr->type]];
while (*mnemonic)
{
*(cur++) = *(mnemonic++);
}
FMT_CONCAT(buf, end, "[");
if (INSTR_HAS_REP(instr))
FMT_CONCAT(buf, end, "rep:");
if (INSTR_HAS_REPNZ(instr))
FMT_CONCAT(buf, end, "repnz:");
if (INSTR_SEGMENT(instr) < 6)
FMT_CONCAT(buf, end, "%cs:", "ecsdfg"[INSTR_SEGMENT(instr)]);
if (INSTR_IS64(instr) && INSTR_ADDRSZ(instr) == 4)
FMT_CONCAT(buf, end, "addr32:");
if (!INSTR_IS64(instr) && INSTR_ADDRSZ(instr) == 2)
FMT_CONCAT(buf, end, "addr16:");
if (INSTR_HAS_LOCK(instr))
FMT_CONCAT(buf, end, "lock:");
FMT_CONCAT(buf, end, "%s", &_mnemonic_str[_mnemonic_offs[instr->type]]);
if (INSTR_WIDTH(instr))
{
*(cur++) = '_';
instr_format_decimal(&cur, INSTR_WIDTH(instr));
}
FMT_CONCAT(buf, end, "_%u", INSTR_WIDTH(instr));
for (int i = 0; i < 4; i++)
{
const struct Operand* operand = &instr->operands[i];
if (operand->type == OT_NONE)
{
break;
}
__builtin_memcpy(cur, " REG IMM MEM" + operand->type * 4 - 4, 4);
cur += 4;
instr_format_decimal(&cur, operand->size);
*(cur++) = ':';
const char* op_type_name = "reg\0imm\0mem" + operand->type * 4 - 4;
FMT_CONCAT(buf, end, " %s%u:", op_type_name, operand->size);
switch (operand->type)
{
size_t immediate;
case OT_REG:
instr_format_decimal(&cur, reg_index(operand->reg));
break;
case OT_IMM:
immediate = instr->immediate;
if (operand->size == 1)
{
immediate &= 0xff;
}
else if (operand->size == 2)
{
immediate &= 0xffff;
}
#if defined(ARCH_X86_64)
else if (operand->size == 4)
{
immediate &= 0xffffffff;
}
#endif
instr_format_hex(&cur, immediate);
break;
case OT_MEM:
if (!reg_is_none(operand->reg))
{
instr_format_decimal(&cur, reg_index(operand->reg));
*(cur++) = ':';
}
if (instr->scale != 0)
{
uint8_t scale = 1 << (instr->scale - 1);
instr_format_decimal(&cur, scale);
*(cur++) = '*';
instr_format_decimal(&cur, reg_index(instr->sreg));
*(cur++) = ':';
}
if (instr->disp < 0)
{
*(cur++) = '-';
instr_format_hex(&cur, -instr->disp);
}
else
{
instr_format_hex(&cur, instr->disp);
}
break;
case OT_NONE:
default:
break;
size_t immediate;
case OT_REG:
if (operand->size == 1 && !INSTR_HAS_REX(instr) &&
operand->reg >= 4 && operand->reg < 8)
FMT_CONCAT(buf, end, "r%uh", operand->reg - 4);
else
FMT_CONCAT(buf, end, "r%u", operand->reg);
break;
case OT_IMM:
immediate = instr->immediate;
if (operand->size == 1)
immediate &= 0xff;
else if (operand->size == 2)
immediate &= 0xffff;
else if (operand->size == 4)
immediate &= 0xffffffff;
FMT_CONCAT(buf, end, "0x%lx", immediate);
break;
case OT_MEM:
if (!reg_is_none(operand->reg))
{
FMT_CONCAT(buf, end, "r%u", operand->reg);
if (instr->scale != 0 || instr->disp > 0)
FMT_CONCAT(buf, end, "+");
}
if (instr->scale != 0)
{
FMT_CONCAT(buf, end, "%u*r%u", 1 << (instr->scale - 1),
instr->sreg);
if (instr->disp > 0)
FMT_CONCAT(buf, end, "+");
}
if (instr->disp < 0)
FMT_CONCAT(buf, end, "-0x%lx", -instr->disp);
else if ((reg_is_none(operand->reg) && instr->scale == 0) ||
instr->disp > 0)
FMT_CONCAT(buf, end, "0x%lx", instr->disp);
break;
case OT_NONE:
default:
break;
}
}
*(cur++) = ']';
*(cur++) = '\0';
#ifndef NDEBUG
if (cur - buffer > 128)
{
__builtin_trap();
}
#endif
FMT_CONCAT(buf, end, "]");
}

20
tests/benchmarks.txt
View File

@@ -1,13 +1,13 @@
decode 90 [NOP]
decode 0fcd [BSWAP REG4:5]
decode 660fcd [BSWAP REG2:5]
decode 6650 [PUSH REG2:0]
decode 0fcd [BSWAP reg4:r5]
decode 660fcd [BSWAP reg2:r5]
decode 6650 [PUSH reg2:r0]
decode a5 [MOVS_4]
decode f3a5 [MOVS_4]
decode f3a5 [rep:MOVS_4]
decode 66a5 [MOVS_2]
decode f366a5 [MOVS_2]
decode f7d7 [NOT REG4:7]
decode f717 [NOT MEM4:7:0x0]
decode f7142f [NOT MEM4:7:1*5:0x0]
decode f7542f12 [NOT MEM4:7:1*5:0x12]
decode f7942f34120000 [NOT MEM4:7:1*5:0x1234]
decode f366a5 [rep:MOVS_2]
decode f7d7 [NOT reg4:r7]
decode f717 [NOT mem4:r7]
decode f7142f [NOT mem4:r7+1*r5]
decode f7542f12 [NOT mem4:r7+1*r5+0x12]
decode f7942f34120000 [NOT mem4:r7+1*r5+0x1234]

18
tests/decode-enter.sh
View File

@@ -1,9 +1,9 @@
decode 66c8000000 "[ENTER_2 IMM4:0x0]"
decode 66c8000f00 "[ENTER_2 IMM4:0xf00]"
decode 66c8000001 "[ENTER_2 IMM4:0x10000]"
decode32 c8000000 "[ENTER_4 IMM4:0x0]"
decode32 c8000f00 "[ENTER_4 IMM4:0xf00]"
decode32 c8000001 "[ENTER_4 IMM4:0x10000]"
decode64 c8000000 "[ENTER_8 IMM4:0x0]"
decode64 c8000f00 "[ENTER_8 IMM4:0xf00]"
decode64 c8000001 "[ENTER_8 IMM4:0x10000]"
decode 66c8000000 [ENTER_2 imm4:0x0]
decode 66c8000f00 [ENTER_2 imm4:0xf00]
decode 66c8000001 [ENTER_2 imm4:0x10000]
decode32 c8000000 [ENTER_4 imm4:0x0]
decode32 c8000f00 [ENTER_4 imm4:0xf00]
decode32 c8000001 [ENTER_4 imm4:0x10000]
decode64 c8000000 [ENTER_8 imm4:0x0]
decode64 c8000f00 [ENTER_8 imm4:0xf00]
decode64 c8000001 [ENTER_8 imm4:0x10000]

4
tests/decode-imul.sh
View File

@@ -1,2 +1,2 @@
decode 69C708010000 "[IMUL3 REG4:0 REG4:7 IMM4:0x108]"
decode 6BC708 "[IMUL3 REG4:0 REG4:7 IMM4:0x8]"
decode 69C708010000 [IMUL3 reg4:r0 reg4:r7 imm4:0x108]
decode 6BC708 [IMUL3 reg4:r0 reg4:r7 imm4:0x8]

26
tests/decode-inc.sh
View File

@@ -1,12 +1,14 @@
decode32 40 "[INC REG4:0]"
decode32 43 "[INC REG4:3]"
decode32 6647 "[INC REG2:7]"
decode fec0 "[INC REG1:0]"
decode fec4 "[INC REG1:4]"
decode ffc0 "[INC REG4:0]"
decode ffc4 "[INC REG4:4]"
decode 66ffc0 "[INC REG2:0]"
decode 66ffc4 "[INC REG2:4]"
decode64 48ffc0 "[INC REG8:0]"
decode64 48ffc4 "[INC REG8:4]"
decode64 49ffc7 "[INC REG8:15]"
decode32 40 [INC reg4:r0]
decode32 43 [INC reg4:r3]
decode32 6647 [INC reg2:r7]
decode fec0 [INC reg1:r0]
decode fec4 [INC reg1:r0h]
decode ffc0 [INC reg4:r0]
decode ffc4 [INC reg4:r4]
decode ff00 [INC mem4:r0]
decode f0ff00 [lock:INC mem4:r0]
decode 66ffc0 [INC reg2:r0]
decode 66ffc4 [INC reg2:r4]
decode64 48ffc0 [INC reg8:r0]
decode64 48ffc4 [INC reg8:r4]
decode64 49ffc7 [INC reg8:r15]

10
tests/decode-movsx.sh
View File

@@ -1,5 +1,5 @@
decode 660fbec2 "[MOVSX REG2:0 REG1:2]"
decode 0fbec2 "[MOVSX REG4:0 REG1:2]"
decode 0fbfc2 "[MOVSX REG4:0 REG2:2]"
decode64 480fbfc2 "[MOVSX REG8:0 REG2:2]"
decode64 4863c2 "[MOVSX REG8:0 REG4:2]"
decode 660fbec2 [MOVSX reg2:r0 reg1:r2]
decode 0fbec2 [MOVSX reg4:r0 reg1:r2]
decode 0fbfc2 [MOVSX reg4:r0 reg2:r2]
decode64 480fbfc2 [MOVSX reg8:r0 reg2:r2]
decode64 4863c2 [MOVSX reg8:r0 reg4:r2]

24
tests/decode-ret.sh
View File

@@ -1,12 +1,12 @@
decode 66c3 "[RET_2]"
decode 66c20000 "[RET_IMM_2 IMM2:0x0]"
decode 66c20d00 "[RET_IMM_2 IMM2:0xd]"
decode 66c20dff "[RET_IMM_2 IMM2:0xff0d]"
decode32 c3 "[RET_4]"
decode32 c20000 "[RET_IMM_4 IMM2:0x0]"
decode32 c20d00 "[RET_IMM_4 IMM2:0xd]"
decode32 c20dff "[RET_IMM_4 IMM2:0xff0d]"
decode64 c3 "[RET_8]"
decode64 c20000 "[RET_IMM_8 IMM2:0x0]"
decode64 c20d00 "[RET_IMM_8 IMM2:0xd]"
decode64 c20dff "[RET_IMM_8 IMM2:0xff0d]"
decode 66c3 [RET_2]
decode 66c20000 [RET_IMM_2 imm2:0x0]
decode 66c20d00 [RET_IMM_2 imm2:0xd]
decode 66c20dff [RET_IMM_2 imm2:0xff0d]
decode32 c3 [RET_4]
decode32 c20000 [RET_IMM_4 imm2:0x0]
decode32 c20d00 [RET_IMM_4 imm2:0xd]
decode32 c20dff [RET_IMM_4 imm2:0xff0d]
decode64 c3 [RET_8]
decode64 c20000 [RET_IMM_8 imm2:0x0]
decode64 c20d00 [RET_IMM_8 imm2:0xd]
decode64 c20dff [RET_IMM_8 imm2:0xff0d]

2
tests/decode-sse-movq.sh
View File

@@ -1 +1 @@
decode f30f7e5c2408 "[SSE_MOVQ_X2X REG8:3 MEM8:4:0x8]"
decode f30f7e5c2408 [SSE_MOVQ_X2X reg8:r3 mem8:r4+0x8]

2
tests/decode-sse-shufpd.sh
View File

@@ -1 +1 @@
decode 660fc6c001 "[SSE_SHUFPD REG16:0 REG16:0 IMM1:0x1]"
decode 660fc6c001 [SSE_SHUFPD reg16:r0 reg16:r0 imm1:0x1]