diff options
| -rw-r--r-- | README.md | 4 | ||||
| -rw-r--r-- | asm.peg | 213 | ||||
| -rw-r--r-- | test/test.sh | 14 |
3 files changed, 141 insertions, 90 deletions
@@ -43,11 +43,11 @@ Bonus features: - [ ] A man page explaining what is supported. - [x] Two pass jump relaxing. -- [ ] Immediate relaxing. +- [x] Immediate relaxing. - [ ] Simple immediate expressions. - [ ] Assemble a libc. - [ ] Test every opcode with all variants in our test suite. -- [ ] Parser that doesn't depend on peg/leg. +- [ ] Parser that doesn't depend on peg/leg or an optimized peg that checks common prefixes. # Notes @@ -353,20 +353,23 @@ movzx = "movz" ( ) add = "add" ( - 'b'? ws s:imm8 ws? ',' ws? d:al { $$ = IMM({.w=0}, -1, 0x04, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:ax { $$ = IMM({.w=0}, 0x66, 0x05, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:eax { $$ = IMM({.w=0}, -1, 0x05, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:rax { $$ = IMM({.w=1}, -1, 0x05, s, d); } - - | 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = IMMREG({.w=0}, -1, 0x80, 0x00, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:r16 { $$ = IMMREG({.w=0}, 0x66, 0x81, 0x00, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:r32 { $$ = IMMREG({.w=0}, -1, 0x81, 0x00, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:r64 { $$ = IMMREG({.w=1}, -1, 0x81, 0x00, s, d); } - - | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x00, s, d); } - | 'w' ws s:imm16 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, 0x81, 0x00, s, d); } - | 'l' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x81, 0x00, s, d); } - | 'q' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, 0x81, 0x00, s, d); } + 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = (d.kind == ASM_AL) ? + IMM ({.w=0}, -1, 0x04, s, d) : + IMMREG({.w=0}, -1, 0x80, 0x00, s, d); } + | 'w'? ws s:imm16-8 ws? ',' ws? d:r16 { $$ = (d.kind == ASM_AX && s.imm.nbytes == 2) ? + IMM ({.w=0}, 0x66, 0x05, s, d) : + IMMREG({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x00, s, d); } + | 'l'? ws s:imm32-8 ws? ',' ws? d:r32 { $$ = (d.kind == ASM_EAX && s.imm.nbytes == 4) ? + IMM ({.w=0}, -1, 0x05, s, d) : + IMMREG({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x00, s, d); } + | 'q'? ws s:imm32-8 ws? ',' ws? d:r64 { $$ = (d.kind == ASM_RAX && s.imm.nbytes == 4) ? + IMM ({.w=1}, -1, 0x05, s, d) : + IMMREG({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x00, s, d); } + + | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x00, s, d); } + | 'w' ws s:imm16-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x00, s, d); } + | 'l' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x00, s, d); } + | 'q' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x00, s, d); } | 'b'? ws s:r8 ws? ',' ws? d:r8 { $$ = REGREG({.w=0}, -1, 0x00, s, d); } | 'w'? ws s:r16 ws? ',' ws? d:r16 { $$ = REGREG({.w=0}, 0x66, 0x01, s, d); } @@ -384,22 +387,24 @@ add = "add" ( | 'q'? ws s:mem ws? ',' ws? d:r64 { $$ = MEMREG({.w=1}, -1, 0x03, s, d); } ) - and = "and" ( - 'b'? ws s:imm8 ws? ',' ws? d:al { $$ = IMM({.w=0}, -1, 0x24, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:ax { $$ = IMM({.w=0}, 0x66, 0x25, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:eax { $$ = IMM({.w=0}, -1, 0x25, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:rax { $$ = IMM({.w=1}, -1, 0x25, s, d); } - - | 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = IMMREG({.w=0}, -1, 0x80, 0x04, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:r16 { $$ = IMMREG({.w=0}, 0x66, 0x81, 0x04, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:r32 { $$ = IMMREG({.w=0}, -1, 0x81, 0x04, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:r64 { $$ = IMMREG({.w=1}, -1, 0x81, 0x04, s, d); } - - | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x04, s, d); } - | 'w' ws s:imm16 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, 0x81, 0x04, s, d); } - | 'l' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x81, 0x04, s, d); } - | 'q' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, 0x81, 0x04, s, d); } + 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = (d.kind == ASM_AL) ? + IMM ({.w=0}, -1, 0x24, s, d) : + IMMREG({.w=0}, -1, 0x80, 0x04, s, d); } + | 'w'? ws s:imm16-8 ws? ',' ws? d:r16 { $$ = (d.kind == ASM_AX && s.imm.nbytes == 2) ? + IMM ({.w=0}, 0x66, 0x25, s, d) : + IMMREG({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x04, s, d); } + | 'l'? ws s:imm32-8 ws? ',' ws? d:r32 { $$ = (d.kind == ASM_EAX && s.imm.nbytes == 4) ? + IMM ({.w=0}, -1, 0x25, s, d) : + IMMREG({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x04, s, d); } + | 'q'? ws s:imm32-8 ws? ',' ws? d:r64 { $$ = (d.kind == ASM_RAX && s.imm.nbytes == 4) ? + IMM ({.w=1}, -1, 0x25, s, d) : + IMMREG({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x04, s, d); } + + | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x04, s, d); } + | 'w' ws s:imm16-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x04, s, d); } + | 'l' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x04, s, d); } + | 'q' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x04, s, d); } | 'b'? ws s:r8 ws? ',' ws? d:r8 { $$ = REGREG({.w=0}, -1, 0x20, s, d); } | 'w'? ws s:r16 ws? ',' ws? d:r16 { $$ = REGREG({.w=0}, 0x66, 0x21, s, d); } @@ -418,20 +423,23 @@ and = "and" ( ) cmp = "cmp" ( - 'b'? ws s:imm8 ws? ',' ws? d:al { $$ = IMM({.w=0}, -1, 0x3c, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:ax { $$ = IMM({.w=0}, 0x66, 0x3d, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:eax { $$ = IMM({.w=0}, -1, 0x3d, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:rax { $$ = IMM({.w=1}, -1, 0x3d, s, d); } - - | 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = IMMREG({.w=0}, -1, 0x80, 0x07, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:r16 { $$ = IMMREG({.w=0}, 0x66, 0x81, 0x07, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:r32 { $$ = IMMREG({.w=0}, -1, 0x81, 0x07, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:r64 { $$ = IMMREG({.w=1}, -1, 0x81, 0x07, s, d); } - - | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x07, s, d); } - | 'w' ws s:imm16 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, 0x81, 0x07, s, d); } - | 'l' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x81, 0x07, s, d); } - | 'q' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, 0x81, 0x07, s, d); } + 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = (d.kind == ASM_AL) ? + IMM ({.w=0}, -1, 0x3c, s, d) : + IMMREG({.w=0}, -1, 0x80, 0x07, s, d); } + | 'w'? ws s:imm16-8 ws? ',' ws? d:r16 { $$ = (d.kind == ASM_AX && s.imm.nbytes == 2) ? + IMM ({.w=0}, 0x66, 0x3d, s, d) : + IMMREG({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x07, s, d); } + | 'l'? ws s:imm32-8 ws? ',' ws? d:r32 { $$ = (d.kind == ASM_EAX && s.imm.nbytes == 4) ? + IMM ({.w=0}, -1, 0x3d, s, d) : + IMMREG({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x07, s, d); } + | 'q'? ws s:imm32-8 ws? ',' ws? d:r64 { $$ = (d.kind == ASM_RAX && s.imm.nbytes == 4) ? + IMM ({.w=1}, -1, 0x3d, s, d) : + IMMREG({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x07, s, d); } + + | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x07, s, d); } + | 'w' ws s:imm16-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x07, s, d); } + | 'l' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x07, s, d); } + | 'q' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x07, s, d); } | 'b'? ws s:r8 ws? ',' ws? d:r8 { $$ = REGREG({.w=0}, -1, 0x38, s, d); } | 'w'? ws s:r16 ws? ',' ws? d:r16 { $$ = REGREG({.w=0}, 0x66, 0x39, s, d); } @@ -450,20 +458,23 @@ cmp = "cmp" ( ) or = "or" ( - 'b'? ws s:imm8 ws? ',' ws? d:al { $$ = IMM({.w=0}, -1, 0x0c, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:ax { $$ = IMM({.w=0}, 0x66, 0x0d, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:eax { $$ = IMM({.w=0}, -1, 0x0d, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:rax { $$ = IMM({.w=1}, -1, 0x0d, s, d); } - - | 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = IMMREG({.w=0}, -1, 0x80, 0x01, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:r16 { $$ = IMMREG({.w=0}, 0x66, 0x81, 0x01, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:r32 { $$ = IMMREG({.w=0}, -1, 0x81, 0x01, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:r64 { $$ = IMMREG({.w=1}, -1, 0x81, 0x01, s, d); } - - | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x01, s, d); } - | 'w' ws s:imm16 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, 0x81, 0x01, s, d); } - | 'l' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x81, 0x01, s, d); } - | 'q' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, 0x81, 0x01, s, d); } + 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = (d.kind == ASM_AL) ? + IMM ({.w=0}, -1, 0x0c, s, d) : + IMMREG({.w=0}, -1, 0x80, 0x01, s, d); } + | 'w'? ws s:imm16-8 ws? ',' ws? d:r16 { $$ = (d.kind == ASM_AX && s.imm.nbytes == 2) ? + IMM ({.w=0}, 0x66, 0x0d, s, d) : + IMMREG({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x01, s, d); } + | 'l'? ws s:imm32-8 ws? ',' ws? d:r32 { $$ = (d.kind == ASM_EAX && s.imm.nbytes == 4) ? + IMM ({.w=0}, -1, 0x0d, s, d) : + IMMREG({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x01, s, d); } + | 'q'? ws s:imm32-8 ws? ',' ws? d:r64 { $$ = (d.kind == ASM_RAX && s.imm.nbytes == 4) ? + IMM ({.w=1}, -1, 0x0d, s, d) : + IMMREG({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x01, s, d); } + + | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x01, s, d); } + | 'w' ws s:imm16-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x01, s, d); } + | 'l' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x01, s, d); } + | 'q' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x01, s, d); } | 'b'? ws s:r8 ws? ',' ws? d:r8 { $$ = REGREG({.w=0}, -1, 0x08, s, d); } | 'w'? ws s:r16 ws? ',' ws? d:r16 { $$ = REGREG({.w=0}, 0x66, 0x09, s, d); } @@ -482,20 +493,23 @@ or = "or" ( ) sub = "sub" ( - 'b'? ws s:imm8 ws? ',' ws? d:al { $$ = IMM({.w=0}, -1, 0x2c, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:ax { $$ = IMM({.w=0}, 0x66, 0x2d, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:eax { $$ = IMM({.w=0}, -1, 0x2d, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:rax { $$ = IMM({.w=1}, -1, 0x2d, s, d); } - - | 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = IMMREG({.w=0}, -1, 0x80, 0x05, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:r16 { $$ = IMMREG({.w=0}, 0x66, 0x81, 0x05, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:r32 { $$ = IMMREG({.w=0}, -1, 0x81, 0x05, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:r64 { $$ = IMMREG({.w=1}, -1, 0x81, 0x05, s, d); } - - | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x05, s, d); } - | 'w' ws s:imm16 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, 0x81, 0x05, s, d); } - | 'l' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x81, 0x05, s, d); } - | 'q' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, 0x81, 0x05, s, d); } + 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = (d.kind == ASM_AL) ? + IMM ({.w=0}, -1, 0x2c, s, d) : + IMMREG({.w=0}, -1, 0x80, 0x05, s, d); } + | 'w'? ws s:imm16-8 ws? ',' ws? d:r16 { $$ = (d.kind == ASM_AX && s.imm.nbytes == 2) ? + IMM ({.w=0}, 0x66, 0x2d, s, d) : + IMMREG({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x05, s, d); } + | 'l'? ws s:imm32-8 ws? ',' ws? d:r32 { $$ = (d.kind == ASM_EAX && s.imm.nbytes == 4) ? + IMM ({.w=0}, -1, 0x2d, s, d) : + IMMREG({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x05, s, d); } + | 'q'? ws s:imm32-8 ws? ',' ws? d:r64 { $$ = (d.kind == ASM_RAX && s.imm.nbytes == 4) ? + IMM ({.w=1}, -1, 0x2d, s, d) : + IMMREG({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x05, s, d); } + + | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x05, s, d); } + | 'w' ws s:imm16-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x05, s, d); } + | 'l' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x05, s, d); } + | 'q' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x05, s, d); } | 'b'? ws s:r8 ws? ',' ws? d:r8 { $$ = REGREG({.w=0}, -1, 0x28, s, d); } | 'w'? ws s:r16 ws? ',' ws? d:r16 { $$ = REGREG({.w=0}, 0x66, 0x29, s, d); } @@ -514,20 +528,23 @@ sub = "sub" ( ) xor = "xor" ( - 'b'? ws s:imm8 ws? ',' ws? d:al { $$ = IMM({.w=0}, -1, 0x34, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:ax { $$ = IMM({.w=0}, 0x66, 0x35, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:eax { $$ = IMM({.w=0}, -1, 0x35, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:rax { $$ = IMM({.w=1}, -1, 0x35, s, d); } - - | 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = IMMREG({.w=0}, -1, 0x80, 0x06, s, d); } - | 'w'? ws s:imm16 ws? ',' ws? d:r16 { $$ = IMMREG({.w=0}, 0x66, 0x81, 0x06, s, d); } - | 'l'? ws s:imm32 ws? ',' ws? d:r32 { $$ = IMMREG({.w=0}, -1, 0x81, 0x06, s, d); } - | 'q'? ws s:imm32 ws? ',' ws? d:r64 { $$ = IMMREG({.w=1}, -1, 0x81, 0x06, s, d); } - - | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x06, s, d); } - | 'w' ws s:imm16 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, 0x81, 0x06, s, d); } - | 'l' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x81, 0x06, s, d); } - | 'q' ws s:imm32 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, 0x81, 0x06, s, d); } + 'b'? ws s:imm8 ws? ',' ws? d:r8 { $$ = (d.kind == ASM_AL) ? + IMM ({.w=0}, -1, 0x34, s, d) : + IMMREG({.w=0}, -1, 0x80, 0x06, s, d); } + | 'w'? ws s:imm16-8 ws? ',' ws? d:r16 { $$ = (d.kind == ASM_AX && s.imm.nbytes == 2) ? + IMM ({.w=0}, 0x66, 0x35, s, d) : + IMMREG({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x06, s, d); } + | 'l'? ws s:imm32-8 ws? ',' ws? d:r32 { $$ = (d.kind == ASM_EAX && s.imm.nbytes == 4) ? + IMM ({.w=0}, -1, 0x35, s, d) : + IMMREG({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x06, s, d); } + | 'q'? ws s:imm32-8 ws? ',' ws? d:r64 { $$ = (d.kind == ASM_RAX && s.imm.nbytes == 4) ? + IMM ({.w=1}, -1, 0x35, s, d) : + IMMREG({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x06, s, d); } + + | 'b' ws s:imm8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, 0x80, 0x06, s, d); } + | 'w' ws s:imm16-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, 0x66, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x06, s, d); } + | 'l' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=0}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x06, s, d); } + | 'q' ws s:imm32-8 ws? ',' ws? d:mem { $$ = IMMMEM({.w=1}, -1, s.imm.nbytes == 1 ? 0x83 : 0x81, 0x06, s, d); } | 'b'? ws s:r8 ws? ',' ws? d:r8 { $$ = REGREG({.w=0}, -1, 0x30, s, d); } | 'w'? ws s:r16 ws? ',' ws? d:r16 { $$ = REGREG({.w=0}, 0x66, 0x31, s, d); } @@ -835,14 +852,34 @@ scale-index-base = | '(' ws? b:r64 ws? ',' ws? i:r64 ')' { $$.memarg = (Memarg){.kind=ASM_MEMARG, .scale = 1, .index=i.kind, .base = b.kind, .disp = {0} }; } -imm8 = i:imm { i.imm.nbytes = 1; $$ = i; } -imm16 = i:imm { i.imm.nbytes = 2; $$ = i; } -imm32 = i:imm { i.imm.nbytes = 4; $$ = i; } imm = '$' ws? val:value { $$ = (Parsev) { .imm = (Imm){ .kind = ASM_IMM, .v = val.value, .nbytes = 0} }; } +imm8 = i:imm { i.imm.nbytes = 1; $$ = i; } +imm16 = i:imm { i.imm.nbytes = 2; $$ = i; } +imm32 = i:imm { i.imm.nbytes = 4; $$ = i; } + +imm16-8 = i:imm { + i.imm.nbytes = 2; + if (i.imm.v.l == NULL) { + if ((int64_t)(int8_t)i.imm.v.c == i.imm.v.c) { + i.imm.nbytes = 1; + } + } + $$ = i; +} +imm32-8 = i:imm { + i.imm.nbytes = 4; + if (i.imm.v.l == NULL) { + if ((int64_t)(int8_t)i.imm.v.c == i.imm.v.c) { + i.imm.nbytes = 1; + } + } + $$ = i; +} + al = "%al" { $$ = REG(AL); } cl = "%cl" { $$ = REG(CL); } ax = "%ax" { $$ = REG(AX); } diff --git a/test/test.sh b/test/test.sh index 4f33634..152ea2b 100644 --- a/test/test.sh +++ b/test/test.sh @@ -284,13 +284,27 @@ do for r in a b do # immediate variants + t "${op}b \$1, (%r${r}x)" t "${op}b \$127, (%r${r}x)" + t "${op}w \$1, (%r${r}x)" t "${op}w \$32767, (%r${r}x)" + t "${op}l \$1, (%r${r}x)" t "${op}l \$2147483647, (%r${r}x)" + t "${op}q \$1, (%r${r}x)" t "${op}q \$2147483647, (%r${r}x)" t "${op}b \$127, %${r}l" t "${op}w \$32767, %${r}x" + + t "${op}w \$1, %${r}x" + + t "${op}l \$1, %e${r}x" + t "${op}l \$1000, %e${r}x" + t "${op}l \$1000000, %e${r}x" t "${op}l \$2147483647, %e${r}x" + + t "${op}q \$1, %r${r}x" + t "${op}q \$1000, %r${r}x" + t "${op}q \$1000000, %r${r}x" t "${op}q \$2147483647, %r${r}x" # r -> m variants |