add 8 bit opcode support

twasm/README.md

@@ -272,33 +272,47 @@ entries are as follows:
 +----------+---------------+----------------------------------+
 
 16 bytes
-+----------------------------------------------+
-| 2 operand operators                          |
-+----------------------------------------------+
-| 127                                       96 |
-+----------------------------------------------+
-| reserved                                     |
-+-------------------+-------+-------+----------+
-| 95             80 | 79 76 | 75 72 | 71    64 |
-+-------------------+-------+-------+----------+
-| reserved          | op3&8 | op2&8 | reserved |
-+-------------------+-------+-------+----------+
-| 63             48 | 47         40 | 39    32 |
-+-------------------+---------------+----------+
-| reserved          | opcode        | opcode   |
-|                   | dst=r/m       | dst=r/m  |
-|                   | src=imm8      | src=imm  |
-+---------+---------+---------------+----------+
-| 31   24 | 23   16 | 15                    0  |
-+---------+---------+--------------------------+
-| opcode  | opcode  | token ID                 |
-| dst=r   | dst=r/m |                          |
-| src=r/m | src=r   |                          |
-+---------+---------+--------------------------+
++-----------------------------------------------+
+| 2 operand operators                           |
++-----------------------------------------------+
+| 127                                        96 |
++-----------------------------------------------+
+| reserved                                      |
++---------+----------+-------+-------+----------+
+| 95   88 | 87    80 | 79 76 | 75 72 | 71    64 |
++---------+----------+-------+-------+----------+
+| flags   | reserved | op3&8 | op2&8 | reserved |
++---------+----------+-------+-------+----------+
+| 63              48 | 47         40 | 39    32 |
++--------------------+---------------+----------+
+| reserved           | opcode        | opcode   |
+|                    | dst=r/m       | dst=r/m  |
+|                    | src=imm8      | src=imm  |
++---------+----------+---------------+----------+
+| 31   24 | 23    16 | 15                    0  |
++---------+----------+--------------------------+
+| opcode  | opcode   | token ID                 |
+| dst=r   | dst=r/m  |                          |
+| src=r/m | src=r    |                          |
++---------+----------+--------------------------+
+
+1 byte
++-----------------+
+| flags byte      |
++----------+------+
+| 95    89 |  88  |
++----------+------+
+| reserved | 8bit |
++----------+------+
+
+; flags key:
+8bit ; tte has opcodes for r/m8 and r8 instead of r/m and r respectively
 
 ; key:
 r/m  ; r/m 16/32/64
+r/m8 ; r/m 8
 r    ; r   16/32/64
+r8   ; r   8
 imm  ; imm 16/32
 imm8 ; imm 8
 rel  ; rel 16/32

twasm/asm/main.asm

@@ -374,8 +374,8 @@ assemble:
 
     and al, 11b ; al = register width
 
-    cmp al, 00b          ; 8 bit
-    je .unexpected_token ; TODO handle 8 bit opcodes
+    cmp al, 00b ; 8 bit
+    je .operator_2_register_8
 
     cmp al, 01b ; 16 bit
     je .operator_2_register_16
@@ -385,11 +385,16 @@ assemble:
 
     cmp al, 11b ; 64 bit
     je .operator_2_register_64
+  .operator_2_register_8:
+    mov bl, 1b ; operator flag 8bit
+    jmp .operator_2_register_continue
   .operator_2_register_16:
+    xor ebx, ebx ; no operator flags
     mov al, 0x66
     call .push_byte
     jmp .operator_2_register_continue
   .operator_2_register_64:
+    xor ebx, ebx ; no operator flags
     mov al, 0x48
     call .push_byte
     jmp .operator_2_register_continue
@@ -430,6 +435,7 @@ assemble:
     push rsi
     mov di, cx ; di = tte of operator
     mov sil, 1 ; dst = reg
+    mov bl, 1  ; bl = operator flag byte
     call get_opcode
     ; al = opcode
     ; dl = op flag
@@ -484,6 +490,7 @@ assemble:
     push rsi
     mov di, cx ; di = tte of operator
     mov sil, 1 ; dst = reg
+               ; bl = operator flag byte
     call get_opcode
     ; al = opcode
     ; dl = op flag
@@ -606,6 +613,17 @@ assemble:
     push rsi
     mov di, cx ; di = tte of operator
     mov sil, 2 ; dst=r/m,src=imm
+               ; bl = operator flag byte
+               ; TODO change sil based on whether bl is 8 bit or not
+    push rbx
+    and ebx, 1
+    cmp bl, 1  ; bit8 flag
+    pop rbx
+    je .operator_2_register_const_get_opcode_8
+    jmp .operator_2_register_const_get_opcode_continue
+  .operator_2_register_const_get_opcode_8:
+    mov sil, 3 ; dst=r/m,src=imm8
+  .operator_2_register_const_get_opcode_continue:
     call get_opcode
     ; al = opcode
     ; dl = op flag
@@ -616,6 +634,12 @@ assemble:
     call .next_token
     jge .break
 
+    push rbx
+    and ebx, 1
+    cmp bl, 1  ; bit8 flag
+    pop rbx
+    je .operator_2_register_const_8
+
     push rdi
     push rsi
     mov edi, .buffer_end - .buffer ; length of buffer
@@ -639,6 +663,11 @@ assemble:
     je .operator_2_register_const_16
 
     jmp .operator_2_register_const_32
+  .operator_2_register_const_8:
+    mov ecx, [.tokens_processed]
+    mov al, [TOKEN_TABLE_ADDR + 2 * rcx] ; get the next byte from the tt
+    call .write_byte                     ; and add it to the buffer
+    jmp .operator_2_register_const_continue
   .operator_2_register_const_16:
     mov ecx, [.tokens_processed]
     mov ax, [TOKEN_TABLE_ADDR + 2 * rcx] ; get the next 2 bytes from the tt
@@ -981,6 +1010,7 @@ get_ModRM:
 ; di = token table entry
 ; sil = offset within opcode entry. 0 is the first opcode, 1 the second, and so
 ;       on
+; bl = flag byte
 ;
 ; returned:
 ; al = opcode; the rest of rax is zeroed.
@@ -989,6 +1019,7 @@ get_ModRM:
 
 get_opcode:
   and edi, 0xFFFF ; di = token table entry
+  and ebx, 0xFF   ; bl = flag byte
 
   add esi, 2
   and esi, 111b ; offset within opcode entry
@@ -1005,10 +1036,20 @@ get_opcode:
     shr eax, 4
 
     cmp cx, di
-    je .found
+    je .maybe_found
 
     inc eax
     jmp .loop
+
+  .maybe_found:
+    shl eax, 4
+    mov cl, [opcodes.by_id + 11 + eax]
+    shr eax, 4
+    cmp cl, bl
+    je .found
+    inc eax
+    jmp .loop
+
   .not_found:
     xor eax, eax
     mov eax, UNRECOGNISED_ID_OPCODE
@@ -2677,6 +2718,22 @@ opcodes:
 
     dd 0x00000000
 
+    ; mov bit8
+    dw 0x0056
+    db 0x88        ; r/m8 <- r8
+    db 0x8A        ; r8   <- r/m8
+
+    db 0x00
+    db 0xC6        ; r/m8 <- imm8
+    dw 0x0000
+
+    dd 0x01000000 ; 000:
+                  ; 0:   r/m8 <- imm8 op flag
+                  ; 00:
+                  ; 01: bit8 flag
+
+    dd 0x00000000
+
     ; add
     dw 0x0057
     db 0x01        ; r/m <- r
@@ -2747,6 +2804,20 @@ opcodes:
 
     dd 0x00000000
 
+    ; cmp bit8
+    dw 0x005B
+    db 0x38       ; r/m8 <- r8
+    db 0x3A       ; r8   <- r/m8
+
+    db 0x00
+    db 0x80       ; r/m8 <- imm8
+    dw 0x0000
+
+    dd 0x01007000 ; 000:
+                  ; 7:   r/m8 <- imm8 op flag
+                  ; 00:
+                  ; 01: bit8 flag
+
     ; jmp
     dw 0x005C
     db 0xFF   ; r/m

twasm/asm/tests.asm

@@ -295,6 +295,7 @@ test_get_opcode:
 
   mov di, 0x0053 ; xor
   mov sil, 0
+  mov bl, 0
   call get_opcode
   cmp al, 0x31
   jne .fail
@@ -303,6 +304,7 @@ test_get_opcode:
 
   mov di, 0x0053 ; xor
   mov sil, 1
+  mov bl, 0
   call get_opcode
   cmp al, 0x33
   jne .fail
@@ -311,6 +313,7 @@ test_get_opcode:
 
   mov di, 0x0053 ; xor
   mov sil, 2
+  mov bl, 0
   call get_opcode
   cmp al, 0x81
   jne .fail
@@ -319,6 +322,7 @@ test_get_opcode:
 
   mov di, 0x0053 ; xor
   mov sil, 3
+  mov bl, 0
   call get_opcode
   cmp al, 0x83
   jne .fail
@@ -327,6 +331,7 @@ test_get_opcode:
 
   mov di, 0x0054  ; inc
   mov sil, 0
+  mov bl, 0
   call get_opcode
   cmp al, 0xFF
   jne .fail
@@ -335,6 +340,7 @@ test_get_opcode:
 
   mov di, 0x0055  ; dec
   mov sil, 0
+  mov bl, 0
   call get_opcode
   cmp al, 0xFF
   jne .fail
@@ -343,6 +349,7 @@ test_get_opcode:
 
   mov di, 0x004F  ; hlt
   mov sil, 0
+  mov bl, 0
   call get_opcode
   cmp al, 0xF4
   jne .fail
@@ -351,6 +358,7 @@ test_get_opcode:
 
   mov di, 0x0059  ; call
   mov sil, 0q0
+  mov bl, 0
   call get_opcode
   cmp al, 0xFF
   jne .fail
@@ -359,6 +367,7 @@ test_get_opcode:
 
   mov di, 0x0003  ; rdx (not an operator)
   mov sil, 0q0
+  mov bl, 0
   call get_opcode
   cmp al, UNRECOGNISED_ID_OPCODE
   jne .fail

twasm/package.nix

@@ -14,7 +14,8 @@ stdenv.mkDerivation {
       asm/main.asm \
       -o out.bin \
       -w+all \
-      -w-reloc-abs
+      -w-reloc-abs \
+      -O0
   '';
   installPhase = ''
     dd if=/dev/zero of=disk bs=512 count=2880