Index
> Main > movdqu illegal operand |
| Author |
|
|
revolution 01 May 2023, 23:45
The data size for 128-bit is dqword.
Code: movdqu xmm0, dqword[$] |
|||
01 May 2023, 23:45 |
|
|
Andy 02 May 2023, 00:23
I was induced in error by the example code I follow. Actually I use a piece a code from Intel Advanced Encryption Standard (AES) New Instructions Set White Paper.
Code: ; Cipher key is stored in “Key”. For example, ; Key 0x0f0e0d0c0b0a09080706050403020100 ; The key scheduled to be stored in the array Key_Schedule. movdqu xmm1, XMMWORD PTR Key movdqu XMMWORD PTR Key_Schedule, xmm1 mov rcx, OFFSET Key_Schedule+16 aeskeygenassist xmm2, xmm1, 0x1 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x2 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x4 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x8 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x10 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x20 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x40 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x80 call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x1b call key_expansion_128 aeskeygenassist xmm2, xmm1, 0x36 call key_expansion_128 jmp END; key_expansion_128: pshufd xmm2, xmm2, 0xff vpslldq xmm3, xmm1, 0x4 pxor xmm1, xmm3 vpslldq xmm3, xmm1, 0x4 pxor xmm1, xmm3 vpslldq xmm3, xmm1, 0x4 pxor xmm1, xmm3 pxor xmm1, xmm2 movdqu XMMWORD PTR [rcx], xmm1 add rcx, 0x10 ret END: It works if I replace XMMWORD with dqword but I got other errors for lines mov rcx, OFFSET Key_Schedule+16 and add rcx, 0x10. I tried even more basic instructions with this register, like xor rcx, rcx and fails also with the same error "Illegal instruction". This line is also strange in this example since rcx is 64 bits in size and xmm1 is 128 bits so I suppose this line will also throw an error eventually. Code: movdqu XMMWORD PTR [rcx], xmm1 BTW what flavor of asm is used in these intel white papers? |
|||
|
02 May 2023, 00:23 |
|
|
revolution 02 May 2023, 00:27
Offset is not part of fasm. just delete it.
To access 64-bit instructions you need to put it into a 64-bit mode. Either use64 or with format ... Code: format pe64 ... ; 64-bit code Code: use64 ; 64-bit |
|||
|
02 May 2023, 00:27 |
|
|
Andy 02 May 2023, 15:10
I almost succeeded to write the desired code but I encounter a strange behavior after the first call of aesenc.
Code: format PE64 console entry start section '.data' data readable writeable msg db 0x32, 0x43, 0xf6, 0xa8, 0x88, 0x5a, 0x30, 0x8d, 0x31, 0x31, 0x98, 0xa4, 0xe0, 0x37, 0x07, 0x34 key db 0x2b, 0x7e, 0x15, 0x16, 0x28, 0xae, 0xd2, 0xa6, 0xab, 0xf7, 0x15, 0x88, 0x09, 0xcf, 0x4f, 0x3c enc db 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 section '.code' code readable executable start: movdqu xmm0, dqword [msg] movdqu xmm1, dqword [key] pxor xmm2, xmm2 movdqu xmm5, xmm1 pxor xmm0, xmm1 aeskeygenassist xmm6, xmm1, 0x01 pshufd xmm6, xmm6, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm6 aesenc xmm0, xmm1 aeskeygenassist xmm7, xmm1, 0x02 pshufd xmm7, xmm7, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm7 aesenc xmm0, xmm1 aeskeygenassist xmm8, xmm1, 0x04 pshufd xmm8, xmm8, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm8 aesenc xmm0, xmm1 aeskeygenassist xmm9, xmm1, 0x08 pshufd xmm9, xmm9, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm9 aesenc xmm0, xmm1 aeskeygenassist xmm10, xmm1, 0x10 pshufd xmm10, xmm10, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm10 aesenc xmm0, xmm1 aeskeygenassist xmm11, xmm1, 0x20 pshufd xmm11, xmm11, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm11 aesenc xmm0, xmm1 aeskeygenassist xmm12, xmm1, 0x40 pshufd xmm12, xmm12, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm12 aesenc xmm0, xmm1 aeskeygenassist xmm13, xmm1, 0x80 pshufd xmm13, xmm13, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm13 aesenc xmm0, xmm1 aeskeygenassist xmm14, xmm1, 0x1b pshufd xmm14, xmm14, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm14 aesenc xmm0, xmm1 aeskeygenassist xmm15, xmm1, 0x36 pshufd xmm15, xmm15, 11111111b shufps xmm2, xmm1, 00010000b pxor xmm1, xmm2 shufps xmm2, xmm1, 10001100b pxor xmm1, xmm2 pxor xmm1, xmm15 aesenclast xmm0, xmm1 I use a tool that simulate each round so I tried to follow in debugger to see if I get the same results for each encryption round and it's pretty much the same except for the first 4 bytes (underlined with green color). I attached two screenshots with what I expect to get and what I get in debugger after the first round of encryption.
|
|||||||||||||||||||
|
02 May 2023, 15:10 |
|
||||||||||||||||||
< Last Thread | Next Thread > |
Forum Rules:
|