Generating and catching invalid instruction

问题

I want to isolate a custom opcode in my program in order to catch invalid opcode exception if there is any. If I write an empty main, I see an assembly sequence as below:

(gdb) list
1   void main()
2   {
3   }
(gdb) disass /r main
Dump of assembler code for function main:
   0x00000000004004ed <+0>: 55  push   %rbp
   0x00000000004004ee <+1>: 48 89 e5    mov    %rsp,%rbp
   0x00000000004004f1 <+4>: 5d  pop    %rbp
   0x00000000004004f2 <+5>: c3  retq   
End of assembler dump.

If I write a single random byte such as __asm__(".byte 0x01"); I see that it automatically append 5d and c3 to that because it assumes the opcode is ADD and so it needs more than one byte.

0x00000000004004f1 <+4>:    01 5d c3    add    %ebx,-0x3d(%rbp)

End of assembler dump.

If I put some NOP (0x90) before and after that, it doesn't also help and it will automatically append some 0x90 to 0x01.

0x00000000004004f1 <+4>:    90  nop
0x00000000004004f2 <+5>:    90  nop
0x00000000004004f3 <+6>:    90  nop
0x00000000004004f4 <+7>:    01 90 90 90 90 90   add    %edx,-0x6f6f6f70(%rax)
0x00000000004004fa <+13>:   90  nop
0x00000000004004fb <+14>:   5d  pop    %rbp
0x00000000004004fc <+15>:   c3  retq   

By executing that ADD, I get segmentation fault error. Therefore, I can not distinguish seg fault and invalid instruction/opcode. I want to isolate that opcode. What about aligning with different pages and handling page faults?

Any comment?

来源：https://stackoverflow.com/questions/56935238/generating-and-catching-invalid-instruction