On Tue, Nov 24, 2020 at 11:19:33AM +0100, Borislav Petkov wrote:
> In any case, at least the case where I give it
> 
> 0x48 0xcf 0x48 0x83
> 
> and say that buf size is 4, should return an error because the second
> insn is incomplete. So I need to go look at that now.

Ok, got it:

./arch/x86/tools/insn_sanity: Success: decoded and checked 10000 random instructions with 0 errors (seed:0x826fdf9c)
insn buffer:
0x48 0xcf 0x48 0x83 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90 
supplied buf size: 15, ret 0
supplied buf size: 2, ret 0
supplied buf size: 3, ret 0
supplied buf size: 4, ret 0
supplied buf size: 1, ret -22

the current decoder simply decodes the *first* insn in the buffer it
encounters and that's it.

When you give it a buffer of size smaller than the first instruction:

supplied buf size: 1, ret -22

while the first insn is 2 bytes long:

0x48 0xcf (IRETQ)

then it signals an error.

Andy, does that work for your use cases?

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette