Re: [PATCH v3 01/13] x86/retpoline: Add initial retpoline support

[Borislav Petkov]

On Sun, Jan 07, 2018 at 09:40:42AM +0000, David Woodhouse wrote:
> Right, so it all tends to work out OK purely by virtue of the fact that
> oldinstr and altinstr end up far enough apart in the image that they're
> 5-byte jumps. Which isn't perfect but we've lived with worse.

Well, the reference point is important. And I don't think we've done
more involved things than jumping back to something in .text proper.
However, I think I know how to fix this so that arbitrary jump offsets
would work but I need to talk to our gcc guys first.

If the jump is close enough for 2 bytes, then it should work as long as
the offset to the target doesn't change.

The main thing recompute_jumps() does is turn 5-byte jumps - which gas
creates because the jump target is in .text but the jump itself is in
.altinstr_replacement - into 2-byte ones. Because when you copy the jump
back into .text, the offset might fit in a signed byte all of a sudden.

There are still some nasties with forcing 5-byte jumps but I think I
know how to fix those. Stay tuned...

> I'm relatively pleased that we've managed to eliminate this as a
> dependency for inverting the X86_FEATURE_RETPOLINE logic though, by
> following Linus' suggestion to just emit the thunk inline instead of
> calling the same one as GCC.
> 
> The other fun one for alternatives is in entry_64.S, where we really
> need the return address of the call instruction to be *precisely* theÂ
> .Lentry_SYSCALL_64_after_fastpath_call label, so we have to eschew the
> normal NOSPEC_CALL there:

So CALL, as the doc says, pushes the offset of the *next* insn onto the
stack and branches to the target address.

So I'm thinking, as long as the next insn doesn't move and gcc doesn't
pad anything, you're fine.

However, I suspect that I'm missing something else here and I guess I'll
have more clue if I look at the whole thing. So can you point me to your
current branch so that I can take a look at the code?

Thx.

-- 
Regards/Gruss,
    Boris.

SUSE Linux GmbH, GF: Felix ImendÃrffer, Jane Smithard, Graham Norton, HRB 21284 (AG NÃrnberg)
--