Re: [v6 PATCH 06/21] x86/insn-eval: Add utility functions to get segment selector

[Ricardo Neri]

On Sun, 2017-04-30 at 19:15 +0200, Borislav Petkov wrote:
> On Wed, Apr 26, 2017 at 01:44:43PM -0700, Ricardo Neri wrote:
> > I regard that the role of this function is to obtain the the segment
> > selector from either of the prefixes or inferred from the operands. It
> > is the role of caller to determine if the segment selector should be
> > ignored.
> 
> No, this is wrong. The function is called resolve_seg_selector() and it
> gives you the segment selector. CS, DS, ES, and SS in 64-bit mode are
> treated as null segments and your function should return/signal exactly
> that, i.e, saying that those should be ignored in that case.
> 
> > I double-checked the latest version of the Intel Software Development
> > manual [2], in the table 3-5 in section 3.7.4 mentions that DS is
> > default segment for all data references, except string destinations. I
> > tested this code with the UMIP-protected instructions and whenever I use
> > %edi the default segment is %ds.
> 
> Yes, all correct. Except that we're adding a more-or-less generic x86
> insn decoder so we should make it so...
> 
> > Is this example valid? The documentation of MOVS specifies that it
> > always moves DS:(E)SI to ES:(E)DI.
> 
> ... that the decoder should do exactly that:
> 
> 	if (MOVS and rDI)
> 		return SEG_ES;
> 
> And you're handing in struct insn * so you can easily check which insn
> you're looking at.

I see. I have submitted v7 of the series and I have implemented all the
changes above. Now I am able to identify string instructions.

Thanks and BR,
Ricardo