Re: KVM: x86: use proper port value when checking io instruction permission

[Gleb Natapov <gleb@redhat.com>]

On Tue, May 24, 2011 at 10:07:48PM +0300, Avi Kivity wrote:
> >>  @@ -2955,6 +2964,15 @@ static int check_perm_out(struct x86_emulate_ctxt *ctxt)
> >>   {
> >>   	struct decode_cache *c =&ctxt->decode;
> >>
> >>  +	switch (c->b) {
> >>  +	case 0x6e: /* outsb */
> >>  +	case 0x6f: /* outsw/outsd */
> >>  +	case 0xee: /* out dx,al */
> >>  +	case 0xef: /* out dx,(e/r)ax */
> >>  +		c->dst.val = c->regs[VCPU_REGS_RDX];
> >>  +		break;
> >>  +	}
> >>  +
> >>   	c->src.bytes = min(c->src.bytes, 4u);
> >>   	if (!emulator_io_permited(ctxt, c->dst.val, c->src.bytes))
> >>   		return emulate_gp(ctxt, 0);
> >I'd rather do it at decoding stage by adding SrcDX/DstDX.
> >
> 
> Note we haven't decoded operands yet.  And this doesn't fix in $imm8, %al.
> 
We haven't? check_perm is called from x86_emulate_insn() and operands are
decode in x86_decode_insn(). So $imm8, %al should work now. Or am I
missing something?

> Maybe we need an additional check site after operands are fetched.
> 
> -- 
> I have a truly marvellous patch that fixes the bug which this
> signature is too narrow to contain.

--
			Gleb.