On Thu, Jul 04, 2013 at 01:06:16PM +0200, Denys Vlasenko wrote:
> On 07/04/2013 11:13 AM, Gleb Natapov wrote:
> > On Thu, Jul 04, 2013 at 10:58:29AM +0200, Denys Vlasenko wrote:
> >> The check sits in switch() statement which itself can check
> >> for opcode 0x90 far more efficiently.
> >>
> >> On assembler level, this change simply eliminates the following
> >> bit of code:
> >>
> >> 4c 8b a3 d8 00 00 00 mov 0xd8(%rbx),%r12
> >> 31 f6 xor %esi,%esi
> >> 48 89 df mov %rbx,%rdi
> >> e8 32 8c ff ff callq 490 <reg_read>
> >> 48 8d 83 78 01 00 00 lea 0x178(%rbx),%rax
> >> 83 8b 70 01 00 00 01 orl $0x1,0x170(%rbx)
> >> 83 8b 74 01 00 00 01 orl $0x1,0x174(%rbx)
> >> 49 39 c4 cmp %rax,%r12
> >> 0f 84 88 fa ff ff je 7304 <x86_emulate_insn+0x204>
> >>
> >> Signed-off-by: Denys Vlasenko <dvlas...@redhat.com>
> >> CC: Paolo Bonzini <pbonz...@redhat.com>
> >> CC: Avi Kivity <avi.kiv...@gmail.com>
> >> ---
> >> arch/x86/kvm/emulate.c | 6 +++---
> >> 1 file changed, 3 insertions(+), 3 deletions(-)
> >>
> >> diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
> >> index 2bc1e81..7e6e74a 100644
> >> --- a/arch/x86/kvm/emulate.c
> >> +++ b/arch/x86/kvm/emulate.c
> >> @@ -4544,9 +4544,9 @@ special_insn:
> >> case 0x8d: /* lea r16/r32, m */
> >> ctxt->dst.val = ctxt->src.addr.mem.ea;
> >> break;
> >> - case 0x90 ... 0x97: /* nop / xchg reg, rax */
> >> - if (ctxt->dst.addr.reg == reg_rmw(ctxt, VCPU_REGS_RAX))
> >> - break;
> >> + case 0x90: /* nop */
> >> + break;
> > This does not work on 64bit and REX prefix.
>
> Can you elaborate?
>
> 0x90 is special-cased in CPU to be a NOP regardless of bit width.
> IOW, xchg %eax,%eax ordinarily would clear upper 32 bits of %rax,
> but 0x90 doesn't do that.
>
> Do you mean that with REX.R==1, 0x90 will refer to R8?
Yes.
--
Gleb.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
