[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] x86_emulate(): Upon LOCK prefix for MOV, should inject #UD rather than #GP?


  • To: "Liu, Eric E" <eric.e.liu@xxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>
  • Date: Thu, 20 Dec 2007 09:42:16 +0000
  • Delivery-date: Thu, 20 Dec 2007 01:43:04 -0800
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AchCtD/3n07kelioRSCaZ5nHelu8BQAOFsFl
  • Thread-topic: [Xen-devel] x86_emulate(): Upon LOCK prefix for MOV, should inject #UD rather than #GP?

Strictly speaking we don't actually generate exceptions yet. We return
X86EMUL_UNHANDLEABLE to the caller, and they will often inject #UD as a
result. But yes, it really should be EXC_UD.

 -- Keir


On 20/12/07 02:58, "Liu, Eric E" <eric.e.liu@xxxxxxxxx> wrote:

> Hi all,
> The Intel SDM(2A) says that if the LOCK prefix is used with some
> instructions such as "mov" and the souce  operand is a memory operand, a
> #UD may be generated, but in the x86_emulate.c there are places such as
> :
>     942     /* Lock prefix is allowed only on RMW instructions.
> */
>     943     generate_exception_if((d & Mov) && lock_prefix,
> EXC_GP);
> here we inject EXC_GP but not EXC_UD, is it a bug?
> 
> --Eric (Liu, Feng)
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.