Re: [Xen-devel] [PATCH] x86: clear guest's EFLAGS.RF after emulating instructions

[Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>]

On 19/6/07 11:09, "Jan Beulich" <jbeulich@xxxxxxxxxx> wrote:

>> You set EFLAGS.RF in a few places when injecting #PF. Why would you do that?
>> It isn't Xen's job to set that flag ever afaics, and since the original trap
>> into Xen was not due to #DB you're not avoiding an infinite loop.
> 
> This is only in handle_mmio() where I cleared the flag earlier - since #PF has
> lower
> priority than #DB, .RF must have been set when the exception was reported, and
> hence I must re-set it when forwarding the exception.

Oh, I see. I'm not too familiar with #DB or EFLAGS.RF, but this does appear
to make sense given the info in the reference manual.

 -- Keir



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