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

Re: [Xen-devel] Re: Crash with c/s 20097 (x86 vmx: Update EIP when appropriate during task switch)



On 02/12/2009 12:09, "Keir Fraser" <keir.fraser@xxxxxxxxxxxxx> wrote:

> Argh, stupid, and probably my mistake. I'll push the
> __get_instruction_length() inside the if()s.

I refactored the code as xen-unstable:20561. If that looks agreeable and
correct to you then I will also backport to 3.4-testing.

> Fortunately we never inject software interrupts or exceptions. I reworked
> the code to avoid that a long time ago, due to just such concerns. Possibly
> we should even ASSERT as such in the event-injection functions.

I decided against this because we write to VM_ENTRY_INTR_INFO in a few
places so it's not nice and centralised. But it's pretty obvious that SWInts
and SWExcs are not in the picture since we never write
VM_ENTRY_INSTRUCTION_LENGTH. So we're obviously safe (or even more utterly
broken :-) ).

 -- Keir



_______________________________________________
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®.