Re: [Xen-devel] [PATCH] x86/amd: fix crash as Xen Dom0 on AMDTrinity systems

[Borislav Petkov]

On Wed, May 30, 2012 at 10:49:29AM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, May 30, 2012 at 07:42:56AM -0700, H. Peter Anvin wrote:
> > On 05/30/2012 07:23 AM, Jan Beulich wrote:
> > > 
> > > I see - the Xen code blindly overwrites pv_cpu_ops, despite not
> > > having initialized all members. That's an obvious oversight of the
> > > patch that introduced the _regs variants.
> > > 
> > > Plus having secondary instances of things like rdmsrl_amd_safe()
> > > in asm/paravirt.h seems pretty strange an approach (which was
> > > why initially I didn't spot how a crash could happen there) - only
> > > the lowest level functions should get re-implemented here.
> > > 
> > 
> > This kinds of things are part of why Xen makes me want to cry regularly.
> 
> It looks like an oversight by Borislav (177fed1ee8d727c39601ce9fc2299b4cb25a718e
> and 132ec92f) and Yinghai (b05f78f5c713eda2c34e495d92495ee4f1c3b5e1) where
> they added these wrappers way back in 2009!

This is exactly why xen has nothing to do in arch/x86/. It is not an
oversight - I simply didn't test it on xen because I don't care about
it.

Remember our last discussion about mcelog?

This current case should be a perfect example for why xen shouldn't be
sprinkling code all over the place.

-- 
Regards/Gruss,
Boris.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/