Re: [Xen-devel] [PATCH 0/3] x86/paravirt: Fix baremetal paravirt MSR ops

[Andy Lutomirski <luto@xxxxxxxxxxxxxx>]

On Thu, Sep 17, 2015 at 8:17 AM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> On Thu, Sep 17, 2015 at 08:17:18AM -0700, Andy Lutomirski wrote:
>
>> > Ah, that would be good news. Andy earlier argued I could not rely on
>> > rdmsr_safe() faulting on unknown MSRs. If practically we can there's
>> > some code I can simplify :-)
>>
>> I was taking about QEMU TCG, not KVM.
>
> Just for my education, TCG is the thing without _any_ hardware assist?

Yes.

> The thing you fear to use because it cannot boot a kernel this side of
> tomorrow etc.. ?

I actually use it on a semi-regular basis.  It appears to boot a
normal kernel correctly and surprisingly quickly.  It's important for
a silly reason.  Asking KVM on an Intel host to emulate an AMD CPU or
vice versa results in a chimera: I get an Opteron that's
"GenuineIntel", which causes Linux to treat it as Intel, which means
it gets the Intel quirks (SYSENTER in long mode, no
X86_BUG_SYSRET_SS_ATTRS, etc) instead of the AMD quirks (SYSCALL in
compat mode, etc).  So, if I want to test compat SYSCALL, I have to
use TCG, which will actually do a decent job of emulating an AMD CPU
for me.

Maybe Paolo can fix QEMU to fail bad MSR accesses for real...

--Andy

-- 
Andy Lutomirski
AMA Capital Management, LLC

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel