Xen devel - Aug 2010 - [PATCH 1/4] XSAVE/XRSTOR: Fix frozen states

If a guest sets a state and dirties the state, but later temporarily clears the
state, and at this time this vcpu is scheduled out, then other vcpus may corrupt
the state before the vcpu is scheduled in again, thus the state cannot be
restored correctly. To solve this issue, this patch save/restore all states
unconditionally on vcpu context switch.

Signed-off-by: Weidong Han <weidong.han@intel.com>


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

>>> On 31.08.10 at 04:59, "Han, Weidong"
<weidong.han@intel.com> wrote:
> If a guest sets a state and dirties the state, but later temporarily clears
> the state, and at this time this vcpu is scheduled out, then other vcpus
may
> corrupt the state before the vcpu is scheduled in again, thus the state 
> cannot be restored correctly. To solve this issue, this patch save/restore 
> all states unconditionally on vcpu context switch.
> 
> Signed-off-by: Weidong Han <weidong.han@intel.com>
There''s a lot of seemingly unrelated variable renaming in this patch,
the essence apparently just being
>--- a/xen/arch/x86/i387.c	Fri Aug 27 07:10:57 2010 -0400
>+++ b/xen/arch/x86/i387.c	Fri Aug 27 12:31:05 2010 -0400
>@@ -35,7 +35,10 @@ void save_init_fpu(struct vcpu *v)
> 
>     if ( cpu_has_xsave && is_hvm_vcpu(v) )
>     {
>+        /* Save all supported feature states */
>+        set_xcr0(xfeature_mask);
>         xsave(v);
>+        set_xcr0(v->arch.hvm_vcpu.xcr0);
>     }
>     else if ( cpu_has_fxsr )
>     {
Now, I wonder what the (performance) price of this is (and will be with
larger register sets in the future), and whether the general deferred
saving scheme can''t be made work here too.

Jan


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

On 31/08/2010 07:56, "Jan Beulich" <JBeulich@novell.com> wrote:
>> --- a/xen/arch/x86/i387.c Fri Aug 27 07:10:57 2010 -0400
>> +++ b/xen/arch/x86/i387.c Fri Aug 27 12:31:05 2010 -0400
>> @@ -35,7 +35,10 @@ void save_init_fpu(struct vcpu *v)
>> 
>>     if ( cpu_has_xsave && is_hvm_vcpu(v) )
>>     {
>> +        /* Save all supported feature states */
>> +        set_xcr0(xfeature_mask);
>>         xsave(v);
>> +        set_xcr0(v->arch.hvm_vcpu.xcr0);
>>     }
>>     else if ( cpu_has_fxsr )
>>     {
> 
> Now, I wonder what the (performance) price of this is (and will be with
> larger register sets in the future), and whether the general deferred
> saving scheme can''t be made work here too.
Well, quite. Seems the cheap easy fix rather than the right fix,
doesn''t it.

 -- Keir



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