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

Re: [Xen-devel] [V3] x86/xsaves: fix overwriting between non-lazy/lazy xsaves



>>> On 09.03.16 at 13:33, <shuai.ruan@xxxxxxxxxxxxxxx> wrote:
> On Wed, Mar 09, 2016 at 03:25:58AM -0700, Jan Beulich wrote:
>> Btw., one more thing: Can't the exclusion of FP and SSE states in the logic
>> determining which state set to save be extended to also include YMM? If 
> saved,
>> YMM will also always live at a fixed place (an ASSERT() or BUG_ON() to 
> verify
>> would of course be desirable). And if the guest didn't touch YMM registers, 
> the
>> respective bit in the mask won't be set anyway.
>> 
> YMM do live at the begining of the xsave extended area. But when we
> exclude YMM like FP/SSE states ,if guest set YMM | XSTATE_NONLAZY respective 
> bitin xcro_accum, the over-writing may happend too. In such case , 
> vcpu_xsave_mask
> will return XSTATE_NONLAZY, the first xstate of XSTATE_NONLAZY will be
> xsaved at the begining of xsave extended area which may over-write YMM
> state.
> Do I miss something ?

Oh, yes, you're right. That would work only if YMM was part of
NOLAZY.

Jan


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

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.