Xen devel - Mar 2007 - ptep_set_access_flags() using UVMF_MULTI?

Why is this so when the touched mm is active, even though the alternative
case and native only use flush_tlb_page() (i.e. only flush the local tlb)?

Thanks, Jan

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

On 13/3/07 11:14, "Jan Beulich" <jbeulich@novell.com> wrote:
> Why is this so when the touched mm is active, even though the alternative
> case and native only use flush_tlb_page() (i.e. only flush the local tlb)?
> 
> Thanks, Jan
We don''t need any more flushing than native. Perhaps the original
native
code did a cross-TLBs flush and that got reduced in more recent kernels but
the Xen code didn''t get modified? Anyway, please send a patch.

 -- Keir


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

>>> Keir Fraser <keir@xensource.com> 13.03.07 13:09 >>>
>On 13/3/07 11:14, "Jan Beulich" <jbeulich@novell.com> wrote:
>
>> Why is this so when the touched mm is active, even though the
alternative
>> case and native only use flush_tlb_page() (i.e. only flush the local
tlb)?
>> 
>> Thanks, Jan
>
>We don''t need any more flushing than native. Perhaps the original
native
>code did a cross-TLBs flush and that got reduced in more recent kernels but
>the Xen code didn''t get modified? Anyway, please send a patch.
My mistake, I looked at the !SMP case (and even half-way knew I did, but
didn''t
bother to look at the SMP one, too.

Sorry for the noise, Jan

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