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

Re: [Xen-devel] [V9 PATCH 6/8] pvh dom0: Add and remove foreign pages



At 09:54 +0100 on 28 Apr (1398675248), Jan Beulich wrote:
> >>> On 26.04.14 at 03:34, <mukesh.rathor@xxxxxxxxxx> wrote:
> > Ok, so far:
> > 
> > option 1, Jan's suggestion: 
> >   - Use my latest patch, but add explicit set of p2m type for intermediate
> >     entries in ept_set_middle_entry() like Jan said.
> > 
> >     Good: we won't need to then worry about checking for leaf entries.
> >     Bad: future code can't use the bits for interm entries.
> > 
> > option 2, Tim's suggestion:
> >   - Use my latest patch, but add sp = 1 in leaf entries. Then use level
> >     to figure if we are in superpage. Will work too, but will change the
> >     definition of sp bit, and superpage check will always require checking
> >     for level. 
> > 
> > option 3:
> >   - Pass level to atomic_write_ept_entry() and check for foreign only in
> >     case of level == 0, ie leaf entries where foreign can be set.
> > 
> >     It appears that information is available at all call sites, and would
> >     be least intrusive in terms of using bits for non-leaf entries.
> > 
> > option 4:
> >   - each call site checks for level and calls either macro for leaf entry
> >     or macro for non-leaf entry. same as 3, but the check is in call site.
> 
> I think my preference would go backwards from 4 to 1, but I think
> Tim would - like you - prefer 3 over 4, which I wouldn't object to.

My order would be 2, 3, 4, 1, but really only objecting to #1; so that
sounds like consensus for option 3.

Tim.

_______________________________________________
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®.