All of lore.kernel.org
 help / color / mirror / Atom feed
* PAE & hypercall interfaces
@ 2005-05-18 19:23 Gerd Knorr
  2005-05-19  8:20 ` Keir Fraser
  0 siblings, 1 reply; 3+ messages in thread
From: Gerd Knorr @ 2005-05-18 19:23 UTC (permalink / raw)
  To: xen-devel

  Hi folks,

As mentioned already in the pae patch mail, the xen hypercall
interfaces needs some changes.  The affected calls (hope I
didn't miss one) are:

 * do_mmuext_op
 * do_update_va_mapping & friends

The problem is that in PAE mode the page table entries are
64bit-sized instead of 32bit, and the current hypercall
interfaces allow to pass 32bit values only.

Current state is that the PAE code simply passes the lower
32 bits only, which will work fine as long as nobody attempts
to address memory above 4GB.  That obviously isn't an option,
the whole point of the PAE attempts is to use the memory above
4GB ;)

The options we have are:

  (1) make page table entrys 64bit wide everythere.
  (2) make page table entrys 64bit wide in PAE mode only.
  (3) pass two values: 32bit page frame number & flags.
  (4) others?

I'd prefeare option (1), option (3) is reasonable as well I
think.  I don't think (2) is a good idea, that will cause some
very nasty problems in case we'll try to allow non-PAE guests in
PAE mode.

Comments?

  Gerd

-- 
-mm seems unusually stable at present.
	-- akpm about 2.6.12-rc3-mm3

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: PAE & hypercall interfaces
  2005-05-18 19:23 PAE & hypercall interfaces Gerd Knorr
@ 2005-05-19  8:20 ` Keir Fraser
  0 siblings, 0 replies; 3+ messages in thread
From: Keir Fraser @ 2005-05-19  8:20 UTC (permalink / raw)
  To: Gerd Knorr; +Cc: xen-devel


On 18 May 2005, at 20:23, Gerd Knorr wrote:

>   (1) make page table entrys 64bit wide everythere.
>   (2) make page table entrys 64bit wide in PAE mode only.
>   (3) pass two values: 32bit page frame number & flags.
>   (4) others?
>
> I'd prefeare option (1), option (3) is reasonable as well I
> think.  I don't think (2) is a good idea, that will cause some
> very nasty problems in case we'll try to allow non-PAE guests in
> PAE mode.
>
> Comments?

I like option 3.

  -- Keir

^ permalink raw reply	[flat|nested] 3+ messages in thread

* RE: PAE & hypercall interfaces
@ 2005-05-19  8:20 Ian Pratt
  0 siblings, 0 replies; 3+ messages in thread
From: Ian Pratt @ 2005-05-19  8:20 UTC (permalink / raw)
  To: Gerd Knorr; +Cc: xen-devel

 

> On 18 May 2005, at 20:23, Gerd Knorr wrote:
> 
> >   (1) make page table entrys 64bit wide everythere.
> >   (2) make page table entrys 64bit wide in PAE mode only.
> >   (3) pass two values: 32bit page frame number & flags.
> >   (4) others?
> >
> > I'd prefeare option (1), option (3) is reasonable as well I 
> think.  I 
> > don't think (2) is a good idea, that will cause some very nasty 
> > problems in case we'll try to allow non-PAE guests in PAE mode.
> >
> > Comments?
> 
> I like option 3.

As do I, cooking down the NX bit.

Ian

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2005-05-19  8:20 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-05-18 19:23 PAE & hypercall interfaces Gerd Knorr
2005-05-19  8:20 ` Keir Fraser
  -- strict thread matches above, loose matches on Subject: below --
2005-05-19  8:20 Ian Pratt

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.