linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Page table sharing
@ 2002-09-25 18:12 Gerold J. Wucherpfennig
  2002-09-25 18:40 ` Dave McCracken
  0 siblings, 1 reply; 6+ messages in thread
From: Gerold J. Wucherpfennig @ 2002-09-25 18:12 UTC (permalink / raw)
  To: linux-kernel

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

What about page table sharing? Does anybody still care about this?

The patch from Daniel Phillips
(http://www.geocrawler.com/mail/msg.php3?msg_id=7855063&list=35)
is a few month old and I can't see any progress.

Sorry, I'm not a kernel expert, so I can't help.
But page table sharing is still listed as betaware at the
Linux Kernel 2.5 Status page (http://kernelnewbies.org/status/latest.html)
and Page Table sharing isn't marked post Halloween.

Some comments from Daniel Phillips or Dave McCracken?


Gerold

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE9kfyY53zyw+ONqrERAlvgAJ0RCNWZm7MEf7wGshZcE09oA5N7LQCfVuhq
L/EX/pUmmjOQAEtXpaDm2yA=
=QdO3
-----END PGP SIGNATURE-----


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Page table sharing
  2002-09-25 18:12 Page table sharing Gerold J. Wucherpfennig
@ 2002-09-25 18:40 ` Dave McCracken
  2002-09-26 18:20   ` Gerold J. Wucherpfennig
  2002-10-01 13:08   ` Daniel Phillips
  0 siblings, 2 replies; 6+ messages in thread
From: Dave McCracken @ 2002-09-25 18:40 UTC (permalink / raw)
  To: Gerold J. Wucherpfennig, linux-kernel


--On Wednesday, September 25, 2002 20:12:36 +0200 "Gerold J. Wucherpfennig"
<gjwucherpfennig@gmx.net> wrote:

> What about page table sharing? Does anybody still care about this?
> 
> The patch from Daniel Phillips
> (http://www.geocrawler.com/mail/msg.php3?msg_id=7855063&list=35)
> is a few month old and I can't see any progress.
> 
> Sorry, I'm not a kernel expert, so I can't help.
> But page table sharing is still listed as betaware at the
> Linux Kernel 2.5 Status page (http://kernelnewbies.org/status/latest.html)
> and Page Table sharing isn't marked post Halloween.
> 
> Some comments from Daniel Phillips or Dave McCracken?

I'm working on it.  I sent out a patch to the mm list a few weeks ago, but
it didn't have the locking right.  I'm in the proces of finishing an
improved version with new locking.  I'll send a snapshot of it out when I
can make it stop oopsing :)

Dave McCracken

======================================================================
Dave McCracken          IBM Linux Base Kernel Team      1-512-838-3059
dmccr@us.ibm.com                                        T/L   678-3059


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Page table sharing
  2002-09-25 18:40 ` Dave McCracken
@ 2002-09-26 18:20   ` Gerold J. Wucherpfennig
  2002-10-01 13:08   ` Daniel Phillips
  1 sibling, 0 replies; 6+ messages in thread
From: Gerold J. Wucherpfennig @ 2002-09-26 18:20 UTC (permalink / raw)
  To: linux-kernel

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wednesday September 25 2002 20:40, Dave McCracken <dmccr@us.ibm.com> wrote:
> --On Wednesday, September 25, 2002 20:12:36 +0200 "Gerold J. Wucherpfennig"
> <gjwucherpfennig@gmx.net> wrote:
...
> > Some comments from Daniel Phillips or Dave McCracken?
>
> I'm working on it.  I sent out a patch to the mm list a few weeks ago, but
> it didn't have the locking right.  I'm in the proces of finishing an
> improved version with new locking.  I'll send a snapshot of it out when I
> can make it stop oopsing :)
>
...

Oh, really geat :-)

I hope you will get it stable till Halloween...
...hopefully somebody will help you.
Page table sharing seems to be very useful,
but I'm no expert. I only read about it at LWN
and Kernel Traffic.


Gerold

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE9k0/Y53zyw+ONqrERAkauAJ0b3ztvgORi6c7rgSbhD3IIQh5v3ACdG2ou
Ds2vW/vuXHlk1NU/Hau3NNk=
=FgOz
-----END PGP SIGNATURE-----


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Page table sharing
  2002-09-25 18:40 ` Dave McCracken
  2002-09-26 18:20   ` Gerold J. Wucherpfennig
@ 2002-10-01 13:08   ` Daniel Phillips
  2002-10-01 16:06     ` Martin J. Bligh
  1 sibling, 1 reply; 6+ messages in thread
From: Daniel Phillips @ 2002-10-01 13:08 UTC (permalink / raw)
  To: Dave McCracken, Gerold J. Wucherpfennig, linux-kernel

On Wednesday 25 September 2002 20:40, Dave McCracken wrote:
> --On Wednesday, September 25, 2002 20:12:36 +0200 "Gerold J. Wucherpfennig"
> <gjwucherpfennig@gmx.net> wrote:
> 
> > What about page table sharing? Does anybody still care about this?
> > 
> > The patch from Daniel Phillips
> > (http://www.geocrawler.com/mail/msg.php3?msg_id=7855063&list=35)
> > is a few month old and I can't see any progress.
> > 
> > Sorry, I'm not a kernel expert, so I can't help.
> > But page table sharing is still listed as betaware at the
> > Linux Kernel 2.5 Status page (http://kernelnewbies.org/status/latest.html)
> > and Page Table sharing isn't marked post Halloween.
> > 
> > Some comments from Daniel Phillips or Dave McCracken?
> 
> I'm working on it.  I sent out a patch to the mm list a few weeks ago, but
> it didn't have the locking right.  I'm in the proces of finishing an
> improved version with new locking.  I'll send a snapshot of it out when I
> can make it stop oopsing :)

Hi Dave,

I'm not sure how relevant page table sharing has to the halloween deadline
since it's not a feature per se, just an optimization.   It has more to do
with getting numa ia32 boxes to survive, so it's an ideal out-of-tree patch.

Anyway, I feel your pain - debugging these deep VM hacks can get very weird.
I just got back home, I'll have a read through your latest.

-- 
Daniel

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Page table sharing
  2002-10-01 13:08   ` Daniel Phillips
@ 2002-10-01 16:06     ` Martin J. Bligh
  2002-10-01 16:28       ` Daniel Phillips
  0 siblings, 1 reply; 6+ messages in thread
From: Martin J. Bligh @ 2002-10-01 16:06 UTC (permalink / raw)
  To: Daniel Phillips, Dave McCracken, Gerold J. Wucherpfennig, linux-kernel

> I'm not sure how relevant page table sharing has to the halloween 
> deadline since it's not a feature per se, just an optimization.   
> It has more to do with getting numa ia32 boxes to survive, so it's 
> an ideal out-of-tree patch.

Any large 32 bit box with significant numbers of processes will need 
it to cope with the bloat that rmap introduced - this has nothing to
do with NUMA (some apps may be saved by large pages, some not). 
Avoiding hangs from ZONE_NORMAL oom is not an "optimisation", and I 
doubt optimisations involving major VM changes would be very welcome
after the freeze. This is something we need to get working ASAP ...

M.


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Page table sharing
  2002-10-01 16:06     ` Martin J. Bligh
@ 2002-10-01 16:28       ` Daniel Phillips
  0 siblings, 0 replies; 6+ messages in thread
From: Daniel Phillips @ 2002-10-01 16:28 UTC (permalink / raw)
  To: Martin J. Bligh, Dave McCracken, Gerold J. Wucherpfennig, linux-kernel

On Tuesday 01 October 2002 18:06, Martin J. Bligh wrote:
> > I'm not sure how relevant page table sharing has to the halloween 
> > deadline since it's not a feature per se, just an optimization.   
> > It has more to do with getting numa ia32 boxes to survive, so it's 
> > an ideal out-of-tree patch.
> 
> Any large 32 bit box with significant numbers of processes will need 
> it to cope with the bloat that rmap introduced - this has nothing to
> do with NUMA (some apps may be saved by large pages, some not). 
> Avoiding hangs from ZONE_NORMAL oom is not an "optimisation", and I 
> doubt optimisations involving major VM changes would be very welcome
> after the freeze. This is something we need to get working ASAP ...

Any reason why page table swapping wouldn't deliver just as much bang
for the buck?  Same argument re VM changes.

Anyway, I don't necessarily buy the VM change argument.  I've already
established that turning sharing off either at compile time or run
time is trivial, so there is no element of risk.  I'd suggest: take
a deep breath, relax, go slow and get it right.

What we should be coding right now is the patch Linus already asked
for, to recover page table pages as soon as possible instead of in
clear_page_tables.  With that in place the locking for shared page
tables gets a lot nicer, as Linus showed.

-- 
Daniel

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2002-10-01 16:35 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-09-25 18:12 Page table sharing Gerold J. Wucherpfennig
2002-09-25 18:40 ` Dave McCracken
2002-09-26 18:20   ` Gerold J. Wucherpfennig
2002-10-01 13:08   ` Daniel Phillips
2002-10-01 16:06     ` Martin J. Bligh
2002-10-01 16:28       ` Daniel Phillips

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).