On Tuesday 20 March 2007 07:00, Nick Piggin wrote:
> On Mon, Mar 19, 2007 at 09:44:28PM +0100, Blaisorblade wrote:
> > On Sunday 18 March 2007 03:50, Nick Piggin wrote:
> > > > > Yes, I believe that is the case, however I wonder if that is going
> > > > > to be a problem for you to distinguish betw
On Mon, Mar 19, 2007 at 09:44:28PM +0100, Blaisorblade wrote:
> On Sunday 18 March 2007 03:50, Nick Piggin wrote:
> > > >
> > > > Yes, I believe that is the case, however I wonder if that is going to
> > > > be a problem for you to distinguish between write faults for clean
> > > > writable ptes, a
On Sunday 18 March 2007 03:50, Nick Piggin wrote:
> On Sat, Mar 17, 2007 at 01:17:00PM +0100, Blaisorblade wrote:
> > On Tuesday 13 March 2007 02:19, Nick Piggin wrote:
> > > On Tue, Mar 13, 2007 at 12:01:13AM +0100, Blaisorblade wrote:
> > > > On Wednesday 07 March 2007 11:02, Nick Piggin wrote:
>
On Sun, Mar 18, 2007 at 03:50:10AM +0100, Nick Piggin wrote:
> Yes, that should be the case. So would this mean that nonlinear protections
> don't work on regular files? I guess that's OK if Oracle and UML both use
> tmpfs/shm?
Sometimes ramfs is also used in the Oracle case. I presume that's even
On Sun, Mar 18, 2007 at 03:50:10AM +0100, Nick Piggin wrote:
> Yes, that should be the case. So would this mean that nonlinear protections
> don't work on regular files? I guess that's OK if Oracle and UML both use
> tmpfs/shm?
It's OK for UML.
Jeff
--
Work email - jdike at linu
On Sat, Mar 17, 2007 at 01:17:00PM +0100, Blaisorblade wrote:
> On Tuesday 13 March 2007 02:19, Nick Piggin wrote:
> > On Tue, Mar 13, 2007 at 12:01:13AM +0100, Blaisorblade wrote:
> > > On Wednesday 07 March 2007 11:02, Nick Piggin wrote:
> > > > > Yeah, tmpfs/shm segs are what I was thinking abou
On Tuesday 13 March 2007 02:19, Nick Piggin wrote:
> On Tue, Mar 13, 2007 at 12:01:13AM +0100, Blaisorblade wrote:
> > On Wednesday 07 March 2007 11:02, Nick Piggin wrote:
> > > > Yeah, tmpfs/shm segs are what I was thinking about. If UML can live
> > > > with that as well, then I think it might be
On Tue, Mar 13, 2007 at 12:01:13AM +0100, Blaisorblade wrote:
> On Wednesday 07 March 2007 11:02, Nick Piggin wrote:
> > >
> > > Yeah, tmpfs/shm segs are what I was thinking about. If UML can live with
> > > that as well, then I think it might be a good option.
> >
> > Oh, hmm if you can trunca
On Wednesday 07 March 2007 11:02, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 10:49:47AM +0100, Nick Piggin wrote:
> > On Wed, Mar 07, 2007 at 01:44:20AM -0800, Bill Irwin wrote:
> > > On Wed, Mar 07, 2007 at 10:28:21AM +0100, Nick Piggin wrote:
> > > > Depending on whether anyone wants it, and wh
On Wednesday 07 March 2007 10:44, Bill Irwin wrote:
> On Wed, Mar 07, 2007 at 10:28:21AM +0100, Nick Piggin wrote:
> > Depending on whether anyone wants it, and what features they want, we
> > could emulate the old syscall, and make a new restricted one which is
> > much less intrusive.
> > For exa
On Wed, Mar 07, 2007 at 02:52:12PM +0100, Peter Zijlstra wrote:
> > Well I don't think UML uses nonlinear yet anyway, does it? Can they
> > make do with restricting nonlinear to mlocked vmas, I wonder? Probably
> > not.
>
> I think it does, but lets ask, Jeff?
Nope, UML needs to be able to change
On Wed, Mar 07, 2007 at 03:34:27PM +0100, Peter Zijlstra wrote:
> On Wed, 2007-03-07 at 14:52 +0100, Peter Zijlstra wrote:
>
> > True. We could even guesstimate the nonlinear dirty pages by subtracting
> > the result of page_mkclean() from page_mapcount() and force an
> > msync(MS_ASYNC) on said m
On Wed, Mar 07, 2007 at 02:53:07PM +0100, Miklos Szeredi wrote:
> > > msync() might never get called and then we're back with the old
> > > behaviour where we can surprise the VM with a ton of dirty pages.
> >
> > But we're root. With your patch, root *can't* do nonlinear writeback
> > well. Ever.
On Wed, 2007-03-07 at 14:52 +0100, Peter Zijlstra wrote:
> True. We could even guesstimate the nonlinear dirty pages by subtracting
> the result of page_mkclean() from page_mapcount() and force an
> msync(MS_ASYNC) on said mapping (or all (nonlinear) mappings of the
> related file) when some thres
> > Well I don't think UML uses nonlinear yet anyway, does it? Can they
> > make do with restricting nonlinear to mlocked vmas, I wonder? Probably
> > not.
>
> I think it does, but lets ask, Jeff?
Looks like it doesn't:
$ grep -r remap_file_pages arch/um/
$
Miklos
-
To unsubscribe from this lis
> On Wed, Mar 07, 2007 at 02:19:22PM +0100, Peter Zijlstra wrote:
> > On Wed, 2007-03-07 at 14:08 +0100, Nick Piggin wrote:
> >
> > > > > The thing is, I don't think anybody who uses these things cares
> > > > > about any of the 'problems' you want to fix, do they? We are
> > > > > interested in d
On Wed, 2007-03-07 at 14:36 +0100, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 02:19:22PM +0100, Peter Zijlstra wrote:
> > On Wed, 2007-03-07 at 14:08 +0100, Nick Piggin wrote:
> >
> > > > > The thing is, I don't think anybody who uses these things cares
> > > > > about any of the 'problems' you
On Wed, Mar 07, 2007 at 02:19:22PM +0100, Peter Zijlstra wrote:
> On Wed, 2007-03-07 at 14:08 +0100, Nick Piggin wrote:
>
> > > > The thing is, I don't think anybody who uses these things cares
> > > > about any of the 'problems' you want to fix, do they? We are
> > > > interested in dirty pages o
On Wed, 2007-03-07 at 14:08 +0100, Nick Piggin wrote:
> > > The thing is, I don't think anybody who uses these things cares
> > > about any of the 'problems' you want to fix, do they? We are
> > > interested in dirty pages only for the correctness issue, rather
> > > than performance. Same as recl
On Wed, Mar 07, 2007 at 01:41:26PM +0100, Peter Zijlstra wrote:
> On Wed, 2007-03-07 at 13:17 +0100, Nick Piggin wrote:
>
> > > Tracking these ranges on a per-vma basis would avoid taking the mm wide
> > > mmap_sem and so would be cheaper than regular vmas.
> > >
> > > Would that still be too exp
On Wed, 2007-03-07 at 13:17 +0100, Nick Piggin wrote:
> > Tracking these ranges on a per-vma basis would avoid taking the mm wide
> > mmap_sem and so would be cheaper than regular vmas.
> >
> > Would that still be too expensive?
>
> Well you can today remap N pages in a file, arbitrarily for
> s
On Wed, Mar 07, 2007 at 04:22:24AM -0800, Bill Irwin wrote:
> On Wed, Mar 07, 2007 at 11:47:42AM +0100, Peter Zijlstra wrote:
> >> Well, now they don't, but it could be done or even exploited as a DoS.
>
> On Wed, Mar 07, 2007 at 12:00:36PM +0100, Nick Piggin wrote:
> > But so could nonlinear page
On Wed, Mar 07, 2007 at 11:47:42AM +0100, Peter Zijlstra wrote:
>> Well, now they don't, but it could be done or even exploited as a DoS.
On Wed, Mar 07, 2007 at 12:00:36PM +0100, Nick Piggin wrote:
> But so could nonlinear page reclaim. I think we need to restrict nonlinear
> mappings to root if
On Wed, Mar 07, 2007 at 12:48:06PM +0100, Peter Zijlstra wrote:
> On Wed, 2007-03-07 at 12:00 +0100, Nick Piggin wrote:
> > On Wed, Mar 07, 2007 at 11:47:42AM +0100, Peter Zijlstra wrote:
> > > On Wed, 2007-03-07 at 11:38 +0100, Nick Piggin wrote:
> > >
> > > > > > There are real users who want th
On Wed, 2007-03-07 at 12:00 +0100, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 11:47:42AM +0100, Peter Zijlstra wrote:
> > On Wed, 2007-03-07 at 11:38 +0100, Nick Piggin wrote:
> >
> > > > > There are real users who want these fast, though.
> > > >
> > > > Yeah, why don't we have a tree per nonl
On Wed, Mar 07, 2007 at 11:47:42AM +0100, Peter Zijlstra wrote:
> On Wed, 2007-03-07 at 11:38 +0100, Nick Piggin wrote:
>
> > > > There are real users who want these fast, though.
> > >
> > > Yeah, why don't we have a tree per nonlinear vma to find these pages?
> > >
> > > wli mentions shadow pa
On Wed, 2007-03-07 at 11:38 +0100, Nick Piggin wrote:
> > > There are real users who want these fast, though.
> >
> > Yeah, why don't we have a tree per nonlinear vma to find these pages?
> >
> > wli mentions shadow page tables..
>
> We could do something more efficient, but I thought that half
On Wed, 2007-03-07 at 11:17 +0100, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 11:05:48AM +0100, Benjamin Herrenschmidt wrote:
> >
> > > > NOPAGE_REFAULT is removed. This should be implemented with ->fault, and
> > > > no users have hit mainline yet.
> > >
> > > Did benh agree with that?
> >
>
On Wed, Mar 07, 2007 at 11:24:45AM +0100, Peter Zijlstra wrote:
> On Wed, 2007-03-07 at 11:21 +0100, Nick Piggin wrote:
> > On Wed, Mar 07, 2007 at 11:13:20AM +0100, Miklos Szeredi wrote:
> > > > *sigh* yes was looking at all that code, thats gonna be darn slow
> > > > though, but I'll whip up a pa
On Wed, 2007-03-07 at 11:21 +0100, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 11:13:20AM +0100, Miklos Szeredi wrote:
> > > *sigh* yes was looking at all that code, thats gonna be darn slow
> > > though, but I'll whip up a patch.
> >
> > Well, if it's going to be darn slow, maybe it's better to
On Wed, Mar 07, 2007 at 11:13:20AM +0100, Miklos Szeredi wrote:
> > *sigh* yes was looking at all that code, thats gonna be darn slow
> > though, but I'll whip up a patch.
>
> Well, if it's going to be darn slow, maybe it's better to go with
> mingo's plan on emulating nonlinear vmas with linear o
On Wed, Mar 07, 2007 at 11:05:48AM +0100, Benjamin Herrenschmidt wrote:
>
> > > NOPAGE_REFAULT is removed. This should be implemented with ->fault, and
> > > no users have hit mainline yet.
> >
> > Did benh agree with that?
>
> I won't use NOPAGE_REFAULT, I use NOPFN_REFAULT and that has hit
> m
> *sigh* yes was looking at all that code, thats gonna be darn slow
> though, but I'll whip up a patch.
Well, if it's going to be darn slow, maybe it's better to go with
mingo's plan on emulating nonlinear vmas with linear ones. That'll be
darn slow as well, but at least it will be much less comp
> > NOPAGE_REFAULT is removed. This should be implemented with ->fault, and
> > no users have hit mainline yet.
>
> Did benh agree with that?
I won't use NOPAGE_REFAULT, I use NOPFN_REFAULT and that has hit
mainline. I will switch to ->fault when I have time to adapt the code,
in the meantime, N
On Wed, 7 Mar 2007 01:29:03 -0800 Bill Irwin <[EMAIL PROTECTED]> wrote:
>> Guess what major real-life application not only uses nonlinear daily
>> but would even be very happy to see it extended with non-vma-creating
>> protections and more?
On Wed, Mar 07, 2007 at 01:39:42AM -0800, Andrew Morton
On Wed, 2007-03-07 at 11:04 +0100, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 10:45:03AM +0100, Nick Piggin wrote:
> > On Wed, Mar 07, 2007 at 10:32:22AM +0100, Peter Zijlstra wrote:
> > >
> > > Can recollect as much, I modelled it after page_referenced() and can't
> > > find any VM_NONLINEAR sp
On Wed, Mar 07, 2007 at 10:45:03AM +0100, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 10:32:22AM +0100, Peter Zijlstra wrote:
> >
> > Can recollect as much, I modelled it after page_referenced() and can't
> > find any VM_NONLINEAR specific code in there either.
> >
> > Will have a hard look, but
On Wed, Mar 07, 2007 at 10:49:47AM +0100, Nick Piggin wrote:
> On Wed, Mar 07, 2007 at 01:44:20AM -0800, Bill Irwin wrote:
> > On Wed, Mar 07, 2007 at 10:28:21AM +0100, Nick Piggin wrote:
> > > Depending on whether anyone wants it, and what features they want, we
> > > could emulate the old syscall
On Wed, Mar 07, 2007 at 10:22:52AM +0100, Ingo Molnar wrote:
>>> ok. What do you think about the sys_remap_file_pages_prot() thing that
>>> Paolo has done in a nicely split up form - does that complicate things
>>> in any fundamental way? That is what is useful to UML.
* Bill Irwin <[EMAIL PROTE
On Wed, Mar 07, 2007 at 10:22:52AM +0100, Ingo Molnar wrote:
>
> * Nick Piggin <[EMAIL PROTECTED]> wrote:
>
> > After these patches, I don't think there is too much burden. The main
> > thing left really is just the objrmap stuff, but that is just handled
> > with a minimal 'dumb' algorithm tha
On Wed, Mar 07, 2007 at 01:44:20AM -0800, Bill Irwin wrote:
> On Wed, Mar 07, 2007 at 10:28:21AM +0100, Nick Piggin wrote:
> > Depending on whether anyone wants it, and what features they want, we
> > could emulate the old syscall, and make a new restricted one which is
> > much less intrusive.
> >
On Wed, Mar 07, 2007 at 10:28:21AM +0100, Nick Piggin wrote:
> Depending on whether anyone wants it, and what features they want, we
> could emulate the old syscall, and make a new restricted one which is
> much less intrusive.
> For example, if we can operate only on MAP_ANONYMOUS memory and speci
On Wed, Mar 07, 2007 at 10:32:22AM +0100, Peter Zijlstra wrote:
> On Wed, 2007-03-07 at 01:07 -0800, Andrew Morton wrote:
> > On Wed, 07 Mar 2007 09:51:57 +0100 Miklos Szeredi <[EMAIL PROTECTED]> wrote:
> >
> > > > > Dirty page accounting doesn't work either on
> > > > > non-linear mappings
> > >
On Wed, Mar 07, 2007 at 01:26:38AM -0800, Andrew Morton wrote:
> On Wed, 7 Mar 2007 10:18:23 +0100 Nick Piggin <[EMAIL PROTECTED]> wrote:
>
> >
> > msync breakage is bad, but otherwise I don't know that we care about
> > dirty page writeout efficiency.
>
> Well. We made so many changes to suppo
On Wed, 7 Mar 2007 01:29:03 -0800 Bill Irwin <[EMAIL PROTECTED]> wrote:
> On Wed, 7 Mar 2007 09:27:55 +0100 Ingo Molnar <[EMAIL PROTECTED]> wrote:
> >> btw., if we decide that nonlinear isnt worth the continuing maintainance
> >> pain, we could internally implement/emulate sys_remap_file_pages()
* Bill Irwin <[EMAIL PROTECTED]> wrote:
> * Nick Piggin <[EMAIL PROTECTED]> wrote:
> >> After these patches, I don't think there is too much burden. The main
> >> thing left really is just the objrmap stuff, but that is just handled
> >> with a minimal 'dumb' algorithm that doesn't cost much.
>
On Wed, 2007-03-07 at 01:07 -0800, Andrew Morton wrote:
> On Wed, 07 Mar 2007 09:51:57 +0100 Miklos Szeredi <[EMAIL PROTECTED]> wrote:
>
> > > > Dirty page accounting doesn't work either on
> > > > non-linear mappings
> > >
> > > It doesn't? Confused - these things don't have anything to do with
* Nick Piggin <[EMAIL PROTECTED]> wrote:
>> After these patches, I don't think there is too much burden. The main
>> thing left really is just the objrmap stuff, but that is just handled
>> with a minimal 'dumb' algorithm that doesn't cost much.
On Wed, Mar 07, 2007 at 10:22:52AM +0100, Ingo Mol
> > But I think we discovered that those msync changes are bogus anyway
> > becuase there is a small race window where pte could be dirtied without
> > page being set dirty?
>
> Dunno, I don't recall that. We dirty the page before the pte...
That's the one I just submitted a fix for ;)
http:/
On Wed, 7 Mar 2007 09:27:55 +0100 Ingo Molnar <[EMAIL PROTECTED]> wrote:
>> btw., if we decide that nonlinear isnt worth the continuing maintainance
>> pain, we could internally implement/emulate sys_remap_file_pages() via a
>> call to mremap() and essentially deprecate it, without breaking the A
On Wed, 7 Mar 2007 10:18:23 +0100 Nick Piggin <[EMAIL PROTECTED]> wrote:
> On Wed, Mar 07, 2007 at 01:07:56AM -0800, Andrew Morton wrote:
> > On Wed, 07 Mar 2007 09:51:57 +0100 Miklos Szeredi <[EMAIL PROTECTED]> wrote:
> >
> > > > > Dirty page accounting doesn't work either on
> > > > > non-linea
On Wed, Mar 07, 2007 at 09:53:23AM +0100, Ingo Molnar wrote:
>
> * Andrew Morton <[EMAIL PROTECTED]> wrote:
>
> > > btw., if we decide that nonlinear isnt worth the continuing
> > > maintainance pain, we could internally implement/emulate
> > > sys_remap_file_pages() via a call to mremap() and
> >
> > Look in page_mkclean(). Where does it handle non-linear mappings?
> >
>
> OK, I'd forgotten about that. It won't break dirty memory accounting,
> but it'll potentially break dirty memory balancing.
>
> If we have the wrong page (due to nonlinear), page_check_address() will
> fail and
* Nick Piggin <[EMAIL PROTECTED]> wrote:
> After these patches, I don't think there is too much burden. The main
> thing left really is just the objrmap stuff, but that is just handled
> with a minimal 'dumb' algorithm that doesn't cost much.
ok. What do you think about the sys_remap_file_page
On Wed, Mar 07, 2007 at 01:07:56AM -0800, Andrew Morton wrote:
> On Wed, 07 Mar 2007 09:51:57 +0100 Miklos Szeredi <[EMAIL PROTECTED]> wrote:
>
> > > > Dirty page accounting doesn't work either on
> > > > non-linear mappings
> > >
> > > It doesn't? Confused - these things don't have anything to
On Wed, Mar 07, 2007 at 09:59:44AM +0100, Nick Piggin wrote:
> Apart from a handful of trivial if (pte_file()) cases throughout mm/,
> our maintainance burden basically now amounts to the following patch.
> Even the rmap.c change looks bigger than it is because I split out
> the nonlinear unmapping
On Wed, 07 Mar 2007 09:51:57 +0100 Miklos Szeredi <[EMAIL PROTECTED]> wrote:
> > > Dirty page accounting doesn't work either on
> > > non-linear mappings
> >
> > It doesn't? Confused - these things don't have anything to do with each
> > other do they?
>
> Look in page_mkclean(). Where does it
On Wed, Mar 07, 2007 at 09:27:55AM +0100, Ingo Molnar wrote:
>
> * Nick Piggin <[EMAIL PROTECTED]> wrote:
>
> > Then 4,5,6 is the fault/nonlinear rewrite, take it or leave it. I
> > thought you would have liked the patches...
>
> btw., if we decide that nonlinear isnt worth the continuing maint
> > Dirty page accounting doesn't work either on
> > non-linear mappings
>
> It doesn't? Confused - these things don't have anything to do with each
> other do they?
Look in page_mkclean(). Where does it handle non-linear mappings?
Miklos
-
To unsubscribe from this list: send the line "unsubsc
* Andrew Morton <[EMAIL PROTECTED]> wrote:
> > btw., if we decide that nonlinear isnt worth the continuing
> > maintainance pain, we could internally implement/emulate
> > sys_remap_file_pages() via a call to mremap() and essentially
> > deprecate it, without breaking the ABI - and remove all
On Wed, 07 Mar 2007 09:38:34 +0100 Miklos Szeredi <[EMAIL PROTECTED]> wrote:
> Dirty page accounting doesn't work either on
> non-linear mappings
It doesn't? Confused - these things don't have anything to do with each
other do they?
-
To unsubscribe from this list: send the line "unsubscribe lin
> > If it doesn't look very impressive, it could be because it leaves all
> > the old crud around for backwards compatibility (the worst offenders
> > are removed in patch 6/6).
> >
> > If you look at the patchset as a whole, it removes about 250 lines,
> > mostly of (non trivial) duplicated co
On Wed, 7 Mar 2007 09:27:55 +0100 Ingo Molnar <[EMAIL PROTECTED]> wrote:
>
> * Nick Piggin <[EMAIL PROTECTED]> wrote:
>
> > If it doesn't look very impressive, it could be because it leaves all
> > the old crud around for backwards compatibility (the worst offenders
> > are removed in patch 6/
* Nick Piggin <[EMAIL PROTECTED]> wrote:
> If it doesn't look very impressive, it could be because it leaves all
> the old crud around for backwards compatibility (the worst offenders
> are removed in patch 6/6).
>
> If you look at the patchset as a whole, it removes about 250 lines,
> mostly
On Wed, Mar 07, 2007 at 08:08:53AM +0100, Nick Piggin wrote:
> On Tue, Mar 06, 2007 at 10:51:01PM -0800, Andrew Morton wrote:
>
> > This patch seems to churn things around an awful lot for minimal benefit.
>
> Well it fixes the whole design of the nonlinear fault path.
If it doesn't look very im
On Tue, Mar 06, 2007 at 10:51:01PM -0800, Andrew Morton wrote:
> Does anybody really pass a NULL `type' arg into filemap_nopage()?
The major vs. minor fault accounting patch that introduced the argument
didn't make non-NULL type arguments a requirement. It's essentially an
optional second return v
On Tue, Mar 06, 2007 at 10:51:01PM -0800, Andrew Morton wrote:
> On Wed, 21 Feb 2007 05:50:17 +0100 (CET) Nick Piggin <[EMAIL PROTECTED]>
> wrote:
>
> > Nonlinear mappings are (AFAIKS) simply a virtual memory concept that
> > encodes the virtual address -> file offset differently from linear
> >
On Wed, 21 Feb 2007 05:50:17 +0100 (CET) Nick Piggin <[EMAIL PROTECTED]> wrote:
> Nonlinear mappings are (AFAIKS) simply a virtual memory concept that
> encodes the virtual address -> file offset differently from linear
> mappings.
>
> I can't see why the filesystem/pagecache code should need to
Nonlinear mappings are (AFAIKS) simply a virtual memory concept that
encodes the virtual address -> file offset differently from linear
mappings.
I can't see why the filesystem/pagecache code should need to know anything
about it, except for the fact that the ->nopage handler didn't quite pass
dow
69 matches
Mail list logo