LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Andrew Morton <akpm@linux-foundation.org>
To: Nick Piggin <npiggin@suse.de>
Cc: Linux Filesystems <linux-fsdevel@vger.kernel.org>,
	Linux Kernel <linux-kernel@vger.kernel.org>,
	Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [rfc][patch 0/3] a faster buffered write deadlock fix?
Date: Fri, 9 Feb 2007 02:52:49 -0800	[thread overview]
Message-ID: <20070209025249.0a87a435.akpm@linux-foundation.org> (raw)
In-Reply-To: <20070209103258.GC14398@wotan.suse.de>

On Fri, 9 Feb 2007 11:32:58 +0100 Nick Piggin <npiggin@suse.de> wrote:

> On Fri, Feb 09, 2007 at 02:09:54AM -0800, Andrew Morton wrote:
> > On Fri, 9 Feb 2007 10:54:05 +0100 Nick Piggin <npiggin@suse.de> wrote:
> > 
> > > 
> > > That's still got a deadlock,
> > 
> > It does?
> 
> Yes, PG_lock vs mm->mmap_sem.

Where?  It requires that someone hold mmap_sem for writing as well as a
page lock (in an order which would require some thought).  Do we ever do
that?

> > >  and also it doesn't work if we want to lock
> > > the page when performing a minor fault (which I want to fix fault vs
> > > invalidate),
> > 
> > It's hard to discuss this without a description of what you want to fix
> > there, and a description of how you plan to fix it.
> 
> http://marc.theaimsgroup.com/?l=linux-mm&m=116865911432667&w=2

mutter.

Could perhaps fix that by running ClearPageUptodate in invalidate, thus
forcing the pagefault code to take the page lock (which we already hold).

That does mean that we'll fleetingly have a non-uptodate page in pagetables
which is a bit nasty.

Or, probably better, we could add a new page flag (heh) telling nopage that
it needs to lock the page even if it's uptodate.

> > > and also assumes nobody's ->nopage locks the page or
> > > requires any resources that are held by prepare_write (something not
> > > immediately clear to me with the cluster filesystems, at least).
> > 
> > The nopage handler is filemap_nopage().  Are there any exceptions to that?
> 
> OCFS2 and GFS2.

So the rule is that ->nopage handlers against pagecache mustn't lock the
page if it's already uptodate.  That's OK.  But it might conflict with the
above invalidate proposal.

Gad.  ocfs2_nopage() diddles with signals.


> > > But that all becomes legacy path, so do we really care? Supposing fs
> > > maintainers like perform_write, then after the main ones have implementations
> > > we could switch over to the slow-but-correct prepare_write legacy path.
> > > Or we could leave it, or we could use Linus's slightly-less-buggy scheme...
> > > by that point I expect I'd be sick of arguing about it ;)
> > 
> > It's worth "arguing" about.  This is write().  What matters more??
> 
> That's the legacy path that uses prepare/commit (ie. supposing that all
> major filesystems did get converted to perform_write).

We'll never, ever, ever update and test all filesytems.  What you're
calling "legacy" code will be there for all time.

I haven't had time to look at the perform_write stuff yet.

> Of course I would still want my correct-but-slow version in that case,
> but I just wouldn't care to argue if you still wanted to keep it fast.

This is write().  We just cannot go and double-copy all the memory or take
mmap_sem and do a full pagetable walk in there.  It just means that we
haven't found a suitable solution yet.


  reply	other threads:[~2007-02-09 10:53 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-02-08 13:07 Nick Piggin
2007-02-08 13:07 ` [patch 1/3] fs: add an iovec iterator Nick Piggin
2007-02-08 19:49   ` Christoph Hellwig
2007-02-09  1:46     ` Nick Piggin
2007-02-09  2:03       ` Nate Diller
2007-02-09  3:31         ` Nick Piggin
2007-02-09 17:28           ` Zach Brown
2007-03-09 10:40         ` Christoph Hellwig
2007-02-08 23:04   ` Mark Fasheh
2007-02-08 13:07 ` [patch 2/3] fs: introduce perform_write aop Nick Piggin
2007-03-09 10:39   ` Christoph Hellwig
2007-03-09 12:52     ` Nick Piggin
2007-03-09 22:01       ` Anton Altaparmakov
2007-03-09 23:33     ` Mark Fasheh
2007-03-10  9:25       ` Christoph Hellwig
2007-03-12  2:13         ` Mark Fasheh
2007-03-14 13:30         ` Nick Piggin
2007-03-14 15:17           ` Christoph Hellwig
2007-02-08 13:07 ` [patch 3/3] ext2: use " Nick Piggin
2007-02-08 14:47   ` Dmitriy Monakhov
2007-02-09 19:14   ` Andrew Morton
2007-02-09 19:45     ` Andrew Morton
2007-02-10  1:34       ` Nick Piggin
2007-02-10  1:50         ` Andrew Morton
2007-02-09  0:38 ` [rfc][patch 0/3] a faster buffered write deadlock fix? Mark Fasheh
2007-02-09  2:04   ` Nick Piggin
2007-02-09  8:41 ` Andrew Morton
2007-02-09  9:54   ` Nick Piggin
2007-02-09 10:09     ` Andrew Morton
2007-02-09 10:32       ` Nick Piggin
2007-02-09 10:52         ` Andrew Morton [this message]
2007-02-09 11:31           ` Nick Piggin
2007-02-09 11:46             ` Andrew Morton
2007-02-09 12:11               ` Nick Piggin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20070209025249.0a87a435.akpm@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=npiggin@suse.de \
    --cc=torvalds@linux-foundation.org \
    --subject='Re: [rfc][patch 0/3] a faster buffered write deadlock fix?' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

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).