ext4_da_block_invalidatepages() question

ext4_da_block_invalidatepages() question

[Wu Fengguang]

Hi Jan,

I noticed that ext4_da_block_invalidatepages() does pagevec_lookup()
without pagevec_release()/put_page(). Is that OK?

Thanks,
Fengguang

Re: ext4_da_block_invalidatepages() question

[Jan Kara]

[-- Attachment #1: Type: text/plain, Size: 497 bytes --]

  Hi,

On Tue 26-01-10 21:36:08, Wu Fengguang wrote:
> I noticed that ext4_da_block_invalidatepages() does pagevec_lookup()
> without pagevec_release()/put_page(). Is that OK?
  Yes, the function looks buggy. Luckily, it is called only in case we are
not able to allocate space for delay-allocated data which is a bug on its
own. So people should never hit it.
  Attached patch should fix the issue. Ted, will you merge it please?
Thanks.

								Honza

-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR

[-- Attachment #2: 0001-ext4-Release-page-references-acquired-in-ext4_da_blo.patch --]
[-- Type: text/x-patch, Size: 1081 bytes --]

>From 47085f1ac03eaca9e4d7a5f8f1e40e87d3879512 Mon Sep 17 00:00:00 2001
From: Jan Kara <jack@suse.cz>
Date: Tue, 26 Jan 2010 16:15:19 +0100
Subject: [PATCH] ext4: Release page references acquired in ext4_da_block_invalidatepages

We forget to release page references we acquire in
ext4_da_block_invalidatepages.  Luckily, this function gets called only if we
are not able to allocate blocks for delay-allocated data so that function
should better never be called.

Also cleanup handling of index variable.

Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jan Kara <jack@suse.cz>
---
 fs/ext4/inode.c |    7 +++----
 1 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index c818972..1680007 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -2127,17 +2127,16 @@ static void ext4_da_block_invalidatepages(struct mpage_da_data *mpd,
 			break;
 		for (i = 0; i < nr_pages; i++) {
 			struct page *page = pvec.pages[i];
-			index = page->index;
-			if (index > end)
+			if (page->index > end)
 				break;
-			index++;

Re: ext4_da_block_invalidatepages() question

[Wu Fengguang]

On Tue, Jan 26, 2010 at 08:32:22AM -0700, Jan Kara wrote:

> @@ -2127,17 +2127,16 @@ static void ext4_da_block_invalidatepages(struct mpage_da_data *mpd,
>  			break;
>  		for (i = 0; i < nr_pages; i++) {
>  			struct page *page = pvec.pages[i];
> -			index = page->index;
> -			if (index > end)
> +			if (page->index > end)
>  				break;
> -			index++;
> -
>  			BUG_ON(!PageLocked(page));
>  			BUG_ON(PageWriteback(page));
>  			block_invalidatepage(page, 0);
>  			ClearPageUptodate(page);
>  			unlock_page(page);
>  		}
> +		index = pvec.pages[nr_pages - 1]->index + 1;
> +		pagevec_release(&pvec);
>  	}
>  	return;
>  }

The patch includes a cleanup and a bug fix, both looks OK to me.
But if we can split it, the bug fix would be good candidate for
the stable kernel?

Thanks,
Fengguang

Re: ext4_da_block_invalidatepages() question

[Jan Kara]

On Wed 27-01-10 09:53:39, Wu Fengguang wrote:
> On Tue, Jan 26, 2010 at 08:32:22AM -0700, Jan Kara wrote:
> 
> > @@ -2127,17 +2127,16 @@ static void ext4_da_block_invalidatepages(struct mpage_da_data *mpd,
> >  			break;
> >  		for (i = 0; i < nr_pages; i++) {
> >  			struct page *page = pvec.pages[i];
> > -			index = page->index;
> > -			if (index > end)
> > +			if (page->index > end)
> >  				break;
> > -			index++;
> > -
> >  			BUG_ON(!PageLocked(page));
> >  			BUG_ON(PageWriteback(page));
> >  			block_invalidatepage(page, 0);
> >  			ClearPageUptodate(page);
> >  			unlock_page(page);
> >  		}
> > +		index = pvec.pages[nr_pages - 1]->index + 1;
> > +		pagevec_release(&pvec);
> >  	}
> >  	return;
> >  }
> 
> The patch includes a cleanup and a bug fix, both looks OK to me.
> But if we can split it, the bug fix would be good candidate for
> the stable kernel?
  I don't think we want to push this to -stable kernel. As I wrote, this
code is called only in case user is going to loose written data which is a
bug on it's own so loosing a few page references is nothing compared to
that.

									Honza
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR

Re: ext4_da_block_invalidatepages() question

[Wu Fengguang]

On Wed, Jan 27, 2010 at 04:25:28AM -0800, Jan Kara wrote:
> On Wed 27-01-10 09:53:39, Wu Fengguang wrote:

> > The patch includes a cleanup and a bug fix, both looks OK to me.
> > But if we can split it, the bug fix would be good candidate for
> > the stable kernel?
>   I don't think we want to push this to -stable kernel. As I wrote, this
> code is called only in case user is going to loose written data which is a
> bug on it's own so loosing a few page references is nothing compared to
> that.

Ah OK. Thanks for the patch.

Cheers,
Fengguang

Re: ext4_da_block_invalidatepages() question

[Jan Kara]

>   Hi,
> 
> On Tue 26-01-10 21:36:08, Wu Fengguang wrote:
> > I noticed that ext4_da_block_invalidatepages() does pagevec_lookup()
> > without pagevec_release()/put_page(). Is that OK?
>   Yes, the function looks buggy. Luckily, it is called only in case we are
> not able to allocate space for delay-allocated data which is a bug on its
> own. So people should never hit it.
>   Attached patch should fix the issue. Ted, will you merge it please?
> Thanks.
  Ted, could you please merge the patch below? Thanks!

 								Honza

---
>From 47085f1ac03eaca9e4d7a5f8f1e40e87d3879512 Mon Sep 17 00:00:00 2001
From: Jan Kara <jack@suse.cz>
Date: Tue, 26 Jan 2010 16:15:19 +0100
Subject: [PATCH] ext4: Release page references acquired in ext4_da_block_invalidatepages

We forget to release page references we acquire in
ext4_da_block_invalidatepages.  Luckily, this function gets called only if we
are not able to allocate blocks for delay-allocated data so that function
should better never be called.

Also cleanup handling of index variable.

Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jan Kara <jack@suse.cz>
---
 fs/ext4/inode.c |    7 +++----
 1 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index c818972..1680007 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -2127,17 +2127,16 @@ static void ext4_da_block_invalidatepages(struct mpage_da_data *mpd,
 			break;
 		for (i = 0; i < nr_pages; i++) {
 			struct page *page = pvec.pages[i];
-			index = page->index;
-			if (index > end)
+			if (page->index > end)
 				break;
-			index++;