linux-next: manual merge of the vfs tree with the gfs2 tree

linux-next: manual merge of the vfs tree with the gfs2 tree

[Stephen Rothwell]

Hi Al,

Today's linux-next merge of the vfs tree got a conflict in
fs/gfs2/ops_super.c between commit
9e6e0a128bca0a151d8d3fbd9459b22fc21cfebb ("GFS2: Merge mount.c and
ops_super.c into super.c") from the gfs2 tree and commits
17af8f24a7750ea3d947904f97eb6dfacf9a88aa ("gfs2: remove ->write_super and
stop maintaining ->s_dirt") and 8123178eb9ca12cde31a95170746e15a79528a62
("push BKL down into ->put_super") from the vfs tree.

The former commit removed the file.  I have mechanically applied the
changes in the latter two commits to fs/gfs2/super.c (see below) and can
carry this as a merge fix as necessary.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index 40bcc37..c8930b3 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -719,6 +719,8 @@ static void gfs2_put_super(struct super_block *sb)
 	int error;
 	struct gfs2_jdesc *jd;
 
+	lock_kernel();
+
 	/*  Unfreeze the filesystem, if we need to  */
 
 	mutex_lock(&sdp->sd_freeze_lock);
@@ -785,17 +787,8 @@ restart:
 
 	/*  At this point, we're through participating in the lockspace  */
 	gfs2_sys_fs_del(sdp);
-}
-
-/**
- * gfs2_write_super
- * @sb: the superblock
- *
- */
 
-static void gfs2_write_super(struct super_block *sb)
-{
-	sb->s_dirt = 0;
+	unlock_kernel();
 }
 
 /**
@@ -807,7 +800,6 @@ static void gfs2_write_super(struct super_block *sb)
 
 static int gfs2_sync_fs(struct super_block *sb, int wait)
 {
-	sb->s_dirt = 0;
 	if (wait && sb->s_fs_info)
 		gfs2_log_flush(sb->s_fs_info, NULL);
 	return 0;
@@ -1324,7 +1316,6 @@ const struct super_operations gfs2_super_ops = {
 	.write_inode		= gfs2_write_inode,
 	.delete_inode		= gfs2_delete_inode,
 	.put_super		= gfs2_put_super,
-	.write_super		= gfs2_write_super,
 	.sync_fs		= gfs2_sync_fs,
 	.freeze_fs 		= gfs2_freeze,
 	.unfreeze_fs		= gfs2_unfreeze,
-- 
1.6.3.1

Re: linux-next: manual merge of the vfs tree with the gfs2 tree

[Steven Whitehouse]

Hi,

On Wed, 2009-05-27 at 11:52 +1000, Stephen Rothwell wrote:
> Hi Al,
> 
> Today's linux-next merge of the vfs tree got a conflict in
> fs/gfs2/ops_super.c between commit
> 9e6e0a128bca0a151d8d3fbd9459b22fc21cfebb ("GFS2: Merge mount.c and
> ops_super.c into super.c") from the gfs2 tree and commits
> 17af8f24a7750ea3d947904f97eb6dfacf9a88aa ("gfs2: remove ->write_super and
> stop maintaining ->s_dirt") and 8123178eb9ca12cde31a95170746e15a79528a62
> ("push BKL down into ->put_super") from the vfs tree.
> 
> The former commit removed the file.  I have mechanically applied the
> changes in the latter two commits to fs/gfs2/super.c (see below) and can
> carry this as a merge fix as necessary.
> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> 
Thanks for fixing this up.

> diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
> index 40bcc37..c8930b3 100644
> --- a/fs/gfs2/super.c
> +++ b/fs/gfs2/super.c
> @@ -719,6 +719,8 @@ static void gfs2_put_super(struct super_block *sb)
>  	int error;
>  	struct gfs2_jdesc *jd;
>  
> +	lock_kernel();
> +
>  	/*  Unfreeze the filesystem, if we need to  */
>  
>  	mutex_lock(&sdp->sd_freeze_lock);
> @@ -785,17 +787,8 @@ restart:
>  
>  	/*  At this point, we're through participating in the lockspace  */
>  	gfs2_sys_fs_del(sdp);
> -}


Al/Christoph, what is the purpose of the lock_kernel() ? I don't see why
this is required. What is it protecting against?

Steve.

Re: linux-next: manual merge of the vfs tree with the gfs2 tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 726 bytes --]

Hi Al,

On Wed, 27 May 2009 11:52:53 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Today's linux-next merge of the vfs tree got a conflict in
> fs/gfs2/ops_super.c between commit
> 9e6e0a128bca0a151d8d3fbd9459b22fc21cfebb ("GFS2: Merge mount.c and
> ops_super.c into super.c") from the gfs2 tree and commits
> 17af8f24a7750ea3d947904f97eb6dfacf9a88aa ("gfs2: remove ->write_super and
> stop maintaining ->s_dirt") and 8123178eb9ca12cde31a95170746e15a79528a62
> ("push BKL down into ->put_super") from the vfs tree.

This conflict is now against Linus's tree as the gfs2 tree has been
merged.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: manual merge of the vfs tree with the gfs2 tree

[Al Viro]

On Fri, Jun 12, 2009 at 11:12:37AM +1000, Stephen Rothwell wrote:
> Hi Al,
> 
> On Wed, 27 May 2009 11:52:53 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> >
> > Today's linux-next merge of the vfs tree got a conflict in
> > fs/gfs2/ops_super.c between commit
> > 9e6e0a128bca0a151d8d3fbd9459b22fc21cfebb ("GFS2: Merge mount.c and
> > ops_super.c into super.c") from the gfs2 tree and commits
> > 17af8f24a7750ea3d947904f97eb6dfacf9a88aa ("gfs2: remove ->write_super and
> > stop maintaining ->s_dirt") and 8123178eb9ca12cde31a95170746e15a79528a62
> > ("push BKL down into ->put_super") from the vfs tree.
> 
> This conflict is now against Linus's tree as the gfs2 tree has been
> merged.

I know, properly ported stuff is in updated for-next (basically, a bunch
of functions moved from ops_super.c to super.c, so changesets needed to
be updated to touch the latter).

Re: linux-next: manual merge of the vfs tree with the gfs2 tree

[Steven Whitehouse]

Hi,

On Mon, 2010-01-18 at 10:55 +1100, Stephen Rothwell wrote:
> Hi Al,
> 
> Today's linux-next merge of the vfs tree got a conflict in
> fs/gfs2/ops_inode.c between commit
> 83c77e8b3457f2ee5dad028dc54cf3ce540104b2 ("GFS2: Fix refcnt leak on
> gfs2_follow_link() error path") from the gfs2 tree and commit
> 261a144ac2b3867c7be70f08925e446430df6937 ("Switch gfs2 to nd_set_link()")
> from the vfs tree.
> 
> I can't figure out if the gfs2 tree fix is required any more, so I just
> used the vfs tree version.
> 
> Al, if that gfs2 patch is standalone, you should probably send it to the
> gfs2 guys.
> 

It looks as if the two patches would be alternatives. The only question
is whether Al's patch should be left for the merge window or whether its
ok to send it ahead of time to fix the issue (which is what I had been
intending to do with the bug fix prior to Al's patch).

Thoughts?

Steve.

linux-next: manual merge of the vfs tree with the gfs2 tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 633 bytes --]

Hi Al,

Today's linux-next merge of the vfs tree got a conflict in
fs/gfs2/ops_inode.c between commit
83c77e8b3457f2ee5dad028dc54cf3ce540104b2 ("GFS2: Fix refcnt leak on
gfs2_follow_link() error path") from the gfs2 tree and commit
261a144ac2b3867c7be70f08925e446430df6937 ("Switch gfs2 to nd_set_link()")
from the vfs tree.

I can't figure out if the gfs2 tree fix is required any more, so I just
used the vfs tree version.

Al, if that gfs2 patch is standalone, you should probably send it to the
gfs2 guys.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]