Re: [PATCH] [3/18] BKL-removal: Convert ext3 to use unlocked_ioctl

Started by Andrew Morton on Jan 28, 2008 · 2 messages from 2 authors · Last reply Jan 28, 2008

open the first message on its own page

Message details and related content
On Sun, 27 Jan 2008 03:17:09 +0100 (CET) Andi Kleen [off-list ref] wrote:
I checked ext3_ioctl and it looked largely safe to not be used
without BKL.  So convert it over to unlocked_ioctl.

The only case where I wasn't quite sure was for the
dynamic fs grow ioctls versus umounting -- I kept the BKL for those. 
Please cpoy linux-ext4 on ext2/3/4 material.

I skippped a lot of these patches because I just got bored of fixing
rejects.  Now is a very optimistic time to be raising patches against
mainline.

I'm going to work on getting a unified devel tree operating: one which
contains everyone's latest stuff and is updated daily.  Basically it'll be
-mm without a couple of the quilt trees.  People can then prepare patches
against that, as it seems that most can't be bothered patching against -mm,
let alone building and testing it.  More later.
quoted hunk
+		/* AK: not sure the BKL is needed, but this might prevent
+		 * races against umount */
+		lock_kernel();
 		err = ext3_group_extend(sb, EXT3_SB(sb)->s_es, n_blocks_count);
 		journal_lock_updates(EXT3_SB(sb)->s_journal);
 		journal_flush(EXT3_SB(sb)->s_journal);
 		journal_unlock_updates(EXT3_SB(sb)->s_journal);
+		unlock_kernel();
 
 		return err;
 	}
@@ -245,11 +249,14 @@ flags_err:
 		if (copy_from_user(&input, (struct ext3_new_group_input __user *)arg,
 				sizeof(input)))
 			return -EFAULT;
-
+		/* AK: not sure the BKL is needed, but this might prevent
+		 * races against umount */
+		lock_kernel();
 		err = ext3_group_add(sb, &input);
 		journal_lock_updates(EXT3_SB(sb)->s_journal);
 		journal_flush(EXT3_SB(sb)->s_journal);
 		journal_unlock_updates(EXT3_SB(sb)->s_journal);
+		unlock_kernel();
 
The ext3_ioctl() caller has an open fd against the fs - should be
sufficient to keep unmount away?

(gets even more rejects, drops all the fasync patches too)

It's all reached the stage of stupid.
Message details and related content
On Monday 28 January 2008 06:33, Andrew Morton wrote:
On Sun, 27 Jan 2008 03:17:09 +0100 (CET) Andi Kleen [off-list ref] wrote:
quoted
I checked ext3_ioctl and it looked largely safe to not be used
without BKL.  So convert it over to unlocked_ioctl.

The only case where I wasn't quite sure was for the
dynamic fs grow ioctls versus umounting -- I kept the BKL for those.
Please cpoy linux-ext4 on ext2/3/4 material.
Ok I'll resubmit those to tytso/ext4-devel (or perhaps he has already seen 
them) 
I skippped a lot of these patches because I just got bored of fixing
rejects.  Now is a very optimistic time to be raising patches against
mainline.
JFS and CIFS are already taken care of by the maintainers. This leaves
remote_llseek which touches a couple of file systems. Could you
perhaps take that one only please? And perhaps Nick's minix 
patchkit which looks safe to me and is unlikely to cause conflicts.
quoted
+		/* AK: not sure the BKL is needed, but this might prevent
+		 * races against umount */
+		lock_kernel();
 		err = ext3_group_add(sb, &input);
 		journal_lock_updates(EXT3_SB(sb)->s_journal);
 		journal_flush(EXT3_SB(sb)->s_journal);
 		journal_unlock_updates(EXT3_SB(sb)->s_journal);
+		unlock_kernel();
The ext3_ioctl() caller has an open fd against the fs - should be
sufficient to keep unmount away?
True. I am still conservative because group_add is a lot of code
which I didn't fully check. But with the open fd it's likely safe
to not take the BKL because there is nothing else (except
readdir?) in ext* that takes it.
It's all reached the stage of stupid.
I'll resubmit ->fasync_unlocked against -mm.

Also I wanted to recheck the ->f_flags locking. I found one bug in those 
already and I can extract the bug fix for that one. 

-Andi
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help