2009-07-10 11:03:55

by Peng Tao

Subject: [PATCH] fix null handler of ioctls in no journal mode

EXT4_IOC_GROUP_ADD and EXT4_IOC_GROUP_EXTEND ioctls should not flush journal
in no_journal mode.

This can be triggered by running resize2fs on a mounted no_journal partition
as follows:
mkfs.ext4 /dev/sdb1 -O ^has_journal
enlarge /dev/sdb1 in partition table
mount /dev/sdb1
resize2fs /dev/sdb1

And it will get these error message:
BUG: unable to handle kernel NULL pointer dereference at 00000014
IP: [<c039d282>] _spin_lock+0x8/0x19
*pde = 00000000
Oops: 0002 [#1] SMP

Signed-off-by: Peng Tao <[email protected]>
---
fs/ext4/ioctl.c | 20 ++++++++++++--------
1 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
index 91e75f7..9e29c0a 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -191,7 +191,7 @@ setversion_out:
case EXT4_IOC_GROUP_EXTEND: {
ext4_fsblk_t n_blocks_count;
struct super_block *sb = inode->i_sb;
- int err, err2;
+ int err, err2=0;

if (!capable(CAP_SYS_RESOURCE))
return -EPERM;
@@ -204,9 +204,11 @@ setversion_out:
return err;

err = ext4_group_extend(sb, EXT4_SB(sb)->s_es, n_blocks_count);
- jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal);
- err2 = jbd2_journal_flush(EXT4_SB(sb)->s_journal);
- jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
+ if (EXT4_SB(sb)->s_journal) {
+ jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal);
+ err2 = jbd2_journal_flush(EXT4_SB(sb)->s_journal);
+ jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
+ }
if (err == 0)
err = err2;
mnt_drop_write(filp->f_path.mnt);
@@ -216,7 +218,7 @@ setversion_out:
case EXT4_IOC_GROUP_ADD: {
struct ext4_new_group_data input;
struct super_block *sb = inode->i_sb;
- int err, err2;
+ int err, err2=0;

if (!capable(CAP_SYS_RESOURCE))
return -EPERM;
@@ -230,9 +232,11 @@ setversion_out:
return err;

err = ext4_group_add(sb, &input);
- jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal);
- err2 = jbd2_journal_flush(EXT4_SB(sb)->s_journal);
- jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
+ if (EXT4_SB(sb)->s_journal) {
+ jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal);
+ err2 = jbd2_journal_flush(EXT4_SB(sb)->s_journal);
+ jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
+ }
if (err == 0)
err = err2;
mnt_drop_write(filp->f_path.mnt);
--
1.6.2-rc2.GIT

2009-07-13 13:29:57

by Theodore Ts'o

Subject: Re: [PATCH] fix null handler of ioctls in no journal mode

On Fri, Jul 10, 2009 at 07:03:46PM +0800, Peng Tao wrote:
> EXT4_IOC_GROUP_ADD and EXT4_IOC_GROUP_EXTEND ioctls should not flush journal
> in no_journal mode.

Thanks, added to the ext4 patch queue, altough I did tighten up up
your patch description slightly:

ext4: fix null handler of ioctls in no journal mode

The EXT4_IOC_GROUP_ADD and EXT4_IOC_GROUP_EXTEND ioctls should not
flush the journal in no_journal mode. Otherwise, running resize2fs on
a mounted no_journal partition triggers the following error messages:

BUG: unable to handle kernel NULL pointer dereference at 00000014
IP: [<c039d282>] _spin_lock+0x8/0x19
*pde = 00000000
Oops: 0002 [#1] SMP

Signed-off-by: Peng Tao <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>

- Ted