Re: [Bug #14472] EXT4 corruption

[Andrew Lutomirski <luto@mit.edu>]

On Thu, Oct 29, 2009 at 5:23 PM, Theodore Tso <tytso@mit.edu> wrote:
> On Thu, Oct 29, 2009 at 03:57:32PM -0400, Andrew Lutomirski wrote:
>>
>> This but is *not* fixed.  I just triggered it a few minutes ago by
>> abusing i915 and drm, which caused a panic.  This is slightly newer
>> than 2.6.32-rc5, with a couple of i915 bugfixes thrown in.
>
> Andrew, can you test to see if this patch helps?
>
> Thanks,
>
>                                                - Ted
>
> commit a8836b1d6f92273e001012c7705ae8f4c3d5fb65
> Author: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
> Date:   Tue Oct 27 15:36:38 2009 +0530
>
>    ext4: discard preallocation during truncate
>
>    We need to make sure when we drop and reacquire the inode's
>    i_data_sem we discard the inode preallocation. Otherwise we
>    could have blocks marked as free in bitmap but still belonging
>    to prealloc space.
>
>    Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
>
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index 5c5bc5d..a1ef1c3 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -209,6 +209,12 @@ static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
>        up_write(&EXT4_I(inode)->i_data_sem);
>        ret = ext4_journal_restart(handle, blocks_for_truncate(inode));
>        down_write(&EXT4_I(inode)->i_data_sem);
> +       /*
> +        * We have dropped i_data_sem. So somebody else could have done
> +        * block allocation. So discard the prealloc space created as a
> +        * part of block allocation
> +        */
> +       ext4_discard_preallocations(inode);
>
>        return ret;
>  }
>


It looks like 2.6.32-rc6 is supposed to fix this bug, but it also
looks like this patch didn't make it in.  Should I still be using this
patch?

Thanks,
Andy