On 6/29/11 8:01 PM, Lachlan McIlroy wrote:
> bd_super is currently reset to NULL in kill_block_super() so we rely on previous
> users of the block_device object to initialise this value for the next user.
> This quirk was exposed on RHEL5 when a third party filesystem did not always use
> kill_block_super() and therefore bd_super wasn't being reset when a block_device
> object was recycled within the cache. This may not be a problem upstream but
> makes sense to be defensive.
>
> Signed-off-by: Lachlan McIlroy <lmcilroy@xxxxxxxxxx>
> ---
This seems reasonable to me, I think it's dangerous to assume the prior user
will "put it away" properly.
blkdev_releasepage() then may try to deref it and go boom, without this
explicit initialization.
And there is already other initialization in bdget()...
Reviewed-by: Eric Sandeen <sandeen@xxxxxxxxxx>
> fs/block_dev.c | 1 +
> 1 files changed, 1 insertions(+), 0 deletions(-)
>
> diff --git a/fs/block_dev.c b/fs/block_dev.c
> index 610e8e0..2b0dc33 100644
> --- a/fs/block_dev.c
> +++ b/fs/block_dev.c
> @@ -547,6 +547,7 @@ struct block_device *bdget(dev_t dev)
>
> if (inode->i_state & I_NEW) {
> bdev->bd_contains = NULL;
> + bdev->bd_super = NULL;
> bdev->bd_inode = inode;
> bdev->bd_block_size = (1 << inode->i_blkbits);
> bdev->bd_part_count = 0;
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
