On Tue, Dec 05, 2017 at 01:51:43PM +0200, Nikolay Borisov wrote:
> add_pending_csums was added as part of the new data=ordered implementation in
> e6dcd2dc9c48 ("Btrfs: New data=ordered implementation"). Even back then it
> called the btrfs_csum_file_blocks which can fail but it never bothered 
> handling
> the failure. In ENOMEM situation this could lead to the filesystem failing to
> write the checksums for a particular extent and not detect this. On read this
> could lead to the filesystem erroring out due to crc mismatch. Fix it by
> propagating failure from add_pending_csums and handling them
> 
> Signed-off-by: Nikolay Borisov <nbori...@suse.com>
> ---
>  fs/btrfs/inode.c | 11 +++++++++--
>  1 file changed, 9 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
> index e87ec11c0986..432bffdbb02f 100644
> --- a/fs/btrfs/inode.c
> +++ b/fs/btrfs/inode.c
> @@ -2039,11 +2039,14 @@ static noinline int add_pending_csums(struct 
> btrfs_trans_handle *trans,
>                            struct inode *inode, struct list_head *list)
>  {
>       struct btrfs_ordered_sum *sum;
> +     int ret;
>  
>       list_for_each_entry(sum, list, list) {
>               trans->adding_csums = true;
> -             btrfs_csum_file_blocks(trans,
> +             ret = btrfs_csum_file_blocks(trans,
>                      BTRFS_I(inode)->root->fs_info->csum_root, sum);
> +             if (ret)
> +                     return ret;

The return should come after the line below, otherwise the transaction
will be left in the "adding csums".

>               trans->adding_csums = false;

...
>       }
>       return 0;
> @@ -3051,7 +3054,11 @@ static int btrfs_finish_ordered_io(struct 
> btrfs_ordered_extent *ordered_extent)
>               goto out;
>       }
>  
> -     add_pending_csums(trans, inode, &ordered_extent->list);
> +     ret = add_pending_csums(trans, inode, &ordered_extent->list);
> +     if (ret) {
> +             btrfs_abort_transaction(trans, ret);

Ok, we can't do better here, this is too late and
add_pending_csums -> btrfs_csum_file_blocks modifies too much of the
state to be rolled back safely.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to