On Mon, Mar 03, 2014 at 04:39:53PM +1100, Dave Chinner wrote:
> From: Dave Chinner <dchinner at redhat.com>
>
> While the verifier reoutines may return EFSBADCRC when a buffer ahs
> a bad CRC, we need to translate that to EFSCORRUPTED so that the
> higher layers treat the error appropriately and so we return a
> consistent error to userspace. This fixes a xfs/005 regression.
>
> Signed-off-by: Dave Chinner <dchinner at redhat.com>
> ---
Looking through again, I don't see any obvious cases where EFSBADCRC
could slip through. The verifier users mostly call
xfs_trans_read_buf_map(). The readahead case looks like it should be
handled by the xfs_buf_iowait() checks therein.
The one case I came across that looked suspicious is the readahead/read
cycle for inode log recovery in xlog_recover_commit_trans(). This calls
into xfs_buf_read() and ultimately uses bp->b_error directly in
xlog_recover_inode_pass2(). That said, I don't see any crc checking in
these verifiers and thus they don't source EFSBADCRC. Looks good to
me...
Reviewed-by: Brian Foster <bfoster at redhat.com>
> fs/xfs/xfs_mount.c | 3 +++
> fs/xfs/xfs_symlink.c | 4 ++++
> fs/xfs/xfs_trans_buf.c | 11 +++++++++++
> 3 files changed, 18 insertions(+)
>
> diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
> index f96c056..993cb19 100644
> --- a/fs/xfs/xfs_mount.c
> +++ b/fs/xfs/xfs_mount.c
> @@ -314,6 +314,9 @@ reread:
> error = bp->b_error;
> if (loud)
> xfs_warn(mp, "SB validate failed with error %d.", error);
> + /* bad CRC means corrupted metadata */
> + if (error == EFSBADCRC)
> + error = EFSCORRUPTED;
> goto release_buf;
> }
>
> diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c
> index 14e58f2..5fda189 100644
> --- a/fs/xfs/xfs_symlink.c
> +++ b/fs/xfs/xfs_symlink.c
> @@ -80,6 +80,10 @@ xfs_readlink_bmap(
> if (error) {
> xfs_buf_ioerror_alert(bp, __func__);
> xfs_buf_relse(bp);
> +
> + /* bad CRC means corrupted metadata */
> + if (error == EFSBADCRC)
> + error = EFSCORRUPTED;
> goto out;
> }
> byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
> diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
> index 647b6f1..b8eef05 100644
> --- a/fs/xfs/xfs_trans_buf.c
> +++ b/fs/xfs/xfs_trans_buf.c
> @@ -275,6 +275,10 @@ xfs_trans_read_buf_map(
> XFS_BUF_UNDONE(bp);
> xfs_buf_stale(bp);
> xfs_buf_relse(bp);
> +
> + /* bad CRC means corrupted metadata */
> + if (error == EFSBADCRC)
> + error = EFSCORRUPTED;
> return error;
> }
> #ifdef DEBUG
> @@ -338,6 +342,9 @@ xfs_trans_read_buf_map(
> if (tp->t_flags & XFS_TRANS_DIRTY)
> xfs_force_shutdown(tp->t_mountp,
> SHUTDOWN_META_IO_ERROR);
> + /* bad CRC means corrupted metadata */
> + if (error == EFSBADCRC)
> + error = EFSCORRUPTED;
> return error;
> }
> }
> @@ -375,6 +382,10 @@ xfs_trans_read_buf_map(
> if (tp->t_flags & XFS_TRANS_DIRTY)
> xfs_force_shutdown(tp->t_mountp, SHUTDOWN_META_IO_ERROR);
> xfs_buf_relse(bp);
> +
> + /* bad CRC means corrupted metadata */
> + if (error == EFSBADCRC)
> + error = EFSCORRUPTED;
> return error;
> }
> #ifdef DEBUG
> --
> 1.9.0
>
> _______________________________________________
> xfs mailing list
> xfs at oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs