From: Eric Sandeen <sandeen@sandeen.net>
To: "Darrick J. Wong" <djwong@alder.djwong.org>,
david@fromorbit.com, darrick.wong@oracle.com
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 03/10] xfs_repair: ignore "repaired" flag after we decide to clear xattr block
Date: Mon, 17 Aug 2015 14:20:32 -0500 [thread overview]
Message-ID: <55D23400.7020106@sandeen.net> (raw)
In-Reply-To: <20150815014357.1839.68720.stgit@birch.djwong.org>
On 8/14/15 8:43 PM, Darrick J. Wong wrote:
> If in the course of examining extended attribute block contents we
> first decide to repair an entry (*repair = 1) but secondly decide to
> clear the whole block, set *repair = 0 because the clearing action
> only happens if *repair == 0. Put another way, if we're nuking a
> block, don't pretend like we've fixed it too.
Hm, or what happens otherwise?
TBH this is making my brain hurt, chasing it all the way back up
the callchain. It's *repair not *repaired; it's not saying we
fixed it, but that we should fix it, right? It'll take me a while
to work out what all the callers do with the "clearit" return from
process_attr_leaf_block and with *repair.
Did you have a specific example of something going wrong?
thanks,
-Eric
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
> repair/attr_repair.c | 14 +++++++++++++-
> 1 file changed, 13 insertions(+), 1 deletion(-)
>
>
> diff --git a/repair/attr_repair.c b/repair/attr_repair.c
> index 62f80e7..2bd9334 100644
> --- a/repair/attr_repair.c
> +++ b/repair/attr_repair.c
> @@ -1311,6 +1311,13 @@ process_leaf_attr_block(
> * we can add it then.
> */
> }
> + /*
> + * If we're just going to zap the block, don't pretend like we
> + * repaired it, because repairing the block stops the clear
> + * operation.
> + */
> + if (clearit)
> + *repair = 0;
> if (*repair)
> xfs_attr3_leaf_hdr_to_disk(mp->m_attr_geo, leaf, &leafhdr);
>
> @@ -1524,6 +1531,7 @@ process_longform_attr(
> xfs_dahash_t next_hashval;
> int repairlinks = 0;
> struct xfs_attr3_icleaf_hdr leafhdr;
> + int error;
>
> *repair = 0;
>
> @@ -1604,12 +1612,16 @@ process_longform_attr(
> libxfs_writebuf(bp, 0);
> } else
> libxfs_putbuf(bp);
> - return (process_node_attr(mp, ino, dip, blkmap)); /* + repair */
> + error = process_node_attr(mp, ino, dip, blkmap); /* + repair */
> + if (error)
> + *repair = 0;
> + return error;
> default:
> do_warn(
> _("bad attribute leaf magic # %#x for dir ino %" PRIu64 "\n"),
> be16_to_cpu(leaf->hdr.info.magic), ino);
> libxfs_putbuf(bp);
> + *repair = 0;
> return(1);
> }
>
>
> _______________________________________________
> xfs mailing list
> xfs@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs
>
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2015-08-17 19:20 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-08-15 1:43 [PATCH 00/10] xfsprogs August 2015 patchbomb Darrick J. Wong
2015-08-15 1:43 ` [PATCH 01/10] libxfs: readahead of dir3 data blocks should use the read verifier Darrick J. Wong
2015-08-17 18:31 ` Eric Sandeen
2015-08-17 20:30 ` Darrick J. Wong
2015-08-15 1:43 ` [PATCH 02/10] xfs_db: don't crash on a corrupt inode Darrick J. Wong
2015-08-17 18:52 ` Eric Sandeen
2015-08-17 20:45 ` Darrick J. Wong
2015-08-15 1:43 ` [PATCH 03/10] xfs_repair: ignore "repaired" flag after we decide to clear xattr block Darrick J. Wong
2015-08-17 19:20 ` Eric Sandeen [this message]
2015-08-17 20:50 ` Darrick J. Wong
2015-08-15 1:44 ` [PATCH 04/10] xfs_repair: fix broken EFSBADCRC/EFSCORRUPTED usage with buffer errors Darrick J. Wong
2015-08-17 19:51 ` Eric Sandeen
2015-08-17 19:57 ` Eric Sandeen
2015-08-15 1:44 ` [PATCH 05/10] xfs_repair: force not-so-bad bmbt blocks back through the verifier Darrick J. Wong
2015-08-17 21:14 ` Eric Sandeen
2015-08-17 23:48 ` Darrick J. Wong
2015-08-15 1:44 ` [PATCH 06/10] xfs_repair: mark unreachable prefetched metadata blocks stale Darrick J. Wong
2015-08-15 1:44 ` [PATCH 07/10] xfs_io: support reflinking and deduping file ranges Darrick J. Wong
2015-08-15 1:44 ` [PATCH 08/10] xfs_db: enable blocktrash for checksummed filesystems Darrick J. Wong
2015-08-18 19:26 ` Eric Sandeen
2015-08-19 15:22 ` Darrick J. Wong
2015-08-15 1:44 ` [PATCH 09/10] xfs_db: trash the block at the top of the cursor stack Darrick J. Wong
2015-08-18 19:59 ` Eric Sandeen
2015-08-19 15:12 ` Darrick J. Wong
2015-08-15 1:44 ` [PATCH 10/10] xfs_db: enable blockget for v5 filesystems Darrick J. Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=55D23400.7020106@sandeen.net \
--to=sandeen@sandeen.net \
--cc=darrick.wong@oracle.com \
--cc=david@fromorbit.com \
--cc=djwong@alder.djwong.org \
--cc=xfs@oss.sgi.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox