linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Eric Sandeen <sandeen@sandeen.net>
To: Dave Chinner <david@fromorbit.com>, linux-xfs@vger.kernel.org
Subject: Re: [PATCH 2/7] repair: don't dirty inodes which are not unlinked
Date: Tue, 30 Oct 2018 15:09:56 -0500	[thread overview]
Message-ID: <a2aa0727-21db-fd2a-6207-0b70d05a7374@sandeen.net> (raw)
In-Reply-To: <539a02d8-f9fd-e9b4-d928-dcdced642376@sandeen.net>

So I'm not stealing commits: ;)

===

repair: don't dirty inodes unconditionally when testing unlinked state

From: Dave Chinner <dchinner@redhat.com>

I noticed phase 4 writing back lots of inode buffers during recent
testing. The recent rework of clear_inode() in commit 0724d0f4cb53
("xfs_repair: clear_dinode should simply clear, not check contents")
accidentally caught a call to clear_inode_unlinked() as well,
resulting in all inodes being marked dirty whether then needed
updating or not.

Fix it by making clear_inode_unlinked unconditionally do the clear
(as was done for clear_inode), and move the test to the caller.
Add warnings as well so that this corruption is no longer silently
fixed.

Reported-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
[sandeen: rework so clear_inode_unlinked is unconditional]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
diff --git a/repair/dinode.c b/repair/dinode.c
index 379f85c..f466e77 100644
--- a/repair/dinode.c
+++ b/repair/dinode.c
@@ -125,23 +125,16 @@ clear_dinode_core(struct xfs_mount *mp, xfs_dinode_t *dinoc, xfs_ino_t ino_num)
 	return;
 }
 
-static int
+static void
 clear_dinode_unlinked(xfs_mount_t *mp, xfs_dinode_t *dino)
 {
 
-	if (be32_to_cpu(dino->di_next_unlinked) != NULLAGINO)  {
-		if (!no_modify)
-			dino->di_next_unlinked = cpu_to_be32(NULLAGINO);
-		return(1);
-	}
-
-	return(0);
+	dino->di_next_unlinked = cpu_to_be32(NULLAGINO);
 }
 
 /*
  * this clears the unlinked list too so it should not be called
  * until after the agi unlinked lists are walked in phase 3.
- * returns > zero if the inode has been altered while being cleared
  */
 static void
 clear_dinode(xfs_mount_t *mp, xfs_dinode_t *dino, xfs_ino_t ino_num)
@@ -2675,9 +2668,16 @@ _("bad (negative) size %" PRId64 " on inode %" PRIu64 "\n"),
 	 * we're going to find.  check_dups is set to 1 only during
 	 * phase 4.  Ugly.
 	 */
-	if (check_dups && !no_modify) {
-		clear_dinode_unlinked(mp, dino);
-		*dirty += 1;
+	if (check_dups && be32_to_cpu(dino->di_next_unlinked) != NULLAGINO) {
+		if (no_modify) {
+			do_warn(
+	_("Would clear next_unlinked in inode %" PRIu64 "\n"), lino);
+		} else  {
+			clear_dinode_unlinked(mp, dino);
+			do_warn(
+	_("Cleared next_unlinked in inode %" PRIu64 "\n"), lino);
+			*dirty += 1;
+		}
 	}
 
 	/* set type and map type info */

  reply	other threads:[~2018-10-31  5:04 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-30 11:20 [PATCH 0/7] xfs_repair: scale to 150,000 iops Dave Chinner
2018-10-30 11:20 ` [PATCH 1/7] Revert "xfs_repair: treat zero da btree pointers as corruption" Dave Chinner
2018-10-30 17:20   ` Darrick J. Wong
2018-10-30 19:35     ` Eric Sandeen
2018-10-30 20:11       ` Dave Chinner
2018-10-30 11:20 ` [PATCH 2/7] repair: don't dirty inodes which are not unlinked Dave Chinner
2018-10-30 17:26   ` Darrick J. Wong
2018-10-30 20:03   ` Eric Sandeen
2018-10-30 20:09     ` Eric Sandeen [this message]
2018-10-30 20:34       ` Dave Chinner
2018-10-30 20:40         ` Eric Sandeen
2018-10-30 20:58           ` Dave Chinner
2018-10-30 11:20 ` [PATCH 3/7] cache: prevent expansion races Dave Chinner
2018-10-30 17:39   ` Darrick J. Wong
2018-10-30 20:35     ` Dave Chinner
2018-10-31 17:13   ` Brian Foster
2018-11-01  1:27     ` Dave Chinner
2018-11-01 13:17       ` Brian Foster
2018-11-01 21:23         ` Dave Chinner
2018-11-02 11:31           ` Brian Foster
2018-11-02 23:26             ` Dave Chinner
2018-10-30 11:20 ` [PATCH 4/7] workqueue: bound maximum queue depth Dave Chinner
2018-10-30 17:58   ` Darrick J. Wong
2018-10-30 20:53     ` Dave Chinner
2018-10-31 17:14       ` Brian Foster
2018-10-30 11:20 ` [PATCH 5/7] repair: Protect bad inode list with mutex Dave Chinner
2018-10-30 17:44   ` Darrick J. Wong
2018-10-30 20:54     ` Dave Chinner
2018-10-30 11:20 ` [PATCH 6/7] repair: protect inode chunk tree records with a mutex Dave Chinner
2018-10-30 17:46   ` Darrick J. Wong
2018-10-30 11:20 ` [PATCH 7/7] repair: parallelise phase 6 Dave Chinner
2018-10-30 17:51   ` Darrick J. Wong
2018-10-30 20:55     ` Dave Chinner
2018-11-07  5:44 ` [PATCH 0/7] xfs_repair: scale to 150,000 iops Arkadiusz Miśkiewicz
2018-11-07  6:48   ` Dave Chinner

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=a2aa0727-21db-fd2a-6207-0b70d05a7374@sandeen.net \
    --to=sandeen@sandeen.net \
    --cc=david@fromorbit.com \
    --cc=linux-xfs@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).