All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: xfs <linux-xfs@vger.kernel.org>
Cc: Christoph Hellwig <hch@infradead.org>,
	Chandan Babu R <chandanrlinux@gmail.com>
Subject: [PATCH v2] xfs: ensure inobt record walks always make forward progress
Date: Tue, 17 Nov 2020 10:14:56 -0800	[thread overview]
Message-ID: <20201117181456.GZ9695@magnolia> (raw)

From: Darrick J. Wong <darrick.wong@oracle.com>

The aim of the inode btree record iterator function is to call a
callback on every record in the btree.  To avoid having to tear down and
recreate the inode btree cursor around every callback, it caches a
certain number of records in a memory buffer.  After each batch of
callback invocations, we have to perform a btree lookup to find the
next record after where we left off.

However, if the keys of the inode btree are corrupt, the lookup might
put us in the wrong part of the inode btree, causing the walk function
to loop forever.  Therefore, we add extra cursor tracking to make sure
that we never go backwards neither when performing the lookup nor when
jumping to the next inobt record.  This also fixes an off by one error
where upon resume the lookup should have been for the inode /after/ the
point at which we stopped.

Found by fuzzing xfs/460 with keys[2].startino = ones causing bulkstat
and quotacheck to hang.

Fixes: a211432c27ff ("xfs: create simplified inode walk function")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
v2: fix idiotic mismerge, sorry about that...
---
 fs/xfs/xfs_iwalk.c |   27 ++++++++++++++++++++++++---
 1 file changed, 24 insertions(+), 3 deletions(-)

diff --git a/fs/xfs/xfs_iwalk.c b/fs/xfs/xfs_iwalk.c
index 233dcc8784db..2a45138831e3 100644
--- a/fs/xfs/xfs_iwalk.c
+++ b/fs/xfs/xfs_iwalk.c
@@ -55,6 +55,9 @@ struct xfs_iwalk_ag {
 	/* Where do we start the traversal? */
 	xfs_ino_t			startino;
 
+	/* What was the last inode number we saw when iterating the inobt? */
+	xfs_ino_t			lastino;
+
 	/* Array of inobt records we cache. */
 	struct xfs_inobt_rec_incore	*recs;
 
@@ -301,6 +304,9 @@ xfs_iwalk_ag_start(
 	if (XFS_IS_CORRUPT(mp, *has_more != 1))
 		return -EFSCORRUPTED;
 
+	iwag->lastino = XFS_AGINO_TO_INO(mp, agno,
+				irec->ir_startino + XFS_INODES_PER_CHUNK - 1);
+
 	/*
 	 * If the LE lookup yielded an inobt record before the cursor position,
 	 * skip it and see if there's another one after it.
@@ -347,15 +353,17 @@ xfs_iwalk_run_callbacks(
 	struct xfs_mount		*mp = iwag->mp;
 	struct xfs_trans		*tp = iwag->tp;
 	struct xfs_inobt_rec_incore	*irec;
-	xfs_agino_t			restart;
+	xfs_agino_t			next_agino;
 	int				error;
 
+	next_agino = XFS_INO_TO_AGINO(mp, iwag->lastino) + 1;
+
 	ASSERT(iwag->nr_recs > 0);
 
 	/* Delete cursor but remember the last record we cached... */
 	xfs_iwalk_del_inobt(tp, curpp, agi_bpp, 0);
 	irec = &iwag->recs[iwag->nr_recs - 1];
-	restart = irec->ir_startino + XFS_INODES_PER_CHUNK - 1;
+	ASSERT(next_agino == irec->ir_startino + XFS_INODES_PER_CHUNK);
 
 	error = xfs_iwalk_ag_recs(iwag);
 	if (error)
@@ -372,7 +380,7 @@ xfs_iwalk_run_callbacks(
 	if (error)
 		return error;
 
-	return xfs_inobt_lookup(*curpp, restart, XFS_LOOKUP_GE, has_more);
+	return xfs_inobt_lookup(*curpp, next_agino, XFS_LOOKUP_GE, has_more);
 }
 
 /* Walk all inodes in a single AG, from @iwag->startino to the end of the AG. */
@@ -396,6 +404,7 @@ xfs_iwalk_ag(
 
 	while (!error && has_more) {
 		struct xfs_inobt_rec_incore	*irec;
+		xfs_ino_t			rec_fsino;
 
 		cond_resched();
 		if (xfs_pwork_want_abort(&iwag->pwork))
@@ -407,6 +416,15 @@ xfs_iwalk_ag(
 		if (error || !has_more)
 			break;
 
+		/* Make sure that we always move forward. */
+		rec_fsino = XFS_AGINO_TO_INO(mp, agno, irec->ir_startino);
+		if (iwag->lastino != NULLFSINO &&
+		    XFS_IS_CORRUPT(mp, iwag->lastino >= rec_fsino)) {
+			error = -EFSCORRUPTED;
+			goto out;
+		}
+		iwag->lastino = rec_fsino + XFS_INODES_PER_CHUNK - 1;
+
 		/* No allocated inodes in this chunk; skip it. */
 		if (iwag->skip_empty && irec->ir_freecount == irec->ir_count) {
 			error = xfs_btree_increment(cur, 0, &has_more);
@@ -535,6 +553,7 @@ xfs_iwalk(
 		.trim_start	= 1,
 		.skip_empty	= 1,
 		.pwork		= XFS_PWORK_SINGLE_THREADED,
+		.lastino	= NULLFSINO,
 	};
 	xfs_agnumber_t		agno = XFS_INO_TO_AGNO(mp, startino);
 	int			error;
@@ -623,6 +642,7 @@ xfs_iwalk_threaded(
 		iwag->data = data;
 		iwag->startino = startino;
 		iwag->sz_recs = xfs_iwalk_prefetch(inode_records);
+		iwag->lastino = NULLFSINO;
 		xfs_pwork_queue(&pctl, &iwag->pwork);
 		startino = XFS_AGINO_TO_INO(mp, agno + 1, 0);
 		if (flags & XFS_INOBT_WALK_SAME_AG)
@@ -696,6 +716,7 @@ xfs_inobt_walk(
 		.startino	= startino,
 		.sz_recs	= xfs_inobt_walk_prefetch(inobt_records),
 		.pwork		= XFS_PWORK_SINGLE_THREADED,
+		.lastino	= NULLFSINO,
 	};
 	xfs_agnumber_t		agno = XFS_INO_TO_AGNO(mp, startino);
 	int			error;

             reply	other threads:[~2020-11-17 18:15 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-11-17 18:14 Darrick J. Wong [this message]
2020-11-18  5:58 ` [PATCH v2] xfs: ensure inobt record walks always make forward progress Chandan Babu R

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=20201117181456.GZ9695@magnolia \
    --to=darrick.wong@oracle.com \
    --cc=chandanrlinux@gmail.com \
    --cc=hch@infradead.org \
    --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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.