linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dave Chinner <david@fromorbit.com>
To: xfs@oss.sgi.com
Subject: [PATCH 9/9] repair: detect and handle attribute tree CRC errors
Date: Thu, 24 Apr 2014 15:02:02 +1000	[thread overview]
Message-ID: <1398315722-20870-10-git-send-email-david@fromorbit.com> (raw)
In-Reply-To: <1398315722-20870-1-git-send-email-david@fromorbit.com>

From: Dave Chinner <dchinner@redhat.com>

Currently the attribute code will not detect and correct errors in
the attribute tree. It also fails to validate the CRCs and headers
on remote attribute blocks. Ensure that all the attribute blocks are
CRC checked and that the processing functions understand the correct
block formats for decoding.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
---
 repair/attr_repair.c | 45 +++++++++++++++++++++++++++++++++++++--------
 1 file changed, 37 insertions(+), 8 deletions(-)

diff --git a/repair/attr_repair.c b/repair/attr_repair.c
index ba85ac2..9b57960 100644
--- a/repair/attr_repair.c
+++ b/repair/attr_repair.c
@@ -604,6 +604,7 @@ verify_da_path(xfs_mount_t	*mp,
 			libxfs_putbuf(bp);
 			return(1);
 		}
+
 		/*
 		 * update cursor, write out the *current* level if
 		 * required.  don't write out the descendant level
@@ -615,6 +616,8 @@ verify_da_path(xfs_mount_t	*mp,
 			libxfs_writebuf(cursor->level[this_level].bp, 0);
 		else
 			libxfs_putbuf(cursor->level[this_level].bp);
+
+		/* switch cursor to point at the new buffer we just read */
 		cursor->level[this_level].bp = bp;
 		cursor->level[this_level].dirty = 0;
 		cursor->level[this_level].bno = dabno;
@@ -624,6 +627,14 @@ verify_da_path(xfs_mount_t	*mp,
 		cursor->level[this_level].n = newnode;
 #endif
 		entry = cursor->level[this_level].index = 0;
+
+		/*
+		 * We want to rewrite the buffer a CRC error seeing as it
+		 * contains what appears to be a valid node block, but only if
+		 * we are fixing errors.
+		 */
+		if (bp->b_error == EFSBADCRC && !no_modify)
+			cursor->level[this_level].dirty++;
 	}
 	/*
 	 * ditto for block numbers
@@ -974,6 +985,10 @@ rmtval_get(xfs_mount_t *mp, xfs_ino_t ino, blkmap_t *blkmap,
 	xfs_dfsbno_t	bno;
 	xfs_buf_t	*bp;
 	int		clearit = 0, i = 0, length = 0, amountdone = 0;
+	int		hdrsize = 0;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		hdrsize = sizeof(struct xfs_attr3_rmt_hdr);
 
 	/* ASSUMPTION: valuelen is a valid number, so use it for looping */
 	/* Note that valuelen is not a multiple of blocksize */
@@ -986,16 +1001,26 @@ rmtval_get(xfs_mount_t *mp, xfs_ino_t ino, blkmap_t *blkmap,
 			break;
 		}
 		bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
-				XFS_FSB_TO_BB(mp, 1), 0, NULL);
+				    XFS_FSB_TO_BB(mp, 1), 0,
+				    &xfs_attr3_rmt_buf_ops);
 		if (!bp) {
 			do_warn(
 	_("can't read remote block for attributes of inode %" PRIu64 "\n"), ino);
 			clearit = 1;
 			break;
 		}
+
+		if (bp->b_error == EFSBADCRC || bp->b_error == EFSCORRUPTED) {
+			do_warn(
+	_("Corrupt remote block for attributes of inode %" PRIu64 "\n"), ino);
+			clearit = 1;
+			break;
+		}
+
 		ASSERT(mp->m_sb.sb_blocksize == XFS_BUF_COUNT(bp));
-		length = MIN(XFS_BUF_COUNT(bp), valuelen - amountdone);
-		memmove(value, XFS_BUF_PTR(bp), length);
+
+		length = MIN(XFS_BUF_COUNT(bp) - hdrsize, valuelen - amountdone);
+		memmove(value, XFS_BUF_PTR(bp) + hdrsize, length);
 		amountdone += length;
 		value += length;
 		i++;
@@ -1143,7 +1168,6 @@ process_leaf_attr_block(
 
 	xfs_attr3_leaf_hdr_from_disk(&leafhdr, leaf);
 	clearit = usedbs = 0;
-	*repair = 0;
 	firstb = mp->m_sb.sb_blocksize;
 	stop = xfs_attr3_leaf_hdr_size(leaf);
 
@@ -1320,13 +1344,16 @@ process_leaf_attr_level(xfs_mount_t	*mp,
 		}
 
 		bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, dev_bno),
-					XFS_FSB_TO_BB(mp, 1), 0, NULL);
+				    XFS_FSB_TO_BB(mp, 1), 0,
+				    &xfs_attr3_leaf_buf_ops);
 		if (!bp) {
 			do_warn(
 	_("can't read file block %u (fsbno %" PRIu64 ") for attribute fork of inode %" PRIu64 "\n"),
 				da_bno, dev_bno, ino);
 			goto error_out;
 		}
+		if (bp->b_error == EFSBADCRC)
+			repair++;
 
 		leaf = bp->b_addr;
 		xfs_attr3_leaf_hdr_from_disk(&leafhdr, leaf);
@@ -1382,9 +1409,9 @@ process_leaf_attr_level(xfs_mount_t	*mp,
 
 		current_hashval = greatest_hashval;
 
-		if (repair && !no_modify) 
+		if (repair && !no_modify)
 			libxfs_writebuf(bp, 0);
-		else 
+		else
 			libxfs_putbuf(bp);
 	} while (da_bno != 0);
 
@@ -1512,6 +1539,8 @@ process_longform_attr(
 			ino);
 		return(1);
 	}
+	if (bp->b_error == EFSBADCRC)
+		(*repair)++;
 
 	/* verify leaf block */
 	leaf = (xfs_attr_leafblock_t *)XFS_BUF_PTR(bp);
@@ -1555,7 +1584,7 @@ process_longform_attr(
 	case XFS_DA_NODE_MAGIC:		/* btree-form attribute */
 	case XFS_DA3_NODE_MAGIC:
 		/* must do this now, to release block 0 before the traversal */
-		if (repairlinks) {
+		if (*repair || repairlinks) {
 			*repair = 1;
 			libxfs_writebuf(bp, 0);
 		} else
-- 
1.9.0

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

  parent reply	other threads:[~2014-04-24  5:02 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-04-24  5:01 [PATCH 0/9 V2] xfs_db, xfs_repair: improve CRC error detection Dave Chinner
2014-04-24  5:01 ` [PATCH 1/9] db: don't claim unchecked CRCs are correct Dave Chinner
2014-04-24  5:01 ` [PATCH 2/9] db: verify buffer on type change Dave Chinner
2014-04-25  5:41   ` Christoph Hellwig
2014-04-24  5:01 ` [PATCH 3/9] repair: ensure prefetched buffers have CRCs validated Dave Chinner
2014-04-25  5:47   ` Christoph Hellwig
2014-04-24  5:01 ` [PATCH 4/9] repair: detect and correct CRC errors in directory blocks Dave Chinner
2014-04-25  5:47   ` Christoph Hellwig
2014-04-24  5:01 ` [PATCH 5/9] repair: detect CRC errors in AG headers Dave Chinner
2014-04-25  5:55   ` Christoph Hellwig
2014-04-24  5:01 ` [PATCH 6/9] repair: report AG btree verifier errors Dave Chinner
2014-04-25  5:55   ` Christoph Hellwig
2014-04-24  5:02 ` [PATCH 7/9] repair: remove more dirv1 leftovers Dave Chinner
2014-04-24  5:02 ` [PATCH 8/9] repair: handle remote symlink CRC errors Dave Chinner
2014-04-25  6:01   ` Christoph Hellwig
2014-04-24  5:02 ` Dave Chinner [this message]
  -- strict thread matches above, loose matches on Subject: below --
2014-04-28 21:04 [PATCH 0/9 v3] xfs_db, xfs_repair: improve CRC error detection Dave Chinner
2014-04-28 21:04 ` [PATCH 9/9] repair: detect and handle attribute tree CRC errors Dave Chinner
2014-04-29 14:06   ` Brian Foster
2014-04-30  3:55     ` Dave Chinner
2014-04-29 18:18   ` Christoph Hellwig
2014-04-15  8:24 [PATCH 0/9] xfs_db, xfs_repair: improve CRC error detection Dave Chinner
2014-04-15  8:25 ` [PATCH 9/9] repair: detect and handle attribute tree CRC errors Dave Chinner
2014-04-16 13:25   ` Brian Foster
2014-04-21 23:27     ` 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=1398315722-20870-10-git-send-email-david@fromorbit.com \
    --to=david@fromorbit.com \
    --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;
as well as URLs for NNTP newsgroup(s).