diff mbox series

[1/2] xfs: skip fscounters comparisons when the scan is incomplete

Message ID 166473480561.1083794.11478098043597895258.stgit@magnolia (mailing list archive)
State Accepted
Headers show
Series xfs: enhance fs summary counter scrubber | expand

Commit Message

Darrick J. Wong Oct. 2, 2022, 6:20 p.m. UTC 
From: Darrick J. Wong <djwong@kernel.org>

If any part of the per-AG summary counter scan loop aborts without
collecting all of the data we need, the scrubber's observation data will
be invalid.  Set the incomplete flag so that we abort the scrub without
reporting false corruptions.  Document the data dependency here too.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/scrub/fscounters.c |   21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)

Comments

Dave Chinner Oct. 14, 2022, 3:51 a.m. UTC | #1 
On Sun, Oct 02, 2022 at 11:20:05AM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> If any part of the per-AG summary counter scan loop aborts without
> collecting all of the data we need, the scrubber's observation data will
> be invalid.  Set the incomplete flag so that we abort the scrub without
> reporting false corruptions.  Document the data dependency here too.
> 
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> ---
>  fs/xfs/scrub/fscounters.c |   21 ++++++++++++++++++++-
>  1 file changed, 20 insertions(+), 1 deletion(-)

Looks good. Added comments are nice :)

Reviewed-by: Dave Chinner <dchinner@redhat.com>
diff mbox series

Patch

diff --git a/fs/xfs/scrub/fscounters.c b/fs/xfs/scrub/fscounters.c
index 3c66523ec212..c869b537fe34 100644
--- a/fs/xfs/scrub/fscounters.c
+++ b/fs/xfs/scrub/fscounters.c
@@ -145,6 +145,18 @@  xchk_setup_fscounters(
 	return xchk_trans_alloc(sc, 0);
 }
 
+/*
+ * Part 1: Collecting filesystem summary counts.  For each AG, we add its
+ * summary counts (total inodes, free inodes, free data blocks) to an incore
+ * copy of the overall filesystem summary counts.
+ *
+ * To avoid false corruption reports in part 2, any failure in this part must
+ * set the INCOMPLETE flag even when a negative errno is returned.  This care
+ * must be taken with certain errno values (i.e. EFSBADCRC, EFSCORRUPTED,
+ * ECANCELED) that are absorbed into a scrub state flag update by
+ * xchk_*_process_error.
+ */
+
 /* Count free space btree blocks manually for pre-lazysbcount filesystems. */
 static int
 xchk_fscount_btreeblks(
@@ -232,8 +244,10 @@  xchk_fscount_aggregate_agcounts(
 	}
 	if (pag)
 		xfs_perag_put(pag);
-	if (error)
+	if (error) {
+		xchk_set_incomplete(sc);
 		return error;
+	}
 
 	/*
 	 * The global incore space reservation is taken from the incore
@@ -274,6 +288,11 @@  xchk_fscount_aggregate_agcounts(
 	return 0;
 }
 
+/*
+ * Part 2: Comparing filesystem summary counters.  All we have to do here is
+ * sum the percpu counters and compare them to what we've observed.
+ */
+
 /*
  * Is the @counter reasonably close to the @expected value?
  *