dm exception store: fix misordered writes

We must zero the next chunk on disk *before* writing out the current chunk, not
after.  Otherwise if the machine crashes at the wrong time, the "end of metadata"
marker may be missing.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Cc: stable@kernel.org

diff --git a/drivers/md/dm-exception-store.c b/drivers/md/dm-exception-store.c
index fe6cef8df203aa697ef3bf82e68cd2cacd1cc818..6179bf70f98a7837a1828638c9f5bac310ae2ad6 100644 (file)
--- a/drivers/md/dm-exception-store.c
+++ b/drivers/md/dm-exception-store.c
@@ -610,18 +610,23 @@ static void persistent_commit(struct exception_store *store,
            (ps->current_committed != ps->exceptions_per_area))
                return;
 
+       /*
+        * If we completely filled the current area, then wipe the next one.
+        */
+       if ((ps->current_committed == ps->exceptions_per_area) &&
+            zero_disk_area(ps, ps->current_area + 1))
+               ps->valid = 0;
+
        /*
         * Commit exceptions to disk.
         */
-       if (area_io(ps, WRITE))
+       if (ps->valid && area_io(ps, WRITE))
                ps->valid = 0;
 
        /*
         * Advance to the next area if this one is full.
         */
        if (ps->current_committed == ps->exceptions_per_area) {
-               if (zero_disk_area(ps, ps->current_area + 1))
-                       ps->valid = 0;
                ps->current_committed = 0;
                ps->current_area++;
                zero_memory_area(ps);