dm raid1: fail writes if errors are not handled and log fails
authorMikulas Patocka <mpatocka@redhat.com>
Tue, 16 Feb 2010 18:42:55 +0000 (18:42 +0000)
committerAlasdair G Kergon <agk@redhat.com>
Tue, 16 Feb 2010 18:42:55 +0000 (18:42 +0000)
If the mirror log fails when the handle_errors option was not selected
and there is no remaining valid mirror leg, writes return success even
though they weren't actually written to any device.  This patch
completes them with EIO instead.

This code path is taken:
do_writes:
bio_list_merge(&ms->failures, &sync);
do_failures:
if (!get_valid_mirror(ms)) (false)
else if (errors_handled(ms)) (false)
else bio_endio(bio, 0);

The logic in do_failures is based on presuming that the write was already
tried: if it succeeded at least on one leg (without handle_errors) it
is reported as success.

Reference: https://bugzilla.redhat.com/show_bug.cgi?id=555197

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
drivers/md/dm-raid1.c

index ad779bd13aec43ac6aba859be57a8e94d01a5001..6c1046df81f6f7738cd74eec6a6d5558a8e23f13 100644 (file)
@@ -724,7 +724,7 @@ static void do_writes(struct mirror_set *ms, struct bio_list *writes)
        /*
         * Dispatch io.
         */
-       if (unlikely(ms->log_failure)) {
+       if (unlikely(ms->log_failure) && errors_handled(ms)) {
                spin_lock_irq(&ms->lock);
                bio_list_merge(&ms->failures, &sync);
                spin_unlock_irq(&ms->lock);