md/raid1: extend spinlock to protect raid1_end_read_request against inconsistencies

raid1_end_read_request() assumes that the In_sync bits are consistent
with the ->degaded count.
raid1_spare_active updates the In_sync bit before the ->degraded count
and so exposes an inconsistency, as does error()
So extend the spinlock in raid1_spare_active() and error() to hide those
inconsistencies.

This should probably be part of
  Commit: 34cab6f42003 ("md/raid1: fix test for 'was read error from
  last working device'.")
as it addresses the same issue.  It fixes the same bug and should go
to -stable for same reasons.

Fixes: 76073054c95b ("md/raid1: clean up read_balance.")
Cc: stable@vger.kernel.org (v3.0+)
Signed-off-by: NeilBrown <neilb@suse.com>

diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 94f5b55069e09610f21ea640f5dff3efd7e580ca..967a4ed73929ff44a38d9475c5e362fc2914c758 100644 (file)
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -1476,6 +1476,7 @@ static void error(struct mddev *mddev, struct md_rdev *rdev)
 {
        char b[BDEVNAME_SIZE];
        struct r1conf *conf = mddev->private;
+       unsigned long flags;
 
        /*
         * If it is not operational, then we have already marked it as dead
@@ -1495,14 +1496,13 @@ static void error(struct mddev *mddev, struct md_rdev *rdev)
                return;
        }
        set_bit(Blocked, &rdev->flags);
+       spin_lock_irqsave(&conf->device_lock, flags);
        if (test_and_clear_bit(In_sync, &rdev->flags)) {
-               unsigned long flags;
-               spin_lock_irqsave(&conf->device_lock, flags);
                mddev->degraded++;
                set_bit(Faulty, &rdev->flags);
-               spin_unlock_irqrestore(&conf->device_lock, flags);
        } else
                set_bit(Faulty, &rdev->flags);
+       spin_unlock_irqrestore(&conf->device_lock, flags);
        /*
         * if recovery is running, make sure it aborts.
         */
@@ -1568,7 +1568,10 @@ static int raid1_spare_active(struct mddev *mddev)
         * Find all failed disks within the RAID1 configuration
         * and mark them readable.
         * Called under mddev lock, so rcu protection not needed.
+        * device_lock used to avoid races with raid1_end_read_request
+        * which expects 'In_sync' flags and ->degraded to be consistent.
         */
+       spin_lock_irqsave(&conf->device_lock, flags);
        for (i = 0; i < conf->raid_disks; i++) {
                struct md_rdev *rdev = conf->mirrors[i].rdev;
                struct md_rdev *repl = conf->mirrors[conf->raid_disks + i].rdev;
@@ -1599,7 +1602,6 @@ static int raid1_spare_active(struct mddev *mddev)
                        sysfs_notify_dirent_safe(rdev->sysfs_state);
                }
        }
-       spin_lock_irqsave(&conf->device_lock, flags);
        mddev->degraded -= count;
        spin_unlock_irqrestore(&conf->device_lock, flags);