drbd: do not reset rs_pending_cnt too early

Fix asserts like
  block drbd0: in got_BlockAck:4634: rs_pending_cnt = -35 < 0 !

We reset the resync lru cache and related information (rs_pending_cnt),
once we successfully finished a resync or online verify, or if the
replication connection is lost.

We also need to reset it if a resync or online verify is aborted
because a lower level disk failed.

In that case the replication link is still established,
and we may still have packets queued in the network buffers
which want to touch rs_pending_cnt.

We do not have any synchronization mechanism to know for sure when all
such pending resync related packets have been drained.

To avoid this counter to go negative (and violate the ASSERT that it
will always be >= 0), just do not reset it when we lose a disk.

It is good enough to make sure it is re-initialized before the next
resync can start: reset it when we re-attach a disk.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>

diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 41ccb580d5ac81d311d102414439a4e1b66d70af..91a4853d9eeb36a8de1f4fd7464ad105ff718a85 100644 (file)
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -1514,6 +1514,13 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 
        /* Do not change the order of the if above and the two below... */
        if (os.pdsk == D_DISKLESS && ns.pdsk > D_DISKLESS) {      /* attach on the peer */
+               /* we probably will start a resync soon.
+                * make sure those things are properly reset. */
+               mdev->rs_total = 0;
+               mdev->rs_failed = 0;
+               atomic_set(&mdev->rs_pending_cnt, 0);
+               drbd_rs_cancel_all(mdev);
+
                drbd_send_uuids(mdev);
                drbd_send_state(mdev, ns);
        }
@@ -1681,10 +1688,6 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
                                 "ASSERT FAILED: disk is %s while going diskless\n",
                                 drbd_disk_str(mdev->state.disk));
 
-                mdev->rs_total = 0;
-                mdev->rs_failed = 0;
-                atomic_set(&mdev->rs_pending_cnt, 0);
-
                if (ns.conn >= C_CONNECTED)
                        drbd_send_state(mdev, ns);
 

diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index 03fc853be2b91a1e7b70d266df3d2ceda1f5771b..a68d9bfb731cd4c6d90cd9ed3002eb6c8694f9c6 100644 (file)
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -959,6 +959,11 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
        /* make sure there is no leftover from previous force-detach attempts */
        clear_bit(FORCE_DETACH, &mdev->flags);
 
+       /* and no leftover from previously aborted resync or verify, either */
+       mdev->rs_total = 0;
+       mdev->rs_failed = 0;
+       atomic_set(&mdev->rs_pending_cnt, 0);
+
        /* allocation not in the IO path, cqueue thread context */
        nbc = kzalloc(sizeof(struct drbd_backing_dev), GFP_KERNEL);
        if (!nbc) {

diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index a35393f2fd1bc4cf9a771b32a3cdef45e52ee016..6bce2cc179d4112980673b1b9d82ae91d70d390f 100644 (file)
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -1501,14 +1501,6 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side)
                return;
        }
 
-       if (mdev->state.conn < C_AHEAD) {
-               /* In case a previous resync run was aborted by an IO error/detach on the peer. */
-               drbd_rs_cancel_all(mdev);
-               /* This should be done when we abort the resync. We definitely do not
-                  want to have this for connections going back and forth between
-                  Ahead/Behind and SyncSource/SyncTarget */
-       }
-
        if (side == C_SYNC_TARGET) {
                /* Since application IO was locked out during C_WF_BITMAP_T and
                   C_WF_SYNC_UUID we are still unmodified. Before going to C_SYNC_TARGET