drbd: Fixes from the drbd-8.3 branch

* drbd-8.3:
  drbd: fix spurious meta data IO "error"
  drbd: Fixed a race condition between detach and start of resync
  drbd: fix harmless race to not trigger an ASSERT
  drbd: Derive sync-UUIDs only from the bitmap-uuid if it is non-zero
  drbd: Fixed current UUID generation (regression introduced recently, after 8.3.11)

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

4 files changed, 26 insertions, 3 deletions

diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index edd0227f4b43..822fb3d42356 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -1087,7 +1087,11 @@ void drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev)
 
 	D_ASSERT(mdev->state.disk == D_UP_TO_DATE);
 
-	uuid = mdev->ldev->md.uuid[UI_BITMAP] + UUID_NEW_BM_OFFSET;
+	uuid = mdev->ldev->md.uuid[UI_BITMAP];
+	if (uuid && uuid != UUID_JUST_CREATED)
+		uuid = uuid + UUID_NEW_BM_OFFSET;
+	else
+		get_random_bytes(&uuid, sizeof(u64));
 	drbd_uuid_set(mdev, UI_BITMAP, uuid);
 	drbd_print_uuids(mdev, "updated sync UUID");
 	drbd_md_sync(mdev);
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index 16c3710e1b9c..c50c1753aa9a 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -1666,7 +1666,9 @@ static int adm_detach(struct drbd_conf *mdev, int force)
 	}
 
 	drbd_suspend_io(mdev); /* so no-one is stuck in drbd_al_begin_io */
+	drbd_md_get_buffer(mdev); /* make sure there is no in-flight meta-data IO */
 	retcode = drbd_request_state(mdev, NS(disk, D_FAILED));
+	drbd_md_put_buffer(mdev);
 	/* D_FAILED will transition to DISKLESS. */
 	ret = wait_event_interruptible(mdev->misc_wait,
 			mdev->state.disk != D_FAILED);
diff --git a/drivers/block/drbd/drbd_state.c b/drivers/block/drbd/drbd_state.c
index 4075bd2d2515..dffc6973e2b5 100644
--- a/drivers/block/drbd/drbd_state.c
+++ b/drivers/block/drbd/drbd_state.c
@@ -1231,6 +1231,11 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	}
 
 	if (ns.pdsk < D_INCONSISTENT && get_ldev(mdev)) {
+		if (os.peer == R_SECONDARY && ns.peer == R_PRIMARY &&
+		    mdev->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) {
+			drbd_uuid_new_current(mdev);
+			drbd_send_uuids(mdev);
+		}
 		/* D_DISKLESS Peer becomes secondary */
 		if (os.peer == R_PRIMARY && ns.peer == R_SECONDARY)
 			/* We may still be Primary ourselves.
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index 34a6065d95e6..bc2cfd6c600f 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -74,10 +74,21 @@ void drbd_md_io_complete(struct bio *bio, int error)
 
 	md_io->error = error;
 
+	/* We grabbed an extra reference in _drbd_md_sync_page_io() to be able
+	 * to timeout on the lower level device, and eventually detach from it.
+	 * If this io completion runs after that timeout expired, this
+	 * drbd_md_put_buffer() may allow us to finally try and re-attach.
+	 * During normal operation, this only puts that extra reference
+	 * down to 1 again.
+	 * Make sure we first drop the reference, and only then signal
+	 * completion, or we may (in drbd_al_read_log()) cycle so fast into the
+	 * next drbd_md_sync_page_io(), that we trigger the
+	 * ASSERT(atomic_read(&mdev->md_io_in_use) == 1) there.
+	 */
+	drbd_md_put_buffer(mdev);
 	md_io->done = 1;
 	wake_up(&mdev->misc_wait);
 	bio_put(bio);
-	drbd_md_put_buffer(mdev);
 	put_ldev(mdev);
 }
 
@@ -1581,12 +1592,13 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side)
 	}
 	clear_bit(B_RS_H_DONE, &mdev->flags);
 
+	write_lock_irq(&global_state_lock);
 	if (!get_ldev_if_state(mdev, D_NEGOTIATING)) {
+		write_unlock_irq(&global_state_lock);
 		mutex_unlock(mdev->state_mutex);
 		return;
 	}
 
-	write_lock_irq(&global_state_lock);
 	ns = drbd_read_state(mdev);
 
 	ns.aftr_isp = !_drbd_may_sync_now(mdev);