drbd: fix hang on local read errors while disconnected

"canceled" w_read_retry_remote never completed, if they have been
canceled after drbd_disconnect connection teardown cleanup has already
run (or we are currently not connected anyways).

Fixed by not queueing a remote retry if we already know it won't work
(pdsk not uptodate), and cleanup ourselves on "cancel", in case we hit a
race with drbd_disconnect.

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

diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c
index e6c4d579eaba0146f0dc06be5aeb7b814f7c4eb2..8915644af722cb0620a474816ec14d00885427d7 100644 (file)
--- a/drivers/block/drbd/drbd_req.c
+++ b/drivers/block/drbd/drbd_req.c
@@ -452,20 +452,21 @@ void __req_mod(struct drbd_request *req, enum drbd_req_event what,
 
                dev_alert(DEV, "Local READ failed sec=%llus size=%u\n",
                      (unsigned long long)req->sector, req->size);
-               /* _req_mod(req,to_be_send); oops, recursion... */
                D_ASSERT(!(req->rq_state & RQ_NET_MASK));
-               req->rq_state |= RQ_NET_PENDING;
-               inc_ap_pending(mdev);
 
                __drbd_chk_io_error(mdev, FALSE);
                put_ldev(mdev);
-               /* NOTE: if we have no connection,
-                * or know the peer has no good data either,
-                * then we don't actually need to "queue_for_net_read",
-                * but we do so anyways, since the drbd_io_error()
-                * and the potential state change to "Diskless"
-                * needs to be done from process context */
 
+               /* no point in retrying if there is no good remote data,
+                * or we have no connection. */
+               if (mdev->state.pdsk != D_UP_TO_DATE) {
+                       _req_may_be_done(req, m);
+                       break;
+               }
+
+               /* _req_mod(req,to_be_send); oops, recursion... */
+               req->rq_state |= RQ_NET_PENDING;
+               inc_ap_pending(mdev);
                /* fall through: _req_mod(req,queue_for_net_read); */
 
        case queue_for_net_read:
@@ -575,6 +576,9 @@ void __req_mod(struct drbd_request *req, enum drbd_req_event what,
                _req_may_be_done(req, m);
                break;
 
+       case read_retry_remote_canceled:
+               req->rq_state &= ~RQ_NET_QUEUED;
+               /* fall through, in case we raced with drbd_disconnect */
        case connection_lost_while_pending:
                /* transfer log cleanup after connection loss */
                /* assert something? */

diff --git a/drivers/block/drbd/drbd_req.h b/drivers/block/drbd/drbd_req.h
index 16119d7056cc65e1eae0e97245eaf7d28f78e884..02d575d245187b301e223cfc8ef5859990a637f6 100644 (file)
--- a/drivers/block/drbd/drbd_req.h
+++ b/drivers/block/drbd/drbd_req.h
@@ -91,6 +91,7 @@ enum drbd_req_event {
        send_failed,
        handed_over_to_network,
        connection_lost_while_pending,
+       read_retry_remote_canceled,
        recv_acked_by_peer,
        write_acked_by_peer,
        write_acked_by_peer_and_sis, /* and set_in_sync */

diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index a12b447bafbd9104d2e42b204ced2ef67b323e9d..67371fcbc5aa63935851ef3e94678dac54189d9a 100644 (file)
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -266,10 +266,8 @@ int w_read_retry_remote(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
         * to give the disk the chance to relocate that block */
 
        spin_lock_irq(&mdev->req_lock);
-       if (cancel ||
-           mdev->state.conn < C_CONNECTED ||
-           mdev->state.pdsk <= D_INCONSISTENT) {
-               _req_mod(req, send_canceled);
+       if (cancel || mdev->state.pdsk != D_UP_TO_DATE) {
+               _req_mod(req, read_retry_remote_canceled);
                spin_unlock_irq(&mdev->req_lock);
                dev_alert(DEV, "WE ARE LOST. Local IO failure, no peer.\n");
                return 1;