drbd: Concurrent write detection fix
Commit 9b1e63e changed the concurrent write detection algorithm to only insert
peer requests into write_requests tree after determining that there is no
conflict. With this change, new conflicting local requests could be added
while the algorithm runs, but this case was not handled correctly. Instead of
making the algorithm deal with this case, switch back to adding peer requests
to the write_requests tree immediately: this improves fairness.
When a peer request is discarded, remove that request from the write_requests
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_receiver.c b/drivers/block/drbd/drbd_receiver.c
index c08a99d..c61bf12 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -1636,16 +1636,10 @@
struct drbd_peer_request *peer_req =
container_of(w, struct drbd_peer_request, w);
struct drbd_conf *mdev = w->mdev;
- int ok = 1;
+ int ok;
D_ASSERT(mdev->tconn->net_conf->wire_protocol == DRBD_PROT_C);
ok = drbd_send_ack(mdev, P_DISCARD_ACK, peer_req);
-
- spin_lock_irq(&mdev->tconn->req_lock);
- D_ASSERT(!drbd_interval_empty(&peer_req->i));
- drbd_remove_epoch_entry_interval(mdev, peer_req);
- spin_unlock_irq(&mdev->tconn->req_lock);
-
dec_unacked(mdev);
return ok;
@@ -1836,6 +1830,12 @@
spin_lock_irq(&mdev->tconn->req_lock);
+ /*
+ * Inserting the peer request into the write_requests tree will
+ * prevent new conflicting local requests from being added.
+ */
+ drbd_insert_interval(&mdev->write_requests, &peer_req->i);
+
first = 1;
for (;;) {
struct drbd_interval *i;
@@ -1844,26 +1844,26 @@
prepare_to_wait(&mdev->misc_wait, &wait,
TASK_INTERRUPTIBLE);
- i = drbd_find_overlap(&mdev->write_requests, sector, size);
- if (i) {
+ drbd_for_each_overlap(i, &mdev->write_requests, sector, size) {
+ struct drbd_request *req2;
+
+ if (i == &peer_req->i || !i->local)
+ continue;
+
/* only ALERT on first iteration,
* we may be woken up early... */
if (first)
- dev_alert(DEV, "%s[%u] Concurrent %s write detected!"
+ dev_alert(DEV, "%s[%u] Concurrent local write detected!"
" new: %llus +%u; pending: %llus +%u\n",
current->comm, current->pid,
- i->local ? "local" : "remote",
(unsigned long long)sector, size,
(unsigned long long)i->sector, i->size);
- if (i->local) {
- struct drbd_request *req2;
-
- req2 = container_of(i, struct drbd_request, i);
- if (req2->rq_state & RQ_NET_PENDING)
- ++have_unacked;
- }
+ req2 = container_of(i, struct drbd_request, i);
+ if (req2->rq_state & RQ_NET_PENDING)
+ ++have_unacked;
++have_conflict;
+ break;
}
if (!have_conflict)
break;
@@ -1872,6 +1872,7 @@
if (first && discard && have_unacked) {
dev_alert(DEV, "Concurrent write! [DISCARD BY FLAG] sec=%llus\n",
(unsigned long long)sector);
+ drbd_remove_epoch_entry_interval(mdev, peer_req);
inc_unacked(mdev);
peer_req->w.cb = e_send_discard_ack;
list_add_tail(&peer_req->w.list, &mdev->done_ee);
@@ -1888,6 +1889,7 @@
}
if (signal_pending(current)) {
+ drbd_remove_epoch_entry_interval(mdev, peer_req);
spin_unlock_irq(&mdev->tconn->req_lock);
finish_wait(&mdev->misc_wait, &wait);
goto out_interrupted;
@@ -1906,12 +1908,11 @@
* there must be none now. */
D_ASSERT(have_unacked == 0);
}
+ /* FIXME: Introduce a timeout here after which we disconnect. */
schedule();
spin_lock_irq(&mdev->tconn->req_lock);
}
finish_wait(&mdev->misc_wait, &wait);
-
- drbd_insert_interval(&mdev->write_requests, &peer_req->i);
}
list_add(&peer_req->w.list, &mdev->active_ee);