From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759476Ab3GaJtr (ORCPT ); Wed, 31 Jul 2013 05:49:47 -0400 Received: from mail09.linbit.com ([212.69.161.110]:39360 "EHLO mail09.linbit.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759302Ab3GaJtU (ORCPT ); Wed, 31 Jul 2013 05:49:20 -0400 From: Philipp Reisner To: linux-kernel@vger.kernel.org, Jens Axboe Cc: drbd-dev@lists.linbit.com Subject: [PATCH 2/5] drbd: Fix an connection drop issue after enabling allow-two-primaries Date: Wed, 31 Jul 2013 11:41:12 +0200 Message-Id: <1375263675-18410-3-git-send-email-philipp.reisner@linbit.com> X-Mailer: git-send-email 1.7.9.5 In-Reply-To: <1375263675-18410-1-git-send-email-philipp.reisner@linbit.com> References: <1375263675-18410-1-git-send-email-philipp.reisner@linbit.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Since drbd-8.4.0 it is possible to change the allow-two-primaries network option while the connection is established. The sequence code used to partially order packets from the data socket with packets from the meta-data socket, still assued that the allow-two-primaries option is constant while the connection is established. I.e. On a node that has the RESOLVE_CONFLICTS bits set, after enabling allow-two-primaries, when receiving the next data packet it timed out while waiting for the necessary packets on the data socket to arrive (wait_for_and_update_peer_seq() function). Fixed that by always tracking the sequence number, but only waiting for it if allow-two-primaries is set. Signed-off-by: Philipp Reisner Signed-off-by: Lars Ellenberg --- drivers/block/drbd/drbd_receiver.c | 39 +++++++++++++++--------------------- 1 file changed, 16 insertions(+), 23 deletions(-) diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c index cc29cd3..12c59eb 100644 --- a/drivers/block/drbd/drbd_receiver.c +++ b/drivers/block/drbd/drbd_receiver.c @@ -1890,29 +1890,11 @@ static u32 seq_max(u32 a, u32 b) return seq_greater(a, b) ? a : b; } -static bool need_peer_seq(struct drbd_conf *mdev) -{ - struct drbd_tconn *tconn = mdev->tconn; - int tp; - - /* - * We only need to keep track of the last packet_seq number of our peer - * if we are in dual-primary mode and we have the resolve-conflicts flag set; see - * handle_write_conflicts(). - */ - - rcu_read_lock(); - tp = rcu_dereference(mdev->tconn->net_conf)->two_primaries; - rcu_read_unlock(); - - return tp && test_bit(RESOLVE_CONFLICTS, &tconn->flags); -} - static void update_peer_seq(struct drbd_conf *mdev, unsigned int peer_seq) { unsigned int newest_peer_seq; - if (need_peer_seq(mdev)) { + if (test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags)) { spin_lock(&mdev->peer_seq_lock); newest_peer_seq = seq_max(mdev->peer_seq, peer_seq); mdev->peer_seq = newest_peer_seq; @@ -1972,22 +1954,31 @@ static int wait_for_and_update_peer_seq(struct drbd_conf *mdev, const u32 peer_s { DEFINE_WAIT(wait); long timeout; - int ret; + int ret = 0, tp; - if (!need_peer_seq(mdev)) + if (!test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags)) return 0; spin_lock(&mdev->peer_seq_lock); for (;;) { if (!seq_greater(peer_seq - 1, mdev->peer_seq)) { mdev->peer_seq = seq_max(mdev->peer_seq, peer_seq); - ret = 0; break; } + if (signal_pending(current)) { ret = -ERESTARTSYS; break; } + + rcu_read_lock(); + tp = rcu_dereference(mdev->tconn->net_conf)->two_primaries; + rcu_read_unlock(); + + if (!tp) + break; + + /* Only need to wait if two_primaries is enabled */ prepare_to_wait(&mdev->seq_wait, &wait, TASK_INTERRUPTIBLE); spin_unlock(&mdev->peer_seq_lock); rcu_read_lock(); @@ -2228,8 +2219,10 @@ static int receive_Data(struct drbd_tconn *tconn, struct packet_info *pi) } goto out_interrupted; } - } else + } else { + update_peer_seq(mdev, peer_seq); spin_lock_irq(&mdev->tconn->req_lock); + } list_add(&peer_req->w.list, &mdev->active_ee); spin_unlock_irq(&mdev->tconn->req_lock); -- 1.7.9.5