All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] libceph: cleanup old messages according to reconnect seq
@ 2017-05-05 11:08 Yan, Zheng
  2017-05-24 15:26 ` Ilya Dryomov
  0 siblings, 1 reply; 2+ messages in thread
From: Yan, Zheng @ 2017-05-05 11:08 UTC (permalink / raw)
  To: ceph-devel; +Cc: jlayton, idryomov, Yan, Zheng

when reopen a connection, use 'reconnect seq' to clean up
messages that have already been received by peer.

Fixes: http://tracker.ceph.com/issues/18690
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
---
 net/ceph/messenger.c | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index f76bb33..af9c5a1 100644
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -2230,10 +2230,17 @@ static void process_ack(struct ceph_connection *con)
 	struct ceph_msg *m;
 	u64 ack = le64_to_cpu(con->in_temp_ack);
 	u64 seq;
-
-	while (!list_empty(&con->out_sent)) {
-		m = list_first_entry(&con->out_sent, struct ceph_msg,
-				     list_head);
+	bool reconnect = (con->in_tag == CEPH_MSGR_TAG_SEQ);
+	struct list_head *list = reconnect ? &con->out_queue : &con->out_sent;
+	/*
+	 * In the reconnect case, con_fault() has requeued messages
+	 * in out_sent. We should cleanup old messages according to
+	 * the reconnect seq.
+	 */
+	while (!list_empty(list)) {
+		m = list_first_entry(list, struct ceph_msg, list_head);
+		if (reconnect && m->needs_out_seq)
+			break;
 		seq = le64_to_cpu(m->hdr.seq);
 		if (seq > ack)
 			break;
@@ -2242,6 +2249,7 @@ static void process_ack(struct ceph_connection *con)
 		m->ack_stamp = jiffies;
 		ceph_msg_remove(m);
 	}
+
 	prepare_read_tag(con);
 }
 
-- 
2.9.3


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH] libceph: cleanup old messages according to reconnect seq
  2017-05-05 11:08 [PATCH] libceph: cleanup old messages according to reconnect seq Yan, Zheng
@ 2017-05-24 15:26 ` Ilya Dryomov
  0 siblings, 0 replies; 2+ messages in thread
From: Ilya Dryomov @ 2017-05-24 15:26 UTC (permalink / raw)
  To: Yan, Zheng; +Cc: Ceph Development, Jeff Layton, idryomov

On Fri, May 5, 2017 at 1:08 PM, Yan, Zheng <zyan@redhat.com> wrote:
> when reopen a connection, use 'reconnect seq' to clean up
> messages that have already been received by peer.
>
> Fixes: http://tracker.ceph.com/issues/18690
> Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
> ---
>  net/ceph/messenger.c | 16 ++++++++++++----
>  1 file changed, 12 insertions(+), 4 deletions(-)
>
> diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
> index f76bb33..af9c5a1 100644
> --- a/net/ceph/messenger.c
> +++ b/net/ceph/messenger.c
> @@ -2230,10 +2230,17 @@ static void process_ack(struct ceph_connection *con)
>         struct ceph_msg *m;
>         u64 ack = le64_to_cpu(con->in_temp_ack);
>         u64 seq;
> -
> -       while (!list_empty(&con->out_sent)) {
> -               m = list_first_entry(&con->out_sent, struct ceph_msg,
> -                                    list_head);
> +       bool reconnect = (con->in_tag == CEPH_MSGR_TAG_SEQ);
> +       struct list_head *list = reconnect ? &con->out_queue : &con->out_sent;
> +       /*
> +        * In the reconnect case, con_fault() has requeued messages
> +        * in out_sent. We should cleanup old messages according to
> +        * the reconnect seq.
> +        */
> +       while (!list_empty(list)) {
> +               m = list_first_entry(list, struct ceph_msg, list_head);
> +               if (reconnect && m->needs_out_seq)
> +                       break;
>                 seq = le64_to_cpu(m->hdr.seq);
>                 if (seq > ack)
>                         break;
> @@ -2242,6 +2249,7 @@ static void process_ack(struct ceph_connection *con)
>                 m->ack_stamp = jiffies;
>                 ceph_msg_remove(m);
>         }
> +
>         prepare_read_tag(con);
>  }
>

Reviewed-by: Ilya Dryomov <idryomov@gmail.com>

Thanks,

                Ilya

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2017-05-24 15:26 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-05-05 11:08 [PATCH] libceph: cleanup old messages according to reconnect seq Yan, Zheng
2017-05-24 15:26 ` Ilya Dryomov

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.