[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <1485648328-2141-49-git-send-email-jsimmons@infradead.org>
Date: Sat, 28 Jan 2017 19:05:16 -0500
From: James Simmons <jsimmons@...radead.org>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
devel@...verdev.osuosl.org,
Andreas Dilger <andreas.dilger@...el.com>,
Oleg Drokin <oleg.drokin@...el.com>
Cc: Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Lustre Development List <lustre-devel@...ts.lustre.org>,
Liang Zhen <liang.zhen@...el.com>,
James Simmons <jsimmons@...radead.org>
Subject: [PATCH 48/60] staging: lustre: ksocklnd: ignore timedout TX on closing connection
From: Liang Zhen <liang.zhen@...el.com>
ksocklnd reaper thread always tries to close the connection for the
first timedout zero-copy TX. This is wrong if this connection is
already being closed, because the reaper will see the same TX again
and again and cannot find out other timedout zero-copy TXs and close
connections for them.
Signed-off-by: Liang Zhen <liang.zhen@...el.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8867
Reviewed-on: https://review.whamcloud.com/23973
Reviewed-by: Doug Oucharek <doug.s.oucharek@...el.com>
Reviewed-by: Oleg Drokin <oleg.drokin@...el.com>
Signed-off-by: James Simmons <jsimmons@...radead.org>
---
drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 16 +++++++++-------
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
index df4f55e..b7043e2 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
@@ -2456,6 +2456,7 @@ void ksocknal_write_callback(struct ksock_conn *conn)
list_for_each_entry(peer, peers, ksnp_list) {
unsigned long deadline = 0;
+ struct ksock_tx *tx_stale;
int resid = 0;
int n = 0;
@@ -2503,6 +2504,7 @@ void ksocknal_write_callback(struct ksock_conn *conn)
if (list_empty(&peer->ksnp_zc_req_list))
continue;
+ tx_stale = NULL;
spin_lock(&peer->ksnp_lock);
list_for_each_entry(tx, &peer->ksnp_zc_req_list, tx_zc_list) {
if (!cfs_time_aftereq(cfs_time_current(),
@@ -2511,26 +2513,26 @@ void ksocknal_write_callback(struct ksock_conn *conn)
/* ignore the TX if connection is being closed */
if (tx->tx_conn->ksnc_closing)
continue;
+ if (!tx_stale)
+ tx_stale = tx;
n++;
}
- if (!n) {
+ if (!tx_stale) {
spin_unlock(&peer->ksnp_lock);
continue;
}
- tx = list_entry(peer->ksnp_zc_req_list.next,
- struct ksock_tx, tx_zc_list);
- deadline = tx->tx_deadline;
- resid = tx->tx_resid;
- conn = tx->tx_conn;
+ deadline = tx_stale->tx_deadline;
+ resid = tx_stale->tx_resid;
+ conn = tx_stale->tx_conn;
ksocknal_conn_addref(conn);
spin_unlock(&peer->ksnp_lock);
read_unlock(&ksocknal_data.ksnd_global_lock);
CERROR("Total %d stale ZC_REQs for peer %s detected; the oldest(%p) timed out %ld secs ago, resid: %d, wmem: %d\n",
- n, libcfs_nid2str(peer->ksnp_id.nid), tx,
+ n, libcfs_nid2str(peer->ksnp_id.nid), tx_stale,
cfs_duration_sec(cfs_time_current() - deadline),
resid, conn->ksnc_sock->sk->sk_wmem_queued);
--
1.8.3.1
Powered by blists - more mailing lists