[PATCH 5.8 320/464] RDMA/rtrs-clt: add an additional random 8 seconds before reconnecting

From: Greg Kroah-Hartman
Date: Mon Aug 17 2020 - 15:24:15 EST


From: Danil Kipnis <danil.kipnis@xxxxxxxxxxxxxxx>

[ Upstream commit 09e0dbbeed82e35ce2cd21e086a6fac934163e2a ]

In order to avoid all the clients to start reconnecting at the same time
schedule the reconnect dwork with a random jitter of +[0,8] seconds.

Fixes: 6a98d71daea1 ("RDMA/rtrs: client: main functionality")
Link: https://lore.kernel.org/r/20200724111508.15734-2-haris.iqbal@xxxxxxxxxxxxxxx
Signed-off-by: Danil Kipnis <danil.kipnis@xxxxxxxxxxxxxxx>
Signed-off-by: Md Haris Iqbal <haris.iqbal@xxxxxxxxxxxxxxx>
Signed-off-by: Jason Gunthorpe <jgg@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/infiniband/ulp/rtrs/rtrs-clt.c | 14 ++++++++++++--
1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c
index 564388a85603f..5b31d3b03737c 100644
--- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c
+++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c
@@ -12,6 +12,7 @@

#include <linux/module.h>
#include <linux/rculist.h>
+#include <linux/random.h>

#include "rtrs-clt.h"
#include "rtrs-log.h"
@@ -23,6 +24,12 @@
* leads to "false positives" failed reconnect attempts
*/
#define RTRS_RECONNECT_BACKOFF 1000
+/*
+ * Wait for additional random time between 0 and 8 seconds
+ * before starting to reconnect to avoid clients reconnecting
+ * all at once in case of a major network outage
+ */
+#define RTRS_RECONNECT_SEED 8

MODULE_DESCRIPTION("RDMA Transport Client");
MODULE_LICENSE("GPL");
@@ -306,7 +313,8 @@ static void rtrs_rdma_error_recovery(struct rtrs_clt_con *con)
*/
delay_ms = clt->reconnect_delay_sec * 1000;
queue_delayed_work(rtrs_wq, &sess->reconnect_dwork,
- msecs_to_jiffies(delay_ms));
+ msecs_to_jiffies(delay_ms +
+ prandom_u32() % RTRS_RECONNECT_SEED));
} else {
/*
* Error can happen just on establishing new connection,
@@ -2503,7 +2511,9 @@ static void rtrs_clt_reconnect_work(struct work_struct *work)
sess->stats->reconnects.fail_cnt++;
delay_ms = clt->reconnect_delay_sec * 1000;
queue_delayed_work(rtrs_wq, &sess->reconnect_dwork,
- msecs_to_jiffies(delay_ms));
+ msecs_to_jiffies(delay_ms +
+ prandom_u32() %
+ RTRS_RECONNECT_SEED));
}
}

--
2.25.1