Commit fcf2959d authored by Jack Wang's avatar Jack Wang Committed by Jason Gunthorpe
Browse files

RDMA/rtrs-clt: Avoid run destroy_con_cq_qp/create_con_cq_qp in parallel

It could happen two kworkers race with each other:

        CPU0                             CPU1
    addr_resolver kworker           reconnect kworker
    rtrs_clt_rdma_cm_handler
    rtrs_rdma_addr_resolved
    create_con_cq_qp: s.dev_ref++
    "s.dev_ref is 1"
                                    wait in create_cm fails with TIMEOUT
                                    destroy_con_cq_qp: --s.dev_ref
                                    "s.dev_ref is 0"
                                    destroy_con_cq_qp: sess->s.dev = NULL
     rtrs_cq_qp_create -> create_qp(con, sess->dev->ib_pd...)
    sess->dev is NULL, panic.

To fix the problem using mutex to serialize create_con_cq_qp and
destroy_con_cq_qp.

Fixes: 6a98d71d ("RDMA/rtrs: client: main functionality")
Link: https://lore.kernel.org/r/20201023074353.21946-4-jinpu.wang@cloud.ionos.com


Signed-off-by: default avatarJack Wang <jinpu.wang@cloud.ionos.com>
Reviewed-by: default avatarGioh Kim <gi-oh.kim@cloud.ionos.com>
Signed-off-by: default avatarJason Gunthorpe <jgg@nvidia.com>
parent 73385fdb
Loading
Loading
Loading
Loading
+13 −2
Original line number Diff line number Diff line
@@ -1499,6 +1499,7 @@ static int create_con(struct rtrs_clt_sess *sess, unsigned int cid)
	con->c.cid = cid;
	con->c.sess = &sess->s;
	atomic_set(&con->io_cnt, 0);
	mutex_init(&con->con_mutex);

	sess->s.con[cid] = &con->c;

@@ -1510,6 +1511,7 @@ static void destroy_con(struct rtrs_clt_con *con)
	struct rtrs_clt_sess *sess = to_clt_sess(con->c.sess);

	sess->s.con[con->c.cid] = NULL;
	mutex_destroy(&con->con_mutex);
	kfree(con);
}

@@ -1520,6 +1522,7 @@ static int create_con_cq_qp(struct rtrs_clt_con *con)
	int err, cq_vector;
	struct rtrs_msg_rkey_rsp *rsp;

	lockdep_assert_held(&con->con_mutex);
	if (con->c.cid == 0) {
		/*
		 * One completion for each receive and two for each send
@@ -1593,7 +1596,7 @@ static void destroy_con_cq_qp(struct rtrs_clt_con *con)
	 * Be careful here: destroy_con_cq_qp() can be called even
	 * create_con_cq_qp() failed, see comments there.
	 */

	lockdep_assert_held(&con->con_mutex);
	rtrs_cq_qp_destroy(&con->c);
	if (con->rsp_ius) {
		rtrs_iu_free(con->rsp_ius, DMA_FROM_DEVICE,
@@ -1625,7 +1628,9 @@ static int rtrs_rdma_addr_resolved(struct rtrs_clt_con *con)
	struct rtrs_sess *s = con->c.sess;
	int err;

	mutex_lock(&con->con_mutex);
	err = create_con_cq_qp(con);
	mutex_unlock(&con->con_mutex);
	if (err) {
		rtrs_err(s, "create_con_cq_qp(), err: %d\n", err);
		return err;
@@ -1938,8 +1943,9 @@ static int create_cm(struct rtrs_clt_con *con)

errr:
	stop_cm(con);
	/* Is safe to call destroy if cq_qp is not inited */
	mutex_lock(&con->con_mutex);
	destroy_con_cq_qp(con);
	mutex_unlock(&con->con_mutex);
destroy_cm:
	destroy_cm(con);

@@ -2046,7 +2052,9 @@ static void rtrs_clt_stop_and_destroy_conns(struct rtrs_clt_sess *sess)
		if (!sess->s.con[cid])
			break;
		con = to_clt_con(sess->s.con[cid]);
		mutex_lock(&con->con_mutex);
		destroy_con_cq_qp(con);
		mutex_unlock(&con->con_mutex);
		destroy_cm(con);
		destroy_con(con);
	}
@@ -2213,7 +2221,10 @@ static int init_conns(struct rtrs_clt_sess *sess)
		struct rtrs_clt_con *con = to_clt_con(sess->s.con[cid]);

		stop_cm(con);

		mutex_lock(&con->con_mutex);
		destroy_con_cq_qp(con);
		mutex_unlock(&con->con_mutex);
		destroy_cm(con);
		destroy_con(con);
	}
+1 −0
Original line number Diff line number Diff line
@@ -72,6 +72,7 @@ struct rtrs_clt_con {
	struct rtrs_iu		*rsp_ius;
	u32			queue_size;
	unsigned int		cpu;
	struct mutex		con_mutex;
	atomic_t		io_cnt;
	int			cm_err;
};