netdev - [PATCH net-next 3/8] net/smc: improve abnormal termination locking

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20191021141315.58969-4-kgraul@linux.ibm.com>
Date:   Mon, 21 Oct 2019 16:13:10 +0200
From:   Karsten Graul <kgraul@...ux.ibm.com>
To:     davem@...emloft.net
Cc:     netdev@...r.kernel.org, linux-s390@...r.kernel.org,
        gor@...ux.ibm.com, heiko.carstens@...ibm.com, raspl@...ux.ibm.com,
        ubraun@...ux.ibm.com
Subject: [PATCH net-next 3/8] net/smc: improve abnormal termination locking

From: Ursula Braun <ubraun@...ux.ibm.com>

Locking hierarchy requires that the link group conns_lock can be
taken if the socket lock is held, but not vice versa. Nevertheless
socket termination during abnormal link group termination should
be protected by the socket lock.
This patch reduces the time segments the link group conns_lock is
held to enable usage of lock_sock in smc_lgr_terminate().

Signed-off-by: Ursula Braun <ubraun@...ux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@...ux.ibm.com>
---
 net/smc/smc_core.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/net/smc/smc_core.c b/net/smc/smc_core.c
index b53ba8f0a833..1f58cd82928c 100644
--- a/net/smc/smc_core.c
+++ b/net/smc/smc_core.c
@@ -491,23 +491,26 @@ static void __smc_lgr_terminate(struct smc_link_group *lgr)
 	if (!lgr->is_smcd)
 		smc_llc_link_inactive(&lgr->lnk[SMC_SINGLE_LINK]);
 
-	write_lock_bh(&lgr->conns_lock);
+	/* kill remaining link group connections */
+	read_lock_bh(&lgr->conns_lock);
 	node = rb_first(&lgr->conns_all);
 	while (node) {
+		read_unlock_bh(&lgr->conns_lock);
 		conn = rb_entry(node, struct smc_connection, alert_node);
 		smc = container_of(conn, struct smc_sock, conn);
+		lock_sock(&smc->sk);
 		sock_hold(&smc->sk); /* sock_put in close work */
 		conn->killed = 1;
 		conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
-		__smc_lgr_unregister_conn(conn);
+		smc_lgr_unregister_conn(conn);
 		conn->lgr = NULL;
-		write_unlock_bh(&lgr->conns_lock);
 		if (!schedule_work(&conn->close_work))
 			sock_put(&smc->sk);
-		write_lock_bh(&lgr->conns_lock);
+		release_sock(&smc->sk);
+		read_lock_bh(&lgr->conns_lock);
 		node = rb_first(&lgr->conns_all);
 	}
-	write_unlock_bh(&lgr->conns_lock);
+	read_unlock_bh(&lgr->conns_lock);
 	if (!lgr->is_smcd)
 		wake_up(&lgr->lnk[SMC_SINGLE_LINK].wr_reg_wait);
 	smc_lgr_schedule_free_work(lgr);
-- 
2.17.1