[PATCH 5.15 031/277] net/smc: Send directly when TCP_CORK is cleared

From: Greg Kroah-Hartman
Date: Tue Apr 12 2022 - 03:15:49 EST


From: Tony Lu <tonylu@xxxxxxxxxxxxxxxxx>

[ Upstream commit ea785a1a573b390a150010b3c5b81e1ccd8c98a8 ]

According to the man page of TCP_CORK [1], if set, don't send out
partial frames. All queued partial frames are sent when option is
cleared again.

When applications call setsockopt to disable TCP_CORK, this call is
protected by lock_sock(), and tries to mod_delayed_work() to 0, in order
to send pending data right now. However, the delayed work smc_tx_work is
also protected by lock_sock(). There introduces lock contention for
sending data.

To fix it, send pending data directly which acts like TCP, without
lock_sock() protected in the context of setsockopt (already lock_sock()ed),
and cancel unnecessary dealyed work, which is protected by lock.

[1] https://linux.die.net/man/7/tcp

Signed-off-by: Tony Lu <tonylu@xxxxxxxxxxxxxxxxx>
Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
net/smc/af_smc.c | 4 ++--
net/smc/smc_tx.c | 25 +++++++++++++++----------
net/smc/smc_tx.h | 1 +
3 files changed, 18 insertions(+), 12 deletions(-)

diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c
index 5c4c0320e822..183b122807b6 100644
--- a/net/smc/af_smc.c
+++ b/net/smc/af_smc.c
@@ -2430,8 +2430,8 @@ static int smc_setsockopt(struct socket *sock, int level, int optname,
sk->sk_state != SMC_CLOSED) {
if (!val) {
SMC_STAT_INC(smc, cork_cnt);
- mod_delayed_work(smc->conn.lgr->tx_wq,
- &smc->conn.tx_work, 0);
+ smc_tx_pending(&smc->conn);
+ cancel_delayed_work(&smc->conn.tx_work);
}
}
break;
diff --git a/net/smc/smc_tx.c b/net/smc/smc_tx.c
index 738a4a99c827..31ee76131a79 100644
--- a/net/smc/smc_tx.c
+++ b/net/smc/smc_tx.c
@@ -594,27 +594,32 @@ int smc_tx_sndbuf_nonempty(struct smc_connection *conn)
return rc;
}

-/* Wakeup sndbuf consumers from process context
- * since there is more data to transmit
- */
-void smc_tx_work(struct work_struct *work)
+void smc_tx_pending(struct smc_connection *conn)
{
- struct smc_connection *conn = container_of(to_delayed_work(work),
- struct smc_connection,
- tx_work);
struct smc_sock *smc = container_of(conn, struct smc_sock, conn);
int rc;

- lock_sock(&smc->sk);
if (smc->sk.sk_err)
- goto out;
+ return;

rc = smc_tx_sndbuf_nonempty(conn);
if (!rc && conn->local_rx_ctrl.prod_flags.write_blocked &&
!atomic_read(&conn->bytes_to_rcv))
conn->local_rx_ctrl.prod_flags.write_blocked = 0;
+}
+
+/* Wakeup sndbuf consumers from process context
+ * since there is more data to transmit
+ */
+void smc_tx_work(struct work_struct *work)
+{
+ struct smc_connection *conn = container_of(to_delayed_work(work),
+ struct smc_connection,
+ tx_work);
+ struct smc_sock *smc = container_of(conn, struct smc_sock, conn);

-out:
+ lock_sock(&smc->sk);
+ smc_tx_pending(conn);
release_sock(&smc->sk);
}

diff --git a/net/smc/smc_tx.h b/net/smc/smc_tx.h
index 07e6ad76224a..a59f370b8b43 100644
--- a/net/smc/smc_tx.h
+++ b/net/smc/smc_tx.h
@@ -27,6 +27,7 @@ static inline int smc_tx_prepared_sends(struct smc_connection *conn)
return smc_curs_diff(conn->sndbuf_desc->len, &sent, &prep);
}

+void smc_tx_pending(struct smc_connection *conn);
void smc_tx_work(struct work_struct *work);
void smc_tx_init(struct smc_sock *smc);
int smc_tx_sendmsg(struct smc_sock *smc, struct msghdr *msg, size_t len);
--
2.35.1