From 1bc449382b3d5f5a2403c4aea244073c9ed72a7c Mon Sep 17 00:00:00 2001 From: Chengfeng Ye Date: Sun, 27 Sep 2026 15:51:20 +0800 Subject: [PATCH] net/smc: Hold a socket reference for transmit work SMC transmit work is queued without holding a socket reference. After an active close has cancelled tx_work, a received CDC message can queue it again while the socket is in SMC_PEERCLOSEWAIT1. Passive close can then reach SMC_CLOSED, call smc_conn_free() and drop the last socket reference before smc_tx_work() acquires the socket lock. The worker then accesses the freed socket. KASAN reported: BUG: KASAN: slab-use-after-free in lock_sock_nested+0x97/0x180 Write of size 8 by task kworker/0:1/11 Workqueue: smc_tx_wq-00000000 smc_tx_work Call Trace: lock_sock_nested+0x97/0x180 smc_tx_work+0x5d/0x170 process_one_work+0x5ce/0xeb0 worker_thread+0x45b/0xd10 Allocated by task 24: sk_prot_alloc+0x56/0x210 sk_alloc+0x2b/0x6f0 smc_tcp_listen_work+0x16d/0xfc0 Freed by task 182: slab_free_after_rcu_debug+0xa6/0x1e0 rcu_core+0x509/0x1850 Hold a socket reference for each successful enqueue and release it when the worker finishes or a pending invocation is cancelled. Cover all three queue sites and all cancellation sites, including the socket options. Check conn->freed in smc_tx_pending() under the socket lock: retaining the socket does not retain connection resources released by smc_conn_free(). This also covers pending transmit processing from smc_release_cb(). Use queue_delayed_work() for the busy-slot retry as well. All tx_work delays are zero, so an already queued invocation needs no timer update. Unlike mod_delayed_work(), its return value distinguishes a successful enqueue from work disabled temporarily by cancel_delayed_work_sync(), allowing the extra reference to be returned when no work was queued. Cc: stable+noautosel@kernel.org # LLM report + LLM fix, not seen in real life Fixes: e6727f39004b ("smc: send data (through RDMA)") Link: https://lists.openwall.net/netdev/2026/09/09/105 Signed-off-by: Chengfeng Ye Link: https://patch.msgid.link/20260927075120.3695060-1-nicoyip.dev@gmail.com Signed-off-by: Jakub Kicinski --- net/smc/af_smc.c | 6 ++++-- net/smc/smc_close.c | 18 ++++++++++++------ net/smc/smc_core.c | 4 +++- net/smc/smc_tx.c | 18 +++++++++++++----- 4 files changed, 32 insertions(+), 14 deletions(-) diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c index 221643624ace..7f762bb31879 100644 --- a/net/smc/af_smc.c +++ b/net/smc/af_smc.c @@ -3147,7 +3147,8 @@ int smc_setsockopt(struct socket *sock, int level, int optname, if (val) { SMC_STAT_INC(smc, ndly_cnt); smc_tx_pending(&smc->conn); - cancel_delayed_work(&smc->conn.tx_work); + if (cancel_delayed_work(&smc->conn.tx_work)) + sock_put(sk); } } break; @@ -3158,7 +3159,8 @@ int smc_setsockopt(struct socket *sock, int level, int optname, if (!val) { SMC_STAT_INC(smc, cork_cnt); smc_tx_pending(&smc->conn); - cancel_delayed_work(&smc->conn.tx_work); + if (cancel_delayed_work(&smc->conn.tx_work)) + sock_put(sk); } } break; diff --git a/net/smc/smc_close.c b/net/smc/smc_close.c index bb0313ef5f7c..a8d4b4b65f24 100644 --- a/net/smc/smc_close.c +++ b/net/smc/smc_close.c @@ -118,7 +118,8 @@ static void smc_close_cancel_work(struct smc_sock *smc) release_sock(sk); if (cancel_work_sync(&smc->conn.close_work)) sock_put(sk); - cancel_delayed_work_sync(&smc->conn.tx_work); + if (cancel_delayed_work_sync(&smc->conn.tx_work)) + sock_put(sk); lock_sock(sk); } @@ -230,7 +231,8 @@ int smc_close_active(struct smc_sock *smc) case SMC_ACTIVE: smc_close_stream_wait(smc, timeout); release_sock(sk); - cancel_delayed_work_sync(&conn->tx_work); + if (cancel_delayed_work_sync(&conn->tx_work)) + sock_put(sk); lock_sock(sk); if (sk->sk_state == SMC_ACTIVE) { /* send close request */ @@ -264,7 +266,8 @@ int smc_close_active(struct smc_sock *smc) if (!smc_cdc_rxed_any_close(conn)) smc_close_stream_wait(smc, timeout); release_sock(sk); - cancel_delayed_work_sync(&conn->tx_work); + if (cancel_delayed_work_sync(&conn->tx_work)) + sock_put(sk); lock_sock(sk); if (sk->sk_state != SMC_APPCLOSEWAIT1 && sk->sk_state != SMC_APPCLOSEWAIT2) @@ -372,7 +375,8 @@ static void smc_close_passive_work(struct work_struct *work) /* peer has not received all data */ smc_close_passive_abort_received(smc); release_sock(sk); - cancel_delayed_work_sync(&conn->tx_work); + if (cancel_delayed_work_sync(&conn->tx_work)) + sock_put(sk); lock_sock(sk); goto wakeup; } @@ -462,7 +466,8 @@ int smc_close_shutdown_write(struct smc_sock *smc) case SMC_ACTIVE: smc_close_stream_wait(smc, timeout); release_sock(sk); - cancel_delayed_work_sync(&conn->tx_work); + if (cancel_delayed_work_sync(&conn->tx_work)) + sock_put(sk); lock_sock(sk); if (sk->sk_state != SMC_ACTIVE) goto again; @@ -475,7 +480,8 @@ int smc_close_shutdown_write(struct smc_sock *smc) if (!smc_cdc_rxed_any_close(conn)) smc_close_stream_wait(smc, timeout); release_sock(sk); - cancel_delayed_work_sync(&conn->tx_work); + if (cancel_delayed_work_sync(&conn->tx_work)) + sock_put(sk); lock_sock(sk); if (sk->sk_state != SMC_APPCLOSEWAIT1) goto again; diff --git a/net/smc/smc_core.c b/net/smc/smc_core.c index 9974149659c2..6021322253e9 100644 --- a/net/smc/smc_core.c +++ b/net/smc/smc_core.c @@ -1070,7 +1070,9 @@ static int smc_switch_cursor(struct smc_sock *smc, struct smc_cdc_tx_pend *pend, smc->sk.sk_state != SMC_CLOSED) { rc = smcr_cdc_msg_send_validation(conn, pend, wr_buf); if (!rc) { - queue_delayed_work(conn->lgr->tx_wq, &conn->tx_work, 0); + sock_hold(&smc->sk); + if (!queue_delayed_work(conn->lgr->tx_wq, &conn->tx_work, 0)) + sock_put(&smc->sk); smc->sk.sk_data_ready(&smc->sk); } } else { diff --git a/net/smc/smc_tx.c b/net/smc/smc_tx.c index 3144b4b1fe29..c5c00c63ce4c 100644 --- a/net/smc/smc_tx.c +++ b/net/smc/smc_tx.c @@ -570,8 +570,10 @@ static int smcr_tx_sndbuf_nonempty(struct smc_connection *conn) if (conn->killed) return -EPIPE; rc = 0; - mod_delayed_work(conn->lgr->tx_wq, &conn->tx_work, - SMC_TX_WORK_DELAY); + sock_hold(&smc->sk); + if (!queue_delayed_work(conn->lgr->tx_wq, &conn->tx_work, + SMC_TX_WORK_DELAY)) + sock_put(&smc->sk); } return rc; } @@ -667,7 +669,7 @@ void smc_tx_pending(struct smc_connection *conn) struct smc_sock *smc = container_of(conn, struct smc_sock, conn); int rc; - if (smc->sk.sk_err) + if (smc->sk.sk_err || conn->freed) return; rc = smc_tx_sndbuf_nonempty(conn); @@ -690,6 +692,7 @@ void smc_tx_work(struct work_struct *work) lock_sock(&smc->sk); smc_tx_pending(conn); release_sock(&smc->sk); + sock_put(&smc->sk); } void smc_tx_consumer_update(struct smc_connection *conn, bool force) @@ -718,8 +721,13 @@ void smc_tx_consumer_update(struct smc_connection *conn, bool force) return; if ((smc_cdc_get_slot_and_msg_send(conn) < 0) && !conn->killed) { - queue_delayed_work(conn->lgr->tx_wq, &conn->tx_work, - SMC_TX_WORK_DELAY); + struct smc_sock *smc = + container_of(conn, struct smc_sock, conn); + + sock_hold(&smc->sk); + if (!queue_delayed_work(conn->lgr->tx_wq, &conn->tx_work, + SMC_TX_WORK_DELAY)) + sock_put(&smc->sk); return; } }