do_tls_setsockopt_tx_payload_len() refuses to resize records while one is open, but the check is broken in two ways. It reaches for the open record through tls_sw_ctx_tx(), an unchecked cast of ctx->priv_ctx_tx. Under device offload that pointer is a tls_offload_context_tx, so the check reads a field of the wrong struct: it returns EBUSY on whatever happens to be there, and never sees the record that really is open. Dispatch on tx_conf. The socket lock alone is also not enough to tell no sender is in flight. The device tx path drops it in sk_stream_wait_memory() after tls_push_record() cleared open_record, so setsockopt can slip in mid-sendmsg and shrink the limit. tls_push_data() latches the limit once per call, and a record left open by MSG_MORE can then be bigger than the new limit, making the u32 subtraction in the copy clamp wrap around and build an oversized record. Take tx_lock like the sendmsg paths do so setsockopt waits for in-flight senders. Re-reading the limit inside the tls_push_data() loop could close the wrap too, but a parked sender would still finish with the old limit, so serializing on tx_lock is simpler. Fixes: 82cb5be6ad64 ("net/tls: support setting the maximum payload size") Signed-off-by: Jiayuan Chen --- v1 -> v2: - also take tx_lock in setsockopt, the socket lock alone can't stop a device sender parked in sk_stream_wait_memory() (AI review on v1) v1: https://lore.kernel.org/netdev/20260831025607.62927-1-jiayuan.chen@linux.dev/ base on my netdevsim + tls (in progress) https://lore.kernel.org/netdev/20260728125658.390500-1-jiayuan.chen@linux.dev/ Previous finding: b17cf742eaad ("tls: device: fix out-of-bounds write in tls_append_frag()") --- net/tls/tls_main.c | 27 +++++++++++++++++++++++++-- 1 file changed, 25 insertions(+), 2 deletions(-) diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c index fbb274287aa5..114e5113bafd 100644 --- a/net/tls/tls_main.c +++ b/net/tls/tls_main.c @@ -833,15 +833,29 @@ static int do_tls_setsockopt_no_pad(struct sock *sk, sockptr_t optval, return rc; } +/* priv_ctx_tx holds a different structure on each TX path, so tx_conf has to + * say which open record to look at. + */ +static bool tls_tx_record_is_open(struct tls_context *ctx) +{ + switch (ctx->tx_conf) { + case TLS_SW: + return !!tls_sw_ctx_tx(ctx)->open_rec; + case TLS_HW: + return !!tls_offload_ctx_tx(ctx)->open_record; + default: + return false; + } +} + static int do_tls_setsockopt_tx_payload_len(struct sock *sk, sockptr_t optval, unsigned int optlen) { struct tls_context *ctx = tls_get_ctx(sk); - struct tls_sw_context_tx *sw_ctx = tls_sw_ctx_tx(ctx); u16 value; bool tls_13 = ctx->prot_info.version == TLS_1_3_VERSION; - if (sw_ctx && sw_ctx->open_rec) + if (tls_tx_record_is_open(ctx)) return -EBUSY; if (sockptr_is_null(optval) || optlen != sizeof(value)) @@ -862,6 +876,7 @@ static int do_tls_setsockopt_tx_payload_len(struct sock *sk, sockptr_t optval, static int do_tls_setsockopt(struct sock *sk, int optname, sockptr_t optval, unsigned int optlen) { + struct tls_context *ctx; int rc = 0; switch (optname) { @@ -881,9 +896,17 @@ static int do_tls_setsockopt(struct sock *sk, int optname, sockptr_t optval, rc = do_tls_setsockopt_no_pad(sk, optval, optlen); break; case TLS_TX_MAX_PAYLOAD_LEN: + /* Take tx_lock like the sendmsg paths do, the socket lock is + * dropped while a sender waits for memory, with no record open. + */ + ctx = tls_get_ctx(sk); + rc = mutex_lock_interruptible(&ctx->tx_lock); + if (rc) + break; lock_sock(sk); rc = do_tls_setsockopt_tx_payload_len(sk, optval, optlen); release_sock(sk); + mutex_unlock(&ctx->tx_lock); break; default: rc = -ENOPROTOOPT; -- 2.43.0