[PATCH net] net/iucv: reconcile the socket state when connect() severs the path

From: Bryam Vargas <[email protected]>

A connack can land between iucv_sock_wait() returning an error and the
sever at the end of iucv_sock_connect(): iucv_callback_connack() writes
IUCV_CONNECTED from the iucv tasklet with no socket lock, and lock_sock()
does not exclude it. The sever clears iucv->path and leaves the state, so
connect() fails on a socket still claiming a connection it no longer has,
and a later sendmsg() reaches iucv->path->msglim through the NULL pointer
inside iucv_below_msglim(). The sever runs on every failed classic
connect, O_NONBLOCK included.

Reconcile only that case, so an ordinary failed connect still returns the
socket to its pre-connect state. The sever is a barrier against a later
connack: iucv_path_sever() clears the path from iucv_path_table[] under
the lock the tasklet holds across dispatch.

Fixes: 18becbc5479f ("af_iucv: avoid left over IUCV connections from failing connects")
Cc: [email protected]
Signed-off-by: Bryam Vargas <[email protected]>
---
 net/iucv/af_iucv.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/net/iucv/af_iucv.c b/net/iucv/af_iucv.c
index ea047bab65e7..6fb0041c4c98 100644
--- a/net/iucv/af_iucv.c
+++ b/net/iucv/af_iucv.c
@@ -778,8 +778,16 @@ static int iucv_sock_connect(struct socket *sock, struct sockaddr_unsized *addr,
 	if (sk->sk_state == IUCV_DISCONN || sk->sk_state == IUCV_CLOSED)
 		err = -ECONNREFUSED;
 
-	if (err && iucv->transport == AF_IUCV_TRANS_IUCV)
+	if (err && iucv->transport == AF_IUCV_TRANS_IUCV) {
 		iucv_sever_path(sk, 0);
+		/* A connack may have landed while the wait was unwinding; the
+		 * path is gone, so the socket must not still claim it.
+		 */
+		if (sk->sk_state == IUCV_CONNECTED) {
+			sk->sk_state = IUCV_DISCONN;
+			sk->sk_state_change(sk);
+		}
+	}
 
 done:
 	release_sock(sk);

---
base-commit: a59f57e2aa127c5354168d2ec4bac920df1be4f4
change-id: 20260815-b4-disp-e1faf4b7-feddd1575cc2

Best regards,
--  
Bryam Vargas <[email protected]>