mptcp: fix recent failed backports (20260919)

The following patches could not be applied without conflicts in this tree:

- 2b0f561f21b2 ("mptcp: avoid unneeded actions on subflow reset")
- 42064de57fb8 ("mptcp: close race between scheduler and state change")
- f3ef03357396 ("mptcp: fix bad accounting in __mptcp_subflow_push_pending()")

Conflicts (if any) have been resolved, and documented in each patch.

v2: Drop these two patches that have been already backported to this
version, but git didn't complain about that. Instead, resolve a small conflict in patch 1 ("mptcp: avoid unneeded actions on subflow reset"): decrement the '__unused' variable.

Paolo Abeni (3):
mptcp: avoid unneeded actions on subflow reset
mptcp: close race between scheduler and state change
mptcp: fix bad accounting in __mptcp_subflow_push_pending()

 net/mptcp/protocol.c | 13 ++++++++-----
 net/mptcp/protocol.h |  3 ++-
 net/mptcp/subflow.c  | 11 +++++++++++
 3 files changed, 21 insertions(+), 6 deletions(-)

2.55.0

From: Paolo Abeni <[email protected]>

commit 2b0f561f21b27c40c91ea4975268a06092bd7e9c upstream.

Once in a blue moon, the mptcp receive path can recursively call mptcp_data_ready() via state change under unlucky error conditions, and then try to hold the data lock again.

Break the recursion loop explicitly checking for the exceptional condition.

Add a new flag instead of using an existing one like 'closing', to exit early in subflow_state_change(), and explicitly flush the RX queue at reset time.

This avoids unneeded processing to check for available data -- calling get_mapping_status() and more on a dying subflow -- but also in error reporting and worker scheduling.

Note that we must consume the currently peeked skb before invoking mptcp_dss_corruption to avoid consuming it again after the eventual reset has freed it.

[ Note: conflict in protocol.c, because commit e0ca4057e0ec ("mptcp:
micro-optimize __mptcp_move_skb()") is not in this version, and is part of a consequent rx path refactor. The conflict is in the context, and is easy to resolve, "done = true" can be moved along without consequences.
Also a conflict in protocol.h, because __unused is at a different number. Decrement the one from this version and add the new flag above. ]

---
 net/mptcp/protocol.c |  6 +++---
 net/mptcp/protocol.h |  3 ++-
 net/mptcp/subflow.c  | 11 +++++++++++
 3 files changed, 16 insertions(+), 4 deletions(-)

index 2f34e92d9cce..ff63ba5ee464 100644
@@ -747,13 +747,13 @@ static bool __mptcp_move_skbs_from_subflow(struct mptcp_sock *msk,
 				mptcp_dss_corruption(msk, ssk);
 			}
 		} else {
+			sk_eat_skb(ssk, skb);
+			done = true;
+
 			if (unlikely(!fin)) {
 				DEBUG_NET_WARN_ON_ONCE(1);
 				mptcp_dss_corruption(msk, ssk);
 			}
-
-			sk_eat_skb(ssk, skb);
-			done = true;
 		}
 
 		WRITE_ONCE(tp->copied_seq, seq);

index 2ac28f6ddff5..5fdf2b55b3cc 100644
@@ -545,7 +545,8 @@ struct mptcp_subflow_context {
 		is_mptfo : 1,	    /* subflow is doing TFO */
 		close_event_done : 1,       /* has done the post-closed part */
 		mpc_drop : 1,	    /* the MPC option has been dropped in a rtx */
-		__unused : 8;
+		resetting : 1,	    /* subflow is resetting */
+		__unused : 7;
 	bool	data_avail;
 	bool	scheduled;
 	bool	pm_listener;	    /* a listener managed by the kernel PM? */

index 267e0424eea5..cfa029b4d7f3 100644
@@ -440,6 +440,10 @@ void mptcp_subflow_reset(struct sock *ssk)
 	/* must hold: tcp_done() could drop last reference on parent */
 	sock_hold(sk);
 
+	subflow->resetting = 1;
+
+	/* No need to delay the actual close for to-be discarded data. */
+	__skb_queue_purge(&ssk->sk_receive_queue);
 	mptcp_send_active_reset_reason(ssk);
 	tcp_done(ssk);
 	if (!test_and_set_bit(MPTCP_WORK_CLOSE_SUBFLOW, &mptcp_sk(sk)->flags))

@@ -1857,6 +1861,13 @@ static void subflow_state_change(struct sock *sk)
 
 	__subflow_state_change(sk);
 
+	/* Rx queue processing is unneeded, error reporting will take place at
+	 * __mptcp_close_ssk() time and subflow reset can't happen in case of
+	 * fallback: subflow_sched_work_if_closed() would be a no-op.
+	 */
+	if (subflow->resetting)
+		return;
+
 	/* as recvmsg() does not acquire the subflow socket for ssk selection
 	 * a fin packet carrying a DSS can be unnoticed if we don't trigger
 	 * the data available machinery here.

2.55.0

From: Paolo Abeni <[email protected]>

commit 42064de57fb83231fcc89663a94885f228a1ee53 upstream.

The mptcp scheduler may race with subflow sockets state change: data transmission on the selected socket may fail and a later release could try to use mss_now reset to 0 for a divide operation.

Address the issue by explicitly checking for the critical scenario.

---
 net/mptcp/protocol.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index ff63ba5ee464..1304f3cb6d14 100644
@@ -1567,7 +1567,9 @@ struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 
 static void mptcp_push_release(struct sock *ssk, struct mptcp_sendmsg_info *info)
 {
-	tcp_push(ssk, 0, info->mss_now, tcp_sk(ssk)->nonagle, info->size_goal);
+	if (info->mss_now)
+		tcp_push(ssk, 0, info->mss_now, tcp_sk(ssk)->nonagle,
+			 info->size_goal);
 	release_sock(ssk);
 }
 

2.55.0

From: Paolo Abeni <[email protected]>

commit f3ef03357396d4b147d8e76c75fb612c2f264ffc upstream.

If __subflow_push_pending() errors out we should avoid updating the copied byte counters, to avoid mismatch push call later on.

---
 net/mptcp/protocol.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

index 1304f3cb6d14..ee160c783ff5 100644
@@ -1757,7 +1757,8 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk, bool
 			ret = __subflow_push_pending(sk, ssk, &info);
 			if (ret <= 0)
 				keep_pushing = false;
-			copied += ret;
+			else
+				copied += ret;
 		}
 
 		mptcp_for_each_subflow(msk, subflow) {

2.55.0

Now queued up, thanks.

greg k-h