mptcp: really share subflow snd_wnd

As per RFC, mptcp subflows use a "shared" snd_wnd: the effective
window is the maximum among the current values received on all
subflows. Without such feature a data transfer using multiple
subflows could block.

Window sharing is currently implemented in the RX side:
__tcp_select_window uses the mptcp-level receive buffer to compute
the announced window.

That is not enough: the TCP stack will stick to the window size
received on the given subflow; we need to propagate the msk window
value on each subflow at xmit time.

Change the packet scheduler to ignore the subflow level window
and use instead the msk level one

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

1 files changed, 15 insertions, 9 deletions

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index 7a9e2545884f..0e7f556ea2c7 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -1141,19 +1141,20 @@ struct mptcp_sendmsg_info {
 	bool data_lock_held;
 };
 
-static int mptcp_check_allowed_size(struct mptcp_sock *msk, u64 data_seq,
-				    int avail_size)
+static int mptcp_check_allowed_size(const struct mptcp_sock *msk, struct sock *ssk,
+				    u64 data_seq, int avail_size)
 {
 	u64 window_end = mptcp_wnd_end(msk);
+	u64 mptcp_snd_wnd;
 
 	if (__mptcp_check_fallback(msk))
 		return avail_size;
 
-	if (!before64(data_seq + avail_size, window_end)) {
-		u64 allowed_size = window_end - data_seq;
+	mptcp_snd_wnd = window_end - data_seq;
+	avail_size = min_t(unsigned int, mptcp_snd_wnd, avail_size);
 
-		return min_t(unsigned int, allowed_size, avail_size);
-	}
+	if (unlikely(tcp_sk(ssk)->snd_wnd < mptcp_snd_wnd))
+		tcp_sk(ssk)->snd_wnd = min_t(u64, U32_MAX, mptcp_snd_wnd);
 
 	return avail_size;
 }
@@ -1305,7 +1306,7 @@ alloc_skb:
 	}
 
 	/* Zero window and all data acked? Probe. */
-	copy = mptcp_check_allowed_size(msk, data_seq, copy);
+	copy = mptcp_check_allowed_size(msk, ssk, data_seq, copy);
 	if (copy == 0) {
 		u64 snd_una = READ_ONCE(msk->snd_una);
 
@@ -1498,11 +1499,16 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 	 * to check that subflow has a non empty cwin.
 	 */
 	ssk = send_info[SSK_MODE_ACTIVE].ssk;
-	if (!ssk || !sk_stream_memory_free(ssk) || !tcp_sk(ssk)->snd_wnd)
+	if (!ssk || !sk_stream_memory_free(ssk))
 		return NULL;
 
-	burst = min_t(int, MPTCP_SEND_BURST_SIZE, tcp_sk(ssk)->snd_wnd);
+	burst = min_t(int, MPTCP_SEND_BURST_SIZE, mptcp_wnd_end(msk) - msk->snd_nxt);
 	wmem = READ_ONCE(ssk->sk_wmem_queued);
+	if (!burst) {
+		msk->last_snd = NULL;
+		return ssk;
+	}
+
 	subflow = mptcp_subflow_ctx(ssk);
 	subflow->avg_pacing_rate = div_u64((u64)subflow->avg_pacing_rate * wmem +
 					   READ_ONCE(ssk->sk_pacing_rate) * burst,