Revert "udp: do rmem bulk free even if the rx sk queue is empty"

stable inclusion
from linux-4.19.100
commit 4c1c35c01531cda945cd1e0ae994a6c91c88de81

--------------------------------

[ Upstream commit d39ca259 ]

This reverts commit 0d4a6608.

Willem reported that after commit 0d4a6608 ("udp: do rmem bulk
free even if the rx sk queue is empty") the memory allocated by
an almost idle system with many UDP sockets can grow a lot.

For stable kernel keep the solution as simple as possible and revert
the offending commit.
Reported-by: NWillem de Bruijn <willemdebruijn.kernel@gmail.com>
Diagnosed-by: NEric Dumazet <eric.dumazet@gmail.com>
Fixes: 0d4a6608 ("udp: do rmem bulk free even if the rx sk queue is empty")
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Acked-by: NWillem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NLi Aichun <liaichun@huawei.com>
Reviewed-by: Nguodeqing <geffrey.guo@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

net/ipv4/udp.c

@@ -1302,7 +1302,8 @@ static void udp_rmem_release(struct sock *sk, int size, int partial,
 	if (likely(partial)) {
 		up->forward_deficit += size;
 		size = up->forward_deficit;
-		if (size < (sk->sk_rcvbuf >> 2))
+		if (size < (sk->sk_rcvbuf >> 2) &&
+		    !skb_queue_empty(&up->reader_queue))
 			return;
 	} else {
 		size += up->forward_deficit;