tcp: call sk_wmem_schedule before sk_mem_charge in zerocopy path

commit 358ed62 upstream.

sk_wmem_schedule makes sure that sk_forward_alloc has enough
bytes for charging that is going to be done by sk_mem_charge.

In the transmit zerocopy path, there is sk_mem_charge but there was
no call to sk_wmem_schedule. This change adds that call.

Without this call to sk_wmem_schedule, sk_forward_alloc can go
negetive which is a bug because sk_forward_alloc is a per-socket
space that has been forward charged so this can't be negative.

Fixes: f214f91 ("tcp: enable MSG_ZEROCOPY")
Signed-off-by: Talal Ahmad <talalahmad@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Wei Wang <weiwan@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

net/ipv4/tcp.c

@@ -1361,6 +1361,9 @@ int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
 			}
 			pfrag->offset += copy;
 		} else {
+			if (!sk_wmem_schedule(sk, copy))
+				goto wait_for_space;
+
 			err = skb_zerocopy_iter_stream(sk, skb, msg, copy, uarg);
 			if (err == -EMSGSIZE || err == -EEXIST) {
 				tcp_mark_push(tp, skb);