[PATCH net-next] tcp: honor BPF_SOCK_OPS_RWND_INIT on the active connect path

[Tejas Birajdar]

BPF_SOCK_OPS_RWND_INIT lets a sockops BPF program pick the initial TCP
receive window, e.g. to advertise a larger window up front in environments
where that is known to be safe. Today it is only effective for the passive
(listener) side; on the active (connect) side the value is computed and
then silently discarded.

On the passive path tcp_openreq_init_rwin() inflates full_space when the
program returns a non-zero window, so tcp_select_initial_window() can offer
it:

	else if (full_space < rcv_wnd * mss)
		full_space = rcv_wnd * mss;

tcp_select_initial_window() only clamps the requested window *down* to the
available space, so without inflating the space first the BPF reply can
never raise the offered window above tcp_full_space(sk).

tcp_connect_init() calls tcp_rwnd_init_bpf() but never inflates full_space,
so on connect() the requested window is clamped back to tcp_full_space(sk)
(~64KB at the default rcvbuf) and the program's value is ignored.

Mirror the listener-side inflation in tcp_connect_init() so both directions
behave the same. tp->advmss is the mss the listener path uses as well (both
are tcp_mss_clamp(tp, dst_metric_advmss(dst))).

Fixes: 13d3b1ebe287 ("bpf: Support for setting initial receive window")
Signed-off-by: Tejas Birajdar <tejasbirajdar@xxxxxxxx>
---
Functional test: attached a cgroup sockops BPF program that sets
skops->reply for BPF_SOCK_OPS_RWND_INIT on the connecting socket, and
drove it with packetdrill on the active-open (connect) path (wscale 11,
advmss 1440). On the patched kernel the advertised receive window on the
first post-handshake ACK now reflects the requested value; unpatched it
stays clamped at ~64 KB:

  req_segs   unpatched   patched
  256        ~64 KB      360 KB   (256  * 1440)
  1024       ~64 KB      1.41 MB  (1024 * 1440)
  4096       ~64 KB      5.63 MB  (4096 * 1440)

(The SYN window itself is unscaled/capped; the offered window appears on
the first scaled post-handshake ACK.)

Regression: tools/testing/selftests/net/packetdrill run under virtme-ng
on this commit vs its parent produces an identical pass/fail set (no
newly failing tests).

 net/ipv4/tcp_output.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index d7c1444b5e30..478e4951140f 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -4101,6 +4101,7 @@ static void tcp_ca_dst_init(struct sock *sk, const struct dst_entry *dst)
 static void tcp_connect_init(struct sock *sk)
 {
 	const struct dst_entry *dst = __sk_dst_get(sk);
+	int full_space = tcp_full_space(sk);
 	struct tcp_sock *tp = tcp_sk(sk);
 	__u8 rcv_wscale;
 	u16 user_mss;
@@ -4133,14 +4134,16 @@ static void tcp_connect_init(struct sock *sk)
 
 	/* limit the window selection if the user enforce a smaller rx buffer */
 	if (sk->sk_userlocks & SOCK_RCVBUF_LOCK &&
-	    (tp->window_clamp > tcp_full_space(sk) || tp->window_clamp == 0))
-		WRITE_ONCE(tp->window_clamp, tcp_full_space(sk));
+	    (tp->window_clamp > full_space || tp->window_clamp == 0))
+		WRITE_ONCE(tp->window_clamp, full_space);
 
 	rcv_wnd = tcp_rwnd_init_bpf(sk);
 	if (rcv_wnd == 0)
 		rcv_wnd = dst_metric(dst, RTAX_INITRWND);
+	else if (full_space < rcv_wnd * tp->advmss)
+		full_space = rcv_wnd * tp->advmss;
 
-	tcp_select_initial_window(sk, tcp_full_space(sk),
+	tcp_select_initial_window(sk, full_space,
 				  tp->advmss - (tp->rx_opt.ts_recent_stamp ? tp->tcp_header_len - sizeof(struct tcphdr) : 0),
 				  &tp->rcv_wnd,
 				  &tp->window_clamp,
-- 
2.53.0-Meta