bpf: Support TCP_RTO_MAX_MS for bpf_setsockopt

Some applications don't want to wait for too long because the
time of retransmission increases exponentially and can reach more
than 10 seconds, for example. Eric implements the core logic
on supporting rto max feature in the stack previously. Based on that,
we can support it for BPF use.

This patch reuses the same logic of TCP_RTO_MAX_MS in do_tcp_setsockopt()
and do_tcp_getsockopt(). BPF program can call bpf_{set/get}sockopt()
to set/get the maximum value of RTO.

Signed-off-by: Jason Xing <kerneljasonxing@gmail.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250219081333.56378-2-kerneljasonxing@gmail.com

diff --git a/net/core/filter.c b/net/core/filter.c
index 2ec162dd83c463640dcf3c151327206f519b217a..ffec7b4357f9f80e373f49a1333e459d853392cb 100644 (file)
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -5382,6 +5382,7 @@ static int sol_tcp_sockopt(struct sock *sk, int optname,
        case TCP_USER_TIMEOUT:
        case TCP_NOTSENT_LOWAT:
        case TCP_SAVE_SYN:
+       case TCP_RTO_MAX_MS:
                if (*optlen != sizeof(int))
                        return -EINVAL;
                break;