Thread (36 messages) flat view 36 messages, 5 authors, 2022-01-18

Re: [RFC PATCH v2 net-next 4/4] tcp: change SYN ACK retransmit behaviour to account for rehash

From: Eric Dumazet <edumazet@google.com>
Date: 2021-12-02 17:24:24

On Thu, Dec 2, 2021 at 8:41 AM Akhmat Karakotov [off-list ref] wrote:
quoted hunk ↗ jump to hunk
Disabling rehash behavior did not affect SYN ACK retransmits because hash
was forcefully changed bypassing the sk_rethink_hash function. This patch
adds a condition which checks for rehash mode before resetting hash.

Signed-off-by: Akhmat Karakotov <redacted>
---
 net/ipv4/tcp_output.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 6d72f3ea48c4..7d54bbe00cde 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -4108,7 +4108,8 @@ int tcp_rtx_synack(const struct sock *sk, struct request_sock *req)
        struct flowi fl;
        int res;

-       tcp_rsk(req)->txhash = net_tx_rndhash();
+       if (sk->sk_txrehash == SOCK_TXREHASH_ENABLED)
Since you add a new sk->sk_txrehash, you probably want to add
READ_ONCE()/WRITE_ONCE() over the reads/writes,
because at this point the listener socket lock is not held.

      /* Paired with WRITE_ONCE() in sock_setsockopt() */
       if (READ_ONCE(sk->sk_txrehash) == SOCK_TXREHASH_ENABLED)

sock_setsockopt() would need a similar comment
    /* Paired with READ_ONCE() in tcp_rtx_synack() */
     WRITE_ONCE(sk->sk_txrehash, val);


+               tcp_rsk(req)->txhash = net_tx_rndhash();
        res = af_ops->send_synack(sk, NULL, &fl, req, NULL, TCP_SYNACK_NORMAL,
                                  NULL);
        if (!res) {
--
2.17.1
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help