netdev - Re: [PATCH net-next] tcp: clarify tcp_congestion

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <04828827-e664-4fbc-b96f-ce149da7c410@cdn77.com>
Date: Fri, 19 Dec 2025 08:37:18 +0100
From: Daniel Sedlak <daniel.sedlak@...77.com>
To: Neal Cardwell <ncardwell@...gle.com>
Cc: Eric Dumazet <edumazet@...gle.com>, Kuniyuki Iwashima
 <kuniyu@...gle.com>, "David S. Miller" <davem@...emloft.net>,
 Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
 Simon Horman <horms@...nel.org>, netdev@...r.kernel.org
Subject: Re: [PATCH net-next] tcp: clarify tcp_congestion_ops functions
 comments

Hi Neal,

On 12/18/25 2:37 PM, Neal Cardwell wrote:

> 
> Perhaps something like the following.
> 
> diff --git a/include/net/tcp.h b/include/net/tcp.h
> index 10706a1753e96..d35908bc977db 100644
> --- a/include/net/tcp.h
> +++ b/include/net/tcp.h
> @@ -1326,12 +1326,28 @@ struct rate_sample {
>   struct tcp_congestion_ops {
>   /* fast path fields are put first to fill one cache line */
> 
> +       /* A congestion control (CC) must provide one of either:
> +        *
> +        * (a) a cong_avoid function, if the CC wants to use the core TCP
> +        *     stack's default functionality to implement a "classic"
> +        *     (Reno/CUBIC-style) response to packet loss, RFC3168 ECN,
> +        *     idle periods, pacing rate computations, etc.
> +        *
> +        * (b) a cong_control function, if the CC wants custom behavior and
> +        *      complete control of all congestion control behaviors
> +        */
> +       /* (a) "classic" response: calculate new cwnd.
> +        */
> +       void (*cong_avoid)(struct sock *sk, u32 ack, u32 acked);
> +       /* (b) "custom" response: call when packets are delivered to update
> +        * cwnd and pacing rate, after all the ca_state processing.
> +        */
> +       void (*cong_control)(struct sock *sk, u32 ack, int flag,
> +                            const struct rate_sample *rs);
> +
>          /* return slow start threshold (required) */
>          u32 (*ssthresh)(struct sock *sk);
> 
> -       /* do new cwnd calculation (required) */
> -       void (*cong_avoid)(struct sock *sk, u32 ack, u32 acked);
> -
>          /* call before changing ca_state (optional) */
>          void (*set_state)(struct sock *sk, u8 new_state);
> 
> @@ -1347,12 +1363,6 @@ struct tcp_congestion_ops {
>          /* pick target number of segments per TSO/GSO skb (optional): */
>          u32 (*tso_segs)(struct sock *sk, unsigned int mss_now);
> 
> -       /* call when packets are delivered to update cwnd and pacing rate,
> -        * after all the ca_state processing. (optional)
> -        */
> -       void (*cong_control)(struct sock *sk, u32 ack, int flag, const
> struct rate_sample *rs);
> -
> -
>          /* new value of cwnd after loss (required) */
>          u32  (*undo_cwnd)(struct sock *sk);
>          /* returns the multiplier used in tcp_sndbuf_expand (optional) */
> 
> How does that sound?
> 

Thank you for your response & suggestions. This sounds like really nice 
improvement, can I use it for v2 and add you as Co-developed-by (since 
you've done most of the heavy lifting)?

Thanks!
Daniel