Re: [PATCH mptcp-next 1/3] mptcp: sched: penalise a slow subflow by halving its cwnd

Hi Matt,

On Wed, 2026-07-29 at 13:49 +0200, Matthieu Baerts wrote:
> Hi Shardul,
> 
> On 26/07/2026 07:55, Shardul Bankar wrote:
> > 
> 
> Could you reduce the size of the comments, please? I understood that
> LLM
> tends to leave long comments, but too long is not always good:
> explanations can go in the commit message, or be understood by
> reading
> the conditions (or that's a sign the code should be improved).
> Ideally,
> comments should not be needed when reading the code, or limited to
> one,
> with some exceptions for complex cases.
> 
> Here for example, I think you could limit the comment to one line:
> 
> /* Penalise subflows with pacing rate < this fraction of the fastest
> path */
> 

Yes, shortened them to one line each and moved the rationale into the
commit messages.

> > +
> > +/* Halve the congestion window (and ssthresh, if cwnd is past it)
> > of a subflow
> > + * the scheduler flagged. Runs in the push path under the subflow
> > socket lock,
> > + * which protects snd_cwnd. The congestion control grows the
> > window back,
> > + * ACK-clocked, and that regrowth is the built-in probe, so no
> > explicit probing
> > + * is needed.
> > + */
> 
> Same here: I think what is important to mention is that it is under
> subflow socket lock, and you can move the comment about the CC
> growing
> the window back below: no explicit probing required then.
> 

Ack'ed.

> > +static void mptcp_penalise_cwnd(struct sock *ssk)
> > +{
> > +       struct mptcp_subflow_context *subflow =
> > mptcp_subflow_ctx(ssk);
> > +       struct tcp_sock *tp = tcp_sk(ssk);
> > +       u32 cwnd = tcp_snd_cwnd(tp);
> > +
> > +       subflow->penalise = false;
> 
> You probably need to check inet_csk(ssk)->icsk_ca_state ==
> TCP_CA_Open
> again here, under the subflow socket lock, just in case it has been
> modified in between.

Added: mptcp_penalise_cwnd() now re-reads it under the subflow lock and
bails if it is no longer Open.

> > +       subflow->last_penalise = tcp_jiffies32;
> > +       tcp_snd_cwnd_set(tp, max_t(u32, cwnd >> 1, 2));
> > +       if (cwnd >= tp->snd_ssthresh)
> > +               tp->snd_ssthresh = max_t(u32, tp->snd_ssthresh >>
> > 1, 2);
> > +}
> > +
> >  /* implement the mptcp packet scheduler;
> >   * returns the subflow that will transmit the next DSS
> >   * additionally updates the rtx timeout
> > @@ -1565,9 +1603,9 @@ struct sock *mptcp_subflow_get_send(struct
> > mptcp_sock *msk)
> >         struct subflow_send_info send_info[SSK_MODE_MAX];
> >         struct mptcp_subflow_context *subflow;
> >         struct sock *sk = (struct sock *)msk;
> > -       u32 pace, burst, wmem;
> > +       u32 pace, burst, wmem, max_pace = 0;
> 
> Sashiko is (rightly I think) mentioning that the size of these
> variables
> are wrong. I guess a separate fix is needed to change "pace" to
> "unsigned long" is required. "max_pace" will get this type too then.
> 

I am splitting that into its own patch (1/4, Fixes: 3ce0852c86b9).

> >  
> > +       /* Flag the chosen subflow for cwnd halving (applied in the
> > push path)
> > +        * when its delivery rate is a small fraction of the
> > fastest path's and
> > +        * that fast path is saturated (cwnd-limited), so moving
> > load off the
> > +        * slow path is worthwhile. Only penalise a path in
> > TCP_CA_Open, one
> > +        * whose cwnd is not already being shrunk by loss recovery,
> > and at most
> > +        * once per RTT.
> > +        */
> 
> Same here. Maybe a comment is not needed, or limited?
> 

Ack'ed.

> > +       subflow = mptcp_subflow_ctx(ssk);
> > +       subflow->penalise = fastest && ssk != fastest &&
> > +                           (u64)subflow->avg_pacing_rate *
> > MPTCP_PENALISE_RATE_RATIO < max_pace &&
> 
> (or divide max_pace once before?)
> 

Changed to avg_pacing_rate < max_pace / MPTCP_PENALISE_RATE_RATIO.

> > +                           inet_csk(ssk)->icsk_ca_state ==
> > TCP_CA_Open &&
> > +                           tcp_is_cwnd_limited(fastest) &&
> > +                           mptcp_penalise_throttle_ok(subflow);
> > +
> >         burst = min(MPTCP_SEND_BURST_SIZE, mptcp_wnd_end(msk) -
> > msk->snd_nxt);
> >         wmem = READ_ONCE(ssk->sk_wmem_queued);
> >         if (!burst)
> >                 return ssk;
> >  
> > -       subflow = mptcp_subflow_ctx(ssk);
> >         subflow->avg_pacing_rate = div_u64((u64)subflow-
> > >avg_pacing_rate * wmem +
> >                                            READ_ONCE(ssk-
> > >sk_pacing_rate) * burst,
> 
> (here as well, it looks like there is an existing bug, and a cast to
> u64
> is probably needed for 32-bit system, as pointed by Sashiko)
> 

Added in the same fix as the issue reported above.

> >                                            burst + wmem);
> > @@ -1685,6 +1744,9 @@ static int __subflow_push_pending(struct sock
> > *sk, struct sock *ssk,
> >         struct mptcp_data_frag *dfrag;
> >         int len, copied = 0, err = 0;
> >  
> > +       if (mptcp_subflow_ctx(ssk)->penalise)
> > +               mptcp_penalise_cwnd(ssk);
> > +
> >         while ((dfrag = mptcp_send_head(sk))) {
> >                 info->sent = dfrag->already_sent;
> >                 info->limit = dfrag->data_len;
> > diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
> > index da40c6f3705f..2bf801292563 100644
> > --- a/net/mptcp/protocol.h
> > +++ b/net/mptcp/protocol.h
> > @@ -587,6 +587,7 @@ struct mptcp_subflow_context {
> >                 __unused : 8;
> >         bool    data_avail;
> >         bool    scheduled;
> > +       bool    penalise;           /* scheduler flagged this
> > subflow for cwnd halving */
> 
> (If you don't need to read this locklessly, then you can probably use
> one unused bit.)
> 

It is written under two locks (set in the scheduler under the msk lock,
cleared in the push path under the subflow lock), so a shared bitfield
word would race on the read-modify-write, and I kept it a bool for that
reason. If you see a single-lock way to do it, I am glad to move it to
a bit.

Thanks,
Shardul