Re: [net-next,v9,09/15] quic: add congestion control
Simon Horman <[email protected]> Thu, 5 Feb 2026 11:55:32 +0000
| Newsgroups | dev.linux.lists.quic,dev.linux.lists.kernel-tls-handshake,org.kernel.vger.linux-cifs,org.kernel.vger.netdev |
|---|---|
| Message-ID | <[email protected]> |
This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.
For local reproduction steps see https://netdev-ai.bots.linux.dev/ai-local.html
---
quic: add congestion control
This patch introduces quic_cong for RTT measurement and congestion control,
implementing a state machine with slow start, congestion avoidance, and
recovery phases using the New Reno algorithm.
> diff --git a/net/quic/cong.c b/net/quic/cong.c
> --- /dev/null
> +++ b/net/quic/cong.c
[ ... ]
> +static int quic_cong_check_persistent_congestion(struct quic_cong *cong, u64 time)
> +{
> + u32 ssthresh;
> +
> + /* rfc9002#section-7.6.1:
> + * (smoothed_rtt + max(4*rttvar, kGranularity) + max_ack_delay) *
> + * kPersistentCongestionThreshold
> + */
> + ssthresh = cong->smoothed_rtt + max(4 * cong->rttvar, QUIC_KGRANULARITY);
> + ssthresh = (ssthresh + cong->max_ack_delay) * QUIC_KPERSISTENT_CONGESTION_THRESHOLD;
> + if (cong->time - time <= ssthresh)
> + return 0;
[ ... ]
> +/* rfc9002#section-5: Estimating the Round-Trip Time */
> +void quic_cong_rtt_update(struct quic_cong *cong, u64 time, u32 ack_delay)
> +{
> + u32 adjusted_rtt, rttvar_sample;
> +
> + /* Ignore RTT sample if ACK delay is suspiciously large. */
> + if (ack_delay > cong->max_ack_delay * 2)
> + return;
> +
> + /* rfc9002#section-5.1: latest_rtt = ack_time - send_time_of_largest_acked */
> + cong->latest_rtt = cong->time - time;
The field cong->time is described as "Cached current timestamp" in struct
quic_cong, but where is it written? The socket struct is zero-initialized,
so cong->time will always be 0.
When computing RTT with a positive packet send timestamp in the time
parameter, does this cause unsigned integer underflow? For example, if the
packet was sent at time 12345, computing 0 - 12345 wraps to a very large
value.
The same issue appears in quic_cong_check_persistent_congestion() where
cong->time - time is compared against ssthresh, and in quic_reno_on_packet_lost()
and quic_reno_on_process_ecn() where recovery_time is set to cong->time.