[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240403071332.49c219e8@kernel.org>
Date: Wed, 3 Apr 2024 07:13:32 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Hechao Li <hli@...flix.com>
Cc: Eric Dumazet <edumazet@...gle.com>, "David S . Miller"
<davem@...emloft.net>, Paolo Abeni <pabeni@...hat.com>, Soheil Hassas
Yeganeh <soheil@...gle.com>, netdev@...r.kernel.org,
kernel-developers@...flix.com, Tycho Andersen <tycho@...ho.pizza>
Subject: Re: [PATCH net-next] tcp: update window_clamp together with
scaling_ratio
On Tue, 2 Apr 2024 14:54:06 -0700 Hechao Li wrote:
> After commit dfa2f0483360 ("tcp: get rid of sysctl_tcp_adv_win_scale"),
> we noticed an application-level timeout due to reduced throughput. This
> can be reproduced by the following minimal client and server program.
Hi Hechao, nice to e-meet you :)
I suspect Eric may say that SO_RCVBUF = 64k is not very reasonable.
But I'll leave the technical review to him.
What I noticed is that our cryptic CI system appears to point at this
change as breaking BPF tests:
https://netdev.bots.linux.dev/flakes.html?min-flip=0&pw-n=0&tn-needle=gh-bpf
We accumulate all outstanding patches and test together.
BPF broke at net-next-2024-04-03--00-00, and:
$ cidiff origin/net-next-2024-04-02--21-00 \
origin/net-next-2024-04-03--00-00
+tcp: update window_clamp together with scaling_ratio
+tools: ynl: ethtool.py: Output timestamping statistics from tsinfo-get operation
+netlink: specs: ethtool: add header-flags enumeration
+net/mlx5e: Implement ethtool hardware timestamping statistics
+net/mlx5e: Introduce timestamps statistic counter for Tx DMA layer
+net/mlx5e: Introduce lost_cqe statistic counter for PTP Tx port timestamping CQ
+ethtool: add interface to read Tx hardware timestamping statistics
The other patches are all driver stuff..
Here's the BPF CI output:
https://github.com/kernel-patches/bpf/actions/runs/8538300303
Powered by blists - more mailing lists