public inbox for netdev@vger.kernel.org 
 help / color / mirror / Atom feed
From: Daniel Borkmann <daniel@iogearbox•net>
To: Neal Cardwell <ncardwell@google•com>,
	Steve Ibanez <sibanez@stanford•edu>
Cc: Netdev <netdev@vger•kernel.org>, Florian Westphal <fw@strlen•de>,
	Mohammad Alizadeh <alizadeh@csail•mit.edu>,
	Lawrence Brakmo <brakmo@fb•com>
Subject: Re: Linux ECN Handling
Date: Mon, 06 Nov 2017 15:08:54 +0100	[thread overview]
Message-ID: <5A006CF6.1020608@iogearbox.net> (raw)
In-Reply-To: <CADVnQymmMeo=cemUPQCzHpNvFPKCaqY1+VPchmui+HWuCL7CuA@mail.gmail.com>

On 10/24/2017 03:11 AM, Neal Cardwell wrote:
> On Mon, Oct 23, 2017 at 6:15 PM, Steve Ibanez <sibanez@stanford•edu> wrote:
>> Hi All,
>>
>> I upgraded the kernel on all of our machines to Linux
>> 4.13.8-041308-lowlatency. However, I'm still observing the same
>> behavior where the source enters a timeout when the CWND=1MSS and it
>> receives ECN marks.
>>
>> Here are the measured flow rates:
>> <https://drive.google.com/file/d/0B-bt9QS-C3ONT0VXMUt6WHhKREE/view?usp=sharing>
>>
>> Here are snapshots of the packet traces at the sources when they both
>> enter a timeout at t=1.6sec:
>>
>> 10.0.0.1 timeout event:
>> <https://drive.google.com/file/d/0B-bt9QS-C3ONcl9WRnRPazg2ems/view?usp=sharing>
>>
>> 10.0.0.3 timeout event:
>> <https://drive.google.com/file/d/0B-bt9QS-C3ONeDlxRjNXa0VzWm8/view?usp=sharing>
>>
>> Both still essentially follow the same sequence of events that I
>> mentioned earlier:
>> (1) receives an ACK for byte XYZ with the ECN flag set
>> (2) stops sending for RTO_min=300ms
>> (3) sends a retransmission for byte XYZ
>>
>> The cwnd samples reported by tcp_probe still indicate that the sources
>> are reacting to the ECN marks more than once per window. Here are the
>> cwnd samples at the same timeout event mentioned above:
>> <https://drive.google.com/file/d/0B-bt9QS-C3ONdEZQdktpaW5JUm8/view?usp=sharing>
>>
>> Let me know if there is anything else you think I should try.
>
> Sounds like perhaps cwnd is being set to 0 somewhere in this DCTCP
> scenario. Would you be able to add printk statements in
> tcp_init_cwnd_reduction(), tcp_cwnd_reduction(), and
> tcp_end_cwnd_reduction(), printing the IP:port, tp->snd_cwnd, and
> tp->snd_ssthresh?
>
> Based on the output you may be able to figure out where cwnd is being
> set to zero. If not, could you please post the printk output and
> tcpdump traces (.pcap, headers-only is fine) from your tests?

Hi Steve, do you have any updates on your debugging?

> thanks,
> neal
>

  reply	other threads:[~2017-11-06 14:08 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CACJspmLFdy9i8K=TkzXHnofyFMNoJf9HkYE7On8uG+PREc2Dqw@mail.gmail.com>
2017-10-19 12:43 ` Linux ECN Handling Florian Westphal
2017-10-23 22:15   ` Steve Ibanez
2017-10-24  1:11     ` Neal Cardwell
2017-11-06 14:08       ` Daniel Borkmann [this message]
2017-11-06 23:31         ` Steve Ibanez
2017-11-20  7:31           ` Steve Ibanez
2017-11-20 15:05             ` Neal Cardwell
     [not found]             ` <CADVnQy=q4qBpe0Ymo8dtFTYU_0x0q_XKE+ZvazLQE-ULwu7pQA@mail.gmail.com>
2017-11-20 15:40               ` Eric Dumazet
2017-11-21  5:58               ` Steve Ibanez
2017-11-21 15:01                 ` Neal Cardwell
2017-11-21 15:51                   ` Yuchung Cheng
2017-11-21 16:20                     ` Neal Cardwell
2017-11-21 16:52                       ` Eric Dumazet
2017-11-22  3:02                         ` Steve Ibanez
2017-11-22  3:46                           ` Neal Cardwell
2017-11-27 18:49                             ` Steve Ibanez
2017-12-01 16:35                               ` Neal Cardwell
2017-12-05  5:22                                 ` Steve Ibanez
2017-12-05 15:23                                   ` Neal Cardwell
2017-12-05 19:36                                     ` Steve Ibanez
2017-12-05 20:04                                       ` Neal Cardwell
2017-12-19  5:16                                         ` Steve Ibanez
2017-12-19 15:28                                           ` Neal Cardwell
2017-12-19 22:00                                             ` Steve Ibanez
2017-12-20  0:08                                               ` Neal Cardwell
2017-12-20 19:20                                                 ` Steve Ibanez
2017-12-20 20:17                                                   ` Neal Cardwell
2018-01-02  7:43                                                     ` Steve Ibanez
2018-01-02 16:27                                                       ` Neal Cardwell
2018-01-02 23:57                                                         ` Steve Ibanez
2018-01-03 19:39                                                           ` Neal Cardwell
2018-01-03 22:21                                                             ` Steve Ibanez

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=5A006CF6.1020608@iogearbox.net \
    --to=daniel@iogearbox$(echo .)net \
    --cc=alizadeh@csail$(echo .)mit.edu \
    --cc=brakmo@fb$(echo .)com \
    --cc=fw@strlen$(echo .)de \
    --cc=ncardwell@google$(echo .)com \
    --cc=netdev@vger$(echo .)kernel.org \
    --cc=sibanez@stanford$(echo .)edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox