netdev - Re: [PATCH net-next v2 0/3] add tx packets aggregation to ethtool and rmnet

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 for Android: free password hash cracker in your pocket

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

Message-ID: <CAA93jw53CEdZjcfOYXpN91n5n5WkjTr9HPo-jbGSJr-WUUea5w@mail.gmail.com>
Date:   Fri, 2 Dec 2022 09:27:24 -0800
From:   Dave Taht <dave.taht@...il.com>
To:     Daniele Palmas <dnlplm@...il.com>
Cc:     David Miller <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>,
        Eric Dumazet <edumazet@...gle.com>,
        Subash Abhinov Kasiviswanathan <quic_subashab@...cinc.com>,
        Sean Tranchetti <quic_stranche@...cinc.com>,
        Jonathan Corbet <corbet@....net>,
        Alexander Lobakin <alexandr.lobakin@...el.com>,
        Gal Pressman <gal@...dia.com>,
        Bjørn Mork <bjorn@...k.no>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        netdev@...r.kernel.org
Subject: Re: [PATCH net-next v2 0/3] add tx packets aggregation to ethtool and rmnet

On Fri, Dec 2, 2022 at 9:03 AM Daniele Palmas <dnlplm@...il.com> wrote:
>
> Il giorno gio 1 dic 2022 alle ore 16:22 Dave Taht
> <dave.taht@...il.com> ha scritto:
> >
> > A switch to using the "BBR" congestion controller might be a vast
> > improvement over what I figure is your default of cubic. On both
> > server and client....
> >
> > modprobe tcp_bbr
> > sysctl -w net.ipv4.tcp_congestion_control=bbr
> >
> > And rerun your tests.
> >
>
> Done, results available at
> https://drive.google.com/drive/folders/14CRs3ugyX_ANpUhMRJCH3ORy-jR3mbIS?usp=sharing

Mildly better on the up with bbr. till cracks 24 seconds on the rrul
up + down test. bbr on the server also you are using might actually
work out (which of course involves convincing the world to switch to
bbr for cell transfers) - could you try bbr on the server also?

I keep hoping that somewhere in the 4g or 5g usb stack there is some
way to tag packets and tell the OS that the given
packet had actually been sent out over the air. Getting the max
packets in the card at any rate good or bad, down below 20ms would be
marvelous. the 5G folk keep boasting about 1ms mac "latency" and never
seem to have noticed how bad it can get if you actually send real data
over a usb bus to it.

> > Any chance you could share these results with the maker of the test
> > tool you are primarily using?
>
> Mmmhh... Not easy, since I don't have a direct point of contact.

Mind if I rant about this somewhere?

> > My comments on your patchset are not a blocker to being accepted!
> >
>
> Thanks, I must admit that I don't have much knowledge on this topic,
> will try to educate myself.

thx! while I have a lot of entertaining videos about this kind of
stuff the starting point I recommend most is van jacobson's
ietf talk, with a fountain analogy:

video:https://archive.org/details/video1_20191129
slides: https://pollere.net/Pdfdocs/codel.ietf-tsv.jul12.pdf

The abstraction I keep hoping we find (somehow) - for lte/4g/5g - to
me would look a lot like BQL:

https://lwn.net/Articles/469652/
>
> Regards,
> Daniele



-- 
This song goes out to all the folk that thought Stadia would work:
https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-6981366665607352320-FXtz
Dave Täht CEO, TekLibre, LLC

Download attachment "bbr-rrul_-_bad_tx_aggr.png" of type "image/png" (103903 bytes)

Download attachment "bbr-tcp_nup_-_good_tx_aggr_vs-bad_.png" of type "image/png" (149012 bytes)