netdev - Re: [Cake] [PATCH net] sch_cake: diffserv8 CS1 should be bulk

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-Id: <C50A8C7B-4EAE-455B-B27D-B6FAB13B1EA7@gmx.de>
Date:   Thu, 27 Jan 2022 17:02:36 +0100
From:   Sebastian Moeller <moeller0@....de>
To:     Kevin 'ldir' Darbyshire-Bryant <ldir@...byshire-bryant.me.uk>
Cc:     Toke Høiland-Jørgensen <toke@...hat.com>,
        Cake List <cake@...ts.bufferbloat.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Matt Johnston <matt@...econstruct.com.au>,
        "David S . Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
        Dave Täht <dave.taht@...il.com>
Subject: Re: [Cake] [PATCH net] sch_cake: diffserv8 CS1 should be bulk



> On Jan 27, 2022, at 10:00, Kevin 'ldir' Darbyshire-Bryant via Cake <cake@...ts.bufferbloat.net> wrote:
> 
> 
> 
>> On 25 Jan 2022, at 10:58, Toke Høiland-Jørgensen <toke@...hat.com> wrote:
>> 
>> Matt Johnston <matt@...econstruct.com.au> writes:
>> 
>>> The CS1 priority (index 0x08) was changed from 0 to 1 when LE (index
>>> 0x01) was added. This looks unintentional, it doesn't match the
>>> docs and CS1 shouldn't be the same tin as AF1x
>> 
>> Hmm, Kevin, any comments?
>> 
>> -Toke
>> 
> 
> I’ll have to find my thinking head & time machine :-)
> This would be a lot easier if we had ‘diffserv9’, LE could have simply
> been added as the ‘if you’ve really nothing better to do’ class that it
> is.  And it’s why I’ve personally argued for a diffserv5: lowest;low;normal;high;highest
> moving on.
> 
> I think I screwed up when LE was added to diffserv8 - Matt the CS1 change from 0 to 1 IS intentional
> and IIRC I tried to bump everything else up 1 to compensate.. I may have missed some things though.

	But that way, introduction of LE does not fix much... I really think with LE's existence ,CS1 should be put into the same tin as CS0/BE, for the simple reason that currently CS1 is already in use both for priority below and priority above CS0/BE, the only course forward avoiding priority inversions is to treat CS1 like CS0.
	As a case in point I remember that Dave Täht reported seeing oodles of packets marked CS1 in his ingress some time in the past, packets that should not be treated as bulk. Sure we can argue that anybody using DSCPs for priority steering needs to re-map anyways, but that is not the logic behind cake's default mapping.
	And that CS1 to BE mapping should apply to all diffserv modes, that offer a lower priority tier, no?


Regards
	Sebastian



> 
> 
> Cheers,
> 
> Kevin D-B
> 
> gpg: 012C ACB2 28C6 C53E 9775  9123 B3A2 389B 9DE2 334A
> 
> _______________________________________________
> Cake mailing list
> Cake@...ts.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cake