netdev - Re: [PATCH v2] net: sch_generic: Use pfifo_fast as fallback scheduler for CAN hardware

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: <894a5426-b649-fd51-9aad-e66543374b9e@hartkopp.net>
Date:   Tue, 22 Oct 2019 20:21:08 +0200
From:   Oliver Hartkopp <socketcan@...tkopp.net>
To:     Eric Dumazet <eric.dumazet@...il.com>,
        Stephen Hemminger <stephen@...workplumber.org>,
        Marc Kleine-Budde <mkl@...gutronix.de>
Cc:     Vincent Prince <vincent.prince.fr@...il.com>, jiri@...nulli.us,
        jhs@...atatu.com, netdev@...r.kernel.org, dave.taht@...il.com,
        linux-can@...r.kernel.org, kernel@...gutronix.de,
        xiyou.wangcong@...il.com, davem@...emloft.net
Subject: Re: [PATCH v2] net: sch_generic: Use pfifo_fast as fallback scheduler
 for CAN hardware


On 22/10/2019 19.28, Eric Dumazet wrote
> On 10/22/19 9:42 AM, Stephen Hemminger wrote:
> 
>> Why not fix fq_codel to return the same errors as other qdisc?
>>
> 
> I believe the same problem would happen with any qdisc not doing tail drops.
> 
> Do we really want to enforce all qdisc to have a common drop strategy ?

CAN has no drop strategy. There is no transport protocol at this point 
which can handle and compensate drops. CAN is just about PDUs that are 
sent on a special medium.

And that's what this patch was about.

> 
> For example, FQ could implement a strategy dropping the oldest packet in the queue,
> which is absolutely not related to the enqueue order.
> 

We have a CAN ID related ematch rule in em_canid.c to be able to 
configure HTBs, delays or drops. But this is for testing and special 
use-cases. Silently dropped frames are not expected.

Regards,
Oliver