netdev - Re: [PATCH net-next v2] net_sched: restore "mpu xxx" handling

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] [day] [month] [year] [list]

Message-ID: <20220113110801.7c1a6347@kicinski-fedora-PC1C0HJN.hsd1.ca.comcast.net>
Date:   Thu, 13 Jan 2022 11:08:01 -0800
From:   Jakub Kicinski <kuba@...nel.org>
To:     Kevin Bracey <kevin@...cey.fi>
Cc:     Eric Dumazet <edumazet@...gle.com>,
        netdev <netdev@...r.kernel.org>, Jiri Pirko <jiri@...nulli.us>,
        Vimalkumar <j.vimal@...il.com>, maximmi@...dia.com
Subject: Re: [PATCH net-next v2] net_sched: restore "mpu xxx" handling

On Wed, 12 Jan 2022 19:31:54 +0200 Kevin Bracey wrote:
> On 12/01/2022 19:08, Eric Dumazet wrote:
> > On Wed, Jan 12, 2022 at 9:02 AM Kevin Bracey <kevin@...cey.fi> wrote:  
> >> commit 56b765b79e9a ("htb: improved accuracy at high rates") broke
> >> "overhead X", "linklayer atm" and "mpu X" attributes.  

Applied, thanks!

> > Thanks for the nice changelog.
> >
> > I do have a question related to HTB offload.
> >
> > Is this mpu attribute considered there ?  
> 
> I had not considered that. None of these 3 adjustment parameters are 
> passed to its setup - tc_htb_qopt_offload merely contains u64 rate and ceil.
> 
> Dealing with that looks like it might be a bit fiddly. htb_change_class 
> would need reordering. It appears the software case permits parameter 
> changing and it is ordered on that basis, but the offload locks in 
> parameters on creation.
> 
> But in principle, tc_htb_qopt_offload could contain a pair of 
> psched_ratecfg rather than a pair of u64s, and then offload would have 
> all the adjustment information.

CC Maxim, offload should at least be rejected if user request
unsupported params.