netdev - Re: [PATCH net-next 0/3] net: ethtool: Track TX pause storm

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: <1e2d7bb1-517a-401a-9b8c-3201f44b5f43@nvidia.com>
Date: Mon, 26 Jan 2026 08:51:04 +0200
From: Gal Pressman <gal@...dia.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: Oleksij Rempel <o.rempel@...gutronix.de>,
 Mohsin Bashir <mohsin.bashr@...il.com>, netdev@...r.kernel.org,
 alexanderduyck@...com, alok.a.tiwari@...cle.com, andrew+netdev@...n.ch,
 andrew@...n.ch, chuck.lever@...cle.com, davem@...emloft.net,
 donald.hunter@...il.com, edumazet@...gle.com, horms@...nel.org,
 idosch@...dia.com, jacob.e.keller@...el.com, kernel-team@...a.com,
 kory.maincent@...tlin.com, lee@...ger.us, pabeni@...hat.com,
 vadim.fedorenko@...ux.dev, kernel@...gutronix.de
Subject: Re: [PATCH net-next 0/3] net: ethtool: Track TX pause storm

On 26/01/2026 0:30, Jakub Kicinski wrote:
> On Sun, 25 Jan 2026 11:59:07 +0200 Gal Pressman wrote:
>> On 23/01/2026 20:40, Jakub Kicinski wrote:
>>>> - The auto-recovery (service task) enforces a fixed policy. Can we make
>>>>   this configurable? I used devlink health (.recover) to let userspace
>>>>   decide between auto-reset or manual intervention.  
>>>
>>> There is already a tunable for this exact feature but for PFC:
>>> ETHTOOL_PFC_PREVENTION_TOUT. Should be trivial to add the same thing for
>>> non-PFC pause. But we didn't want to open the uAPI can of warms unless
>>> there's a clear ask and consensus. We don't need tuning (or so we
>>> think), and there was some talk about not adding uAPI for fbnic because
>>> it's a "private device".  
>>
>> I had to refresh my memory on this, but I think we've chosen a non-ideal
>> name back then.
>> We use this value for both PFC and global pause, I recommend you do the
>> same (perhaps with better documentation?).
> 
> Excellent, I was wondering if that may indeed be the case. Thanks for
> doing the digging. Let's respin, use that knob and document it better
> as you suggest.

FWIW, I see bnxt also supports this tunable.

> 
>> mlx5 exposes tx_pause_storm_warning_events/tx_pause_storm_error_events
>> through 'ethtool -S', we can probably assign one of them into
>> tx-pause-storm-events.
> 
> Do you prefer to take care of that or should Mohsin do it? According to
> https://enterprise-support.nvidia.com/s/article/understanding-mlx5-ethtool-counters
> sounds like the "error" counter is the one that matches.

Yes, tx_pause_storm_error_events should be used.
I don't have a preference on who adds it.