netdev - Re: [PATCH net-next RFC v2 01/13] devlink: Add reload action option to devlink reload command

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: <cd0e3d7e-4746-d26d-dd0c-eb36c9c8a10f@nvidia.com>
Date:   Wed, 19 Aug 2020 15:18:22 +0300
From:   Moshe Shemesh <moshe@...dia.com>
To:     Jakub Kicinski <kuba@...nel.org>
CC:     Jiri Pirko <jiri@...nulli.us>, Moshe Shemesh <moshe@...lanox.com>,
        "David S. Miller" <davem@...emloft.net>,
        Jiri Pirko <jiri@...lanox.com>, <netdev@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH net-next RFC v2 01/13] devlink: Add reload action option
 to devlink reload command


On 8/19/2020 3:10 AM, Jakub Kicinski wrote:
>
> On Tue, 18 Aug 2020 12:10:36 +0300 Moshe Shemesh wrote:
>> On 8/17/2020 7:36 PM, Jiri Pirko wrote:
>>> Mon, Aug 17, 2020 at 11:37:40AM CEST, moshe@...lanox.com wrote:
>>>> Add devlink reload action to allow the user to request a specific reload
>>>> action. The action parameter is optional, if not specified then devlink
>>>> driver re-init action is used (backward compatible).
>>>> Note that when required to do firmware activation some drivers may need
>>>> to reload the driver. On the other hand some drivers may need to reset
>>> Sounds reasonable. I think it would be good to indicate that though. Not
>>> sure how...
>> Maybe counters on the actions done ? Actually such counters can be
>> useful on debug, knowing what reloads we had since driver was up.
> Wouldn't we need to know all types of reset of drivers may do?


Right, we can't tell all reset types driver may have, but we can tell 
which reload actions were done.

> I think documenting this clearly should be sufficient.
>
> A reset counter for the _requested_ reset type (fully maintained by
> core), however - that may be useful. The question "why did this NIC
> reset itself / why did the link just flap" comes up repeatedly.


I will add counters on which reload were done. reload_down()/up() can 
return which actions were actually done and devlink will show counters.