netdev - Re: [PATCH net-next v8 6/9] devlink: Allow to change priv in devlink-rate from parent

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: <1390589a-77e4-fe3c-853e-9e2fcc5fd57d@intel.com>
Date:   Fri, 4 Nov 2022 15:38:07 +0100
From:   "Wilczynski, Michal" <michal.wilczynski@...el.com>
To:     Jiri Pirko <jiri@...nulli.us>
CC:     <netdev@...r.kernel.org>, <alexandr.lobakin@...el.com>,
        <jacob.e.keller@...el.com>, <jesse.brandeburg@...el.com>,
        <przemyslaw.kitszel@...el.com>, <anthony.l.nguyen@...el.com>,
        <kuba@...nel.org>, <ecree.xilinx@...il.com>
Subject: Re: [PATCH net-next v8 6/9] devlink: Allow to change priv in
 devlink-rate from parent_set callbacks



On 10/31/2022 1:22 PM, Jiri Pirko wrote:
> Fri, Oct 28, 2022 at 12:51:40PM CEST, michal.wilczynski@...el.com wrote:
> >From driver perspective it doesn't make any sense to make any changes to
>> the internal HQoS tree if the created node doesn't have a parent. So a
>> node created without any parent doesn't have to be initialized in the
>> driver. Allow for such scenario by allowing to modify priv in parent_set
>> callbacks.
>>
>> Change priv parameter to double pointer, to allow for setting priv during
>> the parent set phase.
> I fail to understand the reason for this patch, but anyway, it looks
> very hacky. The priv is something the leaf/node is created with.
> Changing it from the callback awfully smells like wrong design. Please
> don't do that.

I was trying to point-out that nodes without any parent, or children
doesn't actually exist in any hierarchy, so in driver internally we don't
really need objects representing them.
Anyway I removed this commit in v9, this involved pre-allocation of
ice_sched_node so it's not ideal for me either, but it solves the problem.

Thanks for reviewing.

BR,
Michał