netdev - Re: [RFC net-next 03/15] devlink: split out netlink code

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]

Date:   Thu, 15 Dec 2022 11:35:57 -0800
From:   Jacob Keller <jacob.e.keller@...el.com>
To:     Jakub Kicinski <kuba@...nel.org>
CC:     <davem@...emloft.net>, <netdev@...r.kernel.org>,
        <edumazet@...gle.com>, <pabeni@...hat.com>, <jiri@...nulli.us>,
        <leon@...nel.org>
Subject: Re: [RFC net-next 03/15] devlink: split out netlink code

On 12/15/2022 11:14 AM, Jakub Kicinski wrote:
> On Thu, 15 Dec 2022 10:45:48 -0800 Jacob Keller wrote:
>> On 12/14/2022 6:01 PM, Jakub Kicinski wrote:
>>> Move out the netlink glue into a separate file.
>>> Leave the ops in the old file because we'd have to export a ton
>>> of functions. Going forward we should switch to split ops which
>>> will let us to put the new ops in the netlink.c file.
>>>   
>> Moving to split ops will also be a requirement for per-op policy right?
> 
> We can mix within one family, tho, IIRC.
> So new ops can have their own families and the old ones can stick to
> the family policy (unless someone takes the risk of converting them).

I would like to convert them at some point, even if we leave the old
commands as non-strict. Either way I think it would be preferable to
begin enforcing that new ops must be strict and must be per-op policy.

I personally think that moving from non-strict to strict should be ok
especially if we plan it over a few releases. A sane user space
application would generally prefer to be told "this is unacceptable"
rather than have the wrong operation done, or have its attributes
silently ignored... but I can understand the argument against breaking
user space that used to work.