netdev - Re: [pull request][net-next 00/15] mlx5 updates 2023-12-20

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: <ZZyBzmqsFy6nTs2R@x130>
Date: Mon, 8 Jan 2024 15:14:22 -0800
From: Saeed Mahameed <saeed@...nel.org>
To: Jakub Kicinski <kuba@...nel.org>
Cc: "David S. Miller" <davem@...emloft.net>,
	Paolo Abeni <pabeni@...hat.com>, Eric Dumazet <edumazet@...gle.com>,
	Saeed Mahameed <saeedm@...dia.com>, netdev@...r.kernel.org,
	Tariq Toukan <tariqt@...dia.com>
Subject: Re: [pull request][net-next 00/15] mlx5 updates 2023-12-20

On 07 Jan 17:19, Jakub Kicinski wrote:
>On Thu, 4 Jan 2024 14:47:21 -0800 Jakub Kicinski wrote:
>> On Wed, 20 Dec 2023 16:57:06 -0800 Saeed Mahameed wrote:
>> > Support Socket-Direct multi-dev netdev
>>
>> There's no documentation for any of it?
>>
>> $ git grep -i 'socket.direct' -- Documentation/
>> $
>>
>> it's a feature many people have talked about for ever.
>> I'm pretty sure there are at least 2 vendors who have
>> HW support to do the same thing. Without docs everyone
>> will implement is slightly differently :(
>
>No replies so far, and v6.8 merge window has just begun,
>so let me drop this from -next for now.
>

But why revert ? what was wrong with the code or the current design? 
The current comments aren't that critical and I am sure you understand
that people are on holiday vacation.

We will provide the docs, but IMHO, docs could have been easily a follow
up.

What's the point of the upstream process if a surprise
revert can be done at any point by a maintainer? This is is not the first
instance, This has happened before with the management PF first iteration,
at least that time you asked for a revert and we approved, but this revert
came as a complete surprise.. 

Can we not do these reverts in such a stealthy way, this makes the whole
acceptance criteria unreliable, many teams rely on things getting accepted
so they plan the next steps, we have an upstream first open source policy
at nVidia networking and predictability is very important to us,
uncertainty especially when things are already accepted is something that
is very hard for us to work with.

Thanks,
Saeed.