Re: [PATCH net-next v3 01/14] virtchnl: create 'include/linux/intel' and move necessary header files

[Alexander Lobakin <aleksander.lobakin@intel.com>]

From: Jakub Kicinski <kuba@kernel.org>
Date: Fri, 22 May 2026 08:40:50 -0700

> On Fri, 22 May 2026 13:08:08 +0200 Alexander Lobakin wrote:
>>>> There are at least
>>>>
>>>> include/linux/mlx4, include/linux/mlx5 and include/linux/bnxt.
>>>>
>>>> Those are per-driver and not per-vendor, but intel ethernet has too many drivers 
>>>> to have separate folders for them.
>>>>
>>>> I just do not think this creates a precedent neccessarily.  
>>>
>>> You just said the other ones are for specific drivers.  
>>
>> Right, but according to your earlier suggestion they belong to
>> include/net, not include/linux.
>>
>> My understanding is that they're under include/linux, not include/net as
>> mlx5 is not only about Ethernet, but also RDMA etc. The same applies to
>> Intel's headers.
>>
>> What's your position after all this? Still include/net/intel? This
>> commit is about stopping scattering Intel headers all over include/linux
>> and set one place for them.
> 
> I strongly dislike the idea there are "intel" headers. Header files
> are not sorted by vendors. That gives off way too much "Intel's corner
> of the kernel" vibe. "net+Intel" is fine, but Intel by itself is too
> broad.
> 
> So IDK. include/net/intel is fine. So is the current layout. Or stick 
> to driver / module by module like other vendors.

We'll stick to the current include/linux/net/intel, a bit different
reasons than yours, but I agree with you in general.

> 
>>>> Folder structure is for you to decide as a maintainer, but it would be nice to 
>>>> have known about such doubts earlier.  
>>>
>>> I'd love to know if you any suggestions for improving the process.
>>> Otherwise please keep your venting off list.  
>>
>> I think Larysa just wanted to say that you disliked this commit after
>> the series went through several iterations on IWL and 3 iterations here,
>> nothing more. It's not about the overall process.
> 
> Intel has a strongly negative reviewer score right now.
> IMHO it's not appropriate for y'all to complain about upstream
> reviews, or how long it takes to get your patches merged...

We don't complain, apologies if this sounded like that.

Re review score -- I noticed that as well and have been focusing on
reviews more than sending patches, *but*:

To Aleksandr:

Please *stop* spamming IWL/netdev with patches. This doesn't help at all.
Or at least do minimum 1 review per each patch you send.

The imbalance you (not only, but mostly) create will only lead to that
we won't be able to send anything at all.

To you and sorta everyone from Intel:

Please focus more on reviews and wait with sending stuff until Intel is
back to the top 10 reviewers and has at least +/- neutral score.

Most of stuff coming from Intel right now is not critical for customers
and users, by doing that you paralyze really important series.

Thanks,
Olek