[PATCH 0/8] DRM Rust abstractions

[Asahi Lina]

On 4/9/25 4:17 AM, Danilo Krummrich wrote:
> On Wed, Apr 09, 2025 at 03:06:38AM +0900, Asahi Lina wrote:
>> On 4/9/25 2:04 AM, Danilo Krummrich wrote:
>>> On Wed, Apr 09, 2025 at 01:29:35AM +0900, Asahi Lina wrote:
>>>> On 3/26/25 8:54 AM, Danilo Krummrich wrote:
>>>
>>> You said "use it/submit it in any way".
>>
>> I thought keeping authorship is an implied part of kernel etiquette.
>> Usually when you submit code someone else wrote, you keep them as
>> primary author... I want you and others to use the code, that doesn't
>> mean I want you to put your name on it as if you wrote most of it.
> 
> The broader context of the private mail was about you stepping back on kernel
> development. You did so with a few more details (which I'm not going to
> disclose), that made it clear to me that you don't want to be bothered with
> kernel development any more.
> 
> In combination with you giving permission to "use it/submit it in any way", I
> thought it's better to just pick a safe path to not misrepresent you given all
> the changes I made.
> 
> I do still credit you on all corresponding patches though.
> 
>>>> I'm really tired of kernel politics and I don't want to spend more brain
>>>> cycles looking at all the other patches or having to argue (in fact I
>>>> usually don't look at patch emails at all recently), but I would
>>>> appreciate if you keep my authorship for files that I did largely author
>>>> myself. After everything I've been going through the past weeks (some of
>>>> the people on Cc know what that's about...) this feels like yet another
>>>> slap in the face.
>>>
>>> As mentioned, please diff the correct thing and then just tell me where you'd
>>> like to have primary authorship changed.
>>
>> I don't even know what tree this series is supposed to apply onto (tried
>> drm-misc next, torvalds/master, v6.15-rc1) so I'm just going to take
>> drm-misc/topic/rust-drm and assume that's what this series includes.
>>
>> $ diff -urN rust/kernel/drm/drv.rs ../uplinux/rust/kernel/drm/driver.rs
>> | grep '^+' | wc -l
>> 45
>>
>> So I'm diffing the correct thing now and the result is essentially
>> identical.
>>
>> Meanwhile, device.rs has many additions... but a big chunk of those is
>> code that was just moved from drv.rs (like drm_legacy_fields and the
>> code that uses it).
> 
> Except drm_legacy_fields! and VTABLE (which is just trival boilerplate code)
> device.rs changed fundamentally, i.e. I switched the device abstraction to use
> the subclassing pattern.
> 
> If you look further you will find that I really changed a lot of things.
> 
> I have *nothing* to hide, here's the overall diff for all the changes I made:
> 
> 	[1] https://pastebin.com/FT4tNn5d
> 
>>
>> Again, I don't have the spoons to make some deep analysis here, you
>> should know how much of the code you changed, added, or just moved
>> around. I'm not going to litigate this further. If you think splitting
>> up a commit into multiple commits and moving code around warrants taking
>> over primary authorship of a project I've been working on for years now,
>> so be it.
> 
> You just said you "don't have the spoons to make some deep analysis here" and
> right below you acuse me of just "moving code around".
> 
> Which means that you do so *without* evidence. And again, I have *nothing* to
> hide, see [1].
> 
> Besides that I also told you that I'm fine to change primary authership, if you
> tell me where you think it would be appropriate (even though I do think my
> changes do justify how things are currently).
> 
>> I'm just disappointed.
> 
> That's where you are maneuvering *yourself* into.
> 
> You could have easily just asked me to change things for patch #X, #Y and #Z.
> 
> Instead you outright started with accusing me of things. I also feel like you
> intentionally try to misrepresent what I am doing and what my intentions are.
> 
> I neither have the time, nor am I willing to deal with random drama like this.
> 
> If you want something changed, just go ahead and tell me what, *without* more
> drama and without more accusing me of things.
> 

Alright, then please remove my authorship entirely from this series,
including Co-developed-by and signoff lines. I hereby release my code as
CC-0, which means you don't need the signoffs, it's yours now. The same
applies to any future code submitted that I originally authored as part
of the Asahi kernel git tree. That way we don't need to argue about any
of this.

I thought asking for patches that I mostly authored to keep my Git
authorship would be an uncontroversial request (and not unreasonable to
ask you to figure out which those are, since you made the
changes/splits, and #3 clearly is one), but apparently even that gets
you flamed on Linux threads these days.

I regret having been part of this community.

~~ Lina