It Looks Like AMDGPU DC (DAL) Will Not Be Accepted In The Linux Kernel

[bridgman]

bridgman and airlied: I hope the above is a sufficiently accurate representation of your positions? Feel free to correct/expand on this post as warranted.

Yep, I think most of the excitement came from ambiguity between DC-the-code and DC-the-specific-interface (aka HAL). Dave said "no" to the interface (which we understand) while the article and nearly all of the subsequent posts suggest that Dave said no to the idea of sharing code at all (which *would* be as bad as the article suggests).

Regarding the interface itself, there is an equally valid counter-argument that migrating code and state from DC into driver or drm should be done carefully over time after up-streaming rather than trying to come up with the right solution immediately and then having to rework a bunch of drivers as we learn what should have been done...

... but even then there needs to be a plan for (a) what the end state will probably look like, (b) how we deal with that internally (impact of changes on other platforms) and (c) who does the work and when.

Harry's initial RFC was only one part of that (proposing focus on new HW only so we could make corresponding decisions in our branch planning now-ish); we still have a bunch of work to do, and IMO what we didn't do was communicate that context sufficiently clearly.

[pal666]

DC/DAL abstraction layer would make future evolution of DRM hard as any change that would require that abstraction layer to change could get pushback "oh, but this will break things on Windows for us".

how did you come up with such bullshit argument? any change would be made by someone who has time to do it. which is the whole point: amd has no time to do it themselves and have to use other teams work.

[pal666]

I think you are wrong!

i think you are idiot

[mattlach]

Why would this even need to be a part of the Kernel? Why not distribute it as module?

[pal666]

but the fact that agdf said (AIUI) that the hal had to stay because otherwise they just don't have the resources to do validation as well

you utterly failed to understand what he said. he said that hal is already written by other team and they don't have people to rewrite it. if some imaginary community contributor jumps in and rewrites it, everybody would be happy. so who is that opensource champion?

[pal666]

Why would this even need to be a part of the Kernel? Why not distribute it as module?

it is a module. it needs to be part of kernel git to ease maintenance and distribution to users

[pal666]

so your theory of bad FOSS code in reality: YOUR SHIT CODE IS NOT ACCEPTED TO THE KERNEL...

reality was more like "your functions are named differently from my functions"

[DonQ]

David megalomania Airlie...

To be honest I don't understand this harsh judgement (and I think the unfriendly wording is simply unnescessary) - all AMD want is to have this code *shipped* with the linux kernel, so that users do not have to download/compile/install 3rd party kernel drivers. I guess nobody seriously expects anybody else than AMD putting serious effort and maintenance into this codebase (development of the open-source radeonsi-driver is also almost exclusivly done by paid AMD OSS developers).

And for AMD's efforts, after I upgraded from nvidia 7600GT, I have bought a 3850, 5850, 7950, 380x, and 390 (I may be off on card numbers, I gave away the 390 because of bad drivers, though). I am also eagerly waiting to spend $600+ on a Vega card when it comes out. AMD, your efforts don't go unnoticed.

[pal666]

There are quality borders that are more important than the development budget of AMD

this thread is starting to look like a census of people who never heard of staging

[pal666]

Dave says "NO! NEVER!"

actually, he didn't say that. he asked "what would happen if someone rewrites half of your code" and the answer is "everyone would be happy"