There's A Lot Of Exciting AMDGPU DRM Code Brewing For Eventual Catalyst Support

[darkbasic]

The posting had an interesting passage which *to me* seems to indicate that these patches add an additional, *internal* API that is being tunneled through the normal kernel API. Is this correct? If so, how does this align with the kernel's code rules?

It sounds like Linus is *not* going to pull amdgpu-DAL. Can someone from AMD please explain better what the hell is happening?

[Nille]

Well not exactly. Only one card from Rx 3yy Series is supported by amdgpu driver - that is Tonga R9 380.

There is a Kernel parameter for some renamed Cards (CIK).

[dungeon]

There is a Kernel parameter for some renamed Cards (CIK).

CIK part there is for developer purpose only - no one else should bother with that. There is no user support there.

And for sure distro kernels shoudn't ever built that, if they are they just give people false hope

[agd5f]

These are just testing branches with some of the stuff we are working on internally. Not all of it is necessarily ready for upstream in it's current form. When it is, we will get it upstream.

[agd5f]

What's the deal with posting on the forums? Every time I post lately, my message disappears and I get a message that says "Unapproved".

[agd5f]

If I try and write a coherent answer, I get "unapproved", if I complain about being "unapproved", it goes through right away. It's like there's a filter blocking all useful content.

[bridgman]

Michael mentioned that anti-spam code seemed to be randomly flagging posts into the moderation queue. Seems to be happening a lot more in the last week or so.

[W.Irrkopf]

These are just testing branches with some of the stuff we are working on internally. Not all of it is necessarily ready for upstream in it's current form. When it is, we will get it upstream.

Great! Actually this is something I'm very happy to see: So now (at least part of) the development with all the trial and error is happening in the open too.
Thanks!

[hwentland]

Interesting to see this pop up here. I'd just like to re-iterate what Alex already said. This is a testing branch that's not ready for upstreaming yet. We've gotten some feedback from the community and are working hard on getting something ready that can be accepted.

[hwentland]

A lot of this code is authored by Harry Wentland, an AMD developer not commonly seen attributed to many open-source AMD Linux patches but has been with the company in Toronto since 2011.

There's actually a whole team working on this. I just volunteered to be the public face, at least for the initial push. On http://cgit.freedesktop.org/~agd5f/l...g/?h=amd-15.31 you can see patches from others on the team.