AMD Ports Open-Source Linux Driver To Windows Embedded

[droidhacker]

Has anybody considered the possibility that the reason it may be easier to use the OSS driver than catalyst is due to CPU ARCHITECTURE? Correct me if I'm wrong, but isn't catalyst x86-only?

[pingufunkybeat]

The fact is that nothing stops AMD to add support for UVD, and close the driver because of this.

Well, except the fact that they have to write a UVD driver from scratch. Because they can't just use the Catalyst code, the drivers are probably way too different, and it would likely be easier to write something new using the Catalyst code as a reference.

And if that is written, then it all depends on the internal review. If they get the green light, they can open source it and add it to the Linux drivers at any tie. So I don't see the Linux driver losing anything because of this.

[Drago]

The fact is that nothing stops AMD to add support for UVD, and close the driver because of this.
Maybe Bridgman, should say clearly what is AMD intention.

[pingufunkybeat]

I'm sure the GPL whiners are quite happy that the driver will be closed-source due to the permissible MIT license.

The good thing about this particular situation is that there is very little reason to believe that AMD has any interest in closing the driver and adding secret sauce to it.

Hopefully this means more manpower working on the driver, which could translate to useful code that can make its way into Mesa/Gallium drivers.

[bridgman]

BTW I stopped being the most prolific poster shortly after Michael raised the edit time limit enough that I didn't have to edit by deleting and re-posting

[scjet]

"...such as some retail systems, medical devices,..."
Ya sure, just what I want, Windows* on a Life-Support system... no thx.

[Setlec]

i hope that with this new project the linux driver will get better OGL support and principally better consumption and memory management!

[duby229]

Being an embedded device I wouldnt be at all surprised if it consisted of nothing more than 1 or 2 heads and a small compute core like maybe 40 spu's.

[AnonymousCoward]

Why do you say the driver will be closed source ?



Nope. This is a fairly recent project, conceived and started years after we got back into supporting open source graphics drivers, and the request was for a driver that could be released in source code form.

Going 100% the open source route without any proprietary sauce means your windows embedded driver won't support UVD, just like the open source Linux driver. I find it hard to believe that you will also leave your windows embedded customers with a crippled driver and device. Unless of course you are developing this driver for a specific device or set of devices that won't require UVD, doubtful though IMO.

[agd5f]

When it comes to efi boot for macs then it does not work

You can thank apple for that. That said, Matthew Garret has mostly sorted out the efi issues.