AMD's Raven Ridge Botchy Linux Support Appears Worse With Some Motherboards/BIOS

[bridgman]

What I meant was that I think a lot of AMD code itself was generated with some sort of C generator.

Well, we do have some pretty awesome C generators. We call them "developers"

[pal666]

Is this a pattern with AMD? Initial Open Source support is there, but incomplete and buggy. We have to wait some months for everything to bake. Vega graphics was like this.

lol, no. pattern with amd is "you can use 5 year old hardware .. you can use 4 year old hardware .. you can use 3 year old hardware .. you can use 2 year old hardware .. you can use 1 year old hardware .. you can use just released hardware but some issues will be sorted in coming weeks(you are here)"

If the above pattern is true it's a shame, but tolerable for me.

if your pattern is true then you live in alternative universe where amd always supported hardware on release day. it's a shame to see such clueless posts

I could even start a crowd funded bounty to get it fixed.

better spend your energy on something amd isn't doing right now

[bridgman]

Well ... I have the HP Envy x360 with Ryzen 2500u ... there is one driver that actually works. ONE. And there is no official AMD driver, only the HP one

I don't understand what you are trying to say here. HP doesn't write their own drivers - AMD writes the drivers and we jointly test them on the hardware prior to launch. That driver (usually with a handful of last-minute fixes specific to the platform) becomes the official driver and is distributed by the vendor.

I don't know of a single tier-1 OEM that wants to invest in testing generic drivers - they all want "their own driver" (typically one which supports all the hardware THEY are shipping this cycle) and generally they only want Windows support, at least for laptops and all-in-ones.

That is slowly starting to change, but emphasis on the slowly. Dell is now offering a "developers edition" of one of their laptops pre-loaded with Ubuntu LTS, but for now that is limited to a single Intel-based design.

I suspect one of the reasons OEMs are biased so strongly towards Windows is that the stable API allows them to test and ship a single driver at launch and have it stay usable for the life of the product, while supporting Linux would require constant re-testing of new drivers each time the evolving ABI required an update.

The obvious solution for this is generic hardware and generic drivers, but most OEMs believe that differentiation and vendor-specific features are essential for them to succeed in the market. I don't believe that customization is actually required any more - IMO people make buying decisions based on things like build & display quality and support, not vendor-specific customizations in the SBIOS.

BTW all of the above applies to laptops and AIO's, but does not apply to the same extent to mobos and processors sold into the channel.

I was a bit surprised by the problems that Michael has been seeing since mobos and their SBIOSes tend to follow reference designs fairly closely, and I *think* the same applies to VBIOS (will check).

[nomadewolf]

I am glad I didn't wait for raven ridge when I built my Ryzen system a couple of weeks ago. I considered to do so but I have been burned with my older Radeon 7870 which IMO has always been in eternal beta support level, plus I don't like the way AMD has went with their multiple drivers that never work 100% properly. So I got a Ryzen 1700 with an 1050 Ti. A bit more expensive combination with closed driver but you know what? It works perfectly and also I didn't have to upgrade half of my OS stack to get it working!

AMD people, I do appreciate your hard work and support for Linux BUT please consider to put the end user experience at the center of your software strategy and try to be competitive with what Nvidia offers in that respect. For me open source is a plus but is not the end goal.

I have a Ryzen 7 1700x and an RX 550. Everything runs fine without the need to install anything, not even drivers.
And i do have Steam and play games.
The distro i use is Mint 18.3

[pal666]

We would not have deal with this mess Linux had stable driver ABIs

since amd driver is open source, abi has zero relevance here. stop spreading bullshit

within the same major release (similar to FreeBSD's driver model).

lol, that is why freebsd has to use linux drivers?

I know Linus and Co. will instantly reject any proposals for stable kABIs,

so why do you continue with your nonsense then?

if you want solid day-one support for new consumer devices similar to what Windows users have been enjoying for years

there is no joy in inserting cdroms or downloading drivers from websites. linux users enjoy system with best hardware support among all operating systems. you install distro and your hardware just works

[pal666]

The best case scenario would probably be development staying in git but with regular backports to LTS. Unfortunately I seriously doubt AMD has a large enough Linux team to make that happen.

backports to lts is canonicals job, not amds. what next, backports to rhel and sles?

[pal666]

Most of our Linux testing/validation is done on reference platforms

do you advertise linux reference platforms so people could direct their purchases?

[duby229]

Well, we do have some pretty awesome C generators. We call them "developers"

Of course. I don't want to downplay AMD in any way. But realistically a lot of AMD's code looks too "mechanical", too abstracted to have been written by a human. (People tend to like elegance, and tend to strive for it. But very little of AMD's code is elegant. Much of it has multiple layers of abstraction that are totally unnecessary that no human would've written.)

[pal666]

I've also come to believe AMD uses code generators to write a lot of it's code

because you use generators to write a lot of your posts?

[duby229]

because you use generators to write a lot of your posts?

Look at AMD's code and then you tell me if you think a human would've done that....