81 Fresh Patches For AMDGPU's DC (DAL) Display Code

[L_A_G]

I find pointing out "bug fixes" to be funny, seeing as pretty much nobody has access to the hardware to experience the bugs in the first place (unless these bugs are why the code has been rejected). It's kind of like selling a used car to someone saying "the tires have been replaced twice" as though the owner will be affected by the first tire change.

The Vega FE, while really expensive, is out and if I'm not mistaken you can get DC/DAL in both AMDGPU-PRO and AMD's development kernels with it included (I think we even have a few people who dabble with those here in the Phoronix community).

[bridgman]

Which sounds justified — so they take the side that cares about the code, and they can use Linux as a main development platform because this is where contributors will be coming from.

Over time most of the contributions will be coming from our hardware teams. The whole idea is to let all of the fiddly HW-specific work be done once and used across multiple platforms. The challenge is that one platform (Linux) tries to share code across vendors where possible and so is opposed to vendor-specific HALs while all of the rest are vendor-specific and depend on HALs to insulate them from constant hardware changes.

[duby229]

Over time most of the contributions will be coming from our hardware teams. The whole idea is to let all of the fiddly HW-specific work be done once and used across multiple platforms. The challenge is that one platform (Linux) tries to share code across vendors where possible and so is opposed to vendor-specific HALs while all of the rest are vendor-specific and depend on HALs to insulate them from constant hardware changes.

I 'm wondering, there is a KMS interface in the kernel, which basically all modern GPU's use for modesetting. Why don't the linux kernel devs just develop a display interface which all GPU's can use for display interfaces? Or am I just misunderstanding badly?

[Hi-Angel]

Over time most of the contributions will be coming from our hardware teams.

Right, I was meaning to oppose e.g. that peoples from Windows will not contribute at all for the simple reason that there's no FOSS driver.

Although now that I think about it, I could imagine a situation when someone from Windows wanted to tweak something in the driver, then they get to know that the driver is shared across Linux and Windows, and so they get themselves involved into Linux development with the motivation to contribute to Windows driver. That's funny.

The whole idea is to let all of the fiddly HW-specific work be done once and used across multiple platforms. The challenge is that one platform (Linux) tries to share code across vendors where possible and so is opposed to vendor-specific HALs while all of the rest are vendor-specific and depend on HALs to insulate them from constant hardware changes.

[smitty3268]

There were several later discussions via email and IRC.

Ok, i might search around for them then.

If you are curious about that changes we are making, take a look at the patches or compare the display code in older and newer amd-staging branches.

I'm really not that interested in what's been done, though (i have skimmed some of those patch sets, but not anything in depth). I was more curious about what's left, and I don't necessarily have the knowledge to extrapolate everything myself without spending quite a bit of time on it, and i'm not sure i care that much.

Hence the wondering if bridgman would give a quick status update off the top of his head. I kind of expected the answer to be no, but there's no harm in asking.

[holunder]

Once Vega is released I hope for mainline support too! So far I think I will stick to Polaris until the DC situation is clarified.

You won’t get any Vega RX cards anyway until the DC situation is clarified…

[R41N3R]

You won’t get any Vega RX cards anyway until the DC situation is clarified…

How is this related? Vega will be released soon... this doesn't depend from the already available display code to be merged to mainline Linux.

[schmidtbag]

The Vega FE, while really expensive, is out and if I'm not mistaken you can get DC/DAL in both AMDGPU-PRO and AMD's development kernels with it included (I think we even have a few people who dabble with those here in the Phoronix community).

Yeah I forgot about Vega FE after I made my post. Though considering how unpopular it is in Windows, I'm sure you could count on two hands the amount of Linux users who own one (especially for those who pay attention to the driver situation).

Why would we leave out the bug fixes ? That would just mean a bunch of extra work to maintain a second copy of the tree containing arch changes but not bug fixes.

I'm not saying to leave them out - I'm aware of their significance. I just think it's funny for Michael to report on them, since the average reader here isn't affected by them. Hence the tire analogy.

[bridgman]

I 'm wondering, there is a KMS interface in the kernel, which basically all modern GPU's use for modesetting. Why don't the linux kernel devs just develop a display interface which all GPU's can use for display interfaces? Or am I just misunderstanding badly?

There is... but it's a moving target because it is constantly evolving. We just happened to be implementing new code against it while it was going through a very high rate of change, maybe the greatest since the introduction of KMS. It's that whole "stable API" discussion again... a rapidly evolving interface makes it easier to stay current with new hardware in principle but in practice raises the bar pretty high when trying to write & upstream new code against it.

[bridgman]

I'm not saying to leave them out - I'm aware of their significance. I just think it's funny for Michael to report on them, since the average reader here isn't affected by them. Hence the tire analogy.

Ahh, OK. Yeah, that occurred to me after I posted, but I had already closed the thread and didn't have time to go back and try to re-find it. Makes sense.