13 Years After Launch, The Open-Source Radeon Linux Driver Sees Occasional ATI R5xx Fix

[libv]

You fixed it in your mind or in an actual driver that is in use at the moment?

It was in the code that the forkers copied from the RadeonHD driver, code that they later stripped out after they partially rewrote it, removing copyrights in the process.

[starshipeleven]

It was in the code that the forkers copied from the RadeonHD driver, code that they later stripped out after they partially rewrote it, removing copyrights in the process.

If they stripped out code after they rewrote it, isn't it ok to remove the copyrights?
They might have added "based off original code from xxxx" but that's not a legal requirement.

[Dr. Righteous]

The old Radeon HD5750 was probably the best all around video card I ever owned. It worked great with linux at the time. It really seemed support for later cards got very spotty. For instance; had a HD6500 series card that wouldn't even boot to a GUI at all no matter what distro I tried. A HD 7870 had rotten 3D performance no matter what driver, or tweak I tried. The latest was a RX 480 which had excellent linux support. Loved that card, but decided to sell it when the bit coiners were going crazy. Got $100 more than I paid for it.
I'm probably going to get a RX 580 or 590 when I do a Ryzen upgrade later this year. My old AMD FX series machines work fine but they are in 19" rack in a roll around case that take up a lot of space. Had some audio gear in the rack as well when I work for a radio ministry connected to my church but those days are over. Going to do a smallish mini ATX system that will sit at the corner of my desk and be done with it.

[Guest]

Actually this is a good example... My T430 notebook in a repair got replaced by a T430 with an Nvidia optimus. Yes, it has a higher resolution display, but in every other respect the HD4000 has better support than the Nvidia NVS-5400M. While the opengl might be faster to some extent on the nvidia, it still is outdated and pinning that notebook in a medieval period, while the HD4000 gets TLC. There is a higher chance of a game working good on the HD4000 than the Nvidia working at all.
So yeah, the power of opensource is that perfectly good systems (the T430 should be faster than my steam machine CPU wise), can still be used, and that means less waste. This also means that cpu prices can rise again, because we don't change them that much anymore.
At least now I know that my next steam machine will not sport an Nvidia.

ThinkPads from that era are awful when it comes to dedicated graphics. The muxing is done in such a way that motherboards with a dGPU can drive the DisplayPort output only via the Nvidia chip. High resolution panels weren't tied to dGPU variants of these laptops, though. The funny thing about these setups is that my T420 can drive 3 displays at once with the Intel GPU after upgrading to Ivy Bridge (Intel HD 3000 can only do 2, Intel HD 4000 can do 3), so it can be done on that motherboard even though it was not designed for it.

In general, the fact that libre software handles legacy hardware well is a huge benefit. It allows you to keep using your hardware after it's deprecated, and there are cases where hardware is adequate after getting dropped by the manufacturer. (The way Nvidia handled deprecating 340 series drivers before EoL (Xorg update required too much work and they started delaying a new release so long I gave up for good on them) made sure I won't buy from them again, especially looking at how ridiculous their Turing lineup is.) I'm using my Nvidia GeForce 9600M GT on my old laptop for gaming, it's still adequate for even some modern games, even with nouveau, which doesn't perform as well as proprietary drivers did (even after setting it to highest clocks possible). I'd love to be able to build a desktop that I don't use for gaming, and use the same GPU for over a decade without issues, web browsers can be still accelerated with ancient GPUs.

[tuxd3v]

If they stripped out code after they rewrote it, isn't it ok to remove the copyrights?
They might have added "based off original code from xxxx" but that's not a legal requirement.

Well, if the code was partially rewritten or even totally rewritten( but Using previous Copywriter owner algorithms...), then it doesn't seem to me very legal..
Its a derived work..

[Guest]

Well, if the code was partially rewritten or even totally rewritten( but Using previous Copywriter owner algorithms...), then it doesn't seem to me very legal..
Its a derived work..

Then, rewrite it from scratch, which in the end would be the same. That code would be different derived from nothing, right?

[starshipeleven]

Well, if the code was partially rewritten or even totally rewritten( but Using previous Copywriter owner algorithms...), then it doesn't seem to me very legal..
Its a derived work..

Did they use previous owner algorithms? To what extent? He said that in this case with his code this PLL issue was already fixed, while with the rewritten code it clearly was not until now. He did say they used different approach than his too.

And also afaik most of that was rewritten to rely on AtomBIOS or whatever the card firmware was, instead of going "pure" and accessing hardware directly like libv wanted.

[marek]

I am confronted with the half-arsed code that these political/power players produced almost all the time. Such as the stupidity i am uncovering in KMS and atomic now that i am bringing up an all-in-one device for FOSDEM video capture. This was also not a one-time thing, but one big event on a long string of nasty games, and the technical deficiencies that are a result of this are still omni-present and do not get solved.

Which Game of Thrones character do you think you resemble the most?

[sundown]

This is awesome! Not because I suffer from the bug discussed in the article, but because I still think of my laptop which runs the mobility X1600 every now and then. In fact, I will be using it for some home lab experiments soon. But I have been thinking about it for another reason as well. Recently I bought an Intel NUC Hades PC and many people on the internet seem to refer to it as "oh It's called Hades cause Intel and AMD came together to build it and now hell is frozen". Funny thing is the laptop that the X1600 came with is the so called "VBI" or Verified by Intel. They sold those barebones that you could upgrade and finetune yourself. This was exactly back in 2006 when AMD bought ATi. On a side note, I remember being so disappointed when AMD announced they will not support the R500 chips anymore with their proprietary drivers. You couldn't even get decent video playback with those. Of course the opensource ones fixed a lot of things later on. Anyway, I will be installing MX Linux on it or something soon Here is a pic of the X1600 logo I took just right now:

https://ibb.co/rknbQYx

[Aisyk]

My X1250 Laptop (MSI U210) had enormous lags and problems in Win 7 with ATI drivers, hopes this could help Linux to have better 3D acceleration (with Elementary OS, there was some problems with 3D effects)