catalyst 9.5 seems to be out.

[Zetbo]

Sorry for noise. I downgraded to 2.6.28.10 and i don't get those errors anymoore. Now I have new problem Neverwinter Nights leaks memory like hell and I have to restart it after playing about 15min, but thats offtopic.

[lordmozilla]

I can confirm that . Bad news for Ubuntu 9.04 users - we can't go back with Catalyst 9.3 for some reason.

The reason is Xserver 1.6.

Personally i'd like to know whats so good about xserver 1.6... cause it pisses me off.

I know whats bad about Cat 9.4 and Xserver 1.6 though!

[monraaf]

The reason is Xserver 1.6.

Personally i'd like to know whats so good about xserver 1.6... cause it pisses me off.

I know whats bad about Cat 9.4 and Xserver 1.6 though!

There's nothing wrong with Xserver 1.6. Your problem is that you're using proprietary closed source drivers.

Let me refer you to a post from bridgman where he makes the case for open source drivers:

The open source drivers will normally give very quick support on new kernels and X servers since they are part of those source trees and are often used by the developers who are making changes to the underlying framework.

I couldn't agree more, now if only AMD ramped up their development effort to get a working open source r6xx/r7xx 3D driver....

[lordmozilla]

There's nothing wrong with Xserver 1.6. Your problem is that you're using proprietary closed source drivers.

I'm not saying there is, I'm just wondering what it actually brings to the end user!

Dont know about you but between Ubuntu 8.10 and 9.04 I dont see what Xserver 1.6 has brought to make 9.04 better. Except for breaking proprietary drivers.

And i'm all for an Opensource driver, except I want a driver with 3d and that can clock my card down now, not in a year, so if the proprietary one works now, i'll use that one for the time being.

[bridgman]

I couldn't agree more, now if only AMD ramped up their development effort to get a working open source r6xx/r7xx 3D driver....

I do feel compelled to point out that the deal was supposed to be "AMD provides documentation and support, community writes the drivers".

Right now most of the community devs are working on other cool things -- kernel modesetting, memory management, Gallium3D etc -- so we're doing more of the new hardware support to keep things moving, but there was never any plan for us to develop the open source drivers ourselves.

I'm fine with having us add the new hardware support ourselves -- that seems to be a natural fit for AMD developers and I think that's how we will handle new hardware in the future -- but I would like it to be clear that adding new hardware support ourselves is *more* than we originally committed, not *less*.

We released sample 3D code and initial EXA/Xv code using the 3D engine almost six months ago, along with register specs and header files. We added a programming guide more recently, but that was primarily intended for new devs trying to understand the hardware, not for the experienced devs already working on the open source 3d drivers.

[Wielkie G]

Xserver 1.6 doesn't break drivers in my case (Counter Strike 1.6). It doesn't work with Catalyst newer than 9.3, even with Xserver 1.4 (Xorg 7.3). This is IMHO a regression, which stops me with upgrading my Debian Sid system. Should I report it via Catalyst feedback?

[bridgman]

The Catalyst Feedback page is a survey we use to help set priorities for future development. Bug reports should go into the Bugzilla tracker.

If there isn't already an applicable bug report at http://ati.cchtml.com please open one, with particular attention to any OpenGL-related messages you may get from Wine or the game. If a similar bug already exists, please add your info to that ticket rather than opening a new one.

In general, we support up to the OpenGL API, not Wine itself.

[Wielkie G]

OK, but this regression is quite complex... This application only crashes, raising C++ Exception (I can see MSVC++ Exception window)... and nothing else. Wine is quiet, so I don't know exactly what causes this unhandled exception to be reised. With Catalyst 9.3 it also crashes in few moments, for example when clicking some keys at the same time (but not always). So IMHO without deep debuging of wine & app & catalyst it would be hard to find out what's wrong.

I'll test it tommorow. Maybe exception window has some valuable information inside. I don't know why I have always problems with downgrade from Catalyst 9.4/9.5 to 9.3... I think I messed up my Debian system, but now I don't want to do it again (it took ~1.5h at all today to downgrade...).

[monraaf]

I do feel compelled to point out that the deal was supposed to be "AMD provides documentation and support, community writes the drivers".

I must have misunderstood then, I was under the impression that AMD was going to write or at least fund the development of the open source driver... With you as an AMD representative promising "improving consumer support" for Linux users, I don't think it's right to tell customers who purchased AMD hardware to write the driver for it themselves.

Right now as r6xx/r7xx users we have the choice between a closed source driver that is in many aspects broken and a working open source driver with missing 3D functionality. How much longer must we as r6xx/r7xx users wait before we can actually do simple things like watch a tear-free video and have 3D desktop effects at the same time? I have a feeling that fglrx is never going to deliver, it just seems FUBAR to me. On the other hand the open source driver looks promising, Alex did a great job on the radeon driver it's just the 3D driver that's missing.

[Melcar]

From the very beginning AMD's plan was made clear; provide the necessary documentation for the community to have the proper tools to develop a driver. In fact, that's exactly what everyone was asking for (give us the info. and we will make a driver). If you understood something else, then you are simply listening to the wrong sources.