Warning to users of Catalyst 8.11

[RealNC]

On 64-bit Gentoo, with X.Org 7.4 (XServer 1.5.3) and kernels 2.6.26 and 2.6.27, running on Catalyst 8.11 (ati-drivers-8.552-r2) results in data corruption that sometimes manifests in segfaulting application. emerging not-so-small applications (mozilla-thunderbird) has gcc, as (assembler) and python segfault at times. That is, if you're lucky. If you're not lucky, broken machine code can be produced without you ever knowing, fscking up your whole system.

Everything went stable again after reverting back to xf86-video-radeonhd.

So I guess I can say "you have been warned" here. Do thorough tests to see if you're affected by this if you want to use Catalyst 8.11 since this is serious.

[octapult]

Maybe I was lucky. I haven't met any segfault since I installed ati-drivers-8.552-r2.

The only thing I'm complaining about ati-drivers-8.552-r2 is that I still have no luck getting crossfire enabled on my ATI 4870x2.

By the way. AIGLX has been enabled successfully after I installed libstdc++-v3 package
There used to be some error messages in my /var/log/Xorg.0.log:

(EE) AIGLX error: fglrx exports no extensions (/usr/lib64/dri/fglrx_dri.so:
undefined symbol: __driDriverExtensions)
(EE) AIGLX: reverting to software rendering

Installing libstdc++-v3 and re-emerging ati-drivers solved the problem.

[RealNC]

Stress test your system before making an assumption. Like a kernel or OpenOffice compile. I don't have crashes either otherwise.

[DeepDayze]

Any other reports of this happening to others, or is it just one user?

OP, better make sure you have all the latest libs and latest stable GCC installed for your system...maybe you might have some libs that are outdated.

[energyman]

8.11 is the most stable driver ever. Even better than all the nvidia drivers I enjoyed over the years.

But you have to give the 'nopat' option at the kernel command line.

[codestation]

I have the same config (Gentoo 64-bit, xorg-server-1.5.3, kernel-2.6.27 and ati-drivers-8.552-r2). I compiled a crossdev gcc, firefox, xulrunner, etc, and using the system under heavy load (virtualbox + eclipse) and have zero problems. Maybe you have a filesystem/hardware issue.

[eddt]

On 64-bit Gentoo, with X.Org 7.4 (XServer 1.5.3) and kernels 2.6.26 and 2.6.27, running on Catalyst 8.11 (ati-drivers-8.552-r2) results in data corruption that sometimes manifests in segfaulting application. emerging not-so-small applications (mozilla-thunderbird) has gcc, as (assembler) and python segfault at times. That is, if you're lucky. If you're not lucky, broken machine code can be produced without you ever knowing, fscking up your whole system.

Everything went stable again after reverting back to xf86-video-radeonhd.

So I guess I can say "you have been warned" here. Do thorough tests to see if you're affected by this if you want to use Catalyst 8.11 since this is serious.

You are not alone. I am also on gentoo amd64 with a 790gx based MB kernel 2.6.27-7. I reverted to 7.3/8.542 since my box was dieing. I would need to install a serial console to see if it was a panic or stall. It did not take much to cause the new x/driver combo to kill the system. Just moving around a kterm eventually would trigger a crash. Here 7.4/8.552/gcc 4.3.2 makes for a very unstable system.

I have had small issues with 8.542 where X stalls. I have always been able to recover by sshing from another box and killing X. Ssh is not working when 8.552 bugs out. On the positive side its 25% faster...

The thread "Locking issues with Catalyst 8.11" looks to describe something very similar to my report. Which makes three of us with the problem. Guess I will go buy a serial card tomorrow and try to get the opps/panic/stall recorded.

[Jade]

Some Gentoo people deliberately sabotage Gentoo. You are left to figure out their motives.

This Catalyst thing might not be deliberate, but it would be worth checking to see.

Some Gentoo folk have clearly sabotaged Reiser3:

I would have usually formated /dev/hdb3 with the Reiser3 filesystem, however, this usually extremely reliable filesystem, has been deliberately sabotaged. This sabotage causes massive corruption of files after a short time. By comparison, I have used the Reiser3 filesystem for nearly ten years now and have never had any corruption, at all, on any of my many Reiser3 boxes.

From: http://linux.50webs.org/gentoo/qemu-gentoo.htm

Some folk have also clearly sabotaged Reiser4. See; The Linux Kernel SABOTEURS.

http://linux.50webs.org/jews/saboteurs.htm
http://www.phoronix.com/forums/showthread.php?t=9509

[RealNC]

Yes, they're SD-6 double agents.

Seriously, I have this problem, and only with 8.11. I never claimed this to be a bug that affected others, only me, but because on my system it's a serious one, I felt obligated to post here so others can check if they're affected. Do a bit of stress testing, and if you're fine, then I'd say no reason for panic.

PS:
As to whether I'm using outdated libs, well, I'm on Gentoo so no. Nothing is outdated. But perhaps this might be a problem since AMD/ATI historically supported distros with outdated stuff.

[energyman]

em, if you are using stable some libs might be pretty old. Also just -u system and -u world does not cover everything.

And hey, I just got news from a nvidia using friend. The latest driver completly fucked up his fonts ....

maybe your hardware is just broken - but the new driver make the bugs appear more readily?