AMD 8.40.4 Driver -- One Bug A Day Keeps AIGLX Away

[Michael]

With Slackware 12, I used the gui install and it worked fine, and this is the way most slackers are using to install the recent fglrx drivers.

Since Slackware 10 I always used the gui installer,

By GUI are you meaning the "automatic installation" or graphically generating the packages? For graphically generating the packages its SuSE and Red Hat.

[Svartalf]

The new driver isn't out yet till later this year.

If then... I'll believe it when I see it.

[chrisr]

No, it's a load of rubbish

World of Warcraft sometimes crashes now instead of hanging while trying to "Enter World", but it still doesn't work.

Still lots of "BUG: sleeping function called from invalid context" errors in the dmesg log.

A staggering 60 fps in glxgears.

I have generated an atisysteminfo-report, if anyone cares.

[opera]

Any kernel > 2.6.20-rc1 makes X segfault (stack trace says it's in an XF86 signal handler). I'm on debian unstable, dual head, 64-bit.
Now, if someone is in my seat and has had better success with this fglrx version [8.40] please let me know.

Starting from this release I won't bother to download it to try, since it's been shit in every version so far.
It's just a massive pile of crap these idiots at ATI are happy to share with us.

If someone knows how to get dual head to work in avivo, please advise me. I don't care at all about speed or acceleration or 3D etc etc, so avivo would be enough for me.

If it's at all possible to get KVM and coretemp fully working (easily) on a 64-bit linux with dual head ATI card (yes, that's really all I'm asking for), don't hesitate to tell me. Heck, I'll give you a dollar for just writing a solution here.

I'd just like to inform any ATI worker reading this: From the deepest part of me, from my heart and my soul, I sincerely despise you.
You suck.

There was, a f*ck, with no brain,
Who's code, was written, in vain!
Worked at ATI,
Should simply just die!?
Or see, he's freaking insane.

Without any driver for X,
I'd just like to stay home, watch Lex.
So sick of this crap,
Please fill in the gap!
And solve it, though it is complex!

I'm thinking of buying a Mac,
Not happy with seeing the stack.
I paid for this card!
Imbecile retard!
Too bad there is no going back...

I'm not even a native english speaker, but I can make up a few rimes in a minute or two. Compare that with being a huge company writing a freaking suck-ass driver... It's just nothing but pathetic. Go Glisse! Let's get it on!

[hobophobe]

My problem isn't fixed yet either, opera. I'm on debian sid 32bit.

Every new release (ever since xserver-xorg-core 1.3) I

• install the new fglrx
• upgrade xserver-xorg-core
• reboot
• get the same 'No Signal' on my display
• hit ctrl-alt-delete
• boot in single user mode
• downgrade xserver-xorg-core
• reboot
• wait another month

I'm assuming the cause of this isn't something that will be solved with the new driver. It also doesn't seem to be a priority for them. It's still not clear to me what the problem is (either why it happens or why it isn't fixed by now).

The last line from my Xorg.0.log, just tested:
"(WW) fglrx(0): More than one displays are connected,so clone mode is enabled" And yet, only one display is connected.

[lenrek]

When I tried Slackware 12, I found that you really should do

sh ./ati...run --buildpkg Slackware/All

That will not only create the driver packages, but also build the kernel module for you. You shouldn't use the official installer unless you use RedHat/Fedora (or OpenSuSE I guess).

Nope... Does not work. It behave the same as before. It just hang on start. Did a strace -p to the process ID of amdcccle, it returns only:

$> strace -p  3395
Process 3395 attached - interrupt to quit
select(10, [9], NULL, NULL, NULL

It seems to be waiting for something...

[yoshi314]

(I) always imagine a poor driver developer at AMD that likes his work, but that has found and read your posts :P Keep 'em coming!

actually i feel sorry for these guys. i'm pissed at the driver, not them.

the development cycle is probably too tight for them to actually do something big and functional within the driver right away. they have to chase the bugs, and their codebase gets forked every two weeks (according to the article)

i doubt i would enjoy my work as ati linux driver developer.i would feel bad/ashamed releasing such an crippled and incomplete driver on a monthly basis with little hope for serious improvement.

they probably have internal issues like too small team, dumb management, too frequent releases, too many branches to work on, etc.

[opera]

..., too frequent releases, too many branches to work on, etc.

Too frequent releases? Ever since I read ESR's view on the American military and its major oversteppings, I just hate the guy, but he's got a few things about software development very well. "Release early and often". That probably means, more often than each month, each time with 4-5 who-gives-a-crap fixes. It means, "for each change you make", basically.

I'm not sure the management would have a problem with the developers releasing parts of the driver as open source. Especially the parts which is about communication with the kernel, and with X. These parts, the community knows much more about, and ATI could continue to release closed internal code.

And two many branches? Have a look at the Nvidia driver. Perhaps they have something to learn from there.

[Xipeos]

Nope... Does not work. It behave the same as before. It just hang on start. Did a strace -p to the process ID of amdcccle, it returns only:

$> strace -p  3395
Process 3395 attached - interrupt to quit
select(10, [9], NULL, NULL, NULL

It seems to be waiting for something...

I have the same problem, remember?

[lenrek]

I have the same problem, remember?

Ooh... Shit... I forgot...

Hey, this happened after I upgraded from Slack11 to Slack12. Slack12 using Xorg 7.2 while Slack11 still use Xorg6.9. Maybe that is the reason, as I can't think of other reason for this to happen.