[SOLVED] X on current broken for ATI Mobile M7

tux_dude · 03-10-2010, 09:12 AM

Does anyone have current running with X on old ATI hardware? I have a laptop with ATI Mobile M7 and X completely fails to load. There are no errors in the Xorg log file. No failed modules or failure to load a module. The last line in the Xorg log says something to the effect of "VGA0 does not support DRI2".

When monitoring the processes while I start X, I can see xinit load, then X. After the line above is added to the log file, X dies. This is so bad that it takes my entire display with it so I cannot switch terminal. I cannot restart a tty session from a remote connection. The only way to get the display back is to reboot.

I tried the suggestions form this thread to no avail. Recompiling the kernel with "CONFIG_PREEMPT_NONE=y" also had no effect. To get a working system I had to roll back to X from Slack 13, which works fine with current.

Did i missed something with the upgrade? Anything else I should try?

System info

Code:

Slackware-current 32 (with X from Slackware 13)
Kernel: generic-smp-2.6.33-smp
01:00.0 VGA compatible controller: ATI Technologies Inc Radeon Mobility M7 LW [Radeon Mobility 7500]

adamk75 · 03-10-2010, 11:38 AM

Can you actually log in remotely? You simply say that you can not restart a tty session from a remote connection, so I'm guessing you can, in fact, log in remotely. If that's the case, you could try running X through gdb and see if you can get a backtrace of the crash. Also, are you using KMS or not?

Adam

tux_dude · 03-10-2010, 12:55 PM

Everything works expect for X and yes, I can log in remotely. Completely forgot about the trace I did. The file is attached. Not using KMS.

adamk75 · 03-10-2010, 01:06 PM

Unfortunately, that gibberish to me. Even if you were to get an actual backtrace from gdb, it'd probably be gibberish even to the developers unless X was build with debugging symbols. I think the only other thing I could really recommend would be to see if a newer version of xf86-video-ati fixed the problem. http://www.phoronix.com/scan.php?pag...item&px=ODAzMA

What about KMS, by the way?

Adam

tux_dude · 03-10-2010, 01:16 PM

Tried 6.12.5 from current. Still no go.

No to KMS. Should I be using KMS? What's needed to get it going?

adamk75 · 03-10-2010, 01:22 PM

I'm doubtful that KMS would solve this problem. But, if you wanted to try, you could boot with radeon.modeset=1 as a kernel argument and that should enable KMS. I *believe* that the rest of -current is KMS ready, at least according to this post:

http://www.linuxquestions.org/questi...ml#post3891167

Adam

tux_dude · 03-10-2010, 01:58 PM

I will give KMS a try once I upgrade X to current. How do I enable KMS permanently, ie, without passing kernel parameter at boot?

adamk75 · 03-10-2010, 02:10 PM

You can either edit /etc/lilo.conf to enable that option by default, or you would need to recompile your kernel and enable radeon KMS in the staging driver section.

Adam