Recent powerpc kernels (most recently 2.6.14-1.1806_FC5) don't come with the radeonfb driver, and I think this causes problems when X tries to exit. Previous configurations using the "UseFBDev" option set to "true" seemed to work ok, but since upgrading to modular, the X server causes a machine crash on exit or a vt switch back to X. This might just be an X ati driver problem though, the device is an RV350 [Mobility Radeon 9600 M10] and I'm using the 'radeon' driver (I'll try the 'ati' wrapper driver next to see if that changes anything).
Nope, I see the same behavior even if I use the 'ati' wrapper driver. And this is with the latest X packages from yesterday's rawhide update. I'll try to mess around a little more tonight and see if it's a dri problem or something.
Upping the severity as this bug causes a crash and generally makes the system unusable.
Disabling DRI (commening out the Load "dri" part of my xorg.conf) makes things work ok again, so I guess the problem isn't fb related after all.
Oops, spoke too soon. It looks like X startup and exit is working ok, but VT switch is still broken. Switching from the X VT to another VT and back still causes a crash.
Updating summary.
Ah, FB_RADEON support is builtin to this kernel, I was wondering how the console worked. Unfortunately, setting "UseFBDev" to "true" still doesn't make the problem go away. Start X, switch back to a console VT, switch back to X and boom, the machine is tottally wedged. No panic, oops, or anything, just rainbow colors or a totally frozen display.
Please report this problem to X.org bugzilla at http://bugs.freedesktop.org in the "xorg" component. Be sure to attach your X server log file, and config to the X.Org bug report and provide as much detail as you can to ensure upstream PPC developers have as much information as possible in order to diagnose the problem. Once you've filed your bug report to X.Org, if you paste the new bug URL here, Red Hat will continue to track the issue in the centralized X.Org bug tracker, and will review any bug fixes that become available for consideration in future updates. Setting status to "NEEDINFO_REPORTER", awaiting X.Org bug URL for tracking.
I'm seeing something similar with the new xorg-x11-drv-ati-6.5.7.3-x rpms, I've tried both release -1 and release -4, unfortunatly I do not have an older version to try. This is a fully up2date rawhide with a radeon 9250 (rv250) on x86_64, this used to work fine with the older Modular driver and server version. Whenever I try to switch back from console to X X-freezes, the rest of the machine still works, I haven't tried remotely killing X, but it did complete a yum update while X was hanging. Killing X from the console and then restarting it also works fine, only when it tries to restore its previous state it hangs. Let me know if you want me to take this upstream, or if you want this in a seperate bugreport.
Notice that the reporter of bug 178897 is experiencing the same problem with a similar config, so now he can no longer do the required tests to advance bug 178897. This all in all makes for 3 seperate people in bugzilla, all with radeon 92xx on either powerpc or x86_64 with the following problem: startx, switch to VT, switch back to X -> hang. Since the radeon 92xx is supposed to be the best card when one wants free (as in speech) drivers, this is an awefull shame. I would like to nominate this bug as an FC5 showstopper bug. Even though I know this will happen I still accidently trigger this bug atleast once a day, which is very anoying. I've been doing some experimenting with packages from my yum cache and this does not seem to be a xorg-x11-drv-ati problem, but rather a xorg-x11-server-Xorg problem, downgrading to 1.0.0-3 or 0.99.3-9 fixes this, since 1.0.0-3 has broken openGL support for radeon 92xx on x86_64 I'm back to 0.99.3-9, which it seems was a rather good release (for me).
For anyone still having this problem, please review comment #7 above and update the bug report accordingly with an upstream bug URL. Once the URL is present, we will track the problem in X.Org bugzilla. Thanks in advance.
I'm happy to report that this is fixed on x86_64 with the latest xorg rawhide packages (and 3D still works).
Jesse: Can you confirm comment #11 on your system? TIA
I'll try to confirm tomorrow or Thu. when I have time (sorry I haven't filed the upstream bug, I've been tracking the problem via mail and irc w/benh trying to get it fixed).
Just upgraded to the latest today, and everything works again (though now I'm tracking a bug in freetype--kdm spins on ps_hints_apply for a long time when it starts up). Guess this one can be closed out! Jeses
Looks like a DUP of 187083. Jesse *** This bug has been marked as a duplicate of 187083 ***