http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c2 --- Comment #2 from Robin Knapp 2010-04-12 11:03:19 CEST --- Yes that's what I meant. I need to test a real installation, I think I can test it this week to figure out how this works. But isn't the module loaded from the initrd to allow a KMS splash? This is not the case with Live CD, it seems to use the normal fb video mode. Maybe that causes problems when switching to KMS later. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c4 --- Comment #4 from Stefan Dirsch 2010-04-12 09:18:40 UTC --- I'm afraid having an initrd with the appropriate KMS modules depending on the graphics hardware is simply not possible. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c6 Stefan Dirsch changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED Component|Live Medium |X.Org AssignedTo|coolo@novell.com |sndirsch@novell.com Summary|Live CD does not load KMS |radeon [M24] DRI gets |drivers correctly |disabled when KMS has been | |disabled. QAContact|qa@suse.de |xorg-maintainer-bugs@forge. | |provo.novell.com --- Comment #6 from Stefan Dirsch 2010-04-12 15:33:11 UTC --- I don't see that issue when KMS is enabled. This issue is completely unrelated to the LiveCD. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c8 Stefan Dirsch changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|radeon [M24/M66] DRI gets |radeon [M24/M66] DRI gets |disabled when KMS has been |disabled for the first |disabled. |Xserver start when KMS has | |been enabled / Screen | |Distortion with KMS enabled --- Comment #8 from Stefan Dirsch 2010-04-12 15:55:58 UTC --- Hmm. It appears to be the other way round. DRI gets disabled with KMS enabled the first time the Xserver starts. From the second time on you don't get this message any more. Anyway the result of enabling KMS is a screen distortion for me. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c10 Stefan Dirsch changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #10 from Stefan Dirsch 2010-04-12 16:20:25 UTC --- xorg-x11-driver-video.changes: ------------------------------------------------------------------- Mon Apr 12 18:17:30 CEST 2010 - sndirsch@suse.de - /etc/X11/kms-blacklist.radeon * blacklisted M24(1002:3150) (bnc #595653) Jeff, your card was already blacklisted. ==> Closing as fixed. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c12 Robin Knapp changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #12 from Robin Knapp 2010-04-12 22:53:31 CEST --- Guess what... I'm reopening because I got it working. Seems to be a complex issue. What I've done to get it working: - updated to latest Factory kernel (but it's the same release, just a newer build) - zypper dup to X11:XOrg -> xorg-x11-driver-video-7.5-39.1.i586 - boot with "nomodeset 3" - rmmod radeon; modprobe --ignore-install radeon modeset=1 -> resolution switches - init 5 -> X Server starts up, KDE loads, Compositing works, s2ram/resume works So KMS definately works on my chipset - blacklisting is the wrong solution here IMHO. What does NOT work: a) start X-Server without radeon module and let radeon be loaded by the xorg driver automatically -> I've seen a patch somewhere to load the radeon module *before* querying KMS capability. So the bug is in the driver, not in my hardware b) load radeon module from initrd - it hangs 60 seconds while trying to load firmware radeon/R300_cp.bin - which is odd, because this file exists in my initrd under /lib/firmware/radeon. Double and triple checked this. After this timeout, boot continues, but X Server does not start up correctly. Also unloading the module does not work because it is in use. So if these issues are fixed, KMS should work fine on M24 (and maybe the other blacklisted chipset). ah.. found the patch which could solve a) but did not test it: http://osdir.com/ml/debian-bugs-dist/2010-02/msg04726.html http://permalink.gmane.org/gmane.comp.freedesktop.xorg.drivers.ati/12853 https://patchwork.kernel.org/patch/81379/ That's just a quick googling, I guess you know better where exactly to look for the proper fix. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c14 --- Comment #14 from Alex Deucher 2010-04-12 22:05:41 UTC --- Both drivers have the same issue. It's basically a race condition, depending how quickly the drm loads, it might work on some systems but not on others. Really with KMS, you need to have the drm loaded before X starts. The drm in non-KMS mode doesn't actually do anything until the ddx initialize it things via ioctls, so it loads immediately. For KMS, the drm initializes the hw and memory manager which takes a while longer, so it's not always done loading by the time the ddx queries it for kms support. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c15 --- Comment #15 from Robin Knapp 2010-04-13 11:22:35 CEST --- fyi, the delay in issue b) might be caused by /sys unmounted at the end of the initrd. If radeon module tries to load firmware after /sys is unmounted, the firmware cannot be loaded, so again a timing issue. I'll try to verify that theory this evening. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c17 --- Comment #17 from Stefan Dirsch 2010-04-13 15:32:18 UTC --- My fault. The new modprobe rule /etc/modprobe.d/99-kms.conf (added automatically by mkinitrd to initrd) tries to run /sbin/kms_loader.sh script, which doesn't exist on initd. Thus modprobing the DRM modules fails in initrd. Also /etc/X11/kms-blacklist.* files are missing on initrd. The commands lspci and cut as well. I'm going to remove this KMS blacklist framework. It's useless anyway, when the module is already loaded in initrd with KMS enabled. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c Stefan Dirsch changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tiwai@novell.com Summary|radeon [M24/M66] DRI |Broken KMS blacklist breaks |disabled/screen distortion |DRM module loading in |due to radeon(KMS) not been |initrd |loaded in initrd | -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c20 --- Comment #20 from Robin Knapp 2010-04-13 19:10:35 CEST --- The backlist framework wasn't so bad I think. Add the missing files and it should work. However, I know why the firmware cannot be loaded: The order of scripts in initrd is: .. 03-kms.sh .. 04-udev.sh .. If nomodeset is set, kms.sh sets "dont_load_modules" - otherwise kms modules will be loaded automatically. BUT: The firmware loader needs a running udev. If I modify the run-all script and insert a "bash" after the udev.sh script, I can manually load them. It's not possible without the udev daemon. I'm not sure if it's enough to just start kms.sh after udev.sh because udev might load the modules automatically (is this the case with this kind of modules?) Otherwise kms would have to pre split in two scripts - the first sets options and always calls dont_load_modules, the seconds (after udev) then explicitely loads the modules. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c22 --- Comment #22 from Stefan Dirsch 2010-04-13 17:49:40 UTC --- (In reply to comment #21)

...

However, I know why the firmware cannot be loaded:

Thanks a lot. I'm going to open a seperate bugreport for that one.

https://bugzilla.novell.com/show_bug.cgi?id=596268 -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c24 --- Comment #24 from Stefan Dirsch 2010-04-13 18:42:50 UTC --- (In reply to comment #23)

For the kms-blacklist in the build service: Theres still M24 and M66 blacklisted - should be removed as they work fine (well, at least the M24).

Right.

Now what's left? This was originally reported against the live CD - how to proceed there? Either set nomodeset as default on the live cd

This would be fatal on intel graphics. Result would be fbdev.

or load the correct modules somehow (the setup-kms.sh already has some hw-detection which could be re-used there...)

I still don't believe that's possible for a LiveCD. Can't you specify nomodeset manually on the kernel boot commandline of the LiveCD? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=595653 http://bugzilla.novell.com/show_bug.cgi?id=595653#c26 --- Comment #26 from Stefan Dirsch 2010-04-14 13:51:02 UTC --- (In reply to comment #25)

manually specifying this should work, but IMHO the live-cd should work fine out of the box without any kernel command line parameter.

Should only be required if the default (KMS) doesn't work.

On my system for example the radeon drm driver was in kms mode but the xorg assumed UMS - this created different problems when logging out and also sometimes when switching to a console (radeon module error messages on console, screen corruption, freeze, ...)

Hmm. This shouldn't happen. As Alex Deucher mentioned this is likely a race.

But that's a low prio imho, because in general it seemed to work somehow and I was able to install the system.

Ok. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.