It occurs to me that Bug 527910 - "KMS should not be the default for radeon DRM module" should be added to the most annoying bugs page for milestone 5 (It is destroying factory installations for a lot of Radeon users even as we speak.) I had to compile a kernel myself with CONFIG_DRM_RADEON_KMS disabled in order to get my Radeon RS200 working again.
(See: https://bugzilla.novell.com/show_bug.cgi?id=527910)
On Monday 10 August 2009 14:31:21 Mark Gray wrote:
It occurs to me that Bug 527910 - "KMS should not be the default for radeon DRM module" should be added to the most annoying bugs page for milestone 5 (It is
Dear Mark,
I will add this to the list including a more simpler workaround. Instead of re-compiling the kernel it would have been sufficient to add radeon.modeset=0 to the kernel parameters in grub. (As described in the bugreport).
However be aware that this setting should be removed as soon as a kernel without KMS is being used. At the moment Factory contains kernel 2.6.31-rc5 which is compiled correctly.
Regards
Raymond
On Mon, Aug 10, 2009 at 8:34 AM, Raymond Woonincktittiatcoke@gmail.com wrote:
On Monday 10 August 2009 14:31:21 Mark Gray wrote:
It occurs to me that Bug 527910 - "KMS should not be the default for radeon DRM module" should be added to the most annoying bugs page for milestone 5 (It is
Dear Mark,
I will add this to the list including a more simpler workaround. Instead of re-compiling the kernel it would have been sufficient to add radeon.modeset=0 to the kernel parameters in grub. (As described in the bugreport).
According to bug 528187 radeon.modeset=0 did not work (but that bug also involved a problem with hald so it might work after all).
See: https://bugzilla.novell.com/show_bug.cgi?id=528187
On Monday 10 August 2009 16:59:47 Mark Gray wrote:
According to bug 528187 radeon.modeset=0 did not work (but that bug also involved a problem with hald so it might work after all).
Dear Mark,
The radeon.modeset=0 will only solve the issue for a non functional DRI radeon driver. This means that Xorg will normally start, but desktop effects can not be enabled.
If Xorg already did not start before you set this cmdline option, then you face the hald bug, which as a different workaround.
Please also remember that this radeon.modeset=o option should be removed as soon as you update the kernel !! This option is only valid for the openSUSE kernel version 2.6.31-rc4 (Milestone 5 only).
Regards
Raymond
On Tue, Aug 11, 2009 at 3:34 AM, Raymond Woonincktittiatcoke@gmail.com wrote:
On Monday 10 August 2009 16:59:47 Mark Gray wrote:
According to bug 528187 radeon.modeset=0 did not work (but that bug also involved a problem with hald so it might work after all).
The radeon.modeset=0 will only solve the issue for a non functional DRI radeon driver. This means that Xorg will normally start, but desktop effects can not be enabled.
I was just trying to say that having seen bug 528187 report that radeon.modeset=0 was not a recognized kernel option, I did not even bother trying it. Recompiling the kernel with CONFIG_DRM_RADEON_KMS did fix the problem, and the latest kernel from factory (kernel-default-base-2.6.31-1.2) does have KMS disabled.
If Xorg already did not start before you set this cmdline option, then you face the hald bug, which as a different workaround.
Actually on my system the kernel hung long before it even tried to start X.
On Tue, Aug 11, 2009 at 5:48 AM, Mark Graymark.t.gray@gmail.com wrote:
On Tue, Aug 11, 2009 at 3:34 AM, Raymond Woonincktittiatcoke@gmail.com wrote:
If Xorg already did not start before you set this cmdline option, then you face the hald bug, which as a different workaround.
Actually on my system the kernel hung long before it even tried to start X.
Spoke too soon -- I just did a zypper dup and I now do in fact have the hald bug (using my self compiled kernel X did work perfectly before the "dup" -- very, very odd -- but I will have to wait to look into the problem because I have to take my father to the his Doctor today.)
On Monday 10 August 2009 14:31:21 Mark Gray wrote:
It occurs to me that Bug 527910 - "KMS should not be the default for radeon DRM module" should be added to the most annoying bugs page for milestone 5 (It is destroying factory installations for a lot of Radeon users even as we speak.) I had to compile a kernel myself with CONFIG_DRM_RADEON_KMS disabled in order to get my Radeon RS200 working again.
Go ahead and add it yourself, it's a wiki ;)
It's fixed already for Milestone 6, Andreas
-
Andreas Jaeger -
Mark Gray -
Raymond Wooninck