[opensuse-kernel] showstopper bug w.r.t. i915 module
[2.6.33-rc6-11-pae] [being forced currently to use a TTY console only, it's not fun to file a bug report into bugzilla, so you see my report here] X no longer starts with this kernel on my Intel 945GM chip: i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16 i915 0000:00:02.0: setting latency timer to 64 [drm:drm_fill_in_dev] *ERROR* Cannot initialize the agpgart module. DRM: Fill_in_dev failed. i915 0000:00:02.0: PCI INT A disabled i915: probe of 0000:00:02.0 failed with error -22 Later on in the log file, I see agpgart-intel 0000:00:00.0: Intel 945GM Chipset agpgart-intel 0000:00:00.0: detected 7932K stolen memory agpgart-intel 0000:00:00.0: AGP aperture is 256M @ 0xd0000000 Looking into the internet, I see that a similar bug has been filed as #359729 for Ubuntu; Somehow the calling order has been mangled so that i915 is loaded before agpgart... IMHO this is a showstopper bug. Note that this also happens with RC5 on my box -- BTW, who the heck had the idea to call the 2.6.33-rc5 Milestone kernel module just `2.6.33', indicating that it is no longer a release candidate but a final kernel version? I consider this a notably bad idea. Werner -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
[2.6.33-rc6-11-pae]
X no longer starts with this kernel on my Intel 945GM chip:
[...]
Looking into the internet, I see that a similar bug has been filed as #359729 for Ubuntu; Somehow the calling order has been mangled so that i915 is loaded before agpgart...
Some digging into /lib/mkinitrd/scripts: Calling setup-kms.sh manually (and providing the right value for the `kernel_version' variable) properly returns kms_modules = "intel_agp i915" gfx_modules = "i915" on my box. However, `kms_modules' is not used in any other script, just `gfx_modules'. Since I know nothing about the internals of mkinitrd I can't decide whether this is correct or not. On the other hand, doing lsinitrd -c initrd shows =========> config/kms.sh <============ [ "$kms_modules" ] || kms_modules='i915' [ "$gfx_modules" ] || gfx_modules='i915' Similarly, /var/log/YaST2/mkinitrd.log shows [MODULES] 03-kms.sh: i915 Werner -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
Calling setup-kms.sh manually (and providing the right value for the `kernel_version' variable) properly returns
kms_modules = "intel_agp i915" gfx_modules = "i915"
on my box.
[...]
Similarly, /var/log/YaST2/mkinitrd.log shows
[MODULES] 03-kms.sh: i915
What confuses me is that a call of mkinitrd shows KMS drivers: i915 on the console while calling setup-kms.sh as described above gives KMS drivers: intel-agp i915 Any idea why there is such a difference? How comes that `intel-agp' isn't properly returned? Werner -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
What confuses me is that a call of mkinitrd shows
KMS drivers: i915
on the console while calling setup-kms.sh as described above gives
KMS drivers: intel-agp i915
Any idea why there is such a difference? How comes that `intel-agp' isn't properly returned?
It's getting very mysterious. I've applied the following patch: --- setup-kms.sh.orig 2010-01-27 15:19:31.000000000 +0100 +++ setup-kms.sh 2010-02-06 17:34:59.000000000 +0100 @@ -37,6 +37,7 @@ " for i in $pcilist do +echo -e "i: $i" >> ~/zzz line=${i//\*/X} eval $line device[$num]=${d#0000} Calling setup-kms.sh directly, I get something like this in `~/zzz': i: v=000010B9 d=* sv=* sd=* bc=06 sc=00 ii=00* i: v=00001022 d=0000700C sv=* sd=* bc=06 sc=00 ii=00* i: v=00001022 d=0000700E sv=* sd=* bc=06 sc=00 ii=00* i: v=00001022 d=00007006 sv=* sd=* bc=06 sc=00 ii=00* i: v=00001002 d=* sv=* sd=* bc=06 sc=00 ii=00* i: v=00001279 d=* sv=* sd=* bc=06 sc=00 ii=00* [...] Calling mkinitrd, I get an empty output! The loop containing the inserted echo line is never executed. Note that $pcilist holds the same value in both incantation sequences. Anyone who can shed some light on this? BTW, is the value `6' in line 153 of setup-kms.sh correct? I would expect value `8'. Werner -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
On Sat, Feb 6, 2010 at 4:19 AM, Werner LEMBERG
[2.6.33-rc6-11-pae]
[being forced currently to use a TTY console only, it's not fun to file a bug report into bugzilla, so you see my report here]
X no longer starts with this kernel on my Intel 945GM chip:
This was also reported in bug #577000. Leonardo -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kernel+help@opensuse.org
participants (2)
-
Leonardo Chiquitto
-
Werner LEMBERG