Bug 577000 - X fails after last upgrade
Summary: X fails after last upgrade
Status: RESOLVED FIXED
: 577806 578512 (view as bug list)
Alias: None
Product: openSUSE 11.3
Classification: openSUSE
Component: X.Org (show other bugs)
Version: Factory
Hardware: i686 openSUSE 11.3
: P5 - None : Critical with 2 votes (vote)
Target Milestone: ---
Assignee: Michal Marek
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-02-04 14:31 UTC by Forgotten User K-uAvjqJ3n
Modified: 2010-03-09 00:57 UTC (History)
9 users (show)

See Also:
Found By: ---
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
xorg log for last update kernel is listed as 2.6.32 but the log showed the same for kernel 2.6.33 before I downgraded the kernel (5.26 KB, text/plain)
2010-02-04 14:56 UTC, Forgotten User K-uAvjqJ3n
Details
Fix. (928 bytes, patch)
2010-02-10 11:44 UTC, Egbert Eich
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Forgotten User K-uAvjqJ3n 2010-02-04 14:31:51 UTC
User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.0) Gecko/20100115 SUSE/3.6.0-1.1 Firefox/3.6

After upgrading to kernel 2.6.33 rc My xorg.0.log reports no usable configurations found. my chip set is Vendor:  Intel Corporation
  Model:  945 GME
  Driver:  intel

Reproducible: Always

Steps to Reproduce:
1.
2.
3.
Comment 1 Forgotten User K-uAvjqJ3n 2010-02-04 14:56:22 UTC
Created attachment 340728 [details]
xorg log for last update kernel is listed as 2.6.32 but the log  showed the same for kernel 2.6.33 before I downgraded the kernel
Comment 2 Clark Tompsett 2010-02-05 15:45:04 UTC
DRM is not being initilized -- from dmesg 


[    6.842899] [drm] Initialized drm 1.1.0 20060810
[    6.901610] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[    6.901622] i915 0000:00:02.0: setting latency timer to 64
[    6.901666] [drm:drm_fill_in_dev] *ERROR* Cannot initialize the agpgart module.
[    6.901742] DRM: Fill_in_dev failed.
[    6.901804] i915 0000:00:02.0: PCI INT A disabled
[    6.901816] i915: probe of 0000:00:02.0 failed with error -22


HP mini 110
Comment 3 Forgotten User K-uAvjqJ3n 2010-02-06 13:22:00 UTC
this workaround , from the forums, works ut must be done at every reboot
After updating 11.3 M1 to Factory today I was getting a problem with the X server not starting on my EEEpc 901 (Intel 945).

The error I got was:
(EE)No kernel modesetting driver detected

For which I found this solution:

su
rmmod i915
modprobe i915 modeset=1
*your screen should refresh*
exit
sax2 -r -m 0=intel
startx
Comment 4 Forgotten User il58cQohHy 2010-02-06 17:13:17 UTC
You can also pass i915.modeset=1 as a boot parameter. However, on my laptop (with an Intel 965 GM card), I don't have backlight control with KSM which greatly reduces the battery life.
Comment 5 Forgotten User K-uAvjqJ3n 2010-02-08 12:08:25 UTC
I tried passing the command i915.modeset=1 as a boot paramater
also tried sticking it in my xorg.conf file and using sysconfig but no go
this applies to kernel 2.6.33-rc6-2 as well
Comment 6 Leonardo Chiquitto 2010-02-08 15:30:50 UTC
*** Bug 577806 has been marked as a duplicate of this bug. ***
Comment 7 Leonardo Chiquitto 2010-02-08 16:12:02 UTC
Please don't set the "Priority" field. The developer working on the bug has exclusive control over this field (ie, the priority of this bug considering all the other bugs he/she is currently working on). The reporter can set the "Severity" field (ie, the negative impact caused by the bug).
Comment 8 Leonardo Chiquitto 2010-02-09 19:51:30 UTC
Adding Egbert and Michal as this seems to be a mkinitrd problem.
Comment 9 Werner Lemberg 2010-02-09 20:49:46 UTC
Yes, it's a mkinitrd problem; I've investigated a bit – citing from an email which I've sent to opensuse-kernel on Feb. 6th:

> 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'.
Comment 10 Leonardo Chiquitto 2010-02-10 11:32:00 UTC
*** Bug 578512 has been marked as a duplicate of this bug. ***
Comment 11 Egbert Eich 2010-02-10 11:41:37 UTC
A fix for this issue has been submitted for sle11 already.
I'm going to add a patch for setup-kms.sh here.
Comment 12 Egbert Eich 2010-02-10 11:44:21 UTC
Created attachment 341734 [details]
Fix.

This patch should fix the error - which can only be seen when 
'shopt -s nullglob' is set.
Comment 13 Egbert Eich 2010-02-10 11:45:52 UTC
Michal, would you please commit this patch to gitorious and update the package?
Thanks!
Comment 14 Matthias Hopf 2010-02-10 18:28:52 UTC
Egbert is this committed for SLE11 SP1 as well? It think of bug 577940 and the like.
Comment 15 Matthias Hopf 2010-02-10 18:31:06 UTC
Ah, I see, this is pushed in bug 566624.
Comment 16 Forgotten User K-uAvjqJ3n 2010-02-13 14:01:57 UTC
if i boot into runlevel 1 and rmmod rmmod i915 then init 5 i915 loads without doing a modprobe could i915 be removed from initrd and let X handle loading i915?
Comment 17 Clark Tompsett 2010-02-16 14:27:31 UTC
Synced to Monday evenings updates and my HP mini is now working and starting X correctly.
Comment 18 gérald Maricau 2010-02-16 15:10:37 UTC
xorg fail to start at boot time for the Intel Mobile Intel® GM45 Express Chipset

Bug 578512 is a duplicate of the same problem but apply to x86-64 architecture !
Comment 19 Forgotten User K-uAvjqJ3n 2010-02-16 15:16:58 UTC
My X still has the no modesetting driver error
however killing xdm(kdm) and reiniting kdm everything is ok
kill ####
kdm
Comment 20 Michal Marek 2010-02-16 15:51:37 UTC
Committed to git and submitted to Factory (created request id 32705).
Comment 21 gérald Maricau 2010-02-16 16:00:56 UTC
It doesn't work for me like Michal Marek, dale ritchey,

Because i've been updating like  Clark Tompsett, i don't know what has Clark Tompsett done to get it working directly without the temporary workaround that is still working :

with or without an /etc/X11/xorg.conf file) login in as root
and run this command :
rmod i915;init 3;sleep 4;init 5;sleep 15;exit

Please Clark Tompsett could you say if you applied the patch and the way you did it and what version of kernel ?

PS: my kernel is kernel-default-2.6.32-41.1.x86_64 or kernel-desktop-2.6.32-3.10.x86_64

What i did
locate setup-kms.sh
I took a backup of the file setup-kms.sh and /boot dir
# cd /
# tar cjvf boot-setup-kms.sh.tar.bz2 /boot /lib/mkinitrd/scripts/setup-kms.sh

# cd /lib/mkinitrd/scripts/
# patch < 0001-Fix-bogus-expansion-of-wildchar-characters-in-pci-de.patch
patching file setup-kms.sh                                                     
Hunk #1 FAILED at 29.                                                          
1 out of 1 hunk FAILED -- saving rejects to file setup-kms.sh.rej
# mkinitrd
...
cp: cannot stat `/usr/lib64/libdirect-1.2.so.0': No such file or directory
cp: cannot stat `/usr/lib64/libdirectfb-1.2.so.0': No such file or directory
cp: cannot stat `/usr/lib64/libfusion-1.2.so.0': No such file or directory
...
SOLUTIONS:
# ln -s /usr/lib64/libdirect-1.2.so.9 /usr/lib64/libdirect-1.2.so.0
# ln -s /usr/lib64/libdirectfb-1.2.so.9 /usr/lib64/libdirectfb-1.2.so.0
# ln -s /usr/lib64/libfusion-1.2.so.9 /usr/lib64/libfusion-1.2.so.0

CONTROL:
# ls -l /usr/lib64/libdirect-1.2.so.0 /usr/lib64/libdirectfb-1.2.so.0 /usr/lib64/libfusion-1.2.so.0
lrwxrwxrwx 1 root root 29 Feb 16 16:40 /usr/lib64/libdirect-1.2.so.0 -> /usr/lib64/libdirect-1.2.so.9
lrwxrwxrwx 1 root root 31 Feb 16 16:44 /usr/lib64/libdirectfb-1.2.so.0 -> /usr/lib64/libdirectfb-1.2.so.9
lrwxrwxrwx 1 root root 29 Feb 16 16:45 /usr/lib64/libfusion-1.2.so.0 -> /usr/lib64/libfusion-1.2.so.9

# mkinitrd
without errors, but when i reboot the machine i've got the same problem !
Comment 22 gérald Maricau 2010-02-16 16:10:12 UTC
I've been reporting this bug 578512 and it's a duplicate (sorry but it was difficult to find it) but it seems to be te same bug !

My version is "opensuse 11.3 Milestone 1" on "x86-64" architecture
with the kernel 2.6.32-3.10.x86_64

I think this bug should be reopened because the fix is not working for some people like me, this bug is not resolved !!!

If not, it should be explained how to solve the problem !

Thanks a lot !
Comment 23 Clark Tompsett 2010-02-16 16:13:32 UTC
All I did was do a zypper dup yesterday afternoon against the 1700 hr copy of the factory repo.  The machine is an HP mini 110-1025dx and is 32 bit only.
Comment 24 Michal Marek 2010-02-16 16:17:32 UTC
(In reply to comment #21)
> # cd /lib/mkinitrd/scripts/
> # patch < 0001-Fix-bogus-expansion-of-wildchar-characters-in-pci-de.patch
> patching file setup-kms.sh                                                     
> Hunk #1 FAILED at 29.                                                          
> 1 out of 1 hunk FAILED -- saving rejects to file setup-kms.sh.rej
> # mkinitrd

You haven't applied the patch, please rather update mkinitrd from http://download.opensuse.org/repositories/Base:/System/openSUSE_Factory/.

> ...
> cp: cannot stat `/usr/lib64/libdirect-1.2.so.0': No such file or directory
> cp: cannot stat `/usr/lib64/libdirectfb-1.2.so.0': No such file or directory
> cp: cannot stat `/usr/lib64/libfusion-1.2.so.0': No such file or directory
> ...
> SOLUTIONS:
> # ln -s /usr/lib64/libdirect-1.2.so.9 /usr/lib64/libdirect-1.2.so.0

That's not a solution, rather update the package that needs the old libraries (I guess splashy).
Comment 25 Forgotten User K-uAvjqJ3n 2010-02-17 13:14:28 UTC
works after a fresh install of milestone2
Comment 26 John Andrew McInnes 2010-02-22 02:00:41 UTC
11.3 milestone 2, I updated mkinitrd from factory. My intel 915 xorg still doesn't work. The modeset=1 as grub boot option doesnt work. the only thing that works is:

sudo rmmod i915
sudo modprobe i915 modeset=1
gdm
Comment 27 Stefan Dirsch 2010-02-22 06:26:12 UTC
See https://bugzilla.novell.com/show_bug.cgi?id=561577#c8
Comment 28 gérald Maricau 2010-02-22 10:48:43 UTC
Everything is ok now that i did a "zypper up" after adding the following repository :
"wget http://download.opensuse.org/repositories/Base:/System/openSUSE_Factory/"
"zypper ar Base\:System.repo" 
or if you don't use a proxy server
"zypper ar http://download.opensuse.org/repositories/Base:/System/openSUSE_Factory/"

My version of OpenSUSE is now:

# cat /etc/SuSE-release
openSUSE 11.3 Milestone 2 (x86_64)
VERSION = 11.3

Thanks a lot for your help and specially to Michal Marek !
Comment 29 Michal Marek 2010-02-22 10:50:18 UTC
I'm only committing Egbert's patches to the mkinitrd repo ;).
Comment 30 John Andrew McInnes 2010-03-09 00:57:41 UTC
11.3 Milestone 3 is working fine now on my intel 915 hardware. No probs.