Intel driver 2.10 & KMS issue

combuster · 2010-02-13 19:12:19

Use intel video driver, all fb modules must be blacklisted...

swiftscythe · 2010-02-13 19:22:29

Yes, the screen goes black... I think you're right, I should report a bug...

combuster · 2010-02-13 19:28:38

http://www.gossamer-threads.com/lists/l … el/1172167

I found this from the lkml archive, but I couldnt find a bug report on the kernel bugzilla. You could report it upstream with the information I've extracted from your log and with the link above for reference.

Ofcourse your hardware (lspci) and uname -a should be mentioned in the report.

Last edited by combuster (2010-02-13 19:31:41)

patroclo7 · 2010-02-13 19:34:02

While the bug is submitted and solved, you could try to recompile the kernel with agp_intel and i915 compiled in (Y instead of M).

swiftscythe · 2010-02-13 19:40:42

I've already done this... but I could try again
What do you mean by Y instead of M?

combuster · 2010-02-13 19:49:07

That it should be compiled in the kernel image and not as module.

What is the output of

zcat /proc/config.gz | grep KMS

patroclo7 · 2010-02-13 19:50:13

When you configure the kernel, choose to compile those drivers into the kernel and not as modules. In the trace there seems to be some conflict when the modules are loaded. If the drivers are already in the kernel perhaps the conflict does not arise, who knows

As a bonus you could get back the arch boot logo

Last edited by patroclo7 (2010-02-13 19:51:21)

swiftscythe · 2010-02-13 19:56:45

Oh I see, thanks for the info
The output is:
CONFIG_DRM_KMS_HELPER=m
CONFIG_DRM_I915_KMS=y
#a comented line about RADEON, which is not my case

combuster · 2010-02-13 20:38:16

Hm, and how about zcat /proc/config.gz | grep DRM

Anyway it's an upstream bug and I don't think that you can do anything about it. You could use the older driver (to compile it yourself or to use the one from AUR) and turn KMS off, but I don't think that will help you forever, this should be reported and fixed as soon as possible... Or to try 2.6.33-rc from aur...

swiftscythe · 2010-02-13 20:45:38

The output:
CONFIG_DRM=m
CONFIG_DRM_KMS_HELPER=m
CONFIG_DRM_TTM=m
CONFIG_DRM_TDFX=m
CONFIG_DRM_R128=m
CONFIG_DRM_RADEON=m
CONFIG_DRM_I810=m
CONFIG_DRM_I830=m
CONFIG_DRM_I915=m
CONFIG_DRM_I915_KMS=y
CONFIG_DRM_MGA=m
CONFIG_DRM_SIS=m
CONFIG_DRM_VIA=m
CONFIG_DRM_SAVAGE=m
# CONFIG_DRM_RADEON_KMS is not set

Okay, I'll think about how to explain this and submit the bug,
Thank you very much for your help

combuster · 2010-02-13 20:59:45

Yep it's compiled as module, could made the difference to compile it in kernel image but it's a longshot.

technocratik · 2010-02-13 23:19:38

I had a similar problem on my Fujitsu Lifebook 4215 (Intel GMA 945G/GM). When booting with KMS enabled, X would start, the display would flicker and/or twitch, and after a few minutes/seconds would go completely blank (switching terminals would not fix it, either, although keyboard input was still clearly working). Disabling powersaving by adding the kernel option i915.powersave=0 seems to have resolved it.

swiftscythe · 2010-02-13 23:26:39

I've just tried with that option but it hasn't worked either...

Thank you very much!!

naphtali · 2010-02-14 01:46:05

I do not know if this helps but I was having the same black screen on boot whenever I turn KMS on via the grub kernel line. But when I installed the vesa & fbdev drivers, KMS on worked without any problem. No kernel recompiled.

d7wang · 2010-02-14 10:37:47

I have a laptop with similar configuration. I just compile the kernel myself with KMS enabled. Most of the things works well except the screen flashes every a few seconds. Now I have to downgrade the driver to 2.9.1.

JoyFM · 2010-02-20 13:26:57

same problem here with my old thinkpad X30 (intel 830M)

I had that problem since several Kernelversions and disabled KMS because of that.
If I pass the video=LVDS-8:1024x768-24@60 it doesn't change anything (also if I use 800x600 etc.).

what drm says:
[~]$ dmesg |grep drm
[drm] Initialized drm 1.1.0 20060810
[<c11376db>] ? sysfs_addrm_start+0x5b/0xb0
[drm] set up 15M of stolen space
[drm] DAC-6: set mode 640x480 0
[drm] LVDS-8: set mode 1024x768 b
fb0: inteldrmfb frame buffer device
[drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
(via ssh)

JoyFM · 2010-02-20 13:49:43

technocratik wrote:

I had a similar problem on my Fujitsu Lifebook 4215 (Intel GMA 945G/GM). When booting with KMS enabled, X would start, the display would flicker and/or twitch, and after a few minutes/seconds would go completely blank (switching terminals would not fix it, either, although keyboard input was still clearly working). Disabling powersaving by adding the kernel option i915.powersave=0 seems to have resolved it.

sadly - it doesn't work here either...

btw. the vesa xorg-driver is painting nice Acid-Alike-Colours and is not refreshing at all

sctincman · 2010-02-21 17:16:31

sctincman wrote:

Perhaps you could try appending this to your boot line, with KMS enabled
video=LVDS-1:1280x800

I was fooling around with this and found you have to take the number off, so it becomes

video=LVDS:1280x800

but it may be that your card is pushing to another output, for example

same problem here with my old thinkpad X30 (intel 830M)
I had that problem since several Kernelversions and disabled KMS because of that.
If I pass the video=LVDS-8:1024x768-24@60 it doesn't change anything (also if I use 800x600 etc.).
what drm says:
[~]$ dmesg |grep drm
[drm] Initialized drm 1.1.0 20060810
[<c11376db>] ? sysfs_addrm_start+0x5b/0xb0
[drm] set up 15M of stolen space
[drm] DAC-6: set mode 640x480 0
[drm] LVDS-8: set mode 1024x768 b
fb0: inteldrmfb frame buffer device
[drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

seems like it's also outputting to DAC, so perhaps try adding this to your boot line

video=DAC:off

or whatever else may be popping up in dmesg as being set

swiftscythe · 2010-02-24 17:08:02

Hi! I've tried everything you said, but still can't get the intel driver to work...
As I've said I'm using the vesa driver without KMS, and therefore can't rotate the screen of the tablet PC

Thanks for your support

DeeCodeUh · 2010-02-24 17:33:46

A temporary fix that works for me...

Install the old intel driver, and put i915.modeset=0 in the menu.lst file

If you are running i686

pacman -U http://arm.konnichi.com/2009/12/31/extra/os/i686/xf86-video-intel-2.9.1-1-i686.pkg.tar.gz

If you are running x86_64

pacman -U http://arm.konnichi.com/2009/12/31/extra/os/x86_64/xf86-video-intel-2.9.1-1-x86_64.pkg.tar.gz

Last edited by DeeCodeUh (2010-02-24 17:36:02)

swiftscythe · 2010-02-24 17:54:26

Thanks, that will do the trick for a while
At least now I can rotate the screen

Thank you very much!

Let's hope they fix it soon

DeeCodeUh · 2010-02-24 17:57:23

swiftscythe wrote:

Thanks, that will do the trick for a while
At least now I can rotate the screen
Thank you very much!
Let's hope they fix it soon

Yeah, I hope they do. I don't want to run old intel drivers forever.

mike_k · 2010-02-24 18:42:00

I have the same issue on another distro: CRUX. To fix I'm temporary using downgraded intel driver v2.9.1 (and xorg to 1.7.3). It isn't clear if upstream is investigating this issue though...

kokoko3k · 2010-02-24 19:11:14

sctincman wrote:

Perhaps you could try appending this to your boot line, with KMS enabled
video=LVDS-1:1280x800
with your desired resolution of course.
you may need to change LVDS-1, depending on your setup, check the output of xrandr -q to see what your display is named

Thanks for the hint and sorry for jumping in in this thread to submit a similar problem.
My eeepc 1005ha doesnt' "see" the lvds panel if the lid is closed at boot time.
The screen will stay just black, no console output, no xorg output.
I'm sure the system is starting because the hd led blinks as usual; also, if i open the lid in the middle of the boot process (before X), the console still stays blank, but X works.
I already tried the string you posted in menu.lst, but things remains the same.
What i suspect is that video output is sent over the vga port; is there a way to force the detection of the lvds panel at boot time, like xrandr?
Thanks in advance.

swiftscythe · 2010-02-27 15:13:10

Hi! I've just downloaded kernel26 from testing, and this problem has been solved:
http://bugs.freedesktop.org/show_bug.cgi?id=22936

Thanks to everyone

Arch Linux

#26 2010-02-13 19:12:19

Re: Intel driver 2.10 & KMS issue

#27 2010-02-13 19:22:29

Re: Intel driver 2.10 & KMS issue

#28 2010-02-13 19:28:38

Re: Intel driver 2.10 & KMS issue

#29 2010-02-13 19:34:02

Re: Intel driver 2.10 & KMS issue

#30 2010-02-13 19:40:42

Re: Intel driver 2.10 & KMS issue

#31 2010-02-13 19:49:07

Re: Intel driver 2.10 & KMS issue

#32 2010-02-13 19:50:13

Re: Intel driver 2.10 & KMS issue

#33 2010-02-13 19:56:45

Re: Intel driver 2.10 & KMS issue

#34 2010-02-13 20:38:16

Re: Intel driver 2.10 & KMS issue

#35 2010-02-13 20:45:38

Re: Intel driver 2.10 & KMS issue

#36 2010-02-13 20:59:45

Re: Intel driver 2.10 & KMS issue

#37 2010-02-13 23:19:38

Re: Intel driver 2.10 & KMS issue

#38 2010-02-13 23:26:39

Re: Intel driver 2.10 & KMS issue

#39 2010-02-14 01:46:05

Re: Intel driver 2.10 & KMS issue

#40 2010-02-14 10:37:47

Re: Intel driver 2.10 & KMS issue

#41 2010-02-20 13:26:57

Re: Intel driver 2.10 & KMS issue

#42 2010-02-20 13:49:43

Re: Intel driver 2.10 & KMS issue

#43 2010-02-21 17:16:31

Re: Intel driver 2.10 & KMS issue

#44 2010-02-24 17:08:02

Re: Intel driver 2.10 & KMS issue

#45 2010-02-24 17:33:46

Re: Intel driver 2.10 & KMS issue

#46 2010-02-24 17:54:26

Re: Intel driver 2.10 & KMS issue

#47 2010-02-24 17:57:23

Re: Intel driver 2.10 & KMS issue

#48 2010-02-24 18:42:00

Re: Intel driver 2.10 & KMS issue

#49 2010-02-24 19:11:14

Re: Intel driver 2.10 & KMS issue

#50 2010-02-27 15:13:10

Re: Intel driver 2.10 & KMS issue

Board footer