[i855] Karmic Koala Alpha 3 live CD give black screen

[Laurent]

On Karmic Koala 9.10 Alpha 3 live Cd, after ISOLINUX menu the screen become definitivly black.

Cannot give you more info, as I cannot access the console (even with CTRL-ALT-Fx)

Toshiba laptop with Intel 855GM graphic card

ProblemType: Bug
Architecture: i386
DistroRelease: Ubuntu 9.04
MediaBuild: Ubuntu 9.10 "Karmic Koala" - Alpha i386 (20090722.2)
Package: xorg 1:7.4~5ubuntu18
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=fr_FR.UTF-8
ProcVersion: Linux version 2.6.30-020630-generic (root@zinc) (gcc version 4.2.3 (Ubuntu 4.2.3-2ubuntu7)) #020630 SMP Wed Jun 10 09:45:40 UTC 2009
SourcePackage: xorg
Uname: Linux 2.6.30-020630-generic i686

[Fatih Aşıcı (fatih-pardus)]

Created an attachment (id=27728)
rom.bin

[Michael Fu (michael-fu-intel)]

need a xorg.log with option "modedebug" "true" in xorg.conf.. thanks.

[Gordon Jin (gordon-jin)]

which version of intel driver ever worked on this machine?

[Fatih Aşıcı (fatih-pardus)]

(In reply to comment #2)
> need a xorg.log with option "modedebug" "true" in xorg.conf.. thanks.
>

The server is already patched to make that option true by default.

(In reply to comment #3)
> which version of intel driver ever worked on this machine?

The user reports that he can install Pardus 2009 Alpha shipping with intel driver 2.7.1.

[Fatih Aşıcı (fatih-pardus)]

(In reply to comment #4)
> (In reply to comment #3)
> > which version of intel driver ever worked on this machine?
>
> The user reports that he can install Pardus 2009 Alpha shipping with intel
> driver 2.7.1.
>

Sorry, he might be installed using fbdev. I am waiting a reply from him.

The user had reported the same issue when we use intel driver 2.4.3. We solved the issue with the patches here:

http://svn.pardus.org.tr/pardus/2008/devel/desktop/freedesktop/xorg/xorg-video/files/intel/

Probably this one solved:

http://svn.pardus.org.tr/pardus/2008/devel/desktop/freedesktop/xorg/xorg-video/files/intel/ubuntu/26_i830-use-lfp-data-ptrs.patch

[Fatih Aşıcı (fatih-pardus)]

(In reply to comment #5)
> (In reply to comment #4)
> > (In reply to comment #3)
> > > which version of intel driver ever worked on this machine?
> >
> > The user reports that he can install Pardus 2009 Alpha shipping with intel
> > driver 2.7.1.
> >
>
> Sorry, he might be installed using fbdev. I am waiting a reply from him.

Yes, he said that he used intel driver 2.7.1 and no problem occurs. It doesn't work after upgrading to 2.7.99.902.

[Michael Fu (michael-fu-intel)]

would you please ask the bug reporter to attach the following:

1) a xorg.log with option "modedebug" "true" in xorg.conf
2) vbios dump. you can get vbios dump via:

# cd /sys/devices/pci0000\:00/0000\:00\:02.0/
# echo 1 > rom
# cat rom > /tmp/rom.bin
# echo 0 > rom

thanks.

[yakuizhao (yakui-zhao)]

> Yes, he said that he used intel driver 2.7.1 and no problem occurs. It doesn't
> work after upgrading to 2.7.99.902.
From the Xorg log it seems that there is no EDID for LVDS. At the same time from the rom.bin it seems that there is no integrated LVDS controller.
   >LVDS config: No LVDS
Will you please ask the bug reporter double check whether the 2.7.1 driver can work well? If it can work, please attach the Xorg log on intel driver 2.7.1.
Please also attach the output of acpidump, dmidecode, lspci -vxxx.
Thanks.

[Michael Fu (michael-fu-intel)]

(In reply to comment #7)
> 2) vbios dump. you can get vbios dump via:
>
> # cd /sys/devices/pci0000\:00/0000\:00\:02.0/
> # echo 1 > rom
> # cat rom > /tmp/rom.bin
> # echo 0 > rom
>
> thanks.
>
Oh, it's been attached. stupid me... never mind..

[Fatih Aşıcı (fatih-pardus)]

Created an attachment (id=27812)
Xorg.0.log with intel driver 2.7.1

[Fatih Aşıcı (fatih-pardus)]

Created an attachment (id=27813)
Output of acpidump

[Fatih Aşıcı (fatih-pardus)]

Created an attachment (id=27814)
Output of dmidecode

[Fatih Aşıcı (fatih-pardus)]

Created an attachment (id=27815)
Output of lspci -vxxx

[yakuizhao (yakui-zhao)]

Created an attachment (id=27837)
try the debug patch in which the DVO timing is obtained by using the DVO timeoffset in LVDS data entry

Will you please try the debug patch and see whether the issue still exists?
Thanks.

[Fatih Aşıcı (fatih-pardus)]

(In reply to comment #14)
> Created an attachment (id=27837) [details]
> try the debug patch in which the DVO timing is obtained by using the DVO
> timeoffset in LVDS data entry
>
> Will you please try the debug patch and see whether the issue still exists?
> Thanks.
>

This patch fixes the issue. Thanks.

[yakuizhao (yakui-zhao)]

*** Bug 22916 has been marked as a duplicate of this bug. ***

[Frédéric Péters (fpeters)]

I am from the duplicated bug report and the patch doesn't fix the issue; I'll attach a serie of files.

[Frédéric Péters (fpeters)]

Created an attachment (id=27967)
fredp - rom.bin

[Frédéric Péters (fpeters)]

Created an attachment (id=27968)
fredp - dmidecode

[Frédéric Péters (fpeters)]

Created an attachment (id=27969)
fredp - acpidump

[Frédéric Péters (fpeters)]

Created an attachment (id=27970)
fredp - lspci -vxxx

[Frédéric Péters (fpeters)]

Created an attachment (id=27971)
fredp - xorg.log (2.8.0+patch from comment 14)

[Frédéric Péters (fpeters)]

Created an attachment (id=27972)
fredp - xorg.log (2.7.99.901)

[yakuizhao (yakui-zhao)]

(In reply to comment #22)
> Created an attachment (id=27971) [details]
> fredp - xorg.log (2.8.0+patch from comment 14)
Will you please double check it again?
I already check the vbios bin and find that the correct preferred mode can be parsed from VBT after applying the patch.

Thanks.
>

[yakuizhao (yakui-zhao)]

(In reply to comment #22)
> Created an attachment (id=27971) [details]
> fredp - xorg.log (2.8.0+patch from comment 14)
Will you please double check it again?
I already check the vbios bin and find that the correct preferred mode can be
parsed from VBT after applying the patch.
But from the xorg log it seems that it still can't be parsed correctly.
Please double check it again.

Thanks.

>

[Frédéric Péters (fpeters)]

Thanks for checking back, indeed I was wrong and somehow the drivers got installed into an incorrect location. Once installed properly X started correctly.

[Remi (remi)]

Can/Should the patch be committed to git?

Thanks

[Michael Fu (michael-fu-intel)]

commit cdaa052b05e26d96a990af5d253fd2af5db2b1fc
Author: Zhao Yakui <email address hidden>
Date: Tue Jul 28 10:54:13 2009 +0800

    drm/I915: Fix offset to DVO timings in LVDS data

[Laurent (l-giuliani)]

On Karmic Koala 9.10 Alpha 3 live Cd, after ISOLINUX menu the screen become definitivly black.

Cannot give you more info, as I cannot access the console (even with CTRL-ALT-Fx)

Toshiba laptop with Intel 855GM graphic card

ProblemType: Bug
Architecture: i386
DistroRelease: Ubuntu 9.04
MediaBuild: Ubuntu 9.10 "Karmic Koala" - Alpha i386 (20090722.2)
Package: xorg 1:7.4~5ubuntu18
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=fr_FR.UTF-8
ProcVersion: Linux version 2.6.30-020630-generic (root@zinc) (gcc version 4.2.3 (Ubuntu 4.2.3-2ubuntu7)) #020630 SMP Wed Jun 10 09:45:40 UTC 2009
SourcePackage: xorg
Uname: Linux 2.6.30-020630-generic i686

[Hadmut Danisch (hadmut)]

I have a similar problem with an Intel 845G graphics chipset:

After ISOLINUX I see the bootsplash, but as soon as X11 starts screen becomes black, except for the mousepoint. Mousepointer still moves, so the machine and the server did not crash.

Please allow switching to console mode.

[Laurent (l-giuliani)]

For me, no mousepoint whole sceen black ..

[Bryce Harrington (bryce)]

https://wiki.ubuntu.com/X/Troubleshooting/BlankScreen has some info on working around and analyzing this class of bug.

[Remi (remi)]

I can't find the corresponding commit in xf86-video-intel. Has it been committed?

Thanks

[yakuizhao (yakui-zhao)]

the commit in comment #28 is for the KMS version.
The commit id for UMS is

commit af45482a52999b52bf41468c458808e30c100e35
Author: Zhao Yakui <email address hidden>
Date: Fri Jul 24 10:44:20 2009 +0800

    Calculate the DVO relative offset in LVDS data entry to get the DVO timing

Thanks.

[Laurent (l-giuliani)]

Further tests:

   - Disable usplash doesn't correct the problem
   - Lower/other resolution doesn't correct the problem

Any other idea of test to do?

[Laurent (l-giuliani)]

 Register dump when it work in Jaunty with intel 2:2.7.1-0ubuntu1~xup~1

I'am still trying too find a way to do the same with Kamic koala when it's not good.

[Laurent (l-giuliani)]

UPDATE : My problem seems linked to KMS as i915.modeset=0 permit to have a console but X don't start because of no valid mode detect for LVDS

Perhaps link this bug is the same as Bug #407804 but as I have a different LID my screen became black?

I attached karmic dmesg and XOrg.0.log

[Bryce Harrington (bryce)]

Laurent - I've forwarded this bug upstream to https://bugs.freedesktop.org/show_bug.cgi?id=23399 - please subscribe yourself to this bug, in case they need further information or wish you to test something. Thanks ahead of time!

[yakuizhao (yakui-zhao)]

Will you please try the latest driver and see whether the issue still exists?
Will you please also attach the vbios.dump by using the following commands?
   echo 1 > /sys/devices/pci0000:00/0000:00:02.0/rom
   cat /sys/devices/pci0000:00/0000:00:02.0/rom > vbios.dump
   echo 0 > /sys/devices/pci0000:00/0000:00:02.0/rom
Thanks.

[Stefan Glasenhardt (glasen)]

Hi,

This is definitely not a xorg-driver bug. It is a kernel bug. I suffer from the same problem (Dell Latitude D505, Intel 855GM) and only disabling all display-modes (Ext. VGA, HDMI, TV-Out, etc.) except the internal display (LVDS) in the kernel source file "drivers/gpu/drm/i915/intel_display.c" in lines 69852-70373 solves the problem :

        list_for_each_entry(connector, &mode_config->connector_list, head) {
                struct intel_output *intel_output = to_intel_output(connector);

                if (!connector->encoder || connector->encoder->crtc != crtc)
                        continue;

                switch (intel_output->type) {
/* case INTEL_OUTPUT_LVDS: */ -> Always enable LVDS
                        is_lvds = true;
/* break;
                case INTEL_OUTPUT_SDVO:
                case INTEL_OUTPUT_HDMI:
                        is_sdvo = true;
                        if (intel_output->needs_tv_clock)
                                is_tv = true;
                        break;
                case INTEL_OUTPUT_DVO:
                        is_dvo = true;
                        break;
                case INTEL_OUTPUT_TVOUT:
                        is_tv = true;
                        break;
                case INTEL_OUTPUT_ANALOG:
                        is_crt = true;
                        break;
                case INTEL_OUTPUT_DISPLAYPORT:
                        is_dp = true;
                        break;
                case INTEL_OUTPUT_EDP:
                        is_edp = true;
                        break;*/
                }

                num_outputs++;
        }

It seems that the kernel does not correcty handle the display pipes of the intel chipset and always enables the VGA-port as first display and oversees the internal display. I tested the kernel with two external display (VGA and DVI-port) and both display work perfectly. Only the internal display isn't initialized correctly.

[yakuizhao (yakui-zhao)]

*** Bug 23399 has been marked as a duplicate of this bug. ***

[yakuizhao (yakui-zhao)]

Hi, Stefan
     This issue is not related with the kernel. The issue can be reproduced regardless of whether KMS or UMS is used.
     In fact the main issue of this bug is that we can't get the fixed mode correclty for LVDS panel in case that there is no EDID for LVDS.
    So please focus on that no fixed mode is obtained for LVDS output device.
    We have fixed the similar issues on another box based on 855GM chipset.
     >http://bugs.freedesktop.org/show_bug.cgi?id=22787

Thanks.

[Stefan Glasenhardt (glasen)]

I just updated to kernel-version 2.6.31-rc9 (vanilla) and KMS is now fully functional on my 855GM-chipset. No more black screen, usplash shows up, etc.

There was no change to the xorg-intel-driver in the last few weeks, so it must have been a kernel problem. Waiting for an official upgrade to the ubuntu-kernel should solve the problem once and forever.

Here is the corresponding bug entry for the kernel :

http://bugs.freedesktop.org/show_bug.cgi?id=20115

[Fabus (fabian-gebert-hh)]

Tried the latest daily karmic ubuntu desktop i386 (as of 2009-09-07) but it still does not work.
The system still reacts however, as I can tell from being able to shut it down using the ACPI shutdown method triggered by a short tap on the power switch.

[Florian Reinhard (freinhard)]

Could you be more specific about your hardware?

Alpha5 boots and installs just fine on my Laptop with 855GM.

[Fabus (fabian-gebert-hh)]

Hardware: mac mini, not the latest release but the one before. lspci says
00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)
In fact, Alpha 4 booted up with a wrong resolution (I already posted that bug).
From Alpha 5 on, nothing shows up at all (display goes to standby).

I thought this bug would be related?

[Khashayar Naderehvandi (khashayar)]

The upstream bug referenced here is marked a dupe of fdo #22787 [1]. And that bug in turn is apparently resolved. Is that commit already in karmic? Or is this, in fact, another bug?

[1] http://bugs.freedesktop.org/show_bug.cgi?id=22787

[Reuben Thomas (rrt)]

I'm experiencing the same problem with linux-image-2.6.31-11-generic 2.6.31-11.36 (i.e. latest karmic as of writing) on a first-gen Intel Mac mini.

[Reuben Thomas (rrt)]

Also, as suggested elsewhere, if I add i915.modeset=0 to my kernel boot line, the problem is resolved (by using UMS rather than KMS).

[Fabus (fabian-gebert-hh)]

I think this might be related to #431812

[Khashayar Naderehvandi (khashayar)]

For what it's worth, I no longer have this problem on a fully up-to-date karmic system.

[Launchpad Janitor (janitor)]

This bug was fixed in the package xserver-xorg-video-intel - 2:2.9.0-1ubuntu1

---------------
xserver-xorg-video-intel (2:2.9.0-1ubuntu1) karmic; urgency=low

  [Tormod Volden]
  * Merge with Debian. Remaining Ubuntu changes:
    + debian/control: Add lpia architecture
  * Drop patches now included upstream:
    + 100_8xx_perf_pict_a8.patch
    + 101_reload_cursors.patch
    + 102_dont_destroy_bufmgr_at_closescreen_time.patch
  * Ubuntu bugs fixed in this release:
    + Calculate the DVO relative offset in LVDS data entry to get the DVO
      timing. Fixes black screen at boot on i855 due to not finding LVDS
      modes.
      (LP: #410455)
    + Fix brightness by enabling BACKLIGHT property for KMS.
      (LP: #397617)

 -- Bryce Harrington <email address hidden> Wed, 07 Oct 2009 20:09:54 -0700