[KM400] Black screen after boot, no text console possible

After some research on launchpad, it seems more likely to be a bug in Plymouth than in xserver-xorg-video-openchrome. Moreover, I've completely removed Xorg and I still can't log in on a TTY

Could you test booting with older kernel. I got also black screen after update to lucid. Older kernel works for me.

Yes, karmic kernel 2.6.31-14-generic works just fine.

The symptoms you describe aren't attributable to plymouth. Sounds like you have a kernel problem.

Update: situation is unchanged after an update to kernel 2.6.32-13-generic.

Update: situation is unchanged after an update to kernel 2.6.32-14-generic.

And I unfortunately also cannot use lucid with kernel 2.6.32-15-generic.

Thanks for reporting this bug and any supporting documentation. Since this bug has enough information provided for a developer to begin work, I'm going to mark it as confirmed and let them handle it from here. Thanks for taking the time to make Ubuntu better!

I tried to append to the information here but I got an error:
""
apport-cli -u 518623
Error connecting to Launchpad: 'NoneType' object has no attribute 'makefile'
You can reset the credentials by removing the file "/home/su/.cache/apport/launchpad.credentials"
rm: cannot remove `/home/su/.cache/apport/launchpad.credentials': No such file or directory
""
I have attached a tar archive of some manually run commands (dmesg, lspci, cat /proc/cpuinfo and Xorg.conf and Xorg.0.log files). Apologies if this comment is a headache(not filed with apport) rather than being helpful.

--------------------------------
My report:

My system shows the symptoms described in this report + some more.
---------
lucid kernel: vmlinuz-2.6.32-15-generic freezes and does not respond to alt+Fn combination -- even in the recovery mode.
I have tried removing plymouth. That did not help.
---------
Using a karmic kernel "2.6.31-19-generic #56-Ubuntu SMP Thu Jan 28 01:26:53 UTC 2010 i686 GNU/Linux", I still get a black screen. But I am able to use alt+Fn in the recovery mode to get to a tty and work from there.

I have tried running "Xorg -configure" and tried using the new configuration(attachement xorg.newconf + Xorg.0.log.newconf), It did not help. But I was able to use the alt+Fn combination again and was able to send a 'ctrl+c' to the "X -c /usr/sbin/xorg.conf.new" command.

But when I run X -c /etc/X11/xorg.conf (attachment xorg.origconf + Xorg.0.log.origconf):
* I lose keyboard control (using a alt+sysrq+REISUB did not help).
* Doing a "kill -9" of X's pid, did not restore keyboard control (numlock led does not toggle for example)
* vncsession still works fine (this comment was added from there.

just for the sake of completeness: also kernel 2.6.32-16-generic shows the same symptoms, as originally described, on my machine

The last boot up message that could be seen in the recovery mode is the fsck output from util-linux-ng. Appears to hang sometime after that.
Is there someway by which I can confirm the exact process that leads to this? (note that every time I try something, it hangs and I have to 'hard reboot' to an older kernel to get to a virtual console -- this might overwrite some logs?).

Marked to triaged again; was changed to confirmed by mistake according to a message received:

From: Nicholas Christian Langkjær Ipsen
To: Charlie Kravetz
Subject: Oops...
Date: Mon, 15 Mar 2010 08:25:59 -0000

Could you please mark Bug518623 as triaged again? I was an idiot and
marked it down as confirmed, thinking it was another bug >.>

I can confirm it's a regression in 2.6.32. Kernel 2.6.31 works well (tested with 2.6.31.12).
I hit this bug when I boot into single mode (recovery mode). Text cursor in top left corner appears instead of recovery menu. Combination "(CTRL) + Alt + Fx" doesn't work, I have to restart using Ctrl + Alt + Delete.
I have no issue in non-single mode. X Windows start and Gnome works well.

I can fix it by parameter radeon.modeset=0 on kernel command line:
-----------
kernel /vmlinuz-2.6.32-16-generic root=UUID=5564d628-8fc2-42b9-ba8d-cb1b7937a3e8 ro radeon.modeset=0 single

Used SW:
Ubuntu 10.04, amd64
kernel 2.6.32-16-generic
xserver-xorg-video-radeon: 1:6.12.191-1ubuntu2

Just for the sake of completeness: also kernel 2.6.32-16-generic shows the same symptoms, as originally described, on my machine. Install of daily image from day 18.03.2010.

Kernel 2.6.32-17-generic shows the same symptom as described by Robert Moerland and gokul. This is probably a kernel problem as this also occurs on a commandline-only installation. Steve Langasek is right that this doesn't have to do with Plymouth. Removing that merely adds errors to the boot messages. However, this definately has to do with this particular chip:

01:00.0 VGA compatible controller: VIA Technologies, Inc. KM400/KN400/P4M800 [S3 UniChrome] (rev 01)

I swap this chip (by inserting a graphics add-on board, this disables the integrated S3 chip) and Ubuntu boots. Take the working chip out and the system hangs again, every time. This is all that's shown:

[screen full of kernel messages, mostly related to the hard drive]
Begin: Running /scripts/local-bottom ...
Done.
Done.
Begin: Running /scripts/init-bottom ...
Done.
fsck from util-linux-ng 2.17
/dev/sdb1: clean, 44079/19365888 files, 49698348/77437308 blocks (check after next mount)

That last line varies.

This is sounding very similar to the problems I'm encountering in bug 499399. I can't boot a Lucid install unless I use the Karmic kernel. If I try a Lucid kernel (or other Ubuntu mainline kernel >= 2.6.32) I get boot messages like those in Paul van Genderen's message above, stopping at the same point. It makes no difference whether it's command-line only or a full install.

Removing Plymouth didn't achieve anything useful.

In my case, lspci shows:
01:00.0 VGA compatible controller: VIA Technologies, Inc. CN896/VN896/P4M900 [Chrome 9 HC] (rev 01)
Not the same, but similar?

I don't have an add-on graphics board to test with.

I can boot other distros which use the 2.6.32 kernel. I conclude therefore that it's not the 2.6.32 kernel per se, but something in the way in which the Ubuntu versions are configured.

I can also confirm this. I have a VIA CX700 graphic kernel 2.6.32-16. One thing I found is if I install Lubuntu 2.6.32-16 the graphics are normal, then install Ubuntu-desktop it still runs good. But if I uninstall Lubuntu-desktop I get the black screen again. I can also confirm plugging external monitor into laptop works.

I had noticed that pressing ctrl+alt+del at some magic moment after the fsck message takes me to gdm. I managed to scrape through today and get some logs.

The mouse was not detected (PS/2).
Looks like there is some issue with "drm"... I saw messages to that effect in dmesg and xorg.0.log. I have no clue what that means.

kern.log:
Mar 24 11:01:18 sea kernel: [ 1747.799491] [drm:via_initialize] *ERROR* called with no agp memory available

xorg.0.log:
(EE) CHROME(0): [drm] Failed to initialize DMA ring-buffer: 14

I will attach the logs that I managed to salvage from that boot. Hope it helps.

Xorg.0.log

daemon.log

syslog

I could not run apport-collect for adding comments #19 to #22 for reasons mentioned in comment #10.

Gokul: That's brilliant! Any ballpark idea at what point in the boot process CAD works?

It was basically just trial and error. You might try under clocking your cpu just in case (might not actually help). Even in verbose mode (i.e, without quiet and splash for the kernel command line), the final message I see is the one from fsck.

I have tried blindly firing away C+A+D after the last udev message till fsck shows something. And it has worked a couple of times. Any earlier or later simply reboots the machine (sometimes it brings up the plymouth splash before it freezes or reboots after the CAD).

I was hoping that:
* the logs would be of better help to narrow down the process
* if upstart could be made non-parallel, it might make it a little easier to narrow down the problem source

I have no expertise here. I will have to do some more whimsical digging.

I've got the same problem here on an Acer Aspire 1353XC Laptop with Beta1, but the Live CD (well, on a USB stick) installer worked OK aside from corrupted graphics during the bootsplash. So what is the difference between the Live CD boot and the normal install boot, that means the Live CD works OK?

I (the original reporter) can confirm the observed behavior by Jon Welsh (#26). I'm typing this response on a version 2.6.32-16 kernel-based lucid, booted from a USB stick. The same kernel version, when booted from hard disk after an upgrade from karmic (see original description) doesn't work at all.

Small addition, the virtual terminals cannot be used by me from within the live cd environment, the screen that is shown is corrupt. Furthermore, I actually used the Xubuntu live cd instead of the Ubuntu live cd.

Just to note that I seem to be suffering from the same issue, however I have found if I boot with the livecd (which works) and then restart (NOT Shutdown) the system will boot normally. I don't no if this will be of any help, or if others can confirm.

Just tested rebooting from livecd, and I still get the black screen on booting the normal install. So this doesn't work for my hardware,

Try disable DRI in following way (create/edit xorg.conf file):
    Section "Module"
       ...
       Disable "dri"
       ...
    EndSection

Try also disable AIGLX. Unfortunately this is only workaround.

Adding:
Section "Module"
       Disable "dri"
EndSection

to my xorg.conf file didn't make a difference, neither did adding 'Disable "glx"'. I also tried removing the xorg.conf file altogether, but also that didn't have any visible effect.

@Bartosz: your involvement in this bug, does that mean you suspect the openchrome driver to be related to this?

update:
On my installation, with the latest packages,

1. with kernel 2.6.32-18-generic, the boot still freezes after fsck message

2. with kernel 2.6.31-19-generic, I get gdm and gnome desktop at low resolution (800x600) and only other mode shown by xrandr was 640x480.
I manually added a monitor section to provide the horizontal and vertical sync (from values I got off the net) and configured a modeline for 1280x1024. (using cvt)
with this:
  * Using the custom xorg.conf, now xrandr (also the "monitors" menu item under system->preferences->monitors) shows a lot more mode options and I have configured it to use 1280x1024.
  * Even after the custom xorg.conf, kernel 2.6.32-18-generic, still freezes.
  * Plymouth splash is not seen at boot or shutdown/restart.

Hardware: Laptop: MiTac 8615P with chipset: VIA VN896
Software:Ubuntu 10.04, the latest release.
Video driver: 1:0.2.904+svn827-1

My laptop freeze while booting, only a black screen is all I can see.
The Ubuntu boot logo appears the left-down corner of the screen, so
probably the resolution is much bigger than required.
If an external monitor is connected to my laptop the system stand up
well, but only the external monitor works properly nevertheless the
laptop screen stills black.
Using Debian 6 testing squeeze alpha 1 (wherein the kernel is also
2.6.32 and video driver) works fine via LiveCD without problems.

Same bug for me -- laptop Fujitsu Siemens Amilo Pro v2010, graphics card: VIA PN800 or PN880 (don't remember which one; I doubt it's important here), kernel 2.6.23-16-generic. Laptop hangs after fsck with completely black screen. Interstingly, when I boot into my old Ubuntu 8.10 and immediately reboot to 10.04 -- it _always_ works. And it does not matter if it's hard reboot or soft. With hard reboot it works if I don't make long break between power down and turing on again. This is strange, as if RAM fading out effects had some influence here? (using uninitalized state? Just guessing...).

A have same problem with ubuntu 10.4 (2.6.32-20-generic), but kernel 2.6.31-20-generic works fine. I tried change driver to "vesa" or disabled module dri, glx, but that didn't have any effect. Now, I have to boot with old kernel.

As mentioned on the duplicate I just filed: adding "blacklist viafb" and "blacklist vga16fb" to /etc/modprobe.d/blacklist.conf makes things work here.

Karmic's kernel (2.6.31-20-generic) also works, but that is because viafb is not loaded there by default, not because viafb works. If I stop gdm and "sudo modprobe viafb" the panel turns off (I'm assuming this is the exact same bug triggering).

There are some changes in the current upstream kernel's viafb. It's possible those help, but I have not gotten around to trying yet.

Would it make sense to blacklist viafb for now, at least on this chipset? Fixing this is a bit of a hassle because neither X nor the console work with this bug present.

Confirming workaround by Marien Zwart. After blacklisting both viafb and vga16fb, I can boot kernel 2.6.32.20 and can use my laptop normally. Don't know if there are any (major) drawbacks involved with blacklisting these modules.

No, the upstream changes in viafb won't help the problem you are experiencing and it is unknown whether viafb and/or openchrome (which should not rely on the output configuration to be correct) will ever be fixed as this would require an amount of documentation that is not yet released by VIA, especially for older IGPs like CLE266. However you can likely make it work with some module parameters as described in viafb.txt:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=Documentation/fb/viafb.txt;h=f3e046a6a987a685910bafc171c2cc9cf243df1b;hb=HEAD
If you also hit some acceleration bugs that where fixed in 2.6.33 you can use viafb_accel=0
This bug is similiar/identical to this one I think
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-via/+bug/538430
I do not object blacklisting viafb for this release as there exists also an even more serious problem:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/539020

However blacklisting vga16fb sounds like a bad idea but that's not something I work on. (but even balcklisting it wouldn't be that much of a problem but it really should simply work)
Perhaps it would be a good idea to do one seperate bugreport only for the vga16fb bug which gives a more verbose description of what the screen looks like?

With the workaround from comment #37 (blacklisting viafb), my boot process completes without freezing [mentioned in my earlier comments].

[I realize that the following may not be related to this bug. But I have mentioned it just in case]:
After the boot, I have to do a Ctrl+Alt+Del to get to gdm [not at random]--- otherwise the monitor (HP M900) displays a message "Invalid Scanning Frequency" and I can see scan lines in the screen(this happens with and without an xorg.conf). After the ctrl+alt+del gdm comes up at the right resolution if an xorg.conf is in place (a low res screen otherwise).
At present:
1) with neither the karmic kernel (2.6.31-19) nor the lucid kernel (2.6.32-20), Xorg is able to set my monitor resolution correct. I need to have a customized xorg.conf.
2) with karmic kernel (2.6.31-19), gdm is reached without the ctrl+alt+del quirk.

My vga16fb problem may be the same one from bug 538893 comment #1, at least in that the garbage I see in text mode seems to be influenced by what was on the screen in X. And yes, that's a different bug. Blacklisting vga16fb is not necessary as long as X works: I get a brief period of garbage between vga16fb and X kicking in, but X works just fine. It's viafb that is the real problem here, because X (xserver-xorg-video-openchrome) does not reset things back to a usable state.

There is the same issue on a desktop computer with "VIA Technologies. CN896/VN896/P4M900 [Chrome 9 HC]"...

9.10 -> 10.4 upgrage (lucid lynx with kernel 2.6.32-22) has been done, and after grub computer freezes...
No CLI by Ctrl-Alt-F1/F2/F?..
No GDM (GNOME/LXDE)...
Just reboot by Ctrl-Alt-Del...

10.4 (lucid) kernel 2.6.32-22 does not work... But 9.10 (karmic) kernels -- 2.6.31.* -- works...

# lspci | grep VGA
01:00.0 VGA compatible controller: VIA Technologies, Inc. CN896/VN896/P4M900 [Chrome 9 HC] (rev 01)

update:
In mavreick beta, there is no bug. I have removed viafb from the blacklist. Splash screen at boot and shutdown/restart are also working.
Not sure if some earlier update to lucid had already fixed it.

Robert Moerland, this bug was reported a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue? If so, could you please test for this with the latest development release of Ubuntu? ISO images are available from http://cdimage.ubuntu.com/daily-live/current/ .

If it remains an issue, could you please run the following command in the development release from a Terminal (Applications->Accessories->Terminal), as it will automatically gather and attach updated debug information to this report:

apport-collect -p linux <replace-with-bug-number>

Also, could you please test the latest upstream kernel available following https://wiki.ubuntu.com/KernelMainlineBuilds ? It will allow additional upstream developers to examine the issue. Please do not test the daily folder, but the one all the way at the bottom. Once you've tested the upstream kernel, please comment on which kernel version specifically you tested. If this bug is fixed in the mainline kernel, please add the following tags:
kernel-fixed-upstream
kernel-fixed-upstream-VERSION-NUMBER

where VERSION-NUMBER is the version number of the kernel you tested. For example:
kernel-fixed-upstream-v3.11

This can be done by clicking on the yellow circle with a black pencil icon next to the word Tags located at the bottom of the bug description. As well, please remove the tag:
needs-upstream-testing

If the mainline kernel does not fix this bug, please add the following tags:
kernel-bug-exists-upstream
kernel-bug-exists-upstream-VERSION-NUMBER

As well, please remove the tag:
needs-upstream-testing

Once testing of the upstream kernel is complete, please mark this bug's Status as Confirmed. Please let us know your results. Thank you for your understanding.

[Expired for linux (Ubuntu) because there has been no activity for 60 days.]