Bug 10892 (SCREENBLACKCOLDBOOT)

Summary: Sometime (often) X come out blank (black screen) on cold boot - Intel chipset
Product: Other Reporter: Romano Giannetti (romano.giannetti)
Component: OtherAssignee: other_other
Status: CLOSED OBSOLETE    
Severity: high CC: alan, bunk, jbarnes, kernel-bugs
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 2.6.26-rc5 Subsystem:
Regression: Yes Bisected commit-id:

Description Romano Giannetti 2008-06-10 05:33:27 UTC
Latest working kernel version: 2.6.25
Earliest failing kernel version: After 2.6.25, exatcly don't know, probably 2.6.26-rc3
Distribution:Ubuntu 8.04 Hardy
Hardware Environment: toshiba U305 intel-based laptop, see http://bugzilla.kernel.org/attachment.cgi?id=16067&action=view
Software Environment:xserver-xorg-video-intel  2:2.2.1-1ubuntu12, see also http://bugzilla.kernel.org/attachment.cgi?id=16066&action=view
Problem Description:

Sometime, on cold boot (after the laptop has been shut off), when the screen switch from the Ubuntu startup to gdm, the screen goes blank. Forever. Rebooting (warm) with the same kernel cures it, sometime. If I reboot 2.6.25 it cures it everytime. 

Otherwise, the system seems to be working.

It was difficult to see because I almost never do a cold boot, the laptop is either sleeping (s2ram) or on, or (warm) rebooted when I test a new kernel. 

Helps / tricks to gather more data when it happens are welcome.
Comment 1 Rafael J. Wysocki 2008-06-10 12:06:50 UTC
This entry is being used for tracking a regression from 2.6.25.  Please don't
close it until the problem is fixed in the mainline.

References : http://lkml.org/lkml/2008/6/10/137
Comment 2 Rafael J. Wysocki 2008-06-15 03:47:53 UTC
On Sunday, 15 of June 2008, you wrote:
> On Sat, 2008-06-14 at 22:12 +0200, Rafael J. Wysocki wrote:
> 
> > 
> > Bug-Entry   : http://bugzilla.kernel.org/show_bug.cgi?id=10892
> > Subject             : Sometime (often) X come out blank (black screen) on
> cold boot - Intel chipset
> > Submitter   : Romano Giannetti <romano.giannetti@gmail.com>
> > Date                : 2008-06-10 05:33 (5 days old)
> > References  : http://lkml.org/lkml/2008/6/10/137
> 
> It happened, again, with the new Ubuntu x-intel driver. Rebooting with
> nousplash solved the problem, but alas, sometime simply rebooting solves
> the problem. It seems a race...
Comment 3 Romano Giannetti 2008-06-24 23:10:56 UTC
Bad luck. Happened again. 

OTOH, it happened with 2.6.25.6 too. 

rebooting with nosplash cured the problem.
Comment 4 Romano Giannetti 2008-06-25 06:24:21 UTC
Again, with -rc8. It seems it happens when switching from the bootsplash (I think it is X in VGA mode, I do not have any fb compiled in) to X intel in full resolution. 

I think I will add nosplash as default, although if anyone has something for me to check, I will. 
Comment 5 Adrian Bunk 2008-06-25 08:22:59 UTC
(In reply to comment #3)
> Bad luck. Happened again. 
> 
> OTOH, it happened with 2.6.25.6 too. 
>...

Is 2.6.25 good and 2.6.25.6 bad or was it simply luck that 2.6.25 worked?
Comment 6 Anonymous Emailer 2008-06-25 09:07:14 UTC
Reply-To: romanol@upcomillas.es


On Wed, 2008-06-25 at 08:22 -0700, bugme-daemon@bugzilla.kernel.org
wrote:
> 
> Is 2.6.25 good and 2.6.25.6 bad or was it simply luck that 2.6.25
> worked?

At this point, I really do not know. It seems to happen quite randomly.
I am testing for it now; after noticing, I think I did a cold boot once
in month, so I cannot say anything for sure. 

Romano
Comment 7 Romano Giannetti 2008-06-25 10:36:50 UTC
Ok, last phrase did not parse. What I tried to say, is that *before* noticing the bug, I hardly did a cold boot (I reboot for a new kernel, and then s2ram everytime), so for what I know, that bug could be here since ever...
Comment 8 Adrian Bunk 2008-06-25 11:18:25 UTC
Thanks for this update.

Until shown otherwise this bug should therefore not be tracked as a regression.
Comment 9 Jeremy Nickurak 2008-06-25 12:58:23 UTC
I do reboot quite regularly under 2.6.24 and recently 2.6.25, and I get this issue VERY regularly under 2.6.25, and have never seen it under 2.6.24.
Comment 10 Adrian Bunk 2008-06-25 13:08:30 UTC
Romano, is nosplash confirmed to help for sure? 
Comment 11 Adrian Bunk 2008-06-25 13:29:51 UTC
Wait, there is no nosplash option that could have any effect for you.
Comment 12 Romano Giannetti 2008-06-25 13:36:15 UTC
Adrian: I do not understand your last phrase. Until now, booting with nosplash (to be exact: editing the grub entry and changing splash to nosplash) did the trick. 
If you were referring to not having fb compiled, I can say that even without fb, I have the splash screen. Not sure how. I suppose Ubuntu starts an X server in VGA 640x480 mode (which by the way has a lot of problems, lost pixels, etc, since ever).
Comment 13 Jeremy Nickurak 2008-06-25 13:45:30 UTC
It certainly seems to me that this is a regression from 2.6.24.

In case it's not clear, the video seems to be knocked for a loop by this. Once the X server starts, even switching back to tty1 or closing the X server doesn't seem to help (although the system is responsive. You can log in and do things blind from the console)
Comment 14 Adrian Bunk 2008-06-25 14:04:57 UTC
Sorry Romano, I thought you were talking about a kernel commandline but you are talking about the grub splashscreen support.
Comment 15 Romano Giannetti 2008-06-26 01:11:58 UTC
Well, I was talking about kernel commandline. What I think it happens is that 
Ubuntu's initrd read the splash/nosplash and acts... it does

for x in $(cat /proc/cmdline); do
        case $x in
        splash*)
                SPLASH=true

and then uses usplash, which I really do not know how to work. It seems that it shouldn't work without framebuffers... hmm... let me dig a bit my .config.

 
Comment 16 Romano Giannetti 2008-06-26 01:16:09 UTC
There is a strange thing that happens to me. VESA/VGA framebuffer option appears ONLY if I choose framebuffer compiled in, not module; my previous configuration was with framebuffer compiled as module, uvesa and intelfb compiled but not used:

# lsmod | grep fb 
 
I am compiling with VESA/VGA framebuffer compiled in, now. 
Comment 17 Romano Giannetti 2008-06-26 02:01:08 UTC
Data points:

VESA compiled in. Booting with video=vesa vga=ask init=/bin/bash and choosing any of the VESA mode results in black display. No console. 

Choosing a normal VGA mode (80x60) gives the splash screen. Trying to switching consoles while the splash screen is in creates a flashing show of rubbish, no usable X starts.

Starting normally (splah, no video option) gives the splash screen, and then a black screen.

Starting with nosplash will give a correct (although ugly) 80x50 console and a working X.

Conclusion: 
1) in early boot, usplash digs the bowels out of the intel graphic card. 
2) VESA modes are busted with the intel graphic card.
Comment 18 Romano Giannetti 2008-07-07 02:57:00 UTC
Hmmm.

2.6.26-rc9: booted two times in a row (no vga nor video options) options, with splash screen, all seems ok. 

Will continue to test, because it seemed fixed another time, but well, something changed.
Comment 19 Romano Giannetti 2008-07-08 01:53:44 UTC
Nope. Happened again. Rebooted and then it was fine, but...
Comment 20 Alan 2012-05-12 16:29:28 UTC
Please re-open if this is incorrect