Natty Alpha 1, i915 has blank screen after boot

Bug #683775 reported by Tim Gardner
136
This bug affects 22 people
Affects Status Importance Assigned to Milestone
Linux
Fix Released
High
linux (Ubuntu)
Fix Released
High
Andy Whitcroft

Bug Description

In only 1 of 5 boots did the display initialize correctly. It appears to make no difference if splash or quiet are removed from the grub command line.

ProblemType: Bug
DistroRelease: Ubuntu 11.04
Package: linux-image-2.6.37-7-generic 2.6.37-7.19
Regression: Yes
Reproducible: Yes
ProcVersionSignature: Ubuntu 2.6.37-7.19-generic 2.6.37-rc3
Uname: Linux 2.6.37-7-generic x86_64
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.23.
AplayDevices: aplay: device_list:235: no soundcards found...
Architecture: amd64
ArecordDevices: arecord: device_list:235: no soundcards found...
AudioDevicesInUse: Error: command ['fuser', '-v', '/dev/snd/by-path', '/dev/snd/controlC0', '/dev/snd/hwC0D0', '/dev/snd/pcmC0D0c', '/dev/snd/pcmC0D0p', '/dev/snd/seq', '/dev/snd/timer'] failed with exit code 1:
CRDA: Error: [Errno 2] No such file or directory
Date: Wed Dec 1 10:47:52 2010
EcryptfsInUse: Yes
HibernationDevice: RESUME=UUID=bd417fbf-4c6d-4540-9cca-e2c46b7fc886
InstallationMedia: Ubuntu 11.04 "Natty Narwhal" - Alpha amd64 (20101201)
MachineType: Hewlett-Packard HP Mini 5102
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.37-7-generic root=UUID=62160825-a474-44c9-a60a-d4bb4f8b59e5 ro drm.debug=0x4
RelatedPackageVersions: linux-firmware 1.40
SourcePackage: linux
StagingDrivers: crystalhd
Title: [STAGING]
dmi.bios.date: 03/11/2010
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: 68PGU Ver. F.05
dmi.board.name: 3653
dmi.board.vendor: Hewlett-Packard
dmi.board.version: KBC Version 43.1D
dmi.chassis.asset.tag: CNU0150MDM
dmi.chassis.type: 10
dmi.chassis.vendor: Hewlett-Packard
dmi.modalias: dmi:bvnHewlett-Packard:bvr68PGUVer.F.05:bd03/11/2010:svnHewlett-Packard:pnHPMini5102:pvr:rvnHewlett-Packard:rn3653:rvrKBCVersion43.1D:cvnHewlett-Packard:ct10:cvr:
dmi.product.name: HP Mini 5102
dmi.sys.vendor: Hewlett-Packard

Revision history for this message
Tim Gardner (timg-tpi) wrote :
Revision history for this message
Tim Gardner (timg-tpi) wrote :

Note that the attached BootDmesg.txt contains the result of booting with drm.debug=0x4

Revision history for this message
Tim Gardner (timg-tpi) wrote :

This appears to be a regression from 2.6.37-6.17 to 2.6.37-7.19. I'm suspecting these commits likely have something to do with it:

UBUNTU: [Config] build in CONFIG_AGP_VIA
UBUNTU: [Config] build in CONFIG_AGP_NVIDIA
UBUNTU: [Config] build in CONFIG_AGP_AMD64
UBUNTU: [Config] build in CONFIG_AGP_AMD
UBUNTU: [Config] build in CONFIG_AGP_INTEL
UBUNTU: [Config] build in CONFIG_AGP

Revision history for this message
Tim Gardner (timg-tpi) wrote :

Well, that wasn't it. Rebasing against 2.6.37-rc4 and adding Dave Airlie's pull request from Tue, Nov 30, 2010 yielded similar results, i.e., zilch.

Revision history for this message
Tim Gardner (timg-tpi) wrote :

Hmm, 2.6.37-6.17 has the same issue. It just doesn't happen as often.

Revision history for this message
Tim Gardner (timg-tpi) wrote :

With 2.6.37-7.19 the display will correctly initialize about 1 in 5 times from a cold boot, and never AFAICT from a warm boot. However, the good news is that it recovers the display 100% of the time after a suspend/resume cycle. I've experimented with unplugging the battery and/or booting with power plugged/unplugged with no discernible effect.

Revision history for this message
Neil Perry (nperry) wrote :

Effecting a couple of people in the ubuntu forums to, marking as confirmed.

Changed in linux (Ubuntu):
status: New → Confirmed
tags: added: iso-testing
Revision history for this message
Andy Whitcroft (apw) wrote :

As Tim indicates there is an element of a race to this issue. On my test system and upgraded (Maverick -> Natty) system the issue was not reproducible in 20 reboots. I scratch installed from the same Natty image and the issue was instantly reproducible. In testing removing the ureadahead pack seems to render the machine bootable for the next boot. All of this tends to indicate that this is indeed a boot race.

It is reminicent of the framebuffer open races we experienced in Maverick. Cirtainly in boots where things work there is no sign of the vesafb being used, whereas it appears to be painted by plymouth in most of the failed boots.

Changed in linux (Ubuntu):
assignee: nobody → Andy Whitcroft (apw)
importance: Undecided → High
status: Confirmed → In Progress
tags: added: regression-release
removed: regression-update
Revision history for this message
Andy Whitcroft (apw) wrote :

Ok this seems to be fallout from grub2 handing off the display in graphics mode (gfxpayload=keep). Switching this back to text resolves the issue.

Revision history for this message
Andy Whitcroft (apw) wrote :

As a work-around we can turn this off by adding the line below to /etc/default/grub and running update-grub:

    GRUB_GFXPAYLOAD_LINUX=text

Revision history for this message
Neil Perry (nperry) wrote :

Workround worked on a warm boot.

Revision history for this message
cariboo (cariboo) wrote :

To me it seems to be more of a detection issue. I have an extra monitor cable sitting on my workbench, as soon as I connect it to my netbook and switch the kvm to the position the monitor cable is connected to I get the login screen displayed. I then disconnect the monitor cable and everything works as it should, as long as I don't reboot.

tags: added: kernel-key-gfxpayload
Revision history for this message
Andy Whitcroft (apw) wrote :

Upstream believes they understand the issue, the pipes are running when we try and reconfigure and our current initialisation can lead to combinations which would be deemed to produce 'undefined behaviour'. Testing with the V2 patch from the upstream bug seems to indicate that the fix there is sufficient for this machine.

Revision history for this message
Andy Whitcroft (apw) wrote :

Further tested this on a number of intel graphics based systems without issue. If those of you affected by this issue could test the kernels below and report back that would be great:

    http://people.canonical.com/~apw/blank-debug2-natty/

Changed in linux (Ubuntu):
status: In Progress → Fix Committed
Revision history for this message
Neil Perry (nperry) wrote :

Andy, thanks for your work. Could you supply i386 image please.

As I'm effected by this bug on my i386 netbook.

Thanks

Revision history for this message
Neil Perry (nperry) wrote :

Ignore last comment, maybe it hadn't completed uploading!

Revision history for this message
Neil Perry (nperry) wrote :

Andy,

20/20 warm reboots, where before its was 3/20.

Thanks for your work.

Neil

Revision history for this message
Daniel Kessel (dkessel) wrote :

Also works for me with a 945GM chipset.

Thanks a lot Andy.

Revision history for this message
Neil Perry (nperry) wrote :

Not working on cold boot, just taken me 15 attempts to boot.

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package linux - 2.6.37-8.20

---------------
linux (2.6.37-8.20) natty; urgency=low

  [ Andy Whitcroft ]

  * Revert "[Upstream] drivers/serial/mfd.c: Fix ARM compile error"
  * Revert "SAUCE: Nouveau: Disable acceleration on MacBook Pros"
  * Revert "SAUCE: Nouveau: Add quirk framework to disable acceleration"
  * Revert "SAUCE: i915 -- disable powersave by default"
  * SAUCE: enable Marvell 9128 PCIe SATA controller
    - LP: #658521
  * [Config] evtchn has been renamed
  * (no-up) add support for installed header files to ubuntu directory
    - LP: #684666
  * ubuntu: AUFS -- include the aufs_types.h file in linux-libc-headers
    - LP: #684666

  [ Tim Gardner ]

  * [Config] MISS: evtchn, NEW : xen-evtchn
  * rebase to v2.6.37-rc4

  [ Upstream Kernel Changes ]

  * drm/i915: Clean conflicting modesetting registers upon init
    - LP: #683775
  * rebase to v2.6.37-rc4
 -- Andy Whitcroft <email address hidden> Fri, 03 Dec 2010 18:42:07 +0000

Changed in linux (Ubuntu):
status: Fix Committed → Fix Released
Revision history for this message
Neil Perry (nperry) wrote :

Hasn't fixed the issue for me, Am able to boot up warm however cold boots seems to take about 10-15 attempts.

Not to sure if this is another issue, however when I hold shift to boot into grub menu, 9/10 it just goes to Grub loading and goes no further.

Revision history for this message
Rick McBride (rmcbride) wrote :

the provided test kernels correct the behavior on a lenovo S10

Thanks for your work!

Revision history for this message
Anton Gladky (gladky-anton) wrote :

Hi,

2.6.37-8.20 should appear in main repository as normal update? Or I need to use "proposed" branch?

Thanks you

Revision history for this message
PeterPall (peterpall) wrote :

In my experience some package mirrors need a few days to update all packets. http://gb.archive.ubuntu.com started to include the new kernel yesterday. Since then every time the kernel was loaded 12 seconds later I got the login screen as I am used to. And this with a slow intel atom processor. Thanks a lot!

Revision history for this message
Anton Gladky (gladky-anton) wrote :

Ok, thanks for information. I have finally got the new kernel also.
But, it seems, I have still black screen after boot. Dell Mini 10v.

Revision history for this message
Tim Gardner (timg-tpi) wrote :

Fixed for me.

Revision history for this message
Anton Gladky (gladky-anton) wrote :

I still can not load Alpha 1 noramlly. With new kernel splashscreen appears, but then .... black screen.

Revision history for this message
PeterPall (peterpall) wrote :

On my system since the kernel update if I get a black screen it is prior to loading the kernel. Filed Bug #685445 for this.

Revision history for this message
Anton Gladky (gladky-anton) wrote :

The problem is also resolved for me. I reinstalled the system. Perharps, Karmic->Lucid->Maverick->Natty update was too much.

Revision history for this message
Jared Klingenberger (itsjareds) wrote :

Is the fix released for the CD/USB image? My screen goes blank after I choose "Try without installing," and I have updated to today's daily image.

Revision history for this message
cedric (cedric) wrote :

workaround fixed it for me on a dell latitude E4310

Revision history for this message
Barry Drake (b-drake) wrote :

I think this is probably the same bug? After clean install from the iso - 64bit desktop downloaded yesterday, it booted into the Unity desktop which seemed to run very slowly and finally stopped responding. On re-boot, the panels had disappeared, and I couldn't make anything happen. I got the startup sound, and the background with an active mouse cursor, but that was all.

The only way I got into it again was to boot in recovery mode, and switch to low definition. Once in, I changed the boot to the gnome desktop and it works just fine. I tried putting a root terminal and Nautilus on the desktop and switching back to Unity. The icons failed to appear, so I couldn't find a way of looking around.

I have done the full update before testing again.

Hardware: AMD ATHLON II X4 640 (3.00GHz/2MB CACHE/AM3), Motherboard ASUS® M4A78LT-M: mATX MAINBOARD, DDR3, USB 2.0, SATA 3.0Gb/s RAM 2GB. Installed on hdb 500 GiB (my hda has 10.04 installed).

Revision history for this message
PeterPall (peterpall) wrote :

Don't like unity. I rather use compiz. Or even the classical gnome session with metacity that is more fit for a user that uses more than a handful applications regularily.

...anyway... If you can see anything on the screen (and you told you did at least get a mouse cursor) you have most likely found a separate bug. On my computer I got a perfectly black screen when I turned on my computer - and the screen stayed perfectly black all the time.

---

Another question: Seems like the fix for this bug has been released now. Is anybody in here against closing it?

Changed in linux:
status: Unknown → Fix Released
Changed in linux:
importance: Unknown → High
Revision history for this message
jhfhlkjlj (fdsuufijjejejejej-deactivatedaccount) wrote :

This issue was fixed for me for a while but it has come back. I'm using daily ISOs. Has anyone else had this problem crop back up?

Revision history for this message
Alejandro Zunino (azunino) wrote :

I have an ACER ASPIRE 4736Z and Ubuntu Natty boots to a blank screen. I did not have problems with Ubuntu Maverick. I have tried lots of boot parameters with no success. Sometimes Natty boots ok... it seems totally random.

Revision history for this message
Jānis Kangarooo (kangarooo) wrote :

I was tryng 11.04 alfa and beta 2months ago and month ago with Multisystem multiple isos. Older versions are fine 11.04 ubuntu and Xubuntu gave black screen. On cold shutdown for 1/5th second i saw screen before power went away. Problem is monitor not showing anything. Its not even making Black Color its just low energy in Laptop monitor.
Ah yes this is like that on Emachines e525 but on Dell D610 with CdbootingUsb all works.
Also i was able to start on Emachines e525 with external monitor and External Works but internal Not.
Its been like that for at least 2months in 11.04 and still on finalized version still exist downloaded Ubuntu and Xubuntu today.

Revision history for this message
Rajesh Verma (verma-r-kumar) wrote :

Having similar Black screen issue when try boot with ubuntu after installation completion of 11.04 version on HP Pavilion dm4 CTO-1100 laptop. But 10.04 works fine. Is there possiblity that Ubuntu 11.04 64-bit tries to install restricted graphics driver (AMD ATI Radeon HD 5xxx graphics driver), cause the same think happens if I try to install the restricted driver.

Kindly let me know if there is any fix available.

Revision history for this message
Arbiel Perlacremaz (arbiel-perlacremaz-g) wrote :

The bug seems to me not to have been fixed. It takes me up to 6 to 8 tries before any successfull cold boot.

I made the modification suggested by Andy in comment #10. Maybe 1 success for every 8 to 10 tries.

I don't know whether it still makes sense to test the kernel linked at in comment #14. If it makes sense, can somebody explain me how to do ?

Revision history for this message
Arbiel Perlacremaz (arbiel-perlacremaz-g) wrote :

I have 3 distros : 1 Natty and 1 Maverick on my internal disk and 1 Natty on an external Usb disk, all 64 bits. Andy's modification helped to boot both Natty, however with better results for my internal Natty (roughly 50% success for warm boots, and 25% for cold ones) than for my external one (maybe 20% for warm boots, and 10% for cold one). Each Natyy has its own swap.

A few days ago, I had to boot Maverick (no problem booting), and this then both Natties refuse to boot (I may have since run up to 20 tries on each one).

Internal Natty's kernel : Linux 2.6.38-11-generic
USB Natty's kernel : Linux 2.6.38-8-generic
Maverik's kernel : Linux 2.6.35-22-generic

The situation is getting very frustrating, as my Maverik is not complete : does not run WIfi and has no sound.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.