Preinstalled Maverick netbook image won't boot on Gumstix Overo

Bug #630825 reported by Vincent Stehlé
10
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Ubuntu CD Images
Invalid
Undecided
Unassigned
linux (Ubuntu)
Expired
Undecided
Unassigned

Bug Description

I tried booting http://cdimage.ubuntu.com/ubuntu-netbook/ports/releases/maverick/beta/ubuntu-netbook-10.10-beta-preinstalled-netbook-armel+omap.img.gz on an Overo earth board but the kernel complains: "Wake up daisy chain activation failed."

This looks similar to bug #594382.

I attached the UART log captured.

I replaced the pre-installed image bootloaders by MLO and u-boot.bin compiled for Overo (from gitorious repositories).

The u-boot command lines used were:

setenv bootargs vram=12M omapfb.mode=dvi:1280x720MR-16@60 root=/dev/mmcblk0p2 fixrtc console=ttyS2,115200n8
mmc init; fatload mmc 0 0x80000000 uImage; fatload mmc 0 0x81600000 uInitrd; bootm 0x80000000 0x81600000

Revision history for this message
Vincent Stehlé (vstehle) wrote :
Revision history for this message
Vincent Stehlé (vstehle) wrote :

(Add MLO.)

Revision history for this message
Vincent Stehlé (vstehle) wrote :

(Add u-boot.)

Revision history for this message
Steve Langasek (vorlon) wrote :

Thank you for taking the time to report this bug and help to improve Ubuntu.

It looks like u-boot-omap3 is the package providing the uboot included on this image. Reassigning to that package.

Changed in ubuntu-cdimage:
status: New → Invalid
Revision history for this message
Vincent Stehlé (vstehle) wrote :

The "Wake up daisy chain activation failed." messages look like kernel messages to me.
Are you sure u-boot-omap3 is the right package?

Revision history for this message
Steve Langasek (vorlon) wrote :

You're right; sorry, I read too quickly and misunderstood what the issue was, I thought you were saying that you had replaced the uboot and MLO and that this /fixed/ the problem. Sounds like this is the kernel, indeed.

affects: u-boot-omap3 (Ubuntu) → linux (Ubuntu)
Revision history for this message
Oliver Grawert (ogra) wrote :

looks like its similar to Bug 594382 ...
which was worked around by disabling CONFIG_CPU_IDLE for beagle a while ago, seems this workaround does not help with gumstix

tags: added: kj-triage
Revision history for this message
Vincent Stehlé (vstehle) wrote :
Revision history for this message
Vincent Stehlé (vstehle) wrote :

FWIW: Recompiling the kernel with CONFIG_PM disabled makes the issue disappear.

Revision history for this message
Vincent Stehlé (vstehle) wrote :

I did a 'git bisect' between Ubuntu-2.6.34-3.10 (good) and Ubuntu-2.6.35-1.1 (bad):

40742fa82ebf53c9367f5807cf1012d518516e74 is first bad commit
commit 40742fa82ebf53c9367f5807cf1012d518516e74
Author: Mike Chan <email address hidden>
Date: Mon May 3 16:04:06 2010 -0700

    OMAP3: PM: Enable IO / IO-CHAIN wakeups for PER

    IO events can also come from GPIO modules, which reside in the PER domain.
    It is possible for the PER to enter RET while CORE is still in ON.
    If GPIO 2-6 are enabled for IO-pad wakeups, the PER domain will not
    wakeup in this case, unless we enable it.

    Signed-off-by: Mike Chan <email address hidden>
    Signed-off-by: Kevin Hilman <email address hidden>

:040000 040000 270b6d5d07c566d7ff1fe6a2d22f7eab3dbf7159 1863e212b840b133e37691c40d3fddabf0f0444e M arch

Revision history for this message
Vincent Stehlé (vstehle) wrote :

This patch can be applied on top of Ubuntu-2.6.35-19.28 to revert 40742fa82ebf53c9367f5807cf1012d518516e74 and see that the "daisy chain" issue goes away.

(Note that it is probably not a definitive solution, just a debug help.)

tags: added: patch
Revision history for this message
Steve Langasek (vorlon) wrote :

Vincent,

Could I ask you to check whether this bug is also present in the linux-image-2.6.35-1005-linaro-omap kernel package?

Revision history for this message
Vincent Stehlé (vstehle) wrote :

Yes, I am afraid the bug is also present for linux-image-2.6.35-1005-linaro-omap_2.6.35-1005.10_armel.deb (see the UART log).

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

@ogra -- could you confirm whether we would expect this image to work unaltered on this machine? I am under the impression that these images are actually Beagle oriented images?

Revision history for this message
Oliver Grawert (ogra) wrote :

right, the images are focused on beagle wrt bootloader and kernel, if there is a fix for the kernel thats going upstream soon that makes it easier using the images with less modification it would be a nice to have. nontheless our focus is on beagle XM here and you will definitely have to alter the bootloader even if the kernel works.

Revision history for this message
Ricardo Salveti (rsalveti) wrote :

CONFIG_MACH_OVERO is set as 'y' at our current omap 3 kernel, so it should work on Overo if the user put the correct x-loader and u-boot at the first disk partition.

Vincent (or any other with an overo board in hands), can you test Natty's image to see if you can still reproduce this issue? x-loader-omap3-overo and u-boot-linaro-omap3-overo should have the needed files to boot the image at Overo.

Revision history for this message
Loïc Minier (lool) wrote :

Ricardo, could you update the package tasks to mention the kernel source package used in the problematic images?

On my side, I've tested Overo with linux-linaro-omap last week and it mostly worked (some issues here and there).

Revision history for this message
Tobin Davis (gruemaster) wrote :

Is this still an issue with the 11.04 release image? If so, please update the bug. Otherwise let it close or see about getting an SRU patch for maverick.

Changed in linux (Ubuntu):
status: New → Incomplete
Revision history for this message
Launchpad Janitor (janitor) wrote :

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

Changed in linux (Ubuntu):
status: Incomplete → Expired
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.