[oe-commits] [Bug 5435] machine htcleo: gcc version etc.

[bugzilla-daemon at git.openembedded.org]

http://bugs.openembedded.org/show_bug.cgi?id=5435





--- Comment #7 from Lukas-David Gorris <lukas at htc-linux.org>  2010-05-22 02:06:00 ---
Hi,
thanks for the quick replies

(comment #3)
> as it seems that udev is causing the problem I would recommend try different
> udev revs out there
> 
> add this to local.conf
> PREFERRED_VERSION_udev_local = "151"
> 
> and then
> 
> bitbake -c clean udev
> bitbake console-image
> 
> that should regenerate the image with udev-151
> 
> likewise keep prodding on different udev versions.
> if nothing works then it has to be debugged. 
> 
> Thanks
> 
> -Khem

tried this along with required usbutils version

PREFERRED_VERSION_udev_local = "151"
PREFERRED_VERSION_usbutils_local      = "0.86"

didn't change situation. tried no further versions.


(Mit Bezug zu comment #4)
> Alternatively, just don't use udev. If you have a recent kernel, devtmpfs works
> like a charm. It's a relief not having to rely on udev.

To try this I did the following:

*Enable devtmpfs in kernel:
+CONFIG_DEVTMPFS=y
+CONFIG_DEVTMPFS_MOUNT=y

*Remove udev from init (ugly?):
rm recipes/udev/udev-151/init

*cleanm & rbeuild kernel-qsd8, udev, console-image 

==>
devtmpfs is being loaded
root mounted
init ran
root remounted
init continiously spits 'segmentation violation, sleeping 30secs' topped with
occasional illegal instruction error. ?!
then I also see things moving on: there was some opkg thing. but it will then
freeze and init will keep spamming the violtion line.
no console is spawned.

things look different when a rootdelay=13 is added in cmdline..
then no errors are shown and it will just attempting to kill init panic
directly after rootfs is remounted line.
I will append dmesg  

no I am worried again if the illegal instruction thing is generic (low level)
problem :(

regards
lukas

(comment #6)
> Lukas
> 
> hmm on second (workaround) thought you could try to disable devtmpfs
> 
> CONFIG_DEVTMPFS is not set
> CONFIG_DEVTMPFS_MOUNT is not set
> 
> in the kernel .config and just rebuild the kernel
> and see if udev is happy

I disabled it and udev still crashes.

-- 
Configure bugmail: http://bugs.openembedded.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.