Bug#690814: [squeeze->wheezy regression] disk activity provokes lockups on VIA EPIA CL-6000

[Jonathan Nieder <jrnieder@gmail.com>]

Frank Lenaerts wrote:

> Tested some of the linux images:
>
> - 2.6.39 seems ok
> - 3.0.0-1 seems ok
> - 3.1.0-1 seems ok
> - 3.2.0-1 does not seem to be ok:
>   - first boot: system hung at "Loading, please wait..."; note that
>     the LED indicating disk activity was not burning
>   - second boot: ok
>   - third boot: ok
>
> Had it running and used it for a few days. Then decided to give
> 3.2.0-3 a try again.
>
> - 3.2.0-3 is not ok:

Thanks much for this.

[...]
> - first boot: crash during the boot process i.e. detected the
>   external USB disk (which I had connected again while working with
>   3.2.0-1), makefile style concurrent boot, hotplug, udev...
>   stacktrace (not sure if it was after or before the hotplug stuff;
>   unfortunately didn't have logging turned on and keyboard was
>   stuck)...

For the future, photographs work fine if you can catch the machine at
the right moment.

[...]
> Question: 3.2.0-3 was the one installed during the installation.
> 3.2.0-1 was installed by me for testing purposes and came from
> 3.2.1-1. I see that the snapshot directory contains other 3.2.0-1
> e.g. under 3.2.2-1. Why is it like that?

Package names like linux-image-3.2.0-1-486 describe the kernel's
ABI, not the package version.  The package version is something like
3.2.1-1.  See <http://kernel-handbook.alioth.debian.org/ch-versions.html>
for more details.

You can get the version of the currently running kernel by running
"cat /proc/version" (it will be in parentheses).  The currently
installed kernel's version number can be retrieved with
"dpkg-query -W linux-image-$(uname -r)".

Which versions were the 3.2.0-1, 3.2.0-2, and 3.2.0-3 kernels
you mentioned testing above?

Hope that helps,
Jonathan