Bug#655944: [2.6.37 -> 2.6.38 regression] gpe18 rising many times per second

[Jonathan Nieder <jrnieder@gmail.com>]

Hi,

James Tocknell wrote:

> This wasn't a regression for me, but I only installed Debian on the
> laptop at the beginning of December. I've tried both the kernel from
> experimental, and the latest one from unstable (3.2), there was no
> change.

Please report this upstream to <http://bugzilla.kernel.org/>, product
ACPI, component Config-Interrupts, and let us know the bug number so
we can track it.  Be sure to mention:

 - steps to reproduce, expected result, actual result, and how the
   difference indicates a bug (should be simple here)

 - which kernel versions you have tested and what happened with each

 - full "dmesg" output from booting an affected and an unaffected
   kernel, as (separate, uncompressed) attachments

 - acpidump output, as another attachment

 - /proc/interrupts

 - a link to http://bugs.debian.org/655944 for the backstory

 - what kind of testing you'd be able to perform if needed (can you
   test debugging patches on this machine if provided?  can you
   bisect?)

 - any workarounds or other weird observations

[...]
>                                               the 2.6.38 kernel had
> slightly higher cpu usage but a high number of interrupts (though
> less than the later kernels), and the 3.* kernels were fairly similar
> in their cpu and interrupts. So this may be a few different
> regressions then.

I'd suggest starting with one for simplicity, and separately solving
any issues that remain once it is addressed.

Thanks very much for your work so far on this.

Sincerely,
Jonathan