The patch "linux-2.6-acpi-fix-sizeof.patch" breaks acpi-cpufreq on my laptop. It has been backported to fedoras 2.6.24 kernels (now removed again due to this bug). I was asked to report it here.
See: https://bugzilla.redhat.com/show_bug.cgi?id=436959#c17 (starting from this comment).
I will attach some debug infos after submitting this.
Created attachment 15377 [details]
dmesg output with cpufreq.debug=7
Created attachment 15378 [details]
Created attachment 15379 [details]
Created attachment 15380 [details]
Created attachment 15381 [details]
By break acpi-cpufreq I mean it fails to load with "no such device" and the cpu stays at max clock. I hope the infos provided are enough if not feel free to ask for more.
Created attachment 15411 [details]
Please try this patch
Mark buffer data valid when copying external buffer object to internal
(In reply to comment #7)
> Created an attachment (id=15411) [details]
> proposed patch
> Please try this patch
> Mark buffer data valid when copying external buffer object to internal
This one fixes the regression. Thx!
what is "linux-2.6-acpi-fix-sizeof.patch"?
it's the patch at Bug 9558 #33
The offending patch from bug 9558 was not included upstream.
Unclear why it was included in fedora before it hit
upstream or even the -mm tree, maybe that was a mistake.
So this isn't an upstream regression, but rather it is feedback
on the patch proposed for bug 9558.
Closing as a duplicate of bug 9558.
*** This bug has been marked as a duplicate of bug 9558 ***