Bugzilla – Bug 134104
Problem with the embedded controller
Last modified: 2005-12-08 18:06:39 UTC
I have a problem with my embedded controller. My system is an Acer Ferrari 3000. The problem is that the events are delayed. For example, if I unplug the adapter, the system doesn't react. I have to replug and unplug again quite a few times then it does. It's the same thing for adjusting the screen brightness (it's a special function key). ec_burst doesn't change anything with the SuSE kernel. However, being the curious that I am, I compiled a vanilla kernel with the latest ACPI patches and when enabling ec_burst, the problem disappears ! And the problem is also inexistant under KUbuntu. Therefore, there's something missing from the SuSE kernel. Maybe someone with more knowledge can pin point it ? Thanx, Georges.
What do you mean "emabling ec_burst"? And does a kernel.org 2.6.14 kernel work properly for you? Or do you require the latest ACPI tree patches too?
I'm gonna install 10.1 alpha3 tonight which comes with 2.6.14. If it doesn't work, I'll compile a vanilla kernel and try it. If it doesn't work, I'll add the latest ACPI patches and this configuration should work (it worked last time I tried). Oh, and by enabling ec_burst, I mean I'm passing ec_burst=1 as a kernel parameter at boot time. As far as I know, the SuSE kernels have a patch that enable ec_burst by default. I'll get back with the results soon. Georges.
ok I just finished installing 10.1 alpha3 and there's improvement. When I'm plugged in, everything works perfectly (no delays). I can adjust the brightness and turn off the backlight no problem. But, when I unplug and go on battery, the delays comes back. It's like if the events are getting buffered and executed 4 or 5 events later. Any idea why ? Georges.
Thomas, you did a lot of work and testing regarding ec_burst, any idea?
I just wanted to let you know that I opened a bug report in the kernel bugzilla : http://bugzilla.kernel.org/show_bug.cgi?id=5633 Thanks, Georges.
I just compiled a vanilla 2.6.14.2 with ec_burst=1 and it works flawlessly. Events are executed instantly whether my laptop is plugged in on running off the battery. So I suppose there's a kernel patch in the SuSE kernel that's screwing up something. Any idea what that might be ?
I hope I find a smaller patch to add to 10.0 soon ... *** This bug has been marked as a duplicate of 61106 ***