Power Management broken in 177.67

[tier]

uhci_hcd has always been a problem when suspending.. unload it and try again. I think i might have it loaded, also, I'm gonna test again when my work is done.

[kcrudup]

Originally Posted by tier

uhci_hcd has always been a problem when suspending

Maybe a while ago, but on Intel chipsets I haven't had to make neither UHCI nor EHCI be modules in two years. (My kernels are 99.9% module-free).

Later kernels also have USB device-side suspend options that need to be turned on, too.

[tier]

Originally Posted by kcrudup

Maybe a while ago, but on Intel chipsets I haven't had to make neither UHCI nor EHCI be modules in two years. (My kernels are 99.9% module-free).

Later kernels also have USB device-side suspend options that need to be turned on, too.

I built monolithic kernels quite often, too. But now I am using the complete modular arch linux kernel, and even without uhci_hcd (which made problem until 2.6.25) my thinkpad wont wake up properly. I checked the logfiles and there's nothing.. they just end without telling anything.

[kcrudup]

Originally Posted by tier

uhci_hcd has always been a problem when suspending

I just realized that you were replying to my post. Nah, the reason I listed the IRQ16 entry is 'cause when I'm not running the NVidia driver with NVreg_EnableMSI=1, the VW IRQ is 16 (which is only co-incidentally shared on my laptop with uhci_hcd):

 16:      22083          0   IO-APIC-fasteoi   uhci_hcd:usb3, nvidia

The "nobody cared" from the kernel meant that no handler returned IRQ_HANDLED when it fired ('cause the NV driver was listening for IRQ 313).

[gfdsa]

Hi guys,

Anybody here with a mobile 8xxx and has backlight control with beta driver ( that mysterious "improved power management" in changelog make me curious ) ?

Thanks!