Power Management

CONFIG_PM
CONFIG_APM
CONFIG_APM_DO_ENABLE
CONFIG_APM_DISPLAY_BLANK
CONFIG_APM_CPU_IDLE
CONFIG_TOSHIBA

The above options are what I use, and basic APM support works (in conjunction with apmd, that is). However, the Toshiba Satellite Pro 6000 has some extra powermanagement features, which are an extension to the basic APM spec. To support these features, you'll want to activate CONFIG_TOSHIBA and download the tools at http://www.buzzard.org.uk/toshiba/; they are also available in the Debian package ``toshutils''. They work pretty well, and make some features supported by Windows available under Linux (and even some that are, AFAICS, *not* supported by Windows, like the alarm power on. Correct me if I'm wrong).

Alternatively, there's the toshset stuff. I have no more information about this than the toshset -q output near the end of this document (thanks go to J. Waalboer mailto:jwaalboer@convolution.nl for that one) and the name of the software. Ask Mr. Waalboer for details.

ACPI support doesn't work; don't try. The kernel mumbles something about ``buggy BIOS'' and disables ACPI. Besides, current Linux ACPI support is incomplete; APM works much better. You're not losing anything.

That doesn't mean it all works flawlessly. Toshiba engineers seem to like putting all possible and impossible hardware so that it uses IRQ 11. My system tends to lock up when suspending, unless I give the extra bootparameter ``pci=biosirq'' (as the kernel suggests if you boot without it). Has some impact on sound, though; it sometimes gives lags, and sometimes doesn't. If it does, you just need to reload the module.

One note: if you suspend your system, you need to reload the usb-ohci module when resuming, else the bus will no longer report when something connects or disconnects. You can do this automatically by creating two shell-scripts. This one comes from my /etc/apm/suspend.d:

#!/bin/sh
/etc/init.d/usbmgr stop
rmmod usb-ohci

and this one from /etc/apm/resume.d:

#!/bin/sh
modprobe usb-ohci
/etc/init.d/usbmgr start

the stopping and starting of usbmgr is because if I don't do that, usb-ohci is still in use, and thus can't be removed. It's not strictly necessary, but then again...

Clever people might notice that I'm not reloading the trident module to fix the sound problem. This is because the sound device may be in use, which means that the module won't be unloaded, which probably means that the script errors out, and thus that the machine doesn't suspend. I don't like that ``feature'' ;-)