Messages in this thread | | | Date | Tue, 13 Jan 2009 23:33:00 +0100 | From | Pavel Machek <> | Subject | Re: hp_accel: do not call ACPI from invalid context |
| |
On Tue 2009-01-13 14:11:16, Andrew Morton wrote: > On Mon, 12 Jan 2009 11:31:55 +0100 > Pavel Machek <pavel@suse.cz> wrote: > > > > > LED on HP notebooks is connected through ACPI. That unfortunately > > means that it needs to be delayed by using schedule_work() to avoid > > calling ACPI interpretter in invalid context. This patch fixes that. > > > > Signed-off-by: Pavel Machek <pavel@suse.cz>
> > -static void hpled_set(struct led_classdev *led_cdev, > > - enum led_brightness value) > > -{ > > - hpled_acpi_write(adev.device->handle, !!value); > > -} > > - > > -static struct led_classdev hpled_led = { > > - .name = "hp:red:hddprotection", > > - .default_trigger = "heartbeat", > > Current mainline+lis3lv02d-merge-with-leds-hp-disk.patch has "none" > here, so the patch didn't apply.
Oops, sorry about that. I was hand-editing patches a bit too much :-(.
> > - .brightness_set = hpled_set, > > +static struct delayed_led_classdev hpled_led = { > > + .led_classdev = { > > + .name = "hp::hddprotect", > > + .default_trigger = "none", > > But I assume we wanted "none" here anyway, so that's what I'll do.
Thanks! Pavel
-- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
| |