Andy Green wrote: > Kernel is always "limiting [userspace] ability to [do everything > itself]" and sometimes that's a sign what you want should be in the kernel.
Or that you're doing too much in the kernel, e.g., the part that does what you'd rather not have done :-) Anyway, if the kernel restores the exact state it was in before suspend, which would seem like reasonable enough behaviour to me, user space could still come up with the backlight off if it just made sure to turn it off before requesting the suspend. This may mean that user space needs a suspend/resume coordinator where there is none yet, but you'd also need this if the kernel did nothing. - Werner
