Re: Dell latitude C810
>
> alexis.guillard@bt.com said:
> > I can not completely shutdown the laptop with the 'halt' command. I
> > put CONFIG_APM=y CONFIG_APM_REAL_MODE_POWER_OFF=y in my kernel.
>
> > I try to test ACPI but it crash the laptop when I plug or unplug the
> > AC power.
>
> I remember someone earlier on this list or on debian-testing mentioning a
> conflict between ACPI and APM.
>
> - Josef
Not so much a conflict, as a replacement of. (or you can think of it
as "conflicts" the same way our package system does, only one works at
a time.)
APM is (to catch anybody up who needs it) power management. It has a very
limited amount of states. Basically, it can yank a few chains, but suspend
is all or nothing, and you can go to RAM, or to disk.
ACPI is a newer standard. In it the subsystems can be given more control
commands.
In the Linux kernel if you have ACPI compiled in, and an ACPI capable chipset
then that's it, you get ACPI, and APM will not be used.
Which means, that your problem is not *directly* related to them conflicting.
What is much more likely is..
On modern systems APM and ACPI are both supported, but it's a crap shoot which
behaves better. Some manufacturers really only test with MSwin, which has
put a lot of effort into ACPI support, and their APM is not much to write home
about, fidgety. Others have done a best effort to render APM per the industry
standards and have just given a shot at ACPI. Or so it seems. At the core
of it, as all the parts get faster timing gets to be a tricky thing, and I
think there have always been few systems which could ever suspend safely
without power-management support active on the OS.
So... you'll need to test which behaves better for you, amd since Linux' ACPI
support is steadily improving, you'll want to check back on it every once in
a while if you went the APM route.
In either case, they won't work with their userland support tools. I prefer
to pick one and build my kernel with only that one in it, so there's no
questions about it. But then, I tend to think that since every laptop is
a mostly-unique combo of parts and most of the parts aren't replaceable,
that most laptops should get a custom kernel when their owner has the time.
* Heather Stern * star@ many places...
Reply to: