> -----Original Message-----
> From: Jamie Lokier
> [mailto:lkd@tantalophile.demon.co.uk]
> Sent: Tuesday, January 18, 2000 3:18 PM
> To: Miles Lane
> Cc: Linux Kernel
> Subject: Re: 2.3.40-5 -- "apm" no longer recognizes
> that I have an
> APM-enabled kernel.
>
>
> Miles Lane wrote:
>> When I run the command "apm", I receive an error
>> stating:
>> No APM support in kernel
>> This in spite of having APM enabled in my .config
>> file:
>
> That's because you've got ACPI enabled too. ACPI
> now overrides APM if it is compiled in.
>
> I don't know why it does that -- APM seems to work
> fine with ACPI in earlier kernels, and now suspend
> takes longer and the clock isn't restored properly
> after a suspend.
APM is disabled by ACPI as both can not be enabled
at once. On some platforms you can get away with
APM engaged (the most common reason is to compensate
for a non-working S5 implementation and fallback to
APM power off) but it's not safe to do. If you want
to use APM, turn off or disable (kernel command
line "acpi=off") ACPI.
> The solution: turn off ACPI. I can't see what the
> point in using the ACPI support is at the moment.
On systems with working ACPI implementations,
ACPI gets you CPU power states (saves power
when CPU is idle), S5 (soft off), and power/sleep
button support. APM additionally provides suspend
and battery status and these features are
currently incomplete for ACPI but on the way.
> If the ACPI user space tools (e.g. acpid) do
> anything useful, it isn't documented.
acpid reads the ACPI tables and uploads important
information about how to do CPU power states,
S-states, and handles power/sleep button presses.
If acpid does not run, the ACPI driver won't
do much useful.
> For me the only effect of acpid is to prevent
> suspends from suspending at all. Nothing useful
> like battery life estimation, clock updating,
> that sort of thing.
Battery support is coming. It's a bit more
complicated to do than APM, hence the delay.
You have to have a completely working AML
interpreter and, for many systems, an
embedded controller driver.
> A lot of work is obviously going into the kernel
> ACPI driver. Of course I'd like to minimise power
> consumption on my laptop. So I'm wondering, how
> is that kernel driver getting tested? Is there a
> user space tool I haven't found?
Most of the problems we are seeing are just
discovered by people running with ACPI on
various systems. There is really no better
way to test things like C-states, S-states,
and buttons other than by hand.
-Andy
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Jan 23 2000 - 21:00:20 EST