Hi , Len 
Thanks for replying
So:

1 - ACPI w/o CONFIG_PM (is not supported)

2 - If accidentally we configure ACPI w/o CONFIG_PM. I tested and
doesn't power off correctly (at least on one Dell Precision 410). 

3 - It is easy make this error in configuration, especial if we try use
.config from early versions.

4 - The patch doesn't hurt at all, just force to one correct
configuration.

On Fri, 2004-05-07 at 05:18, Len Brown wrote:
> Never occurred to me to build ACPI w/o CONFIG_PM...
> There are #ifdef CONFIG_PM in the acpi code, so I guess this was on
> purpose, but it makes ACPI a lot less interesting.
> 
> But I'm inclined to leave 2.4 alone except for real system failures. 
> The only clean-up I'm really interested in doing in 2.4 is when it makes
> maintenance via backporting from 2.6 easier.

I understand very well your point, and I don't check in kernel 2.6.5
Config.in, neither check in others architectures.
But I vote for apply this patch, because can avoid problems with
power-off machines.

Off-topic: I am testing last acpi-2.4.27 patch and it ok, no complains.

I resend the patch just in case.
thanks,
-- 
Sérgio M. B.