Re: CVS commit: src/sys/dev/acpi

[Jukka Ruohonen <jruohonen%iki.fi@localhost>]

On Sun, Feb 28, 2010 at 09:16:38PM +1100, matthew green wrote:
> i'm curious why ignoring the various errors is a good thing.  in
> particular, sysmon_pswitch_register() and pmf_device_register()
> failure mean the device won't work properly, yet there is nothing
> to indicate such...

Well, generally I've tried to reduce the verbosity in the acpi(4) subtree,
as it is quite chatty compared to many other subsystems. The acpi(4) device
drivers are also believed to be robust, having been in production since
NetBSD 1.6.

The sysmon_pswitch_register() is currently a dummy function. If
pmf_device_register() fails, the failure should noticeable without an error
message in the event of e.g. suspend... As for things like AcpiOsExecute(),
the failures typically involve invalid parameters (can be verified not to be
the case by manual code inspection) or lack of memory (in which case there
sure are bigger problems than ACPI drivers).

- Jukka.