[PATCH] fix mapping of chassitype to system.formfactor

[David Zeuthen]

On Fri, 2006-05-05 at 17:48 +0200, Danny Kukawka wrote:
> this patch fix the mapping of smbios.chassis.type to system.formfactor. I 
> added a device_property_atomic_update_begin()/end() around the mapping and a 
> check if the property is already set to avoid overwrite values e.g. set in 
> the ACPI/APM/PMU code. With this patch I changed/moved set a default value 
> ('unkown') for system.formfactor.

So I'm not sure this is the right fix. The idea is to

 1. Synthesize all events we need
 2. Add the computer object (including calling out helpers like
    dmidecode)
 3. Process all the events we synthesized

and I'm not sure why we need locking. 

Maybe the ACPI/PMU/APM code is setting system.formfactor when it
shouldn't (e.g. in step 1 rather than step 2). Also, what about
synthesizing the events we need *after* we added the computer object?
Any chance you can look into that?

> Also added a check for the value of exit_type to skip mapping if execution of 
> the smbios prober failed.

This is a good idea though.

Cheers,
David