Re: [patch] Device model update (with power state transitions)

[Patrick Mochel]

On Fri, 29 Mar 2002, Pavel Machek wrote:

> Hi!
> 
> > (The first is not necessarily related to the other two, but the other two 
> > were created relative to the first, and it's otherwise innocuous).
> > 
> > 1. Implements notion of 'system' bus, so system level devices can be added 
> > in a comon spot. This includes things like CPUs, PICs, timers, etc.
> 
> Good thing, but naming is pretty inconsistent.
> 
> You have device_register() but register_sys_device(). 

Noted. How about sys_device_register()?

> > Testing welcome also, though I wouldn't expect one to get very far, since 
> > they're not actually used. ;) Which, brings up another question - what 
> > would be the proper place to call device_shutdown()? (I haven't looked 
> > very far into that part...)
> 
> Tested, seems to work.
> 									Pavel
> PS: On toshiba 4030cdt, I can suspend once without no apparent ill
> effects. On resume I get 
> 
> utmisc-0373 Ut_acquire_mutex : Invalid acquire order: Thread 5C owns
> [ACPI_MTX_Hardware], wants [ACPI_MTX_Namespace].
> 
> followed by more warnings. Is there easy way to debug this?

Maybe using one of the kernel debuggers? I haven't ever used one, but it 
sounds like you should have enough context where it would work..

	-pat