svn commit: r208152 - in head/sys: dev/ofw powerpc/aim powerpc/ofw

[Nathan Whitehorn]

On 05/16/10 16:47, Nathan Whitehorn wrote:
> On 05/16/10 16:38, Marius Strobl wrote:
>> On Sun, May 16, 2010 at 03:57:00PM +0000, Nathan Whitehorn wrote:
>>> Author: nwhitehorn
>>> Date: Sun May 16 15:56:59 2010
>>> New Revision: 208152
>>> URL: http://svn.freebsd.org/changeset/base/208152
>>>
>>> Log:
>>>    On PowerMac11,2 and (presumably) PowerMac12,1, we need to quiesce 
>>> the
>>>    firmware in order to take over control of the SMU. Without doing 
>>> this,
>>>    the firmware background process doing fan control will run amok 
>>> as we
>>>    take over the system and crash the management chip.
>>>
>>>    This is limited to these two machines because our kernel is heavily
>>>    dependent on firmware accesses, and so quiescing firmware can cause
>>>    nasty problems.
>>>
>> Given that the "quiesce" service isn't part of the Open Firmware
>> standard nor of one of its supplements as far as I can see and
>> given this service isn't available on sun4u I'd prefer this
>> implementation to be entirely moved to MD parts instead (similar
>> to the "SUNW,<foo>" services which we keep MD).
>
> This is tricky on PowerPC, because there are 3 OF implementations in 
> use (native virtual-mode, 32-bit virtual mode, and 32-bit real mode), 
> and it has to go in all of them. I'll try to come up with a 
> non-horrible way to put this in the MD code, and I won't MFC this 
> change in the meantime.
> -Nathan

Immediately after writing this, I realized I had tricked myself into 
thinking it was hard. Since this is called very early on, in the context 
set up by Open Firmware, none of those distinctions matter. It's been 
pulled back in to PPC-specific code in revision 208172.
-Nathan