[PATCH] ppc64: SMU driver update & i2c support

Benjamin Herrenschmidt benh at kernel.crashing.org
Wed Sep 21 09:31:52 EST 2005


On Tue, 2005-09-20 at 16:20 -0700, Andrew Morton wrote:

> Well it is officially extremely sinful of you to attempt to sneak a major
> driver rewrite into -rc2.

Yup, my bad. I should have sent you the patch before going to vacation,
got distracted. I had the rewrite around for some time now. (Bah, I
didn't even notice we were past -rc1 in fact !)

> Does this version have all the functionality of the earlier version?  Will
> it break any existing userland tools?  How well tested is it?  Vendors may
> base releases on 2.6.14.  Will they get grumpy about a half-finished
> subsystem?

The previous driver had pretty much no feature at all and no userland
interface. All it did was read the real time clock at boot and implement
writing to it via genrtc. It's pretty much adding a driver more than
replacing one in fact.

This driver is also only used on some recent Apple machines (iMac G5 and
recent single CPU G5 desktops).

This new version adds the userland interface so tools can be written in
fact :) The previous one had none, so no tool breakage is to be
expected. Some people have already been writing some thermal control
tools based on this new version (well, a slightly earlier one that I
posted to the ppc lists a while ago) which is the main reason why I'm
trying to get it in 2.6.14.

The only possible regression I can see here is more the addition of the
"parent" parameter to the OF platform device creation which might break
build of some configs. Few people use the OF platform device facility
though, it's pretty much a mac only thing. I'll double check ppc32 pmac
and pegasos today just in case.

Ben.




More information about the Linuxppc64-dev mailing list