On Tue, Apr 24, 2012 at 08:54:26AM +0100, Russell King - ARM Linux wrote:
On Tue, Apr 24, 2012 at 09:38:43AM +0800, Shawn Guo wrote:
On Mon, Apr 23, 2012 at 10:45:02AM -0500, Rob Lee wrote:
Let me try last time. What about having a late_initcall hook in machine_desc?
Also fine with me.
Shall I add Shawn's patch to my imx cpuidle patchset or should the arch/arm/kernel/setup.c and arch.h changes be submitted separately? If separately, Shawn, did you want to submit this patch or should I?
Strange. Russell is not in the Cc list. I remember I added Russell into Cc when I propose the idea. Added him again.
I didn't see any message in this thread cc'd to me, but that's not to say I hadn't already read this patch. I don't have any comment against it, but I do wonder how often this hook would be used.
I guess mach-* that use common cpuidle will likely need this hook.
We do seem to have quite a number of late_initcall()s in arch/arm/mach-*, so it seems to be a good idea - provided someone's willing to convert all those users of late_initcall()s.
Agreed. The late_initcall()s in arch/arm/mach-* will not scale for long time, since we are moving toward single build.