Re: [PATCH V6 02/15] ARM: OMAP2+: AM33XX: add lateinit hook for calling pm late init

On 10/10/2013 10:20 AM, Joel Fernandes wrote:

On 10/10/2013 08:32 AM, Nishanth Menon wrote:

quoted

On 00:32-20131010, Joel Fernandes wrote:

quoted

On 10/09/2013 06:24 PM, Nishanth Menon wrote:

quoted

Call OMAP2+ generic lateinit hook from AM specific late init hook.
This allows the generic late initializations such as cpufreq hooks
to be active.

Cc: Benoit Cousson <redacted>
Cc: Kevin Hilman <redacted>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Tony Lindgren <tony@atomide.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
---
 arch/arm/mach-omap2/board-generic.c |    1 +
 arch/arm/mach-omap2/common.h        |    1 +
 arch/arm/mach-omap2/io.c            |    6 ++++++
 3 files changed, 8 insertions(+)

diff --git a/arch/arm/mach-omap2/board-generic.c b/arch/arm/mach-omap2/board-generic.c
index 87162e1..b474498 100644
--- a/arch/arm/mach-omap2/board-generic.c
+++ b/arch/arm/mach-omap2/board-generic.c

@@ -180,6 +180,7 @@ DT_MACHINE_START(AM33XX_DT, "Generic AM33XX (Flattened Device Tree)")
 	.init_irq	= omap_intc_of_init,
 	.handle_irq	= omap3_intc_handle_irq,
 	.init_machine	= omap_generic_init,
+	.init_late	= am33xx_init_late,

Instead of adding a new a *_init_late function for every platform, perhaps
better to just do:
.init_late	= omap2_common_pm_late_init;

since that's the only function you're calling.

Later if more functions are added, then it can be wrapped around.

And what benefit would that give us? we break consistency of functions
available in io.c, considering the work we have already done in
out-of-tree patches on ti-forks, we *do know* that there is more
incoming and has to be done anyways, I prefer having symmetric
functions and a placeholder that folks can add on to.

Ok, sure if you think such placeholders are right way and there's more code
later to be added to init_late callbacks.

Adding 6 lines of identical code for every platform seems redundant, I'd just
define an omap_common_late_init for all platforms for now, and then fork off for
the future differences. That way we save on LOC and readability in the header as
well. Your call.

I see a function of that name already exists. I guess you can leave your patch
as is then and not have to do this.

regards,

-Joel