[PATCH 8/8] gpio/omap: fix missing check in *_runtime_suspend()

[Tony Lindgren]

* Kevin Hilman <khilman at ti.com> [120517 17:00]:
> 
> Argh, then $SUBJECT patch here has caused brokeness in multiple ways.
> It managed to break both runtime suspend and runtime resume at the same
> time. :(
> 
> The change added by this patch to runtime_suspend effectively disables
> the fix I did in 68942edb09 (gpio/omap: fix wakeups on level-triggered GPIOs)
> causing the sluggish network problems to reappear, since that GPIO IRQ
> is no longer causing wakeups.
>
> Simple fix is below, which just moves the check added in $SUBJECT patch
> below the workaround for the edge/level fix.  Can you confirm it works
> on Zoom3 (applies on gpio/next + my previous fix.)

Thanks yes having both of your fixes applied fixes the issues I was
seeing, so for both:

Tested-by: Tony Lindgren <tony at atomide.com>
 
> I didn't notice the same problem on Overo, but I guess it's because
> Overo had some enabled non-wakeup GPIOs in the same bank, so it didn't
> bypass the level-triggered IRQ fix.

Makes sense for why it only appears on some boards.
 
> In the mean time, they're availble in my for_3.5/fixes/gpio-2 branch.

I'll merge that also into l-o master for some more testing before
the merge window.

Regards,

Tony