On Thu, Jan 31, 2019 at 4:44 PM Tony Lindgren <tony@xxxxxxxxxxx> wrote:
-----
> SoC fixes for omaps for v5.0-rc cycle
>
> This series contains two SoC regression fixes and one uninitialized
> variable fix:
>
> - Fix inverted nirq pin handling for omap5 that started producing
> warnings with earlier GIC direction checks and took a while to
> understand and confirm. Basically there are two sys_nirq pins
> that are bypassing peripheral modules and inverted automatically
> by the SoC and need to be handled with a custom irq_set_type()
>
> - Recent ti-sysc changes caused a regression to the pwm-omap-dmtimer
> code where the device tree handling code for timer source clock
> gets confused. It looks like we can remove that code eventually,
> but for now we just drop a bogus pm_runtime_irq_safe() for the
> timers with the related quirks caused by pm_runtime_irq_safe(),
> and have the standard assigned-clocks and assigned-clock-parents
> deal with setting the source clock
>
> - Fix potentially uninitialized value for display init code if
> regmap_read() fails
Pulled into arm/fixes branch, thanks!
Arnd
