Re: [PATCH v3 2/3] gpio/omap: modify wake-up register with interrupt enable.

[Kevin Hilman <khilman@linaro.org>]

Andreas Fenkart <andreas.fenkart@streamunlimited.com> writes:

> OMAP4430 TRM chap. 25.4.5.2
> To reduce dynamic consumption, an efficient idle scheme is based on the
> following:
> • An efficient local autoclock gating for each module
> • The implementation of control sideband signals between the PRCM module
>   and each module
> This enhanced idle control allows clocks to be activated and deactivated
> safely without requiring complex software management. The idle mode
> request, idle acknowledge, and wake-up request are sideband signals
> between the PRCM module and the general-purpose interface
>
> OMAP4430 TRM chap. 25.4.6.2
> There must be a correlation between the wake-up enable and interrupt
> enable register. If a GPIO pin has a wake-up configured on it, it must
> also have the corresponding interrupt enabled. Otherwise, it is possible
> there is a wake-up event, but after exiting the IDLE state, no interrupt
> is generated; the corresponding bit from the interrupt status register is
> not cleared, and the module does not acknowledge a future idle request.
>
> Up to now _set_gpio_triggering() is also handling the wake-up enable
> register. According the TRM this should be in sync with the interrupt
> enable. Wakeup is still enabled by default, since the module would not
> wake from idle otherwise.
> The enabled_wakeup_gpios was introduced to remember an explicit
> _set_gpio_wakeup beyond a mask/unmask cycle. Calling the flag flag
> disabled_wakeup_gpios would spare the problem of initializing it, but
> feels very unnatural to read.

There's a lot of description here, but it still fails to describe the
problem that is being solved and the motiviation for this change.  

> Wakeup functionality is completely untested, since the AM335x
> lacks a IRQWAKEN register.

So in addtion to a better description of the problem, and the changes,
this needs much broader testing, including on platforms with off-mode.

Kevin
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html