Hi Rafael,

On Fri, 27 Feb 2015 23:33:35 +0100
"Rafael J. Wysocki" [off-list ref] wrote:

On Friday, February 27, 2015 04:09:52 PM Boris Brezillon wrote:

quoted

Commit cab303be91dc47942bc25de33dc1140123540800 [1] introduced a WARN_ON
test which triggers a WARNING backtrace on at91 platforms.
While this WARN_ON is absolutely necessary to warn users that they should
not mix request with and without IRQF_NO_SUSPEND flags on shared IRQs,
there is no easy way to solve this issue on at91 platforms.

The main reason is that the init timer is often using a shared irq line
and thus request this irq with IRQF_NO_SUSPEND flag set, while other
peripherals request the same irq line without this flag.

This problem has recently been addressed by this patch [2] which adds
a new IRQF_COND_SUSPEND flag, that authorize mixing IRQF_NO_SUSPEND and
!IRQF_NO_SUSPEND as long as irq handlers setting IRQF_COND_SUSPEND
can safely be called in suspended state.

Doing this also implies taking care of system wakeup in devices handlers
if they tag the IRQ line as a wakeup source.
The first patch of this series exports the pm_system_wakeup symbol so
that drivers can call pm_system_wakeup from their interrupt handler.

This series then patches all at91 drivers that can have devices sharing
their IRQ line with a timer.

This series depends on [2].

Series looks good to me (although admittedly I'm not familiar with the
drivers in question).

Do you want me to apply it on top of [2]?

I'd like to wait for Nicolas' ack.
BTW, I realized I didn't send the series to the relevant subsystem
maintainers and I forgot to increment the patch version (I already
posted a v1 a few weeks ago), so I'll probably resend it.

You can take the first patch though.

Thanks,

Boris

-- 
Boris Brezillon, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com