On Wed, Aug 17, 2016 at 01:47:40PM -0700, Bjorn Andersson wrote:

On Wed 17 Aug 12:43 PDT 2016, Andy Gross wrote:

quoted

On Wed, Aug 17, 2016 at 10:56:22AM +0200, Linus Walleij wrote:

quoted

On Fri, Aug 5, 2016 at 10:38 AM, Linus Walleij [off-list ref] wrote:

quoted

This came out after some frustration with trying to flag a PM8058
interrupt line on the APQ8060 (MSM8660) to trigger on falling
edges. It was conflicting badly with predefined contradictory
flags in the device tree for the entire GPIO block. The same
seems to hold for the MPP GPIO IRQs.

Andy are you funneling these two patches for fixes?

I looked at these last night.  What's the outcome of John's issues with the
patch not fixing his problem?  Do these need to be taken irrespective of his
plight?

John's issue was that the order of operations was changed in the IRQ
framework, so that upon passing through the set_type() in the TLMM
driver I would overwrite, and hence disable, the chained IRQ handler -
as I should.

As the PMIC interrupt is chained from the TLMM no interrupts was
delivered.


So please move forward with these changes.

Ah ok.  I'll queue these up.

Andy