The reality is I've been caught up with other events in my life
and don't see working on the amlogic stuff happening in the near
term. I'm happy if someone wants to take it over and certainly
understand that if the decision is to remove it from the tree.

I've responded to the inline comments.

I've responded to the inline comments.

I've made the suggested changes (as well as adding break to the corresponding code in the m6 usb clock code).

I've made the suggested change (as well as propagating the change to other files which I've piggybacked onto this review).

Typically you would use the gic on the -m6. However it does
also have a pic ... most of my initial development was done
on a -m6 using the pic.

Directly set cpufuncs.cf_sleep rather than using cpu_do_powersave.

I don't believe so and the gic as well as the ti/aintc arm_mask_irq routines
do not currently call arm_irq_memory_barrier.

Reworked as suggested and retested.

The change merely wires up a feature that the comments in cpufunc.c
claim already exists ... though currently cpu_do_powersave don't seem
to do anything useful.

Simply including aml_pic shouldn't be sufficient to enable it as the interrupt controller
(some platforms may require it to correctly disable the PIC hardware so that GIC can
be used).

I've reformatted the line in question.