On Thu, Nov 7, 2013 at 11:38 AM, Markus Pargmann [off-list ref] wrote:
On Thu, Nov 07, 2013 at 10:28:01AM +0100, Lucas Stach wrote:
quoted
quoted
The first version of this series [1] was designed to have a iomux node
with 6 gpio subnodes.
Why was this changed? Having two different DT nodes requesting the same
IO region is certainly the wrong thing to do.
It was suggested to map the same memory range from both drivers, so I
changed the layout completely. Perhaps it would have been better to keep
the DT node structure while not passing the memory to the gpio
subdevices.
From a pinctrl/GPIO point of view anything goes, it's a little bit like
big-endian vs little-endian.
I did merge the i.MX27 driver as it was ACKed by the subarch
maintainers.
So what do we do now, shall I revert this patch or can you refactor
it in accordance with the comments so it should be kept as a
base for the rewrite?
Yours,
Linus Walleij