Hi,
On 16-12-2019 13:16, Linus Walleij wrote:
On Mon, Dec 16, 2019 at 12:11 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
Ugh, taking one last look at the "pinctrl: Export pinctrl_unregister_mappings"
patch it is no good, sorry.
Ooops!
Linus, can you please drop this from your -next ?
Sure, done.
So I see 2 options:
1) Add an orig_map member to maps_node and use that in the comparison,
this is IMHO somewhat ugly
2) Add a new pinctrl_register_mappings_no_dup helper and document in
pinctrl_unregister_mappings kdoc that it can only be used together
with the no_dup variant.
I believe that 2 is by far the best option. Linus do you agree or
do you have any other suggestions?
What about (3) look for all calls to pinctrl_register_mappings()
in the kernel.
Hey it is 2 places in total:
arch/arm/mach-u300/core.c: pinctrl_register_mappings(u300_pinmux_map,
drivers/pinctrl/cirrus/pinctrl-madera-core.c: ret =
pinctrl_register_mappings(pdata->gpio_configs,
Delete __initdata from the u300 table, the other one seems
safe. Fold this into your patch.
Go with the original idea.
That indeed sounds like a cleaner solution I will prepare a new version of
the patch (and this series for the i915 CI) with this approach.
Thanks & Regards,
Hans