13/02/2019 07:14, Petr Štetiar:
Mathias Kresin <d...@kresin.me> [2019-02-12 23:19:24]:
12/02/2019 20:49, Daniel Golle:
The patch was copied to a bunch of platforms and seems to be a quite
useful feature for various things. Move it to generic to avoid code
duplication.
Follow up to the discussion in
https://github.com/openwrt/openwrt/pull/1814#issuecomment-462942022.
We should rather try to get rid of the patch instead of encourage people of
using it. Has to be read as soft NAK.
The patch was a hack for missing kernel functionality and was rejected
upstream a long time ago. The main use case for the patch is pulling reset
GPIOs or enable a USB VCC supply. More and more drivers/subsystems are
getting support to control reset GPIOs or enable the VCC supply on their
own.
The upstream answer to gpio-exports are gpio-hogs. It is just that nobody
migrated the devicetree source files to GPIO hogs so far. Even if gpio-hogs
are still a workaround for drivers not able pull their reset GPIOs/enable
VCC supply, it seem to me the cleaner workaround.
I've sent RFC[1] about userspace replacement as well.
1. http://lists.infradead.org/pipermail/openwrt-devel/2019-February/015715.html
-- ynezz
Hey Petr,
I'm well aware of your mail and I agree that we need to switch to
libgpiod et al. But I don't feel confident to answer any of the question
raised.
I still hope, someone - who has a better idea how/where to integrate the
functionality - picks up your mail.
Mathias
_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
