Hi Linus, On 06.12.2019 14:21, Linus Walleij wrote: > On Fri, Dec 6, 2019 at 12:53 PM Marek Szyprowski > <m.szyprowski@xxxxxxxxxxx> wrote: >> On 06.12.2019 10:58, Linus Walleij wrote: >>> On Fri, Dec 6, 2019 at 10:14 AM Marek Szyprowski >>> <m.szyprowski@xxxxxxxxxxx> wrote: >>> >>> BTW: >>> >>>> I really wonder why >>>> it worked fine with non-descriptor code and the ACTIVE_LOW DT flags... >>> The old code ignored the polarity flags in the device tree and >>> assumed everything was active high, that's how. It could as well >>> be hardcoded to 1337. >> Okay, then to restore current driver behavior after your patch, one has >> to change gpio flags in all dts to ACTIVE_HIGH... > Yeah :/ > > I think we should do a two-stage rocket here, if you make a patch to > all the DTS files I will make sure to add some logic enforcing the > right line levels in this patch as well. > > I'll make sure to assert reset expecting it to be flagged as active low. Frankly, if delay applying this patch one release after the DTS changes are applied, no workarounds in gpio core are needed. In such case we combine your patch with a driver logic cleanup for the reset gpio, so DTS can use ACTIVE_LOW flag then. Best regards -- Marek Szyprowski, PhD Samsung R&D Institute Poland
