On Sat, Apr 13, 2024 at 05:49:13PM +0200, Krzysztof Kozlowski wrote: > On 13/04/2024 17:23, Etienne Buira wrote: > >>> presence of gpio,syscon-dev node (or it will call dev_err() when probed). > >>> Correct rk3328.dtsi and related documentation to follow syscon's > >>> expectations. > >> > >> No, look at gpio-syscon driver. Parent is used. > > > > Parent is used, but the next lines are: > > ret = of_property_read_u32_index(np, "gpio,syscon-dev", 1, &priv->dreg_offset); > > if (ret) > > dev_err(...) > > > > So if gpio,syscon-dev does not have at least 2 items (or is missing), > > dev_err will be called, 3 items for dev_dbg. > > Current tree displays a spurious "can't read the data register offset" > > message. > > Hm, indeed, then I think driver, so > aa1fdda8f7ebf83f678e8d3c2ab4f1638c94195f, should be fixed. Otherwise > please say why binding is not correct and driver is good. I tried fixing the driver first, this was discussed here: https://lore.kernel.org/linux-gpio/ZhptEWb7tD5pummq@Z926fQmE5jqhFMgp6/T/#t you're welcome to comment. I have no strong opinion over what should be fixed, i just wished to shut a spurious error message, that i expected to be straightforward at first (hence good candidate for first kernel patch). Regards
