On Mon, Mar 21, 2022 at 11:34 AM Paul Cercueil <paul@xxxxxxxxxxxxxxx> wrote: > Le dim., mars 20 2022 at 15:20:47 +0000, Jonathan Cameron > <jic23@xxxxxxxxxx> a écrit : > > On Thu, 10 Mar 2022 20:54:50 +0800 > > Zizhuang Deng <sunsetdzz@xxxxxxxxx> wrote: > > > >> The third call to `fwnode_property_read_u32` did not record > >> the return value, resulting in `channel_offstate` possibly > >> being assigned the wrong value. > > Definitely rather odd looking and I think your conclusion is correct. > > +CC Paul for confirmation that this isn't doing something clever.. > > It's been a while, but I don't think there was anything clever going on > here - so the patch is fine. Basically the question here is what value should offstate have when there is no such property. Currently it's the same value as modes (no seeing context other than in the patch). > >> if (!ret) > >> st->channel_modes[reg] = tmp; > >> > >> - fwnode_property_read_u32(child, "adi,off-state", &tmp); > >> + ret = fwnode_property_read_u32(child, "adi,off-state", &tmp); > >> if (!ret) > >> st->channel_offstate[reg] = tmp; -- With Best Regards, Andy Shevchenko
