On Mon, 4 Dec 2023 18:49:21 +0200 Ceclan Dumitru <mitrutzceclan@xxxxxxxxx> wrote: > On 11/25/23 19:21, Jonathan Cameron wrote: > > On Thu, 23 Nov 2023 17:23:22 +0200 > > mitrutzceclan <mitrutzceclan@xxxxxxxxx> wrote: > > > >> From: Dumitru Ceclan <mitrutzceclan@xxxxxxxxx> > > ... > >> + st->regulators[0].supply = ad7173_ref_sel_str[AD7173_SETUP_REF_SEL_EXT_REF]; > >> + st->regulators[1].supply = ad7173_ref_sel_str[AD7173_SETUP_REF_SEL_EXT_REF2]; > >> + st->regulators[2].supply = ad7173_ref_sel_str[AD7173_SETUP_REF_SEL_AVDD1_AVSS]; > >> + > >> + ret = devm_regulator_bulk_get(dev, ARRAY_SIZE(st->regulators), > >> + st->regulators); > > > > If only some of them are supplied, the driver works fine as long as no channel is using them? > > If so should probably get the optionally then check for availability of the ones you > > want. I'm sure someone will just wire up ref. Incidentally, ref2 isn't there for all chips > > I think, so we should not get it on the ones where it can't exist. > > > > > This sets a dummy regulator in place if no proper supply is found. Then > the call regulator_get_voltage() on the dummy will fail. About getting > ref2, sure, I'll set the string only with the right ID. > > ... > >> + ret = ad7173_get_ref_voltage_milli(st, (u8)ref_sel); > >> + if (ret < 0) > >> + return dev_err_probe(dev, ret, > >> + "Cannot use reference %u", ref_sel); > > Here the probe would not continue if a channel selects a supply that is > not available in the DT. Ok. It's a little ugly though - so maybe a comment at the bulk_get to say it will fail cleanly a bit later.
