On Thu, Apr 16, 2020 at 11:23:14AM -0400, Sven Van Asbroeck wrote: > On Thu, Apr 16, 2020 at 8:42 AM Mark Brown <broonie@xxxxxxxxxx> wrote: > > > @@ -0,0 +1,643 @@ > > > +// SPDX-License-Identifier: GPL-2.0-only > > > +/* > > > + * Codec driver for Microsemi ZL38060 Connected Home Audio Processor. > > > + * > > Please make the entire comment a C++ one so things look more > > intentional. > The 'weird' combination of // SPDX and /* Description/copyright */ seems to > be a kernel-wide standard (for C files, at least) ? > E.g.: > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/sound/soc/codecs/wm9090.c?h=v5.7-rc1#n2 > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/sound/soc/codecs/wm8904.c?h=v5.7-rc1#n2 > Ok to keep? Those were all automatically converted, one of these days I might get round to updating them :/ > > This external amplifier support shouldn't be here, if there's other > > devices in the system then they will have their own drivers and the > > machine driver will take care of linking things together. > In our application, the amp is a "dumb" class-D amp with a single enable line: > https://www.onsemi.com/pub/Collateral/FAB3103-D.pdf > I am not sure how I could make this more general. Could you point me to an > example somewhere in the tree? Look for GPIOs in machine drivers, there's quite a lot of examples (eg, rx51 has a speaker amp although it's not the most modern). > > > + priv->regmap = devm_regmap_init(dev, &zl38_regmap_bus, spi, > > > + &zl38_regmap_conf); > > > + if (IS_ERR(priv->regmap)) > > > + return PTR_ERR(priv->regmap); > > devm_regmap_init_spi() > I wish !! This chip has complex SPI addressing, using an "address" which: > - is variable length, depending on the page of the register being accessed; > - contains a field with the length of the data to follow. OK.
Attachment:
signature.asc
Description: PGP signature
