| Message ID | CA+abuVebbr9Cif_SfB+JcNhUOMV70fdWa-Q-nk+gFv9H5MRtgQ@mail.gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, May 28, 2015 at 02:55:48PM -0700, Sergey Kiselev wrote: Please fix your mail client to word wrap within paragraphs, it makes your mail more readable. > On Thu, May 28, 2015 at 12:26 PM, Mark Brown wrote: > OK, I was trying to move the request supplies and reset code to > wm8731_{i2c|spi}_probe(). And I noticed the following weirdness with > the supplies: The current code in wm8731_probe() requests and enables > regulators, resets the codec hardware, tweaks some registers, and then > disables all the regulators: That's totally normal, the new register values will be rewritten as part of resyncing the register cache when the device is powered back up.
On Thu, May 28, 2015 at 09:34:09AM -0700, Sergey Kiselev wrote: > Hi, > > I am working on a machine driver that uses wm8731 codec and I've > noticed that it will not work after rmmod/modprobe of that machine > driver. The issue is that wm8731_probe() resets the hardware, but it > doesn't reinitialize the regmap cache to the default values. > > The patch below fixes this issue. > > I also skimmed through other codecs, and some of them might have > similar issue (wm8510.c, wm8711.c, wm8750.c, wm8753.c, wm8940.c, > wm8960.c, wm8971.c, wm8974.c, wm8988.c, wm8990.c) > > Another way to avoid this issue might be resetting the codec in the > *_spi_probe()/*_i2c_probe(), so that it only gets reset once when > codec is loading. > > Thanks in advance for any comments/suggestions on the patch. > > Thanks, > Sergey > > Signed-off-by: Sergey Kiselev <skiselev@gmail.com> > In addition to the comments from Mark and Liam, if you could CC patches@opensource.wolfsonmicro.com when submitting patches to the Wolfson stuff, ensures we will see them going through, it is easy to miss stuff on the generic lists. Thanks, Charles
diff --git a/sound/soc/codecs/wm8731.c b/sound/soc/codecs/wm8731.c index 2245b6a..76af75b 100644 --- a/sound/soc/codecs/wm8731.c +++ b/sound/soc/codecs/wm8731.c @@ -84,7 +84,7 @@ static bool wm8731_writeable(struct device *dev, unsigned int reg) return reg <= WM8731_RESET; } -#define wm8731_reset(c) snd_soc_write(c, WM8731_RESET, 0) +static int wm8731_reset(struct snd_soc_codec *codec); static const char *wm8731_input_select[] = {"Line In", "Mic"};
Hi, I am working on a machine driver that uses wm8731 codec and I've noticed that it will not work after rmmod/modprobe of that machine driver. The issue is that wm8731_probe() resets the hardware, but it doesn't reinitialize the regmap cache to the default values. The patch below fixes this issue. I also skimmed through other codecs, and some of them might have similar issue (wm8510.c, wm8711.c, wm8750.c, wm8753.c, wm8940.c, wm8960.c, wm8971.c, wm8974.c, wm8988.c, wm8990.c) Another way to avoid this issue might be resetting the codec in the *_spi_probe()/*_i2c_probe(), so that it only gets reset once when codec is loading. Thanks in advance for any comments/suggestions on the patch. Thanks, Sergey Signed-off-by: Sergey Kiselev <skiselev@gmail.com> @@ -665,6 +665,21 @@ static const struct regmap_config wm8731_regmap = { .num_reg_defaults = ARRAY_SIZE(wm8731_reg_defaults), }; +static int wm8731_reset(struct snd_soc_codec *codec) +{ + int ret; + struct wm8731_priv *wm8731 = snd_soc_codec_get_drvdata(codec); + + ret = snd_soc_write(codec, WM8731_RESET, 0); + + if (ret != 0) + return ret; + + regmap_reinit_cache(wm8731->regmap, &wm8731_regmap); + + return 0; +} + #if defined(CONFIG_SPI_MASTER) static int wm8731_spi_probe(struct spi_device *spi) {