On Wed, Dec 14, 2022 at 01:37:41PM +0100, Lukasz Majewski wrote: > Without this change, the wm8940 driver is not working when > set_sysclk callback (wm8940_set_dai_sysclk) is called with > frequency not listed in the switch clause. > > This change adjusts this driver to allow non-standard frequency > set (just after the boot) being adjusted afterwards by the sound > system core code. > > Moreover, support for internal wm8940's PLL is provided, so it > can generate clocks when HOST system is not able to do it. > > Code in this commit is based on previous change done for wm8974 > (SHA1: 51b2bb3f2568e6d9d81a001d38b8d70c2ba4af99). > > Signed-off-by: Lukasz Majewski <lukma@xxxxxxx> > --- > struct snd_soc_component *component = dai->component; > + struct wm8940_priv *priv = snd_soc_component_get_drvdata(component); > u16 iface = snd_soc_component_read(component, WM8940_IFACE) & 0xFD9F; > u16 addcntrl = snd_soc_component_read(component, WM8940_ADDCNTRL) & 0xFFF1; > u16 companding = snd_soc_component_read(component, > WM8940_COMPANDINGCTL) & 0xFFDF; > int ret; > > + priv->fs = params_rate(params); > + ret = wm8940_update_clocks(dai); > + if (ret) > + return ret; > + I think this all looks mostly good, my only slight concern would be the interaction with the manual functions for settings the PLL etc. I guess under this code, whatever manual settings were configured will be overwritten with the new auto settings, I think this should be fine as the PLL wants to be run in a pretty narrow band anyway, so the settings are likely identical. Do you have any thoughts? Thanks, Charles
