On Wed, Feb 07, 2024 at 03:30:16PM +0000, Mark Brown wrote: > On Wed, Feb 07, 2024 at 04:40:38PM +0200, Andy Shevchenko wrote: > > On Wed, Feb 07, 2024 at 01:47:42PM +0800, Shenghao Ding wrote: ... > > > + .max_register = 0x44ffffff, > > > I'm always wondering how this can work in debugfs when one tries to dump all > > registers... > > regmap will filter out non-readable registers from what it shows in > debugfs so sparse regmaps will end up much smaller than they might look. > The debugfs code also has support for seeking into the output (since > each register is a fixed size) so we'll only actually render whatever is > being read in an individual read(). Thanks, good to know. ... > > > +static const struct dev_pm_ops tas2783_sdca_pm = { > > > + SET_SYSTEM_SLEEP_PM_OPS(tas2783_sdca_dev_suspend, > > > + tas2783_sdca_dev_resume) > > > + SET_RUNTIME_PM_OPS(tas2783_sdca_dev_suspend, > > > + tas2783_sdca_dev_resume, NULL) > > > +}; > > > Use new PM macros. > > ...which are? SYSTEM_SLEEP_PM_OPS() RUNTIME_PM_OPS() and pm_ptr() in the .driver.pm assignment. -- With Best Regards, Andy Shevchenko