Hi Tomasz, On Wed, Sep 20, 2017 at 12:32:22PM +0300, sakari.ailus@xxxxxx wrote: > > >> >> @@ -743,11 +770,17 @@ static int at24_probe(struct i2c_client *client, const > > >> >> struct i2c_device_id *id) > > >> >> > > >> >> i2c_set_clientdata(client, at24); > > >> >> > > >> >> + /* enable runtime pm */ > > >> >> + pm_runtime_get_noresume(&client->dev); > > >> >> + pm_runtime_set_active(&client->dev); > > >> >> + pm_runtime_enable(&client->dev); > > >> > > >> Do we need this get_noresume/set_active dance? I remember it was for > > >> some reason needed for PCI devices, but I don't see why for I2C > > >> anything else than just pm_runtime_enable() would be necessary. > > > > > > You specifically do not need (all) this for PCI devices, but AFAIU for I涎 > > > devices you do. The runtime PM status of a device is disabled by default > > > and the use count is zero, but on ACPI based systems the device is still > > > powered on. > > > > Okay, so _get_noresume() and _set_active() would do the thing for ACPI > > indeed, but not sure about other platforms. Perhaps _enable(), > > _get_sync() would be more general? > > What I ended up doing in e.g. the smiapp driver was to explicitly power the > device on first and then enable runtime PM. (See > drivers/media/i2c/smiapp/smiapp-core.c .) This approach works even if > CONFIG_PM is disabled, both on DT and ACPI. pm_runtime_get_noresume() + pm_runtime_put() can be replaced by a single pm_runtime_idle() call (where pm_runtime_put() was). pm_runtime_enable() is required to enable runtime PM for a device. pm_runtime_allow() may be omitted but then to make the runtime PM framework to make the power state transitions this needs to be configured from the user space --- which I don't think is intended. Cc linux-pm, too. -- Regards, Sakari Ailus e-mail: sakari.ailus@xxxxxx
