| Message ID | 20211015133619.4698-9-rf@opensource.cirrus.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ASoC: cs42l42: Collection of bugfixes | expand |
On Fri, Oct 15, 2021 at 02:36:11PM +0100, Richard Fitzgerald wrote: > Driver remove() should assert RESET and disable the supplies. > Previously this assumed that calling pm_runtime_suspend() would result > in a call to cs42l42_runtime_suspend() to power-down. This isn't > guaranteed - pm_runtime can be disabled. > pm_runtime_suspend(&i2c_client->dev); > pm_runtime_disable(&i2c_client->dev); > > + gpiod_set_value_cansleep(cs42l42->reset_gpio, 0); > + regulator_bulk_disable(ARRAY_SIZE(cs42l42->supplies), cs42l42->supplies); Won't this end up with an extra disable of the regulators if they're already disabled?
diff --git a/sound/soc/codecs/cs42l42.c b/sound/soc/codecs/cs42l42.c index 420e16563c45..cf1f68474d21 100644 --- a/sound/soc/codecs/cs42l42.c +++ b/sound/soc/codecs/cs42l42.c @@ -2131,6 +2131,9 @@ static int cs42l42_i2c_remove(struct i2c_client *i2c_client) pm_runtime_suspend(&i2c_client->dev); pm_runtime_disable(&i2c_client->dev); + gpiod_set_value_cansleep(cs42l42->reset_gpio, 0); + regulator_bulk_disable(ARRAY_SIZE(cs42l42->supplies), cs42l42->supplies); + return 0; }
Driver remove() should assert RESET and disable the supplies. Previously this assumed that calling pm_runtime_suspend() would result in a call to cs42l42_runtime_suspend() to power-down. This isn't guaranteed - pm_runtime can be disabled. Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com> Fixes: 2b869e0ea598 ("ASoC: cs42l42: Remove power if the driver is being removed") --- sound/soc/codecs/cs42l42.c | 3 +++ 1 file changed, 3 insertions(+)