Thanks for the review and comments.
I'm going to send v1 patches later.
Bibby
On Fri, 2019-08-30 at 13:58 +0200, Bartosz Golaszewski wrote:
> czw., 29 sie 2019 o 04:04 Bibby Hsieh <bibby.hsieh@xxxxxxxxxxxx> napisał(a):
> >
> > Although in the most platforms, the power of eeprom and i2c
> > are alway on, some platforms disable the eeprom and i2c power
> > in order to meet low power request.
> > This patch add the pm_runtime ops to control power to support
> > all platforms.
> >
> > Signed-off-by: Bibby Hsieh <bibby.hsieh@xxxxxxxxxxxx>
> > ---
> > drivers/misc/eeprom/at24.c | 61 +++++++++++++++++++++++++++++++++++++-
> > 1 file changed, 60 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/misc/eeprom/at24.c b/drivers/misc/eeprom/at24.c
> > index 35bf2477693d..2843e4b4aacd 100644
> > --- a/drivers/misc/eeprom/at24.c
> > +++ b/drivers/misc/eeprom/at24.c
> > @@ -23,6 +23,7 @@
> > #include <linux/nvmem-provider.h>
> > #include <linux/regmap.h>
> > #include <linux/pm_runtime.h>
> > +#include <linux/regulator/consumer.h>
> > #include <linux/gpio/consumer.h>
> >
> > /* Address pointer is 16 bit. */
> > @@ -68,6 +69,12 @@
> > * which won't work on pure SMBus systems.
> > */
> >
> > +static const char * const at24_supply_names[] = {
> > + "power", "i2c",
> > +};
> > +
> > +#define AT24_NUM_SUPPLIES ARRAY_SIZE(at24_supply_names)
> > +
> > struct at24_client {
> > struct i2c_client *client;
> > struct regmap *regmap;
> > @@ -92,6 +99,8 @@ struct at24_data {
> >
> > struct gpio_desc *wp_gpio;
> >
> > + bool has_supplies;
> > + struct regulator_bulk_data supplies[AT24_NUM_SUPPLIES];
> > /*
> > * Some chips tie up multiple I2C addresses; dummy devices reserve
> > * them for us, and we'll use them with SMBus calls.
> > @@ -663,6 +672,17 @@ static int at24_probe(struct i2c_client *client)
> > at24->client[0].client = client;
> > at24->client[0].regmap = regmap;
> >
> > + for (i = 0; i < AT24_NUM_SUPPLIES; i++)
> > + at24->supplies[i].supply = at24_supply_names[i];
>
> Please, add a newline here.
>
> > + err = devm_regulator_bulk_get(&at24->client[0].client->dev,
> > + AT24_NUM_SUPPLIES,
> > + at24->supplies);
>
> This fits nicely in 80 chars - no need to break the line here again.
>
> > + if (err == -EPROBE_DEFER) {
> > + dev_err(dev, "Failed to get power regulators\n");
>
> I assume you want to make the supplies optional by only returning on
> EPROBE_DEFER, but why the error log then?
>
> > + return err;
> > + }
> > + at24->has_supplies = !err ? true : false;
>
> Just a simple at24->has_supplies = !err is enough.
>
> > +
> > at24->wp_gpio = devm_gpiod_get_optional(dev, "wp", GPIOD_OUT_HIGH);
> > if (IS_ERR(at24->wp_gpio))
> > return PTR_ERR(at24->wp_gpio);
> > @@ -705,13 +725,21 @@ static int at24_probe(struct i2c_client *client)
> > /* enable runtime pm */
> > pm_runtime_set_active(dev);
> > pm_runtime_enable(dev);
> > + pm_runtime_get_sync(dev);
> > + if (at24->has_supplies) {
> > + err = regulator_bulk_enable(AT24_NUM_SUPPLIES, at24->supplies);
>
> Unrelated to this patch, but I'm surprised we don't have optional
> variants for bulk regulator operations.
>
> > + if (err) {
> > + dev_err(dev, "Failed to enable power regulators\n");
> > + return err;
> > + }
> > + }
> >
> > /*
> > * Perform a one-byte test read to verify that the
> > * chip is functional.
> > */
> > err = at24_read(at24, 0, &test_byte, 1);
> > - pm_runtime_idle(dev);
> > + pm_runtime_put(dev);
> > if (err) {
> > pm_runtime_disable(dev);
> > return -ENODEV;
> > @@ -726,15 +754,46 @@ static int at24_probe(struct i2c_client *client)
> >
> > static int at24_remove(struct i2c_client *client)
> > {
> > + struct at24_data *at24 = i2c_get_clientdata(client);
> > +
> > pm_runtime_disable(&client->dev);
> > pm_runtime_set_suspended(&client->dev);
> > + if (at24->has_supplies)
> > + regulator_bulk_disable(AT24_NUM_SUPPLIES, at24->supplies);
> >
> > return 0;
> > }
> >
> > +static int __maybe_unused at24_suspend(struct device *dev)
> > +{
> > + struct i2c_client *client = to_i2c_client(dev);
> > + struct at24_data *at24 = i2c_get_clientdata(client);
> > +
> > + if (at24->has_supplies)
> > + return regulator_bulk_disable(AT24_NUM_SUPPLIES,
> > + at24->supplies);
> > +
> > + return 0;
> > +}
> > +
> > +static int __maybe_unused at24_resume(struct device *dev)
> > +{
> > + struct i2c_client *client = to_i2c_client(dev);
> > + struct at24_data *at24 = i2c_get_clientdata(client);
> > +
> > + if (at24->has_supplies)
> > + return regulator_bulk_enable(AT24_NUM_SUPPLIES,
> > + at24->supplies);
> > +
> > + return 0;
> > +}
> > +
> > +static SIMPLE_DEV_PM_OPS(at24_pm_ops, at24_suspend, at24_resume);
> > +
> > static struct i2c_driver at24_driver = {
> > .driver = {
> > .name = "at24",
> > + .pm = &at24_pm_ops,
> > .of_match_table = at24_of_match,
> > .acpi_match_table = ACPI_PTR(at24_acpi_ids),
> > },
> > --
> > 2.18.0
> >
>
> The rest looks good.
>
> Bart
