| Message ID | 20210205132505.20173-8-sakari.ailus@linux.intel.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Support running driver's probe for a device powered off | expand |
On Fri, Feb 5, 2021 at 2:25 PM Sakari Ailus <sakari.ailus@linux.intel.com> wrote: > > In certain use cases (where the chip is part of a camera module, and the > camera module is wired together with a camera privacy LED), powering on > the device during probe is undesirable. Add support for the at24 to > execute probe while being powered off. For this to happen, a hint in form > of a device property is required from the firmware. > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> > Reviewed-by: Tomasz Figa <tfiga@chromium.org> > --- I'll ack this but I still claim that the name acpi_dev_state_low_power() is super misleading for this use-case and I've been saying that for 10 versions now with everyone just ignoring my remarks. :/ Acked-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
On Mon, Feb 8, 2021 at 5:54 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > On Mon, Feb 8, 2021 at 5:44 PM Bartosz Golaszewski > <bgolaszewski@baylibre.com> wrote: > > > > On Fri, Feb 5, 2021 at 2:25 PM Sakari Ailus > > <sakari.ailus@linux.intel.com> wrote: > > > > > > In certain use cases (where the chip is part of a camera module, and the > > > camera module is wired together with a camera privacy LED), powering on > > > the device during probe is undesirable. Add support for the at24 to > > > execute probe while being powered off. For this to happen, a hint in form > > > of a device property is required from the firmware. > > > > > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> > > > Reviewed-by: Tomasz Figa <tfiga@chromium.org> > > > --- > > > > I'll ack this but I still claim that the name > > acpi_dev_state_low_power() is super misleading for this use-case and > > I've been saying that for 10 versions now with everyone just ignoring > > my remarks. :/ > > Well, the function in question simply checks if the current ACPI power > state of the device is different from "full power", so its name > appears to be quite adequate to me. > > If the way in which it is used is confusing, though, I guess > explaining what's going on would be welcome. > Yes, I have explained it multiple time already - last time at v9 of this series: https://www.spinics.net/lists/kernel/msg3816807.html Bartosz > > Acked-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
On Tue, Feb 9, 2021 at 5:23 PM Sakari Ailus <sakari.ailus@linux.intel.com> wrote: > > Hi Bartosz, Rafael, > > On Tue, Feb 09, 2021 at 04:49:37PM +0100, Bartosz Golaszewski wrote: > > On Mon, Feb 8, 2021 at 5:54 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > > > > > On Mon, Feb 8, 2021 at 5:44 PM Bartosz Golaszewski > > > <bgolaszewski@baylibre.com> wrote: > > > > > > > > On Fri, Feb 5, 2021 at 2:25 PM Sakari Ailus > > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > > > In certain use cases (where the chip is part of a camera module, and the > > > > > camera module is wired together with a camera privacy LED), powering on > > > > > the device during probe is undesirable. Add support for the at24 to > > > > > execute probe while being powered off. For this to happen, a hint in form > > > > > of a device property is required from the firmware. > > > > > > > > > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> > > > > > Reviewed-by: Tomasz Figa <tfiga@chromium.org> > > > > > --- > > > > > > > > I'll ack this but I still claim that the name > > > > acpi_dev_state_low_power() is super misleading for this use-case and > > > > I've been saying that for 10 versions now with everyone just ignoring > > > > my remarks. :/ > > > > > > Well, the function in question simply checks if the current ACPI power > > > state of the device is different from "full power", so its name > > > appears to be quite adequate to me. > > > > > > If the way in which it is used is confusing, though, I guess > > > explaining what's going on would be welcome. > > > > > > > Yes, I have explained it multiple time already - last time at v9 of this series: > > > > https://www.spinics.net/lists/kernel/msg3816807.html > > How about adding this to the description of acpi_dev_state_low_power(): > > -----------8<-------------- > * This function is intended to be used by drivers to tell whether the device > * is in low power state (D1--D3cold) in driver's probe or remove function. See > * Documentation/firmware-guide/acpi/low-power-probe.rst for more information. > -----------8<-------------- This information is already there in the kerneldoc description of that function AFAICS. I was thinking about adding an explanation comment to the caller.
On Tue, Feb 9, 2021 at 5:54 PM Sakari Ailus <sakari.ailus@linux.intel.com> wrote: > > On Tue, Feb 09, 2021 at 05:42:45PM +0100, Rafael J. Wysocki wrote: > > On Tue, Feb 9, 2021 at 5:23 PM Sakari Ailus > > <sakari.ailus@linux.intel.com> wrote: > > > > > > Hi Bartosz, Rafael, > > > > > > On Tue, Feb 09, 2021 at 04:49:37PM +0100, Bartosz Golaszewski wrote: > > > > On Mon, Feb 8, 2021 at 5:54 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > > > > > > > > > On Mon, Feb 8, 2021 at 5:44 PM Bartosz Golaszewski > > > > > <bgolaszewski@baylibre.com> wrote: > > > > > > > > > > > > On Fri, Feb 5, 2021 at 2:25 PM Sakari Ailus > > > > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > > > > > > > In certain use cases (where the chip is part of a camera module, and the > > > > > > > camera module is wired together with a camera privacy LED), powering on > > > > > > > the device during probe is undesirable. Add support for the at24 to > > > > > > > execute probe while being powered off. For this to happen, a hint in form > > > > > > > of a device property is required from the firmware. > > > > > > > > > > > > > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> > > > > > > > Reviewed-by: Tomasz Figa <tfiga@chromium.org> > > > > > > > --- > > > > > > > > > > > > I'll ack this but I still claim that the name > > > > > > acpi_dev_state_low_power() is super misleading for this use-case and > > > > > > I've been saying that for 10 versions now with everyone just ignoring > > > > > > my remarks. :/ > > > > > > > > > > Well, the function in question simply checks if the current ACPI power > > > > > state of the device is different from "full power", so its name > > > > > appears to be quite adequate to me. > > > > > > > > > > If the way in which it is used is confusing, though, I guess > > > > > explaining what's going on would be welcome. > > > > > > > > > > > > > Yes, I have explained it multiple time already - last time at v9 of this series: > > > > > > > > https://www.spinics.net/lists/kernel/msg3816807.html > > > > > > How about adding this to the description of acpi_dev_state_low_power(): > > > > > > -----------8<-------------- > > > * This function is intended to be used by drivers to tell whether the device > > > * is in low power state (D1--D3cold) in driver's probe or remove function. See > > > * Documentation/firmware-guide/acpi/low-power-probe.rst for more information. > > > -----------8<-------------- > > > > This information is already there in the kerneldoc description of that > > function AFAICS. > > Ok, the D states are mentioned already. But how to use it is not, nor > there's a reference to the ReST file. I think that wouldn't hurt. > > > > > I was thinking about adding an explanation comment to the caller. > > I think it'd be best if the function name would convey that without a > comment that should then be added to all callers. How about calling the > function e.g. acpi_dev_state_d0() and negating the return value? The D0 > state is well defined and we could do this without adding new terms. That would work for me.
On Tue, Feb 09, 2021 at 05:58:12PM +0100, Rafael J. Wysocki wrote: > On Tue, Feb 9, 2021 at 5:54 PM Sakari Ailus > <sakari.ailus@linux.intel.com> wrote: > > > > On Tue, Feb 09, 2021 at 05:42:45PM +0100, Rafael J. Wysocki wrote: > > > On Tue, Feb 9, 2021 at 5:23 PM Sakari Ailus > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > Hi Bartosz, Rafael, > > > > > > > > On Tue, Feb 09, 2021 at 04:49:37PM +0100, Bartosz Golaszewski wrote: > > > > > On Mon, Feb 8, 2021 at 5:54 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > > > > > > > > > > > On Mon, Feb 8, 2021 at 5:44 PM Bartosz Golaszewski > > > > > > <bgolaszewski@baylibre.com> wrote: > > > > > > > > > > > > > > On Fri, Feb 5, 2021 at 2:25 PM Sakari Ailus > > > > > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > > > > > > > > > In certain use cases (where the chip is part of a camera module, and the > > > > > > > > camera module is wired together with a camera privacy LED), powering on > > > > > > > > the device during probe is undesirable. Add support for the at24 to > > > > > > > > execute probe while being powered off. For this to happen, a hint in form > > > > > > > > of a device property is required from the firmware. > > > > > > > > > > > > > > > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> > > > > > > > > Reviewed-by: Tomasz Figa <tfiga@chromium.org> > > > > > > > > --- > > > > > > > > > > > > > > I'll ack this but I still claim that the name > > > > > > > acpi_dev_state_low_power() is super misleading for this use-case and > > > > > > > I've been saying that for 10 versions now with everyone just ignoring > > > > > > > my remarks. :/ > > > > > > > > > > > > Well, the function in question simply checks if the current ACPI power > > > > > > state of the device is different from "full power", so its name > > > > > > appears to be quite adequate to me. > > > > > > > > > > > > If the way in which it is used is confusing, though, I guess > > > > > > explaining what's going on would be welcome. > > > > > > > > > > > > > > > > Yes, I have explained it multiple time already - last time at v9 of this series: > > > > > > > > > > https://www.spinics.net/lists/kernel/msg3816807.html > > > > > > > > How about adding this to the description of acpi_dev_state_low_power(): > > > > > > > > -----------8<-------------- > > > > * This function is intended to be used by drivers to tell whether the device > > > > * is in low power state (D1--D3cold) in driver's probe or remove function. See > > > > * Documentation/firmware-guide/acpi/low-power-probe.rst for more information. > > > > -----------8<-------------- > > > > > > This information is already there in the kerneldoc description of that > > > function AFAICS. > > > > Ok, the D states are mentioned already. But how to use it is not, nor > > there's a reference to the ReST file. I think that wouldn't hurt. > > > > > > > > I was thinking about adding an explanation comment to the caller. > > > > I think it'd be best if the function name would convey that without a > > comment that should then be added to all callers. How about calling the > > function e.g. acpi_dev_state_d0() and negating the return value? The D0 > > state is well defined and we could do this without adding new terms. > > That would work for me. Bartosz, would that work for you? I'd call the temporary variable in the at24 driver e.g. "full_power".
On Wed, Feb 10, 2021 at 9:41 AM Sakari Ailus <sakari.ailus@linux.intel.com> wrote: > > On Tue, Feb 09, 2021 at 05:58:12PM +0100, Rafael J. Wysocki wrote: > > On Tue, Feb 9, 2021 at 5:54 PM Sakari Ailus > > <sakari.ailus@linux.intel.com> wrote: > > > > > > On Tue, Feb 09, 2021 at 05:42:45PM +0100, Rafael J. Wysocki wrote: > > > > On Tue, Feb 9, 2021 at 5:23 PM Sakari Ailus > > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > > > Hi Bartosz, Rafael, > > > > > > > > > > On Tue, Feb 09, 2021 at 04:49:37PM +0100, Bartosz Golaszewski wrote: > > > > > > On Mon, Feb 8, 2021 at 5:54 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > > > > > > > > > > > > > On Mon, Feb 8, 2021 at 5:44 PM Bartosz Golaszewski > > > > > > > <bgolaszewski@baylibre.com> wrote: > > > > > > > > > > > > > > > > On Fri, Feb 5, 2021 at 2:25 PM Sakari Ailus > > > > > > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > > > > > > > > > > > In certain use cases (where the chip is part of a camera module, and the > > > > > > > > > camera module is wired together with a camera privacy LED), powering on > > > > > > > > > the device during probe is undesirable. Add support for the at24 to > > > > > > > > > execute probe while being powered off. For this to happen, a hint in form > > > > > > > > > of a device property is required from the firmware. > > > > > > > > > > > > > > > > > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> > > > > > > > > > Reviewed-by: Tomasz Figa <tfiga@chromium.org> > > > > > > > > > --- > > > > > > > > > > > > > > > > I'll ack this but I still claim that the name > > > > > > > > acpi_dev_state_low_power() is super misleading for this use-case and > > > > > > > > I've been saying that for 10 versions now with everyone just ignoring > > > > > > > > my remarks. :/ > > > > > > > > > > > > > > Well, the function in question simply checks if the current ACPI power > > > > > > > state of the device is different from "full power", so its name > > > > > > > appears to be quite adequate to me. > > > > > > > > > > > > > > If the way in which it is used is confusing, though, I guess > > > > > > > explaining what's going on would be welcome. > > > > > > > > > > > > > > > > > > > Yes, I have explained it multiple time already - last time at v9 of this series: > > > > > > > > > > > > https://www.spinics.net/lists/kernel/msg3816807.html > > > > > > > > > > How about adding this to the description of acpi_dev_state_low_power(): > > > > > > > > > > -----------8<-------------- > > > > > * This function is intended to be used by drivers to tell whether the device > > > > > * is in low power state (D1--D3cold) in driver's probe or remove function. See > > > > > * Documentation/firmware-guide/acpi/low-power-probe.rst for more information. > > > > > -----------8<-------------- > > > > > > > > This information is already there in the kerneldoc description of that > > > > function AFAICS. > > > > > > Ok, the D states are mentioned already. But how to use it is not, nor > > > there's a reference to the ReST file. I think that wouldn't hurt. > > > > > > > > > > > I was thinking about adding an explanation comment to the caller. > > > > > > I think it'd be best if the function name would convey that without a > > > comment that should then be added to all callers. How about calling the > > > function e.g. acpi_dev_state_d0() and negating the return value? The D0 > > > state is well defined and we could do this without adding new terms. > > > > That would work for me. > > Bartosz, would that work for you? > > I'd call the temporary variable in the at24 driver e.g. "full_power". > Yes, that works, go ahead and thank you. Bartosz
On Wed, Feb 10, 2021 at 01:26:51PM +0100, Bartosz Golaszewski wrote: > On Wed, Feb 10, 2021 at 9:41 AM Sakari Ailus > <sakari.ailus@linux.intel.com> wrote: > > > > On Tue, Feb 09, 2021 at 05:58:12PM +0100, Rafael J. Wysocki wrote: > > > On Tue, Feb 9, 2021 at 5:54 PM Sakari Ailus > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > On Tue, Feb 09, 2021 at 05:42:45PM +0100, Rafael J. Wysocki wrote: > > > > > On Tue, Feb 9, 2021 at 5:23 PM Sakari Ailus > > > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > > > > > Hi Bartosz, Rafael, > > > > > > > > > > > > On Tue, Feb 09, 2021 at 04:49:37PM +0100, Bartosz Golaszewski wrote: > > > > > > > On Mon, Feb 8, 2021 at 5:54 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > > > > > > > > > > > > > > > On Mon, Feb 8, 2021 at 5:44 PM Bartosz Golaszewski > > > > > > > > <bgolaszewski@baylibre.com> wrote: > > > > > > > > > > > > > > > > > > On Fri, Feb 5, 2021 at 2:25 PM Sakari Ailus > > > > > > > > > <sakari.ailus@linux.intel.com> wrote: > > > > > > > > > > > > > > > > > > > > In certain use cases (where the chip is part of a camera module, and the > > > > > > > > > > camera module is wired together with a camera privacy LED), powering on > > > > > > > > > > the device during probe is undesirable. Add support for the at24 to > > > > > > > > > > execute probe while being powered off. For this to happen, a hint in form > > > > > > > > > > of a device property is required from the firmware. > > > > > > > > > > > > > > > > > > > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> > > > > > > > > > > Reviewed-by: Tomasz Figa <tfiga@chromium.org> > > > > > > > > > > --- > > > > > > > > > > > > > > > > > > I'll ack this but I still claim that the name > > > > > > > > > acpi_dev_state_low_power() is super misleading for this use-case and > > > > > > > > > I've been saying that for 10 versions now with everyone just ignoring > > > > > > > > > my remarks. :/ > > > > > > > > > > > > > > > > Well, the function in question simply checks if the current ACPI power > > > > > > > > state of the device is different from "full power", so its name > > > > > > > > appears to be quite adequate to me. > > > > > > > > > > > > > > > > If the way in which it is used is confusing, though, I guess > > > > > > > > explaining what's going on would be welcome. > > > > > > > > > > > > > > > > > > > > > > Yes, I have explained it multiple time already - last time at v9 of this series: > > > > > > > > > > > > > > https://www.spinics.net/lists/kernel/msg3816807.html > > > > > > > > > > > > How about adding this to the description of acpi_dev_state_low_power(): > > > > > > > > > > > > -----------8<-------------- > > > > > > * This function is intended to be used by drivers to tell whether the device > > > > > > * is in low power state (D1--D3cold) in driver's probe or remove function. See > > > > > > * Documentation/firmware-guide/acpi/low-power-probe.rst for more information. > > > > > > -----------8<-------------- > > > > > > > > > > This information is already there in the kerneldoc description of that > > > > > function AFAICS. > > > > > > > > Ok, the D states are mentioned already. But how to use it is not, nor > > > > there's a reference to the ReST file. I think that wouldn't hurt. > > > > > > > > > > > > > > I was thinking about adding an explanation comment to the caller. > > > > > > > > I think it'd be best if the function name would convey that without a > > > > comment that should then be added to all callers. How about calling the > > > > function e.g. acpi_dev_state_d0() and negating the return value? The D0 > > > > state is well defined and we could do this without adding new terms. > > > > > > That would work for me. > > > > Bartosz, would that work for you? > > > > I'd call the temporary variable in the at24 driver e.g. "full_power". > > > > Yes, that works, go ahead and thank you. Thanks! I'll send v11 soon. -- Sakari Ailus
diff --git a/drivers/misc/eeprom/at24.c b/drivers/misc/eeprom/at24.c index 926408b41270c..69a5e4023d9e1 100644 --- a/drivers/misc/eeprom/at24.c +++ b/drivers/misc/eeprom/at24.c @@ -595,6 +595,7 @@ static int at24_probe(struct i2c_client *client) bool i2c_fn_i2c, i2c_fn_block; unsigned int i, num_addresses; struct at24_data *at24; + bool off_during_probe; struct regmap *regmap; bool writable; u8 test_byte; @@ -750,14 +751,16 @@ static int at24_probe(struct i2c_client *client) i2c_set_clientdata(client, at24); - err = regulator_enable(at24->vcc_reg); - if (err) { - dev_err(dev, "Failed to enable vcc regulator\n"); - return err; - } + off_during_probe = acpi_dev_state_low_power(&client->dev); + if (!off_during_probe) { + err = regulator_enable(at24->vcc_reg); + if (err) { + dev_err(dev, "Failed to enable vcc regulator\n"); + return err; + } - /* enable runtime pm */ - pm_runtime_set_active(dev); + pm_runtime_set_active(dev); + } pm_runtime_enable(dev); at24->nvmem = devm_nvmem_register(dev, &nvmem_config); @@ -768,14 +771,17 @@ static int at24_probe(struct i2c_client *client) } /* - * Perform a one-byte test read to verify that the - * chip is functional. + * Perform a one-byte test read to verify that the chip is functional, + * unless powering on the device is to be avoided during probe (i.e. + * it's powered off right now). */ - err = at24_read(at24, 0, &test_byte, 1); - if (err) { - pm_runtime_disable(dev); - regulator_disable(at24->vcc_reg); - return -ENODEV; + if (!off_during_probe) { + err = at24_read(at24, 0, &test_byte, 1); + if (err) { + pm_runtime_disable(dev); + regulator_disable(at24->vcc_reg); + return -ENODEV; + } } pm_runtime_idle(dev); @@ -795,9 +801,11 @@ static int at24_remove(struct i2c_client *client) struct at24_data *at24 = i2c_get_clientdata(client); pm_runtime_disable(&client->dev); - if (!pm_runtime_status_suspended(&client->dev)) - regulator_disable(at24->vcc_reg); - pm_runtime_set_suspended(&client->dev); + if (!acpi_dev_state_low_power(&client->dev)) { + if (!pm_runtime_status_suspended(&client->dev)) + regulator_disable(at24->vcc_reg); + pm_runtime_set_suspended(&client->dev); + } return 0; } @@ -834,6 +842,7 @@ static struct i2c_driver at24_driver = { .probe_new = at24_probe, .remove = at24_remove, .id_table = at24_ids, + .flags = I2C_DRV_FL_ALLOW_LOW_POWER_PROBE, }; static int __init at24_init(void)