| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Wed, 3 Aug 2022 21:18:26 +0200
From: Christophe JAILLET <christophe.jaillet@...adoo.fr>
To: ddrokosov@...rdevices.ru
Cc: andy.shevchenko@...il.com, devicetree@...r.kernel.org,
jic23@...nel.org, kernel@...rdevices.ru, lars@...afoo.de,
linux-iio@...r.kernel.org, linux-kernel@...r.kernel.org,
robh+dt@...nel.org, shawnguo@...nel.org, stano.jakubek@...il.com,
stephan@...hold.net
Subject: Re: [PATCH v4 2/3] iio: add MEMSensing MSA311 3-axis accelerometer
driver
Le 03/08/2022 à 20:39, Dmitry Rokosov a écrit :
> Hello Christophe,
>
> Thank you for quick review
>
> On Wed, Aug 03, 2022 at 08:11:05PM +0200, Christophe JAILLET wrote:
>> Le 03/08/2022 à 15:11, Dmitry Rokosov a écrit :
>>> MSA311 is a tri-axial, low-g accelerometer with I2C digital output for
>>> sensitivity consumer applications. It has dynamic user-selectable full
>>> scales range of +-2g/+-4g/+-8g/+-16g and allows acceleration measurements
>>> with output data rates from 1Hz to 1000Hz.
>>>
>>> Spec: https://cdn-shop.adafruit.com/product-files/5309/MSA311-V1.1-ENG.pdf
>>>
>>> This driver supports following MSA311 features:
>>> - IIO interface
>>> - Different power modes: NORMAL and SUSPEND (using pm_runtime)
>>> - ODR (Output Data Rate) selection
>>> - Scale and samp_freq selection
>>> - IIO triggered buffer, IIO reg access
>>> - NEW_DATA interrupt + trigger
>>>
>>> Below features to be done:
>>> - Motion Events: ACTIVE, TAP, ORIENT, FREEFALL
>>> - Low Power mode
>>>
>>> Signed-off-by: Dmitry Rokosov <ddrokosov-i4r8oA+eLlH99rHkP+FxIw@...lic.gmane.org>
>>> ---
>>> MAINTAINERS | 6 +
>>> drivers/iio/accel/Kconfig | 13 +
>>> drivers/iio/accel/Makefile | 2 +
>>> drivers/iio/accel/msa311.c | 1323 ++++++++++++++++++++++++++++++++++++
>>> 4 files changed, 1344 insertions(+)
>>> create mode 100644 drivers/iio/accel/msa311.c
>>>
>>
>> Hi,
>> a few nits below.
>>
>> [...]
>>
>>> +static int msa311_check_partid(struct msa311_priv *msa311)
>>> +{
>>> + struct device *dev = msa311->dev;
>>> + unsigned int partid;
>>> + int err;
>>> +
>>> + err = regmap_read(msa311->regs, MSA311_PARTID_REG, &partid);
>>> + if (err)
>>> + return dev_err_probe(dev, err,
>>> + "failed to read partid (%d)\n", err);
>>
>> No need for "(%d)" and err.
>>
>
> Do you mean for all dev_err() calls? I think sometimes it's helpful to
> know the actual error value got from external API, isn't? Could you please
> explain your point if possible?
>
No, my comment is only for dev_err_probe() function.
Having ret for dev_err() calls is fine.
See: https://elixir.bootlin.com/linux/v5.19/source/drivers/base/core.c#L4732
dev_err_probe() already has a "error %pe:..., ERR_PTR(err)"
This means that if ret = -ENOMEM:
"(%d)" --> "(-12)"
"error %pe:" --> "error -ENOMEM:"
So there is no real need to duplicate the error code in the message
itself, it is already displayed in a human readable manner.
What your code does would result in:
"error -ENOMEM: failed to read partid (-12)\n"
>>> +
>>> + if (partid == MSA311_WHO_AM_I)
>>> + dev_dbg(dev, "found MSA311 compatible chip[%#x]\n", partid);
>>> + else
>>> + dev_warn(dev, "invalid partid (%#x), expected (%#x)\n",
>>> + partid, MSA311_WHO_AM_I);
>>> +
>>> + return 0;
>>> +}
>>
>> [...]
>>
>>> +static int msa311_probe(struct i2c_client *i2c)
>>> +{
>>> + struct device *dev = &i2c->dev;
>>> + struct msa311_priv *msa311;
>>> + struct iio_dev *indio_dev;
>>> + int err;
>>> +
>>> + indio_dev = devm_iio_device_alloc(dev, sizeof(*msa311));
>>> + if (!indio_dev)
>>> + return dev_err_probe(dev, -ENOMEM,
>>> + "iio device allocation failed\n");
>>> +
>>> + msa311 = iio_priv(indio_dev);
>>> + msa311->dev = dev;
>>> + i2c_set_clientdata(i2c, indio_dev);
>>> +
>>> + err = msa311_regmap_init(msa311);
>>> + if (err)
>>> + return err;
>>> +
>>> + mutex_init(&msa311->lock);
>>> +
>>> + msa311->vdd = devm_regulator_get_optional(dev, "vdd");
>>> + if (IS_ERR(msa311->vdd)) {
>>> + err = PTR_ERR(msa311->vdd);
>>> + if (err == -ENODEV)
>>> + msa311->vdd = NULL;
>>> + else
>>> + return dev_err_probe(dev, PTR_ERR(msa311->vdd),
>>> + "cannot get vdd supply\n");
>>> + }
>>> +
>>> + if (msa311->vdd) {
>>> + err = regulator_enable(msa311->vdd);
>>> + if (err)
>>> + return dev_err_probe(dev, err,
>>> + "cannot enable vdd supply\n");
>>> +
>>> + err = devm_add_action_or_reset(dev, msa311_vdd_disable,
>>> + msa311->vdd);
>>> + if (err) {
>>> + regulator_disable(msa311->vdd);
>>
>> Double regulator_disable(), because of the _or_reset()?
>>
>
> Yep. If devm_add_action_or_reset() returns an error, we will not
> call regulator_disable() by devm subsystem. It means, we have to
> call it directly.
No.
See
https://elixir.bootlin.com/linux/v5.19/source/include/linux/device.h#L249
If devm_add_action_or_reset() fails, "action" is called. This is the
meaning of the _or_reset suffix.
So here, msa311_vdd_disable() would be called and this function is:
+static void msa311_vdd_disable(void *vdd)
+{
+ regulator_disable(vdd);
+}
and "vdd" will be the value of "msa311->vdd"
So, unless I missed something obvious, your code will call twice
regulator_disable(msa311->vdd).
One in devm_add_action_or_reset() and one explicitly after the "if (err)"
Hoping I'm clear and that I didn't miss something obvious.
CJ
>
>>> + return dev_err_probe(dev, err,
>>> + "cannot add vdd disable action\n");
>>> + }
>>> + }
>>> +
>>> + err = msa311_check_partid(msa311);
>>> + if (err)
>>> + return err;
>>> +
>>> + err = msa311_soft_reset(msa311);
>>> + if (err)
>>> + return err;
>>> +
>>> + err = msa311_set_pwr_mode(msa311, MSA311_PWR_MODE_NORMAL);
>>> + if (err)
>>> + return dev_err_probe(dev, err,
>>> + "failed to power on device (%d)\n", err);
>>
>> No need for "(%d)" and err
>
> Asked for the clarification above.
>
>>
>> CJ
>
Powered by blists - more mailing lists