This is a bit involved as the adis library code already has some
sanity checking of the flags of the requested irq that we need
to ensure is happy to pass through the IRQF_NO_AUTOEN flag untouched.
Using this flag avoids us autoenabling the irq in the adis16460 and
adis16475 drivers which cover parts that don't have any means of
masking the interrupt on the device end.
Note, compile tested only!
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Alexandru Ardelean <ardeleanalex@gmail.com>
Reviewed-by: Nuno Sa <nuno.sa@analog.com>
Reviewed-by: Barry Song <song.bao.hua@hisilicon.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Lars-Peter Clausen <lars@metafoo.de>
Link: https://lore.kernel.org/r/20210402184544.488862-7-jic23@kernel.org
Use the adis managed device functions to setup the buffer and the trigger.
The ultimate goal will be to completely drop the non devm version from
the lib.
Since we are here, drop the `.remove` callback by further using devm
functions.
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20200915120258.161587-7-nuno.sa@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
This patch applies the semantic patch:
@@
expression I, P, SP;
@@
I = devm_iio_device_alloc(P, SP);
...
- I->dev.parent = P;
It updates 302 files and does 307 deletions.
This semantic patch also removes some comments like
'/* Establish that the iio_dev is a child of the i2c device */'
But this is is only done in case where the block is left empty.
The patch does not seem to cover all cases. It looks like in some cases a
different variable is used in some cases to assign the parent, but it
points to the same reference.
In other cases, the block covered by ... may be just too big to be covered
by the semantic patch.
However, this looks pretty good as well, as it does cover a big bulk of the
drivers that should remove the parent assignment.
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The IIO core provides a iio_get_debugfs_dentry() helper.
It seems that the ADIS IMU drivers access that field directly.
This change converts them to use iio_get_debugfs_dentry() instead.
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Acked-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
debugfs_create_file_unsafe does not protect the fops handed to it
against file removal. DEFINE_DEBUGFS_ATTRIBUTE makes the fops aware of
the file lifetime and thus protects it against removal.
Signed-off-by: Rohit Sarkar <rohitsarkar5398@gmail.com>
Tested-by Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
All of the actions done in `adis16460_initial_setup()` are now done in
`__adis_initial_startup()` so, there's no need for code duplication.
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
This patch adds a dedicated self_test_reg variable. This is also a step
to let new drivers make use of `adis_initial_startup()`. Some devices
use MSG_CTRL reg to request a self_test command while others use the
GLOB_CMD register.
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The adis library only allows to define a `startup_delay` which for some
devices is enough. However, other devices define different timeouts with
significantly different timings which could lead to devices to not wait
enough time or to wait a lot more than necessary (which is not
efficient). This patch introduces a new timeout struct that must be
passed into `adis_init()`. There are mainly, for now, three timeouts
used. This is also an introductory patch with the goal of refactoring
`adis_initial_startup()`. New driver's (eg: adis16480, adis16460) are
replicating code for the device initial setup. With some changes (being
this the first one) we can pass this to `adis_initial_startup()`.
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The ADIS library functions return zero on success, and negative values for
error. Positive values aren't returned, but we only care about the success
value (which is zero).
This change is mostly needed so that the compiler won't make any inferences
about some about values being potentially un-initialized. This only
triggers after making some functions inline, because the compiler can
better follow return paths.
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Caught via static-analysis checker:
```
drivers/iio/imu/adis16460.c
152 static int adis16460_set_freq(struct iio_dev *indio_dev, int val, int val2)
153 {
154 struct adis16460 *st = iio_priv(indio_dev);
155 unsigned int t;
^^^^^^^^^^^^^^
156
157 t = val * 1000 + val2 / 1000;
158 if (t <= 0)
^^^^^^
Unsigned is not less than zero.
```
The types of `val` && `val2` are obtained from the IIO `write_raw` hook, so
userspace can provide negative values, which can cause weird behavior after
conversion to unsigned.
This patch changes the sign of variable `t` so that -EINVAL will be
returned for negative values as well.
Fixes: db6ed4d23d ("iio: imu: Add support for the ADIS16460 IMU")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The ADIS16460 device is a complete inertial system that includes a triaxial
gyroscope and a triaxial accelerometer. It's more simplified design than
that of the ADIS16480, and does not offer the triaxial magnetometers &
pressure sensors. It does also have a temperature sensor (like the
ADIS16480).
Since it is part of the ADIS16XXX family, it re-uses parts of the ADIS
library.
Naturally, the register map is different and much more simplified than the
ADIS16480 subfamily, so it cannot be integrated into that driver. A major
difference is that the registers are not paged.
One thing that is particularly special about it, is that it requires a
higher delay between CS changes (i.e. when CS goes up, the spec recommends
that it be brought down after a minimum of 16 uS).
Other ADIS chips require (via spec) a minimum of 2 uS between CS changes.
The kernel's 10 uS default should be fine for those other chips; they
haven't been tested with lower CS change delays yet.
Datasheet:
https://www.analog.com/media/en/technical-documentation/data-sheets/adis16460.pdf
Signed-off-by: Dragos Bogdan <dragos.bogdan@analog.com>
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
