IIO new drivers, features and rework for the 3.7 cycle, 4th set.
Here we have
1) a set cleaning up and moving the ad7476 driver out of staging.
Support for a number of additional parts is also added to that driver.
2) cleanups from various people for the in kernel interface code as that
is getting more an more real use and hence people are picking up on
minor issues that made it through review. Also a related useful set
of utility functions to avoid duplicate code for converting IIO
representations to other forms.
3) a new fractional type for our read_raw / write_raw functions.
This allows avoiding loss of accuracy via the in kernel interfaces in some
cases as well as being rather convenient for a lot of range -> scale
conversions.
4) New AD5755 DAC driver.
5) Some Blackfin timer trigger improvements including hardware pulse control
for device triggering.
6) Support for the ad7091r in the ad7476 driver.
Some converters require an external signal to start the conversion. This patch
adds support to the bfintmr trigger driver to generate such a signal.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
This patch hooks up the set_trigger_state callback for the blackfin timer
trigger driver and only enables the timer when a trigger consumer requests it to
be enabled. There really is no reason to keep the timer running and generate
interrupts if nobody is listening to them.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
If the timer frequency has not been configured yet get_gptimer_period() will
return 0. Handle this case instead of blindly dividing by the returned value.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Use the iio_read_channel_processed function to read the sample value in the
proper unit instead of using iio_read_channel_raw and iio_read_channel_scale and
doing the unit conversion manually.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
This pulls in the staging tree fixes in 3.6-rc6 into our branch to resolve the
merge issues.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit- 314be14bb renamed the _st_ functions to loose the bit
that was meant for staging version but forgot to change
the documentation which still have _st_ sprinkled in some of the
places.
Signed-off-by: anish kumar <anish198519851985@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The ad7476 driver is a driver for simple single channel ADCs. The driver does
not export any experimental or custom ABI files nor do the static code check
tools report any issues, so move the driver out of staging.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
After the recent cleanups the buffer support is just a single 23 line function.
This does not really justify a file on its own, so move it to the main driver
file. And with only one source file left the header file containing the device
state struct becomes superflousious so move the content of the header
file to the main driver source file as well.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Slightly rework the reference voltage handling for the ad7476 driver. Now the only
way to specify a external reference voltage is to use the regulator API,
previously it was possible to use either platform_data or the regulator API. The
new way is more consistent with what other drivers do.
Also do not ignore errors when requesting the regulator, since this will cope
very poorly with e.g. deferred probing.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The ad7476 driver has only support for 1 channel ADCs. So the upper limit for
the buffer size is the size of one sample plus the size of the timestamp.
Preallocate a buffer large enough to hold this to avoid having to allocate and
free a new buffer for each sample being captured.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Some of the parts supported by this driver are software compatible. The
difference between them is only in the operating voltage range. So we do not
need extra chip info entries for them.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The datasheet is a bit confusing about this. It says that a dataword has 4
leading zeros, but the first zero is already put on the bus when CS is pulled
low and the second zero is put on the bus on the first leading edge of SCLK, so
when the first bit is sampled on the first trailing edge it will sample what the
datasheet refers to as the second leading zero. Subsequently we only see 3
leading zeros in the 16 bit dataword and the result we get is shifted to the
left by one bit. Fix this by adjusting the channel shift by 1.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The Industrial IO framework supports scaling ADC values by fractions,
but most drivers default to using whole numbers.
This change turns on fractional scaling in the isl29018 driver.
Signed-off-by: Bryan Freed <bfreed@chromium.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Drop timestamp parameter from buffer store_to callback and subsequently from
iio_push_to_buffer. The timestamp parameter is unused and it seems likely that
it will stay unused in the future, so it should be safe to remove it.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Consistently use iio_push_to_buffer instead of manually calling the buffers
store_to callback.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Since iio_trigger_poll() calls generic_handle_irq() it need to be called from
hardirq context. The sysfs trigger is kicked from userspace, so it is obviously
not possible to fulfill this requirement by calling iio_trigger_poll directly.
As a workaround commit 1f785681 ("staging:iio:trigger sysfs userspace trigger
rework.") added iio_trigger_poll_chained() which uses handle_nested_irq instead
of generic_handle_irq. This in itself is a hack and only works by chance.
handle_nested_irq is intended to be called from the threaded interrupt handler
of the parent IRQ. Using handle_nested_irq is also problematic since it will
only call the threaded handler of the IRQ. But quite a few IIO drivers rely on
their hardirq handler being called or undefined behaviour will occur.
This patch uses the irq_work framework to schedule the call to
iio_trigger_poll() from hardirq context, which fixes the issues described above.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Clk_prepare_enable and clk_disable_unprepare combine clk_prepare and
clk_enable, and clk_disable and clk_unprepare. They make the code more
concise, and ensure that clk_unprepare is called when clk_enable fails.
A simplified version of the semantic patch that introduces calls to these
functions is as follows: (http://coccinelle.lip6.fr/)
// <smpl>
@@
expression e;
@@
- clk_prepare(e);
- clk_enable(e);
+ clk_prepare_enable(e);
@@
expression e;
@@
- clk_disable(e);
- clk_unprepare(e);
+ clk_disable_unprepare(e);
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
"val" is used as a divisor later, so we should check for zero here to
avoid a division by zero.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The ad7785 is similar to the ad7792/ad7793, but has 20 bit wide data samples.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The ad7794/ad7795 are similar to the ad7792/ad7793, but have 6 channels instead
of 3.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Convert the ad7192 driver to make use of the new common code for devices from
the Analog Devices Sigma Delta family.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Convert the ad7793 driver to make use of the new common code for devices from
the Analog Devices Sigma Delta family.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Convert the ad7780 driver to make use of the new common code for devices from
the Analog Devices Sigma Delta family.
As a bonus the ad7780 driver gains support for buffered mode. Although this is a
bit tricky. The ad7780 reports in the lower 4 unused bits of the data word the
internal gain used. The driver will update the scale attribute value depending
on the gain accordingly, but obviously this will only work if the gain does not
change while sampling. This is not perfect, but since we store the raw value in
the buffer an application which is aware of this can extract the gain factor
from the buffer as well an apply it accordingly.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The platform data for the device is only used from within the drivers probe
callback, so there is no need to keep it around in the devices state struct.
While we are at it mark the platform data struct as const.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The platform data for the device is only used from within the drivers probe
callback, so there is no need to keep it around in the devices state struct.
While we are at it mark the platform data struct as const.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The values reported by the AD7780 are unsigned with a binary offset:
0x000000 is negative fullscale
0x800000 is zeroscale
0xffffff is positive fullscale
So mark the channel in the channel spec as unsigned rather than signed.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The temperature channel reports values in degree Kelvin with sensitivity of 5630
codes per degree. If the chip is configured in bipolar mode there is an
additional binary offset of 0x800000 and the sensitivity is divided by two.
Currently the driver does the mapping from the raw value to degree Celsius when
doing a manual conversion. This has several disadvantages, the major one being
that it does not work for buffered mode, also by doing the division by the
sensitivity in the driver the precession of the reported value is needlessly
reduced.
Furthermore the current calculation only works in bipolar mode and the current
scale is of by a factor of 1000.
This patch modifies the driver to report correct offset and scale values in
both unipolar and bipolar mode and to report the raw temperature value
for manual conversions.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
In bipolar mode there is a a binary offset of 2**(N-1) (with N being the number
of bits) on the reported value. Currently this value is subtracted when doing a
manual read. While this works for manual channel readings it does not work for
buffered mode. So report the offset in the channels offset property, which will
work in both modes.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
The values reported by the AD7793 are unsigned.
In uniploar mode:
0x000000 is zeroscale
0xffffff is fullscale
In bipolar mode:
0x000000 is negative fullscale
0x800000 is zeroscale
0xffffff is positive fullscale
In bipolar mode there is a binary offset, but the values are still unsigned.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Write to the correct register when setting the ACX bit.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Without the break statement we fall right through to the default case and return
an error value.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The internal reference for the ad7793 and similar is 1.17V
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Make the "in-in_scale_available" attribute follow the new naming spec and
rename it to "in_voltage-voltage_scale_available".
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The temperature channel uses the internal 1.17V reference with 0.81 mv/C. The
reported temperature is in Kevlin, so we need to add the Kelvin to Celcius
offset when reporting the offset for the temperature channel.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
In bipolar mode there is a a binary offset of 2**(N-1) (with N being the number
of bits) on the reported value. Currently this value is subtracted when doing a
manual read. While this works for manual channel readings it does not work for
buffered mode. So report the offset in the channels offset property, which will
work in both modes.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The values reported by the AD7793 are unsigned.
In uniploar mode:
0x000000 is zeroscale
0xffffff is fullscale
In bipolar mode:
0x000000 is negative fullscale
0x800000 is zeroscale
0xffffff is positive fullscale
In bipolar mode there is a binary offset, but the values are still unsigned.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Without the break statement we fall right through to the default case and return
an error value.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
drivers/staging/iio/adc/ad7298_ring.c:97:37: warning: 'time_ns' may
be used uninitialized in this function [-Wmaybe-uninitialized]
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
the code which under _init and _exit does only the platform_driver_register
and platform_driver_unregister, and nothing else,
so its better to use the module_platform_driver macro rather duplicating
its implementation
Signed-off-by: Devendra Naga <develkernel412222@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
The per driver iio_chan_spec arrays are usually shared between multiple device
instances. So a single device instance may not modify the iio_chan_spec array
since this would also affect the other device instances. To make this restriction
explicit mark the per driver iio_chan_spec arrays as const.
Conversion was done automatically using the following coccinelle semantic patch:
// <smpl>
@disable optional_qualifier@
identifier channels;
@@
static
+const
struct iio_chan_spec channels[] = ...;
// </smpl>
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
