]> git.karo-electronics.de Git - mv-sheeva.git/blobdiff - drivers/staging/iio/Documentation/sysfs-bus-iio
Merge branch 'master' into csb1725
[mv-sheeva.git] / drivers / staging / iio / Documentation / sysfs-bus-iio
diff --git a/drivers/staging/iio/Documentation/sysfs-bus-iio b/drivers/staging/iio/Documentation/sysfs-bus-iio
new file mode 100644 (file)
index 0000000..fdb017a
--- /dev/null
@@ -0,0 +1,390 @@
+What:          /sys/bus/iio/devices/device[n]
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Hardware chip or device accessed by on communication port.
+               Corresponds to a grouping of sensor channels.
+
+What:          /sys/bus/iio/devices/trigger[n]
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               An event driven driver of data capture to an in kernel buffer.
+               May be provided by a device driver that also has an IIO device
+               based on hardware generated events (e.g. data ready) or
+               provided by a separate driver for other hardware (e.g.
+               periodic timer, gpio or high resolution timer).
+               Contains trigger type specific elements. These do not
+               generalize well and hence are not documented in this file.
+
+What:          /sys/bus/iio/devices/device[n]:buffer
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Link to /sys/class/iio/device[n]/device[n]:buffer. n indicates
+               the device with which this buffer buffer is associated.
+
+What:          /sys/.../device[n]/name
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Description of the physical chip / device. Typically a part
+               number.
+
+What:          /sys/.../device[n]/sampling_frequency
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Some devices have internal clocks.  This parameter sets the
+               resulting sampling frequency.  In many devices this
+               parameter has an effect on input filters etc rather than
+               simply controlling when the input is sampled.  As this
+               effects datardy triggers, hardware buffers and the sysfs
+               direct access interfaces, it may be found in any of the
+               relevant directories.  If it effects all of the above
+               then it is to be found in the base device directory as here.
+
+What:          /sys/.../device[n]/sampling_frequency_available
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               When the internal sampling clock can only take a small
+               discrete set of values, this file lists those availale.
+
+What:          /sys/.../device[n]/in[m][_name]_raw
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Raw (unscaled no bias removal etc) voltage measurement from
+               channel m. name is used in special cases where this does
+               not correspond to externally available input (e.g. supply
+               voltage monitoring in which case the file is in_supply_raw).
+               If the device supports events on this channel then m must be
+               specified (even on named channels) so as to allow the source
+               of event codes to be identified.
+
+What:          /sys/.../device[n]/in[m][_name]_offset
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               If known for a device, offset to be added to in[m]_raw prior
+               to scaling by in[_name][m]_scale in order to obtain voltage in
+               millivolts.  Not present if the offset is always 0 or unknown.
+               If m is not present, then voltage offset applies to all in
+               channels. May be writable if a variable offset is controlled
+               by the device. Note that this is different to calibbias which
+               is for devices that apply offsets to compensate for variation
+               between different instances of the part, typically adjusted by
+               using some hardware supported calibration procedure.
+
+What:          /sys/.../device[n]/in[m][_name]_offset_available
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               If a small number of discrete offset values are available, this
+               will be a space separated list.  If these are independant (but
+               options the same) for individual offsets then m should not be
+               present.
+
+What:          /sys/.../device[n]/in[m][_name]_offset_[min|max]
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               If a more or less continuous range of voltage offsets are
+               supported then these specify the minimum and maximum.  If shared
+               by all in channels then m is not present.
+
+What:          /sys/.../device[n]/in[m][_name]_calibbias
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Hardware applied calibration offset. (assumed to fix production
+               inaccuracies)
+
+What           /sys/.../device[n]/in[m][_name]_calibscale
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Hardware applied calibration scale factor. (assumed to fix
+               production inaccuracies)
+
+What:          /sys/.../device[n]/in[m][_name]_scale
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               If known for a device, scale to be applied to volt[m]_raw post
+               addition of in[_name][m]_offset in order to obtain the measured
+               voltage in millivolts.  If shared across all in channels then
+               m is not present.
+
+What:          /sys/.../device[n]/in[m]-in[o]_raw
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Raw (unscaled) differential voltage measurement equivalent to
+               channel m - channel o where these channel numbers apply to the
+               physically equivalent inputs when non differential readings are
+               separately available. In differential only parts, then all that
+               is required is a consistent labelling.
+
+What:          /sys/.../device[n]/accel[_x|_y|_z][m]_raw
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Acceleration in direction x, y or z (may be arbitrarily assigned
+               but should match other such assignments on device)
+               channel m (not present if only one accelerometer channel at
+               this orientation). Has all of the equivalent parameters as per
+               in[m]. Units after application of scale and offset are m/s^2.
+
+What:          /sys/.../device[n]/gyro[_x|_y|_z][m]_raw
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Angular velocity about axis x, y or z (may be arbitrarily
+               assigned) channel m (not present if only one gyroscope at
+               this orientation).
+               Data converted by application of offset then scale to
+               radians per second. Has all the equivalent parameters as
+               per in[m].
+
+What:          /sys/.../device[n]/incli[_x|_y|_z][m]_raw
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Inclination raw reading about axis x, y or z (may be arbitarily
+               assigned) channel m (not present if only one inclinometer at
+               this orientation).  Data converted by application of offset
+               and scale to Degrees.
+
+What:          /sys/.../device[n]/magn[_x|_y|_z][m]_raw
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Magnetic field along axis x, y or z (may be arbitrarily
+               assigned) channel m (not present if only one magnetometer
+               at this orientation).  Data converted by application of
+               offset then scale to Gauss. Has all the equivalent modifiers
+               as per in[m].
+
+What:          /sys/.../device[n]/device[n]:event[m]
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Configuration of which hardware generated events are passed up to
+               userspace. Some of these are a bit complex to generalize so this
+               section is a work in progress.
+
+What:          /sys/.../device[n]:event[m]/dev
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               major:minor character device numbers for the event line.
+
+Taking accel_x0 as an example
+
+What:          /sys/.../device[n]:event[m]/accel_x0_thresh[_rising|_falling]_en
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Event generated when accel_x0 passes a threshold in the specfied
+               (_rising|_falling) direction. If the direction is not specified,
+               then either the device will report an event which ever direction
+               a single threshold value is called in (e.g.
+               accel_x0_<raw|input>_thresh_value) or
+               accel_x0_<raw|input>_thresh_rising_value and
+               accel_x0_<raw|input>_thresh_falling_value may take different
+               values, but the device can only enable both thresholds or
+               neither.
+               Note the driver will assume the last p events requested are
+               to be enabled where p is however many it supports (which may
+               vary depending on the exact set requested. So if you want to be
+               sure you have set what you think you have, check the contents of
+               these attributes after everything is configured. Drivers may
+               have to buffer any parameters so that they are consistent when
+               a given event type is enabled a future point (and not those for
+               whatever event was previously enabled).
+
+What:          /sys/.../accel_x0_<raw|input>_thresh[_rising|_falling]_value
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Specifies the value of threshold that the device is comparing
+               against for the events enabled by
+               accel_x0_<raw|input>_thresh[_rising|falling]_en.
+               If seperate exist for the two directions, but direction is
+               not specified for this attribute, then a single threshold value
+               applies to both directions.
+               The raw or input element of the name indicates whether the
+               value is in raw device units or in processed units (as _raw
+               and _input do on sysfs direct channel read attributes).
+
+What:          /sys/.../accel_x0_thresh[_rising|_falling]_meanperiod
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Period of time (in seconds) over which the raw channel value
+               is averaged before being compared with the threshold set in
+               accel_x0_thresh[_rising|_falling]_meanperiod.  If direction is
+               not specified then this mean period applies to both directions.
+
+What:          /sys/.../accel_x0_thresh[_rising|_falling]_period
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Period of time (in seconds) for which the threshold must be
+               passed before an event is generated. If direction is not
+               specified then this period applies to both directions.
+
+What:          /sys/.../device[n]:event[m]/accel_x0_mag[_rising|_falling]_en
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Similar to accel_x0_thresh[_rising|_falling]_en, but here the
+               magnitude of the channel is compared to the threshold, not its
+               signed value.
+
+What:          /sys/.../accel_x0_<raw|input>_mag[_rising|_falling]_value
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               The value to which the magnitude of the channel is compared.
+
+What:          /sys/.../accel_x0_mag[_rising|_falling]_meanperiod
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Period of time (in seconds) over which the value of the channel
+               is averaged before being compared to the threshold
+
+What:          /sys/.../accel_x0_mag[_rising|_falling]_period
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Period of time (in seconds) for which the condition must be true
+               before an event occurs.
+
+What:          /sys/.../device[n]:event[m]/accel_x0_roc[_rising|_falling]_en
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Similar to accel_x0_thresh[_rising|_falling]_en, but here the
+               first differential is compared with the threshold.
+
+What:          /sys/.../accel_x0_<raw|input>_roc[_rising|_falling]_value
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               The value to which the first differential of the channel is
+               compared.
+
+What:          /sys/.../accel_x0_roc[_rising|_falling]_meanperiod
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Period of time (in seconds) over which the value of the channel
+               is averaged before being compared to the threshold
+
+What:          /sys/.../accel_x0_roc[_rising|_falling]_period
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Period of time (in seconds) for which the condition must be true
+               before an event occurs.
+
+What:          /sys/.../device[n]/device[n]:buffer:event/dev
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Buffer for device n event character device major:minor numbers.
+
+What:          /sys/.../device[n]/device[n]:buffer:access/dev
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Buffer for device n access character device o major:minor numbers.
+
+What:          /sys/.../device[n]:buffer/trigger
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               The name of the trigger source being used, as per string given
+               in /sys/class/iio/trigger[n]/name.
+
+What:          /sys/.../device[n]:buffer/length
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Number of scans contained by the buffer.
+
+What:          /sys/.../device[n]:buffer/bytes_per_datum
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Bytes per scan.  Due to alignment fun, the scan may be larger
+               than implied directly by the scan_element parameters.
+
+What:          /sys/.../device[n]:buffer/enable
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Actually start the buffer capture up.  Will start trigger
+               if first device and appropriate.
+
+What:          /sys/.../device[n]:buffer/alignment
+KernelVersion: 2.6.35
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Minimum data alignment.  Scan elements larger than this are
+               aligned to the nearest power of 2 times this.  (may not be
+               true in weird hardware buffers that pack data well)
+
+What:          /sys/.../device[n]/buffer/scan_elements
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Directory containing interfaces for elements that will be
+               captured for a single triggered sample set in the buffer.
+
+What:          /sys/.../device[n]/buffer/scan_elements/accel_x0_en
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Scan element control for triggered data capture.
+
+What:          /sys/.../device[n]/buffer/scan_elements/accel[_x0]_type
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               Description of the scan element data storage within the buffer
+               and hence the form in which it is read from userspace.
+               Form is [s|u]bits/storagebits.  s or u specifies if signed
+               (2's complement) or unsigned. bits is the number of bits of
+               data and storagebits is the space (after padding) that it
+               occupies in the buffer.  Note that some devices will have
+               additional information in the unused bits so to get a clean
+               value, the bits value must be used to mask the buffer output
+               value appropriately.  The storagebits value also specifies the
+               data alignment.  So s48/64 will be a signed 48 bit integer
+               stored in a 64 bit location aligned to a a64 bit boundary.
+               For other storage combinations this attribute will be extended
+               appropriately.
+
+What:          /sys/.../device[n]/buffer/scan_elements/accel[_x0]_index
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               A single positive integer specifying the position of this
+               scan element in the buffer. Note these are not dependant on
+               what is enabled and may not be contiguous. Thus for userspace
+               to establish the full layout these must be used in conjunction
+               with all _en attributes to establish which channels are present,
+               and the relevant _type attributes to establish the data storage
+               format.
+
+What:          /sys/.../device[n]/buffer/scan_elements/accel[_x0]_shift
+KernelVersion: 2.6.37
+Contact:       linux-iio@vger.kernel.org
+Description:
+               A bit shift (to right) that must be applied prior to
+               extracting the bits specified by accel[_x0]_precision.