- 19 Jan, 2020 1 commit
-
-
Greg Kroah-Hartman authored
Merge tag 'iio-for-5.6b' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-next Jonathan writes: Second set of new device support, features and minor fixes for IIO in the 5.6 cycle Just a small set this time. As we are very near the merge window, I've rolled a few fixes in here rather than adding noise just before release. A short delay here will do little harm. New device support * adis16480 - Add support for adis16490. After earlier rework this is simple ID plus chip info. Features * kxcjk1013 - mount matrix support. * lsm_6dsx - mount matrix support. Cleanups / minor or late breaking fixes * ad7124 - add support to ad-sigma-delta and use it in this driver to allow the the interrupt type to be IRQF_TRIGGER_LOW unlike most other devices using this framework. * adis - use delay structure now available in SPI to handle transfer delays - introduce a timeouts structure to allow support of new devices * ak8975 - drop platform data support. No one is using it and it adds complexity. - use device_get_match_data rather than open coding much the same thing. * dht11 - drop meaningless todo * at91-samad2_adc - switch to dma_request_chan * altas-sensor - add a helper function to compute number of channels. Needed for new device support that is under review. * bma400 - add a lower bound check on scale. * inv_mpu6050 - add support for temperature data in the fifos for all chips. - support an odd situation where a board supports only interrupt triggering on both edges. * st_lsm6dsx - check and handle potential error return. * st_sensors - fix some values for the LSM9DS0 which is ever so slightly different from other devices using the same whoami value. - switch over to generic functions from dt ones, avoiding need for separate ACPI support. * stm32-adc - switch to dma_request_chan - suppress an error print in deferred probe case. * stm32-dac - drop private data structure element for reset controller as only used in probe. - reflect more cleanly that the reset controller is optional whilst ensuring that if is specified any errors are caught. * stm32-dfsdm - switch to dma_request_chan - fix missing application of formatting to single conversions. - ensure the sampling rate is updated when the oversampling ratio is changed. * tag 'iio-for-5.6b' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio: (29 commits) iio: dac: stm32-dac: better handle reset controller failures iio: dac: stm32-dac: use reset controller only at probe time dt-bindings: iio: accel: kxcjk1013: Document mount-matrix property iio: accel: kxcjk1013: Support orientation matrix iio: imu: st_lsm6dsx: add mount matrix support iio: adc: stm32-adc: don't print an error on probe deferral dt-bindings: iio: adis16480: add compatible entry for ADIS16490 iio: imu: adis16480: Add support for ADIS16490 iio: accel: bma400: prevent setting accel scale too low iio: imu/mpu6050: support dual-edge IRQ iio: imu: inv_mpu6050: add fifo temperature data support iio: magnetometer: ak8975: Convert to use device_get_match_data() iio: magnetometer: ak8975: Get rid of platform data iio: adc: ad7124: Set IRQ type to falling iio: adc: ad-sigma-delta: Allow custom IRQ flags iio: imu: adis: use new `delay` structure for SPI transfer delays iio: adc: stm32-dfsdm: adapt sampling rate to oversampling ratio iio: adc: stm32-dfsdm: fix single conversion iio: st_sensors: Make use of device properties iio: st_sensors: Drop redundant parameter from st_sensors_of_name_probe() ...
-
- 18 Jan, 2020 15 commits
-
-
Etienne Carriere authored
Use devm_reset_control_get_optional_exclusive() instead of devm_reset_control_get_exclusive() as reset controller is optional. Nevertheless if reset controller is expected but reports an error, propagate the error code to the caller. In such case a nice error trace is emitted unless we're deferring the probe operation. Signed-off-by:
Etienne Carriere <etienne.carriere@st.com> Signed-off-by:
Fabrice Gasnier <fabrice.gasnier@st.com> Signed-off-by:
Jonathan Cameron <Jonathan.Cameron@huawei.com>
-
Etienne Carriere authored
This change removes the reset controller reference from the local DAC instance since it is used only at probe time. Signed-off-by:
-
Dmitry Osipenko authored
The generic IIO mount-matrix property conveys physical orientation of the hardware chip. Signed-off-by:
Dmitry Osipenko <digetx@gmail.com> Acked-by:
Rob Herring <robh@kernel.org> Signed-off-by:
-
Dmitry Osipenko authored
Hardware could be physically mounted in any possible direction and userpspace needs to be aware of the mounting orientation in order to process sensor's data correctly. In particular this helps iio-sensor-proxy to report display's orientation properly on a phone/tablet devices. Signed-off-by:
Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com> Signed-off-by:
-
Martin Kepplinger authored
Allow to read the mount-matrix device tree property and provide the mount_matrix file for userspace to read. Signed-off-by:
Martin Kepplinger <martin.kepplinger@puri.sm> Signed-off-by:
-
Etienne Carriere authored
Do not print an error trace when deferring probe for some resource. Signed-off-by:
-
Alexandru Ardelean authored
This change adds an entry for ADIS16490 in the list of compatible devices defined in the dt-bindings of the adis16480 driver. Signed-off-by:
Alexandru Ardelean <alexandru.ardelean@analog.com> Signed-off-by:
-
Stefan Popa authored
The ADIS16490 is part of the same family with ADIS16495 and ADIS16497, the main difference is the temperature, accelerometer and gyroscope scales. Datasheet: Link: https://www.analog.com/media/en/technical-documentation/data-sheets/adis16490.pdfSigned-off-by:
Stefan Popa <stefan.popa@analog.com> Signed-off-by:
-
Dan Carpenter authored
This puts an upper bound on "val2" but it also needs to have a lower bound (BMA400_SCALE_MIN). Fixes: 465c811f ("iio: accel: Add driver for the BMA400") Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Michał Mirosław authored
Make mpu6050 usable on platforms which provide only any-edge interrupts. One example of this kind of platform is AT91SAM9G45 Signed-off-by:
Michał Mirosław <mirq-linux@rere.qmqm.pl> Acked-by:
Jean-Baptiste Maneyrol <jmaneyrol@invensense.com> Signed-off-by:
-
Jean-Baptiste Maneyrol authored
Add support of temperature data in fifo for all chips. Enable unification of scan elements for icm20602. Add macros for generating scan elements. Signed-off-by:
-
Andy Shevchenko authored
Convert to use device_get_match_data() instead of open coded variant. Signed-off-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by:
-
Andy Shevchenko authored
Since IIO framework supports device property API and driver has been moved already to the use of GPIO descriptors the logical continuation is to get rid of platform data completely. We are on the safe side here since there are no users of it in the kernel. Signed-off-by:
-
Alexandru Tachici authored
Ad7124 data-sheet specifies that the falling edge of the DOUT line should be used for an interrupt. The current irq flag (IRQF_TRIGGER_LOW) used will cause unwanted behaviour. When enabling the interrupt it will fire once because the DOUT line is already low. This will make the driver to read an unfinished conversion from the chip. This patch sets the irq type to the one specified in the data-sheet. Signed-off-by:
Alexandru Tachici <alexandru.tachici@analog.com> Signed-off-by:
-
Alexandru Tachici authored
Before this patch the ad_sigma_delta implementation hardcoded the irq trigger type to low, assuming that all Sigma-Delta ADCs have the same interrupt-type. This patch allows all drivers using the ad_sigma_delta layer to set the irq trigger type to the one specified in the datasheet. Signed-off-by:
-
- 16 Jan, 2020 24 commits
-
-
Jérôme Pouiller authored
Some work has been done :) Signed-off-by:
Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-66-Jerome.Pouiller@silabs.comSigned-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Jérôme Pouiller authored
Usage of the "buf_loc" variable does not simplify the function. Signed-off-by:
-
Jérôme Pouiller authored
wfx_tx_queue_get_num_queued() can take advantage of BIT() instead of maintaining one variable for a counter and another for a mask. In add, wfx_tx_queue_get_num_queued() has no real reason to return a size_t instead of an int. Signed-off-by:
-
Jérôme Pouiller authored
Since enum action has now only two cases, it can be dropped. Then hif_handle_tx_data() can be simplified. Signed-off-by:
-
Jérôme Pouiller authored
Obviously, the value of wvif->state as no reason to be wrong. At least, if it the case, dropping the frame is probably not the bast thing to do. Signed-off-by:
-
Jérôme Pouiller authored
The identifier do_probe is unused since "staging: wfx: remove workaround to send probe requests" Signed-off-by:
-
Jérôme Pouiller authored
In current code, packet_id is derivated from mac802.11 packet sequence number, but this number is only unique for a station. It is not sufficient. Signed-off-by:
-
Jérôme Pouiller authored
wfx_ps_notify_sta() is too complex compared to the task it do. Signed-off-by:
-
Jérôme Pouiller authored
When mac80211 ask for a frame to be sent after a DTIM, driver should: 1. Update TIM with multicast bit set (using update_ie). This function can be called whenever. 2. Keep buffered all frames marked "after dtim" 3. When it receive a suspend_resume indication (see wfx_suspend_resume_mc()), send all the buffered frames. This indication is sent by the firmware 4ms before the dtim. 4. If one of the frames returns status "REQUEUE", it means that the DTIM period was ended before to be able to send the frame. 5. When all the buffered frames were sent or if DTIM period was ended, driver should update the TIM with multicast bit reset. All the mess with the asynchronous works can be dropped. Signed-off-by: -
Jérôme Pouiller authored
Indeed, it is not necessary to pass whole hif_ind_suspend_resume_tx to wfx_suspend_resume_mc(). In add, the structure hif_ind_suspend_resume_tx come from hardware API. It is not intended to be manipulated in upper layers of the driver. Signed-off-by:
-
Jérôme Pouiller authored
At the beginning, firmware could send suspend_resume indication to notify that a station wake up or sleep down. However, mac80211 already handles power save status of stations and this behavior has been removed from the firmware. So now, when suspend_resume indication is received, it is always to notify that a DTIM is about to be sent. So, it is possible to simply wfx_suspend_resume(). Signed-off-by:
-
Jérôme Pouiller authored
wfx_ps_notify() is called for any changes in the TIM. However, association ID 0 is a very special case that should be handled independently. Signed-off-by:
-
Jérôme Pouiller authored
pspoll_mask is here to send data buffered in driver. But since station is marked buffered, TIM for this station is 1 and mac80211 will call sta_notify when a ps-poll is received. So pspoll_mask is useless and sta_alseep_mask is sufficient. Signed-off-by:
-
Jérôme Pouiller authored
wfx_tx_get_tid() was used as a wrapper around ieee80211_get_tid(). It did sometime return WFX_MAX_TID to ask to upper layers to not include the frame in "buffered" counter. The objective of this behavior is not clear, but tests has shown that wfx_tx_get_tid() can be replaced by ieee80211_get_tid() without any regressions. BTW, it is not necessary to save the tid in tx_rpiv since it can be retrieved from the 802.11 header. Signed-off-by:
-
Jérôme Pouiller authored
Just for sanity. Signed-off-by:
-
Jérôme Pouiller authored
The "link-id" is a slot number provided to the chip. A link-id is allocated to every station associated with the chip (mainly when the chip is in AP mode). It is more or less the same thing than the association ID, but it is limited to 14 values. Firmware uses the link-id to track the power save status of the stations. The current code try to associate a link-id as soon as data are exchanged with station. It is far easier to rely on sta_add() and sta_remove(). Until now the value WFX_LINK_ID_NO_ASSOC, was only used when no more link-id was available. Now, we also use this value for not-yet-associated stations (that was its primary behavior). Signed-off-by:
-
Jérôme Pouiller authored
Argument provided to wfx_set_tim_impl() is always wvif->aid0_bit_set and there is no reason to provide another argument. Also rename wfx_set_tim_impl() into wfx_update_tim() to reflect the new behavior. Signed-off-by:
-
Jérôme Pouiller authored
Indeed, sta was used after call to rcu_unlock() Signed-off-by:
-
Jérôme Pouiller authored
wfx_tx_confirm_cb() retrieves the station associated with a frame using the MAC address from the 802.11 header. In the other side wfx_tx() retrieves the station using sta field from the ieee80211_tx_control argument. In wfx_cqm_bssloss_sm(), wfx_tx() was called directly without valid sta field, but with a valid MAC address in 802.11 header. So there the processing of this packet was unbalanced and may produce weird bugs. Signed-off-by:
-
Jérôme Pouiller authored
It simplify the code if field buffered is hosted in the struct sta_priv instead of in the struct wfx_link_entry. More globally, struct wfx_link_entry has no real reasons to exist and should be dropped soon. Signed-off-by:
-
Jérôme Pouiller authored
It seems that purpose of "early_data" was to prevent sending data to mac80211 before station was completely associated. It is a useless precaution. Signed-off-by:
-
Jérôme Pouiller authored
MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It is recommended to use function time_*() to compare jiffies. Signed-off-by:
-
Jérôme Pouiller authored
If RTS threshold is 0, it currently disables RTS. It should mean "enabled for every frames". Signed-off-by:
-
Jérôme Pouiller authored
wfx_conf_tx() is called for each queue. On every call, the function updates UAPSD mask and PM mode for all queues. It is a pity since the UAPSD configuration very rarely changes and it makes exchanges between the host and the chip more difficult to track. This patch avoid to update UAPSD and Power Mode in most usual cases. Signed-off-by:
-
