]> www.infradead.org Git - users/jedix/linux-maple.git/log
users/jedix/linux-maple.git
8 months agoMerge tag 'mhi-for-v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mani...
Greg Kroah-Hartman [Sun, 12 Jan 2025 12:57:24 +0000 (13:57 +0100)]
Merge tag 'mhi-for-v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mani/mhi into char-misc-next

Pull MHI update from Manivannan:

MHI Host
========

- Free the "img_info::mhi_buf" in the error path of mhi_alloc_bhie_table() to
  avoid memory leak.

- Add support for Qualcomm X100 5G RAN Accelerator card with relevant channel
  and event config.

- Enable MSI-X if the endpoint supports.

* tag 'mhi-for-v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mani/mhi:
  bus: mhi: host: pci_generic: Enable MSI-X if the endpoint supports
  bus: mhi: host: pci_generic: Add support for QDU100 device
  bus: mhi: host: Free mhi_buf vector inside mhi_alloc_bhie_table()

8 months agoMerge tag 'iio-for-6.14a' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/jic23...
Greg Kroah-Hartman [Sun, 12 Jan 2025 12:45:22 +0000 (13:45 +0100)]
Merge tag 'iio-for-6.14a' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/jic23/iio into char-misc-next

Pull IIO updaate from Jonathan:

IIO: 1st set of new device support, features and cleanup for 6.14

Fairly quiet cycle. Usual mix of new drivers, device support in
existing drivers, features and more general rework and cleanup.
There are a few late breaking or long standing but complex fixes
in here as well.

There is one expected merge conflict due to an upstream fix touching
neighboring code in ti-ads1119. The trivial resolution is the right one with
the result ending up as:

struct {
s16 sample;
aligned_s64 timestamp;
} scan;

New device support
==================

adi,ad4000
- Add support for many Pulsar ADC devices: AD7685, AD7686, AD7687, AD7688,
  AD7690, AD7691, AD7693, AD7942, AD7946, AD7980, AD7982, AD7983, AD7984,
  AD7988-1 and AD7988-5 ADCs. Generally similar to the AD4000 series
  but with lower sampling rates and no configuration registers.
  Includes addition of timestamp channels.
adi,adis16480
- Add support for ADIS16486, ADIS16487 and ADIS16489 IMUs. Required a few tweaks
  to existing driver and addition of tables.
kionix,kx022a
- Add support for KX134ACR-LBZ accelerometer that is similar to the
  KX132ACR-LBZ but with a wider (+-64G) sensor range.
- Add support for KX134-1211 accelerometer that is similar to the
  KX132-1211 but with a wider (+-16G) sensor range.
nxp,fxls8962af
- Add support for fxls8974cf and fxls8967af accelerometers, Both are
  compatible with fxls8962af but with different device IDs which are used in
  presence checks.
renesas,rzg2l
- Add support for Renesas RZ/GS3 SoC ADCs (various driver refactors
  precede this to allow for chip differences).
rohm,bd79704
- New driver for this 6 channel DAC.
st,mpu6050
- Support he IAM20380 which is effectively a cut down IAM20608 IMU with only
  a gyroscope (no accelerometer).
st,stm-timmer-trigger
- Add support for ADC trigger use case for the STM32MP25 SOC. Do not support
  the counter functionality in this driver as that is handled by the
  counter subsystem.
ti,opt4060
- New driver for this RGBW color sensor.

Driver drop
===========

rohm,bu20008
- Drop as decision was made to not mass produce this light sensor after
  Matti had done all the work to get a driver upstream.

Features
========

adi,ad_sigma_delta library + ad7124
- Allow for GPIO to check interrupt status, enabling this device on
  more platforms that don't obey prior (non general) assumptions on
  how the interrupt chips work.
- Allow variation in reset sequence length allowing chip specific
  optimizations rather than always using worst case.
adi,ad7124
- Add temperature channel support.
adi,ad7173
- Add support calibration modes for this family of ADCs.
adi,adxl345
- Binding update to allow specification of which interrupt line is
  connected (or none).
- Support interrupts and FIFO based data capture.
bosch,bme680
- Add regulators support. Note this required a new binding doc rather than
  use of trivial-devices
- Runtime PM support.
microchip,pac1921
- Add ACPI support including _DSM for shunt value and label.
renesas,rzg2l
- Enable runtime autosuspend.
- Add suspend and resume support.
tyhx,hx9023s
- Add loading of a firmware file used to set defaults for some
  configuration registers.
vishay,veml6030
- Support triggered buffers allowing efficient data capture at
  higher speeds.
- Add regmap cache to reduce access to device.

Cleanup and minor fixes
=======================

cross-tree
- Another batch of conversions to devm_regulator_get_enable_read_voltage()
  helper and related conversions to full devm that this enables.
- Various patches using guard() to allow early returns and simpler
  code flow.
- Various conversions from s64 timestamp __aligned(8) to aligned_s64 type.
  Includes a few cleanups where this unsigned and it should have been
  signed.
- Fix up some missing types for drive-open-drain in dt-binding docs.
core
- Add missing documentation for iio_dmaengine_buffer_setup_ext()
- Add check that all buffers passed to iio_read_channel_ext_info()
  and iio_read_channel_label() are page sized and page aligned.
  Done this way because the callbacks are almost always only used
  to fill sysfs attributes. The check covers the tiny percentage
  of cases where use is made of this data in a consumer driver.
- Mark scan_timestamp memory of struct iio_dev private ensuring no
  drivers change the value which belongs to the IIO core.
documentation
- Various missing ABI docs added.
- ABI docs made to use Y consistently as the wildcard for channel
  number.
- Combine duplicate in_currentY_raw entries in ABI docs.
iio-mux
- Fix alignment of buffers passed to iio_channel_read_ext_info().
adi,ad_sigma_delta library
- Respect keep_cs_asserted flag in read path.
- Close a race condition around irq enabling and disabling.
- Use explicit unsigned int in place of unsigned.
adi,ad6695
- Move dt-binding header under adc sub-directory and fix include path in dt
  example.
adi,ad7124
- Check number of channels in DT doesn't exceed what the driver can handle.
- Check input specified in DT are possible.
- Improved error reporting during probe.
adi,ad7173
- Drop unused structure element.
adi,ad7293
- Ensure power is turned on before resetting.
adi,adxl345
- Some documentation simplification and parameter renames.
- Add a function than unifies handling of power up and power down.
- Add defines to have a complete set of registers defined.
- Add missing \n to end of error messages.
amlogic,meson_saradc
- Simplify handling of the REG11 register access.
awinic,aw96104
- Constify iio_info structure.
bosch,bmp085
- Add to dt-binding to indicate devices support SPI.
bosch,bmp280
- Use sizeof() to replace a somewhat magic 2.
- Rename sleep related variables so the unit is included and use
  fsleep() to replace usleep_range() calls.
bosch,bno055
- Constify struct bin_attribute
capella,cm3232
- Reset device before checking hardware ID inline with suggested flow
  from datasheet.
diolan,dln2
- Simplify zeroing of structure used to gather up data by just
  clearing the whole thing before writing rather than trying to
  clear out he padding after write.
freescale,vf610
- Use devm_ and dev_error_probe() to simplify code and allow dropping
  of explicit remove() callback.
invensense,timestamp library
- Use a cast to remove possibility of integer overflow.
kionix,kx022a
- Increase reset delay a little.
maxim,max1363
- Use a buffer of sufficient size in iio_priv() rather than allocating
  variable sized buffer at use time.
microchip,mcp4725
- Replace of_property_read_bool() with of_property_present() for
  detecting presence of regulator which is obviously not a bool.
nxp,fxls8962af
- Add wakeup-source property to the dt binding to allow these sensors
  to wake the system up from suspend.
- Enable finer grained build when not all bus types need to be supported.
renesas,rzg2l
- Use dev_err_probe(), improving handling of probe errors and simplifying
  code.
- Convert to devm_ based cleanup.
- Remove unnecessary runtime PM complexity as clocks are managed through
  PM domains.
- Switch pm_ptr() removing need for __maybe_unused markings.
- use read_poll_timeout() to replace open coded equivalent.
samsung, ssp_sensors
- Simplify code by always providing timestamp whether or not it
  is enabled.
st,lsm6dsx
- Avoid need to include linux/i3c/master by using i3cdev_to_dev() to
  get to the contained struct device.
st,stm32-timer-trigger
- Check for clk_enable() fails.
vishay,veml6030
- Use new gts-helper functions and fix the _scale attribute to take into
  account changes in gain and integration time.

Various other typo fixes in variable names + documentation and help text.
A few whitespace cleanup patches.

* tag 'iio-for-6.14a' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/jic23/iio: (142 commits)
  iio: iio-mux: kzalloc instead of devm_kzalloc to ensure page alignment
  iio: adc: ad7625: Add ending newlines to error messages
  iio: accel: adxl345: complete the list of defines
  iio: accel: adxl345: add FIFO with watermark events
  iio: accel: adxl345: initialize FIFO delay value for SPI
  iio: accel: adxl345: introduce interrupt handling
  iio: light: veml3235: fix scale to conform to ABI
  iio: gts-helper: add helpers to ease searches of gain_sel and new_gain
  iio: light: veml3235: extend regmap to add cache
  iio: light: veml3235: fix code style
  dt-bindings: iio: accel: adxl345: add interrupt-names
  dt-bindings: iio: accel: adxl345: make interrupts not a required property
  dt-bindings: iio: imu: bmi323: add boolean type for drive-open-drain
  dt-bindings: iio: imu: bmi270: add boolean type for drive-open-drain
  dt-bindings: iio: imu: bmi160: add boolean type for drive-open-drain
  iio: adc: meson: simplify MESON_SAR_ADC_REG11 register access
  iio: adc: meson: use tabs instead of spaces for some REG11 bit fields
  iio: adc: meson: fix voltage reference selection field name typo
  iio: adc: rockchip: correct alignment of timestamp
  iio: imu: inv_icm42600: switch timestamp type from int64_t __aligned(8) to aligned_s64
  ...

8 months agoMerge tag 'extcon-next-for-6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel...
Greg Kroah-Hartman [Sun, 12 Jan 2025 12:44:27 +0000 (13:44 +0100)]
Merge tag 'extcon-next-for-6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/chanwoo/extcon into char-misc-next

Pull extcon updates from Chanwoo:

Update extcon next for v6.14

Detailed description for this pull request:
- Fix null pointer check of memory allocation on extcon-rtk-type-c.c.
- Add EXTCON subsystem documentation including the detailed description/example.
- Drop unneeded init of struct i2c_device_id:driver_data on extcon-fsa9480/pth5150.c.

* tag 'extcon-next-for-6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/chanwoo/extcon:
  extcon: Drop explicit initialization of struct i2c_device_id::driver_data to 0
  Documentation: extcon: add documentation for Extcon subsystem
  extcon: realtek: fix NULL deref check in extcon_rtk_type_c_probe

8 months agoMerge tag 'coresight-next-v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel...
Greg Kroah-Hartman [Sun, 12 Jan 2025 12:41:43 +0000 (13:41 +0100)]
Merge tag 'coresight-next-v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/coresight/linux into char-misc-next

Pull coresight updates from Suzuki:

coresight: Updates for Linux v6.14

Coresight self-hosted tracing subsystem updates for v6.14 includes:

  - Support for static traceid allocation for devices
  - Support for impdef, static trace filtering in Qualcomm replicators
  - Miscellaneous fixes

Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
* tag 'coresight-next-v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/coresight/linux:
  coresight-tpda: Optimize the function of reading element size
  coresight: Add support for trace filtering by source
  coresight: Add a helper to check if a device is source
  dt-bindings: arm: qcom,coresight-static-replicator: Add property for source filtering
  coresight: Fix dsb_mode_store() unsigned val is never less than zero
  coresight: dummy: Add static trace id support for dummy source
  coresight: Add support to get static id for system trace sources
  dt-bindings: arm: Add arm,static-trace-id for coresight dummy source
  coresight: Drop atomics in connection refcounts
  Coresight: Narrow down the matching range of tpdm

8 months agomisc: fastrpc: Fix copy buffer page size
Ekansh Gupta [Fri, 10 Jan 2025 13:42:39 +0000 (13:42 +0000)]
misc: fastrpc: Fix copy buffer page size

For non-registered buffer, fastrpc driver copies the buffer and
pass it to the remote subsystem. There is a problem with current
implementation of page size calculation which is not considering
the offset in the calculation. This might lead to passing of
improper and out-of-bounds page size which could result in
memory issue. Calculate page start and page end using the offset
adjusted address instead of absolute address.

Fixes: 02b45b47fbe8 ("misc: fastrpc: fix remote page size calculation")
Cc: stable@kernel.org
Signed-off-by: Ekansh Gupta <quic_ekangupt@quicinc.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20250110134239.123603-4-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc: fastrpc: Fix registered buffer page address
Ekansh Gupta [Fri, 10 Jan 2025 13:42:38 +0000 (13:42 +0000)]
misc: fastrpc: Fix registered buffer page address

For registered  buffers, fastrpc driver sends the buffer information
to remote subsystem. There is a problem with current implementation
where the page address is being sent with an offset leading to
improper buffer address on DSP. This is leads to functional failures
as DSP expects base address in page information and extracts offset
information from remote arguments. Mask the offset and pass the base
page address to DSP.

This issue is observed is a corner case when some buffer which is registered
with fastrpc framework is passed with some offset by user and then the DSP
implementation tried to read the data. As DSP expects base address and takes
care of offsetting with remote arguments, passing an offsetted address will
result in some unexpected data read in DSP.

All generic usecases usually pass the buffer as it is hence is problem is
not usually observed. If someone tries to pass offsetted buffer and then
tries to compare data at HLOS and DSP end, then the ambiguity will be observed.

Fixes: 80f3afd72bd4 ("misc: fastrpc: consider address offset before sending to DSP")
Cc: stable@kernel.org
Signed-off-by: Ekansh Gupta <quic_ekangupt@quicinc.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20250110134239.123603-3-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc: fastrpc: Deregister device nodes properly in error scenarios
Anandu Krishnan E [Fri, 10 Jan 2025 13:42:37 +0000 (13:42 +0000)]
misc: fastrpc: Deregister device nodes properly in error scenarios

During fastrpc_rpmsg_probe, if secure device node registration
succeeds but non-secure device node registration fails, the secure
device node deregister is not called during error cleanup. Add proper
exit paths to ensure proper cleanup in case of error.

Fixes: 3abe3ab3cdab ("misc: fastrpc: add secure domain support")
Cc: stable@kernel.org
Signed-off-by: Anandu Krishnan E <quic_anane@quicinc.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20250110134239.123603-2-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: core: improve range check for nvmem_cell_write()
Jennifer Berringer [Mon, 30 Dec 2024 14:19:01 +0000 (14:19 +0000)]
nvmem: core: improve range check for nvmem_cell_write()

When __nvmem_cell_entry_write() is called for an nvmem cell that does
not need bit shifting, it requires that the len parameter exactly
matches the nvmem cell size. However, when the nvmem cell has a nonzero
bit_offset, it was skipping this check.

Accepting values of len larger than the cell size results in
nvmem_cell_prepare_write_buffer() trying to write past the end of a heap
buffer that it allocates. Add a check to avoid that problem and instead
return -EINVAL when len doesn't match the number of bits expected by the
nvmem cell when bit_offset is nonzero.

This check uses cell->nbits in order to allow providing the smaller size
to cells that are shifted into another byte by bit_offset. For example,
a cell with nbits=8 and nonzero bit_offset would have bytes=2 but should
accept a 1-byte write here, although no current callers depend on this.

Fixes: 69aba7948cbe ("nvmem: Add a simple NVMEM framework for consumers")
Cc: stable@vger.kernel.org
Signed-off-by: Jennifer Berringer <jberring@redhat.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230141901.263976-7-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: qcom-spmi-sdam: Set size in struct nvmem_config
Luca Weiss [Mon, 30 Dec 2024 14:19:00 +0000 (14:19 +0000)]
nvmem: qcom-spmi-sdam: Set size in struct nvmem_config

Let the nvmem core know what size the SDAM is, most notably this fixes
the size of /sys/bus/nvmem/devices/spmi_sdam*/nvmem being '0' and makes
user space work with that file.

  ~ # hexdump -C -s 64 /sys/bus/nvmem/devices/spmi_sdam2/nvmem
  00000040  02 01 00 00 04 00 00 00  00 00 00 00 00 00 00 00  |................|
  00000050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
  *
  00000080

Fixes: 40ce9798794f ("nvmem: add QTI SDAM driver")
Cc: stable@vger.kernel.org
Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
Reviewed-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230141901.263976-6-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: imx-ocotp-ele: set word length to 1
Sascha Hauer [Mon, 30 Dec 2024 14:18:59 +0000 (14:18 +0000)]
nvmem: imx-ocotp-ele: set word length to 1

The ELE hardware internally has a word length of 4. However, among other
things we store MAC addresses in the ELE OCOTP. With a length of 6 bytes
these are naturally unaligned to the word length. Therefore we must
support unaligned reads in reg_read() and indeed it works properly when
reg_read() is called via nvmem_reg_read(). Setting the word size to 4
has the only visible effect that doing unaligned reads from userspace
via bin_attr_nvmem_read() do not work because they are rejected by that
function.

Given that we have to abstract from word accesses to byte accesses in
the driver, set the word size to 1. This allows bytewise accesses from
userspace to be able to test what the driver has to support anyway.

Fixes: 22e9e6fcfb50 ("nvmem: imx: support i.MX93 OCOTP")
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Cc: stable <stable@kernel.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230141901.263976-5-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: imx-ocotp-ele: fix MAC address byte order
Sascha Hauer [Mon, 30 Dec 2024 14:18:58 +0000 (14:18 +0000)]
nvmem: imx-ocotp-ele: fix MAC address byte order

According to the i.MX93 Fusemap the two MAC addresses are stored in
words 315 to 317 like this:

315 MAC1_ADDR_31_0[31:0]
316 MAC1_ADDR_47_32[47:32]
MAC2_ADDR_15_0[15:0]
317 MAC2_ADDR_47_16[31:0]

This means the MAC addresses are stored in reverse byte order. We have
to swap the bytes before passing them to the upper layers. The storage
format is consistent to the one used on i.MX6 using imx-ocotp driver
which does the same byte swapping as introduced here.

With this patch the MAC address on my i.MX93 TQ board correctly reads as
00:d0:93:6b:27:b8 instead of b8:27:6b:93:d0:00.

Fixes: 22e9e6fcfb50 ("nvmem: imx: support i.MX93 OCOTP")
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Cc: stable <stable@kernel.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230141901.263976-4-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: imx-ocotp-ele: fix reading from non zero offset
Sascha Hauer [Mon, 30 Dec 2024 14:18:57 +0000 (14:18 +0000)]
nvmem: imx-ocotp-ele: fix reading from non zero offset

In imx_ocotp_reg_read() the offset comes in as bytes and not as words.
This means we have to divide offset by 4 to get to the correct word
offset.

Also the incoming offset might not be word aligned. In order to read
from the OCOTP the driver aligns down the previous word boundary and
reads from there. This means we have to skip this alignment offset from
the temporary buffer when copying the data to the output buffer.

Fixes: 22e9e6fcfb50 ("nvmem: imx: support i.MX93 OCOTP")
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Cc: stable <stable@kernel.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230141901.263976-3-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: imx-ocotp-ele: simplify read beyond device check
Sascha Hauer [Mon, 30 Dec 2024 14:18:56 +0000 (14:18 +0000)]
nvmem: imx-ocotp-ele: simplify read beyond device check

Do the read beyond device check on function entry in bytes instead of
32bit words which is easier to follow.

Fixes: 22e9e6fcfb50 ("nvmem: imx: support i.MX93 OCOTP")
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Cc: stable <stable@kernel.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230141901.263976-2-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agopps: adjust references to actual name of uapi header file
Lukas Bulwahn [Fri, 10 Jan 2025 10:50:00 +0000 (11:50 +0100)]
pps: adjust references to actual name of uapi header file

Commit 86b525bed275 ("drivers pps: add PPS generators support") adds a file
entry in MAINTAINERS and a reference in the ioctl-number documentation
referring to the file pps-gen.h, whereas the file added in this commit is
named pps_gen.h.

Adjust the two references to the actual name of the uapi header file.
While at it, put the entry in MAINTAINERS at the right place for
alphabetical ordering.

Fixes: 86b525bed275 ("drivers pps: add PPS generators support")
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@redhat.com>
Link: https://lore.kernel.org/r/20250110105000.56228-1-lukas.bulwahn@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoVMCI: fix reference to ioctl-number.rst
Alyssa Ross [Thu, 9 Jan 2025 19:38:07 +0000 (20:38 +0100)]
VMCI: fix reference to ioctl-number.rst

There has never been an ioctl-number.h — this must have been a typo
for ioctl-number.txt (which later become ioctl-number.rst).

At the time this comment was written, the note didn't actually end up
appearing anywhere, but I fixed the omission from ioctl-number.rst in
0a8e4dc1d353 ("Documentation: ioctl: document 0x07 ioctl code").

Fixes: 20259849bb1a ("VMCI: Some header and config files.")
Signed-off-by: Alyssa Ross <hi@alyssa.is>
Link: https://lore.kernel.org/r/re3xng4uwull2cu53xnu5dtv3wlstfiv3v7rmbwtw2qbvj5mo3@q45iujse5ovc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agovirtio: console: Replace deprecated kmap_atomic with kmap_local_page
David Reaver [Thu, 9 Jan 2025 03:59:02 +0000 (19:59 -0800)]
virtio: console: Replace deprecated kmap_atomic with kmap_local_page

kmap_atomic() is deprecated and should be replaced with kmap_local_page()
[1][2]. kmap_local_page() is faster in kernels with HIGHMEM enabled, can
take page faults, and allows preemption.

According to [2], this replacement is safe as long as the code between
kmap_atomic() and kunmap_atomic() does not implicitly depend on disabling
page faults or preemption. In this patch, the only thing happening between
mapping and unmapping the page is a memcpy, and I don't suspect it depends
on disabling page faults or preemption.

[1] https://lwn.net/Articles/836144/
[2] https://docs.kernel.org/mm/highmem.html#temporary-virtual-mappings

Signed-off-by: David Reaver <me@davidreaver.com>
Reviewed-by: Amit Shah <amit@kernel.org>
Link: https://lore.kernel.org/r/20250109035904.168345-1-me@davidreaver.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agopps: clients: gpio: Bypass edge's direction check when not needed
Bastien Curutchet [Wed, 8 Jan 2025 15:30:12 +0000 (16:30 +0100)]
pps: clients: gpio: Bypass edge's direction check when not needed

In the IRQ handler, the GPIO's state is read to verify the direction of
the edge that triggered the interruption before generating the PPS event.
If a pulse is too short, the GPIO line can reach back its original state
before this verification and the PPS event is lost.

This check is needed when info->capture_clear is set because it needs
interruptions on both rising and falling edges. When info->capture_clear
is not set, interruption is triggered by one edge only so this check can
be omitted.

Add a warning if irq_handler is left without triggering any PPS event.
Bypass the edge's direction verification when info->capture_clear is not
set.

Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Rodolfo Giometti <giometti@enneenne.com>
Link: https://lore.kernel.org/r/20250108153012.514925-1-bastien.curutchet@bootlin.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agocdx: disable cdx bus from bus shutdown callback
Abhijit Gangurde [Tue, 3 Dec 2024 08:44:09 +0000 (14:14 +0530)]
cdx: disable cdx bus from bus shutdown callback

disable cdx bus when bus shutdown is called.

Signed-off-by: Abhijit Gangurde <abhijit.gangurde@amd.com>
Link: https://lore.kernel.org/r/20241203084409.2747897-2-abhijit.gangurde@amd.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodrivers/card_reader/rtsx_usb: Restore interrupt based detection
Sean Rhodes [Tue, 19 Nov 2024 08:58:15 +0000 (08:58 +0000)]
drivers/card_reader/rtsx_usb: Restore interrupt based detection

This commit reintroduces interrupt-based card detection previously
used in the rts5139 driver. This functionality was removed in commit
00d8521dcd23 ("staging: remove rts5139 driver code").

Reintroducing this mechanism fixes presence detection for certain card
readers, which with the current driver, will taken approximately 20
seconds to enter S3 as `mmc_rescan` has to be frozen.

Fixes: 00d8521dcd23 ("staging: remove rts5139 driver code")
Cc: stable@vger.kernel.org
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sean Rhodes <sean@starlabs.systems>
Link: https://lore.kernel.org/r/20241119085815.11769-1-sean@starlabs.systems
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc: fastrpc: Rename tgid and pid to client_id
Ekansh Gupta [Fri, 10 Jan 2025 13:43:08 +0000 (13:43 +0000)]
misc: fastrpc: Rename tgid and pid to client_id

The information passed as request tgid and pid is actually the
client id of the process. This client id is used as an
identifier by DSP to identify the DSP PD corresponding to the
process. Currently process tgid is getting passed as the
identifier which is getting replaced by a custom client id.
Rename the data which uses this client id.

Signed-off-by: Ekansh Gupta <quic_ekangupt@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20250110134308.123739-3-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc: fastrpc: Add support for multiple PD from one process
Ekansh Gupta [Fri, 10 Jan 2025 13:43:07 +0000 (13:43 +0000)]
misc: fastrpc: Add support for multiple PD from one process

Memory intensive applications(which requires more tha 4GB) that wants
to offload tasks to DSP might have to split the tasks to multiple
user PD to make the resources available.

For every call to DSP, fastrpc driver passes the process tgid which
works as an identifier for the DSP to enqueue the tasks to specific PD.
With current design, if any process opens device node more than once
and makes PD init request, same tgid will be passed to DSP which will
be considered a bad request and this will result in failure as the same
identifier cannot be used for multiple DSP PD.

Assign and pass a client ID to DSP which would be assigned during device
open and will be dependent on the index of session allocated for the PD.
This will allow the same process to open the device more than once and
spawn multiple dynamic PD for ease of processing.

Signed-off-by: Ekansh Gupta <quic_ekangupt@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20250110134308.123739-2-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc: keba: Fix kernfs warning on module unload
Gerhard Engleder [Sat, 14 Dec 2024 21:57:59 +0000 (22:57 +0100)]
misc: keba: Fix kernfs warning on module unload

Unloading the cp500 module leads to the following warning:

kernfs: can not remove 'eeprom', no directory
WARNING: CPU: 1 PID: 1610 at fs/kernfs/dir.c:1683 kernfs_remove_by_name_ns+0xb1/0xc0

The parent I2C device of the nvmem devices is freed before the nvmem
devices. The reference to the nvmem devices is put by devm after
cp500_remove(), but at this time the parent I2C device does not exist
anymore as the I2C controller and its devices have already been freed in
cp500_remove(). Thus, nvmem tries to remove an entry from an already
deleted directory.

Free nvmem devices before I2C controller auxiliary device.

Signed-off-by: Gerhard Engleder <eg@keba.com>
Link: https://lore.kernel.org/r/20241214215759.60811-1-gerhard@engleder-embedded.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc: Kconfig: Make MCHP_LAN966X_PCI depend on OF_OVERLAY
Ricardo Ribalda [Fri, 29 Nov 2024 18:29:44 +0000 (18:29 +0000)]
misc: Kconfig: Make MCHP_LAN966X_PCI depend on OF_OVERLAY

Drivers should depend on configurations that can be user-configurable
instead of selecting them.

Without this patch, OF cannot be disabled this way:
make allyesconfig
scripts/config -d OF
make olddefconfig

Which is a typical test in CI systems like media-ci.

Now that we are at it, remove the dependency on OF, it will come
automatically from OF_OVERLAY.

Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
Link: https://lore.kernel.org/r/20241129-lan966x-depend-v2-1-72bb9397f421@chromium.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agouio: uio_dmem_genirq: check the return value of devm_kasprintf()
Bartosz Golaszewski [Mon, 2 Dec 2024 18:17:03 +0000 (19:17 +0100)]
uio: uio_dmem_genirq: check the return value of devm_kasprintf()

devm_kasprintf() can fail so check its return value and bail-out on no
memory.

Fixes: 52e2dc2ce2d8 ("uio: Convert a few more users to using %pOFn instead of device_node.name")
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Link: https://lore.kernel.org/r/20241202181703.28546-1-brgl@bgdev.pl
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agouio: Fix return value of poll
Angus Chen [Tue, 26 Nov 2024 12:42:58 +0000 (20:42 +0800)]
uio: Fix return value of poll

The return type of __poll_t can't be negative,
-EIO will be translate to __poll_t,and will return to caller.

Fixes: beafc54c4e2f ("UIO: Add the User IO core code")
Signed-off-by: Angus Chen <angus.chen@jaguarmicro.com>
Link: https://lore.kernel.org/r/20241126124259.1367-1-angus.chen@jaguarmicro.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoMerge tag 'fpga-for-6.14-rc1' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git...
Greg Kroah-Hartman [Thu, 9 Jan 2025 09:56:57 +0000 (10:56 +0100)]
Merge tag 'fpga-for-6.14-rc1' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/fpga/linux-fpga into char-misc-next

Xu writes:

FPGA Manager changes for 6.14-rc1

- Peter's change fixes SRIOV problems for Intel DFL device.

All patches have been reviewed on the mailing list, and have been in the
last linux-next releases (as part of our for-next branch).

Signed-off-by: Xu Yilun <yilun.xu@intel.com>
* tag 'fpga-for-6.14-rc1' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/fpga/linux-fpga:
  fpga: dfl: destroy/recreate feature platform device on port release/assign
  fpga: dfl: drop unneeded get_device() and put_device() of feature device
  fpga: dfl: remove unneeded function build_info_create_dev()
  fpga: dfl: allocate platform device after feature device data
  fpga: dfl: store platform device id in feature device data
  fpga: dfl: store platform device name in feature device data
  fpga: dfl: store MMIO resources in feature device data
  fpga: dfl: convert features from flexible array member to separate array
  fpga: dfl: factor out feature device data from platform device data
  fpga: dfl: factor out feature device registration
  fpga: dfl: refactor internal DFL APIs to take/return feature device data
  fpga: dfl: store FIU type in feature platform data
  fpga: dfl: factor out feature data creation from build_info_commit_dev()
  fpga: dfl: pass feature platform data instead of device as argument
  fpga: dfl: afu: define local pointer to feature device
  fpga: dfl: afu: use parent device to log errors on port enable/disable
  fpga: dfl: return platform data from dfl_fpga_inode_to_feature_dev_data()
  fpga: dfl: omit unneeded argument pdata from dfl_feature_instance_init()

8 months agoMerge tag 'socfpga_firmware_update_for_v6.14' of ssh://gitolite.kernel.org/pub/scm...
Greg Kroah-Hartman [Thu, 9 Jan 2025 09:56:11 +0000 (10:56 +0100)]
Merge tag 'socfpga_firmware_update_for_v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux into char-misc-next

Dinh writes:

SoCFPGA Firmware update for v6.14
- Use kthread_run_on_cpu()

* tag 'socfpga_firmware_update_for_v6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux:
  firmware: stratix10-svc: Use kthread_run_on_cpu()

8 months agoMerge tag 'w1-drv-6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/krzk...
Greg Kroah-Hartman [Thu, 9 Jan 2025 09:54:19 +0000 (10:54 +0100)]
Merge tag 'w1-drv-6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/krzk/linux-w1 into char-misc-next

Krzysztof writes:

1-Wire bus drivers for v6.14

1. ds2482: Add support for handling the VCC regulator supply and three
   more minor improvements/cleanups.

2. Constify 'struct bin_attribute' in all drivers.

3. W1 core: use sysfs_emit() instead of sprintf(), as preferred coding
   style.

* tag 'w1-drv-6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/krzk/linux-w1:
  w1: core: use sysfs_emit() instead of sprintf()
  w1: ds28e04: Constify 'struct bin_attribute'
  w1: ds2805: Constify 'struct bin_attribute'
  w1: ds2781: Constify 'struct bin_attribute'
  w1: ds2780: Constify 'struct bin_attribute'
  w1: ds2438: Constify 'struct bin_attribute'
  w1: ds2433: Constify 'struct bin_attribute'
  w1: ds2431: Constify 'struct bin_attribute'
  w1: ds2430: Constify 'struct bin_attribute'
  w1: ds2413: Constify 'struct bin_attribute'
  w1: ds2408: Constify 'struct bin_attribute'
  w1: ds2406: Constify 'struct bin_attribute'
  w1: Constify 'struct bin_attribute'
  w1: ds2482: Fix datasheet URL
  w1: ds2482: Add regulator support
  w1: ds2482: switch to devm_kzalloc() from kzalloc()
  dt-bindings: w1: ds2482: Add vcc-supply property

8 months agobus: mhi: host: pci_generic: Enable MSI-X if the endpoint supports
Vivek Pernamitta [Wed, 8 Jan 2025 16:22:07 +0000 (21:52 +0530)]
bus: mhi: host: pci_generic: Enable MSI-X if the endpoint supports

Enable MSI-X if the endpoint supports.

Signed-off-by: Vivek Pernamitta <quic_vpernami@quicinc.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Link: https://lore.kernel.org/r/20250108-msix-v2-1-dc4466922350@quicinc.com
[mani: added pci_generic prefix to subject]
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
8 months agomisc: microchip: pci1xxxx: Add push-pull drive support for GPIO
Rengarajan S [Thu, 5 Dec 2024 13:49:56 +0000 (19:19 +0530)]
misc: microchip: pci1xxxx: Add push-pull drive support for GPIO

Add support to configure GPIO pins for push-pull drive mode.

Signed-off-by: Rengarajan S <rengarajan.s@microchip.com>
Link: https://lore.kernel.org/r/20241205134956.1493091-1-rengarajan.s@microchip.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoscripts/spdxcheck: Handle license identifiers in Jinja comments
Lukas Bulwahn [Wed, 8 Jan 2025 12:52:07 +0000 (13:52 +0100)]
scripts/spdxcheck: Handle license identifiers in Jinja comments

Commit 4b132aacb076 ("tools: Add xdrgen") adds a tool, which uses Jinja
template files, i.e., files with the j2 file extension, for its lightweight
code generation.

These template files for this tool have proper headers with the SPDX
License information, which are included as Jinja comments by enclosing the
text with '{#' and '#}'. Sofar, the spdxcheck script does not support to
properly parse this license information in Jinja comments and it reports
back with 'Invalid token: #}'.

Parse Jinja comments properly by stripping the known Jinja comment suffix.

Signed-off-by: Lukas Bulwahn <lukas.bulwahn@redhat.com>
Link: https://lore.kernel.org/r/20250108125207.57486-1-lukas.bulwahn@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoscripts/spdxcheck: Parse j2 comments correctly
Thomas Gleixner [Fri, 29 Nov 2024 10:37:45 +0000 (11:37 +0100)]
scripts/spdxcheck: Parse j2 comments correctly

j2 files use '#}' as comment closure, which trips up the SPDX
parser:

 tools/.../definition.j2: 1:36 Invalid token: #}

Handle those comments correctly by removing the closure.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/878qt2xr46.ffs@tglx
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: No longer depend on BROKEN.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:11 +0000 (13:35 -0600)]
ntsync: No longer depend on BROKEN.

f5b335dc025cfee90957efa90dc72fada0d5abb4 ("misc: ntsync: mark driver as "broken"
to prevent from building") was committed to avoid the driver being used while
only part of its functionality was released. Since the rest of the functionality
has now been committed, revert this.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20241213193511.457338-31-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodocs: ntsync: Add documentation for the ntsync uAPI.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:10 +0000 (13:35 -0600)]
docs: ntsync: Add documentation for the ntsync uAPI.

Add an overall explanation of the driver architecture, and complete and precise
specification for its intended behaviour.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-30-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomaintainers: Add an entry for ntsync.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:09 +0000 (13:35 -0600)]
maintainers: Add an entry for ntsync.

Add myself as maintainer, supported by CodeWeavers.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20241213193511.457338-29-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add a stress test for contended waits.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:08 +0000 (13:35 -0600)]
selftests: ntsync: Add a stress test for contended waits.

Test a more realistic usage pattern, and one with heavy contention, in order to
actually exercise ntsync's internal synchronization.

This test has several threads in a tight loop acquiring a mutex, modifying some
shared data, and then releasing the mutex. At the end we check if the data is
consistent.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-28-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for wakeup signaling via alerts.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:07 +0000 (13:35 -0600)]
selftests: ntsync: Add some tests for wakeup signaling via alerts.

Expand the alert tests to cover alerting a thread mid-wait, to test that the
relevant scheduling logic works correctly.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-27-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add tests for alertable waits.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:06 +0000 (13:35 -0600)]
selftests: ntsync: Add tests for alertable waits.

Test the "alert" functionality of NTSYNC_IOC_WAIT_ALL and NTSYNC_IOC_WAIT_ANY,
when a wait is woken with an alert and when it is woken by an object.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-26-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for wakeup signaling with events.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:05 +0000 (13:35 -0600)]
selftests: ntsync: Add some tests for wakeup signaling with events.

Expand the contended wait tests, which previously only covered events and
semaphores, to cover events as well.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-25-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for auto-reset event state.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:04 +0000 (13:35 -0600)]
selftests: ntsync: Add some tests for auto-reset event state.

Test event-specific ioctls NTSYNC_IOC_EVENT_SET, NTSYNC_IOC_EVENT_RESET,
NTSYNC_IOC_EVENT_PULSE, NTSYNC_IOC_EVENT_READ for auto-reset events, and
waiting on auto-reset events.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-24-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for manual-reset event state.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:03 +0000 (13:35 -0600)]
selftests: ntsync: Add some tests for manual-reset event state.

Test event-specific ioctls NTSYNC_IOC_EVENT_SET, NTSYNC_IOC_EVENT_RESET,
NTSYNC_IOC_EVENT_PULSE, NTSYNC_IOC_EVENT_READ for manual-reset events, and
waiting on manual-reset events.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-23-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for wakeup signaling with WINESYNC_IOC_WAIT_ALL.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:02 +0000 (13:35 -0600)]
selftests: ntsync: Add some tests for wakeup signaling with WINESYNC_IOC_WAIT_ALL.

Test contended "wait-for-all" waits, to make sure that scheduling and wakeup
logic works correctly, and that the wait only exits once objects are all
simultaneously signaled.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-22-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for wakeup signaling with WINESYNC_IOC_WAIT_ANY.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:01 +0000 (13:35 -0600)]
selftests: ntsync: Add some tests for wakeup signaling with WINESYNC_IOC_WAIT_ANY.

Test contended "wait-for-any" waits, to make sure that scheduling and wakeup
logic works correctly.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-21-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for NTSYNC_IOC_WAIT_ALL.
Elizabeth Figura [Fri, 13 Dec 2024 19:35:00 +0000 (13:35 -0600)]
selftests: ntsync: Add some tests for NTSYNC_IOC_WAIT_ALL.

Test basic synchronous functionality of NTSYNC_IOC_WAIT_ALL, and when objects
are considered simultaneously signaled.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-20-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for NTSYNC_IOC_WAIT_ANY.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:59 +0000 (13:34 -0600)]
selftests: ntsync: Add some tests for NTSYNC_IOC_WAIT_ANY.

Test basic synchronous functionality of NTSYNC_IOC_WAIT_ANY, when objects are
considered signaled or not signaled, and how they are affected by a successful
wait.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-19-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for mutex state.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:58 +0000 (13:34 -0600)]
selftests: ntsync: Add some tests for mutex state.

Test mutex-specific ioctls NTSYNC_IOC_MUTEX_UNLOCK and NTSYNC_IOC_MUTEX_READ,
and waiting on mutexes.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-18-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoselftests: ntsync: Add some tests for semaphore state.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:57 +0000 (13:34 -0600)]
selftests: ntsync: Add some tests for semaphore state.

Wine has tests for its synchronization primitives, but these are more accessible
to kernel developers, and also allow us to test some edge cases that Wine does
not care about.

This patch adds tests for semaphore-specific ioctls NTSYNC_IOC_SEM_POST and
NTSYNC_IOC_SEM_READ, and waiting on semaphores.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-17-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce alertable waits.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:56 +0000 (13:34 -0600)]
ntsync: Introduce alertable waits.

NT waits can optionally be made "alertable". This is a special channel for
thread wakeup that is mildly similar to SIGIO. A thread has an internal single
bit of "alerted" state, and if a thread is alerted while an alertable wait, the
wait will return a special value, consume the "alerted" state, and will not
consume any of its objects.

Alerts are implemented using events; the user-space NT emulator is expected to
create an internal ntsync event for each thread and pass that event to wait
functions.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20241213193511.457338-16-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_EVENT_READ.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:55 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_EVENT_READ.

This corresponds to the NT syscall NtQueryEvent().

This returns the signaled state of the event and whether it is manual-reset.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-15-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_MUTEX_READ.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:54 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_MUTEX_READ.

This corresponds to the NT syscall NtQueryMutant().

This returns the recursion count, owner, and abandoned state of the mutex.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-14-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_SEM_READ.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:53 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_SEM_READ.

This corresponds to the NT syscall NtQuerySemaphore().

This returns the current count and maximum count of the semaphore.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-13-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_EVENT_PULSE.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:52 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_EVENT_PULSE.

This corresponds to the NT syscall NtPulseEvent().

This wakes up any waiters as if the event had been set, but does not set the
event, instead resetting it if it had been signalled. Thus, for a manual-reset
event, all waiters are woken, whereas for an auto-reset event, at most one
waiter is woken.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20241213193511.457338-12-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_EVENT_RESET.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:51 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_EVENT_RESET.

This corresponds to the NT syscall NtResetEvent().

This sets the event to the unsignaled state, and returns its previous state.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20241213193511.457338-11-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_EVENT_SET.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:50 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_EVENT_SET.

This corresponds to the NT syscall NtSetEvent().

This sets the event to the signaled state, and returns its previous state.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-10-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_CREATE_EVENT.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:49 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_CREATE_EVENT.

This correspond to the NT syscall NtCreateEvent().

An NT event holds a single bit of state denoting whether it is signaled or
unsignaled.

There are two types of events: manual-reset and automatic-reset. When an
automatic-reset event is acquired via a wait function, its state is reset to
unsignaled. Manual-reset events are not affected by wait functions.

Whether the event is manual-reset, and its initial state, are specified at
creation time.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-9-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_MUTEX_KILL.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:48 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_MUTEX_KILL.

This does not correspond to any NT syscall. Rather, when a thread dies, it
should be called by the NT emulator for each mutex, with the TID of the dying
thread.

NT mutexes are robust (in the pthread sense). When an NT thread dies, any
mutexes it owned are immediately released. Acquisition of those mutexes by other
threads will return a special value indicating that the mutex was abandoned,
like EOWNERDEAD returned from pthread_mutex_lock(), and EOWNERDEAD is indeed
used here for that purpose.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-8-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_MUTEX_UNLOCK.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:47 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_MUTEX_UNLOCK.

This corresponds to the NT syscall NtReleaseMutant().

This syscall decrements the mutex's recursion count by one, and returns the
previous value. If the mutex is not owned by the current task, the function
instead fails and returns -EPERM.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-7-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_CREATE_MUTEX.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:46 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_CREATE_MUTEX.

This corresponds to the NT syscall NtCreateMutant().

An NT mutex is recursive, with a 32-bit recursion counter. When acquired via
NtWaitForMultipleObjects(), the recursion counter is incremented by one. The OS
records the thread which acquired it.

The OS records the thread which acquired it. However, in order to keep this
driver self-contained, the owning thread ID is managed by user-space, and passed
as a parameter to all relevant ioctls.

The initial owner and recursion count, if any, are specified when the mutex is
created.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-6-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_WAIT_ALL.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:45 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_WAIT_ALL.

This is similar to NTSYNC_IOC_WAIT_ANY, but waits until all of the objects are
simultaneously signaled, and then acquires all of them as a single atomic
operation.

Because acquisition of multiple objects is atomic, some complex locking is
required. We cannot simply spin-lock multiple objects simultaneously, as that
may disable preëmption for a problematically long time.

Instead, modifying any object which may be involved in a wait-all operation takes
a device-wide sleeping mutex, "wait_all_lock", instead of the normal object
spinlock.

Because wait-for-all is a rare operation, in order to optimize wait-for-any,
this lock is only taken when necessary. "all_hint" is used to mark objects which
are involved in a wait-for-all operation, and if an object is not, only its
spinlock is taken.

The locking scheme used here was written by Peter Zijlstra.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-5-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Introduce NTSYNC_IOC_WAIT_ANY.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:44 +0000 (13:34 -0600)]
ntsync: Introduce NTSYNC_IOC_WAIT_ANY.

This corresponds to part of the functionality of the NT syscall
NtWaitForMultipleObjects(). Specifically, it implements the behaviour where
the third argument (wait_any) is TRUE, and it does not handle alertable waits.
Those features have been split out into separate patches to ease review.

This patch therefore implements the wait/wake infrastructure which comprises the
core of ntsync's functionality.

NTSYNC_IOC_WAIT_ANY is a vectored wait function similar to poll(). Unlike
poll(), it "consumes" objects when they are signaled. For semaphores, this means
decreasing one from the internal counter. At most one object can be consumed by
this function.

This wait/wake model is fundamentally different from that used anywhere else in
the kernel, and for that reason ntsync does not use any existing infrastructure,
such as futexes, kernel mutexes or semaphores, or wait_event().

Up to 64 objects can be waited on at once. As soon as one is signaled, the
object with the lowest index is consumed, and that index is returned via the
"index" field.

A timeout is supported. The timeout is passed as a u64 nanosecond value, which
represents absolute time measured against either the MONOTONIC or REALTIME clock
(controlled by the flags argument). If U64_MAX is passed, the ioctl waits
indefinitely.

This ioctl validates that all objects belong to the relevant device. This is not
necessary for any technical reason related to NTSYNC_IOC_WAIT_ANY, but will be
necessary for NTSYNC_IOC_WAIT_ALL introduced in the following patch.

Some padding fields are added for alignment and for fields which will be added
in future patches (split out to ease review).

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-4-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Rename NTSYNC_IOC_SEM_POST to NTSYNC_IOC_SEM_RELEASE.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:43 +0000 (13:34 -0600)]
ntsync: Rename NTSYNC_IOC_SEM_POST to NTSYNC_IOC_SEM_RELEASE.

Use the more common "release" terminology, which is also the term used by NT,
instead of "post" (which is used by POSIX).

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-3-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agontsync: Return the fd from NTSYNC_IOC_CREATE_SEM.
Elizabeth Figura [Fri, 13 Dec 2024 19:34:42 +0000 (13:34 -0600)]
ntsync: Return the fd from NTSYNC_IOC_CREATE_SEM.

Simplify the user API a bit by returning the fd as return value from the ioctl
instead of through the argument pointer.

Signed-off-by: Elizabeth Figura <zfigura@codeweavers.com>
Link: https://lore.kernel.org/r/20241213193511.457338-2-zfigura@codeweavers.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoscripts/tags.sh: Tag timer definitions
Costa Shulyupin [Mon, 9 Dec 2024 08:29:57 +0000 (10:29 +0200)]
scripts/tags.sh: Tag timer definitions

For timer definitions like
DEFINE_TIMER(mytimer, mytimer_handler);
ctags generates tags `DEFINE_TIMER` and skips `mytimer`
because it doesn't expand the DEFINE_TIMER macro.

Configure ctags to generate tag for `mytimer`
ans skip the `DEFINE_TIMER` tag in such cases.

Signed-off-by: Costa Shulyupin <costa.shul@redhat.com>
Link: https://lore.kernel.org/r/20241209083004.911013-2-costa.shul@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc:minor basic kunit tests
Vimal Agrawal [Mon, 21 Oct 2024 13:39:26 +0000 (13:39 +0000)]
misc:minor basic kunit tests

basic kunit tests for misc minor

Signed-off-by: Vimal Agrawal <vimal.agrawal@sophos.com>
Reviewed-by: Dirk VanDerMerwe <dirk.vandermerwe@sophos.com>
Link: https://lore.kernel.org/r/20241021133926.23774-1-vimal.agrawal@sophos.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agomisc: misc_minor_alloc to use ida for all dynamic/misc dynamic minors
Vimal Agrawal [Mon, 21 Oct 2024 13:38:12 +0000 (13:38 +0000)]
misc: misc_minor_alloc to use ida for all dynamic/misc dynamic minors

misc_minor_alloc was allocating id using ida for minor only in case of
MISC_DYNAMIC_MINOR but misc_minor_free was always freeing ids
using ida_free causing a mismatch and following warn:
> > WARNING: CPU: 0 PID: 159 at lib/idr.c:525 ida_free+0x3e0/0x41f
> > ida_free called for id=127 which is not allocated.
> > <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
...
> > [<60941eb4>] ida_free+0x3e0/0x41f
> > [<605ac993>] misc_minor_free+0x3e/0xbc
> > [<605acb82>] misc_deregister+0x171/0x1b3

misc_minor_alloc is changed to allocate id from ida for all minors
falling in the range of dynamic/ misc dynamic minors

Fixes: ab760791c0cf ("char: misc: Increase the maximum number of dynamic misc devices to 1048448")
Signed-off-by: Vimal Agrawal <vimal.agrawal@sophos.com>
Reviewed-by: Dirk VanDerMerwe <dirk.vandermerwe@sophos.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20241021133812.23703-1-vimal.agrawal@sophos.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoVMCI: remove unused ioctl definitions
Alyssa Ross [Thu, 19 Dec 2024 23:29:57 +0000 (00:29 +0100)]
VMCI: remove unused ioctl definitions

IOCTL_VMCI_SOCKETS_VERSION and IOCTL_VMCI_SOCKETS_GET_AF_VALUE were
never implemented, because VSOCK ended up being implemented as a
generic mechanism with a static AF value.  Likewise,
IOCTL_VMCI_SOCKETS_GET_LOCAL_CID ended up being implemented as
IOCTL_VM_SOCKETS_GET_LOCAL_CID.

This isn't a UAPI header, so it should be fine to remove the unused
values.  I've left a comment noting IOCTL_VM_SOCKETS_GET_LOCAL_CID is
in the VMCI range to avoid unintentional reuse.

Signed-off-by: Alyssa Ross <hi@alyssa.is>
Acked-by: Vishnu Dasa <vishnu.dasa@broadcom.com>
Link: https://lore.kernel.org/r/fzdcrz4yfedokmbm22h2iwsluix4jwejwaltuwcdr6kz3yu6eu@nue5xc6ayevo
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agobinder: fix kernel-doc warning of 'file' member
Carlos Llamas [Mon, 6 Jan 2025 19:26:07 +0000 (19:26 +0000)]
binder: fix kernel-doc warning of 'file' member

The 'struct file' member in 'binder_task_work_cb' definition was renamed
to 'file' between patch versions but its kernel-doc reference kept the
old name 'fd'. Update the naming to fix the W=1 build warning.

Cc: Todd Kjos <tkjos@google.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202501031535.erbln3A2-lkp@intel.com/
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Acked-by: Todd Kjos <tkjos@google.com>
Link: https://lore.kernel.org/r/20250106192608.1107362-1-cmllamas@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agobinderfs: add new binder devices to binder_devices
Li Li [Wed, 18 Dec 2024 21:29:34 +0000 (13:29 -0800)]
binderfs: add new binder devices to binder_devices

When binderfs is not enabled, the binder driver parses the kernel
config to create all binder devices. All of the new binder devices
are stored in the list binder_devices.

When binderfs is enabled, the binder driver creates new binder devices
dynamically when userspace applications call BINDER_CTL_ADD ioctl. But
the devices created in this way are not stored in the same list.

This patch fixes that.

Signed-off-by: Li Li <dualli@google.com>
Acked-by: Carlos Llamas <cmllamas@google.com>
Link: https://lore.kernel.org/r/20241218212935.4162907-2-dualli@chromium.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoDocumentation ABI: add PPS generators documentation
Rodolfo Giometti [Fri, 8 Nov 2024 07:31:15 +0000 (08:31 +0100)]
Documentation ABI: add PPS generators documentation

This patch adds the documentation for the ABI between the Linux kernel
and userspace regarding the PPS generators.

Signed-off-by: Rodolfo Giometti <giometti@enneenne.com>
Link: https://lore.kernel.org/r/20241108073115.759039-5-giometti@enneenne.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoDocumentation pps.rst: add PPS generators documentation
Rodolfo Giometti [Fri, 8 Nov 2024 07:31:14 +0000 (08:31 +0100)]
Documentation pps.rst: add PPS generators documentation

This patch adds some examples about how to register a new PPS
generator in the system, and how to manage it.

Signed-off-by: Rodolfo Giometti <giometti@enneenne.com>
Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com>
Link: https://lore.kernel.org/r/20241108073115.759039-4-giometti@enneenne.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodrivers pps/generators: add dummy PPS generator
Rodolfo Giometti [Fri, 8 Nov 2024 07:31:13 +0000 (08:31 +0100)]
drivers pps/generators: add dummy PPS generator

This dummy PPS generator can be used for debugging and documentation
purposes.

Signed-off-by: Rodolfo Giometti <giometti@enneenne.com>
Link: https://lore.kernel.org/r/20241108073115.759039-3-giometti@enneenne.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodrivers pps: add PPS generators support
Rodolfo Giometti [Fri, 8 Nov 2024 07:31:12 +0000 (08:31 +0100)]
drivers pps: add PPS generators support

Sometimes one needs to be able not only to catch PPS signals but to
produce them also. For example, running a distributed simulation,
which requires computers' clock to be synchronized very tightly.

This patch adds PPS generators class in order to have a well-defined
interface for these devices.

Signed-off-by: Rodolfo Giometti <giometti@enneenne.com>
Link: https://lore.kernel.org/r/20241108073115.759039-2-giometti@enneenne.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agopps: Fix a use-after-free
Calvin Owens [Tue, 12 Nov 2024 04:13:29 +0000 (20:13 -0800)]
pps: Fix a use-after-free

On a board running ntpd and gpsd, I'm seeing a consistent use-after-free
in sys_exit() from gpsd when rebooting:

    pps pps1: removed
    ------------[ cut here ]------------
    kobject: '(null)' (00000000db4bec24): is not initialized, yet kobject_put() is being called.
    WARNING: CPU: 2 PID: 440 at lib/kobject.c:734 kobject_put+0x120/0x150
    CPU: 2 UID: 299 PID: 440 Comm: gpsd Not tainted 6.11.0-rc6-00308-gb31c44928842 #1
    Hardware name: Raspberry Pi 4 Model B Rev 1.1 (DT)
    pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
    pc : kobject_put+0x120/0x150
    lr : kobject_put+0x120/0x150
    sp : ffffffc0803d3ae0
    x29: ffffffc0803d3ae0 x28: ffffff8042dc9738 x27: 0000000000000001
    x26: 0000000000000000 x25: ffffff8042dc9040 x24: ffffff8042dc9440
    x23: ffffff80402a4620 x22: ffffff8042ef4bd0 x21: ffffff80405cb600
    x20: 000000000008001b x19: ffffff8040b3b6e0 x18: 0000000000000000
    x17: 0000000000000000 x16: 0000000000000000 x15: 696e6920746f6e20
    x14: 7369203a29343263 x13: 205d303434542020 x12: 0000000000000000
    x11: 0000000000000000 x10: 0000000000000000 x9 : 0000000000000000
    x8 : 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000
    x5 : 0000000000000000 x4 : 0000000000000000 x3 : 0000000000000000
    x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000000
    Call trace:
     kobject_put+0x120/0x150
     cdev_put+0x20/0x3c
     __fput+0x2c4/0x2d8
     ____fput+0x1c/0x38
     task_work_run+0x70/0xfc
     do_exit+0x2a0/0x924
     do_group_exit+0x34/0x90
     get_signal+0x7fc/0x8c0
     do_signal+0x128/0x13b4
     do_notify_resume+0xdc/0x160
     el0_svc+0xd4/0xf8
     el0t_64_sync_handler+0x140/0x14c
     el0t_64_sync+0x190/0x194
    ---[ end trace 0000000000000000 ]---

...followed by more symptoms of corruption, with similar stacks:

    refcount_t: underflow; use-after-free.
    kernel BUG at lib/list_debug.c:62!
    Kernel panic - not syncing: Oops - BUG: Fatal exception

This happens because pps_device_destruct() frees the pps_device with the
embedded cdev immediately after calling cdev_del(), but, as the comment
above cdev_del() notes, fops for previously opened cdevs are still
callable even after cdev_del() returns. I think this bug has always
been there: I can't explain why it suddenly started happening every time
I reboot this particular board.

In commit d953e0e837e6 ("pps: Fix a use-after free bug when
unregistering a source."), George Spelvin suggested removing the
embedded cdev. That seems like the simplest way to fix this, so I've
implemented his suggestion, using __register_chrdev() with pps_idr
becoming the source of truth for which minor corresponds to which
device.

But now that pps_idr defines userspace visibility instead of cdev_add(),
we need to be sure the pps->dev refcount can't reach zero while
userspace can still find it again. So, the idr_remove() call moves to
pps_unregister_cdev(), and pps_idr now holds a reference to pps->dev.

    pps_core: source serial1 got cdev (251:1)
    <...>
    pps pps1: removed
    pps_core: unregistering pps1
    pps_core: deallocating pps1

Fixes: d953e0e837e6 ("pps: Fix a use-after free bug when unregistering a source.")
Cc: stable@vger.kernel.org
Signed-off-by: Calvin Owens <calvin@wbinvd.org>
Reviewed-by: Michal Schmidt <mschmidt@redhat.com>
Link: https://lore.kernel.org/r/a17975fd5ae99385791929e563f72564edbcf28f.1731383727.git.calvin@wbinvd.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoiio: iio-mux: kzalloc instead of devm_kzalloc to ensure page alignment
Matteo Martelli [Mon, 2 Dec 2024 15:11:08 +0000 (16:11 +0100)]
iio: iio-mux: kzalloc instead of devm_kzalloc to ensure page alignment

During channel configuration, the iio-mux driver allocates a page with
devm_kzalloc(PAGE_SIZE) to read channel ext_info. However, the resulting
buffer points to an offset of the page due to the devres header sitting
at the beginning of the allocated area. This leads to failure in the
provider driver when sysfs_emit* helpers are used to format the ext_info
attributes.

Switch to plain kzalloc version. The devres version is not strictly
necessary as the buffer is only accessed during the channel
configuration phase. Rely on __free cleanup to deallocate the buffer.
Also, move the ext_info handling into a new function to have the page
buffer definition and assignment in one statement as suggested by
cleanup documentation.

Signed-off-by: Matteo Martelli <matteomartelli3@gmail.com>
Fixes: 7ba9df54b091 ("iio: multiplexer: new iio category and iio-mux driver")
Reviewed-by: David Lechner <dlechner@baylibre.com>
Link: https://patch.msgid.link/20241202-iio-kmalloc-align-v1-2-aa9568c03937@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: adc: ad7625: Add ending newlines to error messages
Uwe Kleine-König [Mon, 30 Dec 2024 15:10:31 +0000 (16:10 +0100)]
iio: adc: ad7625: Add ending newlines to error messages

Error messages passed to dev_err_probe() are supposed to end in "\n".
Fix accordingly.

Fixes: b7ffd0fa65e9 ("iio: adc: ad7625: add driver")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
Tested-by: Trevor Gamblin <tgamblin@baylibre.com>
Link: https://patch.msgid.link/20241230151030.3207529-2-u.kleine-koenig@baylibre.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: accel: adxl345: complete the list of defines
Lothar Rubusch [Sat, 28 Dec 2024 23:29:49 +0000 (23:29 +0000)]
iio: accel: adxl345: complete the list of defines

Having interrupts events and FIFO available allows to evaluate the
sensor events. Cover the list of interrupt based sensor events. Keep
them in the header file for readability.

Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>
Link: https://patch.msgid.link/20241228232949.72487-5-l.rubusch@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: accel: adxl345: add FIFO with watermark events
Lothar Rubusch [Sat, 28 Dec 2024 23:29:48 +0000 (23:29 +0000)]
iio: accel: adxl345: add FIFO with watermark events

Add a basic setup for FIFO with configurable watermark. Add a handler
for watermark interrupt events and extend the channel for the
scan_index needed for the iio channel. The sensor is configurable to use
a FIFO_BYPASSED mode or a FIFO_STREAM mode. For the FIFO_STREAM mode now
a watermark can be configured, or disabled by setting 0. Further features
require a working FIFO setup.

Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>
Link: https://patch.msgid.link/20241228232949.72487-4-l.rubusch@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: accel: adxl345: initialize FIFO delay value for SPI
Lothar Rubusch [Sat, 28 Dec 2024 23:29:47 +0000 (23:29 +0000)]
iio: accel: adxl345: initialize FIFO delay value for SPI

Add the possibility to delay FIFO access when SPI is used. According to
the datasheet this is needed for the adxl345. When initialization
happens over SPI the need for delay is to be signalized, and the delay
will be used.

Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>
Link: https://patch.msgid.link/20241228232949.72487-3-l.rubusch@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: accel: adxl345: introduce interrupt handling
Lothar Rubusch [Sat, 28 Dec 2024 23:29:46 +0000 (23:29 +0000)]
iio: accel: adxl345: introduce interrupt handling

Add the possibility to claim an interrupt. Init the state structure
with an interrupt line obtained from the DT. The adxl345 can use
two different interrupt lines for event handling. Only one is used.

Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>
Link: https://patch.msgid.link/20241228232949.72487-2-l.rubusch@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: light: veml3235: fix scale to conform to ABI
Javier Carrasco [Mon, 30 Dec 2024 15:13:53 +0000 (16:13 +0100)]
iio: light: veml3235: fix scale to conform to ABI

The current scale is not ABI-compliant as it is just the sensor gain
instead of the value that acts as a multiplier to be applied to the raw
value (there is no offset).

Use the iio-gts helpers to obtain the proper scale values according to
the gain and integration time to match the resolution tables from the
datasheet. When at it, use 'scale' instead of 'gain' consistently for
the get/set functions to avoid misunderstandings.

Fixes: c5a23f80c164 ("iio: light: add support for veml3235")
Reviewed-by: Matti Vaittinen <mazziesaccount@gmail.com>
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241230-veml3235_scale-v3-2-48a5795e2f64@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: gts-helper: add helpers to ease searches of gain_sel and new_gain
Javier Carrasco [Mon, 30 Dec 2024 15:13:52 +0000 (16:13 +0100)]
iio: gts-helper: add helpers to ease searches of gain_sel and new_gain

This helper functions reduce the burden in the drivers that want to
fetch a gain and time selector for a given scale or a new optimal gain.

The former is currently achieved by calling
iio_gts_find_gain_sel_for_scale_using_time() for the current time
selector, and then iterating over the rest of time selectors if the
gain selector was not found.

The latter requires a combination of multiple iio-gts helpers to find
the new gain, look for an optimal gain if there was no exact match, and
set a minimum gain if the optimal gain is not in the range of available
gains.

Provide simpler workflows by means of functions that address common
patterns in the users of the iio-gts helpers.

Acked-by: Matti Vaittinen <mazziesaccount@gmail.com>
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241230-veml3235_scale-v3-1-48a5795e2f64@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoMAINTAINERS: add slimbus documentation
anish kumar [Mon, 30 Dec 2024 14:33:54 +0000 (14:33 +0000)]
MAINTAINERS: add slimbus documentation

In the commit 202318d37613d264e30d71cc32ef442492d6d279
slimbus documentation was added but it missed
the update in this file. Currently get_maintainer script
is missing the main maintainer.

Signed-off-by: anish kumar <yesanishhere@gmail.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143354.266154-3-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoslimbus: messaging: Reorganize kerneldoc parameter names
Julia Lawall [Mon, 30 Dec 2024 14:33:53 +0000 (14:33 +0000)]
slimbus: messaging: Reorganize kerneldoc parameter names

Reorganize kerneldoc parameter names to match the parameter
order in the function header.

Problems identified using Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143354.266154-2-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodt-bindings: nvmem: qfprom: Add compatible for QCS8300
Jingyi Wang [Mon, 30 Dec 2024 14:30:35 +0000 (14:30 +0000)]
dt-bindings: nvmem: qfprom: Add compatible for QCS8300

Document QFPROM compatible for Qualcomm QCS8300. It provides access
functions for QFPROM data to rest of the drivers via nvmem interface.

Signed-off-by: Jingyi Wang <quic_jingyw@quicinc.com>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-12-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodt-bindings: nvmem: Add compatible for IPQ5424
Manikanta Mylavarapu [Mon, 30 Dec 2024 14:30:34 +0000 (14:30 +0000)]
dt-bindings: nvmem: Add compatible for IPQ5424

Document the QFPROM block found on IPQ5424

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Manikanta Mylavarapu <quic_mmanikan@quicinc.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-11-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodt-bindings: nvmem: Add compatible for MS8917
Barnabás Czémán [Mon, 30 Dec 2024 14:30:33 +0000 (14:30 +0000)]
dt-bindings: nvmem: Add compatible for MS8917

Document the QFPROM block found on MSM8917.

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Barnabás Czémán <barnabas.czeman@mainlining.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-10-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoMAINTAINERS: Update nvmem section
Peng Fan [Mon, 30 Dec 2024 14:30:32 +0000 (14:30 +0000)]
MAINTAINERS: Update nvmem section

Include include/dt-bindings/nvmem into nvmem section

Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-9-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agoMIPS: mobileye: eyeq5: add bootloader config reserved memory
Théo Lebrun [Mon, 30 Dec 2024 14:30:31 +0000 (14:30 +0000)]
MIPS: mobileye: eyeq5: add bootloader config reserved memory

Add a new reserved-memory node, containing bootloader config with MAC
addresses for both ethernet instances of the SoC.

Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-8-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: rmem: add CRC validation for Mobileye EyeQ5 NVMEM
Théo Lebrun [Mon, 30 Dec 2024 14:30:30 +0000 (14:30 +0000)]
nvmem: rmem: add CRC validation for Mobileye EyeQ5 NVMEM

Mobileye EyeQ5 has a non-volatile memory region which
gets used to store MAC addresses. Its format includes
a prefix 12-byte header and a suffix 4-byte CRC.

Add an optional ->checksum() callback inside match data;
it runs CRC32 onto the content.

Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-7-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: rmem: remove unused struct rmem::size field
Théo Lebrun [Mon, 30 Dec 2024 14:30:29 +0000 (14:30 +0000)]
nvmem: rmem: remove unused struct rmem::size field

The private structure used by the rmem driver contains
a `size` field that is unused. Remove it.

Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-6-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: rmem: make ->reg_read() straight forward code
Théo Lebrun [Mon, 30 Dec 2024 14:30:28 +0000 (14:30 +0000)]
nvmem: rmem: make ->reg_read() straight forward code

memory_read_from_buffer() is a weird choice; it:
 - is made for iteration with ppos a pointer.
 - does futile error checking in our case.
 - does NOT ensure we read exactly N bytes.

Replace it by:
1. A check that (offset + bytes) lands inside the region and,
2. a plain memcpy().

Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-5-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: specify ->reg_read/reg_write() expected return values
Théo Lebrun [Mon, 30 Dec 2024 14:30:27 +0000 (14:30 +0000)]
nvmem: specify ->reg_read/reg_write() expected return values

Both ->reg_read() and ->reg_write() return values are not easy to
deduce. Explicit that they should return zero on success (and negative
values otherwise).

Such callbacks, in some alternative world, could return the number of
bytes in the success case. That would be translated to errors in the
nvmem core because of checks like:

ret = nvmem->reg_write(nvmem->priv, offset, val, bytes);
if (ret) {
// error case
}

This mistake is not just theoretical, see commit
28b008751aa2 ("nvmem: rmem: Fix return value of rmem_read()").

Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-4-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agodt-bindings: nvmem: rmem: Add mobileye,eyeq5-bootloader-config
Théo Lebrun [Mon, 30 Dec 2024 14:30:26 +0000 (14:30 +0000)]
dt-bindings: nvmem: rmem: Add mobileye,eyeq5-bootloader-config

On Mobileye EyeQ5, the bootloader will put MAC addresses into memory.
Declare that as reserved memory to be used by the kernel, exposing
nvmem cells. That region has a 12-byte header and a 4-byte trailing CRC.

Acked-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-3-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agonvmem: core: constify 'struct bin_attribute'
Thomas Weißschuh [Mon, 30 Dec 2024 14:30:25 +0000 (14:30 +0000)]
nvmem: core: constify 'struct bin_attribute'

The sysfs core now allows instances of 'struct bin_attribute' to be
moved into read-only memory. Make use of that to protect them against
accidental or malicious modifications.

Also adapt the dynamic sysfs cell logic to handle the const attributes.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20241230143035.265518-2-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 months agobus: mhi: host: pci_generic: Add support for QDU100 device
Vivek Pernamitta [Thu, 12 Dec 2024 11:57:27 +0000 (17:27 +0530)]
bus: mhi: host: pci_generic: Add support for QDU100 device

Add MHI controller configuration for QDU100 device.

The Qualcomm X100 5G RAN Accelerator card is designed to enhance Open vRAN
servers by offloading CPUs from intensive 5G baseband functions.

Link: https://docs.qualcomm.com/bundle/publicresource/87-79371-1_REV_A_Qualcomm_X100_5G_RAN_Accelerator_Card_Product_Brief.pdf
Signed-off-by: Vivek Pernamitta <quic_vpernami@quicinc.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Link: https://lore.kernel.org/r/20241212-qdu100_us-v5-1-3349fb23512a@quicinc.com
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
8 months agoiio: light: veml3235: extend regmap to add cache
Javier Carrasco [Tue, 24 Dec 2024 10:59:02 +0000 (11:59 +0100)]
iio: light: veml3235: extend regmap to add cache

The configuration and ID registers are not volatile and are not affected
by read operations (i.e. not precious), making them suitable to be
cached in order to reduce the number of accesses to the device.

Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241224-veml3235_scale-v2-3-2e1286846c77@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agoiio: light: veml3235: fix code style
Javier Carrasco [Tue, 24 Dec 2024 10:59:01 +0000 (11:59 +0100)]
iio: light: veml3235: fix code style

Trivial fixes to drop double spacings.

Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241224-veml3235_scale-v2-2-2e1286846c77@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agodt-bindings: iio: accel: adxl345: add interrupt-names
Lothar Rubusch [Wed, 25 Dec 2024 18:13:34 +0000 (18:13 +0000)]
dt-bindings: iio: accel: adxl345: add interrupt-names

Add interrupt-names INT1 and INT2 for the two interrupt lines of the
sensor.

When one of the two interrupt lines is connected, the interrupt as its
interrupt-name, need to be declared in the devicetree. The driver then
configures the sensor to indicate its events on either INT1 or INT2.

If no interrupt is configured, then no interrupt-name should be
configured, and vice versa. In this case the sensor runs in FIFO BYPASS
mode. This allows sensor measurements, but none of the sensor events.

Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://patch.msgid.link/20241225181338.69672-4-l.rubusch@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agodt-bindings: iio: accel: adxl345: make interrupts not a required property
Lothar Rubusch [Wed, 25 Dec 2024 18:13:33 +0000 (18:13 +0000)]
dt-bindings: iio: accel: adxl345: make interrupts not a required property

Remove interrupts from the list of required properties. The ADXL345
does not need interrupts for basic accelerometer functionality.

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>
Link: https://patch.msgid.link/20241225181338.69672-3-l.rubusch@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
8 months agodt-bindings: iio: imu: bmi323: add boolean type for drive-open-drain
Vasileios Amoiridis [Thu, 19 Dec 2024 23:47:45 +0000 (00:47 +0100)]
dt-bindings: iio: imu: bmi323: add boolean type for drive-open-drain

Add missing type description "boolean" for the drive-open-drain property.

Signed-off-by: Vasileios Amoiridis <vassilisamir@gmail.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://patch.msgid.link/20241219234745.58723-4-vassilisamir@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>