| Age | Commit message (Collapse) | Author |
|---|
|
Add debugfs register access to the ads112c14 driver. This is a complex
chip and being able to poke registers is useful for debugging and
diagnostic/calibration purposes.
Signed-off-by: David Lechner (TI) <dlechner@baylibre.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Add support for parsing devicetree properties for measurement channels
and doing direct reads on these.
There are quite a lot of conditions that have to be met for each
measurement to be made, so quite a bit of state and algorithms are
required to handle it.
Channels are created dynamically since the number of possibilities is
unreasonably large.
Signed-off-by: David Lechner (TI) <dlechner@baylibre.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Implement support for the programmable gain amplifier on the internal
short SYS_MON channel. This channel is used for calibration, so it is
useful to be able to set the PGA to the same gain as the external
channels. The gain setting is implemented via the `_scale` attribute.
In the future, we may want to support different reference voltages for
this channel, so the scale_available table is populated during probe
rather than being a static table.
Signed-off-by: David Lechner (TI) <dlechner@baylibre.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Add a new driver for the TI ADS112C14/ADS122C14 ADC chips.
This first step is adding a very basic driver that only supports power
on/reset and reading the system monitor channels.
ADS112C14_SYS_MON_CHANNEL_SHORT is the last channel rather than being in
logical order by address to keep the voltage channels together and in
case we find we need to add variants of this channel with different
voltage reference later.
Signed-off-by: David Lechner (TI) <dlechner@baylibre.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Babanpreet Singh <bbnpreetsingh@gmail.com> says:
Two small cleanups for the Axiado SPI driver, both defects present
since the driver was added in e75a6b00ad79 ("spi: axiado: Add driver
for Axiado SPI DB controller").
Patch 1 addresses the coccinelle cond_no_effect warning reported by
the kernel test robot on 2026-07-12 [1]: the RX-only/full-duplex
else-if arm and the trailing else arm in ax_transfer_one() have
identical bodies, so the second condition has no effect. The arms are
merged; no functional change (on x86_64/gcc 14, only
ax_transfer_one()'s object code changes, dropping the dead condition
evaluation).
Patch 2 fixes the kernel-doc comments in the driver: a copy-pasted
function name on ax_spi_get_rx_byte_for_irq(), a /** block with no
identifier line on ax_spi_process_rx_and_finalize(), and stale/missing
member documentation on struct ax_spi. This silences the driver's two
W=1 kernel-doc warnings plus five more visible when kernel-doc is run
on the header directly.
Both patches are against spi/for-next; they are independent of each
other except for touching the same file.
[1] https://lore.kernel.org/r/202607121827.djB0zLAj-lkp@intel.com/
Link: https://patch.msgid.link/20260713060807.7-1-bbnpreetsingh@gmail.com
|
|
Running kernel-doc -Wall on the driver reports 7 warnings, all present
since the driver was added in commit e75a6b00ad79 ("spi: axiado: Add
driver for Axiado SPI DB controller"). Two are in spi-axiado.c and
also show up in W=1 builds of drivers/spi:
Warning: drivers/spi/spi-axiado.c:226 expecting prototype for
ax_spi_get_rx_byte(). Prototype was for ax_spi_get_rx_byte_for_irq()
instead
Warning: drivers/spi/spi-axiado.c:248 This comment starts with
'/**', but isn't a kernel-doc comment. Refer to
Documentation/doc-guide/kernel-doc.rst
The other five are for struct ax_spi in spi-axiado.h, seen when
kernel-doc is invoked on the header directly (headers are not scanned
by the build-time kernel-doc checks):
Warning: drivers/spi/spi-axiado.h:130 Excess struct member 'rxbuf'
description in 'ax_spi'
Warning: drivers/spi/spi-axiado.h:130 Excess struct member 'txbuf'
description in 'ax_spi'
Warning: drivers/spi/spi-axiado.h:130 struct member 'clk_rate' not
described in 'ax_spi'
Warning: drivers/spi/spi-axiado.h:130 struct member 'rx_buf' not
described in 'ax_spi'
Warning: drivers/spi/spi-axiado.h:130 struct member 'tx_buf' not
described in 'ax_spi'
Fix the three underlying defects:
- The comment on ax_spi_get_rx_byte_for_irq() carries the name of
ax_spi_get_rx_byte(), from which it was evidently copied when the
IRQ variant was split out; fix the function name.
- The comment on ax_spi_process_rx_and_finalize() is opened with the
kernel-doc /** marker but has no identifier line; turn it into
proper kernel-doc (identifier line, @ctlr, Return:) since it
already describes the behavior and the return value.
- The struct ax_spi comment documents @txbuf and @rxbuf while the
members are named tx_buf and rx_buf, and does not document
clk_rate; fix the member names and describe clk_rate.
No functional change.
Assisted-by: Claude:claude-fable-5 [kernel-doc]
Signed-off-by: Babanpreet Singh <bbnpreetsingh@gmail.com>
Acked-By: Tzu-Hao Wei <twei@axiado.com>
Link: https://patch.msgid.link/20260713060807.7-3-bbnpreetsingh@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
The else-if arm taken for RX-only and full-duplex transfers and the
trailing else arm in the RX bookkeeping setup of ax_transfer_one()
have identical bodies, so the second condition has no effect:
drivers/spi/spi-axiado.c:433:8-10: WARNING: possible condition with
no effect (if == else)
The trailing else arm (neither TX nor RX buffer) is also unreachable:
the SPI core only calls the ->transfer_one() callback for transfers
that carry at least one buffer, see spi_transfer_one_message().
Merge the two arms into a single else branch and fold their comments.
No functional change. The redundant condition has been present since
the driver was added in commit e75a6b00ad79 ("spi: axiado: Add driver
for Axiado SPI DB controller").
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Julia Lawall <julia.lawall@inria.fr>
Closes: https://lore.kernel.org/r/202607121827.djB0zLAj-lkp@intel.com/
Assisted-by: Claude:claude-fable-5 [coccinelle]
Signed-off-by: Babanpreet Singh <bbnpreetsingh@gmail.com>
Link: https://patch.msgid.link/20260713060807.7-2-bbnpreetsingh@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Extend the kunit module for platform devices with test cases verifying
that the same software node can be added to platform devices repeatedly.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20260716-swnode-remove-on-dev-unreg-v8-5-5c2b8cc38c28@oss.qualcomm.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
|
|
With no drivers that manually assign software nodes to platform devices
created with platform_device_alloc(), we can now unify the release path
and remove platform_device_release_full().
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20260716-swnode-remove-on-dev-unreg-v8-4-5c2b8cc38c28@oss.qualcomm.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
|
|
This driver intentionally uses the fine-grained approach to creating
platform devices. It assigns a software node as the primary firmware
node of the device it creates. Ahead of improving the reference counting
of platform device software nodes, switch to using
device_create_managed_software_node(). This way, we create a dynamic
software node whose life-time is tied to the device to which it's
assigned.
Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20260716-swnode-remove-on-dev-unreg-v8-3-5c2b8cc38c28@oss.qualcomm.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
|
|
Creating a software node for a given set of properties and adding it to
a platform device can be achieved with a single call to
platform_device_register_full(). There's nothing in this driver that
suggests using the more fine-grained interfaces was intentional so
switch to using the high-level helper.
Acked-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20260716-swnode-remove-on-dev-unreg-v8-2-5c2b8cc38c28@oss.qualcomm.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
|
|
Calling platform_device_set_fwnode() (and by extension:
platform_device_set_of_node()) of platform_device_set_of_node_from_dev()
for a dynamically allocated platform device whose primary firmware node
is already assigned and is a software node leads to a resource leak as we
never perform the corresponding call to software_node_notify_remove(). As
there are currently no users for this special case - just disallow it and
warn loudly if it ever happens.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Link: https://patch.msgid.link/20260716-swnode-remove-on-dev-unreg-v8-1-5c2b8cc38c28@oss.qualcomm.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
|
|
Allow buffer blocks flagged as cyclic to be submitted as repeating DMA
transfers. For cyclic blocks, use DMA_PREP_REPEAT so the engine keeps
replaying the descriptor.
This is useful for output buffers where the same data should be driven
continuously without userspace having to requeue it. Examples include
continuous RF transmit paths replaying a calibration, test or beacon
pattern.
Skip installing the completion callback for cyclic blocks. Since the
transfer is continuously replayed, the callback would fire on every
period, throwing off the block refcount.
Because nothing prevents a new cyclic transfer from replacing an
already active cyclic one, always set DMA_PREP_LOAD_EOT so the engine
correctly terminates the active transfer before loading the new
descriptor.
Limit the DMA buffer queue to one cyclic DMABUF at a time. There is
currently no known use case for queueing multiple cyclic blocks, and
cyclic blocks stay referenced until the buffer is disabled.
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Replace sprintf() with sysfs_emit() for single-value sysfs show
functions, and use sysfs_emit_at() for lux_table_show() which
concatenates multiple values. This ensures buffer safety and
follows the modern kernel sysfs API.
Signed-off-by: Chengrui Liu <OrangeBlack0765@outlook.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Most of the supported devices rely on a GAIN pin to control a 2x
multiplier applied to the output voltage. Other devices, e.g. the
single-channel ones, provides a gain control through a bit field in
the control register. Some designs might have the GAIN pin hardwired
to VDD/VLOGIC or GND, which would have no "gain-gpios" device property,
being able to set "adi,range-double" if it is hardwired to VDD. The
vref_mv field is moved down in the struct ad5686_state, so that the
overall size increase is reduced.
Reviewed-by: David Lechner <dlechner@baylibre.com>
Signed-off-by: Rodrigo Alencar <rodrigo.alencar@analog.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Implement trigger handler by leveraging the LDAC gpio to update all DAC
channels at once when it is available. Also, the multiple channel writes
can be flushed at once with the sync() operation.
Reviewed-by: David Lechner <dlechner@baylibre.com>
Signed-off-by: Rodrigo Alencar <rodrigo.alencar@analog.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Use guarded mutex lock to facilitate code review when adding new
attributes. This will allow for early returns, avoiding error-prone
locking and unlocking in error paths. This also adds missing include
linux/cleanup.h. Gain-control support will allow the scale attribute
to be configurable.
Reviewed-by: Maxwell Doose <m32285159@gmail.com>
Reviewed-by: Joshua Crofts <joshua.crofts1@gmail.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Reviewed-by: David Lechner <dlechner@baylibre.com>
Signed-off-by: Rodrigo Alencar <rodrigo.alencar@analog.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Use of local SPI bus data to manage a collection of SPI transfers and
flush them to the SPI platform driver with the sync() operation. This
allows for faster handling of multiple channel DAC writes, avoiding kernel
overhead per spi_sync() call, which will be helpful when enabling
triggered buffer support. For consistency, the read operation leverages
the same transfer data rather than a stack-allocated array.
Reviewed-by: David Lechner <dlechner@baylibre.com>
Signed-off-by: Rodrigo Alencar <rodrigo.alencar@analog.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Add sync() to operation to ad5686_bus_ops, which can be used to flush
multiple pending data transfers at once. This is going to be used when
implementing triggered buffer support.
Reviewed-by: David Lechner <dlechner@baylibre.com>
Signed-off-by: Rodrigo Alencar <rodrigo.alencar@analog.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Replace usage of bit shifting macros for FIELD_PREP(), which would not
ignore bit masking when preparing SPI/I2C commands. This change is a code
hardening measure to be paired with the upcoming triggered buffer support.
For the AD5310 regmap case, 16-bit data coming from the buffer may overlap
with command bits if the data field is unmasked.
*_REF_BIT_MSK and *_PD_MSK bit position macros are renamed (with a DATA
prefix) so to indicate that they are relative to the DATA field.
Reported-by: sashiko-bot@kernel.org
Closes: https://lore.kernel.org/all/20260628143026.EC6CA1F000E9@smtp.kernel.org/
Reviewed-by: David Lechner <dlechner@baylibre.com>
Signed-off-by: Rodrigo Alencar <rodrigo.alencar@analog.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Verify that chip_info pointer is not NULL. If a user binds the driver
using driver_override via sysfs with a device name not present in the
id_table or of_match_table, match data will be NULL.
Fixes: 0eb1728461a1 ("iio: dac: ad5686: drop enum id")
Reported-by: sashiko-bot@kernel.org
Closes: https://lore.kernel.org/all/20260710113149.53EC51F000E9@smtp.kernel.org/
Signed-off-by: Rodrigo Alencar <rodrigo.alencar@analog.com>
Reviewed-by: Joshua Crofts <joshua.crofts1@gmail.com>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
Robert Marko <robert.marko@sartura.hr> says:
This series adds QSPI support for the Microchip LAN969x which has two QSPI
controllers based on SAMA7G5.
It requires pad calibration, supports DMA, and supports 100 MHz operation.
It requires a different init sequence to SAMA7G5 so support for custom
.init per controller caps is added, and existing controllers relying on
checking gclk are converted to it.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Link: https://patch.msgid.link/20260709112006.390742-1-robert.marko@sartura.hr
|
|
Microchip LAN969x has two QSPI controllers based on SAMA7G5 QSPI.
It requires pad calibration, supports DMA, and supports 100 MHz operation.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Link: https://patch.msgid.link/20260709112006.390742-5-robert.marko@sartura.hr
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Assign the existing SAMA7G5 initialization routine to every generic
clock variant and dispatch initialization exclusively through the
capability callback.
This keeps hardware capabilities separate from initialization selection
and lets variants override the sequence explicitly.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Link: https://patch.msgid.link/20260709112006.390742-4-robert.marko@sartura.hr
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Allow controller variants to provide a custom initialization callback
through their capability data.
This prepares the driver for variants which require a different hardware
initialization sequence without adding SoC checks to the common path.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Link: https://patch.msgid.link/20260709112006.390742-3-robert.marko@sartura.hr
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
In k1_spi_write, the count variable is intended to represent the number
of slots available for writing into the TX FIFO.
The current implementation uses FIELD_GET(SSP_STATUS_TFL, val) in an
attempt to determine this count, but this register field returns the
number of occupied slots, not the available space. The previous
implementation attempted to handle this via a ternary operator (? :
K1_SPI_FIFO_SIZE), which incorrectly assumed that the hardware returned
0 when the FIFO was empty (meaning all slots were available), leading to
incorrect accounting of the buffer space.
Fix this by calculating the free slots: count = K1_SPI_FIFO_SIZE -
FIELD_GET(SSP_STATUS_TFL, val);
The associated comment has been updated to reflect the logic change: The
old comment reflected an incorrect assumption about the hardware
behavior, which was the root cause of the previous buggy logic.
This patch accurately and concisely describes the purpose of the new
calculation.
Signed-off-by: Peixin Xie <peixin.xie@spacemit.com>
Signed-off-by: Zhengyu He <hezhy472013@gmail.com>
Link: https://patch.msgid.link/20260715-k1-spi-tx-fifo-fix-v1-for-next-v1-1-02024223b08a@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
rzg2l_gpt_calculate_prescale()
The rzg2l_gpt parameter was passed to rzg2l_gpt_calculate_prescale() but
never used inside the function. Remove it and update the sole call site
accordingly.
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Link: https://patch.msgid.link/20260604095647.108654-6-biju.das.jz@bp.renesas.com
Signed-off-by: Uwe Kleine-König <ukleinek@kernel.org>
|
|
Single characters should occasionally be put into a sequence.
Thus use the corresponding function `seq_putc()`.
The source code was transformed by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Link: https://patch.msgid.link/5cc95f1d-a9f4-4ba4-8411-56cfb16d996a@web.de
[ukleinek: fixup for a third instance, originally addressed differently]
Link: https://patch.msgid.link/03062fdb-cc4f-47c7-a007-4ac67ded6377@web.de
Signed-off-by: Uwe Kleine-König <ukleinek@kernel.org>
|
|
On some SoCs (e.g. R-Car S4) the endpoint-side eDMA raises a single
fixed SPI that is hardwired to CPU0 and covers every read and write
channel. Handling channel events directly in that hard IRQ context
serializes the completion processing of all channels on one CPU:
descriptor recycling and refill, client callbacks (the vchan tasklet
runs on the scheduling CPU) and the doorbell writes all funnel through
CPU0, while the handler additionally spins on each channel's vc.lock.
Especially under heavy multichannel load, this contention becomes a
performance bottleneck.
Keep the hard IRQ handler minimal: clear the status, dispatch channel
events, and defer per-channel processing to work items. A work item per
channel preserves ordering while allowing different channels to run in
parallel on any CPU.
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-10-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
The v0 interrupt handler reads the interrupt status register twice per
invocation, once through the DONE accessor and once through the ABORT
accessor, although both fields live in the same 32-bit register. On
remote setups (dw-edma-pcie) each read is a non-posted round trip across
the PCIe link costing on the order of a microsecond, and with one
completion interrupt per element the duplicate adds up. As an example,
profiling the R-Car S4 remote path put the handler at ~7us per
invocation, dominated by such reads.
Read the register once and derive the DONE and ABORT views from the
snapshot. No abort is lost to this because the pass only clears status
bits it observed, so an abort raised after the snapshot keeps its status
and its own interrupt delivery brings it to the next pass. An abort on
an observed channel cannot race the clear either. Software can restart
the halted channel only after abort() runs, and abort() is called after
dw_edma_v0_core_clear_abort_int().
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-9-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
dw_edma_pcie enables the PCI device with pcim_enable_device(), so IRQ
vectors allocated by pci_alloc_irq_vectors() are released by
pcim_msi_release() on device release. The driver should not call
pci_free_irq_vectors() manually.
Drop the redundant remove-time cleanup and rely on the managed PCI
device lifetime instead, as documented by commit 03e4905402ae ("PCI/MSI:
Clarify pci_free_irq_vectors() usage for managed devices").
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-8-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
terminate_all() can finish immediately when the channel is unconfigured,
paused, idle, or already stopped in hardware. A pending PAUSE request can
survive these paths and block issue_pending() even after termination.
Clear the request whenever termination leaves the channel idle. A running
channel keeps its STOP request until the interrupt handler consumes it.
Fixes: e63d79d1ffcd ("dmaengine: Add Synopsys eDMA IP core driver")
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-7-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
pause() and resume() read and update channel state without holding vc.lock,
while the interrupt handlers update the same state under it. Take the same
lock around those state checks so that request, status, and configured stay
consistent.
For example, pause() can observe EDMA_ST_BUSY right before the interrupt
handler completes the final descriptor and moves the channel to
EDMA_ST_IDLE, and then record EDMA_REQ_PAUSE on an already idle channel. No
further interrupt will acknowledge the request, and since issue_pending()
requires EDMA_REQ_NONE, the channel is wedged for good: terminate_all()
leaves the stale request behind, so even reconfiguring the channel does not
recover it.
issue_pending() already runs under vc.lock, but it tests configured before
taking it. Move that test under the lock as well, so configured, request,
and status are evaluated as one channel-state snapshot.
Fixes: e63d79d1ffcd ("dmaengine: Add Synopsys eDMA IP core driver")
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-6-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
If PAUSE is requested while the final burst of a descriptor is in
flight, the DONE interrupt takes the PAUSE path without checking whether
the descriptor has been depleted. The depleted descriptor remains on the
issued list and the channel enters EDMA_ST_PAUSE.
On resume, dw_edma_start_transfer() can select that depleted descriptor
again even though no burst remains, leaving the channel in an invalid
busy state.
Check for descriptor completion before acknowledging PAUSE. If there is
no work to start on resume, leave the channel idle. Also ignore DONE
interrupts while the channel is paused so a stale or repeated interrupt
cannot change its state or start queued work.
Fixes: e63d79d1ffcd ("dmaengine: Add Synopsys eDMA IP core driver")
Cc: stable@vger.kernel.org
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-5-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
dw_edma_abort_interrupt() drops vc.lock before changing request and
status. issue_pending() can acquire the lock in that small window,
observe the old busy state, and skip starting queued descriptors. Then
the abort handler overwrites the channel status as idle, leaving the new
descriptors stranded for good.
Keep descriptor completion and the state transition in the same critical
section.
Fixes: e63d79d1ffcd ("dmaengine: Add Synopsys eDMA IP core driver")
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-4-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
The DMA Engine client documentation says in the "Terminate APIs" section
of Documentation/driver-api/dmaengine/client.rst:
"No callback functions will be called for any incomplete transfers."
dw-edma instead calls vchan_cookie_complete() when a deferred STOP reaches
the interrupt handler. This schedules a callback for the active descriptor
and leaves other issued or submitted descriptors queued. A late callback
after dmaengine_terminate_sync() can dereference client state that has
already been freed, while leftover descriptors may later restart into
reused buffers or leak.
Move all issued and submitted descriptors to the terminated list whenever
termination completes. For a pending STOP, do this from both the DONE and
ABORT paths. Complete their cookies in order without scheduling callbacks.
A STOP can remain pending until the running transfer raises an
interrupt. Make device_synchronize() wait for such a pending STOP to
complete before releasing terminated descriptors. Reuse it from
free_chan_resources(), then release the remaining virt-dma resources.
Sleep instead of busy-polling while waiting, and warn if the existing
timeout expires.
Fixes: e63d79d1ffcd ("dmaengine: Add Synopsys eDMA IP core driver")
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-3-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
GET_CH_32() takes the direction before the channel ID, but
dw_hdma_v0_core_ch_status() passed them in the opposite order. This can
make the status callback read another HDMA channel status register.
Use the same argument order as the other HDMA register accesses.
Fixes: e74c39573d35 ("dmaengine: dw-edma: Add support for native HDMA")
Cc: stable@vger.kernel.org
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Link: https://patch.msgid.link/20260717180639.2643243-2-den@valinux.co.jp
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux
Pull block fixes from Jens Axboe:
- Fixes for the dio bounce buffer helpers: correct the alignment of
bounced dio read bios to avoid a double unpin, handle huge zero
folios in bio_free_folios(), and don't warn on the larger-order folio
attempts in the greedy allocation path.
- Try a slab allocation in bio_alloc_bioset() before falling back to
the mempool, restoring the previous behavior for non-sleeping
allocations from a cache-enabled bioset.
- Serialize elevator changes for the same queue using the writer lock.
- Fix a race in blk_time_get_ns() where a task preempted between
setting PF_BLOCK_TS and the cached-timestamp reload could return 0.
- blk-cgroup fix for leaks and the online flag on a radix_tree_insert()
failure in blkg_create().
- Free the copied pages when blk_rq_map_kern() fails after
blk_rq_append_bio() rejects the bio.
- Remove manually added partitions on loop device detach, fixing dead
partition devices left behind and a subsequent LOOP_CONFIGURE -EBUSY
- Bound the AIX partition lvd scan to the sector that was actually
read.
- Show the block operation in error injection rules (Jackie)
* tag 'block-7.2-20260717' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
block: fix aligning of bounced dio read bios
block: handle huge zero folios in bio_free_folios
block: try slab allocation in bio_alloc_bioset() before mempool
block: show operation in error injection rules
block: serialize elevator changes for the same queue using a writer lock
block: free copied pages when blk_rq_map_kern() fails
block: do not warn when doing greedy allocation in folio_alloc_greedy()
partitions: aix: bound the lvd scan to one sector
blk-cgroup: fix leaks and online flag on radix_tree_insert failure
loop: remove manually added partitions on detach
block: fix race in blk_time_get_ns() returning 0
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Pull spi fixes from Mark Brown:
"A couple of fairly routine driver fixes, nothing too remarkable"
* tag 'spi-fix-v7.2-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
spi: cadence-quadspi: Fix indirect write timeout when DMA read mode is enabled
spi: dw-dma: Wait for controller idle before completing Tx
|
|
ub->completion is only re-armed by a successful START_USER_RECOVERY. If
the ublk server sends END_USER_RECOVERY without one - e.g. its START
failed with -EBUSY and the error was ignored - the wait is satisfied by
the stale completion of the previous recovery cycle, and the device is
marked LIVE and the requeue list kicked while the FETCH stream is still
running and ubq->canceling is still set. The kick redispatches a
previously requeued request, __ublk_queue_rq_common() sees ->canceling
and parks it again via __ublk_abort_rq(), and after the last FETCH
clears ->canceling nothing ever kicks the requeue list again: the
request is stranded there while holding its tag. If it is the flush
machinery's flush_rq, every subsequent fsync piles up in uninterruptible
sleep and teardown hangs on tag draining. This matches a report of a
lost PREFLUSH with ext4 on top of ublk after daemon crash recovery.
ub->completion is an edge-triggered latch used as a proxy for the level
condition "every queue has fetched all I/O commands", which can regress
(F_BATCH's UNPREP, daemon death) and whose re-arm can be skipped. Drop
it and wait on the real condition instead: the new helper
ublk_wait_dev_ready_and_lock() waits on ublk_dev_ready() via
wait_var_event_interruptible(), woken from ublk_mark_io_ready(), then
re-checks it under ub->mutex, waiting again on regression, and returns
with the mutex held and readiness guaranteed.
Readiness becomes true in the same ub->mutex critical section that
clears the last queue's ->canceling, so END_USER_RECOVERY marks the
device LIVE and kicks the requeue list strictly after ->canceling
clears. The wait stays interruptible, so a server whose daemon died can
still be signalled out. For ublk_ctrl_start_dev() this replaces the
fail-fast -EINVAL on an F_BATCH ready->UNPREP regression with waiting
until the device is ready again.
Reported-by: George Salisbury <gsalisbury@apnic.net>
Fixes: 728cbac5fe21 ("ublk: move device reset into ublk_ch_release()")
Cc: stable@vger.kernel.org
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Link: https://patch.msgid.link/20260719134540.120269-1-tom.leiming@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator
Pull regulator fix from Mark Brown:
"One straightforward driver fix for some incorrectly described
bitfields in the ltc3676 driver"
* tag 'regulator-fix-v7.2-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
regulator: ltc3676: Fix incorrect IRQSTAT bit offsets
|
|
Add missing return value check in the setup function.
Fixes: d0ddfd241e57 ("hwmon: (asus-ec-sensors) add driver for ASUS EC")
Signed-off-by: Eugene Shalygin <eugene.shalygin@gmail.com>
Link: https://lore.kernel.org/r/20260712130602.1256700-2-eugene.shalygin@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Take INITIAL_JIFFIES into account when setting up next update time.
Fixes: d0ddfd241e57 ("hwmon: (asus-ec-sensors) add driver for ASUS EC")
Signed-off-by: Eugene Shalygin <eugene.shalygin@gmail.com>
Link: https://lore.kernel.org/r/20260712110650.1240071-2-eugene.shalygin@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Do not assume there are only bank 0 and bank 1 available, just use '!='
for bank comparison.
Fixes: d0ddfd241e57 ("hwmon: (asus-ec-sensors) add driver for ASUS EC")
Signed-off-by: Eugene Shalygin <eugene.shalygin@gmail.com>
Link: https://lore.kernel.org/r/20260711074217.554656-1-eugene.shalygin@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
MAX34451 and ADPM chips do not support standard PMBus VIN/IIN limit
registers, manufacturer specific min/max registers, or undercurrent
or undertemperature fault limits. STATUS_BYTE and STATUS_OTHER are also
not available. Accessing these non-existent registers during driver
initialization triggers a CML error and asserts ALERT. Handled by
blocking these functions during read/write.
Fixes: 7a001dbab4ad ("hwmon: (pmbus/max34440) Add support for MAX34451.")
Fixes: 629cf8f6c23a ("hwmon: (pmbus/max34440) Add support for ADPM12160")
Fixes: 2e0b52f1ae88 ("hwmon: (pmbus/max34440): add support adpm12200")
Fixes: 479bfeba2eb6 ("hwmon: (pmbus/max34440): add support adpm12250")
Signed-off-by: Alexis Czezar Torreno <alexisczezar.torreno@analog.com>
Link: https://lore.kernel.org/r/20260716-max34451_fixes-v1-1-a941b27eaecb@analog.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Use dev_err_probe() instead of dev_err() to simplify the error path
and cleanly handle deferred probing.
Signed-off-by: Vojtěch Krátký <vo.kratky@seznam.cz>
Signed-off-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
|
|
The qcom_cc_driver_data and critical CBCRs list are never modified by
common code and are expected to be const. Hence add const qualifier for
these fields.
Signed-off-by: Jagadeesh Kona <jagadeesh.kona@oss.qualcomm.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260714-glymur_camcc_const_fixes-v1-1-c635123ebbeb@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
We have a dedicated function to fetch all the BIOS tables when the opmode
starts, and yet we read a couple of tables directly from
iwl_op_mode_mld_start, which is already a large function that does
multiple things.
Move the reading of the sgom, puncturing, and RFI enablement to the
dedicated iwl_mld_get_bios_tables.
Link: https://patch.msgid.link/20260717172958.b19a33e0b507.I73f6b5e6a81d0f411f12589ceb30afa655c0a16b@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
|
|
Add a radio debug trace when MCC_ALLOWED_AP_TYPE_CMD is sent
successfully during AP type table initialization.
This improves bring-up visibility without changing runtime behavior.
Failures are still reported through the existing error log path.
Signed-off-by: Pagadala Yesu Anjaneyulu <pagadala.yesu.anjaneyulu@intel.com>
Link: https://patch.msgid.link/20260717172958.18e1fc5ec109.I76dd832f62d00a8f358f8e4a705f25184ac53da2@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
|
|
New firmware will support version 2 of the update_mcc notification. The
extra field is used for a new feature, but we does not support it.
Keep the existing payload definition compatible with both versions and
register version 2 in the MLD notification version table so the driver
accepts the newer notification without changing the behavior.
This preserves version 1 support and adds compatibility with firmware
that sends version 2.
Signed-off-by: Pagadala Yesu Anjaneyulu <pagadala.yesu.anjaneyulu@intel.com>
Link: https://patch.msgid.link/20260717172958.9c5a940d37dc.I955800c2377b802ffb99003349552cc4036ca4bd@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
|
