There are no more users of the deprecated is_dma_mapped in struct
spi_message so it can be removed.
References in documentation and comments are also removed.
A few similar checks if xfer->tx_dma or xfer->rx_dma are not NULL are
also removed since these are now guaranteed to be NULL because they
were previously set only if is_dma_mapped was true.
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://msgid.link/r/20240325-spi-remove-is_dma_mapped-v2-1-d08d62b61f1c@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
SPI controller with integrated chip select handling still need to adhere
to SPI device's CS setup, hold and inactive delays. For controller
without set_cs_timing spi core shall handle the delays to avoid
duplicated delay handling in each controller driver.
Fixes a regression for the out of tree SPI controller and SPI HID
transport on Apple M1/M1 Pro/Max notebooks.
Fixes: 4d8ff6b099 ("spi: Add multi-cs memories support in SPI core")
Signed-off-by: Janne Grunau <j@jannau.net>
Link: https://msgid.link/r/20240311-spi-cs-delays-regression-v1-1-0075020a90b2@jannau.net
Signed-off-by: Mark Brown <broonie@kernel.org>
Splitting transfers is an expensive operation so we can potentially
optimize it by doing it only once per optimization of the message
instead of repeating each time the message is transferred.
The transfer splitting functions are currently the only user of
spi_res_alloc() so spi_res_release() can be safely moved at this time
from spi_finalize_current_message() to spi_unoptimize_message().
The doc comments of the public functions for splitting transfers are
also updated so that callers will know when it is safe to call them
to ensure proper resource management.
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://msgid.link/r/20240219-mainline-spi-precook-message-v2-2-4a762c6701b9@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This adds a new spi_optimize_message() function that can be used to
optimize SPI messages that are used more than once. Peripheral drivers
that use the same message multiple times can use this API to perform SPI
message validation and controller-specific optimizations once and then
reuse the message while avoiding the overhead of revalidating the
message on each spi_(a)sync() call.
Internally, the SPI core will also call this function for each message
if the peripheral driver did not explicitly call it. This is done to so
that controller drivers don't have to have multiple code paths for
optimized and non-optimized messages.
A hook is provided for controller drivers to perform controller-specific
optimizations.
Suggested-by: Martin Sperl <kernel@martin.sperl.org>
Link: https://lore.kernel.org/linux-spi/39DEC004-10A1-47EF-9D77-276188D2580C@martin.sperl.org/
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://msgid.link/r/20240219-mainline-spi-precook-message-v2-1-4a762c6701b9@baylibre.com
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Uwe Kleine-König <u.kleine-koenig@pengutronix.de>:
This series finishes off the removal of some of the legacy names for
SPI controllers and devices.
The __spi_split_transfer_maxsize() function has a gpf argument to allow
callers to specify the type of memory allocation that needs to be used.
However, this function only allocates struct spi_transfer and is not
intended to be used from atomic contexts so this type should always be
GFP_KERNEL, so we can just drop the argument.
Some callers of these functions also passed GFP_DMA, but since only
struct spi_transfer is allocated and not any tx/rx buffers, this is
not actually necessary and is removed in this commit.
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://lore.kernel.org/r/20240206200648.1782234-1-dlechner@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This moves splitting transfers for CS_WORD software emulation to the
same place where we split transfers for controller-specific reasons.
This fixes a few subtle bugs.
The calculation for maxsize was wrong for bit sizes between 17 and 24.
This is fixed by making use of spi_split_transfers_maxwords() which
already has the correct calculation.
Also, since this indirectly calls spi_res_alloc(), to avoid leaking
resources, spi_finalize_current_message() would need to be called
on all error paths in __spi_validate() and callers of __spi_validate()
would need to do the same. This is fixed by moving the call to
__spi_pump_transfer_message() where it is already splitting transfers
for other reasons and correctly releases resources in the subsequent
error paths.
Fixes: cbaa62e009 ("spi: add software implementation for SPI_CS_WORD")
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://lore.kernel.org/r/20240126212358.3916280-2-dlechner@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The __spi_sync() function calls __spi_validate() early in the function.
Later, it can call spi_async_locked() which calls __spi_validate()
again. __spi_validate() is an expensive function, so we can improve
performance measurably by avoiding calling it twice.
Instead of calling spi_async_locked(), we can call __spi_async() with
the spin lock held.
spi_async_locked() is removed since there are no more callers.
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://msgid.link/r/20240125234732.3530278-2-dlechner@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In __spi_pump_transfer_message(), the message was not finalized in the
first error return as it is in the other error return paths. Not
finalizing the message could cause anything waiting on the message to
complete to hang forever.
This adds the missing call to spi_finalize_current_message().
Fixes: ae7d2346dc ("spi: Don't use the message queue if possible in spi_sync")
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://msgid.link/r/20240125205312.3458541-2-dlechner@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Previously, __spi_sync() and __spi_async() set message->spi to the spi
device independently after calling __spi_validate(). __spi_validate()
also would conditionally set this if it needed to split the message
since it wasn't set yet.
Since both __spi_sync() and __spi_async() call __spi_validate(), we can
consolidate this into only setting message->spi once (unconditionally)
in __spi_validate(). This will also save any future callers of
__spi_validate() from also needing to set message->spi.
Signed-off-by: David Lechner <dlechner@baylibre.com>
Link: https://msgid.link/r/20240123214946.2616786-1-dlechner@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Amit Kumar Mahapatra <amit.kumar-mahapatra@amd.com>:
This patch series adds support to the SPI framework for using multiple
chip selects.
AMD-Xilinx GQSPI controller has two advanced mode that allows the
controller to consider two flashes as one single device.
One of these two mode is the parallel mode in which each byte of data is
stored in both devices, the even bits in the lower flash & the odd bits in
the upper flash. The byte split is automatically handled by the QSPI
controller.
The other mode is the stacked mode in which both the flashes share the
same SPI bus but each of the device contain half of the data. In this mode,
the controller does not follow CS requests but instead internally wires the
two CS levels with the value of the most significant address bit.
For supporting both these modes SPI core need to be updated for providing
multiple CS for a single SPI device.
For adding multi CS support the SPI device need to be aware of all the CS
values. So, the "chip_select" member in the spi_device structure is now an
array that holds all the CS values.
spi_device structure now has a "cs_index_mask" member. This acts as an
index to the chip_select array. If nth bit of spi->cs_index_mask is set
then the driver would assert spi->chip_select[n].
In parallel mode all the chip selects are asserted/de-asserted
simultaneously and each byte of data is stored in both devices, the even
bits in one, the odd bits in the other. The split is automatically handled
by the GQSPI controller. The GQSPI controller supports a maximum of two
flashes connected in parallel mode. A SPI_CONTROLLER_MULTI_CS flag bit is
added in the spi controller flags, through ctlr->flags the spi core
will make sure that the controller is capable of handling multiple chip
selects at once.
For supporting multiple CS via GPIO the cs_gpiod member of the spi_device
structure is now an array that holds the gpio descriptor for each
chipselect.
CS GPIO is not tested on our hardware, but it has been tested by @Stefan
https://lore.kernel.org/all/005001da1efc$619ad5a0$24d080e0$@opensource.cirrus.com/
Signed-off-by: Amit Kumar Mahapatra <amit.kumar-mahapatra@amd.com>
Tested-by: Stefan Binding <sbinding@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20231125092137.2948-4-amit.kumar-mahapatra@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>
A race condition exists where a synchronous (noqueue) transfer can be
active during a system suspend. This can cause a null pointer
dereference exception to occur when the system resumes.
Example order of events leading to the exception:
1. spi_sync() calls __spi_transfer_message_noqueue() which sets
ctlr->cur_msg
2. Spi transfer begins via spi_transfer_one_message()
3. System is suspended interrupting the transfer context
4. System is resumed
6. spi_controller_resume() calls spi_start_queue() which resets cur_msg
to NULL
7. Spi transfer context resumes and spi_finalize_current_message() is
called which dereferences cur_msg (which is now NULL)
Wait for synchronous transfers to complete before suspending by
acquiring the bus mutex and setting/checking a suspend flag.
Signed-off-by: Mark Hasemeyer <markhas@chromium.org>
Link: https://lore.kernel.org/r/20231107144743.v1.1.I7987f05f61901f567f7661763646cb7d7919b528@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@kernel.org
Export acpi_spi_find_controller_by_adev() so that ACPI glue code which
wants to dynamically create a spi_device using acpi_spi_device_alloc() or
spi_new_device() on a controller, to which the code does not already have
a reference, can find the controller.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20231014205314.59333-2-hdegoede@redhat.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Pull char/misc drivers updates from Greg KH:
"Here is the "big" set of char/misc and other driver subsystems for
6.4-rc1.
It's pretty big, but due to the removal of pcmcia drivers, almost
breaks even for number of lines added vs. removed, a nice change.
Included in here are:
- removal of unused PCMCIA drivers (finally!)
- Interconnect driver updates and additions
- Lots of IIO driver updates and additions
- MHI driver updates
- Coresight driver updates
- NVMEM driver updates, which required some OF updates
- W1 driver updates and a new maintainer to manage the subsystem
- FPGA driver updates
- New driver subsystem, CDX, for AMD systems
- lots of other small driver updates and additions
All of these have been in linux-next for a while with no reported
issues"
* tag 'char-misc-6.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (196 commits)
mcb-lpc: Reallocate memory region to avoid memory overlapping
mcb-pci: Reallocate memory region to avoid memory overlapping
mcb: Return actual parsed size when reading chameleon table
kernel/configs: Drop Android config fragments
virt: acrn: Replace obsolete memalign() with posix_memalign()
spmi: Add a check for remove callback when removing a SPMI driver
spmi: fix W=1 kernel-doc warnings
spmi: mtk-pmif: Drop of_match_ptr for ID table
spmi: pmic-arb: Convert to platform remove callback returning void
spmi: mtk-pmif: Convert to platform remove callback returning void
spmi: hisi-spmi-controller: Convert to platform remove callback returning void
w1: gpio: remove unnecessary ENOMEM messages
w1: omap-hdq: remove unnecessary ENOMEM messages
w1: omap-hdq: add SPDX tag
w1: omap-hdq: allow compile testing
w1: matrox: remove unnecessary ENOMEM messages
w1: matrox: use inline over __inline__
w1: matrox: switch from asm to linux header
w1: ds2482: do not use assignment in if condition
w1: ds2482: drop unnecessary header
...
Pull driver core updates from Greg KH:
"Here is the large set of driver core changes for 6.4-rc1.
Once again, a busy development cycle, with lots of changes happening
in the driver core in the quest to be able to move "struct bus" and
"struct class" into read-only memory, a task now complete with these
changes.
This will make the future rust interactions with the driver core more
"provably correct" as well as providing more obvious lifetime rules
for all busses and classes in the kernel.
The changes required for this did touch many individual classes and
busses as many callbacks were changed to take const * parameters
instead. All of these changes have been submitted to the various
subsystem maintainers, giving them plenty of time to review, and most
of them actually did so.
Other than those changes, included in here are a small set of other
things:
- kobject logging improvements
- cacheinfo improvements and updates
- obligatory fw_devlink updates and fixes
- documentation updates
- device property cleanups and const * changes
- firwmare loader dependency fixes.
All of these have been in linux-next for a while with no reported
problems"
* tag 'driver-core-6.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (120 commits)
device property: make device_property functions take const device *
driver core: update comments in device_rename()
driver core: Don't require dynamic_debug for initcall_debug probe timing
firmware_loader: rework crypto dependencies
firmware_loader: Strip off \n from customized path
zram: fix up permission for the hot_add sysfs file
cacheinfo: Add use_arch[|_cache]_info field/function
arch_topology: Remove early cacheinfo error message if -ENOENT
cacheinfo: Check cache properties are present in DT
cacheinfo: Check sib_leaf in cache_leaves_are_shared()
cacheinfo: Allow early level detection when DT/ACPI info is missing/broken
cacheinfo: Add arm64 early level initializer implementation
cacheinfo: Add arch specific early level initializer
tty: make tty_class a static const structure
driver core: class: remove struct class_interface * from callbacks
driver core: class: mark the struct class in struct class_interface constant
driver core: class: make class_register() take a const *
driver core: class: mark class_release() as taking a const *
driver core: remove incorrect comment for device_create*
MIPS: vpe-cmp: remove module owner pointer from struct class usage.
...
Pull spi updates from Mark Brown:
"A fairly standard release for SPI with the exception of a change to
the API for specifying chip selects done in preparation for supporting
devices with more than one chip select, this required some mechanical
changes throughout the tree which have been cooking in -next happily
for a while.
There's also a new API to allow us to support TPM chips on half duplex
controllers.
Summary:
- Refactoring in preparation for supporting multiple chip selects for
a single device, needed by some flash devices, which required a
change in the SPI device API visible throughout the tree
- Support for hardware assisted interaction with SPI TPMs on half
duplex controllers, implemented on nVidia Tedra210 QuadSPI
- Optimisation for large transfers on fsl-cpm devices
- Cleanups around device property use which fix some sisues with
fwnode
- Use of both void remove() and devm_platform_.*ioremap_resource()
- Support for AMD Pensando Elba, Amlogic A1, Cadence device mode,
Intel MetorLake-S and StarFive J7110 QuadSPI"
* tag 'spi-v6.4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (185 commits)
spi: bcm63xx: use macro DEFINE_SIMPLE_DEV_PM_OPS
spi: tegra210-quad: Enable TPM wait polling
spi: Add TPM HW flow flag
spi: bcm63xx: remove PM_SLEEP based conditional compilation
spi: cadence-quadspi: use macro DEFINE_SIMPLE_DEV_PM_OPS
spi: spi-cadence: Add support for Slave mode
spi: spi-cadence: Switch to spi_controller structure
spi: cadence-quadspi: fix suspend-resume implementations
spi: dw: Add support for AMD Pensando Elba SoC
spi: dw: Add AMD Pensando Elba SoC SPI Controller
spi: cadence-quadspi: Disable the SPI before reconfiguring
spi: cadence-quadspi: Update the read timeout based on the length
spi: spi-loopback-test: Add module param for iteration length
spi: add support for Amlogic A1 SPI Flash Controller
dt-bindings: spi: add Amlogic A1 SPI controller
spi: fsl-spi: No need to check transfer length versus word size
spi: fsl-spi: Change mspi_apply_cpu_mode_quirks() to void
spi: fsl-cpm: Use 16 bit mode for large transfers with even size
spi: fsl-spi: Re-organise transfer bits_per_word adaptation
spi: fsl-spi: Fix CPM/QE mode Litte Endian
...
Pull devicetree fixes from Rob Herring:
- Fix interaction between fw_devlink and DT overlays causing devices to
not be probed
- Fix the compatible string for loongson,cpu-interrupt-controller
* tag 'devicetree-fixes-for-6.2-3' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
treewide: Fix probing of devices in DT overlays
dt-bindings: interrupt-controller: loongarch: Fix mismatched compatible
When loading a DT overlay that creates a device, the device is not
probed, unless the DT overlay is unloaded and reloaded again.
After the recent refactoring to improve fw_devlink, it no longer depends
on the "compatible" property to identify which device tree nodes will
become struct devices. fw_devlink now picks up dangling consumers
(consumers pointing to descendent device tree nodes of a device that
aren't converted to child devices) when a device is successfully bound
to a driver. See __fw_devlink_pickup_dangling_consumers().
However, during DT overlay, a device's device tree node can have
sub-nodes added/removed without unbinding/rebinding the driver. This
difference in behavior between the normal device instantiation and
probing flow vs. the DT overlay flow has a bunch of implications that
are pointed out elsewhere[1]. One of them is that the fw_devlink logic
to pick up dangling consumers is never exercised.
This patch solves the fw_devlink issue by marking all DT nodes added by
DT overlays with FWNODE_FLAG_NOT_DEVICE (fwnode that won't become
device), and by clearing the flag when a struct device is actually
created for the DT node. This way, fw_devlink knows not to have
consumers waiting on these newly added DT nodes, and to propagate the
dependency to an ancestor DT node that has the corresponding struct
device.
Based on a patch by Saravana Kannan, which covered only platform and spi
devices.
[1] https://lore.kernel.org/r/CAGETcx_bkuFaLCiPrAWCPQz+w79ccDp6=9e881qmK=vx3hBMyg@mail.gmail.com
Fixes: 4a032827da ("of: property: Simplify of_link_to_phandle()")
Link: https://lore.kernel.org/r/CAGETcx_+rhHvaC_HJXGrr5_WAd2+k5f=rWYnkCZ6z5bGX-wj4w@mail.gmail.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Mark Brown <broonie@kernel.org>
Acked-by: Wolfram Sang <wsa@kernel.org> # for I2C
Acked-by: Shawn Guo <shawnguo@kernel.org>
Acked-by: Saravana Kannan <saravanak@google.com>
Tested-by: Ivan Bornyakov <i.bornyakov@metrotek.ru>
Link: https://lore.kernel.org/r/e1fa546682ea4c8474ff997ab6244c5e11b6f8bc.1680182615.git.geert+renesas@glider.be
Signed-off-by: Rob Herring <robh@kernel.org>
Add spi_split_transfers_maxwords() function that splits
spi_transfers transparently into multiple transfers
that are below a given number of SPI words.
This function reuses most of its code from
spi_split_transfers_maxsize() and for transfers with
eight or less bits per word actually behaves the same.
Signed-off-by: Leonard Göhrs <l.goehrs@pengutronix.de>
Link: https://lore.kernel.org/r/20230310092053.1006459-1-l.goehrs@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
spi_pcpu_stats_totalize() is a rather large macro, and is instantiated
28 times, causing a large amount of duplication in the amount of
generated code.
Reduce the duplication by replacing spi_pcpu_stats_totalize() by a real
C function, and absorb all other common code from
spi_statistics_##name##_show(). As (a) the old "field" parameter was
the name of a structure member, which cannot be passed to a function,
and (b) passing a pointer to the member is also not an option, due to
the loop over all possible CPUs, the "field" parameter is replaced by an
"offset" parameter, pointing to a location within the structure.
This reduces kernel size by ca. 4 KiB (on arm32 and arm64).
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/cb7690d9d04c06eec23dbb98fbb5444082125cff.1677594432.git.geert+renesas@glider.be
Signed-off-by: Mark Brown <broonie@kernel.org>
Pull driver core updates from Greg KH:
"Here is the large set of driver core changes for 6.3-rc1.
There's a lot of changes this development cycle, most of the work
falls into two different categories:
- fw_devlink fixes and updates. This has gone through numerous review
cycles and lots of review and testing by lots of different devices.
Hopefully all should be good now, and Saravana will be keeping a
watch for any potential regression on odd embedded systems.
- driver core changes to work to make struct bus_type able to be
moved into read-only memory (i.e. const) The recent work with Rust
has pointed out a number of areas in the driver core where we are
passing around and working with structures that really do not have
to be dynamic at all, and they should be able to be read-only
making things safer overall. This is the contuation of that work
(started last release with kobject changes) in moving struct
bus_type to be constant. We didn't quite make it for this release,
but the remaining patches will be finished up for the release after
this one, but the groundwork has been laid for this effort.
Other than that we have in here:
- debugfs memory leak fixes in some subsystems
- error path cleanups and fixes for some never-able-to-be-hit
codepaths.
- cacheinfo rework and fixes
- Other tiny fixes, full details are in the shortlog
All of these have been in linux-next for a while with no reported
problems"
[ Geert Uytterhoeven points out that that last sentence isn't true, and
that there's a pending report that has a fix that is queued up - Linus ]
* tag 'driver-core-6.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (124 commits)
debugfs: drop inline constant formatting for ERR_PTR(-ERROR)
OPP: fix error checking in opp_migrate_dentry()
debugfs: update comment of debugfs_rename()
i3c: fix device.h kernel-doc warnings
dma-mapping: no need to pass a bus_type into get_arch_dma_ops()
driver core: class: move EXPORT_SYMBOL_GPL() lines to the correct place
Revert "driver core: add error handling for devtmpfs_create_node()"
Revert "devtmpfs: add debug info to handle()"
Revert "devtmpfs: remove return value of devtmpfs_delete_node()"
driver core: cpu: don't hand-override the uevent bus_type callback.
devtmpfs: remove return value of devtmpfs_delete_node()
devtmpfs: add debug info to handle()
driver core: add error handling for devtmpfs_create_node()
driver core: bus: update my copyright notice
driver core: bus: add bus_get_dev_root() function
driver core: bus: constify bus_unregister()
driver core: bus: constify some internal functions
driver core: bus: constify bus_get_kset()
driver core: bus: constify bus_register/unregister_notifier()
driver core: remove private pointer from struct bus_type
...
Group some variables based on their sizes to reduce hole and avoid padding.
On x86_64, this shrinks the size from 144 to 128 bytes.
Turn 'timestamped' into a bitfield so that it can be easily merged with
some other bifields and move 'error'.
This should have no real impact on memory allocation because 'struct
spi_transfer' is mostly used on stack, but it can save a few cycles
when the structure is initialized or copied.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/93a051da85a895bc6003aedfb00a13e1c2fc6338.1676370870.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Mark Brown <broonie@kernel.org>
