The function FLEXIO_SPI_MasterTransferNonBlocking has a bug when
works in CS continuous mode. In this mode, both RX and TX
interrupts are enabled, they share the same IRQ line.
In the ISR, the RX event and TX event are not handled well,
so a short delay in ISR is needed.
The function FLEXIO_SPI_MasterTransferNonBlocking issue is fixed,
the delay in ISR can be removed.
Signed-off-by: Jason Yu <zejiang.yu@nxp.com>
Currently there is a mismatch between the naming of the hardware and
the drivers targetting the hardware. nrf2_ is used instead of
the actual bindings names, like nrf2_audiopll instead of
nrfs_audiopll. This makes it hard to map drivers to the hardware
they are targetting.
There is historical reason for some of this, namely the same binding
name was used for different hardware, which is why nrf2_ was used
on newer platforms. This is no longer the case though, so drivers
and configs can be named according to the hardware without conflict.
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
Reset the SPI peripheral to its default state
and register values on init by setting its SWRST bit.
This is important since the driver assumes that certain
registers are at their default values.
Signed-off-by: Pete Dietl <petedietl@gmail.com>
The sam0 SPI driver does not ensure that it clears the 32-bit extension
option during init. The 32-bit extension option, which comprises of a field
in the CTRLC register and the LENGTH register enables better bus
utilization by allowing 32-bit writes to the SPI DATA register
(as opposed to the usual 8-bit writes). The driver breaks down if this
option is enabled by causing each intended byte of output to become
four bytes. We fix this by explicitly disabling the 32-bit extension
option in init.
Signed-off-by: Pete Dietl <petedietl@gmail.com>
To determine whether device runtime PM is enabled on a device, use
`pm_device_runtime_is_enabled`. This results in the same behaviour when
`CONFIG_PM_DEVICE_RUNTIME=n`, but properly controls the clocks on a
per-instance basis when `CONFIG_PM_DEVICE_RUNTIME=y`.
Signed-off-by: Jordan Yates <jordan@embeint.com>
To convey the correct data size, use the 'data->dfs' value instead
of '1' when moving the SPI context to the next buffer.
Signed-off-by: Andreas Wolf <awolf002@gmail.com>
Initial commit for SPI driver support on RSK_RX130@512KB board
with RSPI module
Signed-off-by: Duy Nguyen <duy.nguyen.xa@renesas.com>
Signed-off-by: Minh Tang <minh.tang.ue@bp.renesas.com>
Replace UTIL_AND(DT_INST_NODE_HAS_PROP(... construction with the
DT_INST_PROP_OR helper macro.
Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
This is a follow-up to commit a068709171.
As the SPIM_REQUESTS_CLOCK() macro changed the type of its parameter,
all of the calls to it need to be updated accordingly. Also the call to
DT_NODELABEL() needs to be removed from that macro as DT_CLOCKS_CTLR()
already returns a node identifier.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Several drivers checked for the presense and availability of data cache
through Kconfig symbol. This is supported according to the current
documentation, but the symbol DCACHE masks two types of cache devices: arch
and external caches. The latter is present on some Cortex-M33 chips, like
the STM32U5xx. The external dcache is bypassed when accessing internal
SRAM and only used for external memories.
In commit a2dd232410 ("drivers: adc: stm32: dma support") the rationale
for gating dcache for adc_stm32 behind STM32H7X is only hinted at, but
reason seems to be that it was the only SOC the change was tested on. The
SOC configures DCACHE=y so it is most likely safe to swap the SOC gate for
DCACHE.
The DCACHE ifdefs are now hidden inside the shared stm32_buf_in_nocache()
implementation.
Signed-off-by: Henrik Lindblom <henrik.lindblom@vaisala.com>
These two functions were using the value of ctx->xx_len wrong, the unit
is in words, not bytes, but spi_context_count_xx_buf_lens was iterating
over the length of bytes, not words.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
The unit of spi_context's tx_len and rx_len is byte instead of frame.
Thus, in spi_context_update_tx(), the value of ctx->tx_len should be
reduced by (len * dfs).
Also to fix the update of ctx->tx_len in spi_context_update_rx().
Signed-off-by: cyliang tw <cyliang@nuvoton.com>
Remove the auto setting of the external spi clock if its not there,
also fix the calculation of calucation the spi divisor value.
Signed-off-by: Ryan McClelland <ryanmcclelland@meta.com>
The PM Device callbacks is rather unimplemented. There currently is
no device agnostic clock management api (yet), and the pinctrl isn't
fully implemented in this driver. Remove it all.
Signed-off-by: Ryan McClelland <ryanmcclelland@meta.com>
This adds the missing fifo config from the dts which was missed in
the initial revision. This also adds the spi rtio fallback api.
Signed-off-by: Ryan McClelland <ryanmcclelland@meta.com>
Replace semaphores with proper atomic flags when used without
multithreading enabled.
Signed-off-by: Michal Kozikowski <michal.kozikowski@nordicsemi.no>
The nrxf_spim driver currently resumes itself for the duration of
a transfer, however, in case SPI_LOCK_ON is used, the driver needs
to keep itself resumed until spi_release() is called. Currently,
this results in unbalanced suspend as the bus puts itself both
after transaction is done, and when spi_release() is called.
This patch makes the driver check if SPI_LOCK_ON is set once
transaction is complete, if yes, selv get one more time to
account for the two puts which will follow.
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
This dependency was always there but not explicitly defined.
By adding the explicit dependency it becomes more obvious
what is wrong when SPI is enabled but GPIO disabled.
This was found when building `samples/bluetooth/peripheral`
for `nrf54l15dk/nrf54l15/cpuapp` with `CONFIG_GPIO=n`.
Before we got:
- A linker error in `spi_nrfx_common.c` failing
to reference some nrfx_gpiote APIs.
- A linker error in `spi_nrfx_spim.c` failing to reference
the GPIO dts entry.
Now we will get a warning of that GPIO is not enabled
With this it becomes more obvious that SPI driver is enabled by
default because of the external flash mounted on the DK.
Signed-off-by: Rubin Gerritsen <rubin.gerritsen@nordicsemi.no>
SPI driver for Microchip MEC5 HAL based QSPI controller. QSPI
hardware supports full duplex, dual, and quad operation. MEC5
QSPI controller also includes three local DMA channels per
direction to off load firmware. The driver API supports full
or half-duplex. Due to QSPI hardware not supporting one wire
half-duplex, this driver supports full-duplex only. QSPI hardware
design requires it to control chip select and current hardware
supports up to two chip selects. Zephyr's SPI DT macros store the
child SPI device's reg properity as the "slave" member of the SPI
configuration structure. The driver uses the "slave" value as the
chip select. Additional timing settings specific to SPI flash devices
are in a new SPI device YAM file: "microchip,mec5-qspi-device.yaml"
which includes the standard "spi-device.yaml". If the new YAML is not
used, the QSPI controller will use default timing values for chip
select and I/O line taps.
Signed-off-by: Scott Worley <scott.worley@microchip.com>
There was a race condition where `lpspi_end_xfer` can be called multiple
times per transfer. There was the case where a TX interrupt gets
triggered without the RX interrupt being set, and TX finishes writing
its last byte. Then, `spi_context_rx_len_left() == 0` is true and
`lpspi_end_xfer` happens, but the RX interrupt is still active. Then,
when the RX interrupt happens, `lpspi_end_xfer` will get called again.
To fix that, the architecture was adjusted to only call `lpspi_end_xfer`
once no interrupts are active any more, and the disabling of the
interrupts gets used to signal the end of the TX and RX part.
Minor adjustments were necessary to use the interrupt enable signals for
this purpose; the TX irq handler had its internal order reversed,
otherwise it wasn't guaranteed that the physical transfer is finished
when we disable the interrupt.
Also, the code where the RX interrupt gets disabled had to be moved out
of the RX irq handler, because the RX interrupt also needs to be
disabled if RX is finished but no RX interrupt is currently active.
Signed-off-by: Martin Stumpf <finomnis@gmail.com>
In case the SPI transaction has SPI_HOLD_ON_CS set, we need to keep
SPI resumed until spi_release() is called. This is required as we
now need to keep the CS GPIO port resumed until transaction is
complete.
Suspending CS GPIO is not allowed from ISR in some cases (H20 fast
GPIO instance) so we have to defer CS GPIO suspend to some thread
context (put_async or spi_release()).
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
Implement calling spi_context_cs_get_all() and
spi_context_cs_put_all() in line with pm resume/suspend.
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
Introduce spi_context_cs_get() and spi_context_cs_put() which shall
be used from drivers to get/put the GPIO port the CS GPIO belongs to
before and after a transaction, in line with the SPI drivers pm
action hook being called.
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
Changed to define ambiq spi/i2c dma mode as a binding property
instead of kconfig macros, making it more flexible for different
spi/i2c instances.
Signed-off-by: Hao Luo <hluo@ambiq.com>
1. rework IOM cmdq buffer instantiation
2. rework spi and i2c cache handling as it is incorrect.
3. buffers need to be aligned with DCACHE on
Signed-off-by: Swift Tian <swift.tian@ambiq.com>
The TXFTLR register has 2 major fields which are TFT for triggering
interrupt threshold and TXFTLR for starting transfer threshold. This is
to ensure that sufficient data is ready for starting transfer.
Signed-off-by: Younghyun Park <younghyunpark@google.com>
Support SPI_HOLD_ON_CS flag in the CPU-based driver. To do this we will
set CONTC bit to continue previous command. Technically it may not be
necessary right now, and could just not clear CONT bit...
but in the future in the lpspi driver we
will decouple the config/init of a transfer from the SDK
and therefore have more control over TCR,
and when we write the TCR, we need to take CONTC bit into account
otherwise a new command will be made. So this approach is how
it should be handled in the driver going forward in my opinion, even
if it might be possible without this bit right now, I want to introduce
it's usage now.
This commit also does a minor refactor in the ISR and adds some comments
to make the strange CS behavior and strange handling code more clear to
future readers.
Also, make the early predicted SPI xfer end code only happen for spi
versions where it is necessary, since I think that code is really the
best we can do but might have a race condition, where possible the last
word is not finished sending when we end the xfer. So limit the
potential affect to v1 lpspi where the workaround is actually required
due to stalling behavior.
Lastly, set the LPSPI into master mode at active low in init, due to
it being the most common case, we want the SPI CS lines to be
initialized at init of driver. I don't think it's worth it to make it
configurable at this time, but in the future it could be if needed.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
The LPSPI does support word sizes such as 6 or 7, anything as small as 2
bits. So fix the checks and the math to allow for this in the driver.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
Optimize the TX fill algorithm to have less interrupts by filling the TX
fifo as much as possible during each interrupt handle.
Before, the algorithm was just a very simple, fill the TX fifo with as
much from only the current buffer as possible, then send it and wait for
the next interrupt. Now the algorithm is to fill the TX fifo as much as
possible, even if it means reading from multiple buffers during the
interrupt.
This has the advantage from master mode of having less interrupts. And
it is very important for slave mode because the slave mode does not
control the pacing of the transfer and so therefore should fill as much
as possible whenever possible in order not to miss a deadline.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>