Since commit a2fd4f9fa3 'spi: Support transfer speed checking in the core',
the SPI core validates the desired speed of a given transfer against the
minimum and maximum speeds supported by the controller.
So we can remove the same code in this driver and let spi core handle checking
the desired speed of a given transfer.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Remove unused devdata pointer 'orion_spi' in function orion_spi_write_read().
Detected by Coverity: CID 1077860.
Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Signed-off-by: Mark Brown <broonie@linaro.org>
It's pointless to use of_match_ptr within CONFIG_OF guard.
Use of_match_ptr around davinci_spi_of_match when setting .of_match_table.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
spi_master_get_devdata() never returns NULL when spi_alloc_master() success,
so remove NULL test for dspi.
We have ensured master is not NULL before assigning it to dspi->bitbang.master.
So also remove NULL test for dspi->bitbang.master.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
The Allwinner A31 has a new SPI controller IP compared to the older Allwinner
SoCs.
It supports DMA, but the driver only does PIO for now, and DMA will be
supported eventually.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
We cannot unconditionally use dma_map_single() to map data for use with
SPI since transfers may exceed a page and virtual addresses may not be
provided with physically contiguous pages. Further, addresses allocated
using vmalloc() need to be mapped differently to other addresses.
Currently only the MXS driver handles all this, a few drivers do handle
the possibility that buffers may not be physically contiguous which is
the main potential problem but many don't even do that. Factoring this
out into the core will make it easier for drivers to do a good job so if
the driver is using the core DMA code then generate a scatterlist
instead of mapping to a single address so do that.
This code is mainly based on a combination of the existing code in the MXS
and PXA2xx drivers. In future we should be able to extend it to allow the
core to concatenate adjacent transfers if they are compatible, improving
performance.
Currently for simplicity clients are not allowed to use the scatterlist
when they do DMA mapping, in the future the existing single address
mappings will be replaced with use of the scatterlist most likely as
part of pre-verifying transfers.
This change makes it mandatory to use scatterlists when using the core DMA
mapping so update the s3c64xx driver to do this when used with dmaengine.
Doing so makes the code more ugly but it is expected that the old s3c-dma
code can be removed very soon.
Signed-off-by: Mark Brown <broonie@linaro.org>
It is fairly common for SPI devices to require that one or both transfer
directions is always active. Currently drivers open code this in various
ways with varying degrees of efficiency. Start factoring this out by
providing flags SPI_MASTER_MUST_TX and SPI_MASTER_MUST_RX. These will cause
the core to provide buffers for the requested direction if none are
specified in the underlying transfer.
Currently this is fairly inefficient since we actually allocate a data
buffer which may get large, support for mapping transfers using a
scatterlist will allow us to avoid this for DMA based transfers.
Signed-off-by: Mark Brown <broonie@linaro.org>
As of commit 5ce0ba8865 ("spi: rcar: add
Renesas QSPI support on RSPI") the rspi driver handles Renesas QSPI, too,
but this was not reflected in the Kconfig help text.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
The driver core is now taking care of putting our pins into default
state at probe. Thus we can remove the redundant call for it in probe.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Make use of clk_prepare_enable and clk_disable_unprepare to simplify
code. No functional change.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
There is no meaningful code sharing between the PIO and DMA variants
(just the timeout calculation) so in order to make the code easier to
work with split the two cases.
Looking at the code it is not clear how the PIO version works for large
transmits, greater than FIFO size is only handled for RX.
Signed-off-by: Mark Brown <broonie@linaro.org>
The GPIO enable and disable is done in the core so does not need to be
replicated in the driver, delete the unneeded code. enable_cs() was not
referenced at all.
Signed-off-by: Mark Brown <broonie@linaro.org>
When used via spidev with more than one messages to tranfer via
SPI_IOC_MESSAGE the current implementation would return with
-EINVAL, since bits_per_word and speed_hz are set in all
transfer structs. And in the 2nd loop status will stay at
-EINVAL as its not overwritten again via fsl_spi_setup_transfer().
This patch changes this behavious by first checking if one of
the messages uses different settings. If this is the case
the function will return with -EINVAL. If not, the messages
are transferred correctly.
Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Mark Brown <broonie@linaro.org>
Fix the pointer-integer size mismatch warning below:
drivers/spi/spi-gpio.c: In function ‘spi_gpio_setup’:
drivers/spi/spi-gpio.c:252:8: warning: cast from pointer to integer of
different size [-Wpointer-to-int-cast]
cs = (unsigned int) spi->controller_data;
^
Signed-off-by: SeongJae Park <sj38.park@gmail.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
rspi_wait_for_interrupt() unconditionally enables interrupts, even when the
wait condition is already satisfied. This causes a high interrupt load (2
interrupts/byte for full-duplex Single SPI transfers, 1 interrupt/byte for
RSPI with TX Only mode, or QSPI in unidirectional Dual or Quad Transfer
mode).
Change this to return immediately when the wait condition is satisfied.
This dramatically reduces the interrupt load, especially in high-speed
Quad Transfer mode, and increases transfer speed, as no interrupts need to
be handled when there's space available in the output FIFO, or data
available in the input FIFO.
Benchmark results for QSPI on r8a7791 while reading 1 MiB from 30 MHz SPI
FLASH on the Koelsch development board:
Before:
Single SPI Dual SPI Quad SPI
Interrupts: 2096856 1048592 1048594
Mbps: 0.9 1.6 1.6
After:
Single SPI Dual SPI Quad SPI
Interrupts: 1048569 21295 8
Mbps: 0.7 10.8 12.9
I don't know why Single SPI slowed down a bit.
I've also verified functionality for RSPI-RZ on r7s72100, but don't have
benchmark results as there's no SPI FLASH connected to RSPI on the Genmai
development board. Unlike RSPI and QSPI, RSPI-RZ has separate interrupts
for RX and TX, which shows that Single SPI transfers now generate (mostly)
RX interrupts, as expected.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Currently a divide-by-2^n clock is used, causing a very coarse
clock selection, i.e. a 10MHz device will need to use a 6MHz
clock. The McSPI can also use a divide-by-n clock, this patch
adds support for selecting that when possible.
Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
spi_transfer->speed_hz and spi_transfer->bits_per_word used to only be set when
not using the default settings but are not set on every transfer, causing
omap2_mcspi_setup_transfer to be called on each transfer. This patch changes
the check to only call omap2_mcspi_setup_transfer if the settings needs to be
changed.
Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
This patch adds support for 1 to 8 BPW to driver and removes
excess BPW validation since this is already checked by SPI core.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Mark Brown <broonie@linaro.org>
This patch fixes an issue that the following build error happens when
the CONFIG_OF is not set:
drivers/spi/spi-rspi.c: In function 'rspi_probe':
drivers/spi/spi-rspi.c:1203:26: error: 'rspi_of_match' undeclared (first use in this function)
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
The process of DMA mapping buffers for SPI transfers does not vary between
devices so in order to save duplication of code in drivers this can be
factored out into the core, allowing it to be integrated with the work that
is being done on factoring out the common elements from the data path
including more sharing of dmaengine code.
In order to use this masters need to provide a can_dma() operation and while
the hardware is prepared they should ensure that DMA channels are provided
in tx_dma and rx_dma. The core will then ensure that the buffers are mapped
for DMA prior to calling transfer_one_message().
Currently the cleanup on error is not complete, this needs to be improved.
Signed-off-by: Mark Brown <broonie@linaro.org>
Don't wait indefinitely for transfers to complete but time out after 10ms
more than we expect the transfer to take on the wire.
Signed-off-by: Mark Brown <broonie@linaro.org>
For better readability, and then we can remove the comment for last transfer.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Trivial cleanup to avoid duplicate code in various bits_per_word cases.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
The checking for spi->mode is done in the implementation of spi_setup().
Calling sc18is602_check_transfer(spi, NULL, 0) is pointless because
the code is equivent to checking if spi->max_speed_hz is 0.
Note, sc18is602_check_transfer actually allows spi->max_speed_hz is 0
if t->speed_hz is set. So return error in sc18is602_setup() when
spi->max_speed_hz is 0 does not make sense.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Mark Brown <broonie@linaro.org>
When the driver was converted to clk_prepare() the suspend path didn't
have any changes made so the clock remains prepared throughout the runtime
of the driver. Unprepare it when suspended so that any savings that can be
made as a result are made.
Signed-off-by: Mark Brown <broonie@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
None of these files are actually using any __init type directives
and hence don't need to include <linux/init.h>. Most are just a
left over from __devinit and __cpuinit removal, or simply due to
code getting copied from one driver to the next.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Since 93abe8e4 (clk: add non CONFIG_HAVE_CLK routines) code using clk.h
like this platform driver needs not depend on HAVE_CLK.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Mark Brown <broonie@linaro.org>
Add support for Quad and Dual SPI Transfers on the Renesas Quad Serial
Peripheral Interface, as found in R-Car Gen2 SoCs like R-Car H2 (r8a7790)
and R-Car M2 (r8a7791):
- Add unidirectional transfer methods for Quad/Dual SPI Transfers.
- Program the sequencer to handle SPI messages with multiple transfer
modes when Quad or Dual transfers are enabled for an SPI slave.
Up to 4 transfer modes per SPI message are supported by the hardware.
- Advertise the availability of Quad and Dual SPI modes on QSPI.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
<<
Switch mpc512x to the common clock framework and adapt mpc512x
drivers to use the new clock driver. Old PPC_CLOCK code is
removed entirely since there are no users any more.
>>
This reverts commit e120cc0dcf.
It causes a NULL pointer dereference with drivers using the generic
spi_transfer_one_message(), which always calls
spi_finalize_current_message(), which zeroes master->cur_msg.
Drivers implementing transfer_one_message() theirselves must always call
spi_finalize_current_message(), even if the transfer failed:
* @transfer_one_message: the subsystem calls the driver to transfer a single
* message while queuing transfers that arrive in the meantime. When the
* driver is finished with this message, it must call
* spi_finalize_current_message() so the subsystem can issue the next
* transfer
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Get the driver ready for the migration to the common clock framework by
calling clk_prepare_enable() and clk_disable_unprepare().
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Add support for the RSPI variant in the RZ/A1H (r7s72100) SoC.
Main differences with RSPI on SH are:
- Lack of TX only mode, hence we always have to use full duplex
transfers,
- The Data Register must be accessed used 8-bit operations.
RSPI on RZ is matched using the new "rspi-rz" platform device name.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Add support for multiple interrupts, based on the SDK reference code.
This is needed for RZ/A1H, which supports 3 interrupts.
When using multiple interrupts, they must be called "rx" (SPRI) and "tx"
(SPTI). The error interrupt (SPEI) is not used, as it matters for slave
mode only.
When using a single interrupt, it may be called "mux". If it cannot be
found, the first interrupt in the device's resources will be used.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
qspi_send_pio() and qspi_receive_pio() are very similar: they both send
and receive full duplex data to/from the hardware, but ignore the data
stream in the unused direction.
Merge them into qspi_transfer_out_in(), now supporting real full duplex.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
rspi_send_pio() and rspi_receive_pio() are very similar:
- the former only sends data, using TX Only Mode,
- the latter sends and receives full duplex data to/from the hardware,
but uses dummy transmit data.
Merge them into rspi_transfer_out_in(), now supporting full duplex if
needed.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Split off qspi_transfer_one() (which doesn't support DMA yet) from
rspi_transfer_one().
Replace the abstraction of send_pio()/receive_pio() by the abstracrion of
transfer_one().
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Add helpers rspi_data_{out,in,out_in}() to write, read, or write and
read data to/from the Data Register, taking care of waiting until data
or space is available in the buffers.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Add rspi_{write,read}_data(), to abstract 8-bit (QSPI, and RSPI on RZ/A1H)
versus 16-bit (RSPI) Data Register access.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Let the generic SPI core handle SPI messages, calling into our
rspi_transfer_one() method.
rspi_assert_ssl() and rspi_negate_ssl() are absorbed into
rspi_prepare_message() and rspi_unprepare_message(), as they actually
enable/disable the whole SPI function, instead of just (de)asserting the
chip select line.
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Pull spi updates from Mark Brown:
"A respun version of the merges for the pull request previously sent
with a few additional fixes. The last two merges were fixed up by
hand since the branches have moved on and currently have the prior
merge in them.
Quite a busy release for the SPI subsystem, mostly in cleanups big and
small scattered through the stack rather than anything else:
- New driver for the Broadcom BC63xx HSSPI controller
- Fix duplicate device registration for ACPI
- Conversion of s3c64xx to DMAEngine (this pulls in platform and DMA
changes upon which the transiton depends)
- Some small optimisations to reduce the amount of time we hold locks
in the datapath, eliminate some redundant checks and the size of a
spi_transfer
- Lots of fixes, cleanups and general enhancements to drivers,
especially the rspi and Atmel drivers"
* tag 'spi-v3.14-2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (112 commits)
spi: core: Fix transfer failure when master->transfer_one returns positive value
spi: Correct set_cs() documentation
spi: Clarify transfer_one() w.r.t. spi_finalize_current_transfer()
spi: Spelling s/finised/finished/
spi: sc18is602: Convert to use bits_per_word_mask
spi: Remove duplicate code to set default bits_per_word setting
spi/pxa2xx: fix compilation warning when !CONFIG_PM_SLEEP
spi: clps711x: Add MODULE_ALIAS to support module auto-loading
spi: rspi: Add missing clk_disable() calls in error and cleanup paths
spi: rspi: Spelling s/transmition/transmission/
spi: rspi: Add support for specifying CPHA/CPOL
spi/pxa2xx: initialize DMA channels to -1 to prevent inadvertent match
spi: rspi: Add more QSPI register documentation
spi: rspi: Add more RSPI register documentation
spi: rspi: Remove dependency on DMAE for SHMOBILE
spi/s3c64xx: Correct indentation
spi: sh: Use spi_sh_clear_bit() instead of open-coded
spi: bitbang: Grammar s/make to make/to make/
spi: sh-hspi: Spelling s/recive/receive/
spi: core: Improve tx/rx_nbits check comments
...
