There is a subtle difference between dma_transfer_direction and
dma_data_direction: the former is used by the dmaengine framework,
while the latter is used by the dma-mapping API. Although the
purpose is comparable, the actual values are different and must
not be mixed. In this case, the driver just wants to use
dma_transfer_direction.

Without this patch, building s3c6400_defconfig results in:

drivers/spi/spi-s3c64xx.c: In function 's3c64xx_spi_dmacb':
drivers/spi/spi-s3c64xx.c:239:21: warning: comparison between
	'enum dma_data_direction' and 'enum dma_transfer_direction' [-Wenum-compare]

As pointed out by Kukjin Kim, this also changes the use of constants
from DMA_FROM_DEVICE/DMA_TO_DEVICE to DMA_DEV_TO_MEM/DMA_MEM_TO_DEV.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NKukjin Kim <kgene.kim@samsung.com>
Cc: Ben Dooks <ben-linux@fluff.org>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: spi-devel-general@lists.sourceforge.net

When a device tree definition os an s3c64xx SPI master is missing
a "controller-data" subnode, the newly added s3c64xx_get_slave_ctrldata
function might use uninitialized memory in place of that node,
which was correctly reported by gcc.

Without this patch, building s3c6400_defconfig results in:

drivers/spi/spi-s3c64xx.c: In function 's3c64xx_get_slave_ctrldata.isra.25':
drivers/spi/spi-s3c64xx.c:841:5: warning: 'data_np' may be used uninitialized in this function [-Wuninitialized]
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NThomas Abraham <thomas.abraham@linaro.org>
Acked-by: NKukjin Kim <kgene.kim@samsung.com>
Cc: Jaswinder Singh <jaswinder.singh@linaro.org>
Cc: Grant Likely <grant.likely@secretlab.ca>

Silences the following sparse warnings:
drivers/spi/spi-s3c64xx.c:1482:32: warning:
symbol 's3c2443_spi_port_config' was not declared. Should it be static?
drivers/spi/spi-s3c64xx.c:1489:32: warning:
symbol 's3c6410_spi_port_config' was not declared. Should it be static?
drivers/spi/spi-s3c64xx.c:1495:32: warning:
symbol 's5p64x0_spi_port_config' was not declared. Should it be static?
drivers/spi/spi-s3c64xx.c:1501:32: warning:
symbol 's5pc100_spi_port_config' was not declared. Should it be static?
drivers/spi/spi-s3c64xx.c:1508:32: warning:
symbol 's5pv210_spi_port_config' was not declared. Should it be static?
drivers/spi/spi-s3c64xx.c:1515:32: warning:
symbol 'exynos4_spi_port_config' was not declared. Should it be static?
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

They have very few users and they're both just doing a single register
write so the advantage of having the macro is a bit limited. An inline
function might make sense but it's as easy to just do the writes directly.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Saves some error handling and a small amount of code.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Reviewed-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

No call was being made by the GPIO driver to put the GPIO into output
mode meaning that the calls to gpio_set_value() which were being done
were not valid. A similar issue appears to exist with the DT GPIO
requests but as they appear to be being used for pinmux it's less clear
to me that we want to configure them.

Without this fix Cragganmore systems can't talk to their SPI devices.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

When gpio_request() fails the driver logged the failure but while it'd
try to print an error code in the non-DT case it didn't pass the error
code in so garbage would be logged and in the DT case the error wasn't
logged.

Further, in the non-DT case the error code was then overwritten with -EBUSY
depriving the caller of information and breaking automatic probe deferral
pushing back from the GPIO level.  Also reformat the non-DT log message
so it's not word wrapped and we can grep for it.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Add support for device based discovery.
Signed-off-by: NThomas Abraham <thomas.abraham@linaro.org>
Acked-by: NJaswinder Singh <jaswinder.singh@linaro.org>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

The set_level callback in the controller data, which is used to configure
the slave select line, cannot be supported when migrating the driver to
device tree based discovery. Since all the platforms currently use gpio
as the slave select line, this callback can be removed from the
controller data and replaced with call to gpio_set_value in the driver.
Signed-off-by: NThomas Abraham <thomas.abraham@linaro.org>
Acked-by: NJaswinder Singh <jaswinder.singh@linaro.org>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

The platform data pointer that is passed to the spi gpio setup functions
is not used. Hence, this parameter is removed from all the spi gpio setup
functions.
Signed-off-by: NThomas Abraham <thomas.abraham@linaro.org>
Acked-by: NJaswinder Singh <jaswinder.singh@linaro.org>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Platform data is used to specify controller hardware specific information
such as the tx/rx fifo level mask and bit offset of rx fifo level. Such
information is not suitable to be supplied from device tree. Instead,
it can be moved into the driver data and removed from platform data.
Signed-off-by: NThomas Abraham <thomas.abraham@linaro.org>
Acked-by: NJaswinder Singh <jaswinder.singh@linaro.org>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

The macro S3C64XX_SPI_ST_TRLCNTZ is not used and hence it is removed.
Signed-off-by: NThomas Abraham <thomas.abraham@linaro.org>
Acked-by: NJaswinder Singh <jaswinder.singh@linaro.org>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Config operation is separated from request operation in
DMA common operation. Because spi driver can change the
DMA config for every transfer. So this patch is using the
separated DMA config operation.
Signed-off-by: NBoojin Kim <boojin.kim@samsung.com>
Signed-off-by: NKyoungil Kim <ki0351.kim@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Commit 054ebc (spi: Compatibility with direction which is used in samsung
DMA operation) does not build as one hunk adds a brace to the first branch
of an if statement without adding at least the correspoding close. Remove
the unwanted brace.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

msg->complete will be called in spi_finalize_current_message().
Signed-off-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Reviewed-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

I found that there are two kind of direction type.
First one is dma_data_direction defined in include/linux/dma-direction.h.
It is used for parameter of dma_map/unmap_single in spi-s3c64xx.
The other one is dma_transter_direction defined in include/linux/dmaengine.h.
It is used for direction of samsung DMA operation
(arch/arm/plat-samsung/dma-ops.c).
This patch is just some changes to use direction defines
which is used in samsung DMA operation.
Signed-off-by: NBoojin Kim <boojin.kim@samsung.com>
Signed-off-by: NKyoungil Kim <ki0351.kim@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Convert the s3c64xx driver to using the new message queue factored out of
the pl022 driver by Linus Walleij, saving us a nice block of code and
getting the benefits of improvements implemented in the core.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Enable and disable the clocks to the SPI controller using runtime PM. This
serves the dual purpose of reducing power consumption a little and letting
the core know when the device is idle.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NHeiko Stuebner <heiko@sntech.de>

In preparation for the addition of runtime PM ops.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Although the hardware supports interrupts we're not currently using them
at all since for small transfers the overhead is greater than that for
busy waiting and for large transfers we have interrupts from the DMA.
This means that if the hardware reports an error (especially one which
might not stall transfer) we might miss it.

Take a first pass at dealing with such errors by enabling the interrupt
if we can and logging the errors if they happen. Ideally we'd report the
error via the affected transfer but since we're in master mode it's very
difficult to trigger errors at present and this code is much simpler.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>

This patch modifies the driver to stop depending on the
clock names being passed from platform and switch over
to lookup clocks generic names using clkdev
Signed-off-by: NPadmavathi Venna <padma.v@samsung.com>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

This patch modifies to merge the dma control code. Original
s3c64xx spi driver has each dma control code for rx and tx
channel. This patch merges these dma control codes into one.
With this patch, a dma setup function and callback function
handle for both rx and tx channel.
Signed-off-by: NBoojin Kim <boojin.kim@samsung.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NVinod Koul <vinod.koul@intel.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

This patch adds to support DMA generic API to transfer raw
SPI data. Basiclly the spi driver uses DMA generic API if
architecture supports it. Otherwise, uses Samsung specific
S3C-PL330 APIs.
Signed-off-by: NBoojin Kim <boojin.kim@samsung.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NVinod Koul <vinod.koul@intel.com>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Sort the SPI makefile and enforce the naming convention spi_*.c for
spi drivers.

This change also rolls the contents of atmel_spi.h into the .c file
since there is only one user of that particular include file.

v2: - Use 'spi-' prefix instead of 'spi_' to match what seems to be
      be the predominant pattern for subsystem prefixes.
    - Clean up filenames in Kconfig and header comment blocks
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-by: NWolfram Sang <w.sang@pengutronix.de>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>

We can't do without setting channel and bus width to
same size. In order to do that, use loop read/writes in
polling mode and appropriate burst size in DMA mode.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Newer SoCs have the SPI clock scaling control in platform's
clock management unit. Inorder for such SoCs to work, we need
to check the flag clk_from_cmu before making any clock changes.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Since we use DMA mode only for xfers bigger than FIFO size,
do not map/unmap buffers for polling mode transfers.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

When using PIO we have a timeout for the TX and RX FIFOs to ensure that
the data actually gets transferred. Warn if we hit that timeout - it
should never happen, but this makes sure we'll find out if it does.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Instead of, wrongly, reusing the 'val' variable, use a dedicated
one for reading the status register.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Fix compilation warning by typecasting the tx_buf pointer.

[I'm not thrilled with resorting to a cast; but I cannot see a better
way to go about this.  I don't want to drop the const from struct
spi_transfer ~~glikely]
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

For small transfers at high speeds the expected transfer time can easily
be well under 1ms, causing the delay in wait_for_xfer() to be only the
dead reckoning fudge factor of 5ms currently included. Experiments on
some of my systems shows that this is marginal for some transfers so
double it to 10ms.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Allow the use of the S3C64xx SPI controller with things like PMICs by
moving the init up to subsys_initcall().
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NJassi Brar <jassisinghbrar@gmail.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The S3C64xx SPI driver requires the machine to call s3c64xx_spi_set_info()
to select a few options, including the clock to use for the SPI controller.
If this is not done then a NULL will be passed as the clock name for
clk_get(), causing an obscure crash. Guard against this and other missing
configuration by validating that the clock name has been filled in in
the platform data that ets passed in.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

String constants that are continued on subsequent lines with \
are not good.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Since most of the chip-selects are simply going to be like
gpio_set_value, it would do good to have the same callback type
so that it could simply be made to point at gpio_set_value.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Header for platform specific stuff has been rename to include the SoC
type. Include the new header instead.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Add precautionary check before releasing memory region.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The pointer to SPI rate source clock had better be the member of
driver local data structure rather than platform specific.
Also, remove definitions of variable 'sci' that are rendered
useless as a consequence.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The instance of SPI clock for controller and that used for generating
signals ought to be independently handled.
Signed-off-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>