After having removed all code dependencies we can make fsl-espi
completely independent of fsl-lib now.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

This patch adds support for ESPI RXSKIP mode. This mode is optimized
for flash reads:
- sends a number of bytes and then reads a number of bytes
- shifts out zeros automatically when reading

Supporting RXSKIP mode is a prerequisite for supporting dual output
read mode.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Now that we introduced element tx_len in struct mpc8xxx_spi let's
rename element len to rx_len as it actually is the number of bytes to
receive. In addition make it unsigned.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

This change addresses two issues:
- If the TX FIFO is full the ISR polls until there's free space again.
  An ISR should never wait for something.
- Currently the number of bytes to transfer is rounded up to the next
  multiple of 4. For most transfers therefore few bytes remain in the
  TX FIFO after end of transfer.
  This would cause the next transfer to fail and as a workaround the
  ESPI block is disabled / re-enabled in fsl_espi_change_mode.
  This seems to clear the FIFO's (although it's not mentioned in the
  spec).

With this change the TX FIFO is filled as much as possible initially
and whenever the ISR is called. Also the exact number of bytes is
transferred.
The spinlock protects against a potential race if the first interrupt
occurs whilst the TX FIFO is still being initially filled.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Change register access to the method used in other drivers too.
- use register names as in the chip spec for constants
- avoid hard to read statements like
  __be32 __iomem *espi_mode = &reg_base->mode
- get rid of old powerpc-specific functions like in_8

In addition annotate reg_base in struct mpc8xxx_spi as __iomem.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently the driver allocates a 64kb buffer for each single message.
On systems with little and fragmented memory this can result in
memory allocation errors. Solve this by pre-allocating a buffer.

This patch was developed in OpenWRT long ago, however it never
made it upstream.

I slightly modified the original patch to re-initialize the buffer
at the beginning of each transfer.
Signed-off-by: NGabor Juhos <juhosg@openwrt.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Now that most cleanup is done automatically the remove functions
can be significantly simplified.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Signed-off-by: NEsben Haabendal <eha@deif.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Migrates the fsl-(e)spi driver to use the generic master queuing.
Avoids the "master is unqueued, this is deprecated" warning.
Signed-off-by: NHeiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

In current code, master->cleanup and master->setup are not set in the same
function. This makes it hard to read and not good for code maintain.
One example is in fsl-spi.c, master->cleanup is overrided in mpc8xxx_spi_probe()
which leads to a memory leak.

This patch removes mpc8xxx_spi_cleanup() and converts fsl_espi_setup to use
devm_kzalloc so we don't need to take care of freeing memory.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

This relies upon of_spi_register_master to find out which gpios to use.
Acked-by: NAnton Vorontsov <anton@enomsg.org>
Signed-off-by: NAndreas Larsson <andreas@gaisler.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Acked-by: NAnton Vorontsov <anton@enomsg.org>
Signed-off-by: NAndreas Larsson <andreas@gaisler.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

For being able to distinguishing between the regular type of cores and others
with different entries in of_fsl_spi_match.
Acked-by: NAnton Vorontsov <anton@enomsg.org>
Signed-off-by: NAndreas Larsson <andreas@gaisler.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

spi/spi-fsl-spi: Move setting non-zero tx and rx shifts to a function accessed by a function pointer
Acked-by: NAnton Vorontsov <anton@enomsg.org>
Signed-off-by: NAndreas Larsson <andreas@gaisler.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This makes the spi-fsl-spi driver usable in CPU mode outside of an FSL_SOC and
even an powerpc environment by moving CPM mode functionality to a separate file
that is only compiled and linked in an FSL_SOC environment and adding some
ifdefs to hide types and functions or provide alternatives.

For devicetree probing a "clock-frequency" property is used for clock frequency
instead of calls to FSL_SOC-specific functions.
Acked-by: NAnton Vorontsov <anton@enomsg.org>
Signed-off-by: NAndreas Larsson <andreas@gaisler.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

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>

Get rid of users of of_platform_driver in drivers/spi.  The
of_platform_{,un}register_driver functions are going away, so the
users need to be converted to using the platform_bus_type directly.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

We get the following when building on ppc64 due to lack of include of
<asm/io.h>:

In file included from drivers/spi/spi_fsl_espi.c:25:0:
drivers/spi/spi_fsl_lib.h: In function 'mpc8xxx_spi_write_reg':
drivers/spi/spi_fsl_lib.h:88:2: error: implicit declaration of function 'out_be32'
drivers/spi/spi_fsl_lib.h: In function 'mpc8xxx_spi_read_reg':
drivers/spi/spi_fsl_lib.h:93:2: error: implicit declaration of function 'in_be32'
drivers/spi/spi_fsl_espi.c: In function 'fsl_espi_remove':
drivers/spi/spi_fsl_espi.c:571:2: error: implicit declaration of function 'iounmap'
drivers/spi/spi_fsl_espi.c: In function 'fsl_espi_probe':
drivers/spi/spi_fsl_espi.c:602:2: error: implicit declaration of function 'ioremap'
drivers/spi/spi_fsl_espi.c:602:24: warning: assignment makes pointer from integer without a cast
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Add eSPI controller support based on the library code spi_fsl_lib.c.

The eSPI controller is newer controller 85xx/Pxxx devices supported.
There're some differences comparing to the SPI controller:

1. Has different register map and different bit definition
   So leave the code operated the register to the driver code, not
   the common code.

2. Support 4 dedicated chip selects
   The software can't controll the chip selects directly, The SPCOM[CS]
   field is used to select which chip selects is used, and the
   SPCOM[TRANLEN] field is set to tell the controller how long the CS
   signal need to be asserted. So the driver doesn't need the chipselect
   related function when transfering data, just set corresponding register
   fields to controll the chipseclect.

3. Different Transmit/Receive FIFO access register behavior
   For SPI controller, the Tx/Rx FIFO access register can hold only
   one character regardless of the character length, but for eSPI
   controller, the register can hold 4 or 2 characters according to
   the character lengths. Access the Tx/Rx FIFO access register of the
   eSPI controller will shift out/in 4/2 characters one time. For SPI
   subsystem, the command and data are put into different transfers, so
   we need to combine all the transfers to one transfer in order to pass
   the transfer to eSPI controller.

4. The max transaction length limitation
   The max transaction length one time is limitted by the SPCOM[TRANSLEN]
   field which is 0xFFFF. When used mkfs.ext2 command to create ext2
   filesystem on the flash, the read length will exceed the max value of
   the SPCOM[TRANSLEN] field.
Signed-off-by: NMingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Refactor the common code in file spi_fsl_spi.c to spi_fsl_lib.c used
by SPI/eSPI controller driver as a library, and leave the QE/CPM SPI
controller code in the SPI controller driver spi_fsl_spi.c.

Because the register map of the SPI controller and eSPI controller
is so different, also leave the code operated the register to the
driver code, not the common code.
Signed-off-by: NMingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>