The hardware seems to have a race condition when the inactive
channels are in slave mode. We support master mode only, so
we can just switch all channels to master mode.
Signed-off-by: NSascha Hauer <s.hauer@pengutronix.de>
[ukleinek: add more verbose comment about the race]
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NVirupax Sadashivpetimath <virupax.sadashivpetimath@stericsson.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This utilizes the new core voltage switch to power off the PL022
core voltage when it's not in use transmitting packets, if and
only if a core voltage switch is available.

Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This simplifies the DMA code a bit by using the dmaengine helpers.
The cookie from descriptor submission can be ignored in this case
as has been established in review of the MMCI/PL180 driver.

Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

setup_transfer is mandatory if spi_bitbang_transfer is used, so
check for it during initialization and not each time during runtime.
Signed-off-by: NSascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

If there is an error with setting up a transfer, we need to return
immediately rather than trying to continue to process things.  We
already set up the error states for the caller at this point.
Signed-off-by: NSonic Zhang <sonic.zhang@analog.com>
Signed-off-by: NMike Frysinger <vapier@gentoo.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

We can do multiples of 8bit transfers when using the hardware CS and a
little bit of magic, so make it work.
Signed-off-by: NBob Liu <lliubbo@gmail.com>
Signed-off-by: NMike Frysinger <vapier@gentoo.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Add the compat_ioctl for operations on /dev/spi* so that 32 bit
userspace applications can access SPI. As far as I can see all data
structure are already prepared for that, so no additional conversion has
to be done.

My use case is MIPS with N32 userspace ABI and toolchain, and that was
also the platform where I tested it successfully (Cavium Octeon).
Signed-off-by: NBernhard Walle <walle@corscience.de>
Reviewed-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Instead of sending data 8 bits at a time in 8-bit SPI mode, swap bytes
and send and receive them 32 bits at a time. Tested with an SD-card,
with which this patch reduced the number of interrupts by 50%, when
reading 5MiB of data (there are also small service packets, the number
of interrupts, produced by 512-byte sectors should, of course, drop by
75%), and improved throughput by more than 40%.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

1. sort headers alphabetically
2. use fixed-size types u8, u16, u32 for register values and transferred data
3. simplify some arithmetic operations
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
[grant.likely@secretlab.ca: removed label indentation change]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Having conditional around the of_match_table and the of_node pointers
turns out to make driver code use ugly #ifdef blocks.  Drop the
conditionals and remove the #ifdef blocks from the affected drivers.

Also tidy up minor whitespace issues within the same hunks.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>

The Atheros AR71XX/AR724X/AR913X SoCs have a built-in SPI controller. This
patch implements a driver for that.
Signed-off-by: NGabor Juhos <juhosg@openwrt.org>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Cc: spi-devel-general@lists.sourceforge.net
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: linux-mips@linux-mips.org
Cc: Imre Kaloz <kaloz@openwrt.org>
Cc: Luis R. Rodriguez <lrodriguez@atheros.com>
Cc: Cliff Holden <Cliff.Holden@Atheros.com>
Cc: Kathy Giori <Kathy.Giori@Atheros.com>
Patchwork: https://patchwork.linux-mips.org/patch/1960/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Without this fix reloading of the driver is impossible.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

clk_freq is used uninitialized in pl022_setup routine. This patch
fix compilation warning for using uninitialized variable
Signed-off-by: NViresh Kumar <viresh.kumar@st.com>
Acked-by: NLinus Walleij <linus.walleij@stericsson.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This patch enables the existing S3C64XX series SPI driver for S5P64X0
and removed dependency on EXPERIMENTAL because we don't need it now.

v3: Changed dependency of S3C64XX_DMA
v2: Removed dependency on EXPERIMENTAL
Signed-off-by: NAbhilash Kesavan <a.kesavan@samsung.com>
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Acked-by: NJassi Brar <jassi.brar@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

1. Change the Kconfig to include i.MX53
2. add devtype entry for i.MX53
Signed-off-by: NYong Shen <yong.shen@freescale.com>
Acked-by: NSascha Hauer <s.hauer@pengutronix.de>
Acked-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

clk_get() returns a struct clk cookie to the driver and some platforms
may return NULL if they only support a single clock.  clk_get() has only
failed if it returns a ERR_PTR() encoded pointer.
Signed-off-by: NJamie Iles <jamie@jamieiles.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Some platforms have been known to return NULL from clk_get() if they
support only a single struct clk.  Whilst tegra doesn't do this, make
the drivers consistent with others.
Signed-off-by: NJamie Iles <jamie@jamieiles.com>
Acked-by: NRussell King <linux@arm.linux.org.uk>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Running the latest kernel on the 4430SDP board with DMA API debugging
enabled results in this:

WARNING: at lib/dma-debug.c:803 check_unmap+0x19c/0x6f0()
NULL NULL: DMA-API: device driver tries to free DMA memory it has not allocated
[device address=0x000000008129901a] [size=260 bytes]
Modules linked in:
Backtrace:
[<c003cbe0>] (dump_backtrace+0x0/0x10c) from [<c0278da8>] (dump_stack+0x18/0x1c)
 r7:c1839dc0 r6:c0198578 r5:c0304b17 r4:00000323
[<c0278d90>] (dump_stack+0x0/0x1c) from [<c005b158>] (warn_slowpath_common+0x58/0x70)
[<c005b100>] (warn_slowpath_common+0x0/0x70) from [<c005b214>] (warn_slowpath_fmt+0x38/0x40)
 r8:c1839e40 r7:00000000 r6:00000104 r5:00000000 r4:8129901a
[<c005b1dc>] (warn_slowpath_fmt+0x0/0x40) from [<c0198578>] (check_unmap+0x19c/0x6f0)
 r3:c03110de r2:c0304e6b
[<c01983dc>] (check_unmap+0x0/0x6f0) from [<c0198cd8>] (debug_dma_unmap_page+0x74/0x80)
[<c0198c64>] (debug_dma_unmap_page+0x0/0x80) from [<c01d5ad8>] (omap2_mcspi_work+0x514/0xbf0)
[<c01d55c4>] (omap2_mcspi_work+0x0/0xbf0) from [<c006dfb0>] (process_one_work+0x294/0x400)
[<c006dd1c>] (process_one_work+0x0/0x400) from [<c006e50c>] (worker_thread+0x220/0x3f8)
[<c006e2ec>] (worker_thread+0x0/0x3f8) from [<c00738d0>] (kthread+0x88/0x90)
[<c0073848>] (kthread+0x0/0x90) from [<c005e924>] (do_exit+0x0/0x5fc)
 r7:00000013 r6:c005e924 r5:c0073848 r4:c1829ee0
---[ end trace 1b75b31a2719ed20 ]---

I've no idea why this driver uses NULL for dma_unmap_single instead of
the &spi->dev that is laying around just waiting to be used in that
function - but it's an easy fix.

Also replace this comment with a FIXME comment:
                /* Do DMA mapping "early" for better error reporting and
                 * dcache use.  Note that if dma_unmap_single() ever starts
                 * to do real work on ARM, we'd need to clean up mappings
                 * for previous transfers on *ALL* exits of this loop...
                 */
as the comment is not true - we do work in dma_unmap() functions,
particularly on ARMv6 and above.  I've corrected the existing unmap
functions but if any others are required they must be added ASAP.
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The test "if (spi_imx->irq <= 0)" is not testing the IRQ value, but
the return value of platform_get_irq().  As platform_get_irq() can
return an error (-ENXIO) or the IRQ value it found, the test should be
"if (spi_imx->irq < 0)"

[grant.likely: Note: In general, Linux irq number 0 should also mean
no irq, but arm still allows devices to be assigned 0, and the imx
platform uses 0 for one of the spi devices, so this patch is needed
for the device to work]
Signed-off-by: NRichard Genoud <richard.genoud@gmail.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NJustin P. Mattock <justinmattock@gmail.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NJustin P. Mattock <justinmattock@gmail.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

After grabbing a msg from the msgq, the mcfqspi_work function calls
list_del_init on the mcfqspi->msgq which unintentionally deletes the rest
of the list before it can be processed. If qspi call was made using
spi_sync, this can result in a process hang.
Signed-off-by: NJate Sujjavanich <jsujjavanich@syntech-fuelmaster.com>
Acked-by: NSteven King <sfking@fdwdc.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The user must read N bytes of SPIRF (1 <= N <= 4) that do not exceed the
amount of data in the receive FIFO, so read the SPIRF byte by byte when
the data in receive FIFO is less than 4 bytes.

On Simics, when read N bytes that exceed the amout of data in receive
FIFO, we can't read the data out, that is we can't clear the rx FIFO,
then the CPU will loop on the espi rx interrupt.
Signed-off-by: NMingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

When SPI wake up from OFF mode, CS is in the wrong state: force it to the
inactive state.

During the system life, I monitored the CS behavior using a oscilloscope.
I also activated debug in omap2_mcspi, so I saw when driver disable the clocks
and restore context when device is not used.Each time the CS was in the correct
state.  It was only when system was put suspend to ram with off-mode activated
that on resume the CS was in wrong state( ie activated).

Changelog:
* Change from v1 to v2:
  - Rebase on linus/master (after 2.6.37-rc1)
  - Do some clean-up and fix indentation on both patches
  - Add more explanations for patch 2

* Change from v2 to v3:
  - Use directly resume function of spi_master instead of using function
  - from spi_device as Grant Likely pointed it out.
  - Force this transition explicitly for each CS used by a device.

* Change from v3 to v4:
  - Patch clean-up according to Kevin Hilman and checkpatch.
  - Now force CS to be in inactive state only if it was inactive when it was
    suspended.

* Change from v4 to v5:
  - Rebase on linus/master (after 2.6.37-rc3)
  - Collapse some lines as pointed by Grant Likely
  - Fix a spelling

* Change from v5 to v6:
  - Rebase on linus/master (after 2.6.37-rc7)
  - Use CONFIG_SUSPEND instead of CONFIG_PM
  - Didn't use legacy PM methods anymore. Instead, add a struct dev_pm_ops and
    add the resume method there.
  - Fix multi-line comment style

* Change from v6 to v7:
  - Rebase on linus/master (after 2.6.37-rc8)
  - Drop an extra line
Signed-off-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Reviewed-by: NKevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Allow SPI drivers to use runtime PM and other dev_pm_ops features by
implementing dev_pm_ops for the bus. The existing bus specific suspend
and resume functions will be called if a driver does not provide dev_pm_ops
allowing for transition to the new model.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Probe/remove should be in __devinit/__devexit. Found by a section-mismatch
warning. Also, copy do_remove() over to remove().
Signed-off-by: NWolfram Sang <w.sang@pengutronix.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NWolfram Sang <w.sang@pengutronix.de>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Or else we can't operate on the right address when the trans length
is greater than 65535.
Signed-off-by: NMingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The user must read N bytes of SPIRF (1 <= N <= 4) that do not exceed the
amount of data in the receive FIFO, so read the SPIRF byte by byte when
the data in receive FIFO is less than 4 bytes.

On Simics, when read N bytes that exceed the amount of data in receive
FIFO, we can't read the data out, that is we can't clear the rx FIFO,
then the CPU will loop on the espi rx interrupt.
Signed-off-by: NMingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

dw_spi driver in upstream only supports PIO mode, and this patch
will support it to cowork with the Designware dma controller used
on Intel Moorestown platform, at the same time it provides a general
framework to support dw_spi core to cowork with dma controllers on
other platforms

It has been tested with a Option GTM501L 3G modem and Infenion 60x60
modem. To use DMA mode, DMA controller 2 of Moorestown has to be enabled

Also change the dma interface suggested by Linus Walleij.
Acked-by: NLinus Walleij <linus.walleij@stericsson.com>
Signed-off-by: NFeng Tang <feng.tang@intel.com>
[Typo fix and renames to match intel_mid_dma renaming]
Signed-off-by: NVinod Koul <vinod.koul@intel.com>
Signed-off-by: NAlan Cox <alan@linux.intel.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NFeng Tang <feng.tang@intel.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The SPI polling loop timeout only works with HZ=100 as the loop was
actually too short.

Also add appropriate cpu_relax() in the busy wait loops...
Signed-off-by: NArjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NAlan Cox <alan@linux.intel.com>
Signed-off-by: NFeng Tang <feng.tang@intel.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This variable is a bool but defined an int and defined completely
backwards. This makes the code more readable.
Signed-off-by: NLinus Walleij <linus.walleij@stericsson.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Signed-off-by: NLinus Walleij <linus.walleij@stericsson.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The sglen return by the dma_map_sg() should be passed to the DMA
engine, not the one passed in. If we one day have a DMA mapper
that can coalesce entries, this will bug due to a too large
number of entries being passed in.
Reported-by: NRussell King <linux@arm.linux.org.uk>
Signed-off-by: NLinus Walleij <linus.walleij@stericsson.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The struct device for the DMA engine is the apropriate one to use
when mapping/unmapping buffers. This is because the memory which
is addressable by DMA is determined by the DMA engine rather than
the device.
Reported-by: NRussell King <linux@arm.linux.org.uk>
Signed-off-by: NLinus Walleij <linus.walleij@stericsson.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

It seems spi_topcliff_pch of linux-2.6.37-rc6 degraded by previous patch.
In fact, data transfer fails on evaluation board testing.
I found like the following register miss-setting line.
Using this patch, I have confirmed data transfer can work well.
Signed-off-by: NTomoya MORINAGA <tomoya-linux@dsn.okisemi.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

There are no machines in-tree that still use the driver
name as device name.  So save a few bytes and remove it.
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: NJason Wang <jason77.wang@gmail.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>