The TX_FIFO register is 10 bits wide.  The lower 8 bits are the data to be
written, while the upper two bits are flags to indicate stop/start.

The driver apparently attempted to optimize write access, by only writing a
byte in those cases where the stop/start bits are zero.  However, we have
seen cases where the lower byte is duplicated onto the upper byte by the
hardware, which causes inadvertent stop/starts.

This patch changes the write access to the transmit FIFO to always be 16 bits
wide.

Signed off by: Steven A. Falco <sfalco@harris.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Cc: stable@kernel.org

HZ based value is better than a magic number.
Signed-off-by: Eunbong Song <eunb.song@samsung.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

I've been debugging the abnormal operation of i2c on octeon. If a process is
terminated by signal in the middle of i2c operation, next i2c read operation
which is done by another process was failed. So i changed to ignore signal in
the middle of i2c operation. After that the problem was not reproduced.
Signed-off-by: Eunbong Song <eunb.song@samsung.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Driver core already takes care of refcounting, no need to do this on
driver level again.
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Tested-by: Mika Westerberg <mika.westerberg@linux.intel.com>

Driver core already takes care of refcounting, no need to do this on
driver level again.
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Tested-by: Mika Westerberg <mika.westerberg@linux.intel.com>

Driver core already takes care of refcounting, no need to do this on
driver level again.
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Tested-by: Sekhar Nori <nsekhar@ti.com>

Ben has done valuable work for the I2C subsystem, yet he has been
inactive for ~15 months. Let MAINTAINERS reflect that. He will be
most welcome if he wants to participate again.
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Cc: Ben Dooks <ben-linux@fluff.org>

The i2c-arb-gpio-challenge driver implements an I2C arbitration scheme
where masters need to claim the bus with a GPIO before they can start
a transaction.  This should generally only be used when standard I2C
multimaster isn't appropriate for some reason (errata/bugs).

This driver is based on code that Simon Glass added to the i2c-s3c2410
driver in the Chrome OS kernel 3.4 tree.  The current incarnation as a
mux driver is as suggested by Grant Likely.  See
<https://patchwork.kernel.org/patch/1877311/> for some history.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Naveen Krishna Chatradhi <ch.naveen@samsung.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Use generic DMA DT helper. Platforms booting with or without DT populated are
both supported.
Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

In PIO mode we can end up with the same errors as in DMA mode, but as IRQs
are disabled there we have to check for them manually after each command.

Also don't use the big controller reset hammer when receiving a NAK from a
slave. It's sufficient to tell the controller to continue at a clean state.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Marek Vasut <marex@denx.de>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

This commit fixes the three following races in PIO code:

- The CTRL0 register is racy in itself, when programming transfer state and
  run bit in the same cycle the hardware sometimes ends up using the state
  from the last transfer. Fix this by programming state in one cycle, make
  sure the write is flushed down APBX bus by reading back the reg and only
  then trigger the run bit.

- Only clear the DMAREQ bit in DEBUG0 after the read/write to the data reg
  happened. Otherwise we are racing with the hardware about who touches
  the data reg first.

- When checking for completion of a transfer it's not sufficient to check
  if the data engine finished, but also a check for i2c bus idle is needed.
  In PIO mode we are really fast to program the next transfer after a finished
  one, so the controller possibly tries to start a new transfer while the
  clkgen engine is still busy writing the NAK/STOP from the last transfer to
  the bus.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Using autosuspend helps to reduce the resume latency in situations where
another I2C message is going to be started soon. For example with HID over
I2C touch panels we get several messages in a short period of time while
the touch panel is in use.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

This is not an atomic context so there is no need to use mdelay() but
instead use usleep_range().
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

The correct way to disable or enable the controller is to wait until the
DW_IC_ENABLE_STATUS register bit matches the bit we program into DW_IC_ENABLE
register. This procedure is described in the DesignWare I2C databook.

By doing this we can be sure that the controller is in correct state once
the function returns.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

It is not good idea to mix static and dynamic I2C adapter numbering. In
this particular case on Lynxpoint we had graphics I2C adapter which took
the first numbers preventing the designware I2C driver from using the
adapter numbers it preferred.

Since Lynxpoint support was just introduced and there is no hardware available
outside Intel we can fix this by switching to use dynamic adapter numbering
instead of static.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

This makes the error handling much more simpler than open-coding everything
and in addition makes the probe function smaller an tidier.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

With dev_err() we can get the device instance printed as well and is pretty
much standard to use dev_* macros in the drivers anyway. In addition
correct the indentation of probe() arguments.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

This makes the error handling much more simpler than open-coding everything
and in addition makes the probe function smaller and tidier.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

CONFIG_HOTPLUG is going away as an option, cleanup CONFIG_HOTPLUG
ifdefs in i2c files.
Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

SMBus read and write are supported by the emulation layer of i2c
framework if the controller doesn't have SMBus features.

I2C_M_RECV_LEN flag is used to let i2c drivers know rx length is not
yet determined but will be read to the first byte in rx buffer.

s3c2410 doesn't handle this flag. So only one byte is read from slave.
There fore following two features are added to the driver code.

1. skip rx length check if I2C_M_RECV_LEN is set and the length is 1.
2. add actual bytes to the rx length after reading first bytes if
   I2C_M_RECV_LEN.

I2C_M_RECV_LEN is only set for SMBus command. So this code does not
affect legacy codes which only use i2c command for s3c2410.
Signed-off-by: Jaemin Yoo <jmin.yoo@samsung.com>
Tested-by: Prasanna Kumar <prasanna.ps@samsung.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

i2c_del_mux_adapter always returns 0 and none of it current users check its
return value anyway. It is also an essential requirement of the Linux device
driver model, that functions which may be called from a device's remove callback
to free resources provided by the device, are not allowed to fail. This is the
case for i2c_del_mux_adapter(), so make its return type void to make the
fact that it won't fail explicit.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Reviewed-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

i2c_del_mux_adapter() always returns 0. So all checks testing whether it will be
non zero will always evaluate to false and the conditional code is dead code.
This patch updates all callers of i2c_del_mux_adapter() to ignore its return
value and assume that it will always succeed (which it will). A subsequent
patch will make the return type of i2c_del_mux_adapter() void.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

i2c_del_adapter() is usually called from a drivers remove callback. The Linux
device driver model does not allow the remove callback to fail and all resources
allocated in the probe callback need to be freed, as well as all resources which
have been provided to the rest of the kernel(for example a I2C adapter) need to
be revoked. So any function revoking such resources isn't allowed to fail
either. i2c_del_adapter() adheres to this requirement and will never fail. But
i2c_del_adapter()'s return type is int, which may cause driver authors to think
that it can fail. This led to code constructs like:

	ret = i2c_del_adapter(...);
	BUG_ON(ret);

Since i2c_del_adapter() always returns 0 the BUG_ON is never hit and essentially
becomes dead code, which means it can be removed. Making the return type of
i2c_del_adapter() void makes it explicit that the function will never fail and
should prevent constructs like the above from re-appearing in the kernel code.

All callers of i2c_del_adapter() have already been updated in a previous patch
to ignore the return value, so the conversion of the return type from int to
void can be done without causing any build failures.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Reviewed-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

i2c_del_adapter() always returns 0. So all checks testing whether it will be
non zero will always evaluate to false and the conditional code is dead code.
This patch updates all callers of i2c_del_mux_adapter() to ignore the return
value and assume that it will always succeed (which it will). In a subsequent
patch the return type of i2c_del_adapter() will be made void.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Reviewed-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Currently i2c_del_adapter() returns -EINVAL when it gets an adapter which is not
registered. But none of the users of i2c_del_adapter() depend on this behavior,
so for the sake of being able to sanitize the return type of i2c_del_adapter
argue, that the purpose of i2c_del_adapter() is to remove an I2C adapter from
the system. If the adapter is not registered in the first place this becomes a
no-op. So we can return success without having to do anything.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Reviewed-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

The detach_adapter callback has been deprecated for quite some time and has no
user left. Keeping it alive blocks other cleanups, so remove it.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Reviewed-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

This eliminates having an #ifdef returning NULL for the case
when OF is disabled.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Update the code to use devm_* API so that driver core will manage
resources.
Signed-off-by: Vishwanathrao Badarkhe, Manish <manishv.b@ti.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Since we have generic i2c bus recover routines now, these custom ones
need to be renamed to fix the namespace clash. Proper conversion needs
to be done by someone who has access to the hardware.
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

gpio_direction_output() may fail, check for that and deal with it
appropriately. Also log an error message if gpio_request() fails.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Peter Korsgaard <peter.korsgaard@barco.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

GPIOs may not be available immediately when i2c-gpio looks for them.
Implement support for deferred probing so that probing can be
attempted again later when GPIO pins are finally available.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Tested-by: Bo Shen <voice.shen@atmel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Make them conform more to established standards.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

The register definitions are only used in the driver itself.
This also removes the last dependency on plat/ includes from the
i2c driver.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

(SMATCH) drivers/i2c/busses/i2c-ismt.c:186:14: warning: duplicate const
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Tegra only supports, and always enables, device tree. Remove all ifdefs
and runtime checks for DT support from the driver. Platform data is
therefore no longer required. Delete the header that defines it.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

The commit: "i2c-core: dt: Pick i2c bus number from i2c alias if
present" adds support for automatically picking the bus number based
on the alias ID.  Remove the now unnecessary code from i2c-pxa that
did the same thing.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add i2c bus recovery infrastructure to i2c adapters as specified in the i2c
protocol Rev. 03 section 3.1.16 titled "Bus clear".

http://www.nxp.com/documents/user_manual/UM10204.pdf

Sometimes during operation i2c bus hangs and we need to give dummy clocks to
slave device to start the transfer again. Now we may have capability in the bus
controller to generate these clocks or platform may have gpio pins which can be
toggled to generate dummy clocks. This patch supports both.

This patch also adds in generic bus recovery routines gpio or scl line based
which can be used by bus controller. In addition controller driver may provide
its own version of the bus recovery routine.

This doesn't support multi-master recovery for now.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
[wsa: changed gpio type to int and minor reformatting]
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

This allows you to get the equivalent functionality of
i2c_add_numbered_adapter() with all data in the device tree and no
special case code in your driver.  This is a common device tree
technique.

For quick reference, the FDT syntax for using an alias to provide an
ID looks like:
  aliases {
    i2c0 = &i2c_0;
    i2c1 = &i2c_1;
  };
Signed-off-by: Doug Anderson <dianders@chromium.org>
[wsa: removed one check from static function. We know our callers]
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Pull SCSI target fixes from Nicholas Bellinger:
 "These are mostly minor fixes this time around.  The iscsi-target CHAP
  big-endian bugfix and bump FD_MAX_SECTORS=2048 default patch to allow
  1MB sized I/Os for FILEIO backends on >= v3.5 code are both CC'ed to
  stable.

  Also, there is a persistent reservations regression that has recently
  been reported for >= v3.8.x code, that is currently being tracked down
  for v3.9."

* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
  target/pscsi: Reject cross page boundary case in pscsi_map_sg
  target/file: Bump FD_MAX_SECTORS to 2048 to handle 1M sized I/Os
  tcm_vhost: Flush vhost_work in vhost_scsi_flush()
  tcm_vhost: Add missed lock in vhost_scsi_clear_endpoint()
  target: fix possible memory leak in core_tpg_register()
  target/iscsi: Fix mutual CHAP auth on big-endian arches
  target_core_sbc: use noop for SYNCHRONIZE_CACHE