There are two problems with WARN_ON() here. One: It is not ratelimited.
Two: We don't see which adapter was used when trying to transfer
something when already suspended. Implement a custom ratelimit once per
adapter and use dev_WARN there. This fixes both issues. Drawback is that
we don't see if multiple drivers are trying to transfer with the same
adapter while suspended. They need to be discovered one after the other
now. This is better than a high CPU load because a really broken driver
might try to resend endlessly.

Fixes: 9ac6cb5f ("i2c: add suspended flag and accessors for i2c adapters")
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Cc: stable@vger.kernel.org # v5.1+

Mainly some pca954x work, i.e. removal of unused platform data support
and added support for sysfs interface for manipulating/examining the
idle state. And then a mechanical cocci-style patch.

Merge tag 'at24-v5.2-updates-for-wolfram' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux into i2c/for-5.2

at24: updates for linux v5.2

- change my e-mail address
- add a new compatible for Renesas R1EX24016 to the DT binding document

Change the iProc I2C driver to use the 'BIT' macro from all '1 << XXX'
bit operations to get rid of compiler warning and improve readability of
the code
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

- Replace explicit clock handling by Runtime PM calls,
  - Streamline Runtime PM handling in error paths,
  - Enable Runtime PM in .probe(),
  - Disable Runtime PM in .remove(),
  - Make sure the device is runtime-resumed when disabling interrupts in
    .remove().
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Tested-by: Chris Brandt <chris.brandt@renesas.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

One of the more common cases of allocation size calculations is finding
the size of a structure that has a zero-sized array at the end, along
with memory for some number of elements for that array. For example:

struct foo {
    int stuff;
    void *entry[];
};

instance = devm_kzalloc(dev, sizeof(struct foo) + sizeof(void *) * count, GFP_KERNEL);

Instead of leaving these open-coded and prone to type mistakes, we can
now use the new struct_size() helper:

instance = devm_kzalloc(dev, struct_size(instance, entry, count), GFP_KERNEL);

This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Peter Rosin <peda@axentia.se>

The behaviour, by default, to not deselect after each transfer is
unsafe when there is a device with an address that conflicts with
another device on another mux on the same parent bus, and it
may not be convenient to use devicetree to set the deselect mux,
e.g. when running on x86_64 when ACPI is used to discover most of the
device hierarchy.

Therefore, provide the ability to set the idle state behaviour using a
new sysfs file, idle_state as a complement to the method of
instantiating the device via sysfs. The possible behaviours are
disconnect, i.e. to deselect all channels from the mux, as-is (the
default), i.e. leave the last channel selected, and set a
predetermined channel.

The current behaviour of leaving the channel as-is after each
transaction is preserved.
Signed-off-by: Robert Shearman <robert.shearman@att.com>
Signed-off-by: Peter Rosin <peda@axentia.se>

There are no in-tree users of the platform data, so remove it to
simplify the code slightly.

Remove the now unused pca954x.h platform data header.
Signed-off-by: Robert Shearman <robert.shearman@att.com>
Signed-off-by: Peter Rosin <peda@axentia.se>

There are no in-tree users of the pca954x platform data and the
per-channel deselect configuration complicates efforts to export the
configuration to user-space in a way that could be applied to other
muxes. Therefore, remove support for the pca954x platform data.
Signed-off-by: Robert Shearman <robert.shearman@att.com>
Signed-off-by: Peter Rosin <peda@axentia.se>

Add binding documentation of i2c-mtk for MT8516 SoC.
Signed-off-by: Fabien Parent <fparent@baylibre.com>
Acked-by: Rob Herring <robh@kernel.org>
Acked-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Some recent commits to this driver were trying to make sure the TSS
interrupt is not generated on busy system due to 25ms timer expiring
between commands. It can still happen, however if STOP command is not
issued on time at the end of the transmission. If wait_for_completion in
axxia_i2c_xfer_msg() would not return after 25ms of getting an
interrupt, TSS will be generated and idev->err_msg will be set to
-ETIMEDOUT which will be returned from the axxia_i2c_xfer_msg(), even
though the transfer did actually succeed (STOP is automatically issued
when TSS triggers).

Fortunately, apart from already used manual and sequence commands, the
controller also has so called auto command. It works just like manual
mode but it but an automatic STOP is issued when either transfer length
is met or NAK is received from slave device.

This patch changes the axxia_i2c_xfer_msg() function so that auto
command is used for last message in transaction letting hardware manage
issuing STOP. TSS is disabled just after command transferring last
message finishes. Auto command, just like sequence, ends with SS
interrupt instead of SNS so handling of both had to be unified.

The axxia_i2c_stop() is no longer needed as the transfer can only end
with following conditions:
- fully successful - then last message was send by AUTO command and STOP
  was issued automatically
- NAK received - STOP is issued automatically by controller
- arbitration lost - STOP should not be issued as we don't control the
  bus
- IP interrupt received - this is sent when transfer length is set to 0
  for auto/sequence command. The check for that is done before START is
  send so no STOP is required
- TSS received between commands - STOP is issued by the controller
Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Reviewed-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

If switching GPIOs does not sleep, then we can support atomic transfers.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Use the new xfer_atomic callback to check a newly introduced flag to
whitelist atomic transfers. This will report configurations which
worked accidently.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Use the new callback to reject atomic transfers.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

The driver already has the routine in place, tie it to the new callback.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Cc: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

By properly setting up the algorithm at probe time, we can skip the
check at every transfer. This allows us to get rid of the flags
completely.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Cc: Andrew Lunn <andrew@lunn.ch>
Acked-by: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

The driver did handle this internally, convert it to use the new
callbacks.
Reviewed-by: Timo Alho <talho@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add the master_xfer_atomic hook to enable i2c transactions in irq
disabled contexts like the poweroff case.
Signed-off-by: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Keerthy <j-keerthy@ti.com>
[wsa: simplified code a little: 'timeout = !ret']
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

If the parent has an atomic callback, we need to translate it the same
way as the non-atomic callback.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

If the parent adapter has atomic_xfer callbacks, populate them for the
mux adapter as well. We can use the same translation function as for the
non-atomic xfer callback.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Peter Rosin <peda@axentia.se>
Reviewed-by Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

We had the request to access devices very late when interrupts are not
available anymore multiple times now. Mostly to prepare shutdown or
reboot. Allow adapters to specify a specific callback for this case.
Note that we fall back to the generic {master|smbus}_xfer callback if
this new atomic one is not present. This is intentional to preserve the
previous behaviour and avoid regressions. Because there are drivers not
using interrupts or because it might have worked "accidently" before.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Stefan Lengfeld <contact@stefanchrist.eu>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

If I2C transfers are executed in atomic contexts, trylock is used
instead of lock. This behaviour was missing for SMBUS, although a lot of
transfers are of SMBUS type, either emulated or direct. So, factor out
the locking routine into a helper and use it for I2C and SMBUS.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Commit cea443a8 ("i2c: Support i2c_transfer in atomic contexts")
added in_atomic() to the I2C core. However, the use of in_atomic()
outside of core kernel code is discouraged and was already[1] when this
code was added in early 2008. The above commit was a preparation for
commit b7a36701 ("i2c-pxa: Add polling transfer"). Its commit
message says explicitly it was added "for cases where I2C transactions
have to occur at times interrup[t]s are disabled". So, the intention was
'disabled interrupts'. This matches the use cases for atomic I2C
transfers I have seen so far: very late communication (mostly to a PMIC)
to powerdown or reboot the system. For those cases, interrupts are
disabled then. It doesn't seem that in_atomic() adds value.

After a discussion with Peter Zijlstra[2], we came up with a better set
of conditionals to match the use case.

The I2C core will soon gain an extra callback into bus drivers
especially for atomic transfers to make them more generic. The code
deciding which transfer to use (atomic/non-atomic) should mimic the
behaviour which locking to use (trylock/lock). This is why we add a
helper for it.

[1] https://lwn.net/Articles/274695/
[2] http://patchwork.ozlabs.org/patch/1067437/Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Stefan Lengfeld <contact@stefanchrist.eu>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

No further occurences in the driver.

Fixes: dd1aa252 ("i2c: brcmstb: Add Broadcom settop SoC i2c controller driver")
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add i2c compatible for MT8183. Compare to MT2712 i2c controller,
MT8183 has different register offsets. Ltiming_reg is added to
adjust low width of SCL. Arb clock and dma_sync are needed.
Signed-off-by: Qii Wang <qii.wang@mediatek.com>
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

When i2c and apdma use different source clocks, we should enable
synchronization between them.
Signed-off-by: Qii Wang <qii.wang@mediatek.com>
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

When two i2c controllers are internally connected to the same
GPIO pins, the arb clock is needed to ensure that the waveforms
do not interfere with each other. And we also need to enable
the interrupt to find arb lost, old i2c controllers also have
the bit.
Signed-off-by: Qii Wang <qii.wang@mediatek.com>
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add MT8183 i2c binding to binding file. Compare to MT2712 i2c
controller, MT8183 has different registers, offsets, and clock.
Signed-off-by: Qii Wang <qii.wang@mediatek.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

New i2c registers would have different offsets, so we use different
offsets array to distinguish different i2c registers version.
Signed-off-by: Qii Wang <qii.wang@mediatek.com>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Commit 54fb4a05 ("i2c: Check for ACPI resource conflicts") included
<linux/acpi.h> so we could use acpi_check_region(). Commit fd46a006
("i2c: convert i2c-isch to platform_device") removed the use of
acpi_check_region() but not the include.

Remove the now-unnecessary include of <linux/acpi.h>.  No functional change
intended.

Fixes: fd46a006 ("i2c: convert i2c-isch to platform_device")
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jean Delvare <jdelvare@suse.de>
Reviewed-by: Mukesh Ojha <mojha@codeaurora.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

"res" can't be NULL because it's a pointer to somewhere in the middle of
the "adev" struct.  Also probe() succeeded so there is no need to check
here.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add NIC I2C support to the iProc I2C driver. Access to the NIC I2C base
registers requires going through the IDM wrapper to map into the NIC's
address space
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Update iProc I2C binding document to add new compatible string
"brcm,iproc-nic-i2c". Optional property "brcm,ape-hsls-addr-mask" is
also added that allows configuration of the host view into the APE's
address for "brcm,iproc-nic-i2c"
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Use the following wrapper for read/write access of iProc i2c registers:
u32 iproc_i2c_rd_reg(struct bcm_iproc_i2c_dev *iproc_i2c,
                     u32 offset)
void iproc_i2c_wr_reg(struct bcm_iproc_i2c_dev *iproc_i2c, u32 offset,
                      u32 val)

This preps the driver for support of indirect register access required
by certain SoCs with this iProc I2C block integrated
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add polling support to the iProc I2C driver. Polling mode is
activated when the driver fails to obtain an interrupt ID from device
tree
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Update the binding document to make the 'interrupts' property optional.
For certain revisions of the I2C controller (e.g., iProc NIC I2C), I2C
interrupt is unwired to the interrupt controller. In such case, this
'interrupts' property should be left unspecified, and driver will fall
back to polling mode
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add support for more master error status including FIFO underrun and RX
FIFO full
Signed-off-by: Michael Cheng <ccheng@broadcom.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add slave mode support to the iProc I2C driver.
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Signed-off-by: Michael Cheng <ccheng@broadcom.com>
Signed-off-by: Shreesha Rajashekar <shreesha.rajashekar@broadcom.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Add support to allow I2C master read transfer up to 255 bytes.
Signed-off-by: Shreesha Rajashekar <shreesha@broadcom.com>
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>