Switch to use threaded interrupt context in order to avoid checking of
"are we in interrupt?" for the code that may sleep in the IRQ handler.
I2C doesn't require a very low interrupt-handling latency, hence this
change doesn't introduce any noticeable effects.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
VI I2C controller has known hardware bug where immediate multiple
writes to TX_FIFO register gets stuck.
Recommended software work around is to read I2C register after
each write to TX_FIFO register to flush out the data.
This patch implements this work around for VI I2C controller.
Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Upon a communication error, the interrupt handler can call
tegra_i2c_disable_packet_mode. This causes a sleeping poll to happen
unless the current transaction was marked atomic. Fix this by
making the poll happen atomically if we are in an IRQ.
This matches the behavior prior to the patch mentioned
in the Fixes tag.
Fixes: ede2299f71 ("i2c: tegra: Support atomic transfers")
Cc: stable@vger.kernel.org
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Use proper spelling of "NVIDIA" and don't designate driver as Tegra2-only
since newer SoC generations are supported as well.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Some places in the code are missing newlines or have unnecessary
whitespaces and newlines. This creates inconsistency of the code and
hurts readability. This patch removes the unnecessary and adds necessary
whitespaces / newlines, clears indentation of the code.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Make all comments to be consistent in regards to capitalization and
punctuation, correct spelling and grammar errors, improve wording.
Reviewed-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
This patch unifies style of all messages in the driver by starting them
with a lowercase letter and using consistent capitalization and wording
for all messages.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Rename "ret" variables to "err" in order to make code a bit more
expressive, emphasizing that the returned value is an error code.
Same vice versa, where appropriate.
Rename variable "reg" to "val" in order to better reflect the actual
usage of the variable in the code and to make naming consistent with
the rest of the code.
Use briefer names for a few members of the tegra_i2c_dev structure in
order to improve readability of the code.
All dev/&pdev->dev are replaced with i2c_dev->dev in order to have uniform
code style across the driver.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Reorder definition of variables in the code to have them sorted by length
and grouped logically, also replace "unsigned long" with "u32". Do this in
order to make code easier to read.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The driver's code is inconsistent in regards to the error values checking.
The correct way should be to check both positive and negative values.
This patch cleans up the error-checks in the code. Note that the
pm_runtime_get_sync() could return positive value on success, hence only
relevant parts of the code are changed by this patch.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Factor out hardware initialization into a separate function from the probe
function. The only place where runtime PM needs to be resumed during probe
is the place of hardware initialization, hence it makes sense to factor
out it in order to have a bit cleaner error handling in tegra_i2c_probe().
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Factor out register polling into a separate function in order to remove
boilerplate code and make code cleaner.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The code related to packet header setting up is a bit messy and makes
tegra_i2c_xfer_msg() more difficult to read than it could be. Let's
factor the packet header setup from tegra_i2c_xfer_msg() into separate
function in order to make code easier to read and follow.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Factor out error recovery code from tegra_i2c_xfer_msg() in order to
make this function easier to read and follow.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Drop '_timeout' postfix from the wait/poll completion function names in
order to make the names shorter, making code cleaner a tad.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The "dma" variable of tegra_i2c_xfer_msg() function doesn't bring much in
regards to readability and generation of the code.
Besides readability, it's also not very nice that the is_curr_dma_xfer
is initialized in tegra_i2c_xfer_msg() and then could be overridden by
tegra_i2c_config_fifo_trig(). In a result, the "dma" variable creates
slight confusion since it's not instantly obvious why it's set after
tegra_i2c_config_fifo_trig().
Hence should be better to have the variable removed. This makes code
more consistent.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The tegra_i2c_wait_for_config_load() checks for 'has_config_load_reg' by
itself, hence there is no need to duplicate the check.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The likely/unlikely annotations should be used only in a hot paths of
performance-critical code. The I2C driver doesn't have such paths, and
thus, there is no justification for usage of likely/unlikely annotations
in the code. Hence remove them.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The barrier() was intended to reduce possibility of racing with the
interrupt handler, but driver's code evolved significantly and today's
driver enables interrupt only when it waits for completion notification.
Hence barrier() has no good use anymore, let's remove it.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Don't use signed types for unsigned values and use consistent types
for sibling variables.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Reorder location of functions in the code in order to have definition
of functions closer to the place of the invocation. This change makes
easier to navigate around the code and removes the need to have a
prototype for tegra_i2c_init().
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The driver's probe function code is a bit difficult to read. This patch
reorders code of the probe function, forming groups of code that are easy
to work with. The probe tear-down order now matches the driver-removal
order.
Reviewed-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Move out code related to device-tree parsing from the probe function into
tegra_i2c_parse_dt() in order to make code more consistent.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Use clk-bulk helpers and factor out clocks initialization into separate
function in order to make code cleaner.
The clocks initialization now performed after reset-control initialization
in order to avoid a noisy -PROBE_DEFER errors on T186+ from the clk-bulk
helper which doesn't silence this error code. Hence reset_control_get()
now may return -EPROBE_DEFER on newer Tegra SoCs because they use BPMP
driver that provides reset controls and BPMP doesn't come up early during
boot. Previously rst was protected by the clocks retrieval and now this
patch makes dev_err_probe() to be used for the rst error handling.
Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Use common helper for retrieval of the interrupt number in order to make
code cleaner. Note that platform_get_irq() prints error message by itself.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Driver now uses devm_platform_get_and_ioremap_resource() which replaces
the typical boilerplate code and makes code cleaner.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Use a single reset_control_reset() instead of assert/deasset couple in
order to make code cleaner a tad. Note that the reset_control_reset()
uses 1 microsecond delay instead of 2 that was used previously, but this
shouldn't matter because one microsecond is a default reset time for most
of Tegra peripherals and TRM doesn't mention anything special in regards
to I2C controller's reset propagation time.
In addition don't ignore potential error of the reset control by emitting
a noisy warning if it fails, which will indicate an existence of a severe
problem, while still allow machine to boot up.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The error message prints number of vIRQ, which isn't a useful information.
In practice devm_request_irq() never fails, hence let's remove the bogus
message in order to make code cleaner.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The runtime PM is guaranteed to be always available on Tegra after commit
40b2bb1b13 ("ARM: tegra: enforce PM requirement"). Hence let's remove
all the RPM-availability checking and handling from the code.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The "non_hs_mode" divisor value is fixed, thus there is no need to have
the variable i2c_dev.clk_divisor_non_hs_mode struct member. Let's remove
it and move the mode selection into tegra_i2c_init() where it can be
united with the timing selection.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
It doesn't make sense to conditionalize the div-clk rate changes because
rate is fixed and it won't ever change once it's set at the driver's probe
time. All further changes are NO-OPs because CCF caches rate and skips
rate-change if rate is unchanged.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The tegra_i2c_issue_bus_clear() may fail and in this case BUS_CLR_DONE
stays unmasked. Hence let's mask it for consistency. This patch doesn't
fix any known problems.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Technically the tegra_i2c_flush_fifos() may fail and transfer should be
aborted in this case, but this shouldn't ever happen in practice unless
there is a bug somewhere in the driver. Let's add the error check just
for completeness.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The pm_runtime_get_sync() always bumps refcount regardless of whether it
succeeds or fails. Hence driver is responsible for restoring of the RPM
refcounting. This patch adds missing RPM puts which restore refcounting
in a case of pm_runtime_get_sync() error.
Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The tegra_i2c_flush_fifos() shouldn't sleep in atomic transfer and jiffies
are not updating if interrupts are disabled. Let's switch to use iopoll
API helpers for register-polling. The iopoll API provides helpers for both
atomic and non-atomic cases.
Note that this patch doesn't fix any known problem because normally FIFO
is flushed at the time of starting a new transfer.
Reviewed-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
VI I2C is on host1x bus so APB DMA can't be used for Tegra210 VI
I2C and there are no tx and rx dma channels for VI I2C.
So, avoid attempt of requesting DMA channels.
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
VI I2C is on host1x bus and is part of VE power domain.
During suspend/resume VE power domain goes through power off/on.
So, controller reset followed by i2c re-initialization is required
after the domain power up.
This patch fixes it.
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
tegra_i2c_runtime_resume does not disable prior enabled clocks
properly.
This patch fixes it.
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
clk_enable, clk_disable, clk_prepare, and clk_unprepare APIs have
implementation for checking clk pointer not NULL and clock consumers
can safely call these APIs without NULL pointer check.
So, this patch cleans up Tegra i2c driver to remove explicit checks
before these APIs.
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Tegra VI I2C is part of VE power domain and typically used for
camera usecases.
VE power domain is not always on and is non-IRQ safe. So, IRQ safe
device cannot be attached to a non-IRQ safe domain as it prevents
powering off the PM domain and generic power domain driver will warn.
Current driver marks all I2C devices as IRQ safe and VI I2C device
does not require IRQ safe as it will not be used for atomic transfers.
This patch has fix to make VI I2C as non-IRQ safe.
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Tegra210 has an extra instance of the I2C controller that is in the
domain of host1x and usually used for camera use-cases. The programming
model for the VI variant of the controller is roughly the same as for
the other variants, except that the I2C registers start at an offset
and are spaced further apart. VI I2C also doesn't support slave mode.
Signed-off-by: Thierry Reding <treding@nvidia.com>
DMA transfer could be completed, but CPU (which handles DMA interrupt)
may get too busy and can't handle the interrupt in a timely manner,
despite of DMA IRQ being raised. In this case the DMA state needs to
synchronized before terminating DMA transfer in order not to miss the
DMA transfer completion.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Boot CPU0 always handle I2C interrupt and under some rare circumstances
(like running KASAN + NFS root) it may stuck in uninterruptible state for
a significant time. In this case we will get timeout if I2C transfer is
running on a sibling CPU, despite of IRQ being raised. In order to handle
this rare condition, the IRQ status needs to be checked after completion
timeout.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
One of the I2C controllers on Tegra SoCs is typically connected to a
system PMIC, which provides controls for critical power supplies for
most platforms.
Some drivers, such as PCI, need to disable these regulators during a
very late stage during suspend and resume them at a very early stage
during resume.
To support these use-cases, keep interrupts disabled during suspend/
resume.
Suggested-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Depending on the board design, the I2C controllers found on Tegra SoCs
may require pinmuxing in order to function. This is done as part of the
driver's runtime suspend/resume operations. However, the PM core does
not allow devices to go into runtime suspend during system sleep to
avoid potential races with the suspend/resume of their parents.
As a result of this, when Tegra SoCs resume from system suspend, their
I2C controllers may have lost the pinmux state in hardware, whereas the
pinctrl subsystem is not aware of this. To fix this, make sure that if
the I2C controller is not runtime suspended, the runtime suspend code is
still executed in order to disable the module clock (which we don't need
to be enabled during sleep) and set the pinmux to the idle state.
Conversely, make sure that the I2C controller is properly resumed when
waking up from sleep so that pinmux settings are properly restored.
This fixes a bug seen with DDC transactions to an HDMI monitor timing
out when resuming from system suspend.
Signed-off-by: Thierry Reding <treding@nvidia.com>
This reverts commit 9f42de8d4e.
It's not safe to use pm_runtime_force_{suspend,resume}(), especially
during the noirq phase of suspend. See also the guidance provided in
commit 1e2ef05bb8 ("PM: Limit race conditions between runtime PM
and system sleep (v2)").
Signed-off-by: Thierry Reding <treding@nvidia.com>
DMA transfer could be completed, but CPU (which handles DMA interrupt)
may get too busy and can't handle the interrupt in a timely manner,
despite of DMA IRQ being raised. In this case the DMA state needs to
synchronized before terminating DMA transfer in order not to miss the
DMA transfer completion.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
