Usage of regmap_field_alloc becomes much overhead when number of fields
exceed more than 3.
QCOM LPASS driver has extensively converted to use regmap_fields.

Using new bulk api to allocate fields makes it much more cleaner code to read!
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Tested-by: NSrinivasa Rao Mandadapu <srivasam@codeaurora.org>
Link: https://lore.kernel.org/r/20200925164856.10315-2-srinivas.kandagatla@linaro.orgSigned-off-by: NMark Brown <broonie@kernel.org>

While not destroying mutexes doesn't lead to memory leaks, it's still
the correct thing to do for mutex debugging accounting.
Signed-off-by: NBartosz Golaszewski <bgolaszewski@baylibre.com>
Link: https://lore.kernel.org/r/20200928120614.23172-1-brgl@bgdev.plSigned-off-by: NMark Brown <broonie@kernel.org>

Many error paths in __regmap_init rely on ret being pre-initialised to
-EINVAL, add an extra initialisation in after the new call to
regmap_set_name.

Fixes: 94cc89eb ("regmap: debugfs: Fix handling of name string for debugfs init delays")
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20200918152212.22200-1-ckeepax@opensource.cirrus.comSigned-off-by: NMark Brown <broonie@kernel.org>

Non-incrementing writes can fail if register + length crosses page
border. However for non-incrementing writes we should not check for page
border crossing. Fix this by passing additional flag to _regmap_raw_write
and passing length to _regmap_select_page basing on the flag.
Signed-off-by: NDmitry Baryshkov <dmitry.baryshkov@linaro.org>
Fixes: cdf6b11d ("regmap: Add regmap_noinc_write API")
Link: https://lore.kernel.org/r/20200917153405.3139200-2-dmitry.baryshkov@linaro.orgSigned-off-by: NMark Brown <broonie@kernel.org>

Non-incrementing reads can fail if register + length crosses page
border. However for non-incrementing reads we should not check for page
border crossing. Fix this by passing additional flag to _regmap_raw_read
and passing length to _regmap_select_page basing on the flag.
Signed-off-by: NDmitry Baryshkov <dmitry.baryshkov@linaro.org>
Fixes: 74fe7b55 ("regmap: Add regmap_noinc_read API")
Link: https://lore.kernel.org/r/20200917153405.3139200-1-dmitry.baryshkov@linaro.orgSigned-off-by: NMark Brown <broonie@kernel.org>

Fixes: 94cc89eb ("regmap: debugfs: Fix handling of name string for debugfs init delays")
Reported-by: Nkernel test robot <lkp@intel.com>
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20200918112002.15216-1-ckeepax@opensource.cirrus.comSigned-off-by: NMark Brown <broonie@kernel.org>

In regmap_debugfs_init the initialisation of the debugfs is delayed
if the root node isn't ready yet. Most callers of regmap_debugfs_init
pass the name from the regmap_config, which is considered temporary
ie. may be unallocated after the regmap_init call returns. This leads
to a potential use after free, where config->name has been freed by
the time it is used in regmap_debugfs_initcall.

This situation can be seen on Zynq, where the architecture init_irq
callback registers a syscon device, using a local variable for the
regmap_config. As init_irq is very early in the platform bring up the
regmap debugfs root isn't ready yet. Although this doesn't crash it
does result in the debugfs entry not having the correct name.

Regmap already sets map->name from config->name on the regmap_init
path and the fact that a separate field is used to pass the name
to regmap_debugfs_init appears to be an artifact of the debugfs
name being added before the map name. As such this patch updates
regmap_debugfs_init to use map->name, which is already duplicated from
the config avoiding the issue.

This does however leave two lose ends, both regmap_attach_dev and
regmap_reinit_cache can be called after a regmap is registered and
would have had the effect of applying a new name to the debugfs
entries. In both of these cases it was chosen to update the map
name. In the case of regmap_attach_dev there are 3 users that
currently use this function to update the name, thus doing so avoids
changes for those users and it seems reasonable that attaching
a device would want to set the name of the map. In the case of
regmap_reinit_cache the primary use-case appears to be devices that
need some register access to identify the device (for example devices
in the same family) and then update the cache to match the exact
hardware. Whilst no users do currently update the name here, given the
use-case it seemed reasonable the name might want to be updated once
the device is better identified.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20200917120828.12987-1-ckeepax@opensource.cirrus.comSigned-off-by: NMark Brown <broonie@kernel.org>

Devices such as the AD5628 require 32 bits of data divided in 12 bits
for dummy, command and address, and 20 for data and dummy. Eg:

XXXXCCCCAAAADDDDDDDDDDDDDDDDXXXX

Where X is dont care, C is command, A is address and D is data bits.

Which would requierd the following regmap_config:

static const struct regmap_config config_dac = {
	.reg_bits = 12,
	.val_bits = 20,
	.max_register = 0xff,
};
Signed-off-by: NRicardo Ribalda <ribalda@kernel.org>
Link: https://lore.kernel.org/r/20200917114727.1120373-1-ribalda@kernel.orgSigned-off-by: NMark Brown <broonie@kernel.org>

Regmap can't sleep if spinlock is used for the locking protection.
This patch fixes regression caused by a previous commit that switched
regmap to use fsleep() and this broke Amlogic S922X platform.

This patch adds new configuration option for regmap users, allowing to
specify whether regmap operations can sleep and assuming that sleep is
allowed if mutex is used for the regmap locking protection.
Reported-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Fixes: 2b32d2f7 ("regmap: Use flexible sleep")
Signed-off-by: NDmitry Osipenko <digetx@gmail.com>
Tested-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Link: https://lore.kernel.org/r/20200902141843.6591-1-digetx@gmail.comSigned-off-by: NMark Brown <broonie@kernel.org>

The multi-reg write function uses udelay(), which is a busy-loop based
delaying function that is not suitable for a long delays. Hence let's
replace the udelay() with fsleep(), which is flexible sleep function that
selects best delay function based on the delay-time.
Signed-off-by: NDmitry Osipenko <digetx@gmail.com>
Link: https://lore.kernel.org/r/20200830185356.5365-3-digetx@gmail.comSigned-off-by: NMark Brown <broonie@kernel.org>

Make regmap firmware node type agnostic by switching it to use fwnode.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20200708161232.17914-1-andriy.shevchenko@linux.intel.comSigned-off-by: NMark Brown <broonie@kernel.org>

This function is used by dev_get_regmap() to retrieve a regmap for the
specified device. If the device has more than one regmap, the name parameter
can be used to specify one.

The code here uses a pointer comparison to check for equal strings. This
however will probably always fail, as the regmap->name is allocated via
kstrdup_const() from the regmap's config->name.

Fix this by using strcmp() instead.
Signed-off-by: NMarc Kleine-Budde <mkl@pengutronix.de>
Link: https://lore.kernel.org/r/20200703103315.267996-1-mkl@pengutronix.deSigned-off-by: NMark Brown <broonie@kernel.org>

When a register patch is registered the reg_sequence is copied but the
memory allocated is never freed. Add a kfree in regmap_exit to clean it
up.

Fixes: 22f0d90a ("regmap: Support register patch sets")
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20200617152129.19655-1-ckeepax@opensource.cirrus.comSigned-off-by: NMark Brown <broonie@kernel.org>

There are two spaces between arguments in regmap_fields_update_bits_base()
so remove one.
Signed-off-by: NBartosz Golaszewski <bgolaszewski@baylibre.com>
Link: https://lore.kernel.org/r/20200615072507.11303-1-brgl@bgdev.plSigned-off-by: NMark Brown <broonie@kernel.org>

The kerneldoc comment for regmap_test_bits() says that it returns -1 on
regmap_read() failure. This is not true - it will propagate the error
code returned by regmap_read(). Fix it.

Fixes: aa2ff9db ("regmap: provide helpers for simple bit operations")
Signed-off-by: NBartosz Golaszewski <bgolaszewski@baylibre.com>
Link: https://lore.kernel.org/r/20200607093421.22209-1-brgl@bgdev.plSigned-off-by: NMark Brown <broonie@kernel.org>

In many instances regmap_update_bits() is used for simple bit setting
and clearing. In these cases the last argument is redundant and we can
hide it with a static inline function.

This adds three new helpers for simple bit operations: set_bits,
clear_bits and test_bits (the last one defined as a regular function).
Signed-off-by: NBartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The assembly and disassembly of data to be sent to or received from
a device invoke functions regmap_format_XX() and regmap_parse_XX()
that extract or insert data items from or into a buffer, using
assignments. In some cases the functions are called with a buffer
pointer with an odd address. On architectures with strict alignment
requirements this can result in a kernel crash. The assignments
have been replaced by functions that take alignment into account.
Signed-off-by: NJens Thoms Toerring <jt@toerring.de>
Link: https://lore.kernel.org/r/20200531095300.GA27570@toerring.deSigned-off-by: NMark Brown <broonie@kernel.org>

In many instances regmap_update_bits() is used for simple bit setting
and clearing. In these cases the last argument is redundant and we can
hide it with a static inline function.

This adds three new helpers for simple bit operations: set_bits,
clear_bits and test_bits (the last one defined as a regular function).
Signed-off-by: NBartosz Golaszewski <bgolaszewski@baylibre.com>
Link: https://lore.kernel.org/r/20200528154503.26304-2-brgl@bgdev.plSigned-off-by: NMark Brown <broonie@kernel.org>

Add reg_update_bits() support in case some platforms use a special method
to update bits of registers.
Signed-off-by: NBaolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/df32fd0529957d1e7e26ba1465723f16cfbe92c8.1586757922.git.baolin.wang7@gmail.comSigned-off-by: NMark Brown <broonie@kernel.org>

When checking if a register block is writable we must ensure that the
block does not start with or contain a non incrementing register.

Fixes: 8b9f9d4d ("regmap: verify if register is writeable before writing operations")
Signed-off-by: NBen Whitten <ben.whitten@gmail.com>
Link: https://lore.kernel.org/r/20200118205625.14532-1-ben.whitten@gmail.comSigned-off-by: NMark Brown <broonie@kernel.org>

On buses like SlimBus and SoundWire which does not support
gather_writes yet in regmap, A bulk write on paged register
would be silently ignored after programming page.
This is because local variable 'ret' value in regmap_raw_write_impl()
gets reset to 0 once page register is written successfully and the
code below checks for 'ret' value to be -ENOTSUPP before linearising
the write buffer to send to bus->write().

Fix this by resetting the 'ret' value to -ENOTSUPP in cases where
gather_writes() is not supported or single register write is
not possible.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: NMark Brown <broonie@kernel.org>

There were a few files in the regmap code that did not have SPDX
identifiers on them, so fix that up.  At the same time, remove the "free
form" text that specified the license of the file, as that is impossible
for any tool to properly parse.

Also, as Mark loves // comment markers, convert all of the headers to be
the same to make things look consistent :)
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: NMark Brown <broonie@kernel.org>

regmap provides a couple of ways to validate the register range used.
a) maxim allowed register, b) writable/readable register tables,
c) callback function that can be provided by the driver to validate
a register. regmap framework should verify if registers
are writeable before every write operation. However this doesn't
seems to happen in every situation.

The method `_regmap_raw_write_impl` is only using the `writeable_reg`
callback to verify if register is writeable, ignoring the other two.
This can lead to undefined behaviour since this allows to write to
registers that could be declared un-writeable by using any other
option.

Change `_regmap_raw_write_impl` to use the `regmap_writeable` method
to verify if registers are writable before the write operation.
Signed-off-by: NNandor Han <nandor.han@vaisala.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Move the checking of the LOG_DEVICE into a function to reduce the
number of #ifdefs and  ensure more of the code gets compiled/checked,
and make it easier to change this for internal debugging purposes
(such as checking >1 device).
Signed-off-by: NBen Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: NMark Brown <broonie@kernel.org>

The regmap API had a noinc_read function added for instances where devices
supported returning data from an internal FIFO in a single read.

This commit adds the noinc_write variant to allow writing to a non
incrementing register, this is used in devices such as the sx1301 for
loading firmware.
Signed-off-by: NBen Whitten <ben.whitten@lairdtech.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Split regmap_config.use_single_rw into use_single_read and
use_single_write. This change enables drivers of devices which only
support bulk operations in one direction to use the regmap_bulk_*()
functions for both directions and have their bulk operation split into
single operations only when necessary.

Update all struct regmap_config instances where use_single_rw==true to
instead set both use_single_read and use_single_write. No attempt was
made to evaluate whether it is possible to set only one of
use_single_read or use_single_write.
Signed-off-by: NDavid Frey <dpfrey@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The regmap API usually assumes that bulk read operations will read a
range of registers but some I2C/SPI devices have certain registers for
which a such a read operation will return data from an internal FIFO
instead. Add an explicit API to support bulk read without range semantics.

Some linux drivers use regmap_bulk_read or regmap_raw_read for such
registers, for example mpu6050 or bmi150 from IIO. This only happens to
work because when caching is disabled a single regmap read op will map
to a single bus read op (as desired). This breaks if caching is enabled and
reg+1 happens to be a cacheable register.

Without regmap support refactoring a driver to enable regmap caching
requires separate I2C and SPI paths. This is exactly what regmap is
supposed to help avoid.
Suggested-by: NJonathan Cameron <jic23@kernel.org>
Signed-off-by: NCrestez Dan Leonard <leonard.crestez@intel.com>
Signed-off-by: NStefan Popa <stefan.popa@analog.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The handling for the first two cases in regmap_bulk_write is
essentially identical. The first case is just a better implementation of
the second, supporting 8 byte registers and doing the locking manually to
avoid bouncing the lock for each register. Drop some redundant code by
removing the second of these cases and allowing both situations to be
handled by the same code.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Raw writes may need to be split into small chunks if max_raw_write is
set. Tidy up the code implementing this, the new code is slightly
clearer, slightly shorter and slightly more efficient.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently regmap_bulk_write will split a write into chunks before
calling regmap_raw_write if max_raw_write is set. It is more logical
for this handling to be inside regmap_raw_write itself, as this
removes the need to keep re-implementing the chunking code, which
would be the same for all users of regmap_raw_write.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

In the case were the bulk transaction is split up into smaller chunks
data is passed directly to regmap_raw_write. However regmap_bulk_write
uses data in host endian and regmap_raw_write expects data in device
endian. As such if the host and device differ in endian the wrong data
will be written to the device. Correct this issue using a similar
approach to the single raw write case below it, duplicate the data
into a new buffer and use parse_inplace to format the data correctly.

Fixes: adaac459 ("regmap: Introduce max_raw_read/write for regmap_bulk_read/write")
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

This registers regmaps with debugfs even when they do not have an
associated device. For example, this is common for syscon regmaps.
Signed-off-by: NDavid Lechner <david@lechnology.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Bulk reads may potentially read a lot of registers and regmap_read will
take and release the regmap lock for each register. Avoid bouncing
the lock so frequently by holding the lock locally and calling
_regmap_read instead. This also has the nice side-effect that all the
reads will be done atomically so no other threads can sneak a write in
during the regmap_bulk_read.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Raw reads may need to be split into small chunks if max_raw_read is
set.  Tidy up the code implementing this, the new code is slightly
clearer, slightly shorter and slightly more efficient.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently regmap_bulk_read will split a read into chunks before
calling regmap_raw_read if max_raw_read is set. It is more logical for
this handling to be inside regmap_raw_read itself, as this removes the
need to keep re-implementing the chunking code, which would be the
same for all users of regmap_raw_read.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

As a helper function exists for calculating register offsets lets use
that rather than open coding with the reg_stride.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

A bulk read can be implemented either through regmap_raw_read, or
by reading each register individually using regmap_read.  Both
regmap_read and regmap_bulk_read should return values in native
endian. In the individual case the current implementation calls
format_val to put the data into the output array, which can cause
endian issues. The regmap_read will have already converted the data
into native endian, if the hosts endian differs from the device then
format_val will switch the endian back again.

Rather than using format_val simply use the code that is called if
there is no format_val function. This code supports all cases except
24-bit but there don't appear to be any users of regmap_bulk_read for
24-bit. Additionally, it would have to be a big endian host for the
old code to actually function correctly anyway.

Fixes: 15b8d2c4 ("regmap: Fix regmap_bulk_read in BE mode")
Reported-by: NDavid Rhodes <david.rhodes@cirrus.com>
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The cache pointer points to the actual memory used by the cache, as the
comparison here is looking for the type of the cache it should check
against cache_type.

Fixes: 1ea975cf ("regmap: Add a function to check if a regmap register is cached")
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The current implementation is broken for regmaps that have a reg_stride,
since it doesn't take the stride into account. Correct this by using the
helper function to calculate the register offset.

Fixes: f01ee60f ("regmap: implement register striding")
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>