The parameter passed to the regmap lock/unlock callbacks needs to be
map->lock_arg, regcache passes just map. This works fine in the case that no
custom locking callbacks are used, since in this case map->lock_arg equals map,
but will break when custom locking callbacks are used. The issue was introduced
in commit 0d4529c5 ("regmap: make lock/unlock functions customizable") and is
fixed by this patch.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

If range information has been provided then when we allocate a rbnode
within a range allocate the entire range. The goal is to minimise the
number of reallocations done when combining or extending blocks. At
present only readability and yes_ranges are taken into account, this is
expected to cover most cases efficiently.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

In preparation for being slightly smarter about how we allocate memory
factor out the node allocation.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Allow drivers to use an access table as part of their implementation.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Allow drivers to discard parts of the register cache, for example if part
of the hardware has been reset.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

_regmap_raw_write() contains code to call regcache_write() to write
values to the cache. That code calls memcpy() to copy the value data to
the start of the work_buf. However, at least when _regmap_raw_write() is
called from _regmap_bus_raw_write(), the value data is in the work_buf,
and this memcpy() operation may over-write part of that value data,
depending on the value of reg_bytes + pad_bytes. At least when using
reg_bytes==1 and pad_bytes==0, corruption of the value data does occur.

To solve this, remove the memcpy() operation, and modify the subsequent
.parse_val() call to parse the original value buffer directly.

At least in the case of 8-bit register address and 16-bit values, and
writes of single registers at a time, this memcpy-then-parse combination
used to cancel each-other out; for a work-buffer containing xx 89 03,
the memcpy changed it to 89 03 03, and the parse_val changed it back to
89 89 03, thus leaving the value uncorrupted. This appears completely
accidental though. Since commit 8a819ff8 "regmap: core: Split out in
place value parsing", .parse_val only returns the parsed value, and does
not modify the buffer, and hence does not (accidentally) undo the
corruption caused by memcpy(). This caused bogus values to get written
to HW, thus preventing e.g. audio playback on systems with a WM8903
CODEC. This patch fixes that.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This reverts commit bc8ce4 (regmap: don't corrupt work buffer in
_regmap_raw_write()) since it turns out that it can cause issues when
taken in isolation from the other changes in -next that lead to its
discovery.  On the basis that nobody noticed the problems for quite some
time without that subsequent work let's drop it from v3.9.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Fix format specifier in dev_dbg and suppress the following warning

drivers/base/regmap/regcache.c: In function
‘regcache_sync_block_raw_flush’:
drivers/base/regmap/regcache.c:593:2: warning: format ‘%d’ expects
argument of type ‘int’, but argument 4 has type ‘size_t’ [-Wformat]
Signed-off-by: NStratos Karafotis <stratosk@semaphore.gr>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

regcache_sync_block_raw is used only in this file. Hence make it static.
Silences the following warning:
drivers/base/regmap/regcache.c:608:5: warning:
symbol 'regcache_sync_block_raw' was not declared. Should it be static?
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

When syncing blocks of data using raw writes combine the writes into a
single block write, saving us bus overhead for setup, addressing and
teardown.

Currently the block write is done unconditionally as it is expected that
hardware which has a register format which can support raw writes will
support auto incrementing writes, this decision may need to be revised in
future.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Reviewed-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>

For code clarity after implementing block writes split out the raw and
non-raw I/O sync implementations.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Reviewed-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>

The idea of holding blocks of registers in device format is shared between
at least rbtree and lzo cache formats so split out the loop that does the
sync from the rbtree code so optimisations on it can be reused.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Reviewed-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>

The idea of maintaining a bitmap of present registers is something that
can usefully be used by other cache types that maintain blocks of cached
registers so move the code out of the rbtree cache and into the generic
regcache code.

Refactor the interface slightly as we go to wrap the set bit and enlarge
bitmap operations (since we never do one without the other) and make it
more robust for reads of uncached registers by bounds checking before we
look at the bitmap.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Reviewed-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>

Let's only write once...
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This will bring no meaningful benefit by itself, it is done as a separate
commit to aid bisection if there are problems with the following commits
adding support for coalescing adjacent writes.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Mainly useful internally but exported since this is a public API that's
being checked for.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Provide a helper to do the size based index into a block of registers and
use it when reading a value.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This patch aims to bring down the average number of nodes
in the rbtree cache and increase the average number of registers
per node.  This should improve general lookup and traversal times.
This is achieved by setting the minimum size of a block within the
rbnode to the size of the rbnode itself.  This will essentially
cache possibly non-existent registers so to combat this scenario,
we keep a separate bitmap in memory which keeps track of which register
exists.  The memory overhead of this change is likely in the order of
~5-10%, possibly less depending on the register file layout.  On my test
system with a bitmap of ~4300 bits and a relatively sparse register
layout, the memory requirements for the entire cache did not increase
(the cutting down of nodes which was about 50% of the original number
compensated the situation).

A second patch that can be built on top of this can look at the
ratio `sizeof(*rbnode) / map->cache_word_size' in order to suitably
adjust the block length of each block.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This allows the cache to sync values directly to the device when stored
in native format and also allows asynchronous I/O.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Don't grind to a screaming halt, just generate a warning.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

_regmap_raw_write() contains code to call regcache_write() to write
values to the cache. That code calls memcpy() to copy the value data to
the start of the work_buf. However, at least when _regmap_raw_write() is
called from _regmap_bus_raw_write(), the value data is in the work_buf,
and this memcpy() operation may over-write part of that value data,
depending on the value of reg_bytes + pad_bytes. At least when using
reg_bytes==1 and pad_bytes==0, corruption of the value data does occur.

To solve this, remove the memcpy() operation, and modify the subsequent
.parse_val() call to parse the original value buffer directly.

At least in the case of 8-bit register address and 16-bit values, and
writes of single registers at a time, this memcpy-then-parse combination
used to cancel each-other out; for a work-buffer containing xx 89 03,
the memcpy changed it to 89 03 03, and the parse_val changed it back to
89 89 03, thus leaving the value uncorrupted. This appears completely
accidental though. Since commit 8a819ff8 "regmap: core: Split out in
place value parsing", .parse_val only returns the parsed value, and does
not modify the buffer, and hence does not (accidentally) undo the
corruption caused by memcpy(). This caused bogus values to get written
to HW, thus preventing e.g. audio playback on systems with a WM8903
CODEC. This patch fixes that.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Display the name for the chip rather than just the primary IRQ so it is
clearer what exactly has failed.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The last register block, which falls into the specified range, is not handled
correctly. The formula which calculates the number of register which should be
synced is inverse (and off by one). E.g. if all registers in that block should
be synced only one is synced, and if only one should be synced all (but one) are
synced. To calculate the number of registers that need to be synced we need to
subtract the number of the first register in the block from the max register
number and add one. This patch updates the code accordingly.

The issue was introduced in commit ac8d91c8 ("regmap: Supply ranges to the sync
operations").
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@vger.kernel.org

Provide a feel of how much overhead the rbtree cache adds to
the game.

[Slightly reworded output in debugfs -- broonie]
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

In the rbtree code we are exposing statistics relating to the
number of nodes/registers of the rbtree cache for each of the
devices.  Ensure that `map->debugfs' has been initialized before
we attempt to initialize the debugfs entry for the rbtree cache.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@vger.kernel.org

This file lists the register ranges in the register map.  The condition
to split the range is based on whether the block is readable or not.

Ensure that we lock the `debugfs_off_cache' list whenever we access
and modify the list.  There is a possible race otherwise between the
read() operations of the `registers' file and the `range' file.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

We don't need to use any of the file position information
to calculate the base and max register of each block.  Just
use the counter directly.

Set `i = base' at the top to avoid GCC flow analysis bugs.  The
value of `i' can never be undefined or 0 in the if (c) { ... }.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This allows the cached data to be sent directly to the device when
we sync it.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Currently the value parsing operations both return the parsed value and
modify the passed buffer. This precludes their use in places like the cache
code so split out the in place modification into a new parse_inplace()
operation.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Factor things out a little.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

It's more idiomatic to pass the map structure around and this means we
can use other bits of information from the map.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

If we're updating a value in place it's more work to read the value and
compare the value with what we're about to set than it is to just write
the value into the cache; there are no further operations after writing
in the code even though there's an early return here.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Trace when we start and complete async writes, and when we start and
finish blocking for their completion. This is useful for performance
analysis of the resulting I/O patterns.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Even in failed case of pm_runtime_get_sync, the usage_count
is incremented. In order to keep the usage_count with correct
value and runtime power management to behave correctly, call
pm_runtime_put(_sync) in such case.

Signed-off-by Liu Chuansheng <chuansheng.liu@intel.com>
Signed-off-by: NLi Fei <fei.li@intel.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Some mmio devices have a dedicated interface clock that needs
to be enabled to access their registers. This patch optionally
enables a clock before accessing registers in the regmap_bus
callbacks.

I added (devm_)regmap_init_mmio_clk variants of the init
functions that have an added clk_id string parameter. This
is passed to clk_get to request the clock from the clk
framework.
Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

In preparation to support the regmap debugfs ranges functionality
factor this code out to a separate function.  We'll need to ensure
that the value has been correctly calculated from two separate places
in the code.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Optimize this so that we can better guess where to start scanning
from.  We know the length of the register field format, therefore
given the file pointer position align to the nearest register
field and scan from there onwards.  We round down in this calculation
and we let the rest of the code figure out where to start scanning
from.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

We are keeping track of the maximum register as well, this will make
things easier for us in sharing this code with the code implementing
the register ranges functionality.  It also simplifies a bit the
calculations when looking for the relevant block:offset from within
the cache.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

At the moment, if the length of the register field format is
N bytes, we can only get anything meaningful back to userspace
by providing a buffer that is N + 2 bytes large.  Fix this
so we that we only need to provide a buffer of N bytes.
Signed-off-by: NDimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>