If a driver is using regmap directly ensure that we're coherent with
non-ASoC register updates by using the regmap API directly to do our
read/modify/write cycles. This will bypass the ASoC cache but drivers
using regmap directly should not be using the ASoC cache.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

These aren't modules, but they do make use of these macros, so
they will need export.h to get that definition.  Previously,
they got it via the implicit module.h inclusion.
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>

The ASoC code always uses -1 as the error code due to reporting errors in
band with the value. Ensure we don't confuse anything by making sure we
don't pass actual error codes back into the rest of the code on read.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

[This will be used by the ad193x driver to fix the fact that the
original author of the driver put a bodge for their particular chip into
a the generic ASoC register I/O abstraction layer which looked like an
obvious bug which ended up getting fixed in 3.0.  Sadly there were no
comments documenting what was going on.  A minimally invasive correction
to the driver is to remove the register cache support and go direct to
the hardware all the time so we're adding a new feature -- broonie]
Signed-off-by: NScott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Since we changed regmap to be selected and register per bus rather than
via the core only we can't rely on it being enabled by the ASoC core.
Support compiling it out.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Reported-by: NAxel Lin <axel.lin@gmail.com>

When CONFIG_REGMAP_I2C/SPI are m, CONFIG_REGMAP_I2C_MODULE is set in the
pre-processor instead of CONFIG_REGMAP_I2C. This removes SND_SOC_I2C as a
valid option for snd_soc_codec_set_cache_io()'s control parameter, and
causes any ASoC regmap-using codec built as a module to fail to initialize.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

In the case of "make da8xx_omapl_defconfig;make", the SPI support is disabled.
Thus calling regmap_init_spi in soc-io.c has below build error.

ERROR: "regmap_init_spi" [sound/soc/snd-soc-core.ko] undefined!
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2

This patch fixes the build error by adding CONFIG_REGMAP_I2C/CONFIG_REGMAP_SPI
guards for regmap_init_i2c/regmap_init_spi.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Allow drivers to set up their own regmap API structures. This is mainly
useful with MFDs where the core driver will have set up regmap at the
minute, though it may make sense to push the existing regmap setup out
of the core into the drivers.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLiam Girdwood <lrg@ti.com>

Remove all the ASoC specific physical I/O code and replace it with calls
into the regmap API. The bulk write code can only be used safely if all
regmap calls are locked with the CODEC lock, we need to add bulk support
to the regmap API or replace the code with an open coded loop (though
currently it has no users...).
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLiam Girdwood <lrg@ti.com>

In order to facilitate merging with the register map I/O replace the use
of control_data for the bulk writes with direct lookup of the client data
from the device.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLiam Girdwood <lrg@ti.com>

For clarity and to help ongoing refactoring in this area create a new file
to contain the physical I/O functions, separating them out from the cache
operations.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLiam Girdwood <lrg@ti.com>