Ensure that the device pins are configured as soon as possible by moving
the pin configration (including MICBIAS) into the I2C probe() function.
This had been done in the CODEC probe() function when we were relying on
the ASoC register I/O code.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Tested-by: Stephen Warren <swarren@wwwdotorg.org>
This allows the GPIOs to be available as soon as the I2C device has
probed, which in turn enables machine drivers to request the GPIOs in
their probe(), rather than deferring this to their ASoC machine init
function, i.e. after the whole sound card has been constructed, and
hence the WM8903 codec is available.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
In order to support CODEC<->CODEC links remove the assumption that there
is only a single CODEC on a DAI link by removing the use of the CODEC
pointer in the rtd from the CODEC drivers. They are already being passed
their DAI whenever they are passed an rtd and can get the CODEC from
there.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
We can't just pass back the return value of snd_soc_update_bits() as it
will be 1 if a bit changed rather than zero.
Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
This allows the device to be matched against the device tree using the
compatible flag directly, as is standard, rather than falling back to
matching .id_table against the non-vendor portion of the first compatible
property value.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
If we fail to read the IRQ type from the interrupt controller don't
fail, just assume a value and solider on - we may fail later when we try
to request the IRQ but it's possible we'll succeed.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Document the device tree binding for the WM8903 codec, and modify the
driver to extract platform data from the device tree, if present.
Based on work by John Bonesio, but significantly reworked since then.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
If the WM8903 is hooked up to an interrupt, set the irq_active_low flag
in the default platform data based on the IRQ's IRQ_TYPE. Map IRQ_TYPE_NONE
(a lack of explicit configuration/restriction) to irq_active_low = false;
the previous default.
This code is mainly added to support device tree interrupt bindings,
although will work perfectly well in a non device tree system too.
Any interrupt controller that supports only a single IRQ_TYPE could
set each IRQ's type based on that restriction. This applies equally
with and without device tree. To cater for interrupt controllers
that don't do this, for which irqd_get_trigger_type() will return
IRQ_TYPE_NONE, the platform data irq_active_low field may be used
in systems that don't use device tree.
With device tree, every IRQ must have some IRQ_TYPE set.
Controllers that support DT and multiple IRQ_TYPEs must define the
interrupts property (as used in interrupt source nodes) such that it
defines the IRQ_TYPE to use. When the core DT setup code initializes
wm8903->irq, the interrupts property will be parsed, and as a side-
effect, set the IRQ's IRQ_TYPE for the WM8903 probe() function to read.
Controllers that support DT and a single IRQ_TYPE could arrange to
set the IRQ_TYPE somehow during their initialization, or hard-code
it during the processing of the child interrupts property.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The pdata pointer is now always valid. Remove any conditions that check
its validity.
This patch is mostly just removing an indentation level. One variable had
to be moved due to the removal of a scope, and one comment was split into
two. Viewing the patch with git show/diff -b will show that it's actually
very small.
Note that WM8903_MIC_BIAS_CONTROL_0 is now written unconditionally,
whereas it used to be written only if pdata was supplied. Since
defpdata.micdet_cfg = 0, this unconditional write simply echos the HW
defaults in the case where pdata is not supplied.
Based on work by John Bonesio, but significantly reworked since then.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
wm8903_platform_data.gpio_cfg[] was intended to be interpreted as follows:
0: Don't touch this GPIO's configuration register
1..7fff: Write that value to the GPIO's configuration register
8000: Write zero to the GPIO's configuration register
other: Undefined (invalid)
The rationale is that platform data is usually global data, and a value of
zero means that the field wasn't explicitly set to anything (e.g. because
the field was new to the pdata type, and existing users weren't update to
initialize it) and hence the value zero should be ignored. 0x8000 is an
explicit way to get 0 in the register.
The code worked this way until commit 7cfe561 "ASoC: wm8903: Expose GPIOs
through gpiolib", where the behaviour was changed due to my lack of
awareness of the above rationale.
This patch reverts to the intended behaviour, and updates all in-tree users
to use the correct scheme. This also makes WM8903 consistent with other
devices that use a similar scheme.
WM8903_GPIO_NO_CONFIG is also renamed to WM8903_GPIO_CONFIG_ZERO so that
its name accurately reflects its purpose.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Cc: Olof Johansson <olof@lixom.net>
Cc: Colin Cross <ccross@android.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
When no platform data is supplied, point pdata at a default platform
structure. This enables two future changes:
a) Defines the default platform data values in a single place.
b) There is always a valid pdata pointer, so some conditional code can
be simplified by a later patch.
Based on work by John Bonesio, but significantly reworked since then.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Get control of the device earlier and avoid trying to do an ASoC probe
on a card that won't work.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Converting to an rbtree cache as regcache doesn't have a flat cache.
Since the top of the register map is fairly sparse this should be an
overall win.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
We only do this on initial power on so it's at best a waste of time as
the core will have already defaulted to the same values.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
The driver used to use a complicated method to sync the register cache
after having brought the bias level up to standby in resume due to the
use of the write sequencer to manage the initial power up. Now that we
don't use the write sequencer there is no need for this and we can just
use snd_soc_cache_sync() directly.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
The GPIO registers are 15 bits wide. Hence values, higher than 0x7fff are
not legal GPIO register values. Modify the pdata.gpio_cfg handling code
to reject all illegal values, not just WM8903_GPIO_NO_CONFIG (0x8000). This
will allow the later use of 0xffffffff as an invalid value in future device
tree bindings, meaning "don't touch this GPIO's configuration".
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The existence of this parameter is purely historical. None of the CODEC drivers
uses it and we always pass in the same value anyway, so it should be safe to
remove it.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Also rename it to MICBIAS to reflect the pin name and help any out of tree
users notice the change.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Commit 1ee46ebd("ASoC: Make the DAI ops constant in the DAI structure")
introduced the possibility to have constant DAI ops structures, yet this is
barley used in both existing drivers and also new drivers being submitted,
although none of them modifies its DAI ops structure. The later is not
surprising since existing drivers are often used as templates for new drivers.
So this patch just constifies all existing snd_soc_dai_ops structs to eliminate
the issue altogether.
The patch was generated with the following coccinelle semantic patch:
// <smpl>
@@
identifier ops;
@@
-struct snd_soc_dai_ops ops =
+const struct snd_soc_dai_ops ops =
{ ... };
// </smpl>
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Without this, request_irq on subsequent device initialization fails, and
the codec cannot be used.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Commit fafd217 ("ASoC: Store a list of widgets in a DAPM mux/mixer kcontrol")
changed the control private data type that is passed to snd_soc_cnew when
creating dapm mixer and mux controls. Commit did not update a few codec
drivers that are using their own put callbacks and thus are accessing a
wrong data type.
Signed-off-by: Jarkko Nikula <jhnikula@gmail.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Increase the range of the Digital Capture Volume control to be 120 steps.
Each step is 0.75dB, and the range starts at -72dB, giving a max setting
of 18dB, which matches the latest datasheet, to the precision of the step
size.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
In addition to the currently supported analog capture path, the WM8903
also supports digital mics.
The analog and digital capture paths are exclusive; a mux is present to
select the capture source.
Logically, the mux exists to select the decimator's input, from either
the ADC or DMIC block outputs. However, the ADC power domain also
includes the DMIC interface. Consequently, this change represents the
mux as existing immediately before the ADC, and selecting between the
Input PGA and DMIC block outputs.
An alternative might be to represent the mux in its correct location,
and associate the ADC power enable controls with both the real ADC, and
a fake ADC for the DMIC?
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Update the headphone and line out mixers and PGAs use the same logical
set of register bits and sequencing as the speaker mixer/PGA.
This allows ALSA controls for mute and volume on headphone and line out
to operate correctly.
Per conversation on alsa-devel, earlier datasheets indicated that the
POWER_MANAGEMENT_* register bits 0 and 1 were aliases to ANALOG_* register
bits 0 and 4, and hence only one copy of those bits was programmed.
However, later datasheets corrected this.
From: Dilan Lee <dilee@nvidia.com>
[swarren: Applied same change to headphone widgets]
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The WM8903 interrupts are clear on read so if the WM8903 detection is
enabled from platform data when the IRQ is in use (rather than using a
direct signal from a GPIO) status may be lost during startup. Help users
spot this misconfiguration by adding a WARN_ON().
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
The mic detection HW should be enabled when either mic or short detection
is required, not when only both are required.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@kernel.org
This makes no real difference compared to the write sequencer sequence
that was previously used but can run without a clock being provided.
Also remove the write sequencer support code as this was the last use
of it.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
The write sequencer sequencer sequence takes longer than is desirable
as it brings up a full playback path which is not required at this
point. Open coding the sequence cuts the startup time by two thirds.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Explicitly cache the DC servo offsets for digital paths in the driver,
allowing them to be preserved over suspend and resume, and ensure that
we recalibrate analogue outputs paths when they are in use so that we
cover any changes in the input offset.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
The WM8903 register map does not mute the DAC by default at startup
so we need to explicitly do so.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
This simplfies the code and slightly reduces the startup time.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
It causes noisy -codecs to appear in things like .codec_name.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Also, update platform_data GPIO handling to have an explicit "don't
touch this pin" option.
Add #defines for the GPIO pin functions.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The multi-component patch(commit f0fba2ad1) moved the allocation of the
register cache from the driver to the ASoC core. Most drivers where adjusted to
this, but there are quite a few drivers left which now have an unused reg_cache field in
their private device struct.
This patch removes these unused fields.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
