DAPM now has a DAPM-level lock which it manages itself so we don't need
to take the CODEC mutex to call DAPM any more.  Also remove a redundant
call to snd_soc_dapm_sync(), jack reporting also triggers a DAPM sync.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Based on latest production information.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@vger.kernel.org

The optimal management of VMID depends on a number of factors which vary
dynamically at runtime, for example the connection to a system docking
station. In some circumstances it is desirable to keep VMID enabled all
the time, in others it is desirable to aggressively power it up and down.

Provide a callback allowing machine driver to configure either the normal
power up/down mode (WM8994_VMID_MODE_NORMAL) or to maintain VMID even
when idle (WM8994_VMID_MODE_FORCE). This callback, wm8994_vmid_mode(),
should be called with the CODEC lock.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

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

Currently we can the accdet mutex from within DAPM when updating the device
state which means we take accdet then the CODEC mutex but we also do the
locking the other way around when responding to the jackdet IRQ. Move all
the jackdet use of the CODEC mutex out of the accdet lock to avoid this.
Since all the DAPM interactions depend only on a single threaded IRQ this
is still serialised.

The locking improvements in 3.5 allow a better solution there.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

When we are out of system sleep always use audio mode for jack detection
in order to avoid potential performance issues handing off between modes.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Push everything through one function for active use cases, should be
no practical effect.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

If we're not doing jackdet it's not needed.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

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

Normally this will have no effect as we set detection up at system startup
before DAPM syncs take effect, this will only be useful if the system
enables and disables detection at runtime.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The WM1811 DC servo is able to run much faster than previous devices so
the benefit of skipping calibration is not useful.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Using snd_soc_write() means we always write to the register even if it
already contains the newly calculated value.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

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

Don't debounce jack detection for inserted jacks, giving improved
responsiveness.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

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

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

Early revisions of several of the WM8994 variants have register updates
to improve performance. Move these over to using the regmap patch system
instead of open coding them in the audio driver. Since the regmap init
is done by the MFD the code is moved there.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NSamuel Ortiz <sameo@linux.intel.com>

Allows the generic code to set up for that.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

When an external capacitor is connected to MICBIAS2 on devices with
jack detection (which is not required but may be done in some systems)
then the loading may mean that better performance is obtained when
the microphone bias is enabled normally rather than using the low power
mode. Provide platform data allowing systems to indicate if they require
this.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This minimises the chance of any external capacitors that are fitted
being discharged into headphones as they insert.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The latest recommendation for optimal performance.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLiam Girdwood <lrg@ti.com>

Improves performance on power up.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Ensure we're in a known state when we restart.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Raise the ramp time to 50ms to cover corner cases, use the startup bias
generator, explicitly reset the ramp circuit when complete and reorder
things all of which should improve performance somewhat for systems that
are sensitive to noise from VMID.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The enable of the single ended line outputs on wm_hubs devices performs
better if the output is enabled prior to starting VMID. Since inactive
outputs are held at VMID anyway there is little cost to doing this for
unused outputs. Add callbacks into wm_hubs and keep track of which outputs
are really active so we can disable them once we're active.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

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

The VMID ramp rate is supposed to be 0x3, not 11b. Fix that.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@vger.kernel.org

Make the mechanism used for WM8994 more like that for WM1811 and WM8958:
provide the logic to distinguish between headphone and headset and hard
code the reporting of sensible SND_JACK values. Should integration with
other detection mechanisms be required we can add appropriate callbacks
(though some integrations should be able to use the subsystem ones).
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Otherwise we might get an error if the GPIO is configured as an input
since that makes the register volatile and a suspended device can't be
read from.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This is the default state that the runtime PM infrastructure expects so
instead just kick the runtime PM core to suspend us if we're not doing
anything (as is default).
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

In cases where we should enter STANDBY DAPM will power us up, otherwise
there is no need to power up and we can remain in OFF.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

We can enable VMID independently of the bias in some use cases so we need
to ensure that the core device is powered up.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@vger.kernel.org

Currently ASoC can only add kcontrols using codec and platform component device
handles. It's also desirable to add kcontrols for DAIs (i.e. McBSP) and for
SoC card machine drivers too. This allows the kcontrol to have a direct handle to
the parent ASoC component DAI/SoC Card/Platform/Codec device and hence easily
get it's private data.

This change makes snd_soc_add_controls() static and wraps it in the folowing
calls (card and dai are new) :-

snd_soc_add_card_controls()
snd_soc_add_codec_controls()
snd_soc_add_dai_controls()
snd_soc_add_platform_controls()

This patch also does a lot of small mechanical changes in individual codec drivers
to replace snd_soc_add_controls() with snd_soc_add_codec_controls().

It also updates the McBSP DAI driver to use snd_soc_add_dai_controls().

Finally, it updates the existing machine drivers that register controls to either :-

1) Use snd_soc_add_card_controls() where no direct codec control is required.
2) Use snd_soc_add_codec_controls() where there is direct codec control.

In the case of 1) above we also update the machine drivers to get the correct
component data pointers from the kcontrol (rather than getting the machine pointer
via the codec pointer).
Signed-off-by: NLiam Girdwood <lrg@ti.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

For later wm_hubs devices we have much less need to keep the biases up
even when using single ended line outputs so flag idle_bias_off for
everything except the WM8993 and WM8994.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Now we've switched over to regmap the ASoC level cache sync will be
ineffectual and potentially harmful as there is no longer an ASoC level
cache.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Report any accessory with a low impedance as a headphone, previously
anything other than a short or microphone would not be reported at all.
The most likely reason is a microphone with incorrect polarity.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

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

It's still useful to be able to poke around in the register map at
runtime.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

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