Age | Commit message (Collapse) | Author |
|
Create a struct hdmi_priv to store the relevant data of the CPU DAI
driver. As more data is added to the driver, having all the data
in the same location eases its handling. At the moment, only the DMA
configuration parameters are included in the structure.
Also, the required memory is allocated using devm_kzalloc rather than
using a static global variable.
Signed-off-by: Ricardo Neri <ricardo.neri@ti.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
When getting the needed resources fails, return -ENODEV. This is more
in line with other drivers do and it gives a more descriptive error.
Signed-off-by: Ricardo Neri <ricardo.neri@ti.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
In order to utilize the new OMAP HDMI codec and the updated name of
the device of the CPU DAI, update the names at the drivers accordingly.
While there, also update the name of the machine driver to be more
generic and encompass more OMAP processors featuring HDMI and not
only OMAP4.
Signed-off-by: Ricardo Neri <ricardo.neri@ti.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Introduce codec for HDMI. At the moment, this is a dummy codec. In the
future it will parse the EDID to modify the supported parameters, such
as the number of channels and the sample rates. At the moment, it blindly
supports all the sample rates and audio channels described in the HDMI
1.4a specification.
Signed-off-by: Ricardo Neri <ricardo.neri@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
|
|
Headset Jack detect pin need to be added for all the board in order
to get audio record working on PandaBoard. Without this patch the
BIAS of the microphone or Line IN is not set.
Signed-off-by: Sebastien Guirec <s-guiriec@ti.com>
|
|
After calling pm_runtime_put_sync(), the ABE can lose its clocks,
which clear the registers. When pm_runtime_get_sync() is called, the
clocks are restored and the registers are set to their default values.
Since auto-gating is off by default, it needs to be reset every time
the clocks are restored
This patch does this by enforcing the use of the internal
abe_dsp_pm_get() and abe_dsp_pm_put() instead of the raw pm_*
functions. After calling pm_runtime_get_sync(), the clocks are
restored by a new ABE function abe_set_auto_gating().
There are two exceptions to using the abe_dsp_pm_*() functions:
1. In the abe_probe() method, auto-gating is restored when the
firmware is loaded. In addition, the registers can not be
referenced immediately after calling pm_runtime_get_sync().
Therefore, the calls are handled separately.
2. In abe_irq_handler() pm_runtime_put_sync_suspend() is used
instead of pm_runtime_put_sync(). It is left in tact.
Signed-off-by: Gabriel M. Beddingfield <gabrbedd@ti.com>
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Enable OPP scaling for audio domain (device_scale callback).
However, keep in mind that IVA domain OPP50 is deactivated
(so ABE OPP50 currently ties to IVA OPP100).
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
on snd_soc_omap and snd_soc_mcpdm
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
|
|
While there're comments in the code about dinamically loading the firmware,
the blob is always statically #included[1], and we don't have the firmware
as an external file: reuse the fw array.
[1] = see sound/soc/omap/abe/abe_ini.c::aess_firmware_array
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
|
|
Signed-off-by: Sebastien Guiriec <s-guiriec@ti.com>
|
|
Pandaboard does not have DMICs so don't add any routes.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Implementation of EBE integration with the DMIC driver.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
ASoC: omap-dmic: ABE support
Implementation of ABE integration with the DMIC driver.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
|
|
Update generic code for ABE port in order to enable/disable mono
mixer configuration
Signed-off-by: Sebastien Guiriec <s-guiriec@ti.com>
|
|
Make sure we connect the DMIC DAIs to the ABE
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Add stream names for ABE DAIs stream DAI matching.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
When unmuting the "Tones" gains, only the DL1 gain was being un-muted.
This was obviously a typo.
Signed-off-by: Gabriel M. Beddingfield <gabrbedd@ti.com>
|
|
Remove build warning form ABE port initialization
Signed-off-by: Sebastien Guiriec <s-guiriec@ti.com>
|
|
Signed-off-by: Sebastien Guiriec <s-guiriec@ti.com>
|
|
Make sure we set the bias level for all DAPM contexts when
changing level.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Otherwise dmic.2 moans about invalid dma memory from hwmod
Signed-off-by: Andy Green <andy.green@linaro.org>
|
|
Signed-off-by: Andy Green <andy.green@linaro.org>
|
|
Signed-off-by: Andy Green <andy.green@linaro.org>
|
|
Signed-off-by: Andy Green <andy.green@linaro.org>
|
|
Signed-off-by: Andy Green <andy.green@linaro.org>
|
|
Audio Backend requires period time in steps of 250us.
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
|
|
Export 'snd_interval_step' symbol so that ALSA drivers can use it
to add hw rules.
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
|
|
Recent DAI widget updates mean we have to make sure that the DAIs returned
in the valid audio path are all BE type (and not FE).
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
In order to get the legacy mode working (to DAPM power on the path) we need
to connect the legacy stream name to the DACs of the codec.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
|
|
None of the machines uses the gain ramp possibility for HS/HF.
This code path is mostly unused, but it can cause race conditions
in certain scenarios.
Remove the gain ramp, and related features form the driver.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
|
|
Modem is handled by the McBSP DAI and not the voice port.
To be squshed for upstream.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
To be squshed for upstreaming.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Sometimes more than 1 widget per DAI can be created when the
DAI is embedded with a codec or platform component.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
To be squashed for upstreaming.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Check that the DAI can support the stream direction and make sure
we dont try and deref a NULL widget.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Make sure we hold the DAPM mutex for DPCM based stream events.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Guarantee we hold the DAPM mutex for all PCM DAPM stream ops
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
To be squahed into DPCM upstream series.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
If statement not needed after recent updates. To be squashed.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
We should use the capture widget for capture streams. To be squashed.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Bring this up a level to additionally protect the paths and
status. To be squashed for upstream.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Introduced in the 3.5 mutex updates, to be squashed for upstream.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Somehow this was dropped from recent mutex work and will be squashed
into upstream submission.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Deferred probe can cause this to happen more than once. Move to machine init
to fix.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Use the correct DAPM context for pins that ignore suspend.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
Looks like either a merge issue or a missing patch here.
TODO: fixup and squash before upstreaming
Signed-off-by: Liam Girdwood <lrg@ti.com>
|
|
ABE has certain requirements towards the FIFO configuration in McBSP.
Make it posssible to configure the McBSP FIFO threshold from outside of
the McBSP driver stack.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
|
|
We need only one DAI for legacy mode, which provides
the playback, and capture capability.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
|
|
To be squshed prior to release.
Signed-off-by: Liam Girdwood <lrg@ti.com>
|