Replace the hardwired latency definition with platform data
parameter, and simplify the nSample parameter calculation.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Upper threshold is used in mode7 of DAC33.
Instead of hard wired UTHR, add control to change the upper threshold
value.
Changing upper threshold is not allowed when the playback is already
running, since wrongly timed change in the UTHR can cause problems
with the codec.
With this control the length of the burst in mode7 can be changed.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Avoid calling the dac33_hard_power when the codec was
already in BIAS_OFF state.
This could happen in device suspend and module removal
time.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Since the cases when the same power state would be set again
handled gracefully, we do not need to use dev_warn.
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
The core will ensure that the device is in either STANDBY or OFF bias
before suspending, restoring the bias in the driver is unneeded. Some
drivers doing slightly more roundabout things have been left alone
for now.
Tested-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Do not change the codec defaults for the following registers:
0x40, 0x41: Line output gains, do not use amplification
0x42: LOM/LOP Voltage hold, and selection
0x44: LOM inversion control
It has been found, that the values configured to these registers
can cause amplification, which can make the output of DAC33
distorted.
The codec reset values are considered safe in all environmnts.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Let the codec to hit OFF instead of STANDBY, when there is no activity.
When the codec is off, than the associated regulator can be also turned
off (if the number of users on the regulator is 0).
After initialization, the codec remains in power off, it is only turned
on for reading the ID registers (also testing the regulators).
The codec power is enabled, when the codec is moving from BIAS_OFF
to BIAS_STANDBY.
The codec is turned off, when it hits BIAS_OFF.
There are few scenarios, which has to be taken care::
1. Analog bypass caused BIAS_OFF -> BIAS_ON
We need to power on the codec, and do the chip init, but we does not
need to execute the playback related configuration
2. Playback caused BIAS_OFF -> BIAS_ON
We need to power on the codec, and do the chip init, and also we need
to execute the playback related configuration.
3. Playback start, while Analog bypass is on (BIAS_ON -> BIAS_ON)
We need to execute the playback related configuration. The codec is
already on.
4. Analog bypass enable, while playback (BIAS_ON -> BIAS_ON)
Nothing need to be done.
5. Playback start withing soc power down timeout (BIAS_ON -> BIAS_ON)
We need to execute the playback related configuration. The codec is
still on.
Since the power up, and the codec init is optimized, the added overhead
in stream start is minimal.
Withing this patch, the hard_power function is now only doing what it
supposed to: only handle the powers, and GPIO reset line.
The codec initialization and state restore has been moved out.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
As a preparation for supporting codec to be turned off,
when we are in BIAS_STANDBY.
The substream must be easily available in other places than
pcm_* callbacks.
Manage a pointer in _startup, and _shutdown for this.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Optimize the way how tlv320dac33 is powered uppon module and
soc initialization.
Also read the DAC33 ID registers, and update the reg_cache
to reflect it.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
On power up we only need to initialize the codec, and
restore only registers, which are not in either in DAPM
nor in the playback start sequence.
These are mostly gain related registers.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Acked-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Delay reporting for the three implemented DAC33 FIFO modes.
DAC33 has FIFO depth status register(s), but it can not be used, since
inside of pcm_pointer we can not send I2C commands.
Timestamp based estimation need to be used. The method of calculating
the delay depends on the active FIFO mode.
Bypass mode: FIFO is bypassed, report 0 as delay
Mode1: nSample fill mode. In this mode I need to use two timestamp
ts1: taken when the interrupt has been received
ts2: taken before writing to nSample register.
Interrupts are coming when DAC33 FIFO depth goes under alarm threshold.
Phase1: when we received the alarm threshold, but our workqueue has
not been executed (safeguard phase). Just count the played out
samples since ts1 and subtract it from the alarm threshold
value.
Phase2: During nSample burst (after writing to nSample register), count
the played out samples since ts1, count the samples received
since ts2 (in a burst). Estimate the FIFO depth using these and
alarm threshold value.
Phase3: Draining phase (after the burst read), count the played out
samples since ts1. Estimate the FIFO depth using the nSample
configuration and the alarm threshold value.
Mode7: Threshold based fill mode. In this mode one timestamp is enough.
ts1: taken when the interrupt has been received
Interrupts are coming when DAC33 FIFO depth reaches upper threshold.
Phase1: Draining phase (after the burst), counting the played out
samples since ts1, and subtract it from the upper threshold
value.
Phase2: During burst operation. Using the pre calculated time needed to
play out samples from the buffer during the drain period (from
upper to lower threshold), move the time window to cover the
estimated time from the burst start to the current time.
Calculate the samples played out since lower threshold and also
the samples received during the same time.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
When the DAC33 FIFO is in use the dai interface is running in
much higher speed than the sampling frequency.
Calculate the rate based on the internal base frequency and
the bclk divider.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Upper and Lower threshold values are used as magic
numbers. Replace them with defines for later use.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
There is no need for calculations for FIFO bypass mode.
Just in case set the nsample maximum limit, which
has been done in the calculation phase.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Alarm threshold interrupt is triggered right after the
playback start.
This interrupt is recieved during the first burst period,
and caused the state machine to write additional nSample
command, which has to be avoided.
To fix this issue move the DAC33 interrupt unmasking
after we configured the PREFILL register with a small
delay.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
One of the features of the multi CODEC work is that it embeds a struct
device in the CODEC to provide diagnostics via a sysfs class rather than
via the device tree, at which point it's much better to use the struct
device private data rather than having two places to store it. Provide
an accessor function to allow this change to be made more easily, and
update all the CODEC drivers are updated.
To ensure use of the accessor the private data structure member is
renamed, meaning that if code developed with older an older core that
still uses private_data is merged it will fail to build.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
During validation of the internal clocking setup it has
been found that the following settings were not configured
in an optimal way:
ASRC_CTRL_A: SRCLKDIV was incorrect, instad of divide ratio 3,
ratio of 2 has to be used (as the comment stated)
DAC_CTRL_A: Fs = Fsref is the desired configuration instead of
Fs = Fsref / 1.5
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
To make DSP_A mode working correctly the data delay should be
configured to 0. DSP_B mode thus can not be used with DAC33,
so remove it.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Platform data option for the codec to keep the BCLK clock
continuously running in FIFO modes (codec master).
OMAP3 McBSP when in slave mode needs continuous BCLK running
on the serial bus in order to operate correctly.
Since in FIFO mode the DAC33 can also shut down the BCLK clock
and enable it only when it is needed, let the platforms decide
if the CPU side needs the BCLK running or not.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
To avoid race condition especially in FIFO modes the
sequence for enabling and disabling the codec need to
be changed.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
OSCSET calculation was not correct in case of 44.1KHz
sampling rate.
With small adjustment both 48 and 44.1 KHz calculation
now gives the correct value.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
In repeated playback the FIFOFLUSH bit remained set, and
never has been cleared.
Clear it during the setup phase.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Add possibility to configure the burst mode BCLK divider through platform
data structure.
The BCLK divider changes the actual speed of the serial bus in burst mode,
which is faster than the sampling frequency of the running stream.
In this way platforms can experiment with the optimal burst speed without
the need to modify the codec driver itself.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The BCLK divider was not configured in case of mode7.
This leads to unpredictable behavior when switching between FIFO modes.
Configure the BCLK divider depending on the fifo_mode (FIFO is in use,
or FIFO bypass).
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Set the prefill number of samples as the same as the lower
threshold in mode7.
In this way the codec will read the same amount of data on
startup and during the running playback.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The currently available FIFO modes (mode1 and mode7) require master
mode from the codec.
Do not allow the slave configuration when the FIFO is in use.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Mode 7 of tlv320dac33 operates in the following way:
The codec is in master mode.
Host configures upper and lower thresholds in tlv320dac33
During playback the codec will clock in the data until the
upper threshold is reached in FIFO. At this point the codec
stops the colocks on the serial bus.
When the FIFO fill is reaching the lower threshold limit the
codec will enable the clocks on the serial bus, and clocks
in data till the upper threshold is reached.
In this mode, we can also request interrupts for threshold
events (upper, lower and alarm), which could be used for
power management.
At this point the interrupts are not enabled for this mode,
but it can be taken into use in the future, when the surrounding
code makes it possible to use it.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.oc.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Use switch instead of if statements to configure FIFO bypass
and mode1.
With this change adding new FIFO mode is going to be easier,
and cleaner.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Ensure that the code is going to be readable, when new FIFO modes
are introduced later.
Move the prefill and playback state handling to inlined
functions.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
In order to have support for more FIFO modes supported by
tlv320dac33, the switch for enabling/disabling the FIFO
use has to be replaced with an enum.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Take the regulator framework in use for managing the power sources.
Signed-off-by: Ilkka Koskinen <ilkka.koskinen@nokia.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
RT workqueue is going away in the near future, replace it with
singlethread wq for now, which is still supported.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The DAPM widgets are now insntantiated by the core when creating the card
so there is no need for the individual CODEC drivers to do so.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
snd_soc_init_card() is always called as the last part of the CODEC probe
function so we can factor it out into the core card setup rather than
have each CODEC replicate the code to do the initialiastation. This will
be required to support multiple CODECs per card.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Driver for Texas Instruments TLV320DAC33 (SLAS546) low power stereo
audio DAC.
TLV320DAC33 is a stereo audio codec with integrated 24KB FIFO for low
power audio playback.
The digital interface can use I2S, DSP (A or B), Right and Left
justified formats.
DAC33 has stereo analog input, which can be bypassed to the analog
outputs.
Regarding to the internal 24KB FIFO the driver implements 'FIFO bypass'
mode (default) and nSample mode (FIFO is in use).
a) In 'FIFO bypass' mode the internal FIFO is not in use, the codec is
working synchronously as a normal codec (it needs constant stream of
data on the digital interface).
b) The nSample mode implementation uses one interrupt line from DAC33 to
the host:
Alarm threshold is set to 10ms of audio data (limit by the driver
implementation).
DAC33 will signal an interrupt, when the FIFO level goes under the
Alarm threshold.
The host will write to nSample register a value (number of stereo
samples), to tell DAC33 how many samples it should read in a burst from
the host. When the DAC33 received the number of samples, it disables the
clocks on the I2S bus. When the FIFO use again goes under the Alarm
threshold, DAC33 signals the host with an interrupt, and the process is
repeated.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Peter Ujfalusi