An earlier bug fix of mine made the SND_DM365_VOICE_CODEC symbol
tristate to avoid creating an undefined reference from the
davinci-vcif.c driver to the davinci_soc_platform_register
function that may be in a module.
However, this may now lead to a different error on randconfig
kernels:
"warning: SND_DM365_VOICE_CODEC creates inconsistent choice state"
This happens because we now have a choice statement with
one bool and one tristate option, and the latter might not
support being set to 'y' because of dependencies.
This new change turns the other option into 'tristate' as well,
which avoids the problem.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 19926c6de0 ("ASoC: davinci: vcif must be a module if SND_DAVINCI_SOC is")
Signed-off-by: Mark Brown <broonie@kernel.org>
Some SoC, like da850/OMAP-L138 uses one common interrupt request for TX/RX
events.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Here's the set of driver core patches for 3.19-rc1.
They are dominated by the removal of the .owner field in platform
drivers. They touch a lot of files, but they are "simple" changes, just
removing a line in a structure.
Other than that, a few minor driver core and debugfs changes. There are
some ath9k patches coming in through this tree that have been acked by
the wireless maintainers as they relied on the debugfs changes.
Everything has been in linux-next for a while.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iEYEABECAAYFAlSOD20ACgkQMUfUDdst+ylLPACg2QrW1oHhdTMT9WI8jihlHVRM
53kAoLeteByQ3iVwWurwwseRPiWa8+MI
=OVRS
-----END PGP SIGNATURE-----
Merge tag 'driver-core-3.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core update from Greg KH:
"Here's the set of driver core patches for 3.19-rc1.
They are dominated by the removal of the .owner field in platform
drivers. They touch a lot of files, but they are "simple" changes,
just removing a line in a structure.
Other than that, a few minor driver core and debugfs changes. There
are some ath9k patches coming in through this tree that have been
acked by the wireless maintainers as they relied on the debugfs
changes.
Everything has been in linux-next for a while"
* tag 'driver-core-3.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (324 commits)
Revert "ath: ath9k: use debugfs_create_devm_seqfile() helper for seq_file entries"
fs: debugfs: add forward declaration for struct device type
firmware class: Deletion of an unnecessary check before the function call "vunmap"
firmware loader: fix hung task warning dump
devcoredump: provide a one-way disable function
device: Add dev_<level>_once variants
ath: ath9k: use debugfs_create_devm_seqfile() helper for seq_file entries
ath: use seq_file api for ath9k debugfs files
debugfs: add helper function to create device related seq_file
drivers/base: cacheinfo: remove noisy error boot message
Revert "core: platform: add warning if driver has no owner"
drivers: base: support cpu cache information interface to userspace via sysfs
drivers: base: add cpu_device_create to support per-cpu devices
topology: replace custom attribute macros with standard DEVICE_ATTR*
cpumask: factor out show_cpumap into separate helper function
driver core: Fix unbalanced device reference in drivers_probe
driver core: fix race with userland in device_add()
sysfs/kernfs: make read requests on pre-alloc files use the buffer.
sysfs/kernfs: allow attributes to request write buffer be pre-allocated.
fs: sysfs: return EGBIG on write if offset is larger than file size
...
An underrun (playback) event occurs when the serializer transfer
data from the XRBUF buffer to the XRSR shift register, but the
XRBUF hasn't been filled. Similarly, the overrun (capture) event
occurs when data from the XRSR shift register is transferred to
the XRBUF but it hasn't been read yet.
These events are handled as XRUN events that cause the pcm to stop.
The stream has to be explicitly restarted by the userspace which
ensures that after stopping/starting McASP the data transfer is
aligned with DMA. The other possibility was to internally stop and
start McASP without DMA even knowing about it.
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Active slots count depends on the number of channels in the stream
and the number of active serializers. Each serializer will handle
at most the number of channels specified via 'tdm-slots' parameter
in DT.
There are two possible scenarios:
- Single serializer: channel count fits in the max slots supported by
McASP serializers, active slots is same as channel count
- Multiple serializers: channel count is bigger than max slots supported
by a serializer. Channel count determines how many serializers are
needed at their max slot count configuration
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
In IIS (I2S, TDM, etc) mode the maximum number of allowed channels for
either direction can be:
number of serializers for the direction * tdm_slots.
This constraint applicable for the first stream, while consequent stream
should not have more channels then the first stream.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Instead of validating the tdm_slots parameter every time at hw_params we
can do it once during probe. If the parameter is not valid (<2 or >32)
print an error and fix it up.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
When the bus is configured to have more BCLK then the data type demands we
need to use the rotation to move the data to correct place.
Reported-by: Misael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
In IIS mode the tx and rx configuration is symmetric, the BCLK and FSYNC is
shared.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
If probe fails then we need to call pm_runtime_disable() to balance
out the previous pm_runtime_enable() call.
Signed-off-by: Anil Kumar <anilk4.v@gmail.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
In this way the start code for tx/rx going to be located at the same place.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The AFIFO should not be stopped (or started for that matter) when McASP is
running since it can cause unpredictable issues because we are switching off
AFIFO for the direction which was handling the requests from McASP and was
generating DMA request toward the system DMA.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Follow the sequence described in the TRMs when starting RX.
Write to RXBUF register was not correct and there is no need to release the
RX state machine/Receive frame sync generator twice.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Follow the sequence described in the TRMs when starting TX. This sequence
will make sure that we are not facing with initial channel swap caused by
no data available in McASP for transmit.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
If the board is sent to suspend (deep sleep) the McASP context will be lost.
In case when suspend happens during active audio we need to save and restore
more registers, which was configured during hw_param times as well.
We need to add more config registers, AFIFO control registers and we also
need to save and restore the serializer configuration as well.
Since the number of serializers depends on the SoC we need to allocate the
memory for it based on the num_serializer for the given McASP instance.
With this patch the ongoing stream will resume after resuming from deep
sleep.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Instead of individual values use an array to store the registers need to be
saved on suspend and restored on resume.
It is going to be easier to add more registers to save and restore.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
It is possible to configure a kernel with SND_DAVINCI_SOC=m and
SND_DM365_VOICE_CODEC=y, which results in a link error:
sound/built-in.o: In function `davinci_vcif_probe':
sound/soc/davinci/davinci-vcif.c:223: undefined reference to `davinci_soc_platform_register'
The best way to avoid this is to make SND_DM365_VOICE_CODEC
a tristate option that depends on SND_DAVINCI_SOC, so it
can only be a module or disabled when the base driver is
a loadable module
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
In case of capture we should not use rotation. The reverse and mask is
enough to get the data align correctly from the bus to MCU:
Format data from bus after reverse (XRBUF)
S16_LE: |LSB|MSB|xxx|xxx| |xxx|xxx|MSB|LSB|
S24_3LE: |LSB|DAT|MSB|xxx| |xxx|MSB|DAT|LSB|
S24_LE: |LSB|DAT|MSB|xxx| |xxx|MSB|DAT|LSB|
S32_LE: |LSB|DAT|DAT|MSB| |MSB|DAT|DAT|LSB|
With this patch all supported formats will work for playback and capture.
Reported-by: Jyri Sarha <jsarha@ti.com> (broken S24_3LE capture)
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
edma_pcm_platform_register() is declared in edma-pcm.h and defined in
edma-pcm.c. To make sure that the function signature matches for both
edma-pcm.c should include edma-pcm.h
Fixes the following sparse warning:
sound/soc/davinci/edma-pcm.c:48:5: warning: symbol
'edma_pcm_platform_register' was not declared. Should it be static?
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@linaro.org>
The implicit BLCK divider setting was broken by "ASoC: mcasp: don't
override bclk divider if it was provided by the machine"-patch. After
the BCLK divider is implicitly set for the first time the
mcasp->bclk_div gets a non zero value and the implicit setting is
"turned off".
Signed-off-by: Jyri Sarha <jsarha@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
In order to be able to use simple-card with McASP on TI SoC based boards we
need to be able to select the McASP via menuconfig.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
If a machine driver provides an BCLK divider to the McASP driver, skip the
automatic calculation.
This fixes machines on which the physical sample transport always works
in 32 bits, even though not all of them are actually used.
snd_soc_params_to_bclk() will fail to address such cases properly.
Signed-off-by: Daniel Mack <zonque@gmail.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Use the edma-pcm with AM335x and AM437x SoCs.
Keep using the davinci-pcm for daVinci devices, they can be switched to use
the dmaengine based driver later when they are verified to work correctly.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Daniel Mack <daniel@zonque.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
So drivers mixing with other platform drivers, like davinci-mcasp do not
need to fiddle with CONFIG_SND_EDMA_SOC in their code.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Daniel Mack <daniel@zonque.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
The SNDRV_PCM_INFO_BATCH will be added back based on the dmaengine driver's
capabilities.
Patches for edma dmaengine driver has been prepared to suppress the
interrupts. We can add this flag right away and have the benefit of not
having interrupts during audio activity when the edma and ASoC patches
are in the same tree.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Daniel Mack <daniel@zonque.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
In case of S24_LE/U24_LE modes we expect 24bits on the bus while the samples
are stored and transferred in memory on 32bits (lower 3 bytes of the 4
bytes).
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Daniel Mack <daniel@zonque.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Do not fail if the exact BLCK rate can not be produced, just print a
warning. Check that sysclk frequency is set before implicitly setting
the BCLK divider.
Signed-off-by: Jyri Sarha <jsarha@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Fixes build with SND_DAVINCI_SOC or SND_OMAP_SOC alone and adds build
dependecy to SND_DAVINCI_SOC or SND_OMAP_SOC.
Signed-off-by: Jyri Sarha <jsarha@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
In DIT (S/PDIF) mode, program the transmitted user bits to reflect the
configured sample rate, along with some other details.
Signed-off-by: Daniel Mack <zonque@gmail.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
No need to go via the CODEC to get a pointer to the card. This will help to
eventually remove the card field from the snd_soc_codec struct.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Make including the omap-pcm.h outside sound/soc/omap more convenient.
Signed-off-by: Jyri Sarha <jsarha@ti.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
This codec requires I2C to be enabled, so any other option
that selects it should also depend on I2C.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Xia Kaixu <kaixu.xia@linaro.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Remove the cleanup code related to the platform from the DAI drivers at the
same time to avoid breakage.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Version 4 of McASP is using omap-pcm as platform driver and the omap-pcm
platform need to be registered using the cpu dai's device.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
This register is not defined in TI81xx and on AM335x/AM437x it is the
SYSCONFIG register which should not be touched by drivers since it is
related to PM and handled by the generic PM code.
This register write was there since the first time the davinci-mcasp driver
was appeared in the kernel.
The reason why it did not caused any issues on AM335x/AM437x is that it sets
bit 1 in SYSCONFIG register which in turn will enable the smart-idle mode.
This is the default mode and this is the mode McASP should be in also when
in use.
On TI81xx the register is not defined.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
The FS needs to be inverted in McASP compared to other supported formats.
Use a flag to indicate if the FS needs to be inverted.
At the same time fail when non supported format is asked since the default
case was anyways configuring McASP to a not valid format.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
FS polarity can be either rising or falling edge in McASP. Instead of
accessing the registers in every switch/case set a flag and write the
registers after the switch for the invert configuration.
This change will help when adding support for more formats also.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
DSP_A is like DSP_B mode but with one bit delay after the FS.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Use intermediate variable for the data delay needed for the specific format
and write the register after the format configuration at once.
This will help to control the number of lines as support for more formats
going to be added.
Also fixes a case when we switch between two formats with different delay
requirements.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
When the AFIFO is not enabled but more than one serializers are used the DMA
need to transfer number of words equal to active serializers when a DMA
request is generated.
When configuring the burst for the DMA avoid using value '1' for the burst
since it is going to enable additional logic in the DMA drivers. Burst '1'
means that the DMA should send/receive one word per DMA requests.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
