This goto is after the call to clk_get, so it should go to the label that
includes a call to clk_put.
A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)
// <smpl>
@r exists@
expression e1,e2;
statement S;
@@
e1 = clk_get@p1(...);
... when != e1 = e2
when != clk_put(e1)
when any
if (...) { ... when != clk_put(e1)
when != if (...) { ... clk_put(e1) ... }
* return@p3 ...;
} else S
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Really this should be something the IRQ core can cope with for us but since
it doesn't currently do so (at least for threaded interrupts like this) do
so in the driver. This allows us to run with interrupt controllers that
only support edge triggered interrupts.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Current implementation set max98095->sysclk/max98088->sysclk to freq twice.
Set it once is enough, this patch removes the first assignment in case
we may set invalid clock frequency to max98095->sysclk/max98088->sysclk.
Signed-off-by: Axel Lin <axel.lin@gmail.com>
Acked-by: Peter Hsiang <peter.hsiang@maxim-ic.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
We want the default state of the HP_MUTE signal to be asserted, so that
the headphones are muted before the first audio playback. Without this,
the headphones are left unmuted until shortly after the first audio
playback completes.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Run the data through cpu_to_be16() so it's at least clear what we're up to.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Commit af46800 ("ASoC: Implement mux control sharing") introduced
function dapm_is_shared_kcontrol.
When this function returns true, the naming of DAPM controls is derived
from the kcontrol_new. Otherwise, the name comes from the widget (and
possibly a widget's naming prefix).
A bug in the implementation of dapm_is_shared_kcontrol made it return 1
in all cases. Hence, that commit caused a change in control naming for
all controls instead of just shared controls.
Specifically, a control is always considered shared because it is always
compared against itself. Solve this by never comparing against the widget
containing the control being created.
Equally, controls should never be shared between DAPM contexts; when the
same codec is instantiated multiple times, the same kcontrol_new will be
used. However, the control should no be shared between the multiple
instances.
I tested that with the Tegra WM8903 driver:
* Shared is now mostly 0 as expected, and sometimes 1.
* The expected controls are still generated after this change.
However, I don't have any systems that have a widget/control naming
prefix, so I can't test that aspect.
Thanks for Jarkko Nikula for pointing out how to fix this.
Reported-by: Liam Girdwood <lrg@ti.com>
Tested-by: Jarkko Nikula <jhnikula@gmail.com>
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 52ba67b ("ASoC: Force all DAPM contexts into the same bias state")
powers up all the DAPM contexts in a card if any DAPM context becomes
active. Unfortunately power down newer happens if per-card DAPM context
doesn't have any widgets.
Reason for this is that power state of per-card DAPM context without
widgets is never cleared and thus all the DAPM contexts remain permanently
active. Test for widgetless calling DAPM context in dapm_power_widgets()
doesn't work for per-card DAPM context since power change is never
originating from widgetless per-card DAPM context.
Fix this by pre-clearing power state flag of non-codec DAPM context at the
beginning of power sequence.
Signed-off-by: Jarkko Nikula <jhnikula@gmail.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Codec output pin should be defined with SND_SOC_DAPM_OUTPUT.
Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
It's enough to include linux/delay.h just once in
sound/soc/codecs/wm8915.c, so remove the duplicate.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
at91sam9g20 is providing master clock to wm8731: not using a crystal but an
external MCLK. We can avoid conflict and save power using WM8731_SYSCLK_MCLK as
we do not need oscillator to be powered.
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The crystal oscillator is only enabled if the WM8731_SYSCLK_XTAL master clock
is specified. Fix the connected() struct snd_soc_dapm_route function to take
this into account. Oscillator is not enabled on machine that need it otherwise.
Machine drivers have to make sure that they use the proper SYSCLK value.
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
We do not have to free a resource that is not allocated yet.
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
For cards that have two or more DAIs, snd_soc_resume's loop over all
DAIs ends up calling schedule_work(deferred_resume_work) once per DAI.
Since this is the same work item each time, the 2nd and subsequent
calls return 0 (work item already queued), and trigger the dev_err
message below stating that a work item may have been lost.
Solve this by adjusting the loop to simply calculate whether to run the
resume work immediately or defer it, and then call schedule work (or not)
one time based on that.
Note: This has not been tested in mainline, but only in chromeos-2.6.38;
mainline doesn't support suspend/resume on Tegra, nor does the mainline
Tegra ASoC driver contain multiple DAIs. It has been compile-checked in
mainline.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The comment does not reflect reality anymore since the multi-component
monster patch landed. Things are matched by names now, and not by
exporting and referencing a struct. Fix it to avoid confusion.
Signed-off-by: Daniel Mack <zonque@gmail.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Acked-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
In the previous commit 'ASoC: davinci-pcm: convert to BATCH mode', the phase
offset of 2 was mentioned in the commit message but not well commented in the
source.
Add descriptive comments of the phase offset with and without ping-pong
buffers enabled.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The davinci-pcm driver's snd_pcm_ops pointer function currently calls into
the edma controller driver to read the current positions of the edma channels
to determine pos to return to the ALSA framework. In particular,
davinci_pcm_pointer() calls edma_get_position() and the latter has a comment
indicating that "Its channel should not be active when this is called" whereas
the channel is surely active when snd_pcm_ops.pointer is called.
The operation of davinci-pcm in capture and playback appears to follow close
the other pcm drivers who export SNDRV_PCM_INFO_BATCH except that davinci-pcm
does not report it's positions from pointer() using the last transferred
chunk. Instead it peeks directly into the edma controller to determine the
current position as discussed above.
Convert the davinci-pcm driver to BATCH mode: count the periods elapsed in the
prtd->period member and use its value to report the 'pos' to the alsa
framework in the davinci_pcm_pointer function.
There is a phase offset of 2 periods between the position used by dma setup
and the position reported in the pointer function. Either +2 in the dma
setup or -2 in the pointer function (with wrapping, both) accounts for this
offset -- I opted for the latter since it makes the first-time setup clearer.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Reviewed-by: Steven Faludi <stevenfaludi@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Extract functions that modify the prtd->period member in preparation for
conversion to BATCH mode playback.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Reviewed-by: Steven Faludi <stevenfaludi@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The release of the dma channels was being performed in prepare and there was a
edma_resume call for the asp-channel only being executed on START, RESUME and
PAUSE_RELEASE.
The mcasp on da850evm with ping-pong buffers enabled was exhibiting an audible
glitch on every playback after the first. It was determined through trial and
error that the following two changes fix this problem:
1) Move the edma_start calls from prepare to trigger and 2) reverse the order
of starting the asp and ram channels.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Reviewed-by: Steven Faludi <stevenfaludi@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Based on the registration of davinci-mcasp.1 in the davinci-evm platform
setup for da830 and dm6467, davinci-pcm can handle more than the currently
reported maximum channels of 2.
Increase the maximum channels to 384 to match the maximum reported by
davinci-mcasp.1.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Reviewed-by: Steven Faludi <stevenfaludi@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Based on the data_type test in ping_pong_dma_setup, davinci-pcm is capable of
handling data of width up to and including 32bits.
"
if ((data_type == 0) || (data_type > 4)) {
printk(KERN_ERR "%s: data_type=%i\n", __func__, data_type);
return -EINVAL;
}
"
Update the .format member of the snd_pcm_hardware instances it registers to
reflect this capability.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Reviewed-by: Steven Faludi <stevenfaludi@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The setup of the pong channel uses EDMA_CHAN_SLOT instead of & 0x3f as the
setup of the ping channel does.
Make the setup of ping and pong symmetric. There is no functional change
introduced by this patch.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Reviewed-by: Steven Faludi <stevenfaludi@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Currently CODEC and platform drivers have their module reference count
incremented soc_probe_dai_link() whilst CPU DAI drivers have their reference
count incremented in soc_bind_dai_link().
CPU DAIs should have their reference count incremented in soc_probe_dai_link()
just like the CODEC and platform drivers.
Signed-off-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Commit f0fba2ad (ASoC: multi-component - ASoC Multi-Component Support)
broke support for Raumfeld platforms as it didn't take into account the
different hardware features on individual devices.
In particular, Raumfeld speakers have no S/PDIF output, so the members
of the snd_soc_card struct must be set dynamically.
Signed-off-by: Daniel Mack <zonque@gmail.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@kernel.org
This patch is preparation of cleanup suspend/resume patch.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Current FSI driver was using saved_xxx variable for suspend/resume.
OTOH, the start and stop of power/clock are controlled by
fsi_hw_startup/fsi_hw_shutdown in current FSI driver.
The all necessary registers value are set by fsi_hw_startup.
So, if fsi_hw_shutdown is called when "suspend" is generated,
and fsi_hw_startup is called at "resume",
the saved_xxx are not needed.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
FSIA/B ports is enabled by default when power-on,
and current FSI is supporting RuntimePM.
In addition, current fsi_module_init/kill doesn't care
simultaneous playback/recorde.
This mean FSI port control is not needed.
This patch remove fsi_module_init/kill
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
fsi_stream_push/pop might be called in same time.
This patch protect it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
pm_runtime_get/put_sync were used to access FSI register in fsi_dai_set_fmt
which is called when ALSA probe.
But this register value will disappear after pm_runtime_put_sync
if platform is supporting RuntimePM.
To solve this issue, this patch adds new variable for format,
and remove pm_runtime_get/put_sync from fsi_dai_set_fmt.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Some variables on this driver were a unclear naming,
and were different unit (byte, frame, sample).
And some functions had wrong name
(ex. it returned "sample width" but name was "fsi_get_frame_width").
This patch tidy-up this issue, and the minimum unit become "sample".
Special thanks to Takashi YOSHII.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Using fsi_irq_enable/disable in fsi_port_start/stop is very natural.
This patch is preparation of cleanup suspend/resume patch.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Current FSI driver is using set_rate call back function which is for
master mode.
By this patch, it is used from fsi_set_master_clk.
This patch is preparation of cleanup suspend/resume patch.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
It is possible to create buff_len and period_len
from substream->runtime.
This patch is preparation of tidyup unclear variable naming patch.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Commit 22de71b ("ASoC: core - allow ASoC more flexible machine name")
writes "(null)" to driver name string in struct snd_card if card->driver_name
is NULL. This causes segmentation faults with some user space ALSA utilities
like aplay and arecord.
Fix this by using the card->name if no driver name is specified.
Signed-off-by: Jarkko Nikula <jhnikula@gmail.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Whenever we are doing a read or a write through the rbtree code, we'll
cache a pointer to the rbnode. To avoid looking up the register
everytime we do a read or a write, we first check if it can be found in
the cached register block, otherwise we traverse the rbtree and finally
cache the rbnode for future use.
Signed-off-by: Dimitris Papastamos <dp@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
This patch prepares the ground for the actual rbtree optimization patch
which will save a pointer to the last accessed rbnode that was used
in either the read() or write() functions.
Each rbnode manages a variable length block of registers. There can be no
two nodes with overlapping blocks. Each block has a base register and a
currently top register, all the other registers, if any, lie in between these
two and in ascending order.
The reasoning behind the construction of this rbtree is simple. In the
snd_soc_rbtree_cache_init() function, we iterate over the register defaults
provided by the driver. For each register value that is non-zero we
insert it in the rbtree. In order to determine in which rbnode we need
to add the register, we first look if there is another register already
added that is adjacent to the one we are about to add. If that is the case
we append it in that rbnode block, otherwise we create a new rbnode
with a single register in its block and add it to the tree.
In the next patch, where a cached rbnode is used by both the write() and the
read() functions, we also check if the register we are about to add is in the
cached rbnode (the least recently accessed one) and if so we append it in that
rbnode block.
Signed-off-by: Dimitris Papastamos <dp@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The prototype of the inline dummy version of tegra_i2s_debug_add
was not consistent with the real version.
Reported-by: Rhyland-Klein <rklein@nvidia.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
MODULE_ALIAS is required so that the module will auto-load based on a
platform_device registration in the board file.
While we're at it, add some other MODULE_*.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The davinci-i2s driver copies the platform data for playback and capture
sram sizes which is in turn used by davinci-pcm to allocate ping-pong
buffers.
Copy also the platform data in davinci-mcasp probe.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Acked-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
AK4641 connected via I2S and I2C, jack detection via GPIO.
Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com>
Signed-off-by: Dmitry Artamonow <mad_soft@inbox.ru>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
