soc-topology adds extra dai_link by using snd_soc_add_dai_link(),
and removes it by snd_soc_romove_dai_link().
This snd_soc_add/remove_dai_link() and/or its related
functions are unbalanced before, and now, these are balance-uped.
But, it finds the random operation issue, and it is reported by
Pierre-Louis.
When card was released, topology will call snd_soc_remove_dai_link()
via (A).
/* This should be called before snd_card_free() */
(A) soc_remove_link_components(card);
/* free the ALSA card at first; this syncs with pending operations */
if (card->snd_card) {
(B) snd_card_free(card->snd_card);
card->snd_card = NULL;
}
/* remove and free each DAI */
(X) soc_remove_link_dais(card);
/* need to sync the delayed work before releasing resources */
flush_delayed_work(&rtd->delayed_work);
snd_soc_pcm_component_free(rtd);
}
Here, it gets rtd via pcm->private_data.
But, topology related rtd are already freed at (A).
Normal sound card has no damage, becase it frees rtd at (C).
These are finalizing rtd related data.
Thus, these should be called when rtd was freed, not sound card
was freed. It is very natural and understandable.
In other words, pcm->private_free = soc_pcm_private_free()
is no longer needed.
Extra issue is that there is zero chance to call
soc_remove_dai() for topology related dai at (X).
Because (A) removes rtd connection from card too, and,
(X) is based on card connected rtd.
This means, (X) need to be called before (C) (= for normal sound)
and (A) (= for topology).
Now, I want to focus this patch which is the reason why
snd_card_free() = (B) is located there.
Original snd_card_free() was called last of this function.
But moved to top to avoid use-after-free issue.
The issue was happen at soc_pcm_free() which was pcm->private_free,
today it is updated/renamed to soc_pcm_private_free().
In other words, (B) need to be called before (C) (= for normal sound)
and (A) (= for topology), because it needs (not yet freed) rtd.
But, (A) need to be called before (B),
because it needs card->snd_card pointer.
If we call flush_delayed_work() and snd_soc_pcm_component_free()
(= same as soc_pcm_private_free()) when rtd was freed (= (C), (A)),
there is no reason to call snd_card_free() at top of this function.
It can be called end of this function, again.
But, in such case, it will likely break unbind again, as Takashi-san
reported. When unbind is performed in a busy state, the code may
release still-in-use resources.
At least we need to call snd_card_disconnect_sync() at the first place.
Li Xu [Fri, 15 Nov 2019 19:54:13 +0000 (13:54 -0600)]
ASoC: wm_adsp: Expose mixer control API
Expose mixer control API for reading and writing controls from the kernel.
This API can be used by ALSA kernel drivers with ADSP support to
read and write firmware-defined memory regions.
Lucas Stach [Mon, 18 Nov 2019 15:12:06 +0000 (16:12 +0100)]
ASoC: tlv320aic31xx: configure output common-mode voltage
The tlv320aic31xx devices allow to adjust the output common-mode voltage
for best analog performance. The datasheet states that the common mode
voltage should be set to be <= AVDD/2.
This changes allows to configure the output common-mode voltage via a DT
property. If the property is absent the voltage is automatically chosen
as the highest voltage below/equal to AVDD/2.
Vinod Koul [Fri, 15 Nov 2019 10:27:03 +0000 (15:57 +0530)]
ALSA: compress: add flac decoder params
The current design of sending codec parameters assumes that decoders
will have parsers so they can parse the encoded stream for parameters
and configure the decoder.
But this assumption may not be universally true and we know some DSPs
which do not contain the parsers so additional parameters are required
to be passed.
So add these parameters starting with FLAC decoder. The size of
snd_codec_options is still 120 bytes after this change (due to this
being a union)
ASoC: soc-core: care card_probed at soc_cleanup_card_resources()
soc_cleanup_card_resources() will call card->remove(), but it should be
called if card->probe() or card->late_probe() are called.
snd_soc_bind_card() might be error before calling
card->probe() / card->late_probe().
In that time, card->remove() will be called.
This patch adds card_probed parameter to judge it.
Fixes: bfce78a559655 ("ASoC: soc-core: tidyup soc_init_dai_link()") Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Tested-by: Jon Hunter <jonathanh@nvidia.com> Link: https://lore.kernel.org/r/87o8xg4ltr.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: call snd_soc_dapm_shutdown() at soc_cleanup_card_resources()
It is easy to read code if it is cleanly using paired function/naming,
like start <-> stop, register <-> unregister, etc, etc.
But, current ALSA SoC code is very random, unbalance, not paired, etc.
It is easy to create bug at the such code, and it will be difficult to
debug.
snd_soc_bind_card() is calling snd_soc_dapm_init() for both
card and component.
Let's call paired snd_soc_dapm_shutdown() at paired
soc_cleanup_card_resources().
Rob Herring [Fri, 8 Nov 2019 15:35:38 +0000 (09:35 -0600)]
dt-bindings: vendor-prefixes: Add 'fsi[ab]' to the list of non-vendor prefixes
'fsia' and 'fsib' property prefixes were added as schema in commit 2f52475bac7e ("ASoC: fsi: switch to yaml base Documentation").
Unfortunately to do checks on actual vendor prefixes, we have to track
the handful of prefixes which are not vendors like 'fsia' and 'fsib'.
Fixes: 2f52475bac7e ASoC: fsi: switch to yaml base Documentation Cc: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Cc: Mark Brown <broonie@kernel.org> Signed-off-by: Rob Herring <robh@kernel.org> Link: https://lore.kernel.org/r/20191108153538.11970-1-robh@kernel.org Signed-off-by: Mark Brown <broonie@kernel.org>
snd_soc_bind_card() is calling many initialize functions
for each card / link / dai / aux etc, etc, etc...
When error happen, the message is indicated at snd_soc_bind_card(),
not at each functions.
But, only soc_probe_aux_devices() case is indicating error at functions,
not at snd_soc_bind_card().
It is not an issue, but unbalanced.
This patch moves error message to snd_soc_bind_card().
Also avoids deep-nested code.
ASoC: soc-core: merge snd_soc_add_dai_link() and soc_bind_dai_link()
We don't need to separete snd_soc_add_dai_link() and
soc_bind_dai_link() anymore. Let's merge these.
One note is that before this patch, it adds list (A)
eventhough if it had dai_link->ignore (1), or already bounded dai_link (2).
But I guess it is wrong. This patch also solve this issue.
/* BEFORE */
int soc_bind_dai_link(...)
{
...
(1) if (dai_link->ignore)
return 0;
(2) if (soc_is_dai_link_bound(...))
return 0;
...
}
int snd_soc_add_dai_link(...)
{
...
=> ret = soc_bind_dai_link(...);
=> if (ret < 0)
=> return ret;
Because complex separeted "card pre-listed component" and
"topology added component" duplicated operation is now
becoming simple, we don't need to check already bound dai_link
which is not exist anymore.
This patch removes soc_is_dai_link_bound().
Dragos Tarcatu [Mon, 11 Nov 2019 22:20:38 +0000 (16:20 -0600)]
ASoC: SOF: topology: free kcontrol memory on error
The volume and bytes kcontrols are currently not freeing their
memory on initialization failures. When an error occurs, all the
widgets loaded so far are unloaded via sof_widget_unload().
But this only happens for the widgets that got successfully loaded.
Fix that by kfree()-ing the allocated memory on load error.
Fixes: 311ce4fe7637d ("ASoC: SOF: Add support for loading topologies") Reviewed-by: Paul Olaru <paul.olaru@nxp.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Signed-off-by: Dragos Tarcatu <dragos_tarcatu@mentor.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191111222039.19651-1-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Keyon Jie [Mon, 11 Nov 2019 22:33:43 +0000 (16:33 -0600)]
ASoC: SOF: PM: only suspend to D0I3 when needed
We should suspend audio to D3 by default, for the sake of power saving,
change the condition of D0I3 suspending here to that when there is
stream with suspend_ignored specified.
Keyon Jie [Mon, 11 Nov 2019 22:33:42 +0000 (16:33 -0600)]
ASoC: SOF: add helper to check if we should enter d0i3 suspend
Add helper to check if the DSP should be put in D0i3. This function
returns true if a stream has ignored the SUSPEND trigger to keep the
pipelines running in the DSP.
ASoC: SOF: Intel: hda: use fallback for firmware name
We have platforms such as CFL with no known I2S codec being used, and
the ACPI tables are currently empty, so fall-back to using the
firmware filename used in nocodec mode
ASoC: Intel: kbl_rt5663_rt5514_max98927: Add dmic format constraint
On KBL platform, the microphone is attached to external codec(rt5514)
instead of PCH. However, TDM slot between PCH and codec is 16 bits only.
In order to avoid setting wrong format, we should add a constraint to
force to use 16 bits format forever.
This commit adds the Dialog DA7213 audio codec as a selectable option
in the kernel config. Currently the driver can only be selected for
Intel Baytrail/Cherrytrail devices or if SND_SOC_ALL_CODECS is enabled.
Curtis Malainey [Wed, 6 Nov 2019 01:13:34 +0000 (17:13 -0800)]
ASoC: bdw-rt5677: Turn on MCLK1 for DSP via DAPM
The RT5677 DSP needs the I2S MCLK1 to run its DSP. Add a dapm route to
SSP0 CODEC IN so the clock is turned on automatically when the DSP is
turned on.
Curtis Malainey [Wed, 6 Nov 2019 01:13:33 +0000 (17:13 -0800)]
ASoC: rt5677: Allow VAD to be shut on/off at all times
Due to limitations of the clocking configuration, we have no way of
scheduling our hibernation before the bdw dsp hibernates. This causes
issues when the system suspends with an open stream. We need userspace
to toggle the kcontrol before we are suspended so that any writes on
suspend are not lost and we don't corrupt the regmap.
Ben Zhang [Wed, 6 Nov 2019 01:13:32 +0000 (17:13 -0800)]
ASoC: rt5677: Disable irq at suspend
The irq is disabled at suspend to avoid running the threaded irq
handler after the codec has been powered off. At resume, codec irq is
re-enabled and the interrupt status register is checked to see if
headphone has been pluggnd/unplugged while the device is suspended.
There is still a chance that the headphone gets enabled or disabled
after the codec is suspended. disable_irq syncs the threaded irq
handler, but soc-jack's threaded irq handler schedules a delayed
work to poll gpios (for debounce). This is still OK. The codec won't
be powered back on again because all audio paths have been suspended,
and there are no force enabled supply widgets (MICBIAS1 is disabled).
The gpio status read after codec power off could be wrong, so the
gpio values are checked again after resume.
Ben Zhang [Wed, 6 Nov 2019 01:13:31 +0000 (17:13 -0800)]
ASoC: rt5677: Stop and restart DSP over suspend/resume
MCLK1 gets disabled at suspend and re-enabled at resume. Before
MCLK1 is re-enabled, if the DSP is already on (either the DSP was
left on during suspend, or the DSP is turned on early at resume),
i2c register read returns garbage and corrupts the regmap cache.
This patch stops the DSP before suspend and restarts it after
resume with a dalay to ensure MCLK is on while loading firmware.
Ben Zhang [Wed, 6 Nov 2019 01:13:30 +0000 (17:13 -0800)]
ASoC: rt5677: Mark reg RT5677_PWR_ANLG2 as volatile
The codec dies when RT5677_PWR_ANLG2(MX-64h) is set to 0xACE1
while it's streaming audio over SPI. The DSP firmware turns
on PLL2 (MX-64 bit 8) when SPI streaming starts. However regmap
does not believe that register can change by itself. When
BST1 (bit 15) is turned on with regmap_update_bits(), it doesn't
read the register first before write, so PLL2 power bit is
cleared by accident.
Marking MX-64h as volatile in regmap solved the issue.
Ben Zhang [Wed, 6 Nov 2019 01:13:28 +0000 (17:13 -0800)]
ASoC: rt5677: Enable jack detect while DSP is running
Before a hotword is detected, GPIO1 pin is configured as IRQ
output so that jack detect works. When a hotword is detected,
the DSP firmware configures the GPIO1 pin as GPIO1 and
drives a 1. rt5677_irq() is called after a rising edge on
the GPIO1 pin, due to either jack detect event or hotword
event, or both. All possible events are checked and handled
in rt5677_irq() where GPIO1 pin is configured back to IRQ
output if a hotword is detected.
Ben Zhang [Wed, 6 Nov 2019 01:13:26 +0000 (17:13 -0800)]
ASoC: rt5677: Load firmware via SPI using delayed work
The firmware rt5677_elf_vad is an ELF binary obtained from
request_firmware(). Sections of the ELF are loaded to
the DSP via SPI. A model (e.g. en_us.mmap) can optionally be
loaded to the DSP at RT5677_MODEL_ADDR to overwrite the
baked-in model in rt5677_elf_vad.
Then we switch to DSP mode, load firmware, and let DSP run.
When a hotword is detected, an interrupt is fired and
rt5677_irq() is called. When 'DSP VAD Switch' is turned off,
the codec is set back to normal mode.
The kcontrol 'DSP VAD Switch' is automatically enabled/disabled
when the hotwording PCM stream is opened/closed.
Shengjiu Wang [Mon, 11 Nov 2019 07:50:48 +0000 (15:50 +0800)]
ASoC: fsl_audmix: Add spin lock to protect tdms
Audmix support two substream, When two substream start
to run, the trigger function may be called by two substream
in same time, that the priv->tdms may be updated wrongly.
The expected priv->tdms is 0x3, but sometimes the
result is 0x2, or 0x1.
Dragos Tarcatu [Wed, 6 Nov 2019 14:58:16 +0000 (08:58 -0600)]
ASoC: SOF: topology: Fix bytes control size checks
When using the example SOF amp widget topology, KASAN dumps this
when the AMP bytes kcontrol gets loaded:
[ 9.579548] BUG: KASAN: slab-out-of-bounds in
sof_control_load+0x8cc/0xac0 [snd_sof]
[ 9.588194] Write of size 40 at addr ffff8882314559dc by task
systemd-udevd/2411
Fix that by rejecting the topology if the bytes data size > max_size
Fixes: 311ce4fe7637d ("ASoC: SOF: Add support for loading topologies") Reviewed-by: Jaska Uimonen <jaska.uimonen@intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Signed-off-by: Dragos Tarcatu <dragos_tarcatu@mentor.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191106145816.9367-1-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: fixup dead-lock at snd_soc_unregister_component()
snd_soc_unregister_component() is calling snd_soc_lookup_component()
under mutex_lock(). But, snd_soc_lookup_component() itself is using
mutex_lock(), thus it will be dead-lock.
This patch adds _nolocked version of it, and avoid dead-lock issue.
Fixes: ac6a4dd3e9f0("ASoC: soc-core: use snd_soc_lookup_component() at snd_soc_unregister_component()") Reported-by: "kernelci.org bot" <bot@kernelci.org>" Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Link: https://lore.kernel.org/r/87bltph4da.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
Dragos Tarcatu [Wed, 6 Nov 2019 14:58:16 +0000 (08:58 -0600)]
ASoC: SOF: topology: Fix bytes control size checks
When using the example SOF amp widget topology, KASAN dumps this
when the AMP bytes kcontrol gets loaded:
[ 9.579548] BUG: KASAN: slab-out-of-bounds in
sof_control_load+0x8cc/0xac0 [snd_sof]
[ 9.588194] Write of size 40 at addr ffff8882314559dc by task
systemd-udevd/2411
Fix that by rejecting the topology if the bytes data size > max_size
Fixes: 311ce4fe7637d ("ASoC: SOF: Add support for loading topologies") Reviewed-by: Jaska Uimonen <jaska.uimonen@intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Signed-off-by: Dragos Tarcatu <dragos_tarcatu@mentor.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191106145816.9367-1-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: remove topology specific operation
soc-core has some API which is used from topology, but it is doing
topology specific operation at soc-core.
soc-core should care about core things, and topology should care
about topology things, otherwise, it is very confusable.
For example topology type is not related to soc-core,
it is topology side issue.
This patch removes meaningless check from soc-core.
This patch keeps extra initialization/destruction at
snd_soc_add_dai_link() / snd_soc_remove_dai_link()
which were for topology.
From this patch, non-topology card can use it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/87pni6251h.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: call snd_soc_register_dai() from snd_soc_register_dais()
ALSA SoC has 2 functions.
snd_soc_register_dai() is used from topology
snd_soc_register_dais() is used from snd_soc_add_component()
In general, people think like _dai() is called from _dais()
with for loop. But in reality, these are very similar
but different implementation.
We shouldn't have duplicated and confusing implementation.
This patch calls snd_soc_register_dai() from snd_soc_register_dais()
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/87r22m251l.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: don't call snd_soc_dapm_new_dai_widgets() at snd_soc_register_dai()
ALSA SoC has 2 functions.
snd_soc_register_dai() is used from topology
snd_soc_register_dais() is used from snd_soc_add_component()
In general, people think like _dai() is called from _dais()
with for loop. But in reality, these are very similar
but different implementation.
We shouldn't have duplicated and confusing implementation.
snd_soc_register_dai() is now used from topology.
But to reduce duplicated code, it should be used from _dais(), too.
Because of topology side specific reason,
it is calling snd_soc_dapm_new_dai_widgets(),
but it is not needed _dais() side.
This patch factorizes snd_soc_register_dai() to
topology / _dais() common part, and topology specific part.
And do topology specific part at soc-topology.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/87sgn2251p.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: have legacy_dai_naming at snd_soc_register_dai()
ALSA SoC has 2 functions.
snd_soc_register_dai() is used from topology
snd_soc_register_dais() is used from snd_soc_add_component()
In general, people think like _dai() is called from _dais()
with for loop. But in reality, these are very similar
but different implementation.
We shouldn't have duplicated and confusing implementation.
snd_soc_register_dai() is now used from topology.
But to reduce duplicated code, it should be used from _dais(), too.
To prepare it, this patch adds missing parameter legacy_dai_naming
to snd_soc_register_dai().
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/87tv7i251u.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
It is easy to read code if it is cleanly using paired function/naming,
like start <-> stop, register <-> unregister, etc, etc.
But, current ALSA SoC code is very random, unbalance, not paired, etc.
It is easy to create bug at the such code, and is difficult to debug.
This patch adds missing soc_del_dai() and snd_soc_unregister_dai().
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/87v9ry251z.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: use snd_soc_lookup_component() at snd_soc_unregister_component()
snd_soc_unregister_component() is now finding component manually,
but we already have snd_soc_lookup_component() to find component;
Let's use existing function.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/87zhha252c.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
soc-core has
snd_soc_add_component(), snd_soc_component_add(),
snd_soc_del_component(), snd_soc_component_del().
These are very confusing naming.
snd_soc_component_xxx() are called from snd_soc_xxx_component(),
and these are very small.
Let's merge these into snd_soc_xxx_component(), and
remove snd_soc_component_xxx().
It is easy to read code if it is cleanly using paired function/naming,
like start <-> stop, register <-> unregister, etc, etc.
But, current ALSA SoC code is very random, unbalance, not paired, etc.
It is easy to create bug at the such code, and is difficult to debug.
Now ALSA SoC has snd_soc_add_component(), but there is no paired
snd_soc_del_component(). Thus, snd_soc_unregister_component() is
calling cleanup function randomly. it is difficult to read.
This patch adds missing snd_soc_del_component_unlocked() and
balance up code.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/8736f23jn4.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
It is easy to read code if it is cleanly using paired function/naming,
like start <-> stop, register <-> unregister, etc, etc.
But, current ALSA SoC code is very random, unbalance, not paired, etc.
It is easy to create bug at the such code, and it will be difficult to
debug.
ALSA SoC has soc_bind_dai_link(), but its paired soc_unbind_dai_link()
is not implemented.
More confusable is that soc_remove_pcm_runtimes() which should be
soc_unbind_dai_link() is implemented without synchronised
to soc_bind_dai_link().
This patch cleanup this unbalance.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/877e4e3jni.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: soc-core: call soc_bind_dai_link() under snd_soc_add_dai_link()
If we focus to soc_bind_dai_link() at snd_soc_instantiate_card(),
we will notice very complex operation.
static int snd_soc_instantiate_card(...)
{
...
/*
* (1) Bind dai_link via card pre-linked dai_link
*
* Bind dai_link via card pre-linked.
* 1 dai_link will be 1 rtd, and connected to card.
* for_each_card_prelinks() is for card pre-linked dai_link.
*
* Image
*
* card
* - rtd(A)
* - rtd(A)
*/
for_each_card_prelinks(card, i, dai_link) {
ret = soc_bind_dai_link(card, dai_link);
...
}
...
/*
* (2) Connect card pre-linked dai_link to card list
*
* Connect all card pre-linked dai_link to *card list*.
* Here, (A) means from card pre-linked.
*
* Image
*
* card card list
* - rtd(A) - dai_link(A)
* - rtd(A) - dai_link(A)
* - ... - ...
*/
for_each_card_prelinks(card, i, dai_link) {
ret = snd_soc_add_dai_link(card, dai_link);
...
}
...
/*
* (3) Probe binded component
*
* Each rtd has many components.
* Here probes each rtd connected components.
* rtd(A) in Image is the probe target.
*
* During this component probe, topology may add new dai_link to
* *card list* by using snd_soc_add_dai_link() which is
* used at (2).
* Here, (B) means from topology
*
* Image
*
* card card list
* - rtd(A) - dai_link(A)
* - rtd(A) - dai_link(A)
* - ... - ...
* - dai_link(B)
* - dai_link(B)
*/
ret = soc_probe_link_components(card);
...
/*
* (4) Bind dai_link again
*
* Bind dai_link again for topology.
* Note, (1) used for_each_card_prelinks(),
* here is using for_each_card_links()
*
* This means from card list.
* As Image indicating, it has dai_link(A) (from card pre-link)
* and dai_link(B) (from topology).
* main target here is dai_link(B).
* soc_bind_dai_link() ignores already used
* dai_link (= dai_link(A))
*
* Image
*
* card card list
* - rtd(A) - dai_link(A)
* - rtd(A) - dai_link(A)
* - ... - ...
* - rtd(B) - dai_link(B)
* - rtd(B) - dai_link(B)
*/
for_each_card_links(card, dai_link) {
ret = soc_bind_dai_link(card, dai_link);
...
}
...
}
As you see above, it is doing very complex method.
The problem is binding dai_link via "card pre-linked" (= (1)) and
"topology added dai_link" (= (3)) are separated.
The code can be simple if we can bind dai_link when dai_link
is connected to *card list*.
This patch do it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/878sou3jnn.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
soc_is_dai_link_bound() check will be called both
*before* soc_bind_dai_link() (A), and
*under* soc_bind_dai_link() (B).
These are very verbose code. Let's remove one of them.
* static int soc_bind_dai_link(...)
{
...
(B) if (soc_is_dai_link_bound(...)) {
...
return 0;
}
...
}
static int snd_soc_instantiate_card(...)
{
...
for_each_card_links(...) {
(A) if (soc_is_dai_link_bound(...))
continue;
* ret = soc_bind_dai_link(...);
if (ret)
goto probe_end;
}
...
}
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/87a79a3jns.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
Maxime Ripard [Tue, 5 Nov 2019 10:56:15 +0000 (11:56 +0100)]
ASoC: adau7118: Fix example warning
The ADAU7118 has an example where the codec has an i2c address of 14, and
the unit address set to 14 as well.
However, while the address is expressed in decimal, the unit-address is
supposed to be in hexadecimal, which ends up with two different addresses
that trigger a DTC warning. Fix this by setting the address to 0x14.
Cc: Nuno Sá <nuno.sa@analog.com> Cc: Lars-Peter Clausen <lars@metafoo.de> Fixes: 969d49b2cdc8 ("dt-bindings: asoc: Add ADAU7118 documentation") Signed-off-by: Maxime Ripard <maxime@cerno.tech> Acked-by: Rob Herring <robh@kernel.org> Link: https://lore.kernel.org/r/20191105105615.21391-1-maxime@cerno.tech Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: SOF: topology: set trigger order for FE DAI link
Set trigger order for FE DAI links to SND_SOC_DPCM_TRIGGER_POST
to trigger the BE DAI's before the FE DAI's. This prevents the
xruns seen on playback pipelines using the link DMA.
ASoC: pcm: update FE/BE trigger order based on the command
Currently, the trigger orders SND_SOC_DPCM_TRIGGER_PRE/POST
determine the order in which FE DAI and BE DAI are triggered.
In the case of SND_SOC_DPCM_TRIGGER_PRE, the FE DAI is
triggered before the BE DAI and in the case of
SND_SOC_DPCM_TRIGGER_POST, the BE DAI is triggered before
the FE DAI. And this order remains the same irrespective of the
trigger command.
In the case of the SOF driver, during playback, the FW
expects the BE DAI to be triggered before the FE DAI during
the START trigger. The BE DAI trigger handles the starting of
Link DMA and so it must be started before the FE DAI is started
to prevent xruns during pause/release. This can be addressed
by setting the trigger order for the FE dai link to
SND_SOC_DPCM_TRIGGER_POST. But during the STOP trigger,
the FW expects the FE DAI to be triggered before the BE DAI.
Retaining the same order during the START and STOP commands,
results in FW error as the DAI component in the FW is still
active.
The issue can be fixed by mirroring the trigger order of
FE and BE DAI's during the START and STOP trigger. So, with the
trigger order set to SND_SOC_DPCM_TRIGGER_PRE, the FE DAI will be
trigger first during SNDRV_PCM_TRIGGER_START/STOP/RESUME
and the BE DAI will be triggered first during the
STOP/SUSPEND/PAUSE commands. Conversely, with the trigger order
set to SND_SOC_DPCM_TRIGGER_POST, the BE DAI will be triggered
first during the SNDRV_PCM_TRIGGER_START/STOP/RESUME commands
and the FE DAI will be triggered first during the
SNDRV_PCM_TRIGGER_STOP/SUSPEND/PAUSE commands.
This option is only required with the Skylake platform driver, there
is no reason to have this option in machine drivers. This is
e.g. useless for SOF-based solutions.
updated solution to the problem reported with randconfig:
CONFIG_SND_SOC_SOF_IMX depends on CONFIG_SND_SOC_SOF, but is in
turn referenced by the sof-of-dev driver. This creates a reverse
dependency that manifests in a link error when CONFIG_SND_SOC_SOF_OF
is built-in but CONFIG_SND_SOC_SOF_IMX=m:
sound/soc/sof/sof-of-dev.o:(.data+0x118): undefined reference to `sof_imx8_ops'
use def_trisate to propagate the right settings without select.
ASoC: SOF: Kconfig: add EXPERT dependency for developer options, clarify help
Some distros select all possible options, despite existing warnings to
be careful. This leads to e.g. user reports that the HDaudio codec and
DMIC are not handled by SOF.
Add an explicit menu item to unlock developer options, and make them
dependent on CONFIG_EXPERT. Hopefully with this double-lock these
options will only be selected by developers.
ASoC: SOF: Intel: Broadwell: clarify mutual exclusion with legacy driver
Some distros select all options blindly, which leads to confusion and
bug reports. SOF does not fully support Broadwell due to firmware
dependencies, the machine drivers can only support one option, and
UCM/topology files are still being propagated to downstream distros,
so make SOF on Broadwell an opt-in option that first require distros
to opt-out of existing defaults.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=204237 Fixes: f35bf70f61d3 ('ASoC: Intel: Make sure BDW based machine drivers build for SOF') Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191101173045.27099-3-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: SOF: Intel: Baytrail: clarify mutual exclusion with Atom/SST driver
Some distros select all options blindly, which leads to confusion and
bug reports. Since SOF does not support Baytrail-CR for now, and
UCM/topology files are still being propagated to downstream distros,
make SOF on Baytrail an opt-in option that first require distros to
opt-out of existing defaults.
Naveen Manohar [Fri, 1 Nov 2019 17:18:47 +0000 (12:18 -0500)]
ASoC: Intel: boards: Add CML m/c using RT1011 and RT5682
Machine driver to enable
RT5682 on SSP0, DMIC, HDMI and
RT1011 AMP on SSP1 with
2 CH / 24 bit TDM Playback over 4 individual codecs and
4 CH / 24 bit Capture to provide feedback.
Signed-off-by: Naveen Manohar <naveen.m@intel.com> Signed-off-by: Sathya Prakash M R <sathya.prakash.m.r@intel.com> Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20191101171847.26767-3-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Kai Vehmanen [Fri, 1 Nov 2019 17:06:35 +0000 (12:06 -0500)]
ASoC: hdac_hda: fix race in device removal
When ASoC card instance is removed containing a HDA codec,
hdac_hda_codec_remove() may run in parallel with codec resume.
This will cause problems if the HDA link is freed with
snd_hdac_ext_bus_link_put() while the codec is still in
middle of its resume process.
To fix this, change the order such that pm_runtime_disable()
is called before the link is freed. This will ensure any
pending runtime PM action is completed before proceeding
to free the link.
This issue can be easily hit with e.g. SOF driver by loading and
unloading the drivers.