snd_pcm_ops is no longer needed.
Let's use component driver callback.
*Note*
Only Intel skl-pcm has .get_time_info implementation, but ALSA SoC
framework doesn't call it so far.
To keep its implementation, this patch keeps .get_time_info,
but it is still not called.
Intel guy need to support it in the future.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87o8yzaf2f.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Current snd_soc_component_driver has pcm_new/pcm_free, but,
it doesn't have "component" at parameter.
Thus, each callback can't know it is called for which component.
Each callback currently is getting "component" by using
snd_soc_rtdcom_lookup() with driver name.
It works today, but, will not work in the future if we support multi
CPU/Codec/Platform, because 1 rtd might have multiple same driver
name component.
To solve this issue, each callback need to be called with component.
This patch adds new pcm_construct/pcm_destruct with "component"
parameter.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87sgobaf3g.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Current snd_soc_component_driver has snd_pcm_ops, and each driver can
have callback via it (1).
But, it is mainly created for ALSA, thus, it doesn't have "component"
as parameter for ALSA SoC (1)(2).
Thus, each callback can't know it is called for which component.
Thus, each callback currently is getting "component" by using
snd_soc_rtdcom_lookup() with driver name (3).
--- ALSA SoC ---
...
if (component->driver->ops &&
component->driver->ops->open)
(1) return component->driver->ops->open(substream);
...
--- driver ---
(2) static int xxx_open(struct snd_pcm_substream *substream)
{
struct snd_soc_pcm_runtime *rtd = substream->private_data;
(3) struct snd_soc_component *component = snd_soc_rtdcom_lookup(..);
...
}
It works today, but, will not work in the future if we support multi
CPU/Codec/Platform, because 1 rtd might have multiple components which
have same driver name.
To solve this issue, each callback needs to be called with component.
We already have many component driver callback.
This patch copies each snd_pcm_ops member under component driver,
and having "component" as parameter.
--- ALSA SoC ---
...
if (component->driver->open)
=> return component->driver->open(component, substream);
...
--- driver ---
=> static int xxx_open(struct snd_soc_component *component,
struct snd_pcm_substream *substream)
{
...
}
*Note*
Only Intel skl-pcm has .get_time_info implementation, but ALSA SoC
framework doesn't call it so far.
To keep its implementation, this patch keeps .get_time_info,
but it is still not called.
Intel guy need to support it in the future.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87tv8raf3r.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Before regmap_reinit_cache we must reset audio regs as default values.
So we use reset controller unit(toprgu) to reset audio hw.
Signed-off-by: Jiaxin Yu <jiaxin.yu@mediatek.com>
Reviewed-by: Yingjoe Chen <yingjoe.chen@mediatek.com>
Link: https://lore.kernel.org/r/1569580317-21181-5-git-send-email-jiaxin.yu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The snd_pcm_hw_constraint_minmax() works fine when a single codec is
connected to a single CPU DAI, but in multicodec or DPCM setup the
constraints placed by the driver will apply to the whole PCM stream (FE
included) and thus prevents more than 8 playback channels for example.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20191008115720.7135-1-peter.ujfalusi@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
According the documentation for snd_soc_update_bits the API
will return a 1 if the update was successful with a value change,
a 0 if the update was successful with no value change or a negative
if the command just failed.
So the value of return in the driver needs to be checked for being less
then 0 or the caller may indicate failure when the value actually
changed.
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Link: https://lore.kernel.org/r/20191007171157.17813-1-dmurphy@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Remove unused defines and structure variables that are not
referenced by the code. If these are needed for future
enhancements then they should be added at that time.
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Link: https://lore.kernel.org/r/20191007171157.17813-3-dmurphy@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Remove the unneeded and incorrect read of the TDM_CFG3 register.
The read is done but the value is never used.
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Link: https://lore.kernel.org/r/20191007171157.17813-2-dmurphy@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
There are several problems in the error handling in fsl_mqs_probe().
1) "ret" isn't initialized on some paths. GCC has a feature which
warns about uninitialized variables but the code initializes "ret"
to zero at the start of the function so the checking is turned off.
2) "gpr_np" is a pointer so initializing it to zero is confusing and
generates a Sparse warning.
3) of_parse_phandle() doesn't return error pointers on error, it returns
NULL.
4) If devm_snd_soc_register_component() fails then the function should
free the "gpr_np".
Fixes: 9e28f6532c ("ASoC: fsl_mqs: Add MQS component driver")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20191004102208.GB823@mwanda
Signed-off-by: Mark Brown <broonie@kernel.org>
Fixes gcc '-Wunused-but-set-variable' warning:
sound/soc/fsl/fsl_mqs.c: In function fsl_mqs_hw_params:
sound/soc/fsl/fsl_mqs.c:54:6: warning: variable bclk set but not used [-Wunused-but-set-variable]
It is never used, so can be removed.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20191006105522.58560-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
If GPIOLIB is not set, building fails:
sound/soc/codecs/tas2770.c: In function tas2770_reset:
sound/soc/codecs/tas2770.c:38:3: error: implicit declaration of function gpiod_set_value_cansleep; did you mean gpio_set_value_cansleep? [-Werror=implicit-function-declaration]
gpiod_set_value_cansleep(tas2770->reset_gpio, 0);
^~~~~~~~~~~~~~~~~~~~~~~~
gpio_set_value_cansleep
sound/soc/codecs/tas2770.c: In function tas2770_i2c_probe:
sound/soc/codecs/tas2770.c:749:24: error: implicit declaration of function devm_gpiod_get_optional; did you mean devm_regulator_get_optional? [-Werror=implicit-function-declaration]
tas2770->reset_gpio = devm_gpiod_get_optional(tas2770->dev,
^~~~~~~~~~~~~~~~~~~~~~~
devm_regulator_get_optional
sound/soc/codecs/tas2770.c:751:13: error: GPIOD_OUT_HIGH undeclared (first use in this function); did you mean GPIOF_INIT_HIGH?
GPIOD_OUT_HIGH);
^~~~~~~~~~~~~~
GPIOF_INIT_HIGH
Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 1a476abc72 ("tas2770: add tas2770 smart PA kernel driver")
Suggested-by: Ladislav Michl <ladis@linux-mips.org>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20191006104631.60608-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This probe function is only called if the device is backed by a DT node,
so switch this call to of_device_get_match_data() to reduce code size
and simplify a bit. This also avoids needing to reference a potentially
undefined variable because of_device_get_match_data() doesn't need to
know anything beyond the struct device to find the match table.
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Paul Cercueil <paul@crapouillou.net>
Cc: Liam Girdwood <lgirdwood@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Jaroslav Kysela <perex@perex.cz>
Cc: Takashi Iwai <tiwai@suse.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Frank Rowand <frowand.list@gmail.com>
Cc: <alsa-devel@alsa-project.org>
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20191004214334.149976-8-swboyd@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>
snd_soc_limit_volume() is finding snd_kcontrol by using original coding,
but we already have snd_soc_card_get_kcontrol().
Let's use existing function.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87y2y3afgd.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
card->rtd_num is used to count rtd. Initialize it at
snd_soc_instantiate_card() is very natural and less confusion.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87zhijafgk.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
dai / component related resources are created when component is
registered, and it will be freed when component was unregistered.
These resources are not re-used after that.
This means, we can use devm_xxx for dai / component, without
thinking about kfree().
This patch uses devm_xxx for these.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/8736gbbu1a.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
ALSA needs to setup shortname, longname, and driver.
These methods are very similar.
This patch adds new soc_setup_card_name() and setup these.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/874l0rbu1i.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.
soc_cleanup_card_resources() (a) which is paired function of
snd_soc_instantiate_card() (A) is calling soc_remove_dai_links() (*)
to remove card related resources, but it is breaking
add/remove balance (B)(b)(C)(c)(D)(d), in other words
these should be called from soc_cleanup_card_resources() (a)
from balance point of view.
More headacke is that it is using original removing method for
dai_link even though we already have snd_soc_remove_dai_link()
which is the function for it (d).
This patch removes snd_soc_remove_dai_links() and balance up code.
static void soc_remove_dai_links(...)
{
...
(b) soc_remove_link_dais(card);
(c) soc_remove_link_components(card);
for_each_card_links_safe(card, link, _link) {
...
/* it should use snd_soc_remove_dai_link() here */
(d) list_del(&link->list);
}
}
(a) static int soc_cleanup_card_resources(...)
{
...
/* remove and free each DAI */
(*) soc_remove_dai_links(card);
...
}
(A) static int snd_soc_instantiate_card(struct snd_soc_card *card)
{
...
/* add predefined DAI links to the list */
for_each_card_prelinks(card, i, dai_link)
(B) snd_soc_add_dai_link(card, dai_link);
...
/* probe all components used by DAI links on this card */
(C) ret = soc_probe_link_components(card);
...
/* probe all DAI links on this card */
(D) ret = soc_probe_link_dais(card);
...
}
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/875zl7bu1r.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Current rtd, rtd->dev, rtd->codec_dais are created by normal kzalloc(),
but we want to use devm_kzalloc() as much as possible.
Created rtd->dev is registered by device_register() at
soc_new_pcm_runtime(), and it will be freed at
soc_free_pcm_runtime() by device_unregister().
This means, if we can use devm_kzalloc(rtd->dev, xxx) for
rtd / rtd->codec_dais, all these are automatically freed
via soc_free_pcm_runtime().
This patch uses devm_kzalloc(rtd->dev, xxx) for rtd / rtd->codec_dais.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/877e5nbu1z.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
sound/soc/codecs/tas2770.c:796:3-8: No need to set .owner here. The core will do it.
Remove .owner field if calls are used which set it automatically
Generated by: scripts/coccinelle/api/platform_no_drv_owner.cocci
Fixes: 1a476abc72 ("tas2770: add tas2770 smart PA kernel driver")
CC: Frank Shi <shifu0704@thundersoft.com>
Signed-off-by: kbuild test robot <lkp@intel.com>
Link: https://lore.kernel.org/r/20191001175504.ufhyxh66psrht42k@332d0cec05f4
Signed-off-by: Mark Brown <broonie@kernel.org>
Rename arndale_rt5631.c to just arnddale.c as we support other CODECs
than RT5631. While at it replace spaces in Kconfig with tabs.
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Link: https://lore.kernel.org/r/20191002105652.24821-3-s.nawrocki@samsung.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The Arndale boards come with different types of the audio daughter
board. In order to support the WM1811 one we add new definition of
an ASoC card which will be registered when the driver matches on
"samsung,arndale-wm1811" compatible. There is no runtime detection of
the audio daughter board type at the moment, compatible string of the
audio card needs to be adjusted in DT, e.g. by the bootloader,
depending on actual audio board (CODEC) used.
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Link: https://lore.kernel.org/r/20191002105652.24821-2-s.nawrocki@samsung.com
Signed-off-by: Mark Brown <broonie@kernel.org>
There is only one DAI link so we can drop an unnecessary loop statement.
Use card->dai_link in place of direct static arndale_rt5631_dai[] array
dereference as a prerequisite for adding support for other CODECs.
Unnecessary assignment of dai_link->codecs->name to NULL is removed.
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Link: https://lore.kernel.org/r/20191002105652.24821-1-s.nawrocki@samsung.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Helps with FW debug as it provides DSP IPC processing context.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-13-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch prevents the reset of host period bytes
and uses no_stream_position to record requests
for stream position.
Signed-off-by: Marcin Rajwa <marcin.rajwa@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-10-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
fw_ready should be a mandatory op. Make sure fw_ready ops is set
during probe.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-8-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add config option to prevent DSP entering D3 after any FW exception.
This can then be used to dump FW context for debug.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-5-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In a number of debug cases, the DMA-based trace can add problems
(e.g. with HDaudio channel allocation). It also generates additional
traffic on the bus and if the DMA handling is unreliable will prevent
audio use-cases from working normally. Using the trace also requires
tools to be installed on the target.
The trace can be instead handled as dynamic debug. We can use a
Kconfig to force the trace to be enabled in all cases, or use a module
parameter to enable it on a need-basis, e.g. by setting "options
snd_sof sof_debug=0x1" in a /etc/modprobe.d file.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
When removing sof module the support_button_press function will oops
because hp_jack pointer is not checked for NULL. So add a check to fix
this.
Signed-off-by: Jaska Uimonen <jaska.uimonen@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927201408.925-2-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
When removing sof module the rt5682 jack handler will oops
if jack detection is not disabled. So add remove function,
which disables the jack detection.
Signed-off-by: Jaska Uimonen <jaska.uimonen@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927201408.925-5-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
There is a known issue on some Intel platforms which causes
pause/release to run into xrun's during capture usecases.
The suggested workaround to address the issue is to
disable the entry of lower power L1 state in the physical
DMI link when there is a capture stream open.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-14-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
FW mailbox offset was not set before use and HDR size was not validated.
Fix this.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-12-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The "snd_pcm_substream" handle was not initialized properly
in hda-loader.c for firmware load.
When the HDA DMAs were used to load the firmware,
the interrupts related to firmware load also triggered
calls to snd_sof_pcm_period_elapsed() on a non-existent ALSA
PCM stream.
This caused runtime kernel warnings from
pcm_lib.c:snd_pcm_period_elapsed().
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/20190927200538.660-11-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The old STOP sequence is: 1. stop DMA 2. send STOP ipc
If delay happen before the steps 1 and 2, the DMA buffer will be empty in
short time and cause pipeline xrun then stop the pipeline.
Then the step 2 ipc stop will return error as pipeline is already stopped.
Suggested change to avoid the issue is to switch the order of steps 1 and 2
for the stop sequence.
Signed-off-by: Pan Xiuli <xiuli.pan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-7-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Fix a bug in sof_pcm_hw_free() where some cleanup actions were
skipped if STREAM_PCM_FREE IPC was already successfully sent to
DSP when the stream was stopped or suspended. This is incorrect
as hw_free should clean up also other resources, including pcm
lib page allocations, period elapsed work queue and call to
platform hw_free.
Fixes: c29d96c3b9 ("ASoC: SOF: reset DMA state in prepare")
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/20190927200538.660-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
We are using sof_parse_word_tokens() to parse tokens with
bool/byte/short/word tuple types, here add the missing check, to fix the
parsing failure at byte/bool tuple types.
Signed-off-by: Keyon Jie <yang.jie@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
When we fail to boot the firmware, we encounter a kernel oops in
hda_dsp_get_registers(), which is called conditionally in
hda_dsp_dump() when the sdev_>boot_complete flag is set.
Setting this flag _after_ dumping the data fixes the issue and does
not change the programming flow.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927200538.660-2-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Fix potential NULL pointer dereference for alg_region in
wm_adsp_buffer_parse_legacy. In practice this can never happen as
loading the firmware should have failed at the wm_adsp2_setup_algs
stage, however probably better for the code to be robust against
future changes and this is more helpful for static analysis.
Signed-off-by: Li Xu <li.xu@cirrus.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20191001130911.19238-1-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Below Oops is caused in a system which uses ACPI instead of device node:
of_get_named_gpiod_flags: can't parse 'maxim,reset-gpio' property of node '(null)[0]'
BUG: kernel NULL pointer dereference, address: 0000000000000010
This patch avoids NULL pointer deferencing by adding a check before parsing
and initializes to make reset-gpio pin as invalid.
Signed-off-by: Sathyanarayana Nujella <sathyanarayana.nujella@intel.com>
Signed-off-by: Jairaj Arava <jairaj.arava@intel.com>
Link: https://lore.kernel.org/r/1569702150-11976-1-git-send-email-sathyanarayana.nujella@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
On stream stop currently we stop the DMA first followed by the CPU DAI.
This can cause underflow (playback) or overflow (capture) on the DAI side
as the DMA is no longer feeding data while the DAI is still active.
It can be observed easily if the DAI side does not have FIFO (or it is
disabled) to survive the time while the DMA is stopped, but still can
happen on relatively slow CPUs when relatively high sampling rate is used:
the FIFO is drained between the time the DMA is stopped and the DAI is
stopped.
It can only fixed by using different sequence within trigger for 'stop' and
'start':
case SNDRV_PCM_TRIGGER_START:
case SNDRV_PCM_TRIGGER_RESUME:
case SNDRV_PCM_TRIGGER_PAUSE_RELEASE:
Trigger order: dai_link, DMA, CPU DAI then the codec
case SNDRV_PCM_TRIGGER_STOP:
case SNDRV_PCM_TRIGGER_SUSPEND:
case SNDRV_PCM_TRIGGER_PAUSE_PUSH:
Trigger order: codec, CPU DAI, DMA then dai_link
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190927071646.22319-1-peter.ujfalusi@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
