Create the card debugfs directory at the begining of the initilization
rather then the end as various steps in the initilization sequence will try
to register files and sub-directories in the card directory.

Fixes: 4e2576bd ("ASoC: soc-core: initialize debugfs in snd_soc_instantiate_card()")
Reported-by: NFabio Estevam <festevam@gmail.com>
Reported-by: NNicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Current soc_init_card_debugfs() is called from snd_soc_register_card()
but, soc_cleanup_card_debugfs() is called from soc_cleanup_card_resources(),
not from paired function.

This differences don't matter for now. But if anyone wants to implement
a proper hotplug/unplug, this difference would become clearer.

Now, we can assume that snd_soc_instantiate_card() and
soc_cleanup_card_resources() are paired function.
soc_init_card_debugfs() / soc_cleanup_card_debugfs() paired function
should be called from these.
Signed-off-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Since commit d29697dc ("ASoC: Add sysfs entries via static attribute
groups") the sysfs attributes of the rtd are manged by the device core and
there is no need to manually call device_remove_file() anymore.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Any access to the component_list, codec_list and platform_list needs to be
properly locked by the client_mutex. Otherwise undefined behavior can occur
if the list is modified in one thread and concurrently accessed from another
thread.

This patch adds the missing locking to the debugfs file handlers that
display the registered components, as well as the various components
unregister functions.

Furthermore the client_lock is now held for the whole
snd_soc_instantiate_card() sequence to make sure that component removal does
not race against the card registration.
Reported-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Current snd_soc_unregister_card() indicates unregistered debug
message when it was called. But, it should be called only when
it was really unregistered.
Signed-off-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Instead of calling device_create_file() manually, assign the static
attribute group entries at the device registration.  This simplifies
the error handling and avoids the possible races.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

There have been some conflicting interpretations of how
snd_soc_dai_set_tdm_slot() is supposed to work. This patch updates the
documentation to be more specific on the exact semantics to avoid such
problems in the future.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Export snd_soc_runtime_set_dai_fmt() so it can be used in modules.

Fixes: ce64c8b9 ("ASoC: Add helper function for changing the DAI link format")
Reported-by: Nkbuild test robot <fengguang.wu@intel.com>
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

For some setups it is necessary to change the DAI link format at runtime.
This patch factors out the code that does the initial static DAI link format
configuration into a separate helper function which can be used board
drivers as well.

This allows board drivers that have to change the DAI link format at runtime
to reuse it instead of having to manually change the format on all DAIs.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

For legacy reasons the ASoC framework assumes that a CODEC INPUT or OUTPUT
widget that is not explicitly connected to a external source or sink is
potentially connected to a source or a sink and hence the framework treats
the widget itself as source (for INPUT) or sink (for OUTPUT). For this
reason a INPUT or OUTPUT widget that is really not connected needs to be
explicitly marked as so.

Setting the card's fully_routed flag will cause the ASoC core, once that all
widgets and routes have been registered, to go through the list of all
widgets and mark all INPUT and OUTPUT that are not externally connected as
non-connected. This essentially negates the default behaviour of treating
INPUT or OUTPUT widgets without external routes as sources or sinks.

This patch takes a different approach while getting the same result. Instead
of first marking INPUT and OUTPUT widgets as sinks/sources and then later
marking them as non-connected, just never mark them as a sink or a source if
the fully_routed flag is set on a card.

This requires a lot less code and also results in a slightly faster card
initialization since there is no need to iterate over all widgets and check
whether the INPUT and OUTPUT widgets are connected or not.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

This reverts commit f8781db8 (ASoC: dapm: Augment existing card
DAPM routes in snd_soc_of_parse_audio_routing) since it is broken for
deferred probing as it ends up storing data allocated with devm_ over
multiple instantiations of the device.
Reported-by: NRussell King <rmk+kernel@arm.linux.org.uk>
Tested-by: NRussell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: NMark Brown <broonie@kernel.org>

If a snd_soc_card has any DAPM routes when it calls
snd_soc_of_parse_audio_routing, those are clobbered without this change.
Signed-off-by: NPeter Rosin <peda@axentia.se>
Signed-off-by: NMark Brown <broonie@kernel.org>

If regmap is disabled there will be no users of the ASoC regmap helpers.
Furthermore regmap_exit() will no be defined causing the following compile
error:
	sound/soc/soc-core.c: In function 'snd_soc_component_exit_regmap':
	sound/soc/soc-core.c:2645:2: error: implicit declaration of function
		'regmap_exit' [-Werror=implicit-function-declaration]

So disable the helpers if regmap is disabled.
Reported-by: Nkbuild test robot <fengguang.wu@intel.com>
Fixes: 20feb881 ASoC: Add helper functions for deferred regmap setup")
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

This patch exports a core function which handles the DT description
of multi-codec links (as: "sound-dai = <&hdmi 0>, <&spdif_codec>;")
and creates a CODEC component array in the DAI link.
Signed-off-by: NJean-Francois Moine <moinejf@free.fr>
Signed-off-by: NMark Brown <broonie@kernel.org>

Some drivers (most notably the AC'97 drivers) do not have access to their
regmap struct when the component/codec is registered. For those drivers the
automatic regmap setup will not work and needs to be done manually,
typically from the component/CODEC drivers probe callback.

This patch adds a set of helper function to handle deferred regmap
initialization as well as early regmap tear-down.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Setting the ac97_control field on a CPU DAI tells the ASoC core that this
DAI in addition to audio data also transports control data to the CODEC.
This causes the core to suspend the DAI after the CODEC and resume it before
the CODEC so communication to the CODEC is still possible. This is not
necessarily something that is specific to AC'97 and can be used by other
buses with the same requirement. This patch renames the flag from
ac97_control to bus_control to make this explicit.

While we are at it also change the type from int to bool.

The following semantich patch was used for automatic conversion of the
drivers:
// <smpl>
@@
identifier drv;
@@
struct snd_soc_dai_driver drv = {
-	.ac97_control
+	.bus_control
	=
-	1
+	true
};
// </smpl>
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

We have all the information and dependencies we need to initialize and
register the device available in snd_soc_new_ac97_codec(). So there is no
need to delay the device registration until after the card itself as been
registered.

This makes the code significantly simpler and also makes it possible to use
the AC'97 device in the CODECs probe function. The later will be required to
be able to convert the AC'97 CODEC drivers to regmap.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

This has no users since commit f0fba2ad ("ASoC: multi-component - ASoC
Multi-Component Support") which was almost 5 years ago. Given that this runs
after CODEC probe functions have been run it also doesn't seem to be that
useful.

So drop it altogether to make the code simpler.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently the AC'97 support is splattered all throughout soc-core.c. Some
parts are #ifdef'd some parts are not. This patch moves the AC'97 support to
its own file, this should make the code a bit more clearer and also makes it
possible to easily not compile it into the kernel when not needed.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

The memory that a struct device is contained in must not be freed except
from within the device's release callback. The ASoC code currently does not
adhere to this rule for the AC'97 device. This patch fixes it by moving the
freeing of the AC'97 to the release callback and splitting up the
registration and unregistration of the device into separate steps for
getting/putting the reference to the device and adding/removing it to the
device hierarchy.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

The CODEC mutex is now unused and can be removed.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

This function has no more non regmap user, which means we can remove the
implementation of the function and associated functions and structure
fields.

For convenience we keep a static inline version of the function that
forwards calls to regcache_sync() unconditionally.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

CPU and CODEC DAI probe are performed in exactly the same way. Which means
we can reuse the snd_soc_codec_dai_probe() for probing CPU DAIs as well.

While we are at it also drop the unused card parameter form the function.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

The card field of the snd_soc_dai field is very rarely used. We can use
dai->component->card instead and remove the card field from the snd_soc_dai
struct.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Typically a DAI does not need direct access to the platform. Currently the
only user of this field is in a platform driver where we have a more direct
way of getting a pointer to the platform. This patch updates the driver to
use the more direct way and then removes the platform field from the
snd_soc_dai struct.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Those are unused and new drivers should use device driver suspend/resume.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Commit 14621c7e ("ASoC: Consolidate CPU and CODEC DAI lookup")
consolidated the lookup of CPU DAIs and CODEC DAIs into a single function.
When matching a component by name for CODEC DAIs the code previous to the
patch compared the name in the DAI link table with component->name. For CPU
DAIs the code compared to dev_name(component->dev). The newly introduced
function ended up using the later as well.

For most components dev_name(component->dev) and component->name are the
same. The main notable exception are I2C devices where the driver name and
the device name are concatenated to form the component name. By using
dev_name(component->dev) instead of component->name the patch broke the
matching of I2C CODECs by name.

This patch restores the original behavior by using component->name instead
of dev_name(component->dev). This will be safe even for CPU DAIs since for
CPU DAIs both are the same.

Fixes: 14621c7e ("ASoC: Consolidate CPU and CODEC DAI lookup")
Reported-by: NDmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

The main ASoC source file is getting quite large and the standard ops don't
really have anything to do with the rest of the file so split them out into
a separate file.
Signed-off-by: NMark Brown <broonie@kernel.org>

Commit f227b88f ("ASoC: core: Add signed register volume control logic")
added support for signed control to the generic volsw control handler.
This makes it possible to use them for the S8 control as well, rather than
having to use a custom control handler implementation.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

The state of endpoint widgets is affected by that card's power state.
Endpoint widgets that do no have the ignore_suspend flag set will be
considered inactive during suspend. So they have to be re-checked and marked
dirty after the card's power state changes. Currently the input and output
widgets are marked dirty instead, this works most of the time since
typically a path from one endpoint to another will go via a input or output
widget. But marking the endpoints dirty is technically more correct and will
also work for odd corner cases.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

... and reduce the open codes.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

A platform_driver does not need to set an owner, it will be populated by the
driver core.
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

Coverity spotted an use-after-free condition in snd_soc_remove_platform().
Fix this by moving snd_soc_component_cleanup() after the debug print
statement which uses the component's string.
Signed-off-by: NDaniel Mack <daniel@zonque.org>
Signed-off-by: NMark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org

The only remaining user of the CODEC pointer in the DAPM struct is to
initialize the CODEC pointer in the widget struct. The later is scheduled
for removal, but has still a few users left. For now use
dapm->component->codec to initialize it.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Since we cannot make sure the 'params->num_regs' will always be none
zero here, and then if it equals to zero, the kmemdup() will return
ZERO_SIZE_PTR, which equals to ((void *)16).

So this patch fix this with just doing the zero check before calling
kmemdup().
Signed-off-by: NXiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org

This fixes a bug when using the SOC_DOUBLE_R_RANGE_TLV macro in
the invert mode. In the non-invert case, e.g.

SOC_DOUBLE_R_RANGE_TLV("<name>", <reg_l>, <reg_r>,
					0, 40, 255, 0, <tlv>)

the range sent to the hardware is 40..255, but in the invert case:

SOC_DOUBLE_R_RANGE_TLV("<name>", <reg_l>, <reg_r>,
					0, 40, 255, 1, <tlv>)

the range 215..0 was being sent to the hardware. This commit
corrects this to 255..40 so it is consistent with the non-invert
case.
Signed-off-by: NHoward Mitchell <hm@hmbedded.co.uk>
Signed-off-by: NMark Brown <broonie@kernel.org>

Commit e3f205a7 ("ASoC: Remove locking in
snd_soc_{new,free}_ac97_codec()") overlooked a unlock on one of the error
paths.

Fixes: e3f205a7 ("ASoC: Remove locking in snd_soc_{new,free}_ac97_codec()")
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

snd_soc_new_ac97_codec() and snd_soc_free_ac97_codec() are called from
within a CODEC's probe() and remove() callbacks. Those will not run
concurrently against each other for the same CODEC instance, hence it is not
necessary to protect the two functions with a mutex.

This removes the last user in the ASoC core of the snd_soc_codec mutex field
and will allow us to eventually remove it.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

There are no users left and new users should rather use the component_driver
struct embedded in the snd_soc_platform_driver struct to do this. E.g.:

static const struct snd_soc_platform_driver foobar_driver = {
	.component_driver = {
		.dapm_widgets = ...,
		.num_dapm_widgets = ...,
		...,
	},
	...
};

instead of

static const struct snd_soc_platform_driver foobar_driver = {
	.dapm_widgets = ...,
	.num_dapm_widgets = ...,
	...
};

This also allows us to remove the steal_sibling_dai_widgets hack.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@linaro.org>

We do a bit more than just running the callbacks during suspend and resume
these days (e.g. call regcache_mark_dirty() during suspend). But this is
only when suspend and resume callbacks are specified for the driver,
otherwise nothing is done. This means that drivers which don't want to do
anything special during suspend and resume, but still want the standard
operations to run, need to provide empty suspend and resume callback
functions (rather than no callbacks). This patch updates the suspend and
resume code to always run standard sequence regardless of whether suspend
and resume handlers are provided.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>