ASoC Samsung sub-drivers tried to get access to their DMA engine
controllers as a last step in driver probe. If a DMA engine was not
available yet, samsung_asoc_dma_platform_register() function ended in
-EPROBE_DEFER, but the driver already registered its component to ASoC
core. This patch moves samsung_asoc_dma_platform_register() call before
registering any components, to the common place, where driver was gathering
all needed resources.

In case of Samsung Exynos i2s driver the issue was even worse. The driver
managed already to register its secondary DAI platform device before
even getting the DMA engine access. That together with -EPROBE_DEFER error
code from samsung_i2s_probe() immediately triggered another round of
deferred probe retry and in turn endless loop of driver probing.

This patch fixes broken boot on Odroid XU3 and other Exynos5422-based
boards.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: NJavier Martinez Canillas <javier@osg.samsung.com>
Reviewed-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Add the missing clk_disable_unprepare() before return
from samsung_i2s_probe() in the error handling case.
Signed-off-by: NWei Yongjun <weiyj.lk@gmail.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

struct s3c_dma_params already includes struct snd_dmaengine_dai_dma_data,
there is no need for such an indirection so switch to using struct
snd_dmaengine_dai_dma_data instead of struct s3c_dma_params.  This also
allows us to use snd_soc_dai_init_dma_data() function instead of the
platform specific samsung_asoc_init_dma_data helper.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The DMA channel names are specified through struct snd_dmaengine_pcm_config
rather than using SND_DMAENGINE_PCM_FLAG_CUSTOM_CHANNEL_NAME flag when
booting with devicetree in order to properly support deferred probing.
Without this change the sound machine driver initialization can complete
successfully with unavailable DMA resources.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Ensure they secondary DAI device is freed properly when asoc_dma_platform
registration fails.  This change is needed for proper deferred probe support
and will help preventing situations when the CPU DAI's initialization
completes without required DMA resources.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Simplify the code a little by using a standard function for
getting the match data.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

samsung_dai_type_pri is not referenced anywhere so remove it.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

"samsung-i2sv4" identifier was previously used for the I2S device
of the S5PV210 SoCs, it can be removed now when s5pv210 is a dt-only
platform.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Lockdep warns of a potential lock inversion, i2s->lock is held numerous
times whilst we are under the substream lock (snd_pcm_stream_lock). If
we use the IRQ unsafe spin lock calls, you can also end up locking
snd_pcm_stream_lock whilst under i2s->lock (if an IRQ happens whilst we
are holding i2s->lock). This could result in deadlock.

[   18.147001]        CPU0                    CPU1
[   18.151509]        ----                    ----
[   18.156022]   lock(&(&pri_dai->spinlock)->rlock);
[   18.160701]                                local_irq_disable();
[   18.166622]                                lock(&(&substream->self_group.lock)->rlock);
[   18.174595]                                lock(&(&pri_dai->spinlock)->rlock);
[   18.181806]   <Interrupt>
[   18.184408]     lock(&(&substream->self_group.lock)->rlock);
[   18.190045]
[   18.190045]  *** DEADLOCK ***

This patch changes to using the irq safe spinlock calls, to avoid this
issue.

Fixes: ce8bcdbb ("ASoC: samsung: i2s: Protect more registers with a spinlock")
Signed-off-by: NCharles Keepax <ckeepax@opensource.wolfsonmicro.com>
Tested-by: NAnand Moon <linux.amoon@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org

As we are now passing the filter data as pointers to the drivers,
we can take the final step and also pass the filter function the
same way. I'm keeping this change separate, as there it's less
obvious that this is a net win.

Upsides of this are:

- The ASoC drivers are completely independent from the DMA engine
  implementation, which simplifies the Kconfig logic and in theory
  allows the same sound drivers to be built in a kernel that supports
  different kinds of dmaengine drivers.

- Consistency with other subsystems and drivers

On the other hand, we have a few downsides:

- The s3c24xx-dma driver now needs to be built-in for the ac97 platform
  device to be instantiated on s3c2440.

- samsung_dmaengine_pcm_config cannot be marked 'const' any more
  because the filter function pointer needs to be set at runtime.
  This is safe as long we don't have multiple different DMA engines
  in thet same system at runtime, but is nonetheless ugly.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Reviewed-by: NKrzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

ARM64 allmodconfig produces a bunch of warnings when building the
samsung ASoC code:

sound/soc/samsung/dmaengine.c: In function 'samsung_asoc_init_dma_data':
sound/soc/samsung/dmaengine.c:53:32: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
   playback_data->filter_data = (void *)playback->channel;
sound/soc/samsung/dmaengine.c:60:31: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
   capture_data->filter_data = (void *)capture->channel;

We could easily shut up the warning by adding an intermediate cast,
but there is a bigger underlying problem: The use of IORESOURCE_DMA
to pass data from platform code to device drivers is dubious to start
with, as what we really want is a pointer that can be passed into
a filter function.

Note that on s3c64xx, the pl08x DMA data is already a pointer, but
gets cast to resource_size_t so we can pass it as a resource, and it
then gets converted back to a pointer. In contrast, the data we pass
for s3c24xx is an index into a device specific table, and we artificially
convert that into a pointer for the filter function.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Reviewed-by: NKrzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The platform_device_id is not modified by the driver and core uses it as
const.
Signed-off-by: NKrzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

This patch adds clock provider (currently only for DT platforms) for
the CODECLKO (CDCLK) gate, RCLKSRC mux and RCLK pre-scaler divider
divider clock. Those all tree clock are only available in the IIS
Multi Audio Interface (I2S0), the regular IIS Bus Interface has only
CDCLK gate clock.

The motivation behind this patch is to expose the I2S internal clocks
which are currently controlled through set_sysclk() through the clk
API, so dedicated sound machine driver per each board can be avoided.

The intention is also to fix the CDCLK gating issue reported by
Daniel Drake:
http://mailman.alsa-project.org/pipermail/alsa-devel/2014-September/081753.html

This patch also reverts commit b97c60ab
("ASoC: samsung-i2s: Maintain CDCLK settings across i2s_{shutdown/
startup}") The problem that commit attempted to solve only affects
the Odroid X2/U3, which doesn't configure the CDCLK clock in
struct snd_soc_dai_ops hw_params callback and the issue should be
now resolved by using clk API, i.e. having the codec enabling/
disabling the CDCLK clock as required.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Ensure the I2SMOD, I2SPSR registers, which are also exposed through
clk API are only accessed with the i2s->spinlock spinlock held.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

It seems this driver hasn't been updated for SMP, as local_irq_save/
local_irq_restore don't provide proper protection of read/modify/write
of the device's registers on such systems. Introduce a spinlock
serializing access to the register region, it will be helpful later
when I2SMOD, I2SPSR registers are made also accessible through the
clk API.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The usage of this goto seems unjustified, use if/else statement instead.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The code to get pointer to the other DAI is repeated multiple
times. Add a helper function and use it instead.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The clk_prepare_enable() call on the "iis" clock is moved to happen earlier
in the DAI platform device driver's probe() callback, so the I2S registers
can be safely accessed through the clk API, after the clk supplier is
registered in the platform device probe().

After this patch the "iis" clock is kept enabled since the (primary) I2S
platform device probe() and until the platform device driver remove() call.
This is similar to gating the clock in the snd_soc_dai probe() and remove()
callbacks.
Normally, in addition to that we should mark the device as PM runtime active,
so if runtime PM is enabled it can idle the device by turning off the clock.
Correcting this issue is left for a separate patch series, as we need to
ensure the BUSCLK clock is always enabled when required.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Acquire the I2S interface clock in driver probe() callback
as it's a per-device not a per-DAI clock. While at it switch
to the resource managed clk_get().
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The memory mapped registers region is common for both DAIs so request
it in the I2S platform device driver's probe for the platform device
corresponding to the primary DAI, rather than in the ASoC DAI's probe
callback. While at it switch to devm_ioremap_resource(). This also
drops the hard coded (0x100) register region size in the driver.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

These functions may fail so let's properly report any errors.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Tidy up the samsung_i2s_get_driver_data() function by using
IS_ENABLE() instead of #ifdef and add missing braces for
the 'else' part. Also ensure we are not dereferencing NULL
'match' pointer.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The 'gpios' field in 'struct i2s' is now unused, this change
seems to be missing in commit 0429ffef
("ASoC: samsung: Remove obsolete GPIO based DT pinmuxing").
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

After commit b2b49ccb (PM: Kconfig: Set PM_RUNTIME if PM_SLEEP is
selected) PM_RUNTIME is always set if PM is set, so #ifdef blocks
depending on CONFIG_PM_RUNTIME may now be changed to depend on
CONFIG_PM.

Replace CONFIG_PM_RUNTIME with CONFIG_PM everywhere under sound/.
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: NTakashi Iwai <tiwai@suse.de>
Acked-by: NNicolin Chen <nicoleotsuka@gmail.com>
Acked-by: NBrian Austin <brian.austin@cirrus.com>
Acked-by: NMark Brown <broonie@kernel.org>

Add assignment of the variant_regs field which is missing in commit
a5a56871 ("ASoC: samsung: add support
for exynos7 I2S controller"). Without this attempting to probe the
secondary DAI fails with an error like:

[    1.763026] Unable to handle kernel NULL pointer dereference at virtual address 0000000c
[    1.780895] pgd = c0004000
[    1.783606] [0000000c] *pgd=00000000
[    1.838255] Internal error: Oops: 5 [#1] PREEMPT SMP ARM
[    1.843514] Modules linked in:
[    1.846558] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.18.0-rc1-00009-g5dcb01e-dirty #1521
[    1.854887] task: ee00a800 ti: ee088000 task.ti: ee088000
[    1.860284] PC is at i2s_txctrl+0x40/0x2d4
[    1.864350] LR is at i2s_txctrl+0x28/0x2d4
[    1.868428] pc : [<c036ffd4>]    lr : [<c036ffbc>]    psr: 60000153
[    1.868428] sp : ee089dc0  ip : 00000000  fp : ee21f000
[    1.879883] r10: 00000000  r9 : ee21fb00  r8 : c06406c4
[    1.885091] r7 : ee21fb00  r6 : 00000000  r5 : f00f6000  r4 : ed943410
[    1.891601] r3 : 0000016c  r2 : c0464550  r1 : c055cef8  r0 : ed943610
[    1.898113] Flags: nZCv  IRQs on  FIQs off  Mode SVC_32  ISA ARM  Segment kernel
[    1.905490] Control: 10c5387d  Table: 4000404a  DAC: 00000015
[    1.911218] Process swapper/0 (pid: 1, stack limit = 0xee088240)
[    1.917208] Stack: (0xee089dc0 to 0xee08a000)
...
[    2.068431] [<c036ffd4>] (i2s_txctrl) from [<c03719fc>] (samsung_i2s_dai_probe+0xb8/0x450)
[    2.076676] [<c03719fc>] (samsung_i2s_dai_probe) from [<c03607e0>] (snd_soc_register_card+0xd98/0x1348)
[    2.086044] [<c03607e0>] (snd_soc_register_card) from [<c03726e4>] (odroidx2_audio_probe+0xa8/0x11c)
[    2.095160] [<c03726e4>] (odroidx2_audio_probe) from [<c0249dd0>] (platform_drv_probe+0x48/0xa4)
[    2.103922] [<c0249dd0>] (platform_drv_probe) from [<c0248988>] (driver_probe_device+0x10c/0x22c)
[    2.112773] [<c0248988>] (driver_probe_device) from [<c0248b34>] (__driver_attach+0x8c/0x90)
[    2.121192] [<c0248b34>] (__driver_attach) from [<c02471c8>] (bus_for_each_dev+0x54/0x88)
[    2.129352] [<c02471c8>] (bus_for_each_dev) from [<c0248188>] (bus_add_driver+0xd4/0x1d0)
[    2.137510] [<c0248188>] (bus_add_driver) from [<c024915c>] (driver_register+0x78/0xf4)
[    2.145499] [<c024915c>] (driver_register) from [<c0008924>] (do_one_initcall+0x80/0x1b8)
[    2.153670] [<c0008924>] (do_one_initcall) from [<c05b7d40>] (kernel_init_freeable+0xfc/0x1c8)
[    2.162260] [<c05b7d40>] (kernel_init_freeable) from [<c04146c0>] (kernel_init+0x8/0xec)
[    2.170330] [<c04146c0>] (kernel_init) from [<c000e7f8>] (ret_from_fork+0x14/0x3c)
[    2.177873] Code: e5940000 e59f128c e59f228c e2800010 (e59c700c)
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The changes in commit a5a56871 (ASoC: samsung: add support for exynos7
I2S controller) introduce a new variant_regs structure in the driver data
which is now mandatory for accessing registers. Unfortunately this is only
hooked up for DT platforms so non-DT platforms like my primary development
platform for audio are broken by this change and crash on boot.

Since the only non-DT user of these device is s3c64xx fix this by making
the standard samsung-i2s device be of type I2Sv3 and add a new I2Sv4 name
to the platform data section, currently using the I2Sv5 information which
should be about right.
Signed-off-by: NMark Brown <broonie@kernel.org>

Return the error code we got from clk_get() and check to make sure that
clk_prepare_enable() worked.
Signed-off-by: NMark Brown <broonie@kernel.org>

In the i2s_set_sysclk() callback we are currently clearing all bits
of the IISMOD register in i2s_set_sysclk. It's due to an incorrect
mask used for the AND operation which is introduced in commit
a5a56871 (ASoC: samsung:
add support for exynos7 I2S controller) and also adds the missing
break statement.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NPadmavathi Venna <padma.v@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Exynos7 I2S controller has no internal dma, supports more
no. of root clock sampling frequencies and has more no.of Rx
fifos to support 7.1CH recording in TDM mode. Due to more no.
of root clock frequency values some of the bit offsets got
shifted up by one. Also I2S1 on previous Samsung platforms
uses v3 dai type but on Exynos7 it is upgraded to v5 with
slightly modified register offsets for supporting more no.of
RFS values. Due to the above changes, the driver has to be
modified to handle all versions of I2S controller. For this
I introduced a new structure to hold modified bit offsets and
masks which is passed as dai data.
Signed-off-by: NPadmavathi Venna <padma.v@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Internal DMA is available only on some of Samsung platforms.
So added a quirk for the same and made it optional.
Signed-off-by: NPadmavathi Venna <padma.v@samsung.com>
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>

In a couple of places the driver is missing a check to ensure there is a
secondary DAI before it de-references the pointer to it, causing a null
pointer de-reference. This patch adds a check to avoid this.
Signed-off-by: NCharles Keepax <ckeepax@opensource.wolfsonmicro.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@linaro.org>
Cc: stable@vger.kernel.org

The s3c_dma_client structures and the 'ch' and 'ops' members in
s3c_dma_params were only used by the legacy DMA driver and serve
no function any more. This removes any reference to them.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NMark Brown <broonie@linaro.org>

Currently configuration of the CDCLK pad is being overwritten in
the i2s_shutdown() callback in order to gate the SoC output clock.
However if an ASoC machine driver doesn't restore that clock
settings each time after opening the sound device this results
in the CDCLK pin being permanently configured into input mode.
I.e. the output clock will always stay disabled.
Fix that by saving the CDCLKCON bit state in i2s_shutdown() and
and restoring it in the i2s_startup() callback.
Signed-off-by: NChen Zhen <zhen1.chen@samsung.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

We should save/restore relevant I2S registers regardless of
the dai->active flag, otherwise some settings are being lost
after system suspend/resume cycle. E.g. I2S slave mode set only
during dai initialization is not preserved and the device ends
up in master mode after system resume.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@linaro.org>
Cc: stable@vger.kernel.org

commit 8c5178fc ("ALSA: Add params_width() helpers") introduces
a helper to get the sample width. Updating Samsung related sound
drivers to use this helper.
Signed-off-by: NTushar Behera <tushar.behera@linaro.org>
Signed-off-by: NMark Brown <broonie@linaro.org>

Ensure i2s->op_clk is not used when clk_get() for this clock fails.
This prevents working with an incorrectly configured clock in some
conditions.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

Replaced snd_soc_register_platform with devm_snd_soc_register_platform
in samsung_asoc_dma_platform_register(). This makes the function
samsung_asoc_dma_platform_unregister() redundant. This is removed and
all its users are updated.
Signed-off-by: NTushar Behera <tushar.behera@linaro.org>
Signed-off-by: NMark Brown <broonie@linaro.org>

This patch adds support for setting source clock of the "Core CLK"
of the IIS Multi Audio Interface.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

In exiting kernel, if DAIFMT flags are set in dai_link and I2S is
set to run in master mode, the I2S clocks are not getting configured
resulting in no output.

Existing code clears the current I2S clock settings during i2s_startup
and requires that the clocks are reconfigured. It then assumes that
sound-card driver would call snd_soc_dai_{set_sysclk/set_fmt} to
configure the root clock.

1. Since I2S clock settings remain fixed for a board, it would be better
to set the clocks once during sound-card probe.

2. Also if the DAIFMT flags are set in dai_link, snd_soc_dai_set_fmt is
called during DAI probe.

If both these conditions are true, then I2S clock remains unconfigured
during audio playback. Fix this by removing the code to clear
rclk_srcrate in i2s_startup. Instead, reset this during DAI probe.
Signed-off-by: NTushar Behera <tushar.behera@linaro.org>
Signed-off-by: NMark Brown <broonie@linaro.org>