There is no need to restore and restart PCM streams in case ADSP didn't
reach reset and power off state during system suspend/resume cycle. In that
case stream is still active but paused and firmware doesn't allow allocating
a new stream before paused stream is freed.

ADSP remains active in case suspend sequence didn't go to suspend_late
stage. This can happen when either suspend sequence is aborted by a wakeup
or by letting only devices suspend by "echo devices >/sys/power/pm_test".

Currently stream restoring fails in these suspend cases. Fix this by adding
a flag that indicates is complete stream reinitialization needed or is it
enough to resume paused stream. Flag is set when we know that ADSP reached
suspend_late.

Initial fix to this issue came from Fang Yang. I modified it a little and
forward ported it to top of two other suspend/resume patches from me.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Tested-by: NBorun Fu <borun.fu@intel.com>
Cc: yang fang <yang.a.fang@intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

Remove sst_byt_pcm_dev_resume() and move waiting of firmware boot into
sst_byt_pcm_dev_resume_early(). Now suspend_late and resume_early phases are
in sync with each other so that we know that ADSP was put into reset and was
unpowered after suspend_late and is ready to resume IO after resume_early
during resume stage in sst_byt_pcm_trigger().
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Tested-by: NBorun Fu <borun.fu@intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

Merge DSP reset and cleanup sequence in sst_byt_pcm_dev_suspend_noirq()
into sst_byt_pcm_dev_suspend_late(). First their order was wrong by first
unloading firmware modules in suspend_late and then taking DSP into reset
in suspend_noirq. Second ACPI has put device into OFF state already during
suspend_late so trying to reset the DSP is a no-op at suspend_noirq stage.

Fix these by moving DSP reset and cleanup into
sst_byt_pcm_dev_suspend_late() before firmware unloading.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Tested-by: NBorun Fu <borun.fu@intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

This change removes unsupported formats from System,
Capture and Loopback FE DAIs.
Also it fixes S24_LE support on all DAIs.
While at this fix 24 bit flag for BYT as well.
Signed-off-by: NJie Yang <yang.jie@intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>
Cc: stable@vger.kernel.org

Stored DSP DMA pointer must be cleared before starting the stream since
PCM pointer callback sst_byt_pcm_pointer() can be called before pointer is
updated. In that case last position of previous stream was wronly returned.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

In some cases the pcm stream is closed while context has been
scheduled to be restored, causing a null pointer deref panic.
Cancel work to ensure stream does not get freed while work is
still active/pending.

Also, restoring the pcm context can be safely skipped after the
stream has been stopped. Check if pcm stream is still running
before restoring stream context to help pending work finish
more quickly in stream close path.
Signed-off-by: NKevin Strasser <kevin.strasser@intel.com>
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

I managed to drop a change to stream ID setting from commit 49fee178
("ASoC: Intel: Only export one Baytrail DAI") leading to non-working
simultaneous capture-playback since after one DAI conversion
rtd->cpu_dai->id + 1 will be the same for both playback and capture.

Use substream->stream + 1 like it was in original Liam's patch.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

We don't need more than one DAI for Baytrail SST. Usage becomes also more
straightforward by grouping playback and capture streams under the same PCM
device.

[Jarkko: I made Liam's sst-baytrail-pcm.c change a few lines smaller and
squashed together with my byt-rt5640.c change]
Signed-off-by: NLiam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

Prepare for single Baytrail DAI playback/capture link by accessing PCM data
using stream ID instead of rtd->dev. Now rtd->dev is unique for playback
and capture since they are exported as separate DAIs but not once converted
to single DAI.

[Jarkko: Separated from another commit with updated commit log]
Signed-off-by: NLiam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

This follows the same idea than commit 10df3509
("ASoC: Intel: Fix Audio DSP usage when IOMMU is enabled.") by using only
ACPI device for all DMA allocations. Since DMA masking is already done in
firmware loading it can be removed from here.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

Add suspend and resume support to Baytrail SST DSP. This is implemented by
unloading firmware modules and putting DSP into reset prior suspend and
restarting DSP again in normal boot state after resume.

Context restore for running streams is implemented by scheduling a work from
sst_byt_pcm_trigger() that will allocate a stream with existing parameters
and start it from last known buffer position before suspend.

[Jarkko: Squashed together 5 WIP patches from Liam and 1 from me]
Signed-off-by: NLiam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

Stream start position will be needed in resume code. Prepare for it by
adding start offset argument to sst_byt_stream_start().
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

This is for preparing suspend/resume support but can give also more
safeguard against concurrent timestamp structure access between DSP firmware
and host.

Now DSP DMA pointer is sampled in each pcm pointer callback in
sst_byt_pcm_pointer() but that is unneeded since DSP updates the timestamp
period basis and can potentially be racy if sst_byt_pcm_pointer() is called
when DSP is updating the timestamp.

By taking DSP DMA pointer only after period elapsed IPC messages in
byt_notify_pointer() and returning stored hw pointer in
sst_byt_pcm_pointer() there is less risk for concurrent access.

The same stored hw pointer can be also used in suspend/resume code for
restarting the stream at the same position.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

This adds the Baytrail SST DSP PCM platform driver. It registers itself with
the ALSA SoC layer and uses Intel Baytrail SST DSP IPC for DSP control.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Acked-by: NLiam Girdwood <liam.r.girdwood@intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>