linux.git/sound/soc/sof/ipc.c, branch v5.16

ASoC: SOF: prepare code to allocate IPC messages in fw_ready

2021-10-08T12:51:16+00:00

The fixed maximum size of IPC message does not allow for large
transfers, e.g. for filter data. Currently such messages will
be divided into smaller pieces and sent to firmware in multiple
chunks. For future IPC, this strategy is not suitable.

The maximum IPC message size is limited by host box size which
can be known when firmware is ready, so the fw_ready callback
can allocate IPC messages with platform-specific sizes instead
of the current fixed-size.

To be compatible with released firmware, current platforms will
still use SOF_IPC_MSG_MAX_SIZE. For future platforms, there will
be a new fw_ready function and the platform-specific allocation
will take place there.

Signed-off-by: Rander Wang 
Reviewed-by: Bard Liao 
Reviewed-by: Ranjani Sridharan 
Reviewed-by: Pierre-Louis Bossart 
Signed-off-by: Peter Ujfalusi 
Message-Id: <20211008093836.28210-1-peter.ujfalusi@linux.intel.com>
Signed-off-by: Mark Brown

Merge series "Introduce new SOF helpers" from Daniel Baluta

2021-10-07T21:24:58+00:00

Daniel Baluta :

From: Daniel Baluta 

This patchseries adds new helpers in order to reduce code duplication
and prepare for compress audio support with SOF.

Bud Liviu-Alexandru (1):
  ASoC: SOF: Make Intel IPC stream ops generic

Daniel Baluta (3):
  ASoC: SOF: Introduce snd_sof_mailbox_read / snd_sof_mailbox_write
    callbacks
  ASoC: SOF: imx: Use newly introduced generic IPC stream ops
  ASoC: SOF: Introduce fragment elapsed notification API

 sound/soc/sof/Kconfig                         |  4 ++
 sound/soc/sof/Makefile                        |  3 +-
 sound/soc/sof/compress.c                      | 51 +++++++++++++++++++
 sound/soc/sof/imx/Kconfig                     |  2 +
 sound/soc/sof/imx/imx8.c                      | 39 +++++++-------
 sound/soc/sof/imx/imx8m.c                     | 26 ++++------
 sound/soc/sof/intel/Makefile                  |  3 --
 sound/soc/sof/intel/apl.c                     |  4 ++
 sound/soc/sof/intel/bdw.c                     | 12 +++--
 sound/soc/sof/intel/byt.c                     | 24 ++++++---
 sound/soc/sof/intel/cnl.c                     |  4 ++
 sound/soc/sof/intel/icl.c                     |  4 ++
 sound/soc/sof/intel/pci-tng.c                 | 12 +++--
 sound/soc/sof/intel/tgl.c                     |  4 ++
 sound/soc/sof/ipc.c                           |  6 ++-
 sound/soc/sof/ops.h                           | 15 ++++++
 sound/soc/sof/pcm.c                           |  7 ++-
 sound/soc/sof/sof-audio.h                     | 11 +++-
 sound/soc/sof/sof-priv.h                      | 28 ++++++----
 .../sof/{intel/intel-ipc.c => stream-ipc.c}   | 50 +++++++++---------
 sound/soc/sof/topology.c                      |  6 ++-
 21 files changed, 218 insertions(+), 97 deletions(-)
 create mode 100644 sound/soc/sof/compress.c
 rename sound/soc/sof/{intel/intel-ipc.c => stream-ipc.c} (58%)

--
2.27.0

ASoC: SOF: Introduce fragment elapsed notification API

2021-10-07T15:57:57+00:00

This patch prepares the introduction of the compress API with SOF.

After each fragment is accepted by the DSP we need to inform
the userspace applications that they can send the next fragment.
This is done via snd_compr_fragment_elapsed.

Similar with the PCM case, in order to avoid sending an IPC before
the previous IPC is handled we need to schedule a delayed work to
call snd_compr_fragment_elapsed().

See snd_sof_pcm_period_elapsed.

To sum up this patch offers the following API to SOF code:
	* snd_sof_compr_init_elapsed_work
	* snd_sof_compr_fragment_elapsed

Note that implementation for compressed function is in a new file
selected via CONFIG_SND_SOC_SOF_COMPRESS invisible config option.
This option is automatically selected for platforms that support
the compress interface. For now only i.MX8 platforms support this.

For symmetry we introduce snd_sof_pcm_init_elapsed_work to setup
the work struct for PCM case.

Signed-off-by: Daniel Baluta 
Signed-off-by: Bud Liviu-Alexandru 
Reviewed-by: Pierre-Louis Bossart 
Reviewed-by: Paul Olaru 
Reviewed-by: Péter Ujfalusi 
Link: https://lore.kernel.org/r/20211004152147.1268978-5-daniel.baluta@oss.nxp.com
Signed-off-by: Mark Brown

ASoC: SOF: Introduce macro to set the firmware state

2021-10-07T14:45:36+00:00

Add sof_set_fw_state() macro to wrap the sdev->fw_state management to allow
actions to be taken when certain state is set or when state is changing.

Signed-off-by: Peter Ujfalusi 
Reviewed-by: Pierre-Louis Bossart 
Reviewed-by: Kai Vehmanen 
Link: https://lore.kernel.org/r/20211006110645.26679-15-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown

ASoC: SOF: ipc: Re-enable dumps after successful IPC tx

2021-10-07T14:45:34+00:00

The dumps are silenced after an IPC tx timeout by default.
The IPC timeout can indicate severe error (firmware crash) or in some cases
it is less devastating and the firmware remains operational, the timeout
was due to a scheduling spike or other anomaly.

In any case consequent IPC timeouts will not print dumps but if any IPC do
succeed than we should re-enable the dumps to print dumps the next time
a timeout might happen.

Signed-off-by: Peter Ujfalusi 
Reviewed-by: Pierre-Louis Bossart 
Reviewed-by: Kai Vehmanen 
Link: https://lore.kernel.org/r/20211006110645.26679-13-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown

ASoC: SOF: Introduce widget use_count

2021-10-01T19:48:28+00:00

Add a new field, use_count to struct snd_sof_widget to keep track
of the usage count for each widget. Since widgets can belong to
multiple pipelines, this field will ensure that the widget
is setup only when the first pipeline that needs it is started
and freed when the last pipeline that needs it is stopped. There is
no need to protect the widget use_count access as the core already
handles mutual exclusion at the PCM level.
Add a new helper sof_widget_free() to handle freeing the SOF
widgets and export the sof_widget_setup/free() functions.

Signed-off-by: Ranjani Sridharan 
Reviewed-by: Guennadi Liakhovetski 
Reviewed-by: Pierre-Louis Bossart 
Reviewed-by: Kai Vehmanen 
Signed-off-by: Daniel Baluta 
Signed-off-by: Peter Ujfalusi 
Link: https://lore.kernel.org/r/20210927120517.20505-10-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown

ASoC: SOF: ipc: Make the error prints consistent in tx_wait_done()

2021-09-28T12:16:28+00:00

If we get an error on reply (msg->reply_error) then we should print the
error value out.

At the same time extend the print to include the message size as well and
do the same in case of a timeout.

Signed-off-by: Peter Ujfalusi 
Reviewed-by: Daniel Baluta 
Reviewed-by: Paul Olaru 
Reviewed-by: Pierre-Louis Bossart 
Link: https://lore.kernel.org/r/20210928073615.29574-1-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown

ASoC: SOF: add error handling to snd_sof_ipc_msg_data()

2021-09-28T12:16:26+00:00

If an invalid stream is passed to snd_sof_ipc_msg_data() it won't
fill the provided object with data. The caller has to be able to
recognise such cases to avoid handling invalid data. Make the
function return an error when failing.

Signed-off-by: Guennadi Liakhovetski 
Reviewed-by: Peter Ujfalusi 
Reviewed-by: Kai Vehmanen 
Reviewed-by: Ranjani Sridharan 
Signed-off-by: Peter Ujfalusi 
Link: https://lore.kernel.org/r/20210928103516.8066-1-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown

Merge series "ASoC: SOF: ipc: Small cleanups for message handler functions" from Peter Ujfalusi :

2021-09-17T13:56:43+00:00

Hi,

Rename the parameter for ipc_trace_message() to match it's content and use
%#x" for hexadecimal prints in remaining places.

Regards,
Peter
---
Peter Ujfalusi (2):
  ASoC: SOF: ipc: Clarify the parameter name for ipc_trace_message()
  ASoC: SOF: ipc: Print 0x prefix for errors in
    ipc_trace/stream_message()

 sound/soc/sof/ipc.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

--
2.33.0

ASoC: SOF: ipc: Remove redundant error check from sof_ipc_tx_message_unlocked

2021-09-17T12:17:45+00:00

If the snd_sof_dsp_send_msg() failed then we have already returned from
sof_ipc_tx_message_unlocked() with the error message.

There is no need to check if ret is really 0 after this and we can return
directly the return value from tx_wait_done()

At the same time make the remaining checks for error (ret) to match.

Signed-off-by: Peter Ujfalusi 
Reviewed-by: Guennadi Liakhovetski 
Reviewed-by: Kai Vehmanen 
Reviewed-by: Pierre-Louis Bossart 
Link: https://lore.kernel.org/r/20210916125725.25934-1-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown