ASoC: SOF: ipc4-pcm: Do not reset ChainDMA if it is not allocated
Subsystems (3)
Peter Ujfalusi
maintainercover ASoC: SOF: ipc4-pcm: Do not reset ChainDMA if it is not allocated
Hi,
The current code will reset the ChainDMA on release unconditionally which can result the following error when the CHainDMA is not allocated:
ipc tx : 0xe040000|0x0: GLB_CHAIN_DMA
ipc tx reply: 0x2e000007|0x0: GLB_CHAIN_DMA FW reported error: 7 - Unsupported operation requested ipc error for msg 0xe040000|0x0sof_pcm_stream_free: pcm_ops hw_free failed -22
Background:
Pulseaudio and Pipewire on startup opens all available streams and closes them without triggering a start (after probing it's capabilities).
Regards,
Peter
---
Peter Ujfalusi (4):
ASoC: SOF: ipc4-pcm: Use consistent name for snd_sof_pcm_stream
pointer
ASoC: SOF: ipc4-pcm: Use consistent name for sof_ipc4_timestamp_info
pointer
ASoC: SOF: ipc4-pcm: Introduce generic sof_ipc4_pcm_stream_priv
ASoC: SOF: ipc4-pcm: Do not reset the ChainDMA if it has not been
allocated
sound/soc/sof/ipc4-pcm.c | 115 +++++++++++++++++++++++++++------------ 1 file changed, 79 insertions(+), 36 deletions(-)
Signature · 1 line
2.44.0
Peter Ujfalusi
maintainerPATCH 1/4 ASoC: SOF: ipc4-pcm: Use consistent name for snd_sof_pcm_stream pointer
Throughout the file the pointer for snd_sof_pcm_stream is named either 'stream' or (wrongly) 'spcm' which confuses the reader.
Use 'sps' for the pointer name as it is the most common name used in SOF codebase.
--- sound/soc/sof/ipc4-pcm.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-)
sound/soc/sof/ipc4-pcm.c +11 -11
@@ -764,7 +764,7 @@ static int sof_ipc4_pcm_setup(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcmreturn 0 ; } -static void sof_ipc4_build_time_info (struct snd_sof_dev *sdev,struct snd_sof_pcm_stream *spcm) +static void sof_ipc4_build_time_info (struct snd_sof_dev *sdev,struct snd_sof_pcm_stream *sps) {struct sof_ipc4_copier *host_copier =NULL ;struct sof_ipc4_copier *dai_copier =NULL ;
@@ -775,7 +775,7 @@ static void sof_ipc4_build_time_info(struct snd_sof_dev *sdev, struct snd_sof_pcint i;/* find host & dai to locate info in memory window */ -for_each_dapm_widgets (spcm->list, i, widget) { +for_each_dapm_widgets (sps->list, i, widget) {struct snd_sof_widget *swidget = widget->dobj.private;if (!swidget)
@@ -795,7 +795,7 @@ static void sof_ipc4_build_time_info(struct snd_sof_dev *sdev, struct snd_sof_pcreturn ; } - info = spcm->private; + info = sps->private; info->host_copier = host_copier; info->dai_copier = dai_copier; info->llp_offset = offsetof(struct sof_ipc4_fw_registers, llp_gpdma_reading_slots) +
@@ -864,7 +864,7 @@ static int sof_ipc4_pcm_hw_params(struct snd_soc_component *component,static int sof_ipc4_get_stream_start_offset (struct snd_sof_dev *sdev,struct snd_pcm_substream *substream, -struct snd_sof_pcm_stream *stream, +struct snd_sof_pcm_stream *sps,struct sof_ipc4_timestamp_info *time_info) {struct sof_ipc4_copier *host_copier = time_info->host_copier;
@@ -918,7 +918,7 @@ static int sof_ipc4_pcm_pointer(struct snd_soc_component *component,struct sof_ipc4_timestamp_info *time_info;struct sof_ipc4_llp_reading_slot llp;snd_pcm_uframes_t head_cnt, tail_cnt; -struct snd_sof_pcm_stream *stream; +struct snd_sof_pcm_stream *sps;u64 dai_cnt, host_cnt, host_ptr;struct snd_sof_pcm *spcm;int ret;
@@ -927,8 +927,8 @@ static int sof_ipc4_pcm_pointer(struct snd_soc_component *component,if (!spcm)return -EOPNOTSUPP ; - stream = &spcm->stream[substream->stream]; - time_info = stream->private; + sps = &spcm->stream[substream->stream]; + time_info = sps->private;if (!time_info)return -EOPNOTSUPP ;
@@ -938,7 +938,7 @@ static int sof_ipc4_pcm_pointer(struct snd_soc_component *component,* the statistics is complete. And it will not change after the first initiailization. */ if (time_info->stream_start_offset ==SOF_IPC4_INVALID_STREAM_POSITION ) { - ret =sof_ipc4_get_stream_start_offset (sdev, substream, stream, time_info); + ret =sof_ipc4_get_stream_start_offset (sdev, substream, sps, time_info);if (ret <0 )return -EOPNOTSUPP ; }
@@ -1030,15 +1030,15 @@ static snd_pcm_sframes_t sof_ipc4_pcm_delay(struct snd_soc_component *component, {struct snd_soc_pcm_runtime *rtd =snd_soc_substream_to_rtd (substream);struct sof_ipc4_timestamp_info *time_info; -struct snd_sof_pcm_stream *stream; +struct snd_sof_pcm_stream *sps;struct snd_sof_pcm *spcm; spcm =snd_sof_find_spcm_dai (component, rtd);if (!spcm)return 0 ; - stream = &spcm->stream[substream->stream]; - time_info = stream->private; + sps = &spcm->stream[substream->stream]; + time_info = sps->private;/* * Report the stored delay value calculated in the pointer callback. * In the unlikely event that the calculation was skipped/aborted, the
Signature · 1 line
2.44.0
Peter Ujfalusi
maintainerPATCH 2/4 ASoC: SOF: ipc4-pcm: Use consistent name for sof_ipc4_timestamp_info pointer
In reply to Peter UjfalusiThe pointer to sof_ipc4_timestamp_info named most of the time as 'time_info' only to be named as 'stream_info' or 'info' in two function.
Use the consistent name of 'time_info' throughout the file.
--- sound/soc/sof/ipc4-pcm.c | 40 ++++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-)
sound/soc/sof/ipc4-pcm.c +20 -20
@@ -721,7 +721,7 @@ static int sof_ipc4_pcm_setup(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcm {struct snd_sof_pcm_stream_pipeline_list *pipeline_list;struct sof_ipc4_fw_data *ipc4_data = sdev->private; -struct sof_ipc4_timestamp_info *stream_info; +struct sof_ipc4_timestamp_info *time_info;bool support_info = true;u32 abi_version;u32 abi_offset;
@@ -752,13 +752,13 @@ static int sof_ipc4_pcm_setup(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcmif (!support_info)continue ; - stream_info =kzalloc (sizeof (*stream_info),GFP_KERNEL ); -if (!stream_info) { + time_info =kzalloc (sizeof (*time_info),GFP_KERNEL ); +if (!time_info) {sof_ipc4_pcm_free (sdev, spcm);return -ENOMEM ; } - spcm->stream[stream].private = stream_info; + spcm->stream[stream].private = time_info; }return 0 ;
@@ -769,7 +769,7 @@ static void sof_ipc4_build_time_info(struct snd_sof_dev *sdev, struct snd_sof_pcstruct sof_ipc4_copier *host_copier =NULL ;struct sof_ipc4_copier *dai_copier =NULL ;struct sof_ipc4_llp_reading_slot llp_slot; -struct sof_ipc4_timestamp_info *info; +struct sof_ipc4_timestamp_info *time_info;struct snd_soc_dapm_widget *widget;struct snd_sof_dai *dai;int i;
@@ -795,44 +795,44 @@ static void sof_ipc4_build_time_info(struct snd_sof_dev *sdev, struct snd_sof_pcreturn ; } - info = sps->private; - info->host_copier = host_copier; - info->dai_copier = dai_copier; - info->llp_offset = offsetof(struct sof_ipc4_fw_registers , llp_gpdma_reading_slots) + - sdev->fw_info_box.offset; + time_info = sps->private; + time_info->host_copier = host_copier; + time_info->dai_copier = dai_copier; + time_info->llp_offset = offsetof(struct sof_ipc4_fw_registers , + llp_gpdma_reading_slots) + sdev->fw_info_box.offset;/* find llp slot used by current dai */ for (i =0 ; i <SOF_IPC4_MAX_LLP_GPDMA_READING_SLOTS ; i++) { -sof_mailbox_read (sdev, info->llp_offset, &llp_slot,sizeof (llp_slot)); +sof_mailbox_read (sdev, time_info->llp_offset, &llp_slot,sizeof (llp_slot));if (llp_slot.node_id == dai_copier->data.gtw_cfg.node_id)break ; - info->llp_offset +=sizeof (llp_slot); + time_info->llp_offset +=sizeof (llp_slot); }if (i <SOF_IPC4_MAX_LLP_GPDMA_READING_SLOTS )return ;/* if no llp gpdma slot is used, check aggregated sdw slot */ - info->llp_offset = offsetof(struct sof_ipc4_fw_registers , llp_sndw_reading_slots) + - sdev->fw_info_box.offset; + time_info->llp_offset = offsetof(struct sof_ipc4_fw_registers , + llp_sndw_reading_slots) + sdev->fw_info_box.offset;for (i =0 ; i <SOF_IPC4_MAX_LLP_SNDW_READING_SLOTS ; i++) { -sof_mailbox_read (sdev, info->llp_offset, &llp_slot,sizeof (llp_slot)); +sof_mailbox_read (sdev, time_info->llp_offset, &llp_slot,sizeof (llp_slot));if (llp_slot.node_id == dai_copier->data.gtw_cfg.node_id)break ; - info->llp_offset +=sizeof (llp_slot); + time_info->llp_offset +=sizeof (llp_slot); }if (i <SOF_IPC4_MAX_LLP_SNDW_READING_SLOTS )return ;/* check EVAD slot */ - info->llp_offset = offsetof(struct sof_ipc4_fw_registers , llp_evad_reading_slot) + - sdev->fw_info_box.offset; -sof_mailbox_read (sdev, info->llp_offset, &llp_slot,sizeof (llp_slot)); + time_info->llp_offset = offsetof(struct sof_ipc4_fw_registers , + llp_evad_reading_slot) + sdev->fw_info_box.offset; +sof_mailbox_read (sdev, time_info->llp_offset, &llp_slot,sizeof (llp_slot));if (llp_slot.node_id != dai_copier->data.gtw_cfg.node_id) - info->llp_offset =0 ; + time_info->llp_offset =0 ; } static int sof_ipc4_pcm_hw_params(struct snd_soc_component *component,
Signature · 1 line
2.44.0
Peter Ujfalusi
maintainerPATCH 3/4 ASoC: SOF: ipc4-pcm: Introduce generic sof_ipc4_pcm_stream_priv
In reply to Peter UjfalusiUsing the sof_ipc4_timestamp_info struct directly as sps->private data is too restrictive, add a new generic sof_ipc4_pcm_stream_priv struct containing the time_info to allow new information to be stored in a generic way.
--- sound/soc/sof/ipc4-pcm.c | 43 ++++++++++++++++++++++++++++++++-------- 1 file changed, 35 insertions(+), 8 deletions(-)
sound/soc/sof/ipc4-pcm.c +35 -8
@@ -37,6 +37,22 @@ struct sof_ipc4_timestamp_info {snd_pcm_sframes_t delay; }; +/** +* struct sof_ipc4_pcm_stream_priv - IPC4 specific private data +* @time_info: pointer to time info struct if it is supported, otherwise NULL +*/ +struct sof_ipc4_pcm_stream_priv { +struct sof_ipc4_timestamp_info *time_info; +}; + +static inline struct sof_ipc4_timestamp_info * +sof_ipc4_sps_to_time_info (struct snd_sof_pcm_stream *sps) +{ +struct sof_ipc4_pcm_stream_priv *stream_priv = sps->private; + +return stream_priv->time_info; +} +static int sof_ipc4_set_multi_pipeline_state (struct snd_sof_dev *sdev,u32 state,struct ipc4_pipeline_set_state_data *trigger_list) {
@@ -452,7 +468,7 @@ static int sof_ipc4_trigger_pipelines(struct snd_soc_component *component,* Invalidate the stream_start_offset to make sure that it is * going to be updated if the stream resumes */ - time_info = spcm->stream[substream->stream].private; + time_info =sof_ipc4_sps_to_time_info (&spcm->stream[substream->stream]);if (time_info) time_info->stream_start_offset =SOF_IPC4_INVALID_STREAM_POSITION ;
@@ -706,12 +722,16 @@ static int sof_ipc4_pcm_dai_link_fixup(struct snd_soc_pcm_runtime *rtd,static void sof_ipc4_pcm_free (struct snd_sof_dev *sdev,struct snd_sof_pcm *spcm) {struct snd_sof_pcm_stream_pipeline_list *pipeline_list; +struct sof_ipc4_pcm_stream_priv *stream_priv;int stream;for_each_pcm_streams (stream) { pipeline_list = &spcm->stream[stream].pipeline_list;kfree (pipeline_list->pipelines); pipeline_list->pipelines =NULL ; + + stream_priv = spcm->stream[stream].private; +kfree (stream_priv->time_info);kfree (spcm->stream[stream].private); spcm->stream[stream].private =NULL ; }
@@ -721,6 +741,7 @@ static int sof_ipc4_pcm_setup(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcm {struct snd_sof_pcm_stream_pipeline_list *pipeline_list;struct sof_ipc4_fw_data *ipc4_data = sdev->private; +struct sof_ipc4_pcm_stream_priv *stream_priv;struct sof_ipc4_timestamp_info *time_info;bool support_info = true;u32 abi_version;
@@ -749,6 +770,14 @@ static int sof_ipc4_pcm_setup(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcmreturn -ENOMEM ; } + stream_priv =kzalloc (sizeof (*stream_priv),GFP_KERNEL ); +if (!stream_priv) { +sof_ipc4_pcm_free (sdev, spcm); +return -ENOMEM ; + } + + spcm->stream[stream].private = stream_priv; +if (!support_info)continue ;
@@ -758,7 +787,7 @@ static int sof_ipc4_pcm_setup(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcmreturn -ENOMEM ; } - spcm->stream[stream].private = time_info; + stream_priv->time_info = time_info; }return 0 ;
@@ -795,7 +824,7 @@ static void sof_ipc4_build_time_info(struct snd_sof_dev *sdev, struct snd_sof_pcreturn ; } - time_info = sps->private; + time_info =sof_ipc4_sps_to_time_info (sps); time_info->host_copier = host_copier; time_info->dai_copier = dai_copier; time_info->llp_offset = offsetof(struct sof_ipc4_fw_registers,
@@ -849,7 +878,7 @@ static int sof_ipc4_pcm_hw_params(struct snd_soc_component *component,if (!spcm)return -EINVAL ; - time_info = spcm->stream[substream->stream].private; + time_info =sof_ipc4_sps_to_time_info (&spcm->stream[substream->stream]);/* delay calculation is not supported by current fw_reg ABI */ if (!time_info)return 0 ;
@@ -928,7 +957,7 @@ static int sof_ipc4_pcm_pointer(struct snd_soc_component *component,return -EOPNOTSUPP ; sps = &spcm->stream[substream->stream]; - time_info = sps->private; + time_info =sof_ipc4_sps_to_time_info (sps);if (!time_info)return -EOPNOTSUPP ;
@@ -1030,15 +1059,13 @@ static snd_pcm_sframes_t sof_ipc4_pcm_delay(struct snd_soc_component *component, {struct snd_soc_pcm_runtime *rtd =snd_soc_substream_to_rtd (substream);struct sof_ipc4_timestamp_info *time_info; -struct snd_sof_pcm_stream *sps;struct snd_sof_pcm *spcm; spcm =snd_sof_find_spcm_dai (component, rtd);if (!spcm)return 0 ; - sps = &spcm->stream[substream->stream]; - time_info = sps->private; + time_info =sof_ipc4_sps_to_time_info (&spcm->stream[substream->stream]);/* * Report the stored delay value calculated in the pointer callback. * In the unlikely event that the calculation was skipped/aborted, the
Signature · 1 line
2.44.0
Peter Ujfalusi
maintainerPATCH 4/4 ASoC: SOF: ipc4-pcm: Do not reset the ChainDMA if it has not been allocated
In reply to Peter UjfalusiThe ChainDMA operation differs from normal pipelines that it is only created when the stream started, in fact a PCM using ChainDMA has no pipelines, modules.
To reset a ChainDMA, it needs to be first allocated in firmware. When PulseAudio/PipeWire starts, they will probe the PCMs by opening them, check hw_params and then close the PCM without starting audio. Unconditionally resetting the ChainDMA can result the following error:
ipc tx : 0xe040000|0x0: GLB_CHAIN_DMA
ipc tx reply: 0x2e000007|0x0: GLB_CHAIN_DMA FW reported error: 7 - Unsupported operation requested ipc error for msg 0xe040000|0x0sof_pcm_stream_free: pcm_ops hw_free failed -22
Add a new chain_dma_allocated flag to sof_ipc4_pcm_stream_priv to store the ChainDMA allocation state and use this flag to skip sending the reset if the ChainDMA is not allocated.
--- sound/soc/sof/ipc4-pcm.c | 24 ++++++++++++++++++++---- 1 file changed, 20 insertions(+), 4 deletions(-)
sound/soc/sof/ipc4-pcm.c +20 -4
@@ -40,9 +40,12 @@ struct sof_ipc4_timestamp_info {/** * struct sof_ipc4_pcm_stream_priv - IPC4 specific private data * @time_info: pointer to time info struct if it is supported, otherwise NULL +* @chain_dma_allocated: indicates the ChainDMA allocation state */ struct sof_ipc4_pcm_stream_priv {struct sof_ipc4_timestamp_info *time_info; + +bool chain_dma_allocated; }; static inline struct sof_ipc4_timestamp_info *
@@ -269,14 +272,17 @@ sof_ipc4_update_pipeline_state(struct snd_sof_dev *sdev, int state, int cmd,*/ static int sof_ipc4_chain_dma_trigger (struct snd_sof_dev *sdev, -int direction, +struct snd_sof_pcm *spcm,int direction,struct snd_sof_pcm_stream_pipeline_list *pipeline_list,int state,int cmd) {struct sof_ipc4_fw_data *ipc4_data = sdev->private; +struct sof_ipc4_pcm_stream_priv *stream_priv;bool allocate, enable, set_fifo_size;struct sof_ipc4_msg msg = {{0 }}; -int i; +int ret, i; + + stream_priv = spcm->stream[direction].private;switch (state) {case SOF_IPC4_PIPE_RUNNING :/* Allocate and start chained dma */
@@ -297,6 +303,11 @@ static int sof_ipc4_chain_dma_trigger(struct snd_sof_dev *sdev, set_fifo_size = false;break ;case SOF_IPC4_PIPE_RESET :/* Disable and free chained DMA. */ + +/* ChainDMA can only be reset if it has been allocated */ +if (!stream_priv->chain_dma_allocated) +return 0 ; + allocate = false; enable = false; set_fifo_size = false;
@@ -354,7 +365,12 @@ static int sof_ipc4_chain_dma_trigger(struct snd_sof_dev *sdev,if (enable) msg.primary |=SOF_IPC4_GLB_CHAIN_DMA_ENABLE_MASK ; -return sof_ipc_tx_message_no_reply (sdev->ipc, &msg,0 ); + ret =sof_ipc_tx_message_no_reply (sdev->ipc, &msg,0 ); +/* Update the ChainDMA allocation state */ +if (!ret) + stream_priv->chain_dma_allocated = allocate; + +return ret; } static int sof_ipc4_trigger_pipelines(struct snd_soc_component *component,
@@ -394,7 +410,7 @@ static int sof_ipc4_trigger_pipelines(struct snd_soc_component *component,* trigger function that handles the rest for the substream. */ if (pipeline->use_chain_dma) -return sof_ipc4_chain_dma_trigger (sdev, substream->stream, +return sof_ipc4_chain_dma_trigger (sdev, spcm, substream->stream, pipeline_list, state, cmd);/* allocate memory for the pipeline data */
Signature · 1 line
2.44.0
Mark Brown
maintainerQuoted 10 lines · Peter Ujfalusi
The current code will reset the ChainDMA on release unconditionally which can result the following error when the CHainDMA is not allocated:
ipc tx : 0xe040000|0x0:
GLB_CHAIN_DMA
ipc tx reply: 0x2e000007|0x0:GLB_CHAIN_DMAFW reported error: 7 - Unsupported operation requested ipc error for msg 0xe040000|0x0sof_pcm_stream_free:pcm_opshw_freefailed -22[...]
Applied to
https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next
Thanks!
[1/4] ASoC: SOF: ipc4-pcm: Use consistent name for snd_sof_pcm_stream pointer
commit: 965e49cdf8c19f21b8308adeded3a8139cff5c84
[2/4] ASoC: SOF: ipc4-pcm: Use consistent name for sof_ipc4_timestamp_info pointer
commit: 36e980050b0733829e4e0f97b97f7907ba9f00bb
[3/4] ASoC: SOF: ipc4-pcm: Introduce generic sof_ipc4_pcm_stream_priv
commit: 551af3280c16166244425bbb1d73028f3a907e1f
[4/4] ASoC: SOF: ipc4-pcm: Do not reset the ChainDMA if it has not been allocated
commit: 7211814f2adcf376b8db6321447a9725c33b6ae7All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.
You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.
If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.
Please add any relevant lists and maintainers to the CCs when replying to this mail.
Thanks,
Mark