[PATCH 0/5] ASoC: SOF: use common code for DSP core pm
Series to refactor the DSP core management: - move tracking of powered up DSP cores to common SOF code - add logic filter unnecessary power actions - modify existing implementations to use common code whenever DSP cores are powered, so the state in common code stays in sync
Bard Liao (5): ASoC: SOF: Intel: hda: use snd_sof_dsp_core_power_up/down API ASoC: SOF: Intel: hda-loader: keep init cores alive ASoC: SOF: update dsp core power status in common APIs ASoC: SOF: Filter out unneeded core power up/downs ASoC: SOF: intel: hda-loader: use snd_sof_dsp_core_power_down/up APIs
sound/soc/sof/intel/hda-dsp.c | 2 +- sound/soc/sof/intel/hda-loader.c | 9 +++++---- sound/soc/sof/intel/hda.c | 2 +- sound/soc/sof/loader.c | 6 ------ sound/soc/sof/ops.h | 24 ++++++++++++++++++------ sound/soc/sof/pm.c | 1 - sound/soc/sof/topology.c | 8 -------- 7 files changed, 25 insertions(+), 27 deletions(-)
base-commit: e32df142359fb6f4d27977b7652549f0844aa62f
From: Bard Liao yung-chuan.liao@linux.intel.com
To implement common logic in SOF core, core power up/down flows should use common SOF API and not directly use low-level platform specific helper functions.
Signed-off-by: Bard Liao yung-chuan.liao@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Signed-off-by: Kai Vehmanen kai.vehmanen@linux.intel.com --- sound/soc/sof/intel/hda-dsp.c | 2 +- sound/soc/sof/intel/hda-loader.c | 2 +- sound/soc/sof/intel/hda.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/sound/soc/sof/intel/hda-dsp.c b/sound/soc/sof/intel/hda-dsp.c index 2b001151fe37..4fe4175179d4 100644 --- a/sound/soc/sof/intel/hda-dsp.c +++ b/sound/soc/sof/intel/hda-dsp.c @@ -624,7 +624,7 @@ static int hda_suspend(struct snd_sof_dev *sdev, bool runtime_suspend) #endif
/* power down DSP */ - ret = hda_dsp_core_reset_power_down(sdev, chip->host_managed_cores_mask); + ret = snd_sof_dsp_core_power_down(sdev, chip->host_managed_cores_mask); if (ret < 0) { dev_err(sdev->dev, "error: failed to power down core during suspend\n"); diff --git a/sound/soc/sof/intel/hda-loader.c b/sound/soc/sof/intel/hda-loader.c index 365a79fc7081..4b2ab351e2cc 100644 --- a/sound/soc/sof/intel/hda-loader.c +++ b/sound/soc/sof/intel/hda-loader.c @@ -183,7 +183,7 @@ static int cl_dsp_init(struct snd_sof_dev *sdev, int stream_tag) flags |= SOF_DBG_DUMP_FORCE_ERR_LEVEL;
hda_dsp_dump(sdev, flags); - hda_dsp_core_reset_power_down(sdev, chip->host_managed_cores_mask); + snd_sof_dsp_core_power_down(sdev, chip->host_managed_cores_mask);
return ret; } diff --git a/sound/soc/sof/intel/hda.c b/sound/soc/sof/intel/hda.c index 7e703ce22fcd..d57d484a42d1 100644 --- a/sound/soc/sof/intel/hda.c +++ b/sound/soc/sof/intel/hda.c @@ -928,7 +928,7 @@ int hda_dsp_remove(struct snd_sof_dev *sdev)
/* disable cores */ if (chip) - hda_dsp_core_reset_power_down(sdev, chip->host_managed_cores_mask); + snd_sof_dsp_core_power_down(sdev, chip->host_managed_cores_mask);
/* disable DSP */ snd_sof_dsp_update_bits(sdev, HDA_DSP_PP_BAR, SOF_HDA_REG_PP_PPCTL,
From: Bard Liao yung-chuan.liao@linux.intel.com
init_core_mask should be the available cores mask after fw boot. So we should keep not core 0 but init cores alive.
Signed-off-by: Bard Liao yung-chuan.liao@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Signed-off-by: Kai Vehmanen kai.vehmanen@linux.intel.com --- sound/soc/sof/intel/hda-loader.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/sound/soc/sof/intel/hda-loader.c b/sound/soc/sof/intel/hda-loader.c index 4b2ab351e2cc..07b73fe3c5e5 100644 --- a/sound/soc/sof/intel/hda-loader.c +++ b/sound/soc/sof/intel/hda-loader.c @@ -147,8 +147,9 @@ static int cl_dsp_init(struct snd_sof_dev *sdev, int stream_tag) chip->ipc_ack_mask, chip->ipc_ack_mask);
- /* step 5: power down corex */ - ret = hda_dsp_core_power_down(sdev, chip->host_managed_cores_mask & ~(BIT(0))); + /* step 5: power down cores that are no longer needed */ + ret = hda_dsp_core_power_down(sdev, chip->host_managed_cores_mask & + ~(chip->init_core_mask)); if (ret < 0) { if (hda->boot_iteration == HDA_FW_BOOT_ATTEMPTS) dev_err(sdev->dev,
From: Bard Liao yung-chuan.liao@linux.intel.com
Only manage enabled_cores_mask in common snd_sof_dsp_core_power_up/down APIs to ensure it stays in sync with actual DSP core state.
Signed-off-by: Bard Liao yung-chuan.liao@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Signed-off-by: Kai Vehmanen kai.vehmanen@linux.intel.com --- sound/soc/sof/loader.c | 6 ------ sound/soc/sof/ops.h | 22 ++++++++++++++++------ sound/soc/sof/pm.c | 1 - sound/soc/sof/topology.c | 8 -------- 4 files changed, 16 insertions(+), 21 deletions(-)
diff --git a/sound/soc/sof/loader.c b/sound/soc/sof/loader.c index 08a17abb63ff..4ade55bf6d75 100644 --- a/sound/soc/sof/loader.c +++ b/sound/soc/sof/loader.c @@ -811,7 +811,6 @@ EXPORT_SYMBOL(snd_sof_load_firmware); int snd_sof_run_firmware(struct snd_sof_dev *sdev) { int ret; - int init_core_mask;
init_waitqueue_head(&sdev->boot_wait);
@@ -843,8 +842,6 @@ int snd_sof_run_firmware(struct snd_sof_dev *sdev) return ret; }
- init_core_mask = ret; - /* * now wait for the DSP to boot. There are 3 possible outcomes: * 1. Boot wait times out indicating FW boot failure. @@ -874,9 +871,6 @@ int snd_sof_run_firmware(struct snd_sof_dev *sdev) return ret; }
- /* fw boot is complete. Update the active cores mask */ - sdev->enabled_cores_mask = init_core_mask; - return 0; } EXPORT_SYMBOL(snd_sof_run_firmware); diff --git a/sound/soc/sof/ops.h b/sound/soc/sof/ops.h index f0c9ca8820d2..2e9a8da53d57 100644 --- a/sound/soc/sof/ops.h +++ b/sound/soc/sof/ops.h @@ -76,19 +76,29 @@ static inline int snd_sof_dsp_reset(struct snd_sof_dev *sdev) static inline int snd_sof_dsp_core_power_up(struct snd_sof_dev *sdev, unsigned int core_mask) { - if (sof_ops(sdev)->core_power_up) - return sof_ops(sdev)->core_power_up(sdev, core_mask); + int ret = 0;
- return 0; + if (sof_ops(sdev)->core_power_up) { + ret = sof_ops(sdev)->core_power_up(sdev, core_mask); + if (!ret) + sdev->enabled_cores_mask |= core_mask; + } + + return ret; }
static inline int snd_sof_dsp_core_power_down(struct snd_sof_dev *sdev, unsigned int core_mask) { - if (sof_ops(sdev)->core_power_down) - return sof_ops(sdev)->core_power_down(sdev, core_mask); + int ret = 0;
- return 0; + if (sof_ops(sdev)->core_power_down) { + ret = sof_ops(sdev)->core_power_down(sdev, core_mask); + if (!ret) + sdev->enabled_cores_mask &= ~core_mask; + } + + return ret; }
/* pre/post fw load */ diff --git a/sound/soc/sof/pm.c b/sound/soc/sof/pm.c index c83fb6255961..fd265803f7bc 100644 --- a/sound/soc/sof/pm.c +++ b/sound/soc/sof/pm.c @@ -256,7 +256,6 @@ static int sof_suspend(struct device *dev, bool runtime_suspend)
/* reset FW state */ sdev->fw_state = SOF_FW_BOOT_NOT_STARTED; - sdev->enabled_cores_mask = 0;
return ret; } diff --git a/sound/soc/sof/topology.c b/sound/soc/sof/topology.c index b6b32a7a91f8..b69f493b5faa 100644 --- a/sound/soc/sof/topology.c +++ b/sound/soc/sof/topology.c @@ -1352,10 +1352,6 @@ static int sof_core_enable(struct snd_sof_dev *sdev, int core) core, ret); goto err; } - - /* update enabled cores mask */ - sdev->enabled_cores_mask |= BIT(core); - return ret; err: /* power down core if it is host managed and return the original error if this fails too */ @@ -2603,10 +2599,6 @@ static int sof_widget_unload(struct snd_soc_component *scomp, if (ret < 0) dev_err(scomp->dev, "error: powering down pipeline schedule core %d\n", pipeline->core); - - /* update enabled cores mask */ - sdev->enabled_cores_mask &= ~(1 << pipeline->core); - break; default: break;
From: Bard Liao yung-chuan.liao@linux.intel.com
Exclude cores that are already powered on/off correctly. This allows to simplify dsp_power_up/down() implementations and avoid unexpected error.
Signed-off-by: Bard Liao yung-chuan.liao@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Signed-off-by: Kai Vehmanen kai.vehmanen@linux.intel.com --- sound/soc/sof/ops.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/sound/soc/sof/ops.h b/sound/soc/sof/ops.h index 2e9a8da53d57..5099ad03df72 100644 --- a/sound/soc/sof/ops.h +++ b/sound/soc/sof/ops.h @@ -78,7 +78,8 @@ static inline int snd_sof_dsp_core_power_up(struct snd_sof_dev *sdev, { int ret = 0;
- if (sof_ops(sdev)->core_power_up) { + core_mask &= ~sdev->enabled_cores_mask; + if (sof_ops(sdev)->core_power_up && core_mask) { ret = sof_ops(sdev)->core_power_up(sdev, core_mask); if (!ret) sdev->enabled_cores_mask |= core_mask; @@ -92,7 +93,8 @@ static inline int snd_sof_dsp_core_power_down(struct snd_sof_dev *sdev, { int ret = 0;
- if (sof_ops(sdev)->core_power_down) { + core_mask &= sdev->enabled_cores_mask; + if (sof_ops(sdev)->core_power_down && core_mask) { ret = sof_ops(sdev)->core_power_down(sdev, core_mask); if (!ret) sdev->enabled_cores_mask &= ~core_mask;
From: Bard Liao yung-chuan.liao@linux.intel.com
To manage enabled_cores_mask flag, we should always use snd_sof_dsp_ core_power_down/up APIs to power up/down dsp cores. The APIs do a little bit more than the original functions, but it is harmless.
Suggested-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Signed-off-by: Bard Liao yung-chuan.liao@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Signed-off-by: Kai Vehmanen kai.vehmanen@linux.intel.com --- sound/soc/sof/intel/hda-loader.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/sound/soc/sof/intel/hda-loader.c b/sound/soc/sof/intel/hda-loader.c index 07b73fe3c5e5..d744952f6954 100644 --- a/sound/soc/sof/intel/hda-loader.c +++ b/sound/soc/sof/intel/hda-loader.c @@ -93,7 +93,7 @@ static int cl_dsp_init(struct snd_sof_dev *sdev, int stream_tag) int i;
/* step 1: power up corex */ - ret = hda_dsp_core_power_up(sdev, chip->host_managed_cores_mask); + ret = snd_sof_dsp_core_power_up(sdev, chip->host_managed_cores_mask); if (ret < 0) { if (hda->boot_iteration == HDA_FW_BOOT_ATTEMPTS) dev_err(sdev->dev, "error: dsp core 0/1 power up failed\n"); @@ -148,8 +148,8 @@ static int cl_dsp_init(struct snd_sof_dev *sdev, int stream_tag) chip->ipc_ack_mask);
/* step 5: power down cores that are no longer needed */ - ret = hda_dsp_core_power_down(sdev, chip->host_managed_cores_mask & - ~(chip->init_core_mask)); + ret = snd_sof_dsp_core_power_down(sdev, chip->host_managed_cores_mask & + ~(chip->init_core_mask)); if (ret < 0) { if (hda->boot_iteration == HDA_FW_BOOT_ATTEMPTS) dev_err(sdev->dev,
On Thu, 28 Jan 2021 11:38:45 +0200, Kai Vehmanen wrote:
Series to refactor the DSP core management:
- move tracking of powered up DSP cores to common SOF code
- add logic filter unnecessary power actions
- modify existing implementations to use common code whenever DSP cores are powered, so the state in common code stays in sync
[...]
Applied to
https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next
Thanks!
[1/5] ASoC: SOF: Intel: hda: use snd_sof_dsp_core_power_up/down API commit: f6c246eacb62977dea5c9c65ac6fb4921cad5bcd [2/5] ASoC: SOF: Intel: hda-loader: keep init cores alive commit: cedd502d18b5b7a913fa13fa18a037cc51b1798d [3/5] ASoC: SOF: update dsp core power status in common APIs commit: 42077f08b3f1ba891dca1f8f479810f16b7d6cbd [4/5] ASoC: SOF: Filter out unneeded core power up/downs commit: 30876e2a06f35b525dc71f94dfc3c6f329e55a28 [5/5] ASoC: SOF: intel: hda-loader: use snd_sof_dsp_core_power_down/up APIs commit: 92c6ec606cd12c16091b70442da536bdeddb1f7f
All 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
participants (2)
-
Kai Vehmanen -
Mark Brown