[PATCH 0/3] ASoC: SOF: Intel: mtl: Enable multicore support
Hi,
The following series will enable multicore support on MTL platforms similarly to other Intel platforms.
The TGL patch is included to simplify the core_put implementation.
Multicore support can be enabled by updated topologies, with current set of tplg files this series is not introducing any runtime change.
Regards, Peter --- Rander Wang (3): ASoC: SOF: Intel: mtl: setup primary core info on MeteorLake platform ASoC: SOF: Intel: mtl: add core_get & put support on MeterLake platforms ASoC: SOF: Intel: tgl: unify core_put on IPC3 & IPC4 path
sound/soc/sof/intel/mtl.c | 52 +++++++++++++++++++++++++++++++++++---- sound/soc/sof/intel/tgl.c | 10 +++++--- 2 files changed, 54 insertions(+), 8 deletions(-)
From: Rander Wang rander.wang@intel.com
Set primary core mask and refcount.
Signed-off-by: Rander Wang rander.wang@intel.com Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Reviewed-by: Péter Ujfalusi peter.ujfalusi@linux.intel.com Signed-off-by: Peter Ujfalusi peter.ujfalusi@linux.intel.com --- sound/soc/sof/intel/mtl.c | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-)
diff --git a/sound/soc/sof/intel/mtl.c b/sound/soc/sof/intel/mtl.c index 46caf3ccde66..4788ba3b25aa 100644 --- a/sound/soc/sof/intel/mtl.c +++ b/sound/soc/sof/intel/mtl.c @@ -361,11 +361,17 @@ static int mtl_dsp_core_power_up(struct snd_sof_dev *sdev, int core) ret = snd_sof_dsp_read_poll_timeout(sdev, HDA_DSP_BAR, MTL_DSP2CXCTL_PRIMARY_CORE, dspcxctl, (dspcxctl & cpa) == cpa, HDA_DSP_REG_POLL_INTERVAL_US, HDA_DSP_RESET_TIMEOUT_US); - if (ret < 0) + if (ret < 0) { dev_err(sdev->dev, "%s: timeout on MTL_DSP2CXCTL_PRIMARY_CORE read\n", __func__); + return ret; + }
- return ret; + /* set primary core mask and refcount to 1 */ + sdev->enabled_cores_mask = BIT(SOF_DSP_PRIMARY_CORE); + sdev->dsp_core_ref_count[SOF_DSP_PRIMARY_CORE] = 1; + + return 0; }
static int mtl_dsp_core_power_down(struct snd_sof_dev *sdev, int core) @@ -388,10 +394,15 @@ static int mtl_dsp_core_power_down(struct snd_sof_dev *sdev, int core) !(dspcxctl & MTL_DSP2CXCTL_PRIMARY_CORE_CPA_MASK), HDA_DSP_REG_POLL_INTERVAL_US, HDA_DSP_PD_TIMEOUT * USEC_PER_MSEC); - if (ret < 0) + if (ret < 0) { dev_err(sdev->dev, "failed to power down primary core\n"); + return ret; + }
- return ret; + sdev->enabled_cores_mask = 0; + sdev->dsp_core_ref_count[SOF_DSP_PRIMARY_CORE] = 0; + + return 0; }
static int mtl_power_down_dsp(struct snd_sof_dev *sdev)
From: Rander Wang rander.wang@intel.com
In core_get case, driver can power up primary core and don't need to send ipc message to fw. Non-primary core should be powered up by fw with ipc message.
In core_put case, driver should first send ipc message to fw to disable dsp core then power down primary core if the target is primary core.
Signed-off-by: Rander Wang rander.wang@intel.com Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Reviewed-by: Péter Ujfalusi peter.ujfalusi@linux.intel.com Signed-off-by: Peter Ujfalusi peter.ujfalusi@linux.intel.com --- sound/soc/sof/intel/mtl.c | 33 ++++++++++++++++++++++++++++++++- 1 file changed, 32 insertions(+), 1 deletion(-)
diff --git a/sound/soc/sof/intel/mtl.c b/sound/soc/sof/intel/mtl.c index 4788ba3b25aa..a847ae3ed4e9 100644 --- a/sound/soc/sof/intel/mtl.c +++ b/sound/soc/sof/intel/mtl.c @@ -624,6 +624,36 @@ static u64 mtl_dsp_get_stream_hda_link_position(struct snd_sof_dev *sdev, return ((u64)llp_u << 32) | llp_l; }
+static int mtl_dsp_core_get(struct snd_sof_dev *sdev, int core) +{ + const struct sof_ipc_pm_ops *pm_ops = sdev->ipc->ops->pm; + + if (core == SOF_DSP_PRIMARY_CORE) + return mtl_dsp_core_power_up(sdev, SOF_DSP_PRIMARY_CORE); + + if (pm_ops->set_core_state) + return pm_ops->set_core_state(sdev, core, true); + + return 0; +} + +static int mtl_dsp_core_put(struct snd_sof_dev *sdev, int core) +{ + const struct sof_ipc_pm_ops *pm_ops = sdev->ipc->ops->pm; + int ret; + + if (pm_ops->set_core_state) { + ret = pm_ops->set_core_state(sdev, core, false); + if (ret < 0) + return ret; + } + + if (core == SOF_DSP_PRIMARY_CORE) + return mtl_dsp_core_power_down(sdev, SOF_DSP_PRIMARY_CORE); + + return 0; +} + /* Meteorlake ops */ struct snd_sof_dsp_ops sof_mtl_ops; EXPORT_SYMBOL_NS(sof_mtl_ops, SND_SOC_SOF_INTEL_HDA_COMMON); @@ -660,7 +690,8 @@ int sof_mtl_ops_init(struct snd_sof_dev *sdev) sof_mtl_ops.parse_platform_ext_manifest = NULL;
/* dsp core get/put */ - /* TODO: add core_get and core_put */ + sof_mtl_ops.core_get = mtl_dsp_core_get; + sof_mtl_ops.core_put = mtl_dsp_core_put;
sof_mtl_ops.get_stream_position = mtl_dsp_get_stream_hda_link_position;
From: Rander Wang rander.wang@intel.com
Firmware may do context saving before powering off primary core, so driver needs to send ipc msg by set_core_state. In IPC4 path, firmware needs to save current context to IMR before powering off primary core. Firmware does nothing for set_core_state message in IPC3 path. So IPC4 and IPC3 can share the same operation sequence.
Signed-off-by: Rander Wang rander.wang@intel.com Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com Reviewed-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com Reviewed-by: Péter Ujfalusi peter.ujfalusi@linux.intel.com Signed-off-by: Peter Ujfalusi peter.ujfalusi@linux.intel.com --- sound/soc/sof/intel/tgl.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/sound/soc/sof/intel/tgl.c b/sound/soc/sof/intel/tgl.c index 2713b7dc7931..8e2b07e1612b 100644 --- a/sound/soc/sof/intel/tgl.c +++ b/sound/soc/sof/intel/tgl.c @@ -39,14 +39,18 @@ static int tgl_dsp_core_get(struct snd_sof_dev *sdev, int core) static int tgl_dsp_core_put(struct snd_sof_dev *sdev, int core) { const struct sof_ipc_pm_ops *pm_ops = sdev->ipc->ops->pm; + int ret; + + if (pm_ops->set_core_state) { + ret = pm_ops->set_core_state(sdev, core, false); + if (ret < 0) + return ret; + }
/* power down primary core and return */ if (core == SOF_DSP_PRIMARY_CORE) return hda_dsp_core_reset_power_down(sdev, BIT(core));
- if (pm_ops->set_core_state) - return pm_ops->set_core_state(sdev, core, false); - return 0; }
On Tue, 23 May 2023 13:32:14 +0300, Peter Ujfalusi wrote:
The following series will enable multicore support on MTL platforms similarly to other Intel platforms.
The TGL patch is included to simplify the core_put implementation.
Multicore support can be enabled by updated topologies, with current set of tplg files this series is not introducing any runtime change.
[...]
Applied to
https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next
Thanks!
[2/3] ASoC: SOF: Intel: mtl: add core_get & put support on MeterLake platforms commit: c6d15567a4d5dd51ecccc332d514c6dc21bce652 [3/3] ASoC: SOF: Intel: tgl: unify core_put on IPC3 & IPC4 path commit: 1b167ba8a20152041d3af0c0cbbfd710f1e93e4b
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
Hi Mark,
On Tue, 23 May 2023 13:32:14 +0300, Peter Ujfalusi wrote:
The following series will enable multicore support on MTL platforms similarly to other Intel platforms.
The TGL patch is included to simplify the core_put implementation.
Multicore support can be enabled by updated topologies, with current set of tplg files this series is not introducing any runtime change.
[...]
Applied to
https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next
Thanks!
Looks like patch1 was not applied:
[PATCH 1/3] ASoC: SOF: Intel: mtl: setup primary core info on MeteorLake platform
was there anything wrong with it or is this a b4/script issue?
Please let us know if we need to resend it or modify it.
Thanks, -Pierre
[2/3] ASoC: SOF: Intel: mtl: add core_get & put support on MeterLake platforms commit: c6d15567a4d5dd51ecccc332d514c6dc21bce652 [3/3] ASoC: SOF: Intel: tgl: unify core_put on IPC3 & IPC4 path commit: 1b167ba8a20152041d3af0c0cbbfd710f1e93e4b
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
On Fri, May 26, 2023 at 03:50:49PM -0500, Pierre-Louis Bossart wrote:
Looks like patch1 was not applied:
[PATCH 1/3] ASoC: SOF: Intel: mtl: setup primary core info on MeteorLake platform
was there anything wrong with it or is this a b4/script issue?
Please let us know if we need to resend it or modify it.
I imagine it ended up looking like it was already applied, please resend.
participants (3)
-
Mark Brown -
Peter Ujfalusi -
Pierre-Louis Bossart