ASoC: Intel: add I2S function topology support by bardliao · Pull Request #5657 · thesofproject/linux

bardliao · 2026-01-28T02:56:49Z

Currently, function topology is only used by the SoundWire machines. Extend the support to the I2S machines.

ujfalusi

only checked the first patch

ujfalusi · 2026-01-28T07:34:45Z

 	}

-	dev_err(dev, "%s: No match for SSP%d in NHLT table\n", __func__,
+	dev_dbg(dev, "%s: No match for SSP%d in NHLT table\n", __func__,


I would have this as separate patch?

we are also going to look for DMIC blobs in several nhlt, so change those prints as well?

Not sure about the DMIC blobs. Maybe we can change the DMIC prints when we look for DMIC blobs in several nhlt?

but we do after the second patch...

There could be NHLT blobs w/o DMIC (SSP fragment) and we would print error/warn?

intel_nhlt_get_endpoint_blob() will be called for each NHLT blobs, but it uses dev_dbg already.

you need to call out that in preparation for supporting multiple topology fragments per function which could end up calling this function multiple types with NHLT blobs from different fragments, downgrade the error message to debug level.

Commit message updated. But do you mean to move the commit to before the ASoC: SOF: get nhlt from all topologies commit?

ujfalusi · 2026-01-28T07:45:29Z

+								   channel_count, sample_rate,
+								   dir, dev_type);
+				if (cfg)
+					break;


goto out; and remove the cfg check after the loop?

ujfalusi

@bardliao, looks really nice, I have few nitpicks and one bigger question only

ujfalusi · 2026-01-29T06:04:33Z

 	}

-	dev_err(dev, "%s: No match for SSP%d in NHLT table\n", __func__,
+	dev_dbg(dev, "%s: No match for SSP%d in NHLT table\n", __func__,


but we do after the second patch...

There could be NHLT blobs w/o DMIC (SSP fragment) and we would print error/warn?

ujfalusi · 2026-01-29T06:16:23Z

-						      dai_index);
-		if (dev_type < 0)
+		list_for_each_entry(nhlt, &ipc4_data->nhlt_list, list) {
+			dev_type = intel_nhlt_ssp_device_type(sdev->dev, nhlt->nhlt,


nitpick: nhlt->nhlt, can we do entry->nhlt instead? the outer nhlt is not nhlt as such.

ujfalusi · 2026-01-29T06:30:01Z

+		if (!(*tplg_files)[tplg_num])
+			return -ENOMEM;
+		tplg_num++;
+	}


To my untrained eyes this looks like identical to what sof_sdw_get_tplg_files() would do, just that this lacks the SDW part, no?
Why not rename that function ad use it for all types as universal tool?

Yeah, I thought about it. And yes, they look quite similar, but the existing topologies may use different PCM IDs for different machines. I think it would be better to use separated function to keep the flexibility to have different function topology names for different machines.

CBala21 · 2026-01-30T12:05:50Z

Tested HDMI in capture and I2S functional. observed the expected behavior.

ranj063 · 2026-01-30T16:54:36Z

 	}

-	dev_err(dev, "%s: No match for SSP%d in NHLT table\n", __func__,
+	dev_dbg(dev, "%s: No match for SSP%d in NHLT table\n", __func__,


you need to call out that in preparation for supporting multiple topology fragments per function which could end up calling this function multiple types with NHLT blobs from different fragments, downgrade the error message to debug level.

lgirdwood

LGTM, one nitpick would be it could do with more inline commentary when it comes to how we are searching for files.

ranj063

LGTM thanks @bardliao

ujfalusi

I think this looks good, but not really sure how it is planned to be used in practice ;)

lgirdwood · 2026-02-05T16:13:53Z

CI needs kicked.

lgirdwood · 2026-02-05T16:14:00Z

SOFCI TEST

bardliao · 2026-02-06T13:41:53Z

change:
fix checkpatch issues.

lgirdwood

LGTM, just one question.

Copilot

Pull request overview

This PR extends SOF “function topology” support beyond SoundWire machines to also cover Intel I2S (SSP) machines. It does so by (a) allowing multiple NHLT tables (BIOS + topology manifest) to be tracked and consulted during blob lookups, and (b) adding an I2S-specific function-topology file selection helper and wiring it into PTL machine matching.

Changes:

Replace single ipc4_data->nhlt storage with an nhlt_list and update IPC4 NHLT blob/device-type lookups to iterate across tables.
Add sof_i2s_get_tplg_files() to generate function-topology fragment lists for I2S machines and enable it for PTL I2S entries.
Reduce log severity in intel_nhlt_ssp_device_type() when NHLT is missing/no match.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
sound/soc/sof/ipc4-topology.c	Use a list of NHLT tables for SSP/DMIC endpoint blob lookup; add topology-manifest NHLT entries to the list.
sound/soc/sof/ipc4-priv.h	Introduce `struct snd_ipc4_nhlt` and replace single NHLT pointer with `nhlt_list`.
sound/soc/sof/ipc4-loader.c	Initialize `nhlt_list` during IPC4 firmware parsing.
sound/soc/sof/intel/hda-dai.c	Add BIOS NHLT to `nhlt_list` and attempt to free it during teardown.
sound/soc/intel/common/sof-function-topology-lib.h	Export new `sof_i2s_get_tplg_files()` API.
sound/soc/intel/common/sof-function-topology-lib.c	Implement I2S function-topology selection and refactor common helpers.
sound/soc/intel/common/soc-acpi-intel-ptl-match.c	Enable function topologies for PTL I2S machines via `sof_i2s_get_tplg_files`.
sound/hda/core/intel-nhlt.c	Downgrade NHLT “missing/no match” logs from error to debug.

Comments suppressed due to low confidence (1)

sound/soc/sof/ipc4-loader.c:156

INIT_LIST_HEAD(&ipc4_data->nhlt_list) here re-initializes the list during firmware parsing. If any NHLT entries were already added earlier (e.g. from BIOS during ops init), this will discard them and leave dangling list nodes. The list should be initialized once when sof_ipc4_fw_data is allocated, and this re-init should be removed or guarded so it cannot clobber existing entries.

static size_t sof_ipc4_fw_parse_basefw_ext_man(struct snd_sof_dev *sdev)
{
	struct sof_ipc4_fw_data *ipc4_data = sdev->private;
	struct sof_ipc4_fw_library *fw_lib;
	ssize_t payload_offset;
	int ret;

	INIT_LIST_HEAD(&ipc4_data->nhlt_list);
	if (sdev->dsp_test_mode_enabled)
		fw_lib = devm_kzalloc(sdev->dev, sizeof(*fw_lib), GFP_KERNEL);
	else
		fw_lib = devm_kzalloc(sdev->dev, sizeof(*fw_lib), GFP_KERNEL);

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

bardliao · 2026-05-08T11:19:10Z

 	if (sdev->pdata->ipc_type == SOF_IPC_TYPE_4 && !hda_use_tplg_nhlt) {
 		struct sof_ipc4_fw_data *ipc4_data = sdev->private;
+		struct snd_ipc4_nhlt *entry;

-		ipc4_data->nhlt = intel_nhlt_init(sdev->dev);
+		entry = devm_kzalloc(sdev->dev, sizeof(*entry), GFP_KERNEL);
+		if (!entry)
+			return;
+
+		entry->nhlt = intel_nhlt_init(sdev->dev);
+
+		list_add(&entry->list, &ipc4_data->nhlt_list);
 	}


@copilot INIT_LIST_HEAD(&ipc4_data->nhlt_list); is called in the sof_ipc4_fw_parse_basefw_ext_man() function which will be called when loading the firmware that is guarantee before this function is called.

bardliao · 2026-05-08T11:28:16Z

 void hda_ops_free(struct snd_sof_dev *sdev)
 {
 	if (sdev->pdata->ipc_type == SOF_IPC_TYPE_4) {
 		struct sof_ipc4_fw_data *ipc4_data = sdev->private;
+		struct snd_ipc4_nhlt *entry;

-		if (!hda_use_tplg_nhlt)
-			intel_nhlt_free(ipc4_data->nhlt);
+		if (!hda_use_tplg_nhlt) {
+			entry = list_first_entry(&ipc4_data->nhlt_list,
+						 struct snd_ipc4_nhlt, list);
+			intel_nhlt_free(entry->nhlt);
+			list_del(&entry->list);
+		}


@copilot please suggest a fix

+	int ret;
+
+	ret = sscanf(mach->sof_tplg_filename, "sof-%3s-*.tplg", platform);
+	if (ret != 1) {
+		dev_err(card->dev, "Invalid platform name %s of tplg %s\n",
+			platform, mach->sof_tplg_filename);
+		return -EINVAL;
+	}


+	unsigned int ssp_port;
+
+	if (sscanf(dai_link->name, "SSP%d", &ssp_port) != 1) {
+		dev_err(dev, "Invalid SSP port %d\n", ssp_port);
+		return -EINVAL;
+	}
+	if (strstr(dai_link->name, "Codec")) {
+		/*
+		 * Assume DAI link 0 is jack which is true in all existing
+		 * machine drivers
+		 */
+		if (dai_link->id == 0) {
+			*tplg_dev = TPLG_DEVICE_SSP_JACK;
+			*tplg_dev_name = devm_kasprintf(dev, GFP_KERNEL,
+							"ssp%d-jack", ssp_port);
+		} else {
+			*tplg_dev = TPLG_DEVICE_SSP_AMP;
+			*tplg_dev_name = devm_kasprintf(dev, GFP_KERNEL,
+							"ssp%d-amp", ssp_port);
+		}
+	} else if (strstr(dai_link->name, "BT")) {
+		*tplg_dev = TPLG_DEVICE_SSP_BT;
+		*tplg_dev_name = devm_kasprintf(dev, GFP_KERNEL,
+						"ssp%d-bt", ssp_port);
+	} else if (strstr(dai_link->name, "HDMI")) {


+		if (strstr(dai_link->name, "SSP")) {
+			unsigned int ssp_port;
+
+			if (sscanf(dai_link->name, "SSP%d", &ssp_port) != 1) {
+				dev_err(card->dev, "Invalid SSP port %d\n", ssp_port);
+				return -EINVAL;
+			}
+			if (strstr(dai_link->name, "Codec")) {
+				/*
+				 * Assume DAI link 0 is jack which is true in all existing
+				 * machine driver
+				 */
+				if (dai_link->id == 0) {
+					tplg_dev = TPLG_DEVICE_SSP_JACK;
+					tplg_dev_name = devm_kasprintf(card->dev, GFP_KERNEL,
+								       "ssp%d-jack", ssp_port);
+				} else {
+					tplg_dev = TPLG_DEVICE_SSP_AMP;
+					tplg_dev_name = devm_kasprintf(card->dev, GFP_KERNEL,
+								       "ssp%d-amp", ssp_port);
+				}
+			} else if (strstr(dai_link->name, "BT")) {
+				tplg_dev = TPLG_DEVICE_SSP_BT;
+				tplg_dev_name = devm_kasprintf(card->dev, GFP_KERNEL,
+							       "ssp%d-bt", ssp_port);
+			} else if (strstr(dai_link->name, "HDMI")) {


The existing code assumes there is only one NHLT blob either from BIOS or topology. As function topologies are used and each function topology contains a NHLT blob section, we need to search the matching NHLT blob from all the NHLT blobs. The commit suggests adding a list of NHLTs and search the NHLTs from the list. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

To support multiple topology fragments per function, this function could be called multiple types with different NHLT blobs from different fragments. Downgrade the error message to debug level since the SSP device type could be found from one of the nhlt. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

The existing code supports get_function_tplg_files callback for SoundWire machine driver only. Some common sections can be used to extend the support to other machines. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

…et_tplg_files The Intel SOF SDW machine drive also supports I2S interface. Add related supports for the sof_sdw_get_tplg_files() callback. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

… I2S machines Add sof_i2s_get_tplg_files() callback for Intel SOF I2S machines. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

Use sof_i2s_get_tplg_files() for SOF es83x6 machines. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

bardliao requested review from CBala21, ranj063 and ujfalusi January 28, 2026 02:56

bardliao requested review from dbaluta, kv2019i, lgirdwood and plbossart as code owners January 28, 2026 02:56

ujfalusi reviewed Jan 28, 2026

View reviewed changes

bardliao force-pushed the get-i2s-tplg branch 2 times, most recently from a0fd522 to 8d4c0c3 Compare January 29, 2026 03:30

ujfalusi reviewed Jan 29, 2026

View reviewed changes

bardliao force-pushed the get-i2s-tplg branch from 8d4c0c3 to 1af3d92 Compare January 29, 2026 14:19

ranj063 reviewed Jan 29, 2026

View reviewed changes

Comment thread sound/soc/sof/ipc4-topology.c

ranj063 reviewed Jan 30, 2026

View reviewed changes

bardliao force-pushed the get-i2s-tplg branch from 1af3d92 to 9fdc58f Compare February 2, 2026 05:30

lgirdwood reviewed Feb 3, 2026

View reviewed changes

bardliao force-pushed the get-i2s-tplg branch from 9fdc58f to fd2f6cc Compare February 4, 2026 14:05

ranj063 previously approved these changes Feb 4, 2026

View reviewed changes

ujfalusi previously approved these changes Feb 5, 2026

View reviewed changes

Comment thread sound/soc/intel/common/soc-acpi-intel-ptl-match.c

lgirdwood previously approved these changes Feb 5, 2026

View reviewed changes

bardliao dismissed stale reviews from lgirdwood, ujfalusi, and ranj063 via e6325cc February 6, 2026 13:37

bardliao force-pushed the get-i2s-tplg branch from fd2f6cc to e6325cc Compare February 6, 2026 13:37

lgirdwood reviewed Feb 10, 2026

View reviewed changes

Comment thread sound/soc/sof/intel/hda-dai.c

ranj063 previously approved these changes Feb 11, 2026

View reviewed changes

bardliao requested review from lgirdwood and ujfalusi March 24, 2026 06:08

lgirdwood previously approved these changes Apr 16, 2026

View reviewed changes

bardliao dismissed stale reviews from lgirdwood and ranj063 via f7f0490 May 5, 2026 05:39

bardliao force-pushed the get-i2s-tplg branch from e6325cc to f7f0490 Compare May 5, 2026 05:39

ujfalusi requested a review from Copilot May 7, 2026 11:55

Copilot started reviewing on behalf of ujfalusi May 7, 2026 11:56 View session

Copilot AI reviewed May 7, 2026

View reviewed changes

bardliao added 6 commits May 8, 2026 20:15

ASoC: Intel: sof-function-topology-lib: add I2S support for sof_sdw_g…

b2a0733

…et_tplg_files The Intel SOF SDW machine drive also supports I2S interface. Add related supports for the sof_sdw_get_tplg_files() callback. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

ASoC: Intel: sof-function-topology-lib: add get_function_topology for…

6ce6cdb

… I2S machines Add sof_i2s_get_tplg_files() callback for Intel SOF I2S machines. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

ASoC: soc-acpi-intel-ptl-match: I2S machines: use function topology

1a9ea13

Use sof_i2s_get_tplg_files() for SOF es83x6 machines. Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>

bardliao force-pushed the get-i2s-tplg branch from f7f0490 to 1a9ea13 Compare May 8, 2026 12:16

Conversation

bardliao commented Jan 28, 2026

Uh oh!

ujfalusi left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ujfalusi Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ujfalusi left a comment

Choose a reason for hiding this comment

Uh oh!

ujfalusi Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

CBala21 commented Jan 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

lgirdwood left a comment

Choose a reason for hiding this comment

Uh oh!

ranj063 left a comment

Choose a reason for hiding this comment

Uh oh!

ujfalusi left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

lgirdwood commented Feb 5, 2026

Uh oh!

lgirdwood commented Feb 5, 2026

Uh oh!

bardliao commented Feb 6, 2026

Uh oh!

lgirdwood left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

ujfalusi Jan 29, 2026 •

edited

Loading

ujfalusi Jan 29, 2026 •

edited

Loading

CBala21 commented Jan 30, 2026 •

edited

Loading