Code to decode pipeline create messages payload #10265

jsarha · 2025-09-26T11:52:33Z

Code to decode pipline create messages payload. E.g. this code is written to test this PR: thesofproject/linux#5537

src/include/ipc4/pipeline.h

lgirdwood

LGTM, we can followup with the allocations per pipeline or domain later.

src/include/ipc4/pipeline.h

src/ipc/ipc4/helper.c

lyakh · 2025-11-14T11:41:11Z

src/ipc/ipc4/helper.c

+
+		/* Check if there is space for the object header */
+		if ((char *)(obj + 1) - data > size) {
+			tr_err(&ipc_tr, "obj header overflow, %u > %u",


shouldn't the format be %zu for size here and below?

resolved by mistake?

Argh... I fixed this in one place already, but the copy-pasted version is still in the original form.

src/ipc/ipc4/helper.c

jsarha · 2025-12-03T19:41:39Z

@lgirdwood this is now updated so that it should be almost exact match to #10281 , vpages code only needs to use the params passed to pipeline_new(). But to use this the topology should be built from this PR and you need my linux driver PR #10281 too to pass down the topology info.

jsarha · 2025-12-14T23:19:20Z

This version address the changes requested to thesofproject/linux#5537 . I still need to move the module init payload decoding to happen already before module_adapter_mem_alloc() -call, before this is ready.

jsarha · 2025-12-16T16:01:08Z

Finaly the module init payload decoding is happening where it should be. Ready for review.

jsarha · 2025-12-16T16:03:58Z

@lgirdwood should I rebase this or is it better to keep it on top of older base still?

lgirdwood · 2025-12-16T16:25:34Z

@lgirdwood should I rebase this or is it better to keep it on top of older base still?

Best to rebase and adding @ranj063

Copilot

Pull request overview

This PR implements decoding logic for pipeline creation message payloads in IPC4, enabling the parsing of extended pipeline initialization data such as memory requirements (stack, heap, shared memory). The changes support testing against a corresponding Linux kernel PR that extends pipeline creation capabilities.

Key Changes:

Updated memory requirement tokens in topology configuration from a single heap_bytes_requirement to separate interim_heap_bytes_requirement, lifetime_heap_bytes_requirement, and shared_bytes_requirement
Added ipc4_create_pipeline_payload_decode() function to parse extended pipeline creation payloads
Refactored module initialization to pass extended data through a new module_ext_init_data structure instead of storing values directly in module_config

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
tools/topology/topology2/include/components/widget-common.conf	Updates topology token definitions for new memory requirement types and adds default values
tools/topology/topology2/include/common/tokens.conf	Adds token IDs for the new memory requirement types
test/cmocka/src/audio/pipeline/pipeline_new.c	Updates test to pass NULL for new pipeline parameters argument
src/ipc/ipc4/helper.c	Implements payload decoding logic and updates pipeline creation to accept extended parameters
src/ipc/ipc3/helper.c	Updates pipeline creation call to pass NULL for new parameters argument
src/include/sof/audio/pipeline.h	Adds `pipeline_params` structure and updates `pipeline_new()` signature
src/include/sof/audio/module_adapter/module/generic.h	Declares `module_ext_init_decode()` function with IPC4-specific implementation
src/include/module/module/base.h	Adds `module_ext_init_data` structure and refactors `module_config` to use it
src/include/ipc4/pipeline.h	Defines structures for pipeline extended payload objects
src/include/ipc4/module.h	Updates module initialization structure to include separate heap memory fields
src/audio/pipeline/pipeline-graph.c	Updates `pipeline_new()` signature to accept pipeline parameters
src/audio/module_adapter/module_adapter_ipc4.c	Refactors module extension decoding to populate `module_ext_init_data` instead of `module_config`
src/audio/module_adapter/module_adapter.c	Updates module adapter initialization to decode extended data and manage its lifecycle

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

tools/topology/topology2/include/components/widget-common.conf

src/ipc/ipc4/helper.c

src/include/module/module/base.h

src/audio/module_adapter/module_adapter.c

jsarha · 2025-12-16T21:53:49Z

Fixed copilot spotted typos.

…code() Use %zu in size_t prints in module_ext_init_decode(). Signed-off-by: Jyri Sarha <[email protected]>

jsarha · 2025-12-16T22:14:31Z

... and rebased.

src/ipc/ipc4/helper.c

lyakh · 2025-12-17T08:14:19Z

src/ipc/ipc4/helper.c

+
+		/* Check if there is space for the object header */
+		if ((char *)(obj + 1) - data > size) {
+			tr_err(&ipc_tr, "obj header overflow, %u > %u",


resolved by mistake?

src/ipc/ipc4/helper.c

lyakh · 2025-12-17T08:33:08Z

src/audio/module_adapter/module_adapter_ipc4.c

-				return NULL;
+				comp_warn(dev, "dp_data object too small %zu < %zu",
+					  obj->object_words * sizeof(uint32_t), sizeof(*dp_data));
+				break;


why is this not fatal any more? Also break only takes you out of switch, not out of the loop, is that correct?

Looks a bit suspicious, this goes on to access "dp_data->interim_heap_bytes" from the too small object. This does not look correct.

This is mandatory if we go and redefine "struct ipc4_module_init_ext_obj_dp_data". Otherwise payloads sent by an older linux version would always fail dp module creation.

But fear not, non of the data in the too short data object is used. Its all ignored as all of that code accessing the data is within the case block. And we need to continue the decoding loop in case there is some other objects there that we can decode correctly.

lyakh · 2025-12-17T08:36:34Z

src/audio/module_adapter/module_adapter_ipc4.c

 			const struct ipc4_module_init_ext_obj_dp_data *dp_data =
 				(const struct ipc4_module_init_ext_obj_dp_data *)(obj + 1);

 			if (obj->object_words * sizeof(uint32_t) < sizeof(*dp_data)) {


to the commit message, I understand by "released" you mean it's already a part of the ABI? Maybe better use that wording, since I first thought "released" meant "freed" which was confusing. (unrelated to this commit) just noticing, that module_ext_init_decode() returns a pointer of type struct ipc4_base_module_extended_cfg which should only be appropriate for the base module? Strange

Will fix the commit message.

The reason is that the base config follows ext_init in the payload, and if there is anything left aftere ext_init, it should be base config. Anyway, this is going way when I move module_ext_init_decode()-call to module_adapter_new_ext().

lyakh · 2025-12-17T08:45:34Z

src/audio/module_adapter/module_adapter.c

+	 * words from stack,
+	 */
+	struct ipc_config_process spec =
+		*((struct ipc_config_process *) const_spec);


please use const for type-casting just for the peace of mind of imperfect humans like me :-)

src/audio/module_adapter/module_adapter.c

lgirdwood

A few opens on my end but looking like its almost ready.

lgirdwood · 2025-12-17T12:18:41Z

tools/topology/topology2/include/components/widget-common.conf

+}
+
+## Lifetime heap memory allocation requirement in bytes for this component.
+## This token's value indicates the amount of allocated memory needed at component init phase


and used over the lifetime of the component until the component is destroyed.

lgirdwood · 2025-12-17T12:19:40Z

tools/topology/topology2/include/components/widget-common.conf


-## Heap size requirement in bytes for this component. Zero indicates default heap size.
-DefineAttribute."heap_bytes_requirement" {
+## Interim Heap size requirement in bytes for this component.


used for resources that may change in size or be freed during the lifetime of the component.

This is still a bit hard to understand. Maybe add an example for each type (does not have be real, just a practical example of what an algorithm module could use this heap type for). I think that will make this easier follow.

you could state kcontrols, fast get/put(), as examples.

lgirdwood · 2025-12-17T12:20:45Z

tools/topology/topology2/include/components/widget-common.conf

+}
+
+## Shared memory allocation requirement in bytes for this component.
+## this token's value indicates the amount of shared memory the component may need to allocated


that may be shared to other components.

src/ipc/ipc4/helper.c

lgirdwood · 2025-12-17T12:32:47Z

src/include/sof/audio/pipeline.h

+struct pipeline_params {
+	uint32_t domain_id;
+	size_t stack_bytes;
+	size_t lifetime_bytes;
+	size_t interim_bytes;
+	size_t shared_bytes;
+};


Can we add more doxygen comments here for this struct and members

lgirdwood · 2025-12-17T12:36:01Z

tools/topology/topology2/include/components/widget-common.conf

 	token_ref	"comp.word"
 }

+# These default values are here until we have measured module specific numbers


Will these be in the base class and be default until overridden by each module ?

Yes. This was fastest way to get everything working. I think we should eventually remove them when we have component specific default values in place.

kv2019i

Observations inline

kv2019i · 2025-12-17T13:58:44Z

tools/topology/topology2/include/components/widget-common.conf


-## Heap size requirement in bytes for this component. Zero indicates default heap size.
-DefineAttribute."heap_bytes_requirement" {
+## Interim Heap size requirement in bytes for this component.


This is still a bit hard to understand. Maybe add an example for each type (does not have be real, just a practical example of what an algorithm module could use this heap type for). I think that will make this easier follow.

kv2019i · 2025-12-17T14:05:39Z

src/ipc/ipc4/helper.c

-	dcache_invalidate_region((__sparse_force void __sparse_cache *)MAILBOX_HOSTBOX_BASE,
-				 MAILBOX_HOSTBOX_SIZE);
+	sys_cache_data_invd_range((__sparse_force void __sparse_cache *)MAILBOX_HOSTBOX_BASE,
+				 ipc_config.ipc_config_size);


These are good additions. This is overlapping a bit with my IPC payload access rework #10346 but that's stuck now due to dependencies, so no problem with you modifying these here.

@kv2019i & @lyakh since apparently the library builds do not require cache invalidation (I assume library builds have been tested), shouldn't this whole cache invalidation code be behind #ifndef CONFIG_LIBRARY ? (I can still use data ptr instead of MAILBOX_HOSTBOX_BASE).

kv2019i · 2025-12-17T14:10:03Z

src/audio/module_adapter/module_adapter_ipc4.c

-				return NULL;
+				comp_warn(dev, "dp_data object too small %zu < %zu",
+					  obj->object_words * sizeof(uint32_t), sizeof(*dp_data));
+				break;


Looks a bit suspicious, this goes on to access "dp_data->interim_heap_bytes" from the too small object. This does not look correct.

kv2019i

Observations inline

jsarha · 2025-12-17T13:50:19Z

tools/topology/topology2/include/common/tokens.conf

 		num_input_audio_formats	415
 		num_output_audio_formats 416
 		no_wname_in_kcontrol_name 417
 		scheduler_domain	418


commit message: Proof reader "fixed" my typoed "shared" to "shaded"

src/ipc/ipc4/helper.c

jsarha · 2025-12-17T14:11:37Z

src/ipc/ipc4/helper.c

+
+		/* Check if there is space for the object header */
+		if ((char *)(obj + 1) - data > size) {
+			tr_err(&ipc_tr, "obj header overflow, %u > %u",


Argh... I fixed this in one place already, but the copy-pasted version is still in the original form.

src/ipc/ipc4/helper.c

jsarha · 2025-12-17T14:38:09Z

src/ipc/ipc4/helper.c

-	dcache_invalidate_region((__sparse_force void __sparse_cache *)MAILBOX_HOSTBOX_BASE,
-				 MAILBOX_HOSTBOX_SIZE);
+	sys_cache_data_invd_range((__sparse_force void __sparse_cache *)MAILBOX_HOSTBOX_BASE,
+				 ipc_config.ipc_config_size);


@kv2019i & @lyakh since apparently the library builds do not require cache invalidation (I assume library builds have been tested), shouldn't this whole cache invalidation code be behind #ifndef CONFIG_LIBRARY ? (I can still use data ptr instead of MAILBOX_HOSTBOX_BASE).

src/ipc/ipc4/helper.c

jsarha · 2025-12-17T14:47:25Z

tools/topology/topology2/include/components/widget-common.conf

 	token_ref	"comp.word"
 }

+# These default values are here until we have measured module specific numbers


Yes. This was fastest way to get everything working. I think we should eventually remove them when we have component specific default values in place.

jsarha · 2025-12-17T14:51:40Z

src/audio/module_adapter/module_adapter_ipc4.c

 			const struct ipc4_module_init_ext_obj_dp_data *dp_data =
 				(const struct ipc4_module_init_ext_obj_dp_data *)(obj + 1);

 			if (obj->object_words * sizeof(uint32_t) < sizeof(*dp_data)) {


Will fix the commit message.

The reason is that the base config follows ext_init in the payload, and if there is anything left aftere ext_init, it should be base config. Anyway, this is going way when I move module_ext_init_decode()-call to module_adapter_new_ext().

jsarha · 2025-12-17T14:58:35Z

src/audio/module_adapter/module_adapter_ipc4.c

-				return NULL;
+				comp_warn(dev, "dp_data object too small %zu < %zu",
+					  obj->object_words * sizeof(uint32_t), sizeof(*dp_data));
+				break;


This is mandatory if we go and redefine "struct ipc4_module_init_ext_obj_dp_data". Otherwise payloads sent by an older linux version would always fail dp module creation.

But fear not, non of the data in the too short data object is used. Its all ignored as all of that code accessing the data is within the case block. And we need to continue the decoding loop in case there is some other objects there that we can decode correctly.

lyakh · 2025-12-17T15:39:37Z

I am bit surprised that that castiing a const pointer to non const integer pointer does not cause a warning

@jsarha I think that's exactly how it's designed - the compiler respects the explicit cast and relies on the coder knowing what they're doing. Without a cast it would warn. Then there's a bunch of compiler flags to modify that "naive default" of which I certainly don't know all.

Add lifetime_bytes_requirement and shared_bytes_requirement widget attributes. This token's value indicates the amount of heap memory needed at component initialization phase. Signed-off-by: Jyri Sarha <[email protected]>

Adds structs and definitions for adding a payload to struct ipc4_pipeline_create message. The structure of the payload is very similar to struct ipc4_module_init_ext_init payload for struct ipc4_module_init_instance. Signed-off-by: Jyri Sarha <[email protected]>

Add ipc4_create_pipeline_payload_decode() and call it if struct ipc4_pipeline_create's extension.r.payload bit is set. The function decodes the message payload, logs the contents, but does not store the information anywhere. Signed-off-by: Jyri Sarha <[email protected]>

Invalidate cache for only the module init payload size not the whole mailbox. Signed-off-by: Jyri Sarha <[email protected]>

Pass create pipeline payload parameters down to pipeline_new() by using struct pipeline_params. That is the place where the data will be used. Signed-off-by: Jyri Sarha <[email protected]>

…mory Add default values for stack, lifetime heap, interim heap, and shared memory sizes for all widgets, e.g. module instances. These values were tested with vpages branch and all cases that I tried worked with them. Eventually each module should be added with more accurate values, based on measurements. Signed-off-by: Jyri Sarha <[email protected]>

…data Update struct ipc4_module_init_ext_obj_dp_data to match what is required for latest user space features. This is a tricky change as it chenges already part of ABI. However, as the structure was not really used for anything before, changing it should be safe. That is with one exception. The case where an earlier SOF driver sends us the old smaller struct identifier with IPC4_MOD_INIT_DATA_ID_DP_DATA, then we should not fail on that, but only ignore the struct. This is why "return" is changed to "break" in case IPC4_MOD_INIT_DATA_ID_DP_DATA. Signed-off-by: Jyri Sarha <[email protected]>

Move module_ext_init_decode()-call earlier in call chain so that struct ipc4_module_init_ext_obj_dp_data is available when module_adapter_mem_alloc() is called. That is, if the struct ipc4_module_init_ext_object is present in the struct ipc4_module_init_ext_init payload. To accomplish this I needed to redesign how module_ext_init_decode() works and how its called a bit. Signed-off-by: Jyri Sarha <[email protected]>

jsarha · 2025-12-18T20:00:35Z

Tobe squashed commits squashed.

jsarha mentioned this pull request Sep 26, 2025

New memory domain logic thesofproject/linux#5537

Open

lyakh reviewed Sep 29, 2025

View reviewed changes

src/include/ipc4/pipeline.h Show resolved Hide resolved

lgirdwood reviewed Sep 30, 2025

View reviewed changes

src/include/ipc4/pipeline.h Show resolved Hide resolved

jsarha force-pushed the pipeline_memory_attributes branch from ac3bf18 to 5c3c095 Compare September 30, 2025 20:59

lyakh reviewed Oct 1, 2025

View reviewed changes

src/ipc/ipc4/helper.c Outdated Show resolved Hide resolved

lyakh changed the title ~~Code to decode pipline create messages payload~~ Code to decode pipeline create messages payload Oct 2, 2025

lyakh reviewed Oct 6, 2025

View reviewed changes

src/ipc/ipc4/helper.c Outdated Show resolved Hide resolved

src/ipc/ipc4/helper.c Outdated Show resolved Hide resolved

jsarha force-pushed the pipeline_memory_attributes branch from e43cb7b to 9f5d981 Compare October 7, 2025 21:37

jsarha force-pushed the pipeline_memory_attributes branch 2 times, most recently from 2dc8ac4 to 7e4bb9a Compare November 11, 2025 18:06

lyakh reviewed Nov 14, 2025

View reviewed changes

jsarha force-pushed the pipeline_memory_attributes branch from 7e4bb9a to d9de565 Compare December 3, 2025 19:17

jsarha force-pushed the pipeline_memory_attributes branch from d9de565 to 4d8a457 Compare December 14, 2025 23:14

jsarha marked this pull request as ready for review December 16, 2025 16:02

Copilot AI review requested due to automatic review settings December 16, 2025 16:02

jsarha requested review from RanderWang, dbaluta, jxstelter, kv2019i, lbetlej, marcinszkudlinski, mmaka1, pblaszko, plbossart and ranj063 as code owners December 16, 2025 16:02

Copilot AI reviewed Dec 16, 2025

View reviewed changes

jsarha force-pushed the pipeline_memory_attributes branch 2 times, most recently from 7dd6358 to 7727c82 Compare December 16, 2025 21:52

audio: module_adapter: Use %zu in size_t prints in module_ext_init_de…

ea8b7cc

…code() Use %zu in size_t prints in module_ext_init_decode(). Signed-off-by: Jyri Sarha <[email protected]>

lyakh reviewed Dec 17, 2025

View reviewed changes

lgirdwood reviewed Dec 17, 2025

View reviewed changes

kv2019i reviewed Dec 17, 2025

View reviewed changes

jsarha commented Dec 17, 2025

View reviewed changes

Jyri Sarha added 2 commits December 17, 2025 18:24

jsarha force-pushed the pipeline_memory_attributes branch 2 times, most recently from cb378d1 to 2ef7ab4 Compare December 18, 2025 10:16

Jyri Sarha added 6 commits December 18, 2025 21:51

ipc4: helper: Do not invalidate cache of whole mailbox for module init

436647d

Invalidate cache for only the module init payload size not the whole mailbox. Signed-off-by: Jyri Sarha <[email protected]>

pipeline: Pass create pipeline payload parameters down to pipeline_new()

ee95dc8

Pass create pipeline payload parameters down to pipeline_new() by using struct pipeline_params. That is the place where the data will be used. Signed-off-by: Jyri Sarha <[email protected]>

jsarha force-pushed the pipeline_memory_attributes branch from 2ef7ab4 to e09a0b1 Compare December 18, 2025 20:00

Code to decode pipeline create messages payload #10265

Are you sure you want to change the base?

Code to decode pipeline create messages payload #10265

Uh oh!

Conversation

jsarha commented Sep 26, 2025

Uh oh!

Uh oh!

lgirdwood left a comment

Choose a reason for hiding this comment

Uh oh!

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!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

jsarha commented Dec 3, 2025

Uh oh!

jsarha commented Dec 14, 2025

Uh oh!

jsarha commented Dec 16, 2025

Uh oh!

jsarha commented Dec 16, 2025

Uh oh!

lgirdwood commented Dec 16, 2025

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jsarha commented Dec 16, 2025

Uh oh!

jsarha commented Dec 16, 2025

Uh oh!

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!

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!

lgirdwood 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!

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