[DNM] dai: add support for Intel UAOL DAI #9227

tlissows · 2024-06-14T07:44:28Z

This adds support for Intel USB Audio Offload Link (UAOL) DAI.

This PR needs another (Zephyr) PR drivers: dai: add DAI driver for Intel UAOL to be merged first.

src/audio/base_fw_intel.c

src/audio/copier/copier_dai.c

src/lib/dai.c

plbossart

Not following the directions, this doesn't seem aligned with what we previously did for ALH / SoundWire.

plbossart · 2024-06-17T07:21:12Z

src/ipc/ipc4/dai.c

+#else
+		channel = copier_cfg->gtw_cfg.node_id.f.v_index;
+#endif
+		break;


I don't fully understand this.

for LNL+ all the DMA stuff is handled in the same way with an HDaudio-based solution.

We just moved some of the ALH stuff into the HDaudio prodessing - see line 94 the test for ALH.

so shouldn't the UAOL stuff also be moved under the HDAudio category? Why special case what looks identical?

my comment still stands.

@tlissows any update here ?

Removed support for MTL, so the line is no longer present.
However the concept of handling UAOL by HDA code (and removing of all UAOL related stuff from SOF) seems not to be applicable. This is because of UAOL being a type of DAI handled by separate Zephyr driver with the specific API.

plbossart

we really need to make UAOL support dependent on a version of silicon (pick one), e.g. ACE2.0, and not add code that will never be used on previous generations.

There's really no point in supporting anything older than ACE2.0?

src/ipc/ipc4/dai.c

plbossart · 2024-07-19T11:25:00Z

src/ipc/ipc4/dai.c

+#else
+		channel = copier_cfg->gtw_cfg.node_id.f.v_index;
+#endif
+		break;


my comment still stands.

src/audio/copier/copier_dai.c

src/lib/dai.c

lgirdwood · 2024-09-12T16:24:46Z

@tlissows ping ?

tlissows · 2024-10-02T11:58:02Z

Removed support for ACE1.x platform from this PR. Also aligned to the changes introduced by PR basefw: Add handling of IPC4_DMA_CONTROL messages.

lgirdwood

@tlissows I assume still undergoing testing hence the DNM ? Pls also respond to open from @plbossart

lgirdwood · 2024-10-02T13:55:12Z

src/audio/dai-zephyr.c

@@ -153,25 +154,25 @@ int dai_set_config(struct dai *dai, struct ipc_config_dai *common_config,
 	switch (common_config->type) {
 	case SOF_DAI_INTEL_SSP:
 		cfg.type = is_blob ? DAI_INTEL_SSP_NHLT : DAI_INTEL_SSP;
-		cfg_params = is_blob ? spec_config : &sof_cfg->ssp;
+		cfg_params = is_blob ? (void *)&gtw_cfg->config_data : &sof_cfg->ssp;


Do we need the void* cast here or is the compiler throwing up a warning in this case ? cfg_params is also const too.

The cast was added to avoid compiler warning like this (being not related to cfg_params):

pointer type mismatch ('uint32_t const (*)[1]' and 'const struct sof_ipc_dai_ssp_params *') [-Wpointer-type-mismatch] cfg_params = is_blob ? &gtw_cfg->config_data : &sof_cfg->ssp; ^ ~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~

Changing for two casts for better readability:
cfg_params = is_blob ? (void *)&gtw_cfg->config_data : (void *)&sof_cfg->ssp;

src/audio/base_fw_intel.c

lgirdwood · 2024-10-02T13:59:06Z

src/ipc/ipc4/dai.c

+#else
+		channel = copier_cfg->gtw_cfg.node_id.f.v_index;
+#endif
+		break;


@tlissows any update here ?

tmleman

I have no remarks regarding the code, but I would like to see that the whole thing passes through CI. There is a temporary commit missing here that would indicate the Zephyr version from your second pull request:

--- a/west.yml
+++ b/west.yml
@@ -43,7 +43,7 @@ manifest:

     - name: zephyr
       repo-path: zephyr
-      revision: 99e6280d7e22552de9a94992b626acdcbde00fee
+      revision: pull/69906/head
       remote: zephyrproject

       # Import some projects listed in zephyr/west.yml@revision

lgirdwood · 2024-11-04T15:32:29Z

@tlissows @tmleman it looks like this is good to go once it passes CI - I think @plbossart comments have now been addressed. Are we blocking on any Zephyr PRs before we can merge ?

Currently, the dai_set_config() function is called with copier gateway config data passed but without the config length, possibly making the config not parsable for some DAI drivers. This patch adds passing of the full ipc4_copier_gateway_cfg structure, giving DAI drivers the ability to access config length if needed. Signed-off-by: Tomasz Lissowski <tomasz.lissowski@intel.com>

Add strict use of dai_type enum items as a call parameter for dai_get_device() function. Currently, both the dai_type and sof_ipc_dai_type items are used inconsistently. Signed-off-by: Tomasz Lissowski <tomasz.lissowski@intel.com>

This adds support for Intel USB Audio Offload Link (UAOL) DAI. Signed-off-by: Tomasz Lissowski <tomasz.lissowski@intel.com>

Signed-off-by: Tomasz Lissowski <tomasz.lissowski@intel.com>

lgirdwood

@tlissows looks like we have a panic in a legacy playback use case test reported by CI.

tlissows requested review from RanderWang, abonislawski, pblaszko, marcinszkudlinski, lgirdwood, plbossart, mmaka1, lbetlej, dbaluta and kv2019i as code owners June 14, 2024 07:44

abonislawski reviewed Jun 14, 2024

View reviewed changes

src/audio/base_fw_intel.c Outdated Show resolved Hide resolved

lyakh reviewed Jun 17, 2024

View reviewed changes

src/audio/base_fw_intel.c Show resolved Hide resolved

src/audio/base_fw_intel.c Outdated Show resolved Hide resolved

src/audio/base_fw_intel.c Outdated Show resolved Hide resolved

src/audio/copier/copier_dai.c Outdated Show resolved Hide resolved

src/lib/dai.c Show resolved Hide resolved

plbossart reviewed Jun 17, 2024

View reviewed changes

tlissows force-pushed the uaol-dai-upstream branch from b89f3e5 to f6551e3 Compare July 19, 2024 11:02

tlissows changed the title ~~dai: add support for Intel UAOL DAI~~ [DNM] dai: add support for Intel UAOL DAI Jul 19, 2024

plbossart requested changes Jul 19, 2024

View reviewed changes

tlissows force-pushed the uaol-dai-upstream branch from f6551e3 to 11bbbc5 Compare October 2, 2024 07:58

lgirdwood reviewed Oct 2, 2024

View reviewed changes

tmleman reviewed Nov 4, 2024

View reviewed changes

tlissows force-pushed the uaol-dai-upstream branch from 11bbbc5 to 0c5b63a Compare December 17, 2024 11:25

tlissows added 4 commits December 17, 2024 15:33

dai: use dai_type items in calls to dai_get_device()

0ba3ac2

Add strict use of dai_type enum items as a call parameter for dai_get_device() function. Currently, both the dai_type and sof_ipc_dai_type items are used inconsistently. Signed-off-by: Tomasz Lissowski <tomasz.lissowski@intel.com>

dai: add support for Intel UAOL DAI

2613355

This adds support for Intel USB Audio Offload Link (UAOL) DAI. Signed-off-by: Tomasz Lissowski <tomasz.lissowski@intel.com>

west.yml: update zephyr

d3c99ca

Signed-off-by: Tomasz Lissowski <tomasz.lissowski@intel.com>

tlissows force-pushed the uaol-dai-upstream branch from 0c5b63a to d3c99ca Compare December 17, 2024 14:57

lgirdwood reviewed Dec 18, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[DNM] dai: add support for Intel UAOL DAI #9227

[DNM] dai: add support for Intel UAOL DAI #9227

tlissows commented Jun 14, 2024 •

edited

Loading

plbossart left a comment

plbossart Jun 17, 2024

plbossart Jul 19, 2024

lgirdwood Oct 2, 2024

tlissows Dec 17, 2024

plbossart left a comment

plbossart Jul 19, 2024

lgirdwood commented Sep 12, 2024

tlissows commented Oct 2, 2024 •

edited

Loading

lgirdwood left a comment

lgirdwood Oct 2, 2024

tlissows Dec 17, 2024

lgirdwood Oct 2, 2024

tmleman left a comment

lgirdwood commented Nov 4, 2024

lgirdwood left a comment

[DNM] dai: add support for Intel UAOL DAI #9227

Are you sure you want to change the base?

[DNM] dai: add support for Intel UAOL DAI #9227

Conversation

tlissows commented Jun 14, 2024 • edited Loading

plbossart left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

plbossart left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lgirdwood commented Sep 12, 2024

tlissows commented Oct 2, 2024 • edited Loading

lgirdwood left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tmleman left a comment

Choose a reason for hiding this comment

lgirdwood commented Nov 4, 2024

lgirdwood left a comment

Choose a reason for hiding this comment

tlissows commented Jun 14, 2024 •

edited

Loading

tlissows commented Oct 2, 2024 •

edited

Loading