[BUG] Dynabook Portege X30W-K (ADL): missing RT711@l0 + RT714@l1 + RT1316@l2+l3 SoundWire machine match / topology

[groxaxo]

Summary

Dynabook Portege X30W-K / PDA31A on Alder Lake has silent internal speakers because its actual SoundWire layout is:

• RT711 on link 0 (headset codec)
• RT714 on link 1 (digital mic)
• RT1316 on link 2 (speaker amp)
• RT1316 on link 3 (speaker amp)

Linux 6.17 does not appear to have a matching ADL SoundWire machine entry for this exact layout, so the SoundWire speaker amps are never brought up.

System

• Laptop: Dynabook Portege X30W-K / PDA31A
• OS: Ubuntu 24.04.4 LTS
• Kernel: 6.17.0-1023-oem
• PCI audio: 8086:51c8 (Intel Alder Lake PCH-P HD Audio / SoundWire host)
• HDA subsystem observed on machine: 3100:011f

Evidence

Decoded from the machine DSDT _ADR values:

ACPI dev	SDW link	Codec	Role
SWD0	0	RT711	headset codec
SWD3	1	RT714	digital mic
SWD2	2	RT1316	speaker amp (left)
SWD1	3	RT1316	speaker amp (right)

The DSDT also contains the Realtek smart-amp marker:

"mipi-rtk-part-smartamp-private-prop"

Public artifacts:

• Full diagnosis: https://github.com/groxaxo/dynabook-x30w-k-ubuntu-audio/blob/80ddfeb2367eca3f01673faf0e8d127995f955b8/docs/soundwire-sdca-diagnosis-2026-05-25.md
• Raw DSDT: https://github.com/groxaxo/dynabook-x30w-k-ubuntu-audio/blob/80ddfeb2367eca3f01673faf0e8d127995f955b8/diagnostics/dsdt.dsl.gz
• Decoded SoundWire table: https://github.com/groxaxo/dynabook-x30w-k-ubuntu-audio/blob/80ddfeb2367eca3f01673faf0e8d127995f955b8/diagnostics/soundwire-topology.txt

Why I think the missing piece is the ADL SoundWire machine match

In sound/soc/intel/common/soc-acpi-intel-adl-match.c I could only find these close 4-link RT711/RT714/RT1316 combinations:

• sof-adl-rt711-l0-rt1316-l12-rt714-l3
• sof-adl-rt711-l0-rt1316-l13-rt714-l2
• sof-adl-rt711-l2-rt1316-l01-rt714-l3

My hardware is instead:

• RT711 @ l0
• RT714 @ l1
• RT1316 @ l2 + l3

I could not find a corresponding machine entry or prebuilt topology for:

sof-adl-rt711-l0-rt1316-l23-rt714-l1.tplg

Current behavior

To keep the DMIC working, I currently force an HDA-only topology:

intel/sof-tplg/sof-hda-generic-2ch-pdm1.tplg

That makes the mic usable, but the speakers remain silent because the SoundWire links are never brought up. Relevant dmesg lines:

sof-audio-pci-intel-tgl 0000:00:1f.3: using HDA machine driver skl_hda_dsp_generic now
sof-audio-pci-intel-tgl 0000:00:1f.3: Topology file: intel/sof-tplg/sof-hda-generic-2ch-pdm1.tplg

Earlier HDA codec experiments (GPIO/EAPD/coefficients) did nothing, which now makes sense because the internal speakers are not attached to the HDA codec path at all.

Requested fix

Could you please confirm whether this platform needs:

1. a new snd_soc_acpi_intel_adl_sdw_machines[] match entry for RT711@l0 + RT714@l1 + RT1316@l2+l3, and
2. a matching topology file for that layout

For convenience, I included a proposed skeleton in the diagnosis doc linked above.

If this belongs in another tracker or mailing list instead, I’m happy to mirror it there.

[groxaxo]

Follow-up: deeper analysis of what's needed

After reviewing the match table in Linux 7.1-rc4 (master), I can provide more specifics to help the patch author.

What's already in 7.1-rc4

All four link_mask=0xF entries in adl_sdw_machines[] have RT714 on link 0, 2, or 3:

adl_default           → RT711@l0, RT1308@l1+l2, RT715@l3     (0xF)
adl_sdca_default      → RT711@l0, RT1316@l1+l2, RT714@l3     (0xF)
adl_sdca_3_in_1       → RT711@l0, RT1316@l1+l3, RT714@l2     (0xF)
adl_sdw_rt711_link2…  → RT711@l2, RT1316@l0+l1, RT714@l3     (0xF)

My hardware needs: RT711@l0, RT714@l1, RT1316@l2+l3 — not present in any version.

What's missing in the source

Not only is the machine entry missing, but the ADR structure for RT714 on link 1 is also absent:

/* These exist: */
static const struct snd_soc_acpi_adr_device rt714_0_adr[] = { .adr = 0x000030025D071401ull, ... };
static const struct snd_soc_acpi_adr_device rt714_2_adr[] = { .adr = 0x000230025D071401ull, ... };
static const struct snd_soc_acpi_adr_device rt714_3_adr[] = { .adr = 0x000330025D071401ull, ... };

/* This is MISSING: */
static const struct snd_soc_acpi_adr_device rt714_1_adr[] = { .adr = 0x000130025D071401ull, ... };

Minimal patch skeleton

/* 1. Add new ADR structure (after rt714_3_adr, ~line 280) */
static const struct snd_soc_acpi_adr_device rt714_1_adr[] = {
{
um_endpoints = 1,
dpoints = &single_endpoint,
ame_prefix = "rt714"
}
};

/* 2. Add new link array */
static const struct snd_soc_acpi_link_adr adl_sdca_rt711_l0_rt714_l1_rt1316_l23[] = {
{ .mask = BIT(0), .num_adr = ARRAY_SIZE(rt711_sdca_0_adr),    .adr_d = rt711_sdca_0_adr    },
{ .mask = BIT(1), .num_adr = ARRAY_SIZE(rt714_1_adr),         .adr_d = rt714_1_adr          },
{ .mask = BIT(2), .num_adr = ARRAY_SIZE(rt1316_2_group1_adr), .adr_d = rt1316_2_group1_adr  },
{ .mask = BIT(3), .num_adr = ARRAY_SIZE(rt1316_3_group1_adr), .adr_d = rt1316_3_group1_adr  },
{}
};

/* 3. Add machine entry in snd_soc_acpi_intel_adl_sdw_machines[] */
{
.link_mask = 0xF,
.links = adl_sdca_rt711_l0_rt714_l1_rt1316_l23,
.drv_name = "sof_sdw",
.sof_tplg_filename = "sof-adl-rt711-l0-rt1316-l23-rt714-l1.tplg",
},

The topology filename sof-adl-rt711-l0-rt1316-l23-rt714-l1.tplg would need to be compiled and shipped (tracked separately in thesofproject/sof#10805).

Happy to test any branch or topology file.

[groxaxo]

Follow-up: live forced SoundWire test results

I tried a live forced-SoundWire test on the affected Dynabook X30W-K with Ubuntu OEM 6.17.0-1023-oem, using kernel modules rebuilt from the matching Ubuntu source tag applied/6.17.0-1023.23.

Test changes applied locally:

• added the missing ADL machine entry for RT711@l0 + RT714@l1 + RT1316@l2+l3
• installed an experimental sof-adl-rt711-l0-rt1316-l23-rt714-l1.tplg
• forced snd-intel-sdw-acpi sdw_link_mask=0xf
• temporarily ignored the firmware SoundWire bus-disable quirk in soundwire-intel
• temporarily masked the ALC257 HDA codec with snd-sof-intel-hda codec_mask=0x4 so HDA generic would not win machine selection

Good news: this got past the earlier selection/topology blockers. The kernel selected the intended machine and topology:

sof-audio-pci-intel-tgl 0000:00:1f.3: forcing Dynabook X30W-K SoundWire machine for link mask 0xf
sof-audio-pci-intel-tgl 0000:00:1f.3: Topology file: intel/sof-tplg/sof-adl-rt711-l0-rt1316-l23-rt714-l1-2ch.tplg

Bad news: probe then failed during SoundWire bus startup before an ALSA card was created:

soundwire_intel soundwire_intel.link.0: Bus clash for control word
soundwire_intel soundwire_intel.link.0: Bus clash for data word
soundwire_intel soundwire_intel.link.1: Bus clash for control word
soundwire_intel soundwire_intel.link.1: Bus clash for data word
soundwire_intel soundwire_intel.link.2: Bus clash for control word
soundwire_intel soundwire_intel.link.2: Bus clash for data word
soundwire_intel soundwire_intel.link.3: Bus clash for control word
soundwire_intel soundwire_intel.link.3: Bus clash for data word
soundwire_intel soundwire_intel.link.0: Clock stop failed -110
soundwire_intel soundwire_intel.link.1: Clock stop failed -110
soundwire_intel soundwire_intel.link.2: Clock stop failed -110
soundwire_intel soundwire_intel.link.3: Clock stop failed -110

During the failed forced state, sysfs only showed the four masters and two RT711-like devices:

sdw-master-0-0
sdw-master-0-1
sdw-master-0-2
sdw-master-0-3
sdw:0:0:025d:0711:00
sdw:0:1:025d:0711:00

The RT1316 speaker amps on links 2 and 3 did not enumerate.

Cross-check: when I removed codec_mask=0x4, the system went back to the HDA generic machine:

hda codecs found, mask 5
using HDA machine driver skl_hda_dsp_generic now
Topology file: intel/sof-tplg/sof-hda-generic-2ch.tplg

So the machine entry/topology still look necessary, but not sufficient on this laptop. The next blocker appears to be the firmware/ACPI SoundWire link-enable path: firmware reports the SoundWire bus disabled for Linux, and forcing past that state causes bus clashes/timeouts before the RT1316 amps can appear. This may need a Dynabook-specific lower-level quirk or the missing power/link sequence that Windows performs before exposing the SoundWire amps.