Update 2026-11-16: still broken on 7.0.0-43.43 — toolchain ruled out,
target narrowed to the 7.0.0-31.31 → 7.0.0-38.38 SoundWire delta

  ## Current state: still BAD on the newest resolute kernel
(7.0.0-43.43)

  The built-in PDM DMIC still captures pure digital silence. Direct
capture, bypassing the whole session stack:

      arecord -D hw:CARD=sofsoundwire,DEV=4 -f S16_LE -r 48000 -c 2 -d 2
      per-0.1s peaks: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]   peak=0  rms=0.0

  Same signature as 7.0.0-38.38 and 7.0.0-42.42. Codec routing re-verified on 
-43: `rt713-dmic FU1E Capture Switch` on/on/on/on, FU1E Capture Volume 54, FU15 
Boost 0, ADC 25 Mux = DMIC1, ADC 26 Mux = DMIC2; PipeWire source 
  `HiFi__Mic__source` at 78%, not muted. `journalctl -k -b` shows ZERO 
SoundWire/SOF/rt713 errors — firmware boots, topology loads, all codecs 
enumerate, the stream runs and delivers zeros.

  Updated matrix (all resolute unless noted; build = dpkg version):

  | Kernel build | Userspace | Result |
  |---|---|---|
  | 7.0.0-30.30~24.04.1 (noble-hwe) | 24.04 | GOOD (jack + dmic) |
  | 7.0.0-30.30 (resolute) | 26.04 GA | BAD |
  | 7.0.0-31.31 (resolute) | 26.04 GA | BAD |
  | 7.0.0-31.31 (resolute) | 26.04.1 | GOOD (verified 2026-09-09; package since 
purged from the archive) |
  | 7.0.0-38.38 (resolute) | 26.04.1 | BAD (dmic digital silence) |
  | 7.0.0-42.42 (resolute) | 26.04.1 | BAD (dmic digital silence) |
  | 7.0.0-43.43 (resolute) | 26.04.1 | BAD (dmic digital silence) — verified 
2026-11-16 |

  ## Ruled out: compiler / build toolchain

  The good and the bad *resolute* builds were produced by the same
compiler, so this is not a toolchain regression:

  - `readelf -p .comment` on the shipped modules: 7.0.0-30.30~24.04.1 (noble, 
GOOD) = `(Ubuntu 13.3.0-6ubuntu2~24.04.1) 13.3.0`; 7.0.0-42.42 and 7.0.0-43.43 
(BAD) = `(Ubuntu 15.2.0-16ubuntu1) 15.2.0` (checked on 
  `snd-sof-intel-hda-common`, 
`snd-soc-rt711/rt712-sdca/rt712-sdca-dmic/rt715/rt715-sdca`).
  - Launchpad build 33458794 (`linux` 7.0.0-31.31, resolute amd64 — the GOOD 
build) installed `gcc-15_15.2.0-16ubuntu1`: same GCC 15.2.0 as the bad builds.
  - Kernel config: diffing the whole `CONFIG_SND*` / `CONFIG_SOUND*` / 
`CONFIG_SDW*` / `CONFIG_SOF*` set between `/boot/config-7.0.0-30-generic` and 
`/boot/config-7.0.0-43-generic` yields exactly one difference: 
  `CONFIG_SND_SOC_ACPI_AMD_SDCA_QUIRKS=m` → `=y` (AMD-only; introduced by the 
-38 SRU as "[Config] Adjust config SND_SOC_ACPI_AMD_SDCA_QUIRKS").

  So the audio stack is configured identically and compiled by the same
GCC in the good and bad resolute builds. The breakage must lie in the
source delta between 7.0.0-31.31 and 7.0.0-38.38.

  ## Candidate commits in that delta (generic SoundWire init path)

  The -38 SRU changes SoundWire device initialisation, which the rt713
PDM capture path depends on:

  - `soundwire: don't program SDW_SCP_BUSCLOCK_SCALE on a unattached Peripheral`
  - `soundwire: Add a helper function to wait for device initialisation` (+ 
"Move wait for initialisation helper to header")
  - `soundwire: intel_ace2x: release bpt_stream when close it`
  - `ASoC: SOF: Intel: select SND_SOC_SDW_UTILS=y from 
SND_SOC_SOF_HDA_GENERIC=y`
  - `[Config] Adjust config SND_SOC_ACPI_AMD_SDCA_QUIRKS`
  - plus the 2026-08-26 upstream stable patchset.

  Hypothesis: a bus-clock / device-init timing change in the generic SoundWire 
path leaves the rt713 PDM DMIC streaming but silent — which matches a capture 
path that reports zero errors and delivers zeros. The 
  `SDW_SCP_BUSCLOCK_SCALE` commit is the first thing I would test.

  ## What I can do next

  Happy to test anything: a test kernel, a build with one of the above 
reverted, or to collect traces you prefer (SoundWire debugfs register dumps, 
SOF IPC logs, `sdw` tracepoints, `alsabat` loopback). Tell me which and I will 
  report back with the same level of detail as above.

  Note: the bug title still describes "headphone jack silent on 7.0.0-31", 
which no longer matches the live problem. Suggest retitling to something like 
"[regression] sof-soundwire rt713 (Dell Precision 5690 / Meteor Lake): PDM DMIC
  captures digital silence on resolute kernels; noble-hwe build OK" — happy to 
do that if it helps triage.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2166754

Title:
  [regression] sof-soundwire rt713: headphone jack silent on 7.0.0-31
  (worked on 7.0.0-30, broken on 7.0.0-28) — Dell Precision 5690 /
  Meteor Lake

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2166754/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to