Re: [PATCH v2 00/10] soundwire: Fix Critical and High issues reported by Sashiko

From: Lee Jones

Date: Tue Nov 10 2026 - 04:13:52 EST


On Sat, 07 Nov 2026, HyeongJun An wrote:

> On a Dell Latitude 9420 (Tiger Lake, SOF IPC3, sof_sdw with rt711-sdca
> on link 0, two rt1316 on links 1 and 2 that are aggregated into one
> Speaker stream, and rt714 on link 3), I tested this series on top of
> soundwire next at commit afd88e7ad781. Only patches 1-4 run on this
> machine. The intel_ace2x and qcom patches are not exercised here.
>
> With and without the series:
> - all four peripherals attach at boot
> - playback and capture work on every endpoint
> - 40 start/stop cycles of the multi-link Speaker stream (10 of them
> after the links had runtime suspended), all links streaming at the
> same time, and Jack Out start/stop while Speaker was playing all
> passed
> - ALSA pcm-test: 24 pass, 0 fail
> - the dev-properties, dp0 and dpN sysfs attributes of all four
> peripherals read back identical (the words arrays are empty on these
> codecs, so only the trailing newline goes through the changed code)
>
> With the series, s2idle with Speaker playback and Microphone capture
> running: all four peripherals were attached again after resume, both
> streams continued after aplay/arecord re-prepared them, and playback and
> capture on every endpoint still worked. There were no SoundWire or SOF
> kernel messages. I did not run the same s2idle test without the series.
>
> These runs only take the normal paths, so the failure cases that
> patches 3 and 4 guard against were not hit.
>
> I ran these with Claude Code (Claude Opus 5.5) on the laptop.
>
> I am sending Tested-by on patches 1-4 only, in replies to those
> patches.

I really appreciate you taking the time to test these. Thank you.

--
Lee Jones