[BUG] LNL nocodec alsabat-capture fail to broken pipe -32

[kv2019i]

Describe the bug
Follow-up to #9123 and filing a separate bug to track the case of alsabat-capture test fail, with test failing to an overrun.

To Reproduce
Seen in SOF Intel CI testing on LNL devices.

Reproduction Rate
<1/10

Expected behavior
alsabat test should pass

Impact
audio glitches in output

Environment

Linux Branch: topic/sof-dev
Linux Commit: 2fdeeded4e06
KConfig Branch: master
KConfig Commit: 8189104a4f38
SOF Branch: main
SOF Commit: f383b0f
Zephyr Commit: 689d1edee1d5

Screenshots or console output

2024-09-08 14:48:42 UTC [REMOTE_COMMAND] alsabat -Phw:sofnocodec,0 --standalone -n 240000 -r 48000 -c 2 -f S16_LE -F 997 -k 2.1
2024-09-08 14:48:43 UTC [REMOTE_COMMAND] alsabat -Chw:sofnocodec,0 -c 2 -r 48000 -f S16_LE -F 997 -k 2.1
alsa-utils version 1.2.6

Entering playback thread (ALSA).
Get period size: 3000  buffer size: 24000
Playing generated audio sine wave
Playback completed.

Return value is 0
Overrun: Broken pipe(-32)
 FAIL: Peak freq too low 978.52 Hz
 FAIL: Peak freq too low 982.18 Hz

[lrudyX]

@kv2019i could you add link to CI tests results?

[kv2019i]

@lrudyX here's one fresh occurence in CI https://sof-ci.01.org/sofpr/PR9454/build8003/devicetest/index.html?model=LNLM_RVP_NOCODEC&testcase=check-alsabat-headset-capture-997

alsabat is not strictly needed if you have other means to check the captured audio. I.e. a test with SSP playback + capture with SSP loopback should trigger this problem. But only occurs on some of the DUTs.