Add PipeWire support to AudioRecord and AudioPlayback tests #137

tmoida · 2025-08-08T10:15:55Z

Updated AudioRecord and AudioPlayback scripts to support both PulseAudio and PipeWire
Added AUDIO_BACKEND variable for backend selection
Updated README files with usage instructions for both backends
Modified LAVA test plan to run tests with both PulseAudio and PipeWire

smuppand

@tmoida Could you please squash the commits into a single one?

tmoida · 2025-08-11T06:58:54Z

Hi @smuppand ,
I have squashed the commits into a single one as requested. Kindly review it and let me know if any further changes are needed.

smuppand

few chnages needed like pre-scan dmesg (use common helpers) -> Run playback (possibly loop) -> post-scan dmesg fir any potential audio-related dmesg messages detected.

better to add a daemon check which prevents false failures when the server isn't running

loops + timeout - Lets you simulate longer playback without hanging CI.

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh

tmoida · 2025-08-14T12:05:30Z

Hi @smuppand,
I've updated the Audio test scripts and documentation as per the feedback. Key improvements include:

- Added daemon checks for both PulseAudio and PipeWire backends
- Replaced eval with safer command execution using sh -c
- Introduced support for environment variables like PLAYBACK_VOLUME, PLAYBACK_TIMEOUT, PLAYBACK_LOOPS, RECORD_TIMEOUT, and RECORD_LOOPS
- Implemented pre/post dmesg scanning using get_kernel_log and scan_dmesg_errors
- Updated both README files to reflect the new features and usage instructions

Please let me know if there is anything else that needs to be addressed.

smuppand · 2025-08-20T03:42:33Z

@tmoida , please squash your commits.

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh

tmoida · 2025-08-20T07:34:10Z

Hi @smuppand,

I've squashed the commits and fixed the Shell Check warning by using the loop variable. Please let me know if there is anything else that needs to be addressed.

Also. I have validated these scripts on Yocto builds, and they work fine. However, on Mainline (Nightly sanity) builds, they are being skipped due to a missing timeout binary. I wanted to check: will the timeout binary be included in future builds, or should I update the script to avoid using it?

smuppand · 2025-08-20T10:57:43Z

Hi @smuppand,

I've squashed the commits and fixed the Shell Check warning by using the loop variable. Please let me know if there is anything else that needs to be addressed.

Also. I have validated these scripts on Yocto builds, and they work fine. However, on Mainline (Nightly sanity) builds, they are being skipped due to a missing timeout binary. I wanted to check: will the timeout binary be included in future builds, or should I update the script to avoid using it?

You can use the library function run_with_timeout() if the build does not include the timeout utility.

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh

tmoida · 2025-08-22T06:41:11Z

Hi @smuppand,

Thanks for the suggestions!

I've updated the loop logic to handle early exits cleanly and added clearer return handling to distinguish between success and timeout cases. Also, I've replaced timeout with run_with_timeout() as a fallback to ensure compatibility on builds where the timeout binary is missing.

Please let me know if there is anything else that needs to be addressed.

smuppand · 2025-08-26T05:55:11Z

@tmoida Please update the https://github.com/qualcomm-linux/qcom-linux-testkit/blob/main/Runner/plans/meta-ar-ci-premerge.yaml to include the changes that involve launching or adding any tests.

tmoida · 2025-08-28T07:09:17Z

Hi @smuppand,
I have tested the standalone scripts on both QLI 1.0 and the QLI mainline (nightly sanity) builds.
Could you please review and merge the PR at your convenience?

smuppand

Thanks for putting this together. I’ve got a few suggestions to make the Audio test more modular and CI-friendly, consistent.

run.sh only: parse flags → call functions in libaudio.sh.
No direct calls to amixer/aplay/tinymix from run.sh (they go via lib).
All shared helpers should wrap functestlib.sh
Happy to re-review once these land.

tmoida · 2025-09-04T10:36:12Z

Hi @smuppand,

Thank you for your feedback.

I have implemented the requested modular structure and made the following changes:

1. Moved all core logic for playback and recording into a shared `libaudio.sh`.
2. Created minimal `run.sh` scripts under AudioPlayback and AudioRecord that only parse CLI flags and invoke functions from `libaudio.sh`.
3. Wrapped all binary calls (paplay, pw-play, parec, pw-record) inside `libaudio.sh`, eliminating direct calls in `run.sh`.
4. Added support for both CLI flags and environment variables for configuration.

Why separate run.sh scripts for playback and record?
Regarding the use of separate run.sh scripts for playback and recording: We are utilizing run-test.sh as the main test runner, which expects each test case to be in its own directory and produce a result file named after that directory (e.g., AudioPlayback.res, AudioRecord.res). Using a single run.sh within a common Audio/ directory would prevent run-test.sh from locating the correct .res file, potentially leading to false negatives, even when tests pass.

To address this and ensure compatibility with run-test.sh, we have maintained separate directories and minimal run.sh wrappers for each test case. This approach allows for correct result parsing while centralizing the logic in libaudio.sh.

All usage details, CLI options, and examples are documented in the respective Read_me.md files for easy reference.

Let me know if you’d like any further refinements!

smuppand

Also woth adding

WAV 16-bit/48k, 44.1k; Ogg Vorbis/Opus (if tools available), AAC.
Short (5s), medium (30s), long (120s) smoke tiers.
Mixer dumps (pactl list, pw-dump) on failure.

smuppand · 2025-09-05T04:01:42Z

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh

+  SEARCH=$(dirname "$SEARCH")
 done
+[ -z "$INIT_ENV" ] && echo "[ERROR] init_env not found" && exit 1
+# shellcheck source=/dev/null


using # shellcheck source=/dev/null is a blunt silencer that hides real problems and degrades analysis. It tells ShellCheck “pretend this sourced file exists but is empty, It masks real bugs.

smuppand · 2025-09-05T04:03:03Z

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh

+. "$TOOLS/libaudio.sh"

-test_path=$(find_test_case_by_name "$TESTNAME")
-cd "$test_path" || exit 1


This approach will fail if we run any test from the run-test root, and the .res file will be generated in the directory where the command is executed, rather than in the test directory. Please keep the expected snippet.

smuppand · 2025-09-05T04:05:05Z

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh

-test_path=$(find_test_case_by_name "$TESTNAME")
-cd "$test_path" || exit 1
+# Override with BusyBox-compatible version
+check_network_status() {


check_network_status override isn’t used—drop it to keep things tight.
Use available built-in fiunctions like get_ip_address() to avoid duplication.

smuppand · 2025-09-05T04:05:49Z

Runner/suites/Multimedia/Audio/AudioRecord/run.sh

+  SEARCH=$(dirname "$SEARCH")
 done
+[ -z "$INIT_ENV" ] && echo "[ERROR] init_env not found" && exit 1
+# shellcheck source=/dev/null


smuppand · 2025-09-05T04:05:58Z

Runner/suites/Multimedia/Audio/AudioRecord/run.sh

-    # shellcheck disable=SC1090
-    . "$INIT_ENV"
-fi
-# shellcheck disable=SC1090,SC1091


smuppand · 2025-09-05T04:25:04Z

Runner/utils/libaudio.sh

+    done
+
+    get_kernel_log > "$LOGDIR/dmesg_after.log"
+    scan_dmesg_errors "audio" "$LOGDIR"


Your function signature is scan_dmesg_errors(prefix, module_regex, [exclude]).

Current calls use scan_dmesg_errors "audio" "$LOGDIR" (reversed). This silently weakens the scan. It returns 0 when errors are found, 1 when clean. Runners should only gate FAIL on this if a strict flag is set (to avoid flakiness).

smuppand · 2025-09-05T04:26:19Z

Runner/utils/libaudio.sh

+
+    for i in $(seq 1 "$PLAYBACK_LOOPS"); do
+        log_info "Playback loop $i of $PLAYBACK_LOOPS"
+        execute_with_timeout "$PLAYBACK_TIMEOUT" sh -c "$PLAY_CMD" >> "$LOGDIR/playback_stdout.log" 2>&1


Defaults like 15s are fine for timeout(1) but may break run_with_timeout (usually expects integers). Normalize once, inside execute_with_timeout.

smuppand · 2025-09-05T04:27:00Z

Runner/utils/libaudio.sh

+
+    get_kernel_log > "$LOGDIR/dmesg_before.log"
+    PLAY_CMD=$(setup_playback_command)
+    PLAY_SUCCESS=0


Playback uses PLAY_SUCCESS but sets PLAYBACK_SUCCESS in one branch—typo causes false PASS.

smuppand · 2025-09-05T04:27:51Z

Runner/utils/libaudio.sh

+    if [ "$AUDIO_BACKEND" = "pipewire" ]; then
+        select_pipewire_source
+    fi
+    check_dependencies "$([ "$AUDIO_BACKEND" = "pulseaudio" ] && echo "parec" || echo "pw-record")" pgrep grep


select_pipewire_source uses wpctl but dependency check doesn’t require it.

Minor: pgrep grep in check_dependencies is unnecessary; keep pgrep, drop grep.

smuppand · 2025-09-05T04:31:38Z

Runner/utils/libaudio.sh

+        log_fail_exit "$TESTNAME" "Audio playback test failed"
+        echo "$TESTNAME FAIL" > "$RESULT_FILE"
+    fi
+


Transferring PASS/FAIL/SKIP ownership to the top-level runner. These shared libraries will be used across the audio files.

tmoida · 2025-09-12T11:46:40Z

Hi @smuppand,
I've updated the changes as per the review comments. Please let me know if any further modifications are required.

smuppand · 2025-09-12T15:30:47Z

@tmoida Please rebase

tmoida · 2025-09-15T09:53:26Z

Hi @smuppand, The branch has been rebased and updated as requested. Please let me know if any further changes are needed.

Runner/utils/functestlib.sh

Runner/utils/audio_common.sh

tmoida · 2025-09-16T05:35:46Z

Hi @smuppand,
I have updated the check_network_status() function in functestlib.sh and added the missing copyright header to audio_common.sh.

Please let me know if any additional changes are needed.

Thank you.

smuppand

Please add a brief commit message, as this PR introduces a new flow and enhancements.

Runner/utils/audio_common.sh

- Modular runner supports both PipeWire and PulseAudio backends - Validation improved with evidence-based checks for playback and recording - Automatic handling of audio assets (download and extraction) - Enhanced diagnostics: dmesg scan, mixer dumps, JUnit XML output for CI - Added CLI options and environment variable support for configuration - Utility scripts refactored for better extensibility and maintainability Signed-off-by: Teja Swaroop Moida <[email protected]>

tmoida · 2025-09-16T09:40:24Z

Added a brief commit message summarizing the new flow and enhancements.

smuppand

LGTM

smuppand requested changes Aug 8, 2025

View reviewed changes

tmoida force-pushed the patch-1 branch from 64b13ca to 2e80ef4 Compare August 11, 2025 06:55

smuppand reviewed Aug 12, 2025

View reviewed changes

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh Outdated Show resolved Hide resolved

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh Outdated Show resolved Hide resolved

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh Outdated Show resolved Hide resolved

tmoida force-pushed the patch-1 branch from 2bae3ec to 1998094 Compare August 14, 2025 11:50

smuppand requested changes Aug 20, 2025

View reviewed changes

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh Outdated Show resolved Hide resolved

tmoida force-pushed the patch-1 branch from 1998094 to 1389026 Compare August 20, 2025 07:24

smuppand requested changes Aug 20, 2025

View reviewed changes

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh Outdated Show resolved Hide resolved

Runner/suites/Multimedia/Audio/AudioPlayback/run.sh Outdated Show resolved Hide resolved

tmoida force-pushed the patch-1 branch from c917a58 to 2b19003 Compare August 22, 2025 06:23

smuppand requested changes Aug 28, 2025

View reviewed changes

tmoida force-pushed the patch-1 branch 2 times, most recently from 3996525 to 4a24ae8 Compare September 4, 2025 10:17

smuppand requested changes Sep 5, 2025

View reviewed changes

tmoida force-pushed the patch-1 branch 4 times, most recently from d1cc8fe to da44e9c Compare September 12, 2025 11:38

tmoida force-pushed the patch-1 branch 4 times, most recently from eff1a92 to a8b34f3 Compare September 12, 2025 13:41

tmoida force-pushed the patch-1 branch 3 times, most recently from d39f2d6 to a6ce819 Compare September 15, 2025 07:52

tmoida closed this Sep 15, 2025

tmoida force-pushed the patch-1 branch from a6ce819 to e84268e Compare September 15, 2025 09:06

tmoida reopened this Sep 15, 2025

tmoida force-pushed the patch-1 branch from 387b0e6 to dd399f8 Compare September 15, 2025 09:44

smuppand self-requested a review September 15, 2025 10:39

smuppand requested changes Sep 15, 2025

View reviewed changes

Runner/utils/functestlib.sh Outdated Show resolved Hide resolved

Runner/utils/audio_common.sh Show resolved Hide resolved

tmoida force-pushed the patch-1 branch from 43c7d01 to 3a2d68f Compare September 16, 2025 05:30

smuppand self-requested a review September 16, 2025 06:38

smuppand requested changes Sep 16, 2025

View reviewed changes

Runner/utils/audio_common.sh Outdated Show resolved Hide resolved

Runner/utils/audio_common.sh Outdated Show resolved Hide resolved

tmoida force-pushed the patch-1 branch from 3a2d68f to b251b98 Compare September 16, 2025 09:29

smuppand approved these changes Sep 16, 2025

View reviewed changes

smuppand merged commit 034b054 into qualcomm-linux:main Sep 16, 2025
8 checks passed

Add PipeWire support to AudioRecord and AudioPlayback tests #137

Add PipeWire support to AudioRecord and AudioPlayback tests #137

Uh oh!

Conversation

tmoida commented Aug 8, 2025

Uh oh!

smuppand left a comment

Choose a reason for hiding this comment

Uh oh!

tmoida commented Aug 11, 2025

Uh oh!

smuppand left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tmoida commented Aug 14, 2025

Uh oh!

smuppand commented Aug 20, 2025

Uh oh!

Uh oh!

tmoida commented Aug 20, 2025

Uh oh!

smuppand commented Aug 20, 2025

Uh oh!

Uh oh!

Uh oh!

tmoida commented Aug 22, 2025

Uh oh!

smuppand commented Aug 26, 2025

Uh oh!

tmoida commented Aug 28, 2025

Uh oh!

smuppand left a comment

Choose a reason for hiding this comment

Uh oh!

tmoida commented Sep 4, 2025

Uh oh!

smuppand left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tmoida commented Sep 12, 2025

Uh oh!

smuppand commented Sep 12, 2025

Uh oh!

tmoida commented Sep 15, 2025

Uh oh!

Uh oh!

Uh oh!

tmoida commented Sep 16, 2025

Uh oh!

smuppand left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

tmoida commented Sep 16, 2025

Uh oh!

smuppand left a comment