Root cause: ALSA assertion failure `snd_pcm_writei: Assertion 'pcm' failed`
when pcm_playback_handle became NULL during concurrent write operations.
The crash occurred because:
1. Thread A checks pcm_playback_handle != NULL (passes)
2. Thread B calls jetkvm_audio_playback_close(), sets handle = NULL
3. Thread A calls snd_pcm_writei(NULL, ...) → SIGABRT
Solution: Added pthread mutexes to protect concurrent access:
- playback_mutex protects pcm_playback_handle in decode_write and close
- capture_mutex protects pcm_capture_handle in read_encode and close
All critical sections now acquire mutex before accessing ALSA handles,
preventing the NULL pointer from being passed to ALSA functions.
The sample rate cannot be configured by users - it's determined by the audio
source (HDMI device or USB gadget client). The previous UI gave the false
impression that users could select a sample rate, but the value was always
overridden by hardware detection.
Changes:
- Convert sample rate UI from dropdown to read-only display
- Show "(auto-detected from source)" label next to the value
- Remove sampleRate parameter from setAudioConfig RPC
- Update translations to clarify auto-detection
- Backend sample rate validation remains for backwards compatibility
The C code now automatically detects and adapts to whatever rate the hardware
supports, creating the Opus encoder/decoder with matching parameters to
eliminate pitch/speed distortion.
Root cause: ALSA was silently using a different sample rate than configured,
causing severe pitch/speed distortion (the "cassette player" warping effect).
The bug occurred when:
- User configured 48 kHz in UI
- HDMI source output 44.1 kHz audio
- set_rate() failed, set_rate_near() chose 44.1 kHz
- Code never checked what rate was actually set
- Opus encoder created for 48 kHz but received 44.1 kHz audio
- Result: ~9% pitch shift and timing mismatch
Fix:
- Always use set_rate_near() and check the actual rate returned
- Pass detected rate and frame size to Opus encoder/decoder creation
- Avoid modifying global state to prevent capture/playback interference
- Recalculate frame_size for 20ms at the actual rate
- Verify rate after hw_params application
- Add detailed logging for rate adaptation
This ensures Opus encoder/decoder use the correct rate matching the hardware,
regardless of what the HDMI source outputs.
- Update default EDID with registered manufacturer ID (Dell) and proper 24-inch display dimensions (52x32cm) for better macOS/OS compatibility
- Add configurable sample rate (32/44.1/48/96 kHz) to support different HDMI audio sources
- Add packet loss compensation percentage control for FEC overhead tuning
- Fix config migration to ensure new audio parameters get defaults for existing configs
- Update all language translations for new audio settings
When USB state transitions to "configured" (after reconnection or state
changes), automatically close and reopen HID file handles to ensure they
remain valid.
- Add Devices.Equals() method to compare USB device configurations
- Add GetGadgetDevices() to retrieve current device state
- Skip gadget reconfiguration when device state is unchanged
- Remove 6 unused audio translation keys from all language files
- Add config fields for bitrate, complexity, DTX, FEC, buffer periods
- Add RPC methods for get/set audio config and restart
- Add UI settings page with controls for all audio parameters
- Add Apply Settings button to restart audio with new config
- Add config migration for backwards compatibility
- Add translations for all 9 languages
- Clean up redundant comments and optimize log levels
Previous fix protected WriteMessage() but missed Connect() calls. During
session transfers, multiple goroutines could simultaneously initialize the
audio playback system, causing SIGABRT crashes in jetkvm_audio_playback_init.
Extended inputSourceMutex to serialize both Connect() and WriteMessage()
operations, ensuring atomic state transitions and preventing concurrent
C library initialization.
During session transfers, multiple handleInputTrackForSession goroutines
could call WriteMessage() concurrently on the shared inputSource, causing
Opus SILK decoder state corruption and assertion failures.
Added inputSourceMutex to serialize WriteMessage() calls and prevent
race conditions in both input and output audio paths.
When an old connection closed while a new one started, the audio cleanup
would hold audioMutex for up to 37 seconds during CGO disconnect calls,
blocking the new session from initializing.
Use capture-clear-release pattern to minimize mutex hold time:
- Capture references to sources/relays while holding mutex
- Clear globals immediately so new sessions can proceed
- Release mutex before calling blocking Stop/Disconnect operations
This eliminates the 37-second hang during connection transitions.
- Use mono audio instead of stereo for microphone input
- Add cleanup to stop audio track on component unmount
- Explicitly set AudioTrack to nil when creation fails
- Replace mutex-protected inputSource with atomic.Pointer for lock-free reads
- Eliminate ~100 mutex operations per second in audio packet hot path
- Add defer pattern for safer mutex unlock in enable/disable functions
- Add error logging for audio write failures
- Consolidate Makefile lint commands for brevity
- Prevent concurrent getUserMedia requests to fix toggle flakiness
- Add automatic track recovery when microphone track ends unexpectedly
- Fix auto-enable to only trigger on session start, not when toggling setting
- Ensure backend RPC is sent when auto-enable triggers
- Properly clean up old tracks before requesting new ones
* feat: add web route for sending WOL package to given mac addr
```
adds a new route /device/send-wol/:mac-addr to send the magic WOL package
to the specified mac-addr.
Method is POST and is protected.
Useful for custom wake up scripts: example is sending HTTP request through iOS shortcut
Test plan:
calling the API with curl
```
$ curl -X POST http://<jetkvm-ip>/device/send-wol/xx:xx:xx:xx:xx:xx
WOL sent to xx:xx:xx:xx:xx:xx
```
and observing the magic packet on my laptop/PC:
```
$ ncat -u -l 9 -k | xxd
00000000: ffff ffff ffff d050 9978 a620 d050 9978 .......P.x. .P.x
00000010: a620 d050 9978 a620 d050 9978 a620 d050 . .P.x. .P.x. .P
00000020: 9978 a620 d050 9978 a620 d050 9978 a620 .x. .P.x. .P.x.
00000030: d050 9978 a620 d050 9978 a620 d050 9978 .P.x. .P.x. .P.x
00000040: a620 d050 9978 a620 d050 9978 a620 d050 . .P.x. .P.x. .P
00000050: 9978 a620 d050 9978 a620 d050 9978 a620 .x. .P.x. .P.x.
```
calling the api with invalid mac addr returns HTTP 400 error
```
$ curl -X POST -v http://<jetkvm-ip>/device/send-wol/abcd
...
* Request completely sent off
< HTTP/1.1 400 Bad Request
...
...
Invalid mac address provided
* Resolve golint complaint
---------
Co-authored-by: Marc Brooks <IDisposable@gmail.com>
When browsers recreate audio tracks during long sessions, the backend
now properly handles the new track instead of ignoring it. Old track
handlers detect when they've been superseded and exit cleanly to
prevent goroutine leaks.
Alex P
Alex
Nitish Agarwal
Adam Shiervani
Aveline
Marc Brooks
tadic-luka