fix(alsa): hanging double-opens with BufferSize::Fixed om some drivers

[roderickvd]

The range validation for BufferSize::Fixed was calling default_output_config() before opening the stream, which opened and closed the PCM device once just to get the buffer size range. This may leave the driver in a bad state for the subsequent streaming open.

Second, to defend against hanging processes in such an event, the worker now checks PCM state when poll() times out.

Fixes #1157

[Copilot]

Pull request overview

This PR updates the ALSA backend to avoid an extra open/close cycle when validating BufferSize::Fixed (which can leave some drivers in a bad state), and adds a defensive PCM-state check when poll() times out to prevent hangs when error events aren’t delivered.

Changes:

• Move BufferSize::Fixed range validation to set_hw_params_from_format() so it uses the same PCM handle as the streaming open (avoids a second open/close).
• On poll() timeout, query PCM state and surface Disconnected / XRun / Suspended conditions even if POLLERR/POLLHUP wasn’t delivered.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[DanielHabenicht]

As I wrote in the issue there are no apparent changes for me, but it also does not negatively effect the current behavior on the rasperry pi zero w, so I would say its better than before.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

Comments suppressed due to low confidence (1)

src/host/alsa/mod.rs:595

• default_config treats ErrorKind::UnsupportedConfig from supported_configs(stream_t) as “device does not support {dir}”, but UnsupportedConfig can also mean other configuration/parameter failures. Since Device already tracks direction via self.direction/supports_input()/supports_output(), consider checking that first and returning ErrorKind::UnsupportedOperation without relying on errno/kind mapping from snd_pcm_open(). This would also make behavior consistent for supported_*_configs() callers on the wrong direction.

        let mut formats: Vec<_> = {
            match self.supported_configs(stream_t) {
                // EINVAL when querying direction the device does not support (input-only or output-only)
                Err(err) if err.kind() == ErrorKind::UnsupportedConfig => {
                    let dir = match stream_t {
                        alsa::Direction::Capture => "input",
                        alsa::Direction::Playback => "output",
                    };
                    return Err(Error::with_message(
                        ErrorKind::UnsupportedOperation,
                        format!("Device does not support {dir}"),
                    ));

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.