Support pipewire as host by Decodetalkers · Pull Request #1093 · RustAudio/cpal

Decodetalkers · 2026-01-06T12:20:27Z

Add support to pipewire

You can test it with pipewire feature open

Pipewire support use config to define rates. So the default config of cpal with pipewire can be changed through config like following.

cat ~/.config/pipewire/pipewire.conf.d/custom-rates.conf
context.properties = {
    default.clock.rate          = 48000
    default.clock.allowed-rates = [ 44100 48000 88200 96000 176400 192000 ]
}

Still problems left:

*~~once we do 'pipewire::init', we can only dequeue it after the whole thread. even put the function to another thread, the function still works.. But seems we can run init many times..~~ (Ok, seems it is not a problem, because in when we call init, the init action will only be called once. I think it will be ok)
*~~The crates by pipewire need edition 2024. I think that should be another pr.~~

Decodetalkers · 2026-01-06T14:59:35Z

Seems pipewire need edition 2024..

Decodetalkers · 2026-01-07T11:32:35Z

@roderickvd can you help review this pr? Thanks, and when can this crate be upgraded to edition 2024? I would also want to help

roderickvd · 2026-01-07T19:55:44Z

@roderickvd can you help review this pr?

Definitely will help you review it. Need a bit more time.

Thanks, and when can this crate be upgraded to edition 2024? I would also want to help

Actually cpal itself doesn't need to be upgraded to Rust 2024, it just needs a MSRV of Rust of 1.85 or higher to support dependencies that are Rust 2024 already.

When we can I'd like to stick cpal to Rust 2021 so we keep our MSRV down.

SuperKenVery · 2026-01-08T18:12:30Z

Super cool man! Happy to see cpal having better linux support, opening the possibility of loopback recording on Linux!

Decodetalkers · 2026-01-09T22:35:42Z

ok, only one ci that I cannot fix

Decodetalkers · 2026-01-10T16:25:00Z

ok cross-rs based on ubuntu20.04, so

Be-ing · 2026-01-15T22:56:15Z

How does this differ from #692?

it is meaningful, but not used in this repo

no use to keep an unnecessary function

@roderickvd

as @roderickvd suggested

PipeWire does support U64 and I64, but libspa doesn't yet. so I add a todo and a note

make the flag pipewire, beep and pulseaudio enabled the same time

and cross --all-features also with pulseaudio

it should not be here

and add some documents

we do not need object_id and node_id in this place and no need for functions channels, node_name and direction

…ments

add documents add CHANGELOG

roderickvd

Really almost there 😆
Thanks for sticking through all this!

src/host/pipewire/device.rs

roderickvd · 2026-02-27T20:49:48Z

src/host/pipewire/stream.rs

+                let current_channels = user_data.format.channels();
+                let current_rate = user_data.format.rate();
+                if current_channels != channels || rate != current_rate {
+                    (user_data.error_callback)(StreamError::BackendSpecific {


We may want to call stream.disconnect() or at least stream.set_active(false) on such an event. Because in the callback the n_channels is queried, and if we leave that running, then the data slice will be different from what the user expects.

ok, I set inactive when the values do not match.

roderickvd · 2026-02-27T20:50:21Z

src/host/pipewire/mod.rs

+mod stream;
+mod utils;
+
+// just init the pipewire the check if it is available


Comment can be removed; function is self-explanatory.

roderickvd · 2026-02-27T20:50:39Z

src/host/pipewire/device.rs

+    channels: ChannelCount,
+    rate: SampleRate,
+    allow_rates: Vec<SampleRate>,
+    quantum: u32,


This should be of type FrameCount too.

roderickvd · 2026-02-27T20:52:24Z

src/host/pipewire/device.rs

+        &self.node_name
+    }
+
+    pub fn quantum(&self) -> FrameCount {


I still think we should remove rate() and allow_rates() because they can be gotten from supported_*_configs already. quantum() is OK for now but I want to replace with the buffer size querying PR soon after.

roderickvd · 2026-02-27T20:55:06Z

src/host/pipewire/stream.rs

+//
+// This ensures positive values that are compatible with our `StreamInstant` representation.
+#[inline]
+fn stream_timestamp_fallback(


It's called "fallback" but do we ever use true PipeWire timestamps elsewhere with https://pipewire.pages.freedesktop.org/pipewire-rs/pipewire_sys/fn.pw_stream_get_time_n.html? We should: it's RT-safe and more accurate.

LLM-assisted sketch:

/// Hardware timestamp from a PipeWire graph cycle. struct PwTime { /// CLOCK_MONOTONIC nanoseconds, stamped at the start of the graph cycle. now_ns: i64, /// Pipeline delay converted to nanoseconds. /// For output: how far ahead of the driver our next sample will be played. /// For input: how long ago the data in the buffer was captured. delay_ns: i64, } /// Returns a hardware timestamp for the current graph cycle, or `None` if /// the driver has not started yet or the rate is unavailable. fn pw_stream_time(stream: &pw::stream::StreamRc) -> Option<(StreamInstant, i64)> { let mut t: pw_sys::pw_time = unsafe { mem::zeroed() }; let rc = unsafe { pw_sys::pw_stream_get_time_n( stream.as_raw_ptr(), &mut t, mem::size_of::<pw_sys::pw_time>(), ) }; if rc != 0 || t.now == 0 || t.rate.denom == 0 { return None; } debug_assert_eq!(t.rate.num, 1, "unexpected pw_time rate.num"); let delay_ns = t.delay * 1_000_000_000i64 / t.rate.denom as i64; let callback = crate::StreamInstant::from_nanos(t.now); Some((callback, delay_ns)) }

This may require features = ["v0_3_50"].

Then the StreamInstant docs should be updated to include PipeWire support.

roderickvd · 2026-02-27T21:41:09Z

One more thing I just thought of: before passing the data buffer to the output callback, that buffer needs to be zeroed. More precisely: it needs to be set to Sample::EQUILIBRIUM (which is not "0" for unsigned formats). You could look at how ALSA does that with its silence_template.

In case that something happens to pipewire, we cannot use expect

Decodetalkers · 2026-02-28T03:56:53Z

One more thing I just thought of: before passing the data buffer to the output callback, that buffer needs to be zeroed. More precisely: it needs to be set to Sample::EQUILIBRIUM (which is not "0" for unsigned formats). You could look at how ALSA does that with its silence_template.

for example

let Some(samples) = buf_data.data() else {
       return;
};
let len = samples.len();
 let silence_template = vec![0_u8; len];
samples.copy_from_slice(&silence_template);

like this way? I do not know how to get the length of sample before the stream start..

yara-blue · 2026-02-28T18:09:04Z

 let silence_template = vec![0_u8; len];

No that would allocate instead try calling https://doc.rust-lang.org/stable/std/primitive.slice.html#method.fill on the slice with Sample::EQUILIBRIUM as value.

Decodetalkers · 2026-03-01T04:22:53Z

 let silence_template = vec![0_u8; len];
No that would allocate instead try calling https://doc.rust-lang.org/stable/std/primitive.slice.html#method.fill on the slice with Sample::EQUILIBRIUM as value.

thank you very much, done

Decodetalkers marked this pull request as draft January 6, 2026 12:20

Decodetalkers force-pushed the pipewire branch from 19070c2 to 1eca9ce Compare January 6, 2026 12:26

Decodetalkers changed the title ~~Pipewire support~~ Support pipewire as host Jan 6, 2026

Decodetalkers marked this pull request as ready for review January 6, 2026 14:41

Decodetalkers force-pushed the pipewire branch 4 times, most recently from 9e2ef9f to fd59a29 Compare January 6, 2026 14:57

Decodetalkers force-pushed the pipewire branch 3 times, most recently from 07bc999 to 6cea2ad Compare January 9, 2026 22:31

Decodetalkers force-pushed the pipewire branch 3 times, most recently from a81e104 to 5a3817c Compare January 10, 2026 16:19

Decodetalkers force-pushed the pipewire branch 9 times, most recently from 721beb0 to 424d78f Compare January 11, 2026 09:53

Decodetalkers added 16 commits February 26, 2026 09:53

chore: split the logic of parse rates to function and add unit tests

3508e4f

chore: remove limit_quantum

efb9dd7

it is meaningful, but not used in this repo

chore: rename init_roundtrip to init_devices

cce2ea7

no use to keep an unnecessary function

chore: just check the pipewire service to confirm if it support pipewire

b81da72

as @roderickvd suggested

feat: support Stream/Output/Audio and Stream/Input/Audio "stream.0"

7a57d62

chore: update the comment in stream.rs, about the format change

9c01815

PipeWire does support U64 and I64, but libspa doesn't yet. so I add a todo and a note

chore: remove blank in examples/beep.rs

17252d3

chore: sync the written way in beep.rs

c1dcc4c

make the flag pipewire, beep and pulseaudio enabled the same time

chore: sync the comment of pipewire in the commands in examples

7075e37

chore: make the ci same logic with pulseaudio

1bc5358

and cross --all-features also with pulseaudio

chore: remove the file added by typos-lsp

ddbed8c

it should not be here

chore: not commit allow(dead_code) to whole struct

70ddfab

and add some documents

chore: rename all rv to rx

6f5d1c0

chore: remove useless code

c1ee289

we do not need object_id and node_id in this place and no need for functions channels, node_name and direction

chore: clippy fix for examples

da111e4

chore:fix typos in examples and ci, format in Cargo.toml and add docu…

0ba60e9

…ments

Decodetalkers force-pushed the pipewire branch from 092ec59 to 0ba60e9 Compare February 26, 2026 00:53

chore: add pipewire information to CHANGELOG and README

de85918

add documents add CHANGELOG

Decodetalkers requested a review from roderickvd February 27, 2026 01:36

roderickvd requested changes Feb 27, 2026

View reviewed changes

Decodetalkers added 5 commits February 28, 2026 12:09

chore: remove all expect

042788a

In case that something happens to pipewire, we cannot use expect

chore: stop stream when input params do not match

28fcfcc

chore: remove useless comment

65a9e00

chore: remove useless function

616ee50

chore: use timestamp of pipewire

5ae8f23

chore: reset samples to "zero" when doing output

b609389

Conversation

Decodetalkers commented Jan 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Still problems left:

Uh oh!

Decodetalkers commented Jan 6, 2026

Uh oh!

Decodetalkers commented Jan 7, 2026

Uh oh!

roderickvd commented Jan 7, 2026

Uh oh!

SuperKenVery commented Jan 8, 2026

Uh oh!

Decodetalkers commented Jan 9, 2026

Uh oh!

Decodetalkers commented Jan 10, 2026

Uh oh!

Be-ing commented Jan 15, 2026

Uh oh!

roderickvd left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

roderickvd Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

Decodetalkers Mar 1, 2026

Choose a reason for hiding this comment

Uh oh!

roderickvd Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

roderickvd Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

roderickvd Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

roderickvd Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

Decodetalkers Mar 1, 2026

Choose a reason for hiding this comment

Uh oh!

roderickvd commented Feb 27, 2026

Uh oh!

Decodetalkers commented Feb 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

yara-blue commented Feb 28, 2026

Uh oh!

Decodetalkers commented Mar 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

Decodetalkers commented Jan 6, 2026 •

edited

Loading

Decodetalkers commented Feb 28, 2026 •

edited

Loading

Decodetalkers commented Mar 1, 2026 •

edited

Loading