protocols: Add fifo-v1 and commit-timing-v1 #11703

vaxerski · 2025-09-14T17:53:42Z

Fixes #11386

Needs some testing, I didn't test this because I dont know what apps use it.

If anyone has a test app, do lmk.

@Justsnoopy30 @fxzzi

vaxerski · 2025-09-14T17:56:11Z

Also @ikalco @gulafaran I touched your surface commit things. Might wanna verify I didn't exactly fuk something up

vaxerski · 2025-09-14T18:10:35Z

@fufexan wp bump needed tooooo.... idk which version, fifo and ct

fxzzi · 2025-09-14T18:12:19Z

Games using SDL3 (SDL2 as well, not sure?) should be taking advantage of these two protos. As indicated, SDL should automatically pick Wayland as its default backend now that these are available to it. CS2 will still default to XWL for now as they've manually added an env var to their launch script to use x11 SDL backend.

src/protocols/Fifo.hpp

vaxerski · 2025-09-14T18:49:43Z

I tried PPSSPP with wayland backend and it didnt bind to either fifo or ct

fufexan · 2025-09-14T21:35:57Z

@fufexan wp bump needed tooooo.... idk which version, fifo and ct

not needed, they've been available since 1.38. We have 1.45.

vaxerski · 2025-09-14T22:28:21Z

why is ci failing then huh

oh I forgor mesonnn

src/protocols/Fifo.cpp

src/protocols/CommitTiming.cpp

src/protocols/Fifo.cpp

vaxerski · 2025-09-15T19:09:16Z

@ikalco check now

fxzzi · 2025-09-15T19:23:09Z

hyprlandCrashReport105723.txt

crash to tty when trying to launch CS2 with SDL_VIDEO_DRIVER=wayland, nvidia
wleave (gtk4, GSK_RENDERER default vulkan) also crashing to tty

vaxerski · 2025-09-15T19:32:28Z

oops fixer

fxzzi · 2025-09-15T19:40:27Z

PXL_20250915_193818758.TS-scaled.mp4

cs2 now launches, though lots of flickering is present as shown by above video (sorry for shitty camera angle)

explicit sync does seem to be in use as indicated by WAYLAND_DEBUG logs

[ 915462.444] {Default Queue} wl_registry#2.global(56, "wp_fifo_manager_v1", 1)
[ 915462.445] {Default Queue} wl_registry#2.global(57, "wp_commit_timing_manager_v1", 1)

also present in logs, does this mean it's attempting to use them?

Zebra2711 · 2025-09-16T15:36:12Z

If anyone has a test app, do lmk.

mpv has implemented fifo-v1 (mpv-player/mpv@39b53dd)

vaxerski · 2025-09-16T19:09:29Z

also present in logs, does this mean it's attempting to use them?

No, it has to be bind.

vaxerski · 2025-09-16T19:09:59Z

check flicker now

vaxerski · 2025-09-16T19:13:35Z

@Zebra2711 I can't get mpv to bind to fifo. mpv ./file.mp4 --gpu-api=vulkan

fxzzi · 2025-09-16T19:40:19Z

check flicker now

naw, still cooked. The flicker looks exactly like the flicker from pre-explicit-sync days. Discord flickering, steam flickering, cs2 flickering, etc etc, in the same ways they used to.

Zebra2711 · 2025-09-16T19:49:52Z

@Zebra2711 I can't get mpv to bind to fifo. mpv ./file.mp4 --gpu-api=vulkan

mpv can use fifo, ~~but it seems like something is wrong here line 130-L147~~. Log failed when run with mpv log.txt

vaxerski · 2025-09-16T19:54:42Z

how are you launching it to get it to use fifo?

Zebra2711 · 2025-09-16T20:08:55Z

how are you launching it to get it to use fifo?

i run with mpv --gpu-api=vulkan file.mp4 but mpv cannot open with this patch and fifo appeared in the log

...
[2975156.711] wl_callback#79.done(2344)
[2975156.715]  -> zwp_linux_dmabuf_v1#86.get_surface_feedback(new id zwp_linux_dmabuf_feedback_v1#79, wl_surface#7)
[2975156.717]  -> wl_display#1.sync(new id wl_callback#87)
[2975156.783] {Display Queue} wl_display#1.delete_id(76)
[2975156.787] {Display Queue} wl_display#1.delete_id(81)
[2975156.789] {Display Queue} wl_display#1.delete_id(83)
[2975156.791] {Display Queue} wl_display#1.delete_id(87)
[2975156.793] discarded wl_drm#85.device("/dev/dri/renderD128")
[2975156.796] discarded wl_drm#85.capabilities(1)
[2975156.798] discarded wl_drm#85.format(1211384385)
[2975156.800] discarded wl_drm#85.format(1211384408)
[2975156.802] discarded wl_drm#85.format(942948929)
[2975156.804] discarded wl_drm#85.format(942948952)
[2975156.806] discarded wl_drm#85.format(808669761)
[2975156.808] discarded wl_drm#85.format(808669784)
[2975156.810] discarded wl_drm#85.format(808665665)
[2975156.812] discarded wl_drm#85.format(875713089)
[2975156.814] discarded wl_drm#85.format(875708993)
[2975156.816] discarded wl_drm#85.format(875713112)
[2975156.818] discarded wl_drm#85.format(875709016)
[2975156.820] discarded wl_drm#85.format(892424769)
...
[2975185.181] {Default Queue}  -> xdg_toplevel#34.set_title("No file - mpv")
[2975185.201] {Default Queue}  -> wp_content_type_v1#37.set_content_type(0)
[2975186.448]  -> frog_color_managed_surface#52.set_known_container_color_volume(0)
[2975186.458]  -> frog_color_managed_surface#52.set_known_transfer_function(0)
[2975186.461]  -> frog_color_managed_surface#52.set_hdr_metadata(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)
[2975186.474]  -> wp_presentation#80.feedback(wl_surface#7, new id wp_presentation_feedback#92)
[2975186.869]  -> wl_surface#7.attach(wl_buffer#83, 0, 0)
[2975186.876]  -> wl_surface#7.damage_buffer(0, 0, 2147483647, 2147483647)
[2975186.879]  -> wp_fifo_v1#82.set_barrier()
[2975186.881]  -> wp_fifo_v1#82.wait_barrier()
[2975186.883]  -> wl_surface#7.commit()
[2975186.885]  -> wp_fifo_v1#82.wait_barrier()
[2975186.887]  -> wl_surface#7.commit()
[2975187.022] {Default Queue}  -> wl_pointer#10.set_cursor(0, nil, 0, 0)
[2975187.414]  -> wp_presentation#80.feedback(wl_surface#7, new id wp_presentation_feedback#93)
[2975187.807]  -> wl_surface#7.attach(wl_buffer#76, 0, 0)
[2975187.813]  -> wl_surface#7.damage_buffer(0, 0, 2147483647, 2147483647)
[2975187.816]  -> wp_fifo_v1#82.set_barrier()
[2975187.818]  -> wp_fifo_v1#82.wait_barrier()
[2975187.820]  -> wl_surface#7.commit()
[2975187.822]  -> wp_fifo_v1#82.wait_barrier()
[2975187.824]  -> wl_surface#7.commit()
[2975200.712]  -> wp_presentation#80.feedback(wl_surface#7, new id wp_presentation_feedback#94)
[2975201.205]  -> wl_surface#7.attach(wl_buffer#89, 0, 0)
[2975201.213]  -> wl_surface#7.damage_buffer(0, 0, 2147483647, 2147483647)
[2975201.217]  -> wp_fifo_v1#82.set_barrier()
[2975201.228]  -> wp_fifo_v1#82.wait_barrier()
[2975201.231]  -> wl_surface#7.commit()
[2975201.234]  -> wp_fifo_v1#82.wait_barrier()
[2975201.238]  -> wl_surface#7.commit()
[2975209.328]  -> wp_presentation#80.feedback(wl_surface#7, new id wp_presentation_feedback#95)
[2975209.820]  -> wl_surface#7.attach(wl_buffer#91, 0, 0)
[2975209.828]  -> wl_surface#7.damage_buffer(0, 0, 2147483647, 2147483647)
[2975209.831]  -> wp_fifo_v1#82.set_barrier()
[2975209.833]  -> wp_fifo_v1#82.wait_barrier()
[2975209.835]  -> wl_surface#7.commit()
[2975209.837]  -> wp_fifo_v1#82.wait_barrier()
[2975209.839]  -> wl_surface#7.commit()

vaxerski · 2025-09-16T20:14:18Z

it doesnt bind for me :(

src/protocols/CommitTiming.cpp

ikalco · 2025-09-16T20:33:45Z

src/protocols/core/Compositor.cpp

+void CWLSurfaceResource::lockState() {
+    m_stateLocks++;
+}
+
+void CWLSurfaceResource::unlockState(std::optional<size_t> seq) {
+    RASSERT(!!m_stateLocks, "Tried to unlock an unlocked wl_surface state");
+    m_stateLocks--;
+
+    if (m_stateLocks)
+        return;
+
+    while (!m_pendingStates.empty() && (!seq || m_pendingStates.front()->seq <= *seq)) {
+        commitState(*m_pendingStates.front());
+        m_pendingStates.pop();
+    }
+}
+


This should be in SSurfaceState

currently we can skip a frame/buffer:

we get commit1 with buffer1, locks = 1

we get commit2 with buffer2, locks = 2

buffer1 is readable, locks = 1 .... this should've applied commit1

buffer2 is readable, locks = 0 .... this applies commit1 followed by commit2

commit1 buffer was never applied and scanned out, so a frame is skipped

I wrote this before you added the seq mechanism, but I still think this applies cause buffer unlock isn't the only thing locking/unlocking every queued surface state

right... I'll rethink this tomorrow.

@ikalco what about having a queue with fences?

Say we have a queue with states:
A B C D

commit comes, we insert a fence on the buffer F1

A B C D F1

more commits come:

A B C D F1 E

fifo and a new buffer is used, F2 for dma and F3 for fifo:

A B C D F1 E G F2 F3

when F1 signals, all commits before it get applied:

E G F2 F3

Now the only problem is that for E and G, as far as I understand, both F2 and F3 need to signal. This is fine though and easy code-wise.

yeah I think that can work
just make sure that commits without constraints are applied immediately

ikalco · 2025-09-16T20:41:55Z

src/protocols/Fifo.hpp

+    SState m_current, m_pending;
+
+    struct {
+        CHyprSignalListener surfacePrecommit;
+    } m_listeners;


this should be in SSurfaceState, cause this bypasses the m_pendingStates queue

also as a clarifying note, surface commit's aren't really "double buffered"
the first "buffer" is m_pending, and then the second "buffer" is actually that m_pending is added into the m_pendingStates queue until it can be applied depending on constraints (buffer readability, commit timing timeout, fifo_barrier is unset, more stuff?)

ikalco · 2025-09-16T20:44:16Z

src/protocols/Fifo.cpp

+        // Signal all surfaces
+        // TODO: should we wait for the correct monitor(s)?
+
+        fifo->presented();


I think the monitor probably would matter because fifo proto guarantees that a surface buffer will be active for at least one refresh cycle, which depends on each monitors refresh rate

what if its on two?

what if its on two?

?

it should only unlock for the corresponding monitor (PHLMONITOR m), otherwise when refresh cycles are desynced bad stuff happens (mon1 has 60hz, mon2 144hz)

vaxerski · 2025-09-22T16:56:17Z

@ikalco check now, I've done locks per-state, should be better

Zebra2711 · 2025-09-22T20:28:49Z

mpv is able to open now, but it immediately closes when playing a video

WAYLAND_DEBUG=1 mpv --player-operation-mode=pseudo-gui

[2552434.075]  -> wl_surface#7.commit()
[2552437.929] {Display Queue} wl_display#1.delete_id(64)
[2552438.317] {Display Queue} wl_display#1.delete_id(43)
[2552438.326] {Default Queue} wl_callback#43.done(8494166)
[2552438.329] {Default Queue}  -> wl_surface#7.frame(new id wl_callback#43)
[2552438.869] wl_buffer#78.release()
[2552438.871] wl_buffer#62.release()
[2552438.872] wl_buffer#74.release()
[2552440.100] wp_presentation_feedback#64.sync_output(wl_output#29)
[2552440.111] wp_presentation_feedback#64.presented(0, 8492, 166950046, 6944299, 0, 1222610, 7)
[2552440.122]  -> wp_presentation#59.feedback(wl_surface#7, new id wp_presentation_feedback#64)
[2552440.124]  -> wp_commit_timer_v1#65.set_timestamp(0, 8495, 319201537)
[2552440.714]  -> wl_surface#7.attach(wl_buffer#62, 0, 0)
[2552440.718]  -> wl_surface#7.damage_buffer(0, 0, 2147483647, 2147483647)
[2552440.720]  -> wp_fifo_v1#63.set_barrier()
[2552440.736]  -> wp_fifo_v1#63.wait_barrier()
[2552440.738]  -> wl_surface#7.commit()
[2552440.739]  -> wp_fifo_v1#63.wait_barrier()
[2552440.740]  -> wl_surface#7.commit()
[2552444.875] {Display Queue} wl_display#1.delete_id(64)
[2552447.284] wp_presentation_feedback#64.sync_output(wl_output#29)
[2552447.293] wp_presentation_feedback#64.presented(0, 8494, 173895014, 6944299, 0, 1222611, 7)
[2552447.307]  -> wp_presentation#59.feedback(wl_surface#7, new id wp_presentation_feedback#64)
[2552447.310]  -> wp_commit_timer_v1#65.set_timestamp(0, 8495, 326148648)
[2552447.915]  -> wl_surface#7.attach(wl_buffer#74, 0, 0)
[2552447.917]  -> wl_surface#7.damage_buffer(0, 0, 2147483647, 2147483647)
[2552447.919]  -> wp_fifo_v1#63.set_barrier()
[2552447.921]  -> wp_fifo_v1#63.wait_barrier()
[2552447.922]  -> wl_surface#7.commit()
[2552447.924]  -> wp_fifo_v1#63.wait_barrier()
[2552447.926]  -> wl_surface#7.commit()
[2552447.976] {Display Queue} wl_display#1.error(wp_commit_timer_v1#65, 1, "Timestamp is already set")
wp_commit_timer_v1#65: error 1: Timestamp is already set
[2552683.804] {Default Queue}  -> ext_data_control_device_v1#3.destroy()
[2552683.818] {Default Queue}  -> ext_data_control_manager_v1#5.destroy()

full log log.txt

vaxerski · 2025-09-22T20:31:56Z

ew

vaxerski · 2025-09-22T20:36:54Z

now?

Zebra2711 · 2025-09-22T20:48:10Z

now?

now it just freezes :)

log.txt

vaxerski · 2025-09-22T20:57:15Z

stellar

vaxerski · 2025-09-22T20:58:15Z

this should help

Zebra2711 · 2025-09-22T21:08:59Z

it plays a few frames and then freezes again

log.txt

vaxerski · 2025-09-23T20:18:57Z

@ikalco any idea? I still can't test

src/protocols/CommitTiming.cpp

vaxerski · 2025-09-23T22:05:01Z

check now @fxzzi

fxzzi · 2025-09-23T23:55:40Z

check now @fxzzi

rip, still bugged

vaxerski · 2025-09-24T20:17:06Z

well I dunno then, and it's really annoying to dev without being able to test or debug

Kommynct · 2025-09-26T09:12:14Z

Let him remote innnn, dooo ittt

M-UnityDev · 2025-10-02T14:37:02Z

well I dunno then, and it's really annoying to dev without being able to test or debug

I noticed that Godot gives a warning about this protocol when launching the game so maybe you can use a wayland native build of my tech demo to test how your protocol implementation works

(there is also an auto generated sh script that will output debug info to the terminal)

protocols: Add fifo-v1 and commit-timing-v1

13f37be

github-actions bot added core managers protocols helpers labels Sep 14, 2025

vaxerski added 3 commits September 14, 2025 18:57

e

c607bfe

e

8ad89f8

e

b650e64

fxzzi reviewed Sep 14, 2025

View reviewed changes

src/protocols/Fifo.hpp Show resolved Hide resolved

mesooon

cc73f25

ikalco reviewed Sep 15, 2025

View reviewed changes

src/protocols/Fifo.cpp Outdated Show resolved Hide resolved

ikalco reviewed Sep 15, 2025

View reviewed changes

src/protocols/CommitTiming.cpp Show resolved Hide resolved

ikalco reviewed Sep 15, 2025

View reviewed changes

src/protocols/Fifo.cpp Outdated Show resolved Hide resolved

stuff

fa60759

stuff

5080794

stuff

913f820

fix flicker

8657548

ikalco reviewed Sep 16, 2025

View reviewed changes

ehgerhbryugher

19963d3

ugj

d90cfd1

boop

313b868

ikalco reviewed Sep 23, 2025

View reviewed changes

src/protocols/CommitTiming.cpp Outdated Show resolved Hide resolved

fix

226df2b

Uh oh!

protocols: Add fifo-v1 and commit-timing-v1 #11703

Are you sure you want to change the base?

protocols: Add fifo-v1 and commit-timing-v1 #11703

Uh oh!

Conversation

vaxerski commented Sep 14, 2025

Uh oh!

vaxerski commented Sep 14, 2025

Uh oh!

vaxerski commented Sep 14, 2025

Uh oh!

fxzzi commented Sep 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

vaxerski commented Sep 14, 2025

Uh oh!

fufexan commented Sep 14, 2025

Uh oh!

vaxerski commented Sep 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

vaxerski commented Sep 15, 2025

Uh oh!

fxzzi commented Sep 15, 2025

Uh oh!

vaxerski commented Sep 15, 2025

Uh oh!

fxzzi commented Sep 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Zebra2711 commented Sep 16, 2025

Uh oh!

vaxerski commented Sep 16, 2025

Uh oh!

vaxerski commented Sep 16, 2025

Uh oh!

vaxerski commented Sep 16, 2025

Uh oh!

fxzzi commented Sep 16, 2025

Uh oh!

Zebra2711 commented Sep 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

vaxerski commented Sep 16, 2025

Uh oh!

Zebra2711 commented Sep 16, 2025

Uh oh!

vaxerski commented Sep 16, 2025

Uh oh!

Uh oh!

ikalco Sep 16, 2025

Choose a reason for hiding this comment

Uh oh!

vaxerski Sep 16, 2025

Choose a reason for hiding this comment

Uh oh!

vaxerski Sep 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ikalco Sep 19, 2025

Choose a reason for hiding this comment

Uh oh!

ikalco Sep 16, 2025

Choose a reason for hiding this comment

Uh oh!

ikalco Sep 16, 2025

Choose a reason for hiding this comment

Uh oh!

vaxerski Sep 16, 2025

Choose a reason for hiding this comment

Uh oh!

ikalco Sep 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

fxzzi commented Sep 14, 2025 •

edited

Loading

vaxerski commented Sep 14, 2025 •

edited

Loading

fxzzi commented Sep 15, 2025 •

edited

Loading

Zebra2711 commented Sep 16, 2025 •

edited

Loading

vaxerski Sep 18, 2025 •

edited

Loading

ikalco Sep 16, 2025 •

edited

Loading

M-UnityDev commented Oct 2, 2025 •

edited

Loading