marfrit/marfrit-packages
0
0
Fork 1
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

daedalus-v4l2{,-dkms}: 79256dc/6ffe92b -> 5d8b436 — revert parking design #71

Merged
marfrit merged 1 commits from claude-noether/marfrit-packages:noether/daedalus-revert-bump-5d8b436 into main 2026-05-21 14:54:19 +00:00
Conversation 0 Commits 1 Files Changed 6
+72 -32
marfrit commented 2026-05-21 13:42:42 +00:00
Owner
Copy Link
Copy Source

Lock-step downgrade of both packages to the revert tip of daedalus-v4l2 (reauktion/daedalus-v4l2 PR #10, which reverted PRs #7 + #8).

Why

After daedalus-v4l2 0.1.0+r28+g79256dc-1 + daedalus-v4l2-dkms 0.1.0+r30+g6ffe92b-1 landed in production, mpv (--hwdec=vaapi-copy) failed pre-playing on higgs with:

v4l2-request: Unable to dequeue buffer: Resource temporarily unavailable
[ffmpeg/video] h264: Failed to end picture decode issue: 1 (operation failed).
Attempting next decoding method after failure of h264-vaapi-copy.

Daemon journal confirmed decode was working — but PR #7's design parked CAPTURE buffers for cookies whose pixels libavcodec hadn't yet released for display-order output. That violated the V4L2 stateless API's 1:1 OUTPUT-to-CAPTURE contract. Firefox tolerated the resulting stale-pixel mess (the visible "2 1 4 3" pair-swap); strict V4L2 stateless clients (mpv, ffmpeg-vaapi) bailed.

Full diagnosis: reauktion/daedalus-v4l2#9.

What changes

  • DAEDALUS_PROTO_VERSION drops 1 → 0. Install both .debs in the same apt transaction.
  • Userspace daemon ABI returns to f0d4186-equivalent behaviour.
  • Kernel module returns to buf_done_and_job_finish completion (pre-#7).
  • PR #4 (H.264 DECODE_MODE / START_CODE menu controls — cosmetic warning fix) stays.
  • daedalus-v4l2-dkms carries forward the #64 multi-kernel postinst fix.

Visible regression

H.264 B-frame streams in Firefox return to the "2 1 4 3 6 5" pair-swap visual. This is the explicit trade-off ("unblock strict V4L2 stateless clients at the cost of restoring the visible-but-not-breaking Firefox artifact") that closed the design loop with the maintainer.

Proper fix tracked at reauktion/daedalus-v4l2#11: concurrent in-flight requests in the daemon + display-order reorder moved into libva-v4l2-request-fourier where the V4L2 stateless API expects it. Multi-day effort.

Refs

Verified

  • git diff f0d4186 5d8b436 -- kernel/ daemon/ include/ shows only the PR #4 menu-controls delta (the change we keep). All PR #7 / #8 deltas are gone.
  • Cross-built on higgs against 6.18.29+rpt-rpi-2712: clean, no errors.
Lock-step downgrade of both packages to the revert tip of daedalus-v4l2 (reauktion/daedalus-v4l2 PR #10, which reverted PRs #7 + #8). ## Why After `daedalus-v4l2 0.1.0+r28+g79256dc-1` + `daedalus-v4l2-dkms 0.1.0+r30+g6ffe92b-1` landed in production, mpv (`--hwdec=vaapi-copy`) failed pre-playing on higgs with: ``` v4l2-request: Unable to dequeue buffer: Resource temporarily unavailable [ffmpeg/video] h264: Failed to end picture decode issue: 1 (operation failed). Attempting next decoding method after failure of h264-vaapi-copy. ``` Daemon journal confirmed decode was working — but PR #7's design parked CAPTURE buffers for cookies whose pixels libavcodec hadn't yet released for display-order output. That violated the V4L2 stateless API's 1:1 OUTPUT-to-CAPTURE contract. Firefox tolerated the resulting stale-pixel mess (the visible "2 1 4 3" pair-swap); strict V4L2 stateless clients (mpv, ffmpeg-vaapi) bailed. Full diagnosis: reauktion/daedalus-v4l2#9. ## What changes - `DAEDALUS_PROTO_VERSION` drops `1 → 0`. Install both .debs in the same apt transaction. - Userspace daemon ABI returns to f0d4186-equivalent behaviour. - Kernel module returns to `buf_done_and_job_finish` completion (pre-#7). - PR #4 (H.264 DECODE_MODE / START_CODE menu controls — cosmetic warning fix) stays. - `daedalus-v4l2-dkms` carries forward the #64 multi-kernel postinst fix. ## Visible regression H.264 B-frame streams in Firefox return to the "2 1 4 3 6 5" pair-swap visual. This is the explicit trade-off ("unblock strict V4L2 stateless clients at the cost of restoring the visible-but-not-breaking Firefox artifact") that closed the design loop with the maintainer. Proper fix tracked at `reauktion/daedalus-v4l2#11`: concurrent in-flight requests in the daemon + display-order reorder moved into `libva-v4l2-request-fourier` where the V4L2 stateless API expects it. Multi-day effort. ## Refs - reauktion/daedalus-v4l2 issue #9 (diagnostic): https://git.reauktion.de/reauktion/daedalus-v4l2/issues/9 - reauktion/daedalus-v4l2 PR #10 (revert, merged): https://git.reauktion.de/reauktion/daedalus-v4l2/pulls/10 - reauktion/daedalus-v4l2 issue #11 (proper-fix design): https://git.reauktion.de/reauktion/daedalus-v4l2/issues/11 - Predecessor bumps superseded: marfrit-packages #67 (PR #7 deploy) + #69 (PR #8 deploy) ## Verified - `git diff f0d4186 5d8b436 -- kernel/ daemon/ include/` shows only the PR #4 menu-controls delta (the change we keep). All PR #7 / #8 deltas are gone. - Cross-built on higgs against `6.18.29+rpt-rpi-2712`: clean, no errors.
marfrit added 1 commit 2026-05-21 13:42:43 +00:00
daedalus-v4l2{,-dkms}: 79256dc/6ffe92b -> 5d8b436 — revert parking design 34972ae9c1 
Lock-step downgrade of both packages to the revert tip of
daedalus-v4l2 (PR #10 closed PRs #7 + #8).  After
0.1.0+r28+g79256dc-1 / 0.1.0+r30+g6ffe92b-1 landed in production,
mpv (--hwdec=vaapi-copy) failed pre-playing with "Unable to dequeue
buffer: Resource temporarily unavailable" because the daemon
parked CAPTURE buffers waiting for libavcodec's display-order
reorder, violating libva's V4L2 stateless 1:1 contract.  See
daedalus-v4l2#9 for the diagnostic, #10 for the revert PR.

DAEDALUS_PROTO_VERSION drops 1 → 0; install both .debs in the same
apt transaction.  Userspace ABI returns to the f0d4186-equivalent
behaviour, plus PR #4 (cosmetic H.264 menu controls).  The
daedalus-v4l2-dkms #64 multi-kernel postinst behaviour stays in
build-deb.sh.

Visible regression: H.264 B-frame streams in Firefox return to the
"2 1 4 3 6 5" pair-swap visual.  Proper fix (concurrent in-flight
requests in daemon + display-order reorder moved into libva-v4l2-
request-fourier) tracked at daedalus-v4l2#11.

Refs:
  * reauktion/daedalus-v4l2#9
  * reauktion/daedalus-v4l2#10  (merged)
  * reauktion/daedalus-v4l2#11
marfrit merged commit 736b6da176 into main 2026-05-21 14:54:19 +00:00
marfrit deleted branch noether/daedalus-revert-bump-5d8b436 2026-05-21 14:54:19 +00:00
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
marfrit (Markus Fritsche)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: marfrit/marfrit-packages#71
Delete Branch "claude-noether/marfrit-packages:noether/daedalus-revert-bump-5d8b436"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?