marfrit/libva-v4l2-request-fourier
1
0
Fork 1
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

request: route VP9/AV1/H.264 to daedalus_v4l2 on Pi 5 mixed deploy #6

Merged
claude-noether merged 1 commits from claude-noether/libva-v4l2-request-fourier:noether/libva-1-per-codec-dispatch into master 2026-05-20 08:53:05 +00:00
Conversation 0 Commits 1 Files Changed 1
+94 -16
claude-noether commented 2026-05-20 08:47:34 +00:00
Collaborator
Copy Link
Copy Source

LIBVA-1 — per-codec dispatch when rpi-hevc-dec + daedalus_v4l2 both present.

Finishes the Pi 5 mixed-decoder story so HEVC goes to rpi-hevc-dec (existing p override) and VP9 / AV1 / H.264 go to the daedalus daemon (new d kind).

Problem: known_decoder_drivers[] orders rpi-hevc-dec before daedalus_v4l2, so find_codec_device() on a Pi 5 with both modules loaded picked rpi-hevc-dec as primary and alt_driver = NULL. daedalus_v4l2 never opened → VP9/AV1/H.264 frames hit rpi-hevc-dec S_FMT and failed.

Changes:

  • request.c multi-device-probe: when primary = rpi-hevc-dec, alt = daedalus_v4l2 (gated on HAVE_DAEDALUS_V4L2); symmetric in the daedalus_v4l2 primary branch.
  • request.c alt-driver opening block: generalised from the iter38 rkvdec/hantro pair to also dispatch into video_fd_rpi_hevc_dec and video_fd_daedalus slots. Defensive close on unknown alt-driver name.
  • request_switch_device_for_profile: new d kind handler + profile-override block. When daedalus is open, VP9 / AV1 / H.264* route to it. HEVC stays on p. AV1 a kind (RK3588 vpu981) only wins if video_fd_vpu981 >= 0, preserving RK3588 priority.
  • RequestTerminate: close the daedalus_v4l2 fd pair on teardown (was leaking — caught while reviewing the alt-driver expansion).

Build: meson + ninja clean on boltzmann (only pre-existing GStreamer H265 parser warnings). Behaviour on RK3399/3588 boxes unchanged — new branches are gated by HAVE_DAEDALUS_V4L2 and video_fd_daedalus >= 0, both false in those deployments.

Companion changes: daedalus-v4l2 upstream pin 481279c (Phase 8.13 systemd unit landed); marfrit-packages branch noether/daedalus-v4l2-kernel-6.18-compat carries the package-side systemd integration.

Test plan (LIBVA-2, follow-up): On higgs (Pi CM5, kernel 6.18.29, rpi-hevc-dec + daedalus_v4l2 both loaded): ffmpeg -hwaccel vaapi -i h264_small.mp4 lands frames in the daedalus daemon log; ffmpeg -hwaccel vaapi -i hevc_small.mp4 lands frames on rpi-hevc-dec (daemon log stays silent for HEVC). Both produce a valid first frame.

🤖 Generated with Claude Code

LIBVA-1 — per-codec dispatch when rpi-hevc-dec + daedalus_v4l2 both present. Finishes the Pi 5 mixed-decoder story so HEVC goes to rpi-hevc-dec (existing p override) and VP9 / AV1 / H.264 go to the daedalus daemon (new d kind). Problem: known_decoder_drivers[] orders rpi-hevc-dec before daedalus_v4l2, so find_codec_device() on a Pi 5 with both modules loaded picked rpi-hevc-dec as primary and alt_driver = NULL. daedalus_v4l2 never opened → VP9/AV1/H.264 frames hit rpi-hevc-dec S_FMT and failed. Changes: - request.c multi-device-probe: when primary = rpi-hevc-dec, alt = daedalus_v4l2 (gated on HAVE_DAEDALUS_V4L2); symmetric in the daedalus_v4l2 primary branch. - request.c alt-driver opening block: generalised from the iter38 rkvdec/hantro pair to also dispatch into video_fd_rpi_hevc_dec and video_fd_daedalus slots. Defensive close on unknown alt-driver name. - request_switch_device_for_profile: new d kind handler + profile-override block. When daedalus is open, VP9 / AV1 / H.264* route to it. HEVC stays on p. AV1 a kind (RK3588 vpu981) only wins if video_fd_vpu981 >= 0, preserving RK3588 priority. - RequestTerminate: close the daedalus_v4l2 fd pair on teardown (was leaking — caught while reviewing the alt-driver expansion). Build: meson + ninja clean on boltzmann (only pre-existing GStreamer H265 parser warnings). Behaviour on RK3399/3588 boxes unchanged — new branches are gated by HAVE_DAEDALUS_V4L2 *and* video_fd_daedalus >= 0, both false in those deployments. Companion changes: daedalus-v4l2 upstream pin 481279c (Phase 8.13 systemd unit landed); marfrit-packages branch noether/daedalus-v4l2-kernel-6.18-compat carries the package-side systemd integration. Test plan (LIBVA-2, follow-up): On higgs (Pi CM5, kernel 6.18.29, rpi-hevc-dec + daedalus_v4l2 both loaded): ffmpeg -hwaccel vaapi -i h264_small.mp4 lands frames in the daedalus daemon log; ffmpeg -hwaccel vaapi -i hevc_small.mp4 lands frames on rpi-hevc-dec (daemon log stays silent for HEVC). Both produce a valid first frame. 🤖 Generated with Claude Code
claude-noether added 1 commit 2026-05-20 08:47:34 +00:00
request: route VP9/AV1/H.264 to daedalus_v4l2 on Pi 5 mixed deploy 6173a8da8e 
LIBVA-1 — when both rpi-hevc-dec and daedalus_v4l2 are loaded, finish
the per-codec dispatch so HEVC goes to rpi-hevc-dec (existing 'p'
override) and VP9 / AV1 / H.264 go to the daedalus daemon ('d').

Before this change the multi-device-probe accepted only ONE driver
plus a fixed alt slot (rkvdec↔hantro-vpu); on a Pi 5 with both decoders
the find_codec_device() walk preferred rpi-hevc-dec by known_decoder_
drivers[] order and never opened daedalus_v4l2, so VP9/AV1/H.264 frames
hit rpi-hevc-dec's S_FMT and failed.

Changes:

  - request.c multi-device-probe: when primary = rpi-hevc-dec, alt =
    daedalus_v4l2 (when HAVE_DAEDALUS_V4L2 is on); symmetric handling
    in the daedalus_v4l2 primary branch so alt = rpi-hevc-dec.  This
    preserves the iter40 fallback (no daedalus → alt = NULL) when the
    build option is off.

  - request.c alt-driver opening block: generalized from the iter38
    rkvdec/hantro pair to also dispatch into video_fd_rpi_hevc_dec and
    video_fd_daedalus slots.  Defensive close on unknown alt-driver
    name (shouldn't happen — primary_driver branches gate the choices —
    but keeps the slot tally clean if a future driver name is added
    above without wiring up the dispatch here).

  - request_switch_device_for_profile: added 'd' kind handler +
    profile override block.  When daedalus is open, VP9 / AV1 / H.264*
    route to it.  HEVC stays on rpi-hevc-dec via the existing 'p'
    override.  AV1 'a' kind (RK3588 vpu981) wins ONLY if vpu981 was
    probed, so the override only fires on hosts where vpu981 stayed
    -1 (i.e. Pi 5).

  - RequestTerminate: close the daedalus_v4l2 fd pair on teardown
    (was leaking — caught while reviewing the alt-driver expansion).

Build: meson + ninja clean on boltzmann (only pre-existing GStreamer
H265 parser noise).  Behaviour on RK3399/3588 boxes unchanged — the
new branches are gated by HAVE_DAEDALUS_V4L2 *and* video_fd_daedalus
≥ 0, both of which stay false in those deployments.

Companion to daedalus-v4l2 481279c (Phase 8.13 systemd unit) and
marfrit-packages noether/daedalus-v4l2-kernel-6.18-compat branch.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
claude-noether merged commit c332d34643 into master 2026-05-20 08:53:05 +00:00
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
Clear labels
EWONTFIX
Triaged 2026-05-18: scoped out indefinitely; reopen-able if real need emerges. (HEVC-on-ampere first batch.)
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
claude-noether (Claude (noether)) marfrit (Markus Fritsche)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: marfrit/libva-v4l2-request-fourier#6
Delete Branch "claude-noether/libva-v4l2-request-fourier:noether/libva-1-per-codec-dispatch"

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?