marfrit/libva-multiplanar
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
41 Commits 1 Branch 0 Tags
7c54d164d9d4054713127db845e7744fef136eba
Commit Graph

4 Commits

Author SHA1 Message Date
marfrit f36c6b040d Iteration 5 Track G complete + Phase 7G verified 
firefox-fourier 150.0.1-1.1 rebuilt without --enable-profile-generate=cross
on boltzmann firefox-fourier container (single-pass, ~2h27m). Pkg
68.7 MB, libxul.so 169 MB stripped — 21× smaller than iter3
PGO-instrumented 3.6 GB binary. Installed on ohm via pacman -U.

Phase 7G: 35s autonomous run (no MOZ_DISABLE_RDD_SANDBOX=1):
  ENETDOWN: 0  (sandbox patch holds)
  EINVAL: 0    (iter4 fix holds)
  RDD ProcessDecode: 538 events
  Stream mTime reached: 22.3s
  Decode rate: 0.64× realtime (~2.7× speedup vs PGO-instrumented)

All four iter5 tracks (A+G+B+E) GREEN.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-05-05 17:37:34 +00:00
marfrit 707d6fd3c0 Iteration 5 Phase 7: A+E+B verified post-cleanup on driver 4bed52ec 
After fork commit c8b6ede (Phase 5 follow-up sweep):
- Track A: 2000-frame mpv stress, 0 EINVAL, 1 v4l2-request log line,
  3 KB log (down from 9 lines / 4.4 KB pre-cleanup).
- Track E: 2-process concurrent mpv with 2s stagger, both clean.
- Track B: 35s mpv --vo=gpu, 31s stream pos, 0 segfaults (mpv falls
  back to SW gracefully after init-time cap_pool EBUSY race; the
  race latent caveat stands per Phase 5 C4).

Track G still rebuilding on boltzmann; ETA <1h.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-05-05 16:38:48 +00:00
marfrit bd37b51d8b Iteration 5 Phase 5: sonnet review YELLOW caveats addressed 
Sonnet caught four iter5 Phase 5 review caveats:
- C1: sweep missed 4 more DEBUG sites (3 in surface.c, 1 in h264.c)
- C2: static bool readback_warned was new process-global state
- C3: msync removal not pixel-verified (probably safe; named caveat)
- C4: Track B "implicit fix" overstated — reframe as "doesn't reproduce"

C1 + C2 resolved in fork commit c8b6ede (107 lines removed). C3 and C4
documented in phase4_iter5_plan.md as named caveats.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-05-05 16:05:13 +00:00
marfrit c827140986 Iteration 5 Phase 4: A + E + B all complete; G running on boltzmann 
Track A (DEBUG sweep): 6 commits, ~232 lines removed, per-frame
v4l2-request log noise from ~30+ lines/frame to 0. 2000-frame stress
clean (0 EINVAL, log size 4.4 KB).

Track E (multi-context safety): LAST_OUTPUT_WIDTH/HEIGHT moved from
process-global static to per-driver_data. Two concurrent mpv (2s
stagger) both decode 300 frames clean.

Track B (mpv libplacebo segfault): RE-TEST on iter5-end driver shows
the iter3-era segfault is GONE. 32s of mpv --vo=gpu decode with 0
segfaults / SIGSEGV. Implicit fix from iter4 fresh-request_fd-per-frame
+ DPB semantics + iter5 per-driver-data move closed the race window.

Track G (PGO-disabled Firefox rebuild): single-pass build kicked on
boltzmann; ETA ~60 min. Phase 7G pending deployment.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-05-05 15:22:15 +00:00