test0r
052a692668
build and publish packages / distcc-avahi-aarch64 (push) Failing after 19s
build and publish packages / lmcp-debian (push) Has been skipped
build and publish packages / lmcp-any (push) Has been skipped
build and publish packages / claude-his-any (push) Has been skipped
build and publish packages / ffmpeg-v4l2-request-aarch64 (push) Has been skipped
build and publish packages / claude-his-debian (push) Has been skipped
Adds a brute-force fallback to v4l2request_open_decoder() that enumerates /dev/media[0..15] and /dev/video[0..63] directly when udev_enumerate_scan_devices() returns an error. The fallback uses absolute paths only (no fd-relative openat), which is what makes it work inside firefox's RDD seccomp+broker sandbox where Mozilla's OpenAtTrap rejects fd-relative paths used by systemd's chase() symlink resolver. Same approach Chromium uses in media/gpu/v4l2/stateless/ on ChromeOS, where the sandbox similarly forbids libudev's chase pattern. No regression: the libudev path runs first and the brute-force path only activates on its failure. AV_LOG_INFO line announces the fallback so it's visible in MOZ_LOG=FFmpegLib:5. Validated on RK3399 / Pinebook Pro / rkvdec: libudev probe failed (-2), falling back to brute-force /dev/media* Using V4L2 media driver rkvdec (brute-force) for S264 Reinit context to 1920x1088, pix_fmt: drm_prime RDD CPU = 4.9% Bumps pkgrel=2. Worth submitting to Kwiboo's fork upstream. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>