[17:44:08] step 1: 2G image, mkfs -n 4096, loop + dm-delay (read=15ms write=5ms)
[17:44:11] step 1: populate S1 with 100000 files, snapshot -r to S2 (S2 now shares every S1 block)
[17:44:33] S1 id=256 root level=3 (want >=2; this is the shape from IDEA.md step 1)
[17:44:33] step 2: delete S1 (S2 keeps every one of S1's root children referenced)
[17:44:33] step 3: sync-poll for a non-zero drop key at root level 3, then unmount
[17:44:40] poll 2: drop key (44438 ...) level 3 -- child dropped, unmounting
[17:44:40] step 4: resume/interrupt cycles, looking for drop key to go non-zero -> (0 0)
[17:44:41] attempt 1: drop key (44438 ...) level 3
[17:44:42] attempt 2: drop key (44438 ...) level 3
[17:44:43] attempt 3: drop key (44438 ...) level 3
[17:44:44] attempt 4: drop key (44438 ...) level 3
[17:44:45] attempt 5: drop key (44438 ...) level 3
[17:44:46] attempt 6: drop key (44438 ...) level 3
[17:44:47] attempt 7: drop key (44438 ...) level 3
[17:44:48] attempt 8: drop key (44438 ...) level 3
[17:44:49] attempt 9: drop key (44438 ...) level 3
[17:44:50] attempt 10: drop key (44438 ...) level 3
[17:44:51] attempt 11: drop key (44438 ...) level 3
[17:44:52] attempt 12: drop key (44438 ...) level 3
[17:44:53] attempt 13: drop key (44438 ...) level 3
[17:44:54] attempt 14: drop key (44438 ...) level 3
[17:44:56] attempt 15: drop key (44438 ...) level 3
[17:44:57] attempt 16: drop key (44438 ...) level 3
[17:44:58] attempt 17: drop key (44438 ...) level 3
[17:44:59] attempt 18: drop key (44438 ...) level 3
[17:45:00] attempt 19: drop key (44438 ...) level 3
[17:45:01] attempt 20: drop key (44438 ...) level 3
[17:45:01] step 4: drop key never went to zero in 20 cycles (expected on a patched module)
[17:45:01] step 5: uninterrupted mount, wait up to 300s for completion or abort
[17:47:44] S1's root item is gone after 62s -- drop completed, no abort
[17:47:44] RESULT: no abort in dmesg (full log: ./work-repro-simple-patched-1/dmesg-step5.log)
[17:48:12] btrfs check --readonly rc=0 (full log: ./work-repro-simple-patched-1/check.log)
[17:48:12] RESULT: check clean
[17:48:37] S2 content: identical to S1 before the snapshot (100101 entries; ./work-repro-simple-patched-1/s2.manifest)
[17:48:37] summary: zeroed=0 aborted=0 check_bad=0 drop_done=1 s2_ok=1
[17:48:37] VERDICT: no zeroing, drop completed, no abort, check clean, S2 intact (expected on a patched module)