6.18 subsections next to the misc-next rb1103 ones: Josef v2 1/5, 2/5,
5/5 message-only; v2 3/5 and 4/5 replace v1 in place, with the 6.18 port
of 3/5 (spin_lock_irq() on ordered_tree_lock, which 6.18 takes from bio
completion with IRQs off); Filipe's patch replaces ours in compress-fixes
(btrfs/for-next
75d0e4dd64f2a since today). Gate on 6.18 #313 (bhive,
ahive) equal to rb1101; compress-props PASS, negative control on #311.
3/5's IOERR branch is BLIND on 6.18 too. Not a re-export.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
buildcheck of the topic: 6/6 clean. The other patches' content and
messages are unchanged.
+### rb1103 on the 6.18 lane
+
+6.18 compress-fixes 85cc49ced29b: ours (5baf35ed0476) dropped, Filipe's in
+its place, from the posted mbox (applies with an offset; message as posted
++ Qu's Reviewed-by + Link). Since 2026-11-03 08:57 UTC it is also
+btrfs/for-next 75d0e4dd64f2a; on 6.18 it moves to the topic head as a
+cherry-pick when the lanes rebase onto a base that has it. Per-commit
+buildcheck 8/8 (from v6.18.60) clean.
+
+| test | kernel | result | evidence |
+|---|---|---|---|
+| compress-props-test.sh | 6.18 #313 (zb64 cc5b17dda57d, KASAN inline + lockdep + BTRFS_ASSERT, bzImage md5 7e052fac6664fed02bcaeb36e6c3bc41), bhive and ahive | PASS on both: I1 "-c--", I2 "-c-A", I3 8/8 extents encoded; L1-L3 on the fixture pass; every flags/props/persist case passes | raid56-wh-harness `forensics/rb1103-618/{bhive,ahive}/rb1103-compress-props.log` |
+| same (negative control) | 6.18 #311 (rb1101 lane, our old patch 2) | FAIL in I1-I3 only: "C---", 1 extent, 0 encoded (I2's ioctl succeeds but the flags are "C--A"); legacy on the fixture passes | raid56-wh-harness `forensics/rb1103-618/ahive/compress-props-311.log` |
+
## Not covered / open
- (Obsolete since rb1103: the patch that handled a parent directory
scrub-uncorrectable-rc, failing-member-scrub) PASSed on dhive, as in
rb1101. Evidence: raid56-wh-harness `forensics/rb1103-mn/`.
+### rb1103 on the 6.18 lane
+
+Same change on 6.18 (topic dcf5460a7aaf): message-only, the topic tree
+is byte-identical to rb1101's (a82b570d4787), whose per-commit buildcheck
+(14/14) stands. Kernel 6.18 #313 (zb64 cc5b17dda57d, KASAN inline + lockdep + BTRFS_ASSERT, bzImage md5 7e052fac6664fed02bcaeb36e6c3bc41). bhive: rb1028-regress without qu-p1, every
+test PASS (write-error-cache 7/7, rmw-writeback, scrub-uncorrectable-rc,
+failing-member-scrub), retire-wait-abort SKIP 77 as before; no kernel
+report, taint 0. Same as rb1101. Evidence: raid56-wh-harness `forensics/rb1103-618/`.
+
## Not covered / open
- Plain writers (Zygo 2026-10-20: a write is acknowledged when write(2)
replace-uncorrectable-test.sh rc 0 (readback eio_unexpected=0). Same as
rb1101. Evidence: raid56-wh-harness `forensics/rb1103-mn/ehive/`.
+### rb1103 on the 6.18 lane
+
+Same change on 6.18 (topic e5522f716fab): message-only, topic tree
+byte-identical to rb1101b's (422138aeebb6); its buildcheck stands.
+Kernel 6.18 #313 (zb64 cc5b17dda57d, KASAN inline + lockdep + BTRFS_ASSERT, bzImage md5 7e052fac6664fed02bcaeb36e6c3bc41). ahive: replace-discard-test.sh, the same 11 cases, all PASS
+(fstrim-single/raid1/raid5 and resume-single: 0 discard sectors on the
+target; every case verify_full 20237/20237); replace-status-test.sh rc 0
+(clean, cancel, unc-raid1 3/3, unc-raid5 3/3, werr canceled "failed -5"
+at 15.9%); replace-uncorrectable-test.sh 8 profiles rc 0
+(eio_unexpected=0). No kernel report, taint 0. Same as rb1101.
+Evidence: raid56-wh-harness `forensics/rb1103-618/ahive/`.
+
## Not covered / open
- Patch 1's commit-failure branch (the commit after the cleanup fails; only
So v2 3/5 is build- and regression-checked, and its wait is shown not to
hang. Its fix is not shown by any test here. Evidence: raid56-wh-harness `forensics/rb1103-mn/dhive/`.
+### rb1103 on the 6.18 lane
+
+6.18 stripe-alloc a2590646602f: after the disguise backport, 2/5 message
+refresh (content equal to v2; the 6.18 note stays), v2 3/5 and v2 4/5
+replace v1 in place; 84 later commits replayed without a conflict, every
+per-commit content diff unchanged; low-space and test-knobs restacked.
+6.18 ports (bracketed notes in the commits):
+- 3/5: cow_write_failed() takes ordered_tree_lock with spin_lock_irq().
+ On 6.18 that lock is taken from bio completion with IRQs disabled
+ (btrfs_finish_ordered_extent(), btrfs_mark_ordered_io_finished() and
+ btrfs_dec_test_ordered_pending() use spin_lock_irqsave()); misc-next takes
+ it plain. btrfs_inode.h and ordered-data.c: context only. IO_DONE is
+ woken on ordered->wait by both 6.18 completion paths. This lane's fast
+ fsync already kicks the parked rbios of every collected ordered extent
+ before the writeback wait, so the new wait does not sit out a park.
+- 4/5: v1's 6.18 context (no trace_btrfs_sync_log_exit(), log_commit[]
+ cleared with atomic_set()); content equal to v2.
+
+Builds: per-commit buildcheck stripe-alloc 89/89 (from reclaim-fixes),
+low-space 8/8, test-knobs 3/3, clean. Kernel 6.18 #313 (zb64 cc5b17dda57d, KASAN inline + lockdep + BTRFS_ASSERT, bzImage md5 7e052fac6664fed02bcaeb36e6c3bc41); fs/btrfs warnings only
+the two known KASAN frame sizes.
+
+Gate on bhive (as rb1101): rb1028-regress without qu-p1 all PASS,
+retire-wait-abort SKIP 77; enospc-dataloss fsync/nofsync no loss of acked
+data; enospc-drop-fsync baseline arms BLIND x4 (no drop; 8 fsyncs 0, CLEAN
+and CRASH scrub clean); stripe-split-narrow PASS; log-settle-unwritten
+PASS (truncated OE hit at iteration 0, fsync returned, settle 0 -> 2).
+Kernel: no KASAN/WARN/BUG/lockdep/hung task over the boot, taint 0,
+debug_locks 1, 4/5's DEBUG_WARN never fired, NORESERVE 0; 26
+unwritten-on-end lines, all 0xa1 (retire-wait-abort and write-error-cache
+windows), as in rb1101.
+
+3/5 coverage on 6.18: as above, nothing reaches the IOERR branch (BLIND);
+every fast fsync in the gate ran the IO_DONE wait and the irq-safe lock
+under lockdep with no report, which covers the 6.18 port's lock class.
+Evidence: raid56-wh-harness `forensics/rb1103-618/`.
+
## Not covered / open
Patches without a test of their own (covered at most incidentally by the