The resolve commands never escaped their output, so patch 5 no longer
sets -z against escaping: it ends each path (each -P line) with NUL
instead of a newline, for names that contain newlines. Text only
(usage, rst, commit message, a 074 comment); topic c9ea03d8, testing 71ea1475. Cover 0/5 reworded the same way; TESTING.md and README
updated with the new hashes and the ahive results.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Escaped mount point and directory names were already decoded by the
mountinfo parser; patch 4 gains tests for them. New patch 5 adds -z
(--zero) to logical-resolve and inode-resolve (topic 85ca1f91, testing d541f776). Cover 0/5 and TESTING.md updated; the wrong "escapes not
decoded" note is removed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zygo Blaxell [Fri, 30 Oct 2026 22:58:57 +0000 (18:58 -0400)]
btrfs-progs topic series: logical-resolve-fixes bind mount route, test 074
Patch 2 resolves per path through four routes, including bind mounts of
directories inside the subvolume; patch 3 makes inode-resolve use them
and adds misc-tests/074; patch 4 moves the new messages to stderr
(topic bed8b5eb, testing a32d9351). Cover 0/4 and TESTING.md updated;
the other topics' notes point at the new testing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zygo Blaxell [Fri, 30 Oct 2026 21:30:59 +0000 (17:30 -0400)]
btrfs-progs topic series: logical-resolve-fixes gets the inode-resolve patch
New patch 3 "btrfs-progs: inspect inode-resolve: print paths relative to
the argument" (topic da97c420, testing 14c065a7); the stderr patch is now
patch 4. Cover and TESTING.md updated; the other topics' notes point at
the new testing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zygo Blaxell [Fri, 30 Oct 2026 20:59:26 +0000 (16:59 -0400)]
btrfs-progs topic series: logical-resolve through bind mounts, re-export
logical-resolve-fixes: the subvolume lookup patch now also resolves
references when the path argument is a directory inside their subvolume
(a bind mount of a directory, or a directory in one), relative to the
argument; the stderr patch's misc-tests/042 update gains bind mount
cases (bind is the only way in; argument inside a bind; bind of a
directory of a nested subvolume). Cover letter and TESTING.md updated
with the per-case results on ahive. testing 4ba66afa -> 6b5d8220; the
other covers only get a new Date: header, and every TESTING.md's
full-suite note now names 6b5d8220. Verified by export.sh: patches
rebuild every topic tree; the leaf topics merged give testing's tree
with no fixup.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zygo Blaxell [Fri, 30 Oct 2026 14:26:09 +0000 (10:26 -0400)]
btrfs-progs topic series: drop the hex fix, full test suite results
Zygo dropped "btrfs-progs: accept hexadecimal and octal numbers in
arg_strtou64() again" from logical-resolve-fixes (no complaints in
years; decimal or $((0x...)) works). The logical-resolve cover is now
0/3: defect 6 is left out on purpose, with the workaround, and the
octal note is gone; its TESTING.md has the decimal-only reruns (misc
042 PASS, repro 86/98 with the other 12 really unreachable) and the
lane SHA map. README updated.
Every TESTING.md gets the full test suite results (testing 3bc81a49 on
ahive, harness b243013: 0 failures, environment-only first-pass failures
fixed, not-run tests explained; 3bc81a49 still had the hex patch and the
collapsed 042 line).
Exported from testing 4ba66afacd77 (devel 9d01fb77f46e):
mkfs-remap-tree 4e9916543650 (1), dump-tree-replace-stop-time 41723c4246ca (1), replace-abort-uncorrectable 52210d21fbf4 (1),
logical-resolve-fixes cb9e0be5cb68 (3), scrub-limit-only-with-option c6bdae692814 (2), scrub-progress-socket 37cc6ecb63f7 (2, on
scrub-limit-only-with-option). Verified: patches rebuild every topic
tree; the leaf topics merged give testing's tree with no fixup.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zygo Blaxell [Fri, 30 Oct 2026 13:48:05 +0000 (09:48 -0400)]
btrfs-progs topic series: stack scrub-progress-socket, re-export after the message rewrite
The lane's commit messages were rewritten in place (Fixes:, Link:, the
kernel dependency of replace-abort-uncorrectable, subjects <= 72
columns, trailing blank lines) and misc-tests/042's line continuation
was restored in logical-resolve-fixes patch 4. scrub-progress-socket is
now stacked on scrub-limit-only-with-option, with the combined code in
its patch 2, so the integration needs no fixup:
fixups/post-merge.patch is gone.
scripts/export.sh: stacked topics are exported against their parent
topic with --base at the lane base (prerequisite-patch-id lines), as in
the kernel export; a stacked-on topic need not be a testing parent; the
fixup file is written only when non-empty. README and the two scrub
covers describe the stacking; each TESTING.md maps the old SHAs.
Verified: each directory's patches applied to the base (after the series
it is stacked on) give the topic's tree; the leaf topics merged in
testing's order give testing's tree with no fixup (also with git am /
git merge in a scratch clone); every series applies to the current
origin/devel 55f4cfb9c932.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zygo Blaxell [Fri, 30 Oct 2026 13:38:46 +0000 (09:38 -0400)]
btrfs-progs topic series: cover letters and TESTING.md
Cover letters for the three multi-patch series (scrub-limit-only-with-option,
logical-resolve-fixes, scrub-progress-socket) and a TESTING.md for every
topic, from the recorded runs in raid56-wh-harness forensics
(scrub-limit-20261029, logical-resolve-20261029, replace-resume-20261029,
replace-abort-20261016, replace-retrofit-20261016, t-ahive/t-dhive
topics-20261018, bhive-remap-balance-resume-20261013, r-rebase-20261023;
harness up to 7356cc6). The full progs test suite section is pending in
each file.
Re-exported (scripts/export.sh keeps the cover text) from testing 3bc81a49b73f, devel 9d01fb77f46e: mkfs-remap-tree ae91bb721fd5 (1),
dump-tree-replace-stop-time 7246e18c809f (1), replace-abort-uncorrectable 2ee0ef38412e (1), scrub-limit-only-with-option 02e0c99bb36d (2),
logical-resolve-fixes 08cfeba28ba1 (4), scrub-progress-socket 0d535ece7a9a (2). Verified as in the previous commit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zygo Blaxell [Fri, 30 Oct 2026 13:34:17 +0000 (09:34 -0400)]
btrfs-progs topic series: export script, README and first export
Mirror the kernel upstream branch for the btrfs-progs lane: one
directory per topics/* branch, format-patch against testing^1 (devel 9d01fb77f46e, v7.1-102-g9d01fb77), plus fixups/post-merge.patch for the
one conflict between scrub-limit-only-with-option and
scrub-progress-socket.
scripts/export.sh keeps each cover's title and text and regenerates the
headers, shortlog, diffstat and base-commit; it replaces only *.patch
files and never removes a directory. The three multi-patch series carry
placeholder covers in this commit.
Verified: each directory's patches applied to the base give the topic's
tree; all 6 merged in testing's parent order with -X ours plus
fixups/post-merge.patch give testing's tree (also with git am / git
merge in a scratch clone). Every series also applies to the current
origin/devel 55f4cfb9c932.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>