Hi all, Years ago, Christoph diagnosed a problem where freeing an inode on a totally full filesystem could fail due to finobt expansion not being able to allocate enough blocks. He solved the problem by using the per-AG block reservation system to ensure that there are always enough blocks for finobt expansion, but that came at the cost of having to walk the entire finobt at mount time. This new feature solves that performance regression by adding inode btree block counts to the AGI header. v2: rebase kernel to 5.9 v3: split logical changes into separate patches v4: support inobtcounts && !finobt properly If you're going to start using this mess, you probably ought to just pull from my git trees, which are linked below. This is an extraordinary way to destroy everything. Enjoy! Comments and questions are, as always, welcome. --D kernel git tree: https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=inobt-counters xfsprogs git tree: https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=inobt-counters fstests git tree: https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=inobt-counters --- db/agi.c | 2 + db/sb.c | 78 ++++++++++++++++++++++++++++++++++++++++++++- db/xfs_admin.sh | 4 ++ libxfs/xfs_ag.c | 5 +++ libxfs/xfs_format.h | 21 +++++++++++- libxfs/xfs_ialloc.c | 1 + libxfs/xfs_ialloc_btree.c | 65 +++++++++++++++++++++++++++++++++++--- man/man8/mkfs.xfs.8 | 15 +++++++++ man/man8/xfs_admin.8 | 16 +++++++++ mkfs/xfs_mkfs.c | 34 +++++++++++++++++++- repair/phase5.c | 5 +++ repair/scan.c | 38 ++++++++++++++++++++-- 12 files changed, 272 insertions(+), 12 deletions(-)
