Hello,
Please find this small patch series which fixes one of the issue (causing data
abort exception) identified with fast_commit and flex_bg.
Although I have given -g quick, log group a run and didn't see any surprise
there. But a careful review in Patch-1 & Patch-6 will surely help! :)
Will shortly send out the fstest patch, which could trigger this.
Patch details
==============
Patch-1: Fixes a data abort which could happen during recovery with flex_bg
feature. This might be even needed to cc to stable, right?
Patch-[2-5]: Minor cleanups
Patch[6]: Good to have to avoid any accidental set/clear of critical FS Metadata
Ritesh Harjani (6):
ext4: Fixes ext4_mb_mark_bb() with flex_bg with fast_commit
ext4: Implement ext4_group_block_valid() as common function
ext4: Use in_range() for range checking in ext4_fc_replay_check_excluded
ext4: No need to test for block bitmap bits in ext4_mb_mark_bb()
ext4: Refactor ext4_free_blocks() to pull out ext4_mb_clear_bb()
ext4: Add extra check in ext4_mb_mark_bb() to prevent against possible corruption
fs/ext4/block_validity.c | 31 ++++++
fs/ext4/ext4.h | 3 +
fs/ext4/fast_commit.c | 4 +-
fs/ext4/mballoc.c | 235 +++++++++++++++++++++++----------------
4 files changed, 176 insertions(+), 97 deletions(-)
--
2.31.1