Hi, Jens! 在 2025/02/27 15:54, Zheng Qixing 写道:
From: Zheng Qixing <zhengqix...@huawei.com>Hi, during RAID feature implementation testing, we found several bugs in badblocks. This series contains bugfixes and cleanups for MD RAID badblocks handling code. V2: - patch 4: add a description of the issue - patch 5: add comment of parital setting - patch 6: add fix tag - patch 10: two code style modifications - patch 11: keep original functionality of rdev_clear_badblocks(), functionality was incorrectly modified in V1. - patch 1-10 and patch 12 are reviewed by Yu Kuai <yuku...@huawei.com> - patch 1, 3, 5, 6, 8, 9, 10, 12 are acked by Coly Li <col...@kernel.org> Li Nan (8): badblocks: Fix error shitf ops badblocks: factor out a helper try_adjacent_combine badblocks: attempt to merge adjacent badblocks during ack_all_badblocks badblocks: return error directly when setting badblocks exceeds 512 badblocks: return error if any badblock set fails badblocks: fix the using of MAX_BADBLOCKS badblocks: try can_merge_front before overlap_front badblocks: fix merge issue when new badblocks align with pre+1 Zheng Qixing (4): badblocks: fix missing bad blocks on retry in _badblocks_check() badblocks: return boolean from badblocks_set() and badblocks_clear() md: improve return types of badblocks handling functions badblocks: use sector_t instead of int to avoid truncation of badblocks length
This set contains fixes that are found by testing mdraid, please consider this set for the next merge window, or I can apply it to md-6.15. Thanks, Kuai
block/badblocks.c | 322 +++++++++++++--------------------- drivers/block/null_blk/main.c | 16 +- drivers/md/md.c | 48 ++--- drivers/md/md.h | 14 +- drivers/md/raid1-10.c | 2 +- drivers/md/raid1.c | 10 +- drivers/md/raid10.c | 14 +- drivers/nvdimm/badrange.c | 2 +- drivers/nvdimm/nd.h | 2 +- drivers/nvdimm/pfn_devs.c | 7 +- drivers/nvdimm/pmem.c | 2 +- include/linux/badblocks.h | 10 +- 12 files changed, 183 insertions(+), 266 deletions(-)
