From: Cen Zhang Date: Wed, 1 Apr 2026 02:21:53 +0000 (+0800) Subject: btrfs: annotate lockless read of defrag_bytes in should_nocow() X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=89c0dc3de7a73e8aba5e9bfef543eee047a3d0d2;p=thirdparty%2Flinux.git btrfs: annotate lockless read of defrag_bytes in should_nocow() should_nocow() reads inode->defrag_bytes without holding inode->lock, while btrfs_set_delalloc_extent() and btrfs_clear_delalloc_extent() update it under that spinlock. This is a data race. The read is a quick check used to decide whether to fall back to COW for a NOCOW inode: if defrag_bytes is non-zero and the range is tagged EXTENT_DEFRAG, we force COW so that defragmentation can rewrite the extent. Reading a stale value is harmless because: - A missed increment may skip COW once, but the defrag pass will redo the extent later. - A stale non-zero may force an unnecessary COW, which is a minor efficiency loss, not a correctness issue. On 64-bit platforms an aligned u64 load is naturally atomic so tearing cannot happen. On 32-bit platforms u64 may tear, but we only test for zero vs non-zero, so the heuristic stays correct regardless. Use data_race() annotation. Fixes: 47059d930f0e ("Btrfs: make defragment work with nodatacow option") Signed-off-by: Cen Zhang [ Use data_race() instead of READ_ONCXE() ] Signed-off-by: David Sterba --- diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 9ce6200854672..61b5594c4206f 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2293,7 +2293,7 @@ error: static bool should_nocow(struct btrfs_inode *inode, u64 start, u64 end) { if (inode->flags & (BTRFS_INODE_NODATACOW | BTRFS_INODE_PREALLOC)) { - if (inode->defrag_bytes && + if (data_race(inode->defrag_bytes) && btrfs_test_range_bit_exists(&inode->io_tree, start, end, EXTENT_DEFRAG)) return false; return true;