On 3/2/20 7:58 PM, Qu Wenruo wrote:
On 2020/3/3 上午2:47, Josef Bacik wrote:If we fail to load an fs root, or fail to start a transaction we can bail without unsetting the reloc control, which leads to problems later when we free the reloc control but still have it attached to the file system. Signed-off-by: Josef Bacik <josef@xxxxxxxxxxxxxx> --- fs/btrfs/relocation.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 507361e99316..173fc7628235 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -4678,6 +4678,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) fs_root = read_fs_root(fs_info, reloc_root->root_key.offset); if (IS_ERR(fs_root)) { err = PTR_ERR(fs_root); + unset_reloc_control(rc); list_add_tail(&reloc_root->root_list, &reloc_roots); goto out_free; }Shouldn't the unset_reloc_control() also get called for all related errors after set_reloc_control()? That includes btrfs_join_transaction() (the one after set_reloc_contrl(), which looks missing), the read_fs_root() and the commit transaction error you're addressing.
It's already doing unset in the join_transaction right after calling set. These are the only two missing paths. Thanks,
Josef
