system/linux
system/linux
1
0
Fork 0
mirror of https://github.com/torvalds/linux synced 2024-10-03 01:43:05 +00:00
Code Issues Packages Projects Releases Wiki Activity

bcachefs: Fix I_NEW warning in race path in bch2_inode_insert()

Browse source 
discard_new_inode() is the correct interface for tearing down an indoe
that was fully created but not made visible to other threads, but it
expects I_NEW to be set, which we don't use.

Reported-by: https://github.com/koverstreet/bcachefs/issues/690
Fixes: bcachefs: Fix race path in bch2_inode_insert()
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
This commit is contained in:
Kent Overstreet 2024-06-20 13:20:49 -04:00
parent 504794067f
commit dd9086487c
1 changed files with 10 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

12
fs/bcachefs/fs.c
View file

@ -188,6 +188,12 @@ static struct bch_inode_info *bch2_inode_insert(struct bch_fs *c, struct bch_ino
BUG_ON(!old);
if (unlikely(old != inode)) {
/*
* bcachefs doesn't use I_NEW; we have no use for it since we
* only insert fully created inodes in the inode hash table. But
* discard_new_inode() expects it to be set...
*/
inode->v.i_flags |= I_NEW;
discard_new_inode(&inode->v);
inode = old;
} else {
@ -195,8 +201,10 @@ static struct bch_inode_info *bch2_inode_insert(struct bch_fs *c, struct bch_ino
list_add(&inode->ei_vfs_inode_list, &c->vfs_inodes_list);
mutex_unlock(&c->vfs_inodes_lock);
/*
* we really don't want insert_inode_locked2() to be setting
* I_NEW...
* Again, I_NEW makes no sense for bcachefs. This is only needed
* for clearing I_NEW, but since the inode was already fully
* created and initialized we didn't actually want
* inode_insert5() to set it for us.
*/
unlock_new_inode(&inode->v);
}