提交 8b218b8a 编写于 作者: S Sage Weil

ceph: fix inode removal from snap realm when racing with migration

When an inode was dropped while being migrated between two MDSs,
i_cap_exporting_issued was non-zero such that issue caps were non-zero and
__ceph_is_any_caps(ci) was true.  This prevented the inode from being
removed from the snap realm, even as it was dropped from the cache.

Fix this by dropping any residual i_snap_realm ref in destroy_inode.
Signed-off-by: NSage Weil <sage@newdream.net>
上级 052bb34a
......@@ -378,6 +378,22 @@ void ceph_destroy_inode(struct inode *inode)
ceph_queue_caps_release(inode);
/*
* we may still have a snap_realm reference if there are stray
* caps in i_cap_exporting_issued or i_snap_caps.
*/
if (ci->i_snap_realm) {
struct ceph_mds_client *mdsc =
&ceph_client(ci->vfs_inode.i_sb)->mdsc;
struct ceph_snap_realm *realm = ci->i_snap_realm;
dout(" dropping residual ref to snap realm %p\n", realm);
spin_lock(&realm->inodes_with_caps_lock);
list_del_init(&ci->i_snap_realm_item);
spin_unlock(&realm->inodes_with_caps_lock);
ceph_put_snap_realm(mdsc, realm);
}
kfree(ci->i_symlink);
while ((n = rb_first(&ci->i_fragtree)) != NULL) {
frag = rb_entry(n, struct ceph_inode_frag, node);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册