* [PATCH v2] ceph: fix use-after-free bug for inodes when flushing capsnaps
@ 2023-06-01 1:24 xiubli
0 siblings, 0 replies; only message in thread
From: xiubli @ 2023-06-01 1:24 UTC (permalink / raw)
To: idryomov, ceph-devel; +Cc: jlayton, vshankar, mchangir, Xiubo Li, stable
From: Xiubo Li <xiubli@redhat.com>
There is racy between capsnaps flush and removing the inode from
'mdsc->snap_flush_list' list:
== Thread A == == Thread B ==
ceph_queue_cap_snap()
-> allocate 'capsnapA'
->ihold('&ci->vfs_inode')
->add 'capsnapA' to 'ci->i_cap_snaps'
->add 'ci' to 'mdsc->snap_flush_list'
...
== Thread C ==
ceph_flush_snaps()
->__ceph_flush_snaps()
->__send_flush_snap()
handle_cap_flushsnap_ack()
->iput('&ci->vfs_inode')
this also will release 'ci'
...
== Thread D ==
ceph_handle_snap()
->flush_snaps()
->iterate 'mdsc->snap_flush_list'
->get the stale 'ci'
->remove 'ci' from ->ihold(&ci->vfs_inode) this
'mdsc->snap_flush_list' will WARNING
To fix this we will increase the ihold the inode when adding 'ci'
to the 'mdsc->snap_flush_list' list.
Cc: stable@vger.kernel.org
URL: https://bugzilla.redhat.com/show_bug.cgi?id=2209299
Signed-off-by: Xiubo Li <xiubli@redhat.com>
---
V2:
- Increase the i_count reference instead to fix this issue
fs/ceph/caps.c | 6 ++++++
fs/ceph/snap.c | 4 +++-
2 files changed, 9 insertions(+), 1 deletion(-)
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index feabf4cc0c4f..7c2cb813aba4 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -1684,6 +1684,7 @@ void ceph_flush_snaps(struct ceph_inode_info *ci,
struct inode *inode = &ci->netfs.inode;
struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
struct ceph_mds_session *session = NULL;
+ int put = 0;
int mds;
dout("ceph_flush_snaps %p\n", inode);
@@ -1728,8 +1729,13 @@ void ceph_flush_snaps(struct ceph_inode_info *ci,
ceph_put_mds_session(session);
/* we flushed them all; remove this inode from the queue */
spin_lock(&mdsc->snap_flush_lock);
+ if (!list_empty(&ci->i_snap_flush_item))
+ put++;
list_del_init(&ci->i_snap_flush_item);
spin_unlock(&mdsc->snap_flush_lock);
+
+ if (put)
+ iput(inode);
}
/*
diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c
index 5a4bf0201737..d5ad10d94424 100644
--- a/fs/ceph/snap.c
+++ b/fs/ceph/snap.c
@@ -697,8 +697,10 @@ int __ceph_finish_cap_snap(struct ceph_inode_info *ci,
capsnap->size);
spin_lock(&mdsc->snap_flush_lock);
- if (list_empty(&ci->i_snap_flush_item))
+ if (list_empty(&ci->i_snap_flush_item)) {
+ ihold(inode);
list_add_tail(&ci->i_snap_flush_item, &mdsc->snap_flush_list);
+ }
spin_unlock(&mdsc->snap_flush_lock);
return 1; /* caller may want to ceph_flush_snaps */
}
--
2.40.1
^ permalink raw reply related [flat|nested] only message in thread
only message in thread, other threads:[~2023-06-01 1:27 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-06-01 1:24 [PATCH v2] ceph: fix use-after-free bug for inodes when flushing capsnaps xiubli
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).