[SRU Bionic 0/2] CVE-2020-36322 and CVE-2021-28950

Thadeu Lima de Souza Cascardo cascardo at canonical.com
Wed Oct 13 19:18:54 UTC 2021


[Impact]
Unprivileged users could mount a fuse filesystem and trigger a BUG_ON or
a soft lockup.

[Test case]
A test was found for the first issue, but not the second. The fix works for
that first one.

[Backport]
The backport had a lot of conflicts due to context. As the nature of the fix
was basically replacing is_bad_inode with fuse_is_bad and adding the check
for fuse_is_bad, it is not that hard to review that the end result is as expected.

[Potential regression]
FUSE filesystems may lock up, trigger BUGs, or fail to respond. It is more
likely to happen due to malicious actions. So, there is a chance that
forcing bad inodes may lead to other potential hiccups.

Amir Goldstein (1):
  fuse: fix live lock in fuse_iget()

Miklos Szeredi (1):
  fuse: fix bad inode

 fs/fuse/acl.c    |  6 ++++++
 fs/fuse/dir.c    | 41 ++++++++++++++++++++++++++++++++++++-----
 fs/fuse/file.c   | 21 ++++++++++++---------
 fs/fuse/fuse_i.h | 13 +++++++++++++
 fs/fuse/inode.c  |  2 +-
 fs/fuse/xattr.c  |  9 +++++++++
 6 files changed, 77 insertions(+), 15 deletions(-)

-- 
2.30.2




More information about the kernel-team mailing list