|
|
Log in / Subscribe / Register

net: af_unix: useful handling of LSM denials on SCM_RIGHTS

From:  Jori Koolstra <jkoolstra-AT-xs4all.nl>
To:  Christian Brauner <brauner-AT-kernel.org>, Aleksa Sarai <cyphar-AT-cyphar.com>, Kuniyuki Iwashima <kuniyu-AT-google.com>, "David S . Miller" <davem-AT-davemloft.net>, Eric Dumazet <edumazet-AT-google.com>, Jakub Kicinski <kuba-AT-kernel.org>, Paolo Abeni <pabeni-AT-redhat.com>, Simon Horman <horms-AT-kernel.org>
Subject:  [PATCH net-next v4 0/3] net: af_unix: useful handling of LSM denials on SCM_RIGHTS
Date:  Sun, 05 Jul 2026 14:38:23 +0200
Message-ID:  <20260705123826.3818443-1-jkoolstra@xs4all.nl>
Cc:  netdev-AT-vger.kernel.org, linux-fsdevel-AT-vger.kernel.org, linux-kernel-AT-vger.kernel.org, Jori Koolstra <jkoolstra-AT-xs4all.nl>
Archive-link:  Article

Right now if some LSM denies an AF_UNIX socket peer to receive a
SCM_RIGHTS fd, the SCM_RIGHTS fd array will be cut short at
that point, and MSG_CTRUNC is set on return of recvmsg(2). This is
highly problematic behaviour, because it leaves the receiver
wondering what happened. As per man page MSG_CTRUNC is supposed to
indicate that the control buffer was sized too short, but suddenly
a permission error might result in the exact same flag being set.
Moreover, the receiver has no chance to determine how many fds got
originally sent and how many were suppressed.[1]

Add a SO_RIGHTS_NOTRUNC option to UNIX sockets to enable more useful
handling of LSM denials when receiving SCM_RIGHTS messages: instead of
truncating the message at the first blocked fd, keep every fd slot
and store the LSM errno in the blocked slot.

[1]: https://github.com/uapi-group/kernel-features#useful-hand...

Changes:
v4:
  - Removed the __receive_fd() helper and moved logic into
    scm_recv_one_fd() directly (suggested by Brauner).
  - Moved selftest from Smack to BPF (LLM assisted).
  - Add arch specific socket option values for SO_RIGHTS_NOTRUNC.
  - Undo patch that replaced copy_from_sockptr() with
    copy_safe_from_sockptr().
v3:
  - Separated net and vfs changes.
  - Use kselftest_harness.h and system() to call the test script.
v2: https://lore.kernel.org/netdev/20260616143020.3458085-2-j...
  - Reimplemented as a UNIX socket option instead of a per recvmsg(2) flag.
v1: https://lore.kernel.org/netdev/20260428175125.2705296-1-j...

Jori Koolstra (3):
  net: scm: move scm_detach_fds() from common path to scm_recv_unix()
  net: af_unix: useful handling of LSM denials on SCM_RIGHTS
  selftest: Add tests for useful handling of LSM denials on SCM_RIGHTS

 arch/alpha/include/uapi/asm/socket.h          |   2 +
 arch/mips/include/uapi/asm/socket.h           |   2 +
 arch/parisc/include/uapi/asm/socket.h         |   2 +
 arch/sparc/include/uapi/asm/socket.h          |   2 +
 include/net/af_unix.h                         |   1 +
 include/net/scm.h                             |  13 +-
 include/uapi/asm-generic/socket.h             |   2 +
 net/compat.c                                  |   4 +-
 net/core/scm.c                                |  42 ++-
 net/unix/af_unix.c                            |   9 +
 .../testing/selftests/net/af_unix/.gitignore  |   2 +
 tools/testing/selftests/net/af_unix/Makefile  |   8 +
 .../net/af_unix/scm_rights_denial_lsm.bpf.c   |  36 +++
 .../net/af_unix/scm_rights_denial_lsm.c       | 263 ++++++++++++++++++
 14 files changed, 371 insertions(+), 17 deletions(-)
 create mode 100644 tools/testing/selftests/net/af_unix/scm_rights_denial_lsm.bpf.c
 create mode 100644 tools/testing/selftests/net/af_unix/scm_rights_denial_lsm.c


base-commit: b73bc9ca3686b78b642fb35dcc1fdf874ecb74a1
-- 
2.55.0




Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds