On 7/19/2024 10:08 AM, Paul Moore wrote:

On Jul 11, 2024 Xu Kuohai [off-list ref] wrote:

quoted

To be consistent with most LSM hooks, convert the return value of
hook setprocattr to 0 or a negative error code.

Before:
- Hook setprocattr returns the number of bytes written on success
   or a negative error code on failure.

After:
- Hook setprocattr returns 0 on success or a negative error code
   on failure. An output parameter @wbytes is introduced to hold
   the number of bytes written on success.

Signed-off-by: Xu Kuohai <redacted>
---
  fs/proc/base.c                |  5 +++--
  include/linux/lsm_hook_defs.h |  3 ++-
  include/linux/security.h      |  5 +++--
  security/apparmor/lsm.c       | 10 +++++++---
  security/security.c           |  8 +++++---
  security/selinux/hooks.c      | 11 ++++++++---
  security/smack/smack_lsm.c    | 14 ++++++++++----
  7 files changed, 38 insertions(+), 18 deletions(-)

The security_setprocattr() hook is another odd case that we probably
just want to leave alone for two reasons:

1. With the move to LSM syscalls for getting/setting a task's LSM
attributes we are "freezing" the procfs API and not adding any new
entries to it.

2. The BPF LSM doesn't currently register any procfs entries.

I'd suggest leaving security_setprocattr() as-is and blocking it in
the BPF verifier, I can't see any reason why a BPF LSM would need
this hook.

OK, I'll drop this patch in the next version.