Re: [PATCH V2] audit: log 32-bit socketcalls

On 2017-01-16 15:04, Paul Moore wrote:

On Fri, Jan 13, 2017 at 9:42 AM, Eric Paris [off-list ref] wrote:

quoted

On Fri, 2017-01-13 at 04:51 -0500, Richard Guy Briggs wrote:

quoted

diff --git a/include/linux/audit.h b/include/linux/audit.h
index 9d4443f..43d8003 100644
--- a/include/linux/audit.h
+++ b/include/linux/audit.h

@@ -387,6 +387,18 @@ static inline int audit_socketcall(int nargs,

unsigned long *args)
              return __audit_socketcall(nargs, args);
      return 0;
 }
+static inline int audit_socketcall_compat(int nargs, u32 *args)
+{
+     if (unlikely(!audit_dummy_context())) {

I've always hated these likely/unlikely. Mostly because I think they
are so often wrong. I believe this says that you compiled audit in but
you expect it to be explicitly disabled. While that is (recently) true
in Fedora I highly doubt that's true on the vast majority of systems
that have audit compiled in.

Richard and I have talked about the likely/unlikely optimization
before and I know Richard likes to use them, but I don't for the
reasons Eric has already mentioned.   Richard, since you're respinning
the patch, go ahead and yank out the unlikely() call.

I don't "like to use them".  I'm simply following the use and style of
existing code and the arguments of others in places of critical
performance.  If I "fix" that one, then I would feel compelled to yank
out the one in the function immediately above, audit_socketcall() for
consistency to ease my conscience.  Eric conceded that argument.

paul moore

- RGB

--
Richard Guy Briggs [off-list ref]
Kernel Security Engineering, Base Operating Systems, Red Hat
Remote, Ottawa, Canada
Voice: +1.647.777.2635, Internal: (81) 32635