In revision 1.228, I accidentally broke the "total number of processes in

the system" resource limit code: When checking if the caller has superuser
privileges, we should be checking the *real* user, not the *effective*
user.  (In general, resource limiting is done based on the real user, in
order to avoid resource-exhaustion-by-setuid-program attacks.)

Now that a SUSER_RUID flag to suser_cred exists, use it here to return
this code to its correct behaviour.

Pointed out by:	rwatson

1 files changed, 2 insertions, 1 deletions

diff --git a/sys/kern/kern_fork.c b/sys/kern/kern_fork.c
index e42a436..8ac003d 100644
--- a/sys/kern/kern_fork.c
+++ b/sys/kern/kern_fork.c
@@ -300,7 +300,8 @@ fork1(td, flags, pages, procp)
 	 */
 	sx_xlock(&allproc_lock);
 	uid = td->td_ucred->cr_ruid;
-	if ((nprocs >= maxproc - 10 && suser(td) != 0) ||
+	if ((nprocs >= maxproc - 10 &&
+	    suser_cred(td->td_ucred, SUSER_RUID) != 0) ||
 	    nprocs >= maxproc) {
 		error = EAGAIN;
 		goto fail;