[PATCH] Audit Filter Performance
authorSteve Grubb <sgrubb@redhat.com>
Tue, 11 Apr 2006 12:50:56 +0000 (08:50 -0400)
committerAl Viro <viro@zeniv.linux.org.uk>
Mon, 1 May 2006 10:10:07 +0000 (06:10 -0400)
While testing the watch performance, I noticed that selinux_task_ctxid()
was creeping into the results more than it should. Investigation showed
that the function call was being called whether it was needed or not. The
below patch fixes this.

Signed-off-by: Steve Grubb <sgrubb@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
kernel/auditsc.c

index a300736..1c03a4e 100644 (file)
@@ -168,11 +168,9 @@ static int audit_filter_rules(struct task_struct *tsk,
                              struct audit_context *ctx,
                              enum audit_state *state)
 {
-       int i, j;
+       int i, j, need_sid = 1;
        u32 sid;
 
-       selinux_task_ctxid(tsk, &sid);
-
        for (i = 0; i < rule->field_count; i++) {
                struct audit_field *f = &rule->fields[i];
                int result = 0;
@@ -271,11 +269,16 @@ static int audit_filter_rules(struct task_struct *tsk,
                           match for now to avoid losing information that
                           may be wanted.   An error message will also be
                           logged upon error */
-                       if (f->se_rule)
+                       if (f->se_rule) {
+                               if (need_sid) {
+                                       selinux_task_ctxid(tsk, &sid);
+                                       need_sid = 0;
+                               }
                                result = selinux_audit_rule_match(sid, f->type,
                                                                  f->op,
                                                                  f->se_rule,
                                                                  ctx);
+                       }
                        break;
                case AUDIT_ARG0:
                case AUDIT_ARG1: