On Tue, 2006-08-01 at 11:22 -0400, Alexander Viro wrote:
Killing audit overhead in case when no rules are loaded. More
detailed
log (this is a composite of patch series from audit git tree, see audit.b23..
lspp.b23 in there for individual changesets):
[PATCH] introduce audit rules counter
[PATCH] mark context of syscall entered with no rules as dummy
[PATCH] don't bother with aux entires for dummy context
[PATCH] take filling ->pid, etc. out of audit_get_context()
move that stuff downstream and into the only branch where it'll be
used.
diff --git a/fs/namei.c b/fs/namei.c
index 0ab26cb..55a1312 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -159,7 +159,7 @@ char * getname(const char __user * filen
#ifdef CONFIG_AUDITSYSCALL
void putname(const char *name)
{
- if (unlikely(current->audit_context))
+ if (unlikely(!audit_dummy_context()))
audit_putname(name);
else
__putname(name);
[...]
+extern int audit_n_rules;
#else
#define audit_alloc(t) ({ 0; })
#define audit_free(t) do { ; } while (0)
#define audit_syscall_entry(ta,a,b,c,d,e) do { ; } while (0)
#define audit_syscall_exit(f,r) do { ; } while (0)
+#define audit_dummy_context() 0
#define audit_getname(n) do { ; } while (0)
#define audit_putname(n) do { ; } while (0)
#define __audit_inode(n,i) do { ; } while (0)
This should be:
+#define audit_dummy_context() 1
...no?
--
James Antill - <james.antill(a)redhat.com>
setsockopt(fd, IPPROTO_TCP, TCP_CONGESTION, ...);
setsockopt(fd, IPPROTO_TCP, TCP_DEFER_ACCEPT, ...);
setsockopt(fd, SOL_SOCKET, SO_ATTACH_FILTER, ...);