Hi tech@,
I tried to understand the output of "ifconfig bridge0 addr", especially
the "magic number" after the interface name. Apparently it indicates
the aging of the entry (the value is only ever 0 or 1). Eventually I
looked at the aging cycle code in bridge_rtage() and wondered why there
are 3 different cases (if/else-if/else). Finally I came up with below
diff to make the code easier to understand and shorter (no functional
change intended): Either the conditions are met to remove the entry, or
else the value of brt_age is set to 0.
The current code (with the 3 if-cases) was introduced in if_bridge.c in rev
1.6 - almost 20 years ago. In said rev the aging code was extended by a
3rd case to take into account static entries. However, the same thing
can be achieved with just 2 if-cases.
I tested this on i386 for a couple of weeks.
Regards
Holger
Index: bridgectl.c
===================================================================
RCS file: /cvs/src/sys/net/bridgectl.c,v
retrieving revision 1.13
diff -u -p -u -r1.13 bridgectl.c
--- bridgectl.c 12 Dec 2018 14:19:15 -0000 1.13
+++ bridgectl.c 25 Dec 2018 21:53:06 -0000
@@ -329,20 +329,17 @@ bridge_rtage(void *vsc)
for (i = 0; i < BRIDGE_RTABLE_SIZE; i++) {
n = LIST_FIRST(&sc->sc_rts[i]);
while (n != NULL) {
- if ((n->brt_flags & IFBAF_TYPEMASK) == IFBAF_STATIC) {
- n->brt_age = !n->brt_age;
- if (n->brt_age)
- n->brt_age = 0;
- n = LIST_NEXT(n, brt_next);
- } else if (n->brt_age) {
- n->brt_age = 0;
- n = LIST_NEXT(n, brt_next);
- } else {
+ if ((n->brt_flags & IFBAF_TYPEMASK) != IFBAF_STATIC
+ && n->brt_age == 0) {
+ // remove non-static aged entries
p = LIST_NEXT(n, brt_next);
LIST_REMOVE(n, brt_next);
sc->sc_brtcnt--;
free(n, M_DEVBUF, sizeof *n);
n = p;
+ } else {
+ n->brt_age = 0;
+ n = LIST_NEXT(n, brt_next);
}
}
}
