Hi
I was studying the linux 2.6.1 networking code (via lxr.linux.no) and
stumbled on three minor points.
1. fib_props[] in net/ipv4/fib_semantics.c has size determined by RTA_MAX,
yet is indexed by RTN_* constants. Patch included.
2. Two typos ("xrfm") in comments in include/net/xfrm.h. Patch included.
3. I'm quite new to all this, so I might be missing something, but I
think a dependency is reversed. net/sched/Kconfig says:
config NET_CLS_ROUTE
bool
depends on NET_CLS_ROUTE4
default y
Yet evidence seems to indicate NET_CLS_ROUTE4 actually depends on
NET_CLS_ROUTE instead:
net/sched/Makefile:29:obj-$(CONFIG_NET_CLS_ROUTE4) += cls_route.o
net/sched/cls_route.c:153: id = dst->tclassid;
include/net/dst.h:72:#ifdef CONFIG_NET_CLS_ROUTE
include/net/dst.h:73: __u32 tclassid;
include/net/dst.h:74:#endif
Linux 2.4.22's net/sched/Config.in also seems to support this direction
of the dependency.
regards,
--
Matthijs van Duin -- May the Forth be with you!
--- linux-2.6.1/net/ipv4/fib_semantics.c 2004-01-09 07:59:10.000000000
+0100
+++ linux/net/ipv4/fib_semantics.c 2004-01-31 16:34:13.000000000 +0100
@@ -83,7 +83,7 @@
{
int error;
u8 scope;
-} fib_props[RTA_MAX + 1] = {
+} fib_props[RTN_MAX + 1] = {
{
.error = 0,
.scope = RT_SCOPE_NOWHERE,
--- linux-2.6.1/include/net/xfrm.h 2004-01-09 07:59:56.000000000 +0100
+++ linux/include/net/xfrm.h 2004-01-31 16:36:40.000000000 +0100
@@ -48,10 +48,10 @@
|---. child .-> dst -. xfrm .-> xfrm_state #3
|---. child .-> NULL
- Bundles are cached at xrfm_policy struct (field ->bundles).
+ Bundles are cached at xfrm_policy struct (field ->bundles).
- Resolution of xrfm_tmpl
+ Resolution of xfrm_tmpl
-----------------------
Template contains:
1. ->mode Mode: transport or tunnel
|