PERFORCE change 81408 for review

[Robert Watson]

http://perforce.freebsd.org/chv.cgi?CH=81408

Change 81408 by rwatson at rwatson_zoo on 2005/08/03 16:52:45

	Remove stale comment about the caller needing to acquire the
	if_addr_mtx when walking address lists -- the in_multi macros
	now do this themselves.
	
	Remove stale comment to check whether we need to hold
	in_multi_mtx over the ip_mforward() path -- we don't, it would
	be a bad idea due to recursion.

Affected files ...

.. //depot/projects/netsmp/src/sys/netinet/in_var.h#7 edit
.. //depot/projects/netsmp/src/sys/netinet/ip_output.c#3 edit

Differences ...

==== //depot/projects/netsmp/src/sys/netinet/in_var.h#7 (text+ko) ====

@@ -188,8 +188,6 @@
 /*
  * Macro for looking up the in_multi record for a given IP multicast address
  * on a given interface.  If no matching record is found, "inm" is set null.
- *
- * Caller must hold IF_ADDR_LOCK().
  */
 #define IN_LOOKUP_MULTI(addr, ifp, inm) \
 	/* struct in_addr addr; */ \

==== //depot/projects/netsmp/src/sys/netinet/ip_output.c#3 (text+ko) ====

@@ -291,10 +291,6 @@
 				ip->ip_src = IA_SIN(ia)->sin_addr;
 		}
 
-		/*
-		 * XXXRW: Should the in_multi_mtx be held over
-		 * ip_mloopback() or ip_mforward()?
-		 */
 		IN_MULTI_LOCK();
 		IN_LOOKUP_MULTI(ip->ip_dst, ifp, inm);
 		if (inm != NULL &&