On Tue, Aug 13, 2002 at 02:28:54PM +0900, itojun@iijlab.net wrote:

 > 	byte-swapping constant is good.  i've updated my tree.
 > 
 > > > @@ -764,3 +759,3 @@
 > > >  		 */
 > > > -		if (mff || ip->ip_off) {
 > > > +		if (mff || ntohs(ip->ip_off)) {
 > >Don't need to byte-swap if you're testing != 0.
 > 
 > 	for completeness i'd like to have it.

Then you could so:

			if (mff || ip->ip_off != htons(0)) {

which byte-swaps a constant, which the compiler will optimize away :-)

 > 
 > > > Index: netinet/raw_ip.c
 > > > ===================================================================
 > > > RCS file: /cvsroot/syssrc/sys/netinet/raw_ip.c,v
 > > > retrieving revision 1.61
 > > > diff -u -1 -r1.61 raw_ip.c
 > > > --- netinet/raw_ip.c	2002/06/09 16:33:43	1.61
 > > > +++ netinet/raw_ip.c	2002/08/12 22:57:50
 > > > @@ -165,5 +165,7 @@
 > > >  	 * XXX Compatibility: programs using raw IP expect ip_len
 > > > -	 * XXX to have the header length subtracted.
 > > > +	 * XXX to have the header length subtracted, and in host order.
 > > > +	 * XXX ip_off is also expected to be host order.
 > > >  	 */
 > > > -	ip->ip_len -= ip->ip_hl << 2;
 > > > +	ip->ip_len = ntohs(ip->ip_len) - (ip->ip_hl << 2);
 > > > +	NTOHS(ip->ip_off);
 > >You're converting from host-to-net, so use htons() and HTONS(), right?
 > 
 > 	no, this is in rip_input, so net-to-host.

Ah, sorry, my mistake (that's what I get for reviewing code right before
I go to bed :-)

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>