Re: Minutes of MTU Discovery Working Group Meeting (7 Feb 1990)

[Steve Deering <[e-mail]>]

Jeff,

Oops!  You're right about 4.3BSD.  It does appear to do the right thing.

(However, I believe the line of code that you quoted only applies to
non-initial fragments.  The code that updates the initial fragment,
which is the one of concern for MTU discovery, appears later as follows:

	ip->ip_off = htons((u_short)(ip->ip_off | IP_MF));

My mistake was to assume that ip->ip_off on the RHS contained only the
offset value, rather than the offset and the high-order bits.)

Steve