> So far as I can tell "DF" can't be set directly by an application.

Not under NetBSD.  It is not for us to assume that all other APIs are
comparably crippled.

> What other reasons could packets have the DF bit set?

Atomicity of delivery of data to the receiving host is one that comes
to mind immediately.  But I am by no means convinced there are no
others, and I certainly believe it is not for NetBSD, as a router, to
try to intuit why the endpoint chose to set DF, and think it's
justified in ignoring it.

> If you've got a valid argument then state it -- don't just lean back
> on a dusty old pile of RFCs and say "Because they say so".

But they *do* say so.  Standards, especially ones proven by as much
time and real-world testing as IP, are not to be casually tossed aside
for something that hasn't even been tested under fire.

If you want to implement your proposal, see how it behaves, perhaps put
it into real use and see what (if anything) breaks...get an RFC into
the standards track, modifying the IP spec...*then* it may be time to
roll it into the NetBSD tree.  Most certainly not now.

Or at least such is my opinion.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B