fix for IPSEC-IPV4 breakage

[Andrew Atrens]

>>     I'm not sure if this is correct, but it's worth a shot.  I'm almost
>>     positive that the direct protocol dispatch is illegal.
> 
> It's interesting that the second code path is different. Looks like it's
> been that way since FBSD version 1.8 (taken directly from a KAME
> patchset sometime in 2001).  Incidentally version 1.7 looks very
> different.
> 
> Okay, I'll give the netisr_queue thing a try.

Okay, no panics but the packets are getting dropped. Maybe m_pullup() isn't
working properly in this case - it could be the common denominator?

Back to my day job for a while .. if I have time today I'll dig into the
m_pullup code ... and I guess we'll see what Jeff has to say .. :) :)

Andrew.