natd crashing again with current/amd64

Since a build yesterday I am back to crashing.  Something was fixed
Friday or Saturday that cured it, but now it is failing somewhere else:

Oct 11 07:36:19 server kernel: Fatal trap 12: page fault while in kernel mode
Oct 11 07:36:19 server kernel: fault virtual address    = 0x18
Oct 11 07:36:19 server kernel: fault code               = supervisor read, page
not present
Oct 11 07:36:19 server kernel: instruction pointer      = 0x8:0xffffffff803528c0Oct 11 07:36:19 server kernel: stack pointer            = 0x10:0xffffffffb1eea5f0
Oct 11 07:36:19 server kernel: frame pointer            = 0x10:0xffffffffb1eea650
Oct 11 07:36:19 server kernel: code segment             = base 0x0, limit 0xfffff, type 0x1b
Oct 11 07:36:19 server kernel: = DPL 0, pres 1, long 1, def32 0, gran 1
Oct 11 07:36:19 server kernel: processor eflags = interrupt enabled, resume, IOPL = 0
Oct 11 07:36:19 server kernel: current process          = 249 (natd)
Oct 11 07:36:19 server kernel: trap number              = 12

I don't have a traceback, but I can tell you where the failure is:

0xffffffff803528c0 is in m_copypacket (/usr/src/sys/kern/uipc_mbuf.c:452).
447             if (!m_dup_pkthdr(n, m, how))
448                     goto nospace;
449             n->m_len = m->m_len;
450             if (m->m_flags & M_EXT) {
451                     n->m_data = m->m_data;
452                     n->m_ext = m->m_ext;
453                     n->m_flags |= M_EXT;
454                     MEXT_ADD_REF(m);
455             } else {
456                     n->m_data = n->m_pktdat + (m->m_data - m->m_pktdat );

I am rebuilding world now and if it persists I will attempt to get my
kernel to drop into debug.

Sean