Hi all!

I'm sending an in-kernel mbuf chain - where the external cluster is
not allocated from mclpool, but an own pool - to udp_output
(*(so->so_proto->pr_usrreq)). If the total size of the chain is above the
mtu, the kernel panics in ip_output when defragmenting the chain
(m_copym). However, if I split the chain up in several packets ( < mtu) it
works fine.

(db trace included below)

Doen anyone know what the reason for this can be?

Thank you in advance
P=E5l Halvorsen

-------------------
Netbsd 1.4.2

uvm_fault(0xd7c2db10, 0x0, 0, 3) -> 1
kernel: page fault trap, code=3D0
Stopped in a.out at  _m_copym+0x254: movl %esi,0x34(%eax)

trace:
_m_compm(c0b77100,5dc,5c8,1,c0b41664) at _m_copym+0x254
_ip_output(c0b41600,0,c0b44879,0,0) at _ip_output+0x678
_udp_output(...)=20
_udp_usrreq(...)

=2E..