Hello,

I use iked with ipv6 endpoints.  The prepend mbuf allocation counter in 
netstat -m continues to rise.

len 40 makes me think it needs room for the ipv6 header, but I am not 
sure the proper place to reserve it...  any clues?

m_prepend: len 40
-+- mbuf 0xfffffd800bdce100, dat, off 16, len 556, pktlen 556, clsize 2112
  \\- total chain 1, len 556, size 2112
Starting stack trace...
m_print_debug(fffffd800bdce100,ffffffff81a86f0d,28) at m_print_debug+0x31d
m_prepend(fffffd800bdce100,28,2) at m_prepend+0xee
ipip_output(ffff80003155bc28,ffff80000093f420) at ipip_output+0x24d
ipsp_process_packet(fffffd800bdce100,ffff80000093f420,18,0,0) at 
ipsp_process_packet+0x5c4
sec_send(ffff80000040c000) at sec_send+0x190
taskq_thread(ffffffff81cb55e0) at taskq_thread+0x16c
end trace frame: 0x0, count: 251
End of stack trace.
m_prepend: len 40
-+- mbuf 0xfffffd8003456d00, dat, off 16, len 556, pktlen 556, clsize 2112
  \\- total chain 1, len 556, size 2112
Starting stack trace...
m_print_debug(fffffd8003456d00,ffffffff81a86f0d,28) at m_print_debug+0x31d
m_prepend(fffffd8003456d00,28,2) at m_prepend+0xee
ipip_output(ffff80003155bc28,ffff80000093f420) at ipip_output+0x24d
ipsp_process_packet(fffffd8003456d00,ffff80000093f420,18,0,0) at 
ipsp_process_packet+0x5c4
sec_send(ffff80000040c000) at sec_send+0x190
taskq_thread(ffffffff81cb55e0) at taskq_thread+0x16c
end trace frame: 0x0, count: 251
End of stack trace.
m_prepend: len 40
-+- mbuf 0xfffffd800a71e300, dat, off 16, len 556, pktlen 556, clsize 2112
  \\- total chain 1, len 556, size 2112
Starting stack trace...
m_print_debug(fffffd800a71e300,ffffffff81a86f0d,28) at m_print_debug+0x31d
m_prepend(fffffd800a71e300,28,2) at m_prepend+0xee
ipip_output(ffff80003155bc28,ffff80000093f420) at ipip_output+0x24d
ipsp_process_packet(fffffd800a71e300,ffff80000093f420,18,0,0) at 
ipsp_process_packet+0x5c4
sec_send(ffff80000040c000) at sec_send+0x190
taskq_thread(ffffffff81cb55e0) at taskq_thread+0x16c
end trace frame: 0x0, count: 251
End of stack trace.