Hi Nathan, Interesting. I reviewed RFC 8357 and you seem to be correct, as far as I can tell. Kea ARM says this is supported: https://kea.readthedocs.io/en/kea-3.0.0/arm/dhcp4-srv.html#supported-dhcp-standards but it does not appear to be working (in DHCPv4 anyway):
Used perfdhcp as you did: sudo perfdhcp -4 -r 1 -R 1 -p 2 -L 12345 -o 82,1300 10.1.2.2 2025-07-20 13:11:37.462 DEBUG [kea-dhcp4.packets/974.281473765296000] DHCP4_BUFFER_RECEIVED received buffer from 10.1.2.4:12345 to 10.1.2.2:67 over interface enp0s9 2025-07-20 13:11:37.462 DEBUG [kea-dhcp4.options/974.281473697640224] DHCP4_BUFFER_UNPACK parsing buffer received from 10.1.2.4 to 10.1.2.2 over interface enp0s9 2025-07-20 13:11:37.462 DEBUG [kea-dhcp4.bad-packets/974.281473697640224] DHCP4_PACKET_DROP_0001 failed to parse packet from 10.1.2.4 to 10.1.2.2, received over interface enp0s9, reason: OptionInt 19 truncated, hwaddr=00:0c:01:02:03:04 I did a packet capture (attached) and it seems to show a correctly specified option 82 sub-option 19 (Source Port) (at least Wireshark thinks it does). It appears to be supported in DHCPv6, but that may be because it isn't a 0 length option (the desired port is included in the option): sudo perfdhcp -6 -r 1 -R 1 -p 2 -L 12345 -A 1 --or 135,3039 2001:db8::2 msg_type=SOLICIT (1), trans_id=0x0, options: type=00001, len=00014: 00:01:00:01:30:0f:a9:85:00:0c:01:02:03:04 type=00003(IA_NA), len=00012: iaid=1, t1=3600, t2=5400 type=00006, len=00004: 23(uint16) 24(uint16) type=00008, len=00002: 0 (uint16) 1 relay(s): relay[0]: msg-type=12(RELAY_FORWARD), hop-count=0, link-address=2001:db8::4, peer-address=2001:db8::4, 1 option(s) type=00135, len=00002: 12345 (uint16) 2025-07-20 13:20:05.316 DEBUG [kea-dhcp6.packets/1040.281473129902432] DHCP6_RESPONSE_DATA duid=[00:01:00:01:30:0f:a9:85:00:0c:01:02:03:04], [no hwaddr info], tid=0x0: responding with packet ADVERTISE (type 2), packet details: local_address=[2001:db8::2]:547, remote_address=[2001:db8::4]:12345, Long story short, I've created this issue to report this problem: https://gitlab.isc.org/isc-projects/kea/-/issues/4017 Thank you, Darren Ankney On Sat, Jul 19, 2025 at 4:02 PM Nathan Faulkner <nathanfaulkne...@gmail.com> wrote: > > Unless I'm misreading the spec, the length for sub-option 19 (hex 13) should > be 0. > > On Sat, Jul 19, 2025, 3:54 PM Darren Ankney <darren.ank...@gmail.com> wrote: >> >> Hi Nathan, >> >> I think, In this case, "-o 82,1300" will look like sub-option 19 >> with 0 length and no data at all. This would make sense with 0 >> length, but Kea is expecting you to send some data, I think, which is >> probably why it complained. If you already have network activity with >> this option 82 content, you can shortcut this process by capturing >> packets as I said previously. I will give a hypothetical as an >> example, however. Let us say you want to send the circuit id: VLAN12 >> as option 82 sub option 1 (circuit-id). You would specify like so: >> >> -o 82,0106564C414E3132 >> >> where: >> >> 01 = sub option 1 >> 06 = 6 bytes length >> 564C414E3132 = VLAN12 >> >> This is exactly what would appear on the wire. >> >> Thank you, >> Darren Ankney >> >> On Sat, Jul 19, 2025 at 12:59 PM Nathan Faulkner >> <nathanfaulkne...@gmail.com> wrote: >> > >> > The registration page mentioned that some domains were blocked and this >> > list could be emailed to get one unblocked, though I don't know if that >> > included individual addresses on a domain. If not, I'm not sure how >> > someone would contribute without paying for an email address. >> > >> > For "-o 82,1300" the spec >> > (https://www.rfc-editor.org/rfc/rfc8357.html#section-4.1) says the length >> > is 0. When I tested some other values in place of 13 (don't remember >> > which, I think some random higher value and 01) they didn't seem to have >> > an issue with 0 length data. >> > >> > On Sat, Jul 19, 2025, 8:14 AM Darren Ankney <darren.ank...@gmail.com> >> > wrote: >> >> >> >> Hi Nathan, >> >> >> >> I think the reason you had trouble setting up an account in Gitlab was >> >> that you have an @gmail.com address. There has been some lock-down >> >> needed due to a large amount of spam. If you have a business email, >> >> you could probably use that. >> >> >> >> As far as: "-o 82,1300" not working, when sending option 82 data, you >> >> have to actually specify the entire suboption (id+length+data) as >> >> shown here: https://www.rfc-editor.org/rfc/rfc3046#section-2.0 It also >> >> must be converted to hex as would appear in a DHCP packet. perfdhcp >> >> won't do any of this for you. It will merely take what you provide, >> >> compute a length , and send: 0x52<length><your data> (at least I think >> >> that is correct). The easiest thing to do would be to capture DHCP >> >> traffic that already contains this option, open in wireshark, and copy >> >> out the hex for the sub-option as a hex stream. This should give you >> >> exactly what you need to use on the command line. For some tips >> >> (though they are instructions for templates) see here: >> >> https://kb.isc.org/docs/how-to-use-perfdhcp-benchmarking-tool >> >> >> >> Thank you, >> >> Darren Ankney >> >> >> >> On Thu, Jul 17, 2025 at 9:33 PM Nathan Faulkner >> >> <nathanfaulkne...@gmail.com> wrote: >> >> > >> >> > I was trying to create an account to report a couple of kea issues, but >> >> > I got an "Email is not allowed for sign-up." message. >> >> > >> >> > I'm running kea 2.6.2. >> >> > >> >> > Here are the issues I was planning to open: >> >> > >> >> > perfdhcp should specify the relay agent relay port sub option when a >> >> > local port is specified. Not sure how it'd work otherwise. >> >> > https://gitlab.isc.org/isc-projects/kea/-/issues/663 looks to be a >> >> > related issue. >> >> > When sending the relay agent relay port sub option (by passing "-o >> >> > 82,1300" to perfdhcp), the debug kea logs show the packet failing to >> >> > parse for reason "OptionInt 19 truncated". (as a workaround, you can >> >> > specify junk data "-o 82,13020000") >> >> > It'd be nice if perfdhcp could specify multiple servers. I was trying >> >> > to test HA, and normally I'd use a relay server, but perfdhcp itself >> >> > pretends to be a relay server, which causes conflicts. >> >> > >> >> > -- >> >> > ISC funds the development of this software with paid support >> >> > subscriptions. Contact us at https://www.isc.org/contact/ for more >> >> > information. >> >> > >> >> > To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users. >> >> > >> >> > Kea-users mailing list >> >> > Kea-users@lists.isc.org >> >> > https://lists.isc.org/mailman/listinfo/kea-users >> >> -- >> >> ISC funds the development of this software with paid support >> >> subscriptions. Contact us at https://www.isc.org/contact/ for more >> >> information. >> >> >> >> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users. >> >> >> >> Kea-users mailing list >> >> Kea-users@lists.isc.org >> >> https://lists.isc.org/mailman/listinfo/kea-users >> > >> > -- >> > ISC funds the development of this software with paid support >> > subscriptions. Contact us at https://www.isc.org/contact/ for more >> > information. >> > >> > To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users. >> > >> > Kea-users mailing list >> > Kea-users@lists.isc.org >> > https://lists.isc.org/mailman/listinfo/kea-users >> -- >> ISC funds the development of this software with paid support subscriptions. >> Contact us at https://www.isc.org/contact/ for more information. >> >> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users. >> >> Kea-users mailing list >> Kea-users@lists.isc.org >> https://lists.isc.org/mailman/listinfo/kea-users > > -- > ISC funds the development of this software with paid support subscriptions. > Contact us at https://www.isc.org/contact/ for more information. > > To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users. > > Kea-users mailing list > Kea-users@lists.isc.org > https://lists.isc.org/mailman/listinfo/kea-users
kea-dhcp4-RFC8357.pcap
Description: Binary data
-- ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information. To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users. Kea-users mailing list Kea-users@lists.isc.org https://lists.isc.org/mailman/listinfo/kea-users
