IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#ls2
/2022/06/20
orignal SSU2: PeerTest msg=4 code=70
orignal I have receive error code 70 that's basically Alice unknown
orignal but Alice was me and I do have SSU2 address
orignal any issue with RI block?
zzz <zzz> I was working on that over the weekend.
zzz <zzz> basically I couldn't handle RI arriving after peer test
zzz <zzz> fixed yesterday, I doubt anybody has updated yet
zzz <zzz> it would also help if you put the ri block before the peer test block, or sent it first
zzz otherwise I have to set a timer and wait for it
zzz no big deal for peer test where latency doesn't matter, but important for relay
orignal I didn't receive it
orignal probably was disconnected
orignal what? I always put RI block before peer test
zzz ok, great
zzz I didn't know
orignal that's why I'm confused about 70
orignal previously you said you couldn't find intro key in Alice's RI
zzz that's a different error code I think
zzz maybe RI block was in different packet and got lost
orignal I don't know how Java sends it
orignal it Bob was i2pd it would be in the same packet
orignal my RI was short
zzz doesn't always fit of course
zzz always separate packet right now for java
orignal that might be the reason
orignal and as I said sometimes I receive peer test message
orignal basically msg 5
orignal but wrong conn id
zzz and things get reordered in my queues
orignal let's double check how we make it
zzz also I had the conn ids wrong on hole punch, fixed a few days ago
orignal <orignal> zzz, if nonce is 1 should it look like 01 00 00 00 01 00 00 00 in a packet?
zzz I think I have it right on msg 5?
orignal please conform it
orignal I'm not sure who has it wrong
zzz no, big endian
zzz not sure if it says that in the spec
orignal so, 00 00 00 01 00 00 00 01
zzz correct
orignal it says but it's not clear what we make BE
orignal nonce or connid
zzz either way works ))
orignal not sure
orignal probably
orignal anyway let me print out and see
zzz ok I think I just finished my MTU fixes
zzz first fragment can be 5 bytes bigger than followon fragments, I wasn't doing that
orignal I haven't start it yet
zzz and I wasn't counting the 5 bytes overhead in followon fragments so I was exceeding the MTU
zzz I had one error, ever, of trying to send a 1501 byte packet, but that was enough to chase it down
orignal my mistake it was RelayRequest
orignal can it be sent without session?
orignal never mind it was Peer Test
orignal mixed with block type
zzz 0) Hi
zzz idk and zab are on the road
orignal as usual
zzz I got the slides from the student presentation
zzz I think they looked at i2pd
zzz links on my forum
orignal will check
zzz these students from switzerland, sponsored by diva.exchange, presented at monerocon on saturday
zzz I gave them some advice a month or two ago, but I thought they were looking at java i2p. guess they changed their minds
zzz anyway, what's on the agenda?
zzz ok that's 1)
orignal I don't have anything else for now
zzz I don't really have much else
zzz 1) SSU2
zzz I'll go first... just fixing bugs
zzz nothing big to report
zzz I'm getting maybe 5-10 simultaneous connections now
orignal the same
orignal keep investigating that nonce issue
orignal brb in a minute
zzz so it's getting easier to test
zzz I'm announcing a new dev version about once a week on my forum
zzz will probably do that tomorrow
zzz still making minor updates and fixes to the spec
zzz that's about it, EOT
zzz what's your status?
orignal many things work
orignal peer test
orignal as I said investigating nonce for msg 5
orignal also working on replacemnt SSU by SSU2 if configured
orignal e.g. SSU2 only no SSU1
zzz I logged ack blocks and ranges for a while. what I'm getting from i2pd looks good
zzz here's an example of one that seems to be on a bad connection:
zzz Got new ACK block from bAU~6X ACK 43-18 NACK 17 ACK 16-13 NACK 12 ACK 11 NACK 10-9
zzz ACK 8-4 NACK 3 ACK 2-0
orignal very good
orignal seems I have fixed it finally
zzz also I verified you fixed the expiration on the new token block
orignal what threshold do you use for expiration time?
orignal e.g. if you have token that's about to expire
zzz I don't have any
zzz maybe I should
zzz any other status or issues?
orignal I use 2 seconds for now
orignal no other issues
orignal moving forward well
zzz yes
zzz anything else on 1) ?
zzz anything else for the meeting ?
orignal when idk is back?
zzz not sure, he might be flying back today
orignal so next meeting
zzz for sure
orignal please confirm that (nonce << 32) | nonce is incoming connid for Alice
orignal e.g. Charlie puts it to Destination Conn ID in PeerTest messages
orignal because I receive them inverted
orignal SSU2: PeerTest nonce=429696389 connid=9630275011632012313
orignal SSU2: Unexpected PeerTest message SourceConnID=8816469062077539302 DestConnID= 9630275011632012313
orignal but this is for Alice side
zzz looking...
orignal connid is session->m_SourceConnID = htobe64 (((uint64_t)nonce << 32) | nonce);
orignal so I palce this value into sourceconnid for Alice
orignal but you use it as SourceConnID for Charlie
zzz Connection IDs: The two connection IDs are derived from the test nonce. For messages sent from Charlie to Alice, the Destination Connection ID is two copies of the 4-byte big-endian test nonce, i.e. ((nonce << 32) | nonce). The Source Connection ID is the inverse of the Destination Connection ID, i.e. ~((nonce << 32) | nonce). For messages sent from Alice to Charlie, swap the two connection IDs.
orignal yes, from the specs
orignal but you reverse it
zzz do I
zzz for 5 and 7, dest conn ID is nonce<<32 | nonce
orignal but see what I receive
zzz ok, now time to look at my code :)
orignal you put this nonce<<32 | nonce into SourceConnID
orignal Alice places in into SourceConnID
orignal and expect message from Charlie in DestConnID e.g. in first 8 bytes
orignal it's clear that endianess is right
zzz long rcvId = (nonce << 32) | nonce;
zzz long sendId = ~rcvId;
zzz you're right, I'm wrong
orignal will you fix it?
zzz of course
zzz good catch
orignal thanks
zzz will double-check the others too
orignal spend last 2 days ))
zzz 6 and 7 and hole punch
zzz sorry :(
orignal thought it's endianess
orignal no problem
orignal at least we do receive msg 5
orignal and it looks right
zzz 6 and 7 were wrong also
orignal I haven't reached this state anyway
zzz hole punch was right
orignal I will check also