#32709: hsv3: Support onionbalance keys when handling INTRO2 cells -------------------------------------------------+------------------------- Reporter: asn | Owner: dgoulet Type: enhancement | Status: | needs_review Priority: Medium | Milestone: Tor: | 0.4.3.x-final Component: Core Tor/Tor | Version: Severity: Normal | Resolution: Keywords: tor-hs scaling onionbalance tor- | Actual Points: spec network-team-roadmap-2020Q1 043-must | postfreeze-ok | Parent ID: #26768 | Points: 2 Reviewer: nickm | Sponsor: | Sponsor27-must -------------------------------------------------+-------------------------
Comment (by nickm): Okay, reading through it all again. I'll make comments in the PR, but here are some higher-level questions. 1) Does it really make sense to compute the subcredential for three time periods? It seems to me that we are never close to more than two periods. Maybe we should calculate the subcredentials for "now" and "the closest period other than now". 2) The performance here is going to be needlessly bad. Keep in mind that every time we call hs_ntor_client_get_introduce1_keys(), we're doing a curve25519 calculation... but the curve25519 calculation will be the same here every time! The only input that changes is the subcredential, which is an input only to the XOF() part of the process. This could be a followup branch, I guess, though. 3) We should think about timing side channels here. This could also be a followup, however. 4) Test coverage on these changes should really be higher. -- Ticket URL: <https://trac.torproject.org/projects/tor/ticket/32709#comment:13> Tor Bug Tracker & Wiki <https://trac.torproject.org/> The Tor Project: anonymity online
_______________________________________________ tor-bugs mailing list tor-bugs@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs