This ticket is part of sponsor F deliverable 7. Once we have a recommended approach for an external program that discovers bridge addresses to tell Tor about them, we need to write proposals and start implementing them. The "and start implementation as needed" part of the deliverable text is vague enough to focus on the design discussion and proposals. If there's time left to implement something, great, but if not, that's fine, too. We can still implement proposals between March and November 2012.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information
Child items
0
Show closed items
No child items are currently assigned. Use child items to break down this issue into smaller parts.
Linked items
0
Link issues together to show that they're related.
Learn more.
asn said he said nickm was concerned about non-controllers having control port access.
nickm: Is that true? If so, should we write a third proposal that describes a low-privilege control port mode? Perhaps one that can only use POSTMESSAGE?
asn said he said nickm was concerned about non-controllers having control port access.
nickm: Is that true? If so, should we write a third proposal that describes a low-privilege control port mode? Perhaps one that can only use POSTMESSAGE?
The underlying assumption here being that low-priv controllers would POSTMESSAGE to Vidalia, who would then decide to approve or deny their requests, perhaps after asking for user input.
For posterity (and since nick might miss IRC scrollback):
19:18 < nickm> mikeperry: So, I wouldn't be totally opposed in all conceivable cases to having nonprivileged control port mode, but it makes me super-nervous.
19:19 < nickm> It's hard for me to reason abou the security properties of the POSTMESSAGE thing, since it's kind of pushing the semantics of the operation out to controllers, and I don't know what security properties they would rely on
19:22 < nickm> mikeperry: I need to read your proposal and think about it. If you want to argue for an unprivileged option, I'd like to see a threat model for that in the proposal.
21:28 < mikeperry> nickm: maybe we're putting the cart before the horse. I am not saying we need a low-priv control port. I am just trying to minimize all the ports and IPC channels we are planning to create with all of this pluggable transport and bridge discovery stuff
21:29 < mikeperry> so it is a bit early for a threat model, I think
21:33 < mikeperry> the question to answer before that is "could we reduce the number of extra ports with a better control port protocol"
21:34 < mikeperry> if the answer is "no", or "not without a lot more work", then we can forget about the low-priv mode
21:35 < mikeperry> but if keeping everything in the control port is a more robust design, and/or less rickety, and/or less work, then maybe we should think about how we could do it