Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+-
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  closed
 Priority:  High |  Milestone:  Tor:
 |  0.2.9.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:  fixed
 Keywords:  prop224, tor-hs, 029-backport-   |  Actual Points:
  maybe, 031-unreached-backport  |
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+-
Changes (by teor):

 * status:  merge_ready => closed
 * resolution:   => fixed


Comment:

 #27073, #28096, #25773, #23681, and #23512 were merged to 0.2.9 and later
 after testing the new maint-0.2.9, maint-0.3.4, and master on CI.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+-
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:
 |  merge_ready
 Priority:  High |  Milestone:  Tor:
 |  0.2.9.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs, 029-backport-   |  Actual Points:
  maybe, 031-unreached-backport  |
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+-

Comment (by teor):

 This is now:
 https://github.com/torproject/tor/pull/765

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+-
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:
 |  merge_ready
 Priority:  High |  Milestone:  Tor:
 |  0.2.9.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs, 029-backport-   |  Actual Points:
  maybe, 031-unreached-backport  |
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+-

Comment (by dgoulet):

 Yeah I think backporting to 029 is fine.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+-
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:
 |  merge_ready
 Priority:  High |  Milestone:  Tor:
 |  0.2.9.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs, 029-backport-   |  Actual Points:
  maybe, 031-unreached-backport  |
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+-

Comment (by teor):

 This seems like a small, low-risk change to backport.
 It's been tested in 0.3.2 and later.

 dgoulet, let me know if you think we should backport it to 0.2.9.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+-
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:
 |  merge_ready
 Priority:  High |  Milestone:  Tor:
 |  0.2.9.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs, 031-backport, 029   |  Actual Points:
  -backport-maybe|
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+-
Changes (by nickm):

 * milestone:  Tor: 0.3.1.x-final => Tor: 0.2.9.x-final


--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+-
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:
 |  merge_ready
 Priority:  High |  Milestone:  Tor:
 |  0.3.1.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs, 031-backport, 029   |  Actual Points:
  -backport-maybe|
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+-
Changes (by teor):

 * keywords:  prop224, tor-hs, 031-backport => prop224, tor-hs,
 031-backport, 029-backport-maybe


--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
---+---
 Reporter:  asn|  Owner:  dgoulet
 Type:  defect | Status:  merge_ready
 Priority:  High   |  Milestone:  Tor:
   |  0.3.1.x-final
Component:  Core Tor/Tor   |Version:
 Severity:  Normal | Resolution:
 Keywords:  prop224, tor-hs, 031-backport  |  Actual Points:
Parent ID: | Points:
 Reviewer:  mikeperry  |Sponsor:
---+---

Comment (by nickm):

 3 months have passed -- how do you feel about a backport now? :)

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+-
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:
 |  merge_ready
 Priority:  High |  Milestone:  Tor:
 |  0.3.1.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs, 031-backport,   |  Actual Points:
  030-backport   |
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+-
Changes (by nickm):

 * keywords:  prop224, tor-hs => prop224, tor-hs, 031-backport, 030-backport
 * milestone:  Tor: 0.3.2.x-final => Tor: 0.3.1.x-final


Comment:

 Okay.  So, I've rebased it onto 0.2.9 (since if I understand you right,
 that's as far as we might ''ever'' want to backport) and squashed it as
 `bug23681_029_01_squashed`.  Only taking it in 0.3.2 and later for now
 though.  (Merging to 0.3.2 and forward!)

 (Git note: From a git convenience POV, it's easiest to have a branch
 that's based on the earliest convenient* version that we think we might
 want to backport onto.  That way, I can merge the branch to 0.3.2 now, and
 then later either merge the branch to 0.3.1 or 0.2.9 or not.)

 ([*] Of course, this isn't a good idea when the code has changed a lot
 between branches.  If there are huge conflicts, it makes sense to do
 separate branches.)

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  merge_ready
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+

Comment (by dgoulet):

 Replying to [comment:15 nickm]:
 > Should we consider a backport for this, or  is it too minor and/or too
 risky?

 I'm a bit puzzled about a backport. It has some inconvenience in terms of
 usability but doesn't break the HS feature per-se. I think I would be more
 comfortable keeping that in 032 for now so at least v3 is fixed.

 I wouldn't object going up to 030 with it but definitely not in 029 for
 now.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  merge_ready
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+

Comment (by nickm):

 Should we consider a backport for this, or  is it too minor and/or too
 risky?

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  merge_ready
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+
Changes (by dgoulet):

 * status:  needs_information => merge_ready


Comment:

 So the circuit is *not* closed but it is re-used at the next SOCKS
 connection. I was mistaken, the magic of `circuit_is_acceptable()` returns
 the opened INTRODUCING circuit. So at least it is re-used.

 And I confirm with `CrcuitsAvailableTimeout 10` and `SocksTimeout 5`, the
 circuit gets closed short after the socks timeout.

 Going in `merge_ready` since it confirms what Mike was asking about.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_information
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+
Changes (by dgoulet):

 * status:  needs_review => needs_information


Comment:

 Replying to [comment:12 mikeperry]:
 > Ugh, the other wrinkle to check is if the hidden service comes back
 online after the socks timeout, but while that old intro circuit is still
 laying around. Does Tor try to reuse the intro in that case? Does that
 work, or does it break?

 First, no the `SocksTimeout` does *not* make the intro point circuit to
 close.

 I haven't tested it but I'm pretty sure the client won't find that circuit
 because it is set to `INTRODUCING` purpose where the SOCKS request makes
 the client look for an `INTRODUCE_ACK_WAIT` and if can't find any launches
 a new circuit and sets it to `INTRODUCING`.

 I need to confirm that but I do think we'll lose track of this circuit
 until its idle timeout of 52 minutes hehe.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_review
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+

Comment (by mikeperry):

 Ugh, the other wrinkle to check is if the hidden service comes back online
 after the socks timeout, but while that old intro circuit is still laying
 around. Does Tor try to reuse the intro in that case? Does that work, or
 does it break?

 Sorry this is so annoying to verify.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_review
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+

Comment (by mikeperry):

 The socks timeout is not going to close the circuit. Did it eventually
 close? If you have a test case, you can lower that idle timeout by setting
 CircuitsAvailableTimeout to like 3 (3 minutes - long enough for the socks
 timeout).

 If the circuit does close, then this is merge ready IMO.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_review
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+
Changes (by dgoulet):

 * status:  needs_revision => needs_review


Comment:

 The `SocksTimeout` of 120 seconds kicked in much faster than the idle
 circuit timeout:

 {{{
 [...]
 Nov 16 09:17:03.886 [info] connection_ap_handshake_attach_circuit(): Intro
 2980528860 (id: 6) and rend circuit 0 (id: 0) circuits are not both ready.
 Stalling conn. (119 sec old)
 Nov 16 09:17:04.886 [notice] Tried for 120 seconds to get a connection to
 [scrubbed]. Giving up. (waiting for circuit)
 }}}

 Intro circuit (id = 6), choose this quite huge timeout of 52 minutes:

 {{{
 Nov 16 09:14:21.888 [info] origin_circuit_new(): Circuit 6 chose an idle
 timeout of 3140 based on 3026 seconds of predictive building remaining.
 }}}

 Seems "normal" but also quite long.

 Anyway, here is the updated branch: `bug23681_032_01`

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_revision
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+
Changes (by mikeperry):

 * status:  needs_review => needs_revision


Comment:

 Code-wise this looks good. These intro circs still should be given longer
 than normal since they have 4 hops, but there is a case for that.

 The changes file has a little wonky phrasing though. Instead of "The
 introduction circuit was timeout very fast while waiting" I would say "The
 introduction circuit was being timed out too quickly while waiting".

 Other than that, this all looks good. Did you verify the idle timeout
 applies to intro circs? Maybe try to connect to a hidden service that has
 rend circuit building commented out so it rends don't complete?

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_review
 Priority:  High |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:  mikeperry|Sponsor:
-+
Changes (by dgoulet):

 * priority:  Medium => High
 * reviewer:   => mikeperry
 * status:  needs_revision => needs_review


Comment:

 Branch: `bug23681_032_01`

 Putting Mike has a reviewer since we discussed this and it can be very
 tricky.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_revision
 Priority:  Medium   |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:   |Sponsor:
-+
Changes (by dgoulet):

 * status:  needs_information => needs_revision


Comment:

 After a discussion on IRC with mikeperry, we'll go with making
 `circuit_expire_building()` to not consider `C_INTRODUCING` with the build
 timeout because at that point the circuit is OPENED but the cell hasn't
 been sent yet because we could be waiting for the RP circuit to complete.
 Thus, we should not even try to consider it in that function and let the
 idle circuit timeout kick in if it takes too long.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  dgoulet
 Type:  defect   | Status:  needs_information
 Priority:  Medium   |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:   |Sponsor:
-+
Changes (by dgoulet):

 * cc: mikeperry (added)
 * status:  accepted => needs_information


Comment:

 (CCing Mike here to have an opinion from him)

 In a nutshell, the client is waiting on the RP circuit to established and
 without that, the INTRO1 cell is not sent. In the meantime, the client
 also launches an intro circuit (for performance) so when we get the
 RENDEZVOUS_ESTABLISHED cell, we immediately send the INTRO1 cell on the
 open circuit.

 However, imagine the RP circuit takes a while to build (or maybe we fail
 to handle a state #3733 ?), the client will expire the intro circuit and
 rebuild one while waiting for the RP. And seems that this dance went for a
 while in asn's case.

 I really wonder how to fix this. What would be best, just make sure we
 expire the RP circuit in a reasonable amount of time or make an intro
 circuit expire only if no RP circuit can be found? (All this client side
 ofc).

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  (none)
 Type:  defect   | Status:  needs_information
 Priority:  Medium   |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:   |Sponsor:
-+
Changes (by asn):

 * status:  new => needs_information


--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  (none)
 Type:  defect   | Status:  new
 Priority:  Medium   |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:   |Sponsor:
-+

Comment (by dgoulet):

 Oh ok... so the the RP circuit got stuck for a while waiting for the
 `RENDEZVOUS_ESTABLISHED` which stalls the client to send the INTRO1 cell
 and it took so long that many intro point were opened and closed while
 waiting on that RP...

 See the RP circuit `2943321454` ... I can't tell when that was opened nor
 if a establish rdv has been sent!...

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  (none)
 Type:  defect   | Status:  new
 Priority:  Medium   |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:   |Sponsor:
-+

Comment (by asn):

 Attaching info logs. See timestamps above and grep for [notice] to detect
 the right place!

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  (none)
 Type:  defect   | Status:  new
 Priority:  Medium   |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:   |Sponsor:
-+
Changes (by asn):

 * Attachment "weird_ips.log.gz" added.


--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

Re: [tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
-+
 Reporter:  asn  |  Owner:  (none)
 Type:  defect   | Status:  new
 Priority:  Medium   |  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor |Version:
 Severity:  Normal   | Resolution:
 Keywords:  prop224, tor-hs  |  Actual Points:
Parent ID:   | Points:
 Reviewer:   |Sponsor:
-+

Comment (by dgoulet):

 Yah, the expire time of building HS circuits are different timings from
 the rest.

 {{{
   /* Intro circs have an extra round trip (and are also 4 hops long) */
   SET_CUTOFF(c_intro_cutoff, get_circuit_build_timeout_ms() * (14/6.0) +
 1000);
 }}}

 Basically, the build timeout of what I've observed on a "normal service
 connection" is usually around 2s. This line in your logs will tell you the
 latest value:

 {{{
 [info] circuit_build_times_set_timeout(): Set circuit build timeout to 2s
 (1500.00ms, 6.00ms, Xm: 404, a: 2.882408, r: 0.031000) based
 on 1000 circuit times
 }}}

 Thus taking that into consideration, the cutoff should be 5666msec or
 ~5.6sec but in your case it seems the build circuit timeout was way lower
 that is around 1sec or even below... Those were introduced in commit
 `3458d904f62b2d97` so Mike is probably the person to talk to about this.
 There is this #3733 ticket that touches this expire building function that
 I need to sit down with Mike and go over all those timeouts and try to
 make sure they work well, the conditions in that function are good for
 both v2 and v3.

 If you client becomes slower and slower at building circuit, the timeout
 should change gradually but if the first 1000 circuits were super fast and
 then you move to the airport shitty Wifi for instance, ramping up that
 timeout value will take your client a lot of failures...

 But all in all, this shouldn't break the state of the client? It seems
 your client never got the ACK from the introduction point so it is just
 looping on trying to get it with new circuits... :S

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs

[tor-bugs] #23681 [Core Tor/Tor]: prop224: Clients mark intro circs as timed-out within seconds

[Tor Bug Tracker & Wiki]

#23681: prop224: Clients mark intro circs as timed-out within seconds
--+
 Reporter:  asn   |  Owner:  (none)
 Type:  defect| Status:  new
 Priority:  Medium|  Milestone:  Tor: 0.3.2.x-final
Component:  Core Tor/Tor  |Version:
 Severity:  Normal|   Keywords:  prop224, tor-hs
Actual Points:|  Parent ID:
   Points:|   Reviewer:
  Sponsor:|
--+
 I noticed that my prop224 client had some reconnects and I did some log
 digging. The original disconnect was caused by a truncate cell (probably
 natural causes) but then a whole dance of introduction/rendezvous started
 which ended up marking 3 intro circs as timed out within seconds:
 {{{
 
 Sep 28 03:55:26.000 [notice] Introduction circuit 2276964442 has opened.
 Attaching streams.
 Sep 28 03:55:26.000 [info] connection_ap_handshake_attach_circuit(): Intro
 (2276964442) and rend (2943321454) circs are not both ready. Stalling
 conn. (2 sec old)
 Sep 28 03:55:29.000 [info] circuit_expire_building(): Marking circ
 2276964442 (state 4:open, purpose 6) as timed-out HS circ
 
 Sep 28 03:55:29.000 [info] circuit_get_best(): There is an intro circuit
 being created right now, but it has already taken quite a while. Starting
 one in parallel.
 Sep 28 03:55:29.000 [notice] Introduction circuit 4126528467 has opened.
 Attaching streams.
 Sep 28 03:55:29.000 [info] connection_ap_handshake_attach_circuit(): Intro
 (4126528467) and rend (2943321454) circs are not both ready. Stalling
 conn. (5 sec old)
 Sep 28 03:55:34.000 [info] circuit_expire_building(): Marking circ
 4126528467 (state 4:open, purpose 6) as timed-out HS circ
 
 Sep 28 03:55:29.000 [info] circuit_get_best(): There is an intro circuit
 being created right now, but it has already taken quite a while. Starting
 one in parallel.
 Sep 28 03:55:34.000 [notice] Introduction circuit 428281 has opened.
 Attaching streams.
 Sep 28 03:55:34.000 [info] connection_ap_handshake_attach_circuit(): Intro
 (428281) and rend (2943321454) circs are not both ready. Stalling
 conn. (10 sec old)
 Sep 28 03:55:39.000 [info] circuit_expire_building(): Marking circ
 428281 (state 4:open, purpose 6) as timed-out HS circ
 
 Sep 28 03:57:26.000 [info] circuit_expire_building(): Abandoning circ 232
 137.205.124.35:1720:2276964442 (state 1,4:open, purpose 6, len 4)
 Sep 28 03:57:26.000 [info] circuit_mark_for_close_(): Circuit 2276964442
 (id: 232) marked for close at src/or/circuituse.c:820 (orig reason: 10,
 new reason: 0)
 Sep 28 03:57:30.000 [info] circuit_expire_building(): Abandoning circ 231
 137.205.124.35:1720:4126528467 (state 1,4:open, purpose 6, len 4)
 Sep 28 03:57:30.000 [info] circuit_mark_for_close_(): Circuit 4126528467
 (id: 231) marked for close at src/or/circuituse.c:820 (orig reason: 10,
 new reason: 0)
 Sep 28 03:57:35.000 [info] circuit_expire_building(): Abandoning circ 233
 137.205.124.35:1720:428281 (state 1,4:open, purpose 6, len 4)
 Sep 28 03:57:35.000 [info] circuit_mark_for_close_(): Circuit 428281
 (id: 233) marked for close at src/or/circuituse.c:820 (orig reason: 10,
 new reason: 0)
 }}}

 You can see that we went ahead and marked intro circs as timed-out within
 3 seconds of launching them which makes no sense, and caused a whole lot
 of mess in the prop224 state machine.

 We should try to figure out if there is a bug here.

--
Ticket URL: 
Tor Bug Tracker & Wiki 
The Tor Project: anonymity online
___
tor-bugs mailing list
tor-bugs@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs