#6174: De-kludgify marking circs as unsuitable for new streams
------------------------+---------------------------------------------------
Reporter: rransom | Owner:
Type: defect | Status: new
Priority: normal | Milestone: Tor: 0.2.4.x-final
Component: Tor Client | Version:
Keywords: | Parent:
Points: | Actualpoints:
------------------------+---------------------------------------------------
Comment(by rransom):
Replying to [comment:2 nickm]:
> > Yes, it is, and wrong in several different ways too.
>
> Unless you're planning to do this yourself, you need to say what you
think are the several things that must be fixed for this bug to be closed.
The only things that need to be done for this bug to be closed are (a)
extract that hack into a separate function, and (b) add a separate
`do_not_attach_new_streams` `circuit_t` field (preferably a one-bit
bitfield; there should be room for more of those) to mark a circuit as not
suitable for new streams.
The current hack has at least the following problems:
* If a user increases the value of MaxCircuitDirtiness before the circuit
is closed, new streams could get attached to it.
* This hack could underflow `timestamp_dirty` (known bug, but not
obviously fixed in every occurrence).
* Until `newnym_epoch` was added with the stream-isolation code, if a
user clicked ‘New Identity’, then the system clock jumped backward 10
minutes, new streams could be put on old circuits.
There is no need to find or consider every problem with
`circ->timestamp_dirty -= options->MaxCircuitDirtiness` in order to avoid
all of those problems.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/6174#comment:3>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online