[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-bugs] #18762 [Applications/Tor Browser]: implement first-party isolation for OCSP generated by speculative connect



#18762: implement first-party isolation for OCSP generated by speculative connect
--------------------------------------+--------------------------
 Reporter:  arthuredelstein           |          Owner:  tbb-team
     Type:  defect                    |         Status:  new
 Priority:  Medium                    |      Milestone:
Component:  Applications/Tor Browser  |        Version:
 Severity:  Normal                    |     Resolution:
 Keywords:  tbb-linkability           |  Actual Points:
Parent ID:                            |         Points:
 Reviewer:                            |        Sponsor:
--------------------------------------+--------------------------

Comment (by arthuredelstein):

 Replying to [comment:3 gk]:
 > I was looking a bit closer at a thing which was nagging me while doing
 the review for #16998. There is
 > {{{
 >     // Check for proxy information. If there is a proxy configured then
 a
 >     // speculative connect should not be performed because the potential
 >     // reward is slim with tcp peers closely located to the browser.
 > }}}
 > and this piece of code in `nsIOService.cpp`:
 > {{{
 > NS_IMETHODIMP
 > IOServiceProxyCallback::OnProxyAvailable(nsICancelable *request,
 nsIChannel *channel,
 >                                          nsIProxyInfo *pi, nsresult
 status)
 > {
 >     // Checking proxy status for speculative connect
 >     nsAutoCString type;
 >     if (NS_SUCCEEDED(status) && pi &&
 >         NS_SUCCEEDED(pi->GetType(type)) &&
 >         !type.EqualsLiteral("direct")) {
 >         // proxies dont do speculative connect
 >         return NS_OK;
 >     }
 > }}}
 > And it seems to me we hit this code path with Tor Browser. Retesting
 #16324 by looking at `tcpdump` output confirms my suspicion as well: there
 is no network activity visible even if Torbutton claims isolation is
 happening.
 >
 > So, it seems to me that at least this ticket and #16324 can be closed. I
 am not sure yet what this means for #16998. I guess, we should not have
 been worried by it because there is no speculative connect happening
 anyway?

 I watch for STREAM events in the browser console and I can confirm that
 the speculative connects don't seem to be causing any network activity. So
 I agree that this ticket and #16324 can be closed. However, I did notice
 that under some special situations, a favicon is displayed in the
 searchbar popup which causes a connection over the catchall circuit; see
 #19741.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/18762#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs