783178 - socks_remote_dns=true leaks DNS (in some conditions)

Status: UNCONFIRMED

(Core :: Networking: Proxy, defect, P3)

Description

[lzdn]

User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:14.0) Gecko/20100101 Firefox/14.0.1
Build ID: 20120713225548

Steps to reproduce:

Firefox 14.0.1 x64 Linux
I was testing socks5 proxy + DNS resolving with wireshark.

Test method:
0. Change proxy, exit firefox and launch firefox again;
1. Open wireshark and start capture on eth adapter;
2. Enter 'dns' in filter bar;
3. In firefox, navigate to any url, repeat several times;
4. Stop capture and check the result.
   There should be not dns packet for previous visited urls.

Settings:
Only one socks5 proxy is set.
network.proxy.no_proxies_on: doesn't include 127.0.0.1, or the remote socks5 server
network.proxy.socks_remote_dns=true
network.prefetch-next=false


Actual results:

With proxy=127.0.0.1:port, DNS always leaked.
With proxy=remote-domain:port, DNS leakage was not observed.


Expected results:

socks_remote_dns=true doesn't leak DNS.

Comment 1

[Josh Matthews [:jdm]]

This sounds like the same problem identified in bug 751465.

Comment 2

[Firefox Bug Husbandry Bot]

Bulk change to priority: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258

Comment 3

[Firefox Bug Husbandry Bot]

Bulk change to priority: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258

Comment 4

[BugBot [:suhaib / :marco/ :calixte]]

In the process of migrating remaining bugs to the new severity system, the severity for this bug cannot be automatically determined. Please retriage this bug using the new severity system.

Comment 5

[BMO Automation]

Moving bug to Core/Networking: Proxy.

Comment 6

[Brian]

When network.trr.custom_uri & network.trr.uri are set, socks_remote_dns set to true ignores the EDNS client, and uses the system DNS instead of proxying and querying with the EDNS URI. Imo, this is unexpected behavior and causes leaks. When set to false, it proxies and uses the EDNS set in the browser, as expected.

Comment 7

[Manuel Bucher [:manuel]]

I don't think this was what the original report was about due to trr being introduced in Bug 1434852 much later than this bug report. The behavior you describe might similar to the ones observed in Bug 1540584. It think it might be sufficiently different to warrant opening a new bug for this due to not being UX, but behavior change. Do you mind opening a new bug?

Comment 8

[BugBot [:suhaib / :marco/ :calixte]]

Clear a needinfo that is pending on an inactive user.

Inactive users most likely will not respond; if the missing information is essential and cannot be collected another way, the bug maybe should be closed as INCOMPLETE.

For more information, please visit BugBot documentation.