Bug #16618

Torsocks make totem unusable: Unsupported syscall number 217

Added by hefee 2019-04-03 09:52:46 . Updated 2019-04-13 06:03:07 .

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Target version:
Tails_4.0
Start date:
Due date:
% Done:

100%

Feature Branch:
Type of work:
Code
Blueprint:

Starter:
Affected tool:
Deliverable for:

Description

All totem tests fails, because torsocks is breaking totem. You can test easily either with /usr/bin/local/totem or


$killall totem
$torsocks /usr/bin/totem /home/amnesia/video.mp4

1554284499 WARNING torsocks[5329]: [syscall] Unsupported syscall number 217. Denying the call (in tsocks_syscall() at syscall.c:568)

It works when you run totem binary directly:

$/usr/bin/totem /home/amnesia/video.mp4

The video is attached.

Files

video.mp4 (28057 B) hefee, 2019-04-03 09:49:46

Subtasks

History

#1 Updated by hefee 2019-04-03 13:02:36

  • Category deleted (Test suite)

#2 Updated by intrigeri 2019-04-03 13:16:36

  • Category set to Test suite

I can’t reproduce this on bare metal (HP EliteBook 840 G1).

In any case, to get this fixed in torsocks:

  1. find out which syscall is 217 on amd64
  2. hope that this syscall can be safely let through by torsocks
  3. report a bug upstream

#3 Updated by intrigeri 2019-04-03 13:21:34

  • Category deleted (Test suite)

#4 Updated by mercedes508 2019-04-03 13:26:25

  • Status changed from New to Confirmed

#5 Updated by hefee 2019-04-03 17:29:59

I tested it on a normal vm and could reproduce the issue. It is actually quite wired, if you run one time /usr/bin/totem, every other call afterwards succeed. That’s why I added this killall command, to make sure no other totem is running in background.

#6 Updated by intrigeri 2019-04-03 17:34:19

> It is actually quite wired, if you run one time /usr/bin/totem, every other call afterwards succeed.

D-Bus activation might be related.

#7 Updated by intrigeri 2019-04-03 18:27:48

  • Status changed from Confirmed to In Progress
  • Assignee set to hefee
  • % Done changed from 0 to 10
  • QA Check set to Ready for QA

Fixed with torsocks 2.3.0-1.0tails1 in feature-buster, please confirm this fixes the corresponding test then reassign to me so I handle the Debian paperwork to try & get this fixed in Buster proper.

#8 Updated by intrigeri 2019-04-03 18:45:22

Better test with 2.3.0-1.0tails2.

#9 Updated by hefee 2019-04-05 06:54:21

  • Assignee changed from hefee to intrigeri
  • QA Check changed from Ready for QA to Pass

The corresponding test on sib passes.

You haven’t pushed your changes to salsa, that’s why I used debdiff to look at the changes from torsocks_2.3.0-1.dsc against torsocks_2.3.0-1.0tails2.dsc:

apt get torsocks=2.3.0-1
dget https://deb.tails.boum.org/pool/main/t/torsocks/torsocks_2.3.0-1.0tails2.dsc
debdiff torsocks_2.3.0-1.dsc torsocks_2.3.0-1.0tails2.dsc

The diff looks like I expected it.

You can now go on doing the Debian paperwork.

#10 Updated by intrigeri 2019-04-05 08:43:13

  • Assignee changed from intrigeri to lamby

> You haven’t pushed your changes to salsa

Done.

> You can now go on doing the Debian paperwork.

lamby will do it :)

#11 Updated by lamby 2019-04-05 10:03:40

Uploaded, pending ACCEPTED mail:

Uploading torsocks_2.3.0-2.dsc
Uploading torsocks_2.3.0-2.debian.tar.xz
Uploading torsocks-dbgsym_2.3.0-2_amd64.deb
Uploading torsocks_2.3.0-2_amd64.buildinfo
Uploading torsocks_2.3.0-2_amd64.deb
Uploading torsocks_2.3.0-2_amd64.changes

#13 Updated by lamby 2019-04-05 10:12:46

  • Status changed from In Progress to Fix committed

Unblock request filed as https://bugs.debian.org/926442

Marking as fix committed.

#15 Updated by intrigeri 2019-04-05 16:02:12

> Unblocked: https://bugs.debian.org/926442#10

\o/

#16 Updated by intrigeri 2019-04-05 16:16:21

  • Status changed from Fix committed to Resolved
  • Assignee deleted (lamby)
  • % Done changed from 10 to 100

Let’s optimistically assume 2.3.0-2 will migrate and supersede 2.3.0-1.0tails2 :)

#17 Updated by intrigeri 2019-04-13 06:03:07

intrigeri wrote:
> Let’s optimistically assume 2.3.0-2 will migrate and supersede 2.3.0-1.0tails2 :)

Confirmed: https://jenkins.tails.boum.org/view/RM/job/build_Tails_ISO_feature-buster/lastSuccessfulBuild/artifact/build-artifacts/tails-amd64-feature_buster-3.17-20190412T0933Z-3e1b1d35f9+devel@eaecddc27e.packages/*view*/ has 2.3.0-2. So I’ve deleted 2.3.0-1.0tails2 from our feature-buster overlay APT suite in order to keep things tidy and avoid potential confusion like “OMG our nicely patched version was superseded, we need to forward-port our patch!”.