applications segfault systematically with tsocks
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tsocks (Ubuntu) |
Confirmed
|
Undecided
|
Unassigned |
Bug Description
On Xenial, when an application is used through tsocks, it will systematically segfqult:
$ ssh machine
[connection OK]
$ wget http://
[download succeeds]
$ tsocks ssh machine
Segmentation fault
$ tsocks wget http://
Segmentation fault
$ lsb_release -rd
Description: Ubuntu 16.04 LTS
Release: 16.04
$ apt-cache policy tsocks
tsocks:
Installed: 1.8beta5-9.3
Candidate: 1.8beta5-9.3
Version table:
*** 1.8beta5-9.3 500
500 http://
100 /var/lib/
Expected: connections through tsocks succeed
What happens: connections through tsocks systematically segfault
Here's my tsocks.conf:
$ cat /etc/tsocks.conf
local = 192.168.
server = 127.0.0.1
server_port = 8083
server_type = 5
Running through gdb:
$ gdb ssh machine
[...]
(gdb) run
Starting program: /usr/bin/ssh
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_
Program received signal SIGSEGV, Segmentation fault.
0x000000000
(gdb) bt
#0 0x0000000000000000 in ?? ()
#1 0x00007ffff7670e77 in poll (__timeout=10, __nfds=1, __fds=0x7ffffff
at /usr/include/
#2 RAND_poll () at rand_unix.c:325
#3 0x00007ffff766fbd5 in ssleay_rand_bytes (buf=0x555555804010 "\212\003", num=num@entry=60,
#4 0x00007ffff7670603 in drbg_get_entropy (ctx=<optimized out>, pout=0x7fffffff
#5 0x00007ffff76e6278 in fips_get_entropy (dctx=dctx@
#6 0x00007ffff76e6904 in FIPS_drbg_
at fips_drbg_lib.c:223
#7 0x00007ffff7670993 in RAND_init_fips () at rand_lib.c:302
#8 0x00007ffff75b3f7a in OPENSSL_
#9 0x00007ffff7de74ea in call_init (l=<optimized out>, argc=argc@entry=1,
#10 0x00007ffff7de75fb in call_init (env=0x7fffffff
#11 _dl_init (main_map=
at dl-init.c:120
#12 0x00007ffff7dd7cfa in _dl_start_user () from /lib64/
#13 0x0000000000000001 in ?? ()
#14 0x00007fffffffe4da in ?? ()
#15 0x0000000000000000 in ?? ()
Since Xenial does not have debugging symbols for tsocks, I can't provide much more (although
I'd gladly follow a quick and simple tutorial on doing so).
For the record, it was working in Wily (15.10) and all previous versions.
Status changed to 'Confirmed' because the bug affects multiple users.