Bug 236228 - net-mgmt/etherape: produces high graphics and cpu load, makes display freeze, dumps core
Summary: net-mgmt/etherape: produces high graphics and cpu load, makes display freeze,...
Status: New
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: Any Any
: --- Affects Only Me
Assignee: Ben Woods
URL: https://sourceforge.net/p/etherape/bu...
Keywords:
Depends on:
Blocks:
 
Reported: 2019-03-04 20:57 UTC by Martin Birgmeier
Modified: 2019-03-06 16:24 UTC (History)
0 users

See Also:
bugzilla: maintainer-feedback? (woodsb02)


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Birgmeier 2019-03-04 20:57:59 UTC
Scenario:
- Machine A acting as a gateway to the internet, running mpd5 with PPPoE (interface ng0)
- Machine B used as desktop
- Both machines FreeBSD 12.0 latest patchset (releng/12.0), latest ports compiled using portmaster
- X server on machine B configured to accept remote connections
- $DISPLAY on machine A pointing to B:0
- An xterm/zsh is running on A being displayed on B
- 'etherape -i ng0 &' started on B (should display on A)

Result:
- The etherape window appears on B and starts drawing a few connections
- The display on B becomes sluggish
- One unit of load is added on B, obviously by the X server
- The connection to the xterm on A becomes sluggish

Scenario (continued):
- With great delay, I am able to enter the command 'kill %1' on A

Result:
- After a few seconds, the following appears on the xterm:

write(2) failed in write_all(): Broken pipe

** (etherape:41428): ERROR **: 21:34:51.424: Failed to send message to packet-capture process

- A core dump starts to get written, ultimately reaching nearly 1.5 GB in size:

[0]# 
[1]  + trace trap (core dumped)  etherape -i ng0
[0]# ll *.core
-rw-------  1 root  wheel  1391161344 Mar  4 21:36 etherape.core
[0]# 

Notes:
- The program behaved normally before the upgrade to 0.9.18
- I have the impression that the program should be waiting for a second or so before redrawing the screen but instead runs in a tight loop without any wait, thereby spamming the X server and also the network between A and B.
- The same behavior occurs if I select another interface, including the ethernet between A and B.
Comment 1 Ben Woods freebsd_committer 2019-03-06 16:24:27 UTC
I have reported this upstream here:
https://sourceforge.net/p/etherape/bugs/105/