Project

General

Profile

Actions
Copy link

Bug #4134

closed

Segmentation fault on rule reload when using libmagic

Added by Jeff Lucovsky almost 4 years ago. Updated almost 4 years ago.

Status:
Rejected
Priority:
Normal
Assignee:
Shivani Bhardwaj
Target version:
4.1.10
Affected Versions:
4.1.8
Effort:
medium
Difficulty:
medium
Label:

Description

If libmagic is enabled and there are signature using libmagic, rule reloading causes a segmentation fault on the next file matching the rule.
This is due to an improper reinitialization of the thread contexts.
In attachment there are a sample stack trace and a sample rule file.

Files

Download all files
stack_trace (52.8 KB) stack_trace Angelo Mirabella, 05/26/2020 04:48 PM
test.rules (1.18 KB) test.rules Angelo Mirabella, 05/26/2020 04:50 PM
suricata_testcase.zip (178 KB) suricata_testcase.zip Angelo Mirabella, 05/28/2020 10:54 PM
build-info.txt (3.82 KB) build-info.txt Angelo Mirabella, 05/28/2020 10:55 PM

Related issues 1 (0 open — 1 closed)

Copied from Suricata - Bug #3726: Segmentation fault on rule reload when using libmagicClosedAngelo MirabellaActions
Actions
Copy link
 #1

Updated by Jeff Lucovsky almost 4 years ago

  • Copied from Bug #3726: Segmentation fault on rule reload when using libmagic added
Actions
Copy link
 #2

Updated by Shivani Bhardwaj almost 4 years ago

  • Status changed from Assigned to Rejected

This issue is irreproducible on 4.1.x.
I did the following steps:
1. Started suricata in pcap mode on my network interface
2. reloaded rules with suricatasc
3. replayed the pcap using tcpreplay

I tried this on commit ea15282f4 i.e. the commit prior to its fix in master and was able to reproduce the segfault. After including the fix, the segfault disappeared.
On 4.1.x latest, this issue is not reproducible. Rule reload happens fine. No segfault was observed.

Actions
Copy link

Also available in: Atom PDF