Bug #2491
closedasync-oneside and midstream not working as expected
Description
I'm running into an issue with one sided traffic and the stream.async-oneside and midstream options.
suricata.yaml and build info is attached. It's the suricata 4.0.4 build from the OISF PPA on Ubuntu 16.04 on AWS EC2.
Also attached is "async-oneside-test.rules", two simple rules designed to trigger on a single packet contained in the two sample PCAPs.
CASE 1:
Testing these rules against the onesided.pcap file, we get a hit on one rule, but not the other:
$ sudo rm /var/log/suricata/fast.log; sudo suricata -c /etc/suricata/suricata.yaml -r simple_http_download.onesided.pcap -S async-oneside-test.rules --set "stream.async-oneside=true" --set "stream.midstream=true" && sudo cat /var/log/suricata/fast.log 17/4/2018 -- 12:00:24 - <Notice> - This is Suricata version 4.0.4 RELEASE 17/4/2018 -- 12:00:24 - <Notice> - all 5 packet processing threads, 4 management threads initialized, engine started. 17/4/2018 -- 12:00:24 - <Notice> - Signal Received. Stopping engine. 17/4/2018 -- 12:00:24 - <Notice> - Pcap-file module read 12 packets, 1853 bytes 01/25/2011-18:52:33.268459 [**] [1:1:1] ALERT http to_server [**] [Classification: (null)] [Priority: 3] {TCP} 192.168.3.131:55963 -> 65.54.95.140:80
CASE 2:
If we remove the syn packet (onesided.nosyn.pcap), we get the same output:
$ sudo rm /var/log/suricata/fast.log; sudo suricata -c /etc/suricata/suricata.yaml -r simple_http_download.onesided.nosyn.pcap -S async-oneside-test.rules --set "stream.async-oneside=true" --set "stream.midstream=true" && sudo cat /var/log/suricata/fast.log 17/4/2018 -- 12:03:47 - <Notice> - This is Suricata version 4.0.4 RELEASE 17/4/2018 -- 12:03:47 - <Notice> - all 5 packet processing threads, 4 management threads initialized, engine started. 17/4/2018 -- 12:03:47 - <Notice> - Signal Received. Stopping engine. 17/4/2018 -- 12:03:47 - <Notice> - Pcap-file module read 11 packets, 1787 bytes 01/25/2011-18:52:33.268459 [**] [1:1:1] ALERT http to_server [**] [Classification: (null)] [Priority: 3] {TCP} 192.168.3.131:55963 -> 65.54.95.140:80
CASE 3:
However, now using the onesided.nosyn.pcap AND turn off async-oneside, we get hits on both signatures
$ sudo rm /var/log/suricata/fast.log; sudo suricata -c /etc/suricata/suricata.yaml -r simple_http_download.onesided.nosyn.pcap -S async-oneside-test.rules --set "stream.async-oneside=false" --set "stream.midstream=true" && sudo cat /var/log/suricata/fast.log 17/4/2018 -- 12:01:49 - <Notice> - This is Suricata version 4.0.4 RELEASE 17/4/2018 -- 12:01:50 - <Notice> - all 5 packet processing threads, 4 management threads initialized, engine started. 17/4/2018 -- 12:01:50 - <Notice> - Signal Received. Stopping engine. 17/4/2018 -- 12:01:50 - <Notice> - Pcap-file module read 11 packets, 1787 bytes 01/25/2011-18:52:33.268459 [**] [1:1:1] ALERT http to_server [**] [Classification: (null)] [Priority: 3] {TCP} 192.168.3.131:55963 -> 65.54.95.140:80 01/25/2011-18:52:33.268459 [**] [1:2:1] ALERT http established and to_server [**] [Classification: (null)] [Priority: 3] {TCP} 192.168.3.131:55963 -> 65.54.95.140:80
CASE 4:
If we leave async-oneside turned off and go back to the original onesided.pcap that includes the syn packet, we get NO output:
$ sudo rm /var/log/suricata/fast.log; sudo suricata -c /etc/suricata/suricata.yaml -r simple_http_download.onesided.pcap -S async-oneside-test.rules --set "stream.async-oneside=false" --set "stream.midstream=true" && sudo cat /var/log/suricata/fast.log 17/4/2018 -- 12:10:36 - <Notice> - This is Suricata version 4.0.4 RELEASE 17/4/2018 -- 12:10:36 - <Notice> - all 5 packet processing threads, 4 management threads initialized, engine started. 17/4/2018 -- 12:10:36 - <Notice> - Signal Received. Stopping engine. 17/4/2018 -- 12:10:36 - <Notice> - Pcap-file module read 12 packets, 1853 bytes
EXPECTED BEHAVIOUR:
Based on my understanding of the manual and these options, I expect to get hits on both example signatures with async-oneside=true and midstream=true (case 1 & 2). I don't expect any hits in case 3, but I get both hits. But given the hits in case 3, I would expect to get the same hits in case 4, but get no hits.
Files
Updated by Andreas Herz over 6 years ago
- Assignee set to OISF Dev
- Target version set to TBD
Updated by Victor Julien over 6 years ago
- Status changed from New to Assigned
- Assignee changed from OISF Dev to Victor Julien
- Target version changed from TBD to 4.1rc2
Updated by Victor Julien over 6 years ago
- Status changed from Assigned to Closed