HTTP: faster processing of asymmetric flows (#2198)

3 files changed, 18 insertions, 5 deletions

diff --git a/src/include/ndpi_typedefs.h b/src/include/ndpi_typedefs.h
index 718fec834..e7338b0fc 100644
--- a/src/include/ndpi_typedefs.h
+++ b/src/include/ndpi_typedefs.h
@@ -798,6 +798,7 @@ struct ndpi_flow_tcp_struct {
 
   /* NDPI_PROTOCOL_HTTP */
   u_int32_t http_stage:3;
+  u_int32_t http_asymmetric_stage:2;
 
   /* NDPI_PROTOCOL_GNUTELLA */
   u_int32_t gnutella_stage:2;		       // 0 - 2
diff --git a/src/lib/protocols/http.c b/src/lib/protocols/http.c
index 68ba42561..2d894f345 100644
--- a/src/lib/protocols/http.c
+++ b/src/lib/protocols/http.c
@@ -1411,6 +1411,8 @@ static void ndpi_check_http_tcp(struct ndpi_detection_module_struct *ndpi_struct
 	   requests. The easiest (but costly) idea is to reset the state and
 	   process it (i.e. we keep the metadata of the last request that we
 	   have processed) */
+        if(flow->l4.tcp.http_asymmetric_stage < 2)
+          flow->l4.tcp.http_asymmetric_stage++;
         reset(ndpi_struct, flow);
         process_request(ndpi_struct, flow, filename_start);
 	return;
@@ -1437,6 +1439,8 @@ static void ndpi_check_http_tcp(struct ndpi_detection_module_struct *ndpi_struct
       NDPI_LOG_DBG2(ndpi_struct, "Another piece of response\n");
       if(is_response(ndpi_struct, flow)) {
         /* See the comment above about how we handle consecutive requests/responses */
+        if(flow->l4.tcp.http_asymmetric_stage < 2)
+          flow->l4.tcp.http_asymmetric_stage++;
         reset(ndpi_struct, flow);
         process_response(ndpi_struct, flow);
 	return;
@@ -1467,8 +1471,16 @@ static void ndpi_search_http_tcp(struct ndpi_detection_module_struct *ndpi_struc
   NDPI_LOG_DBG(ndpi_struct, "search HTTP\n");
   ndpi_check_http_tcp(ndpi_struct, flow);
 
-  if(flow->host_server_name[0] != '\0'&&
-     flow->http.response_status_code != 0) {
+  if((flow->host_server_name[0] != '\0'&&
+      flow->http.response_status_code != 0) ||
+    /* We have found 3 consecutive requests (without the reply) or 3
+       consecutive replies (without the request). If the traffic is really
+       asymmetric, stop here, because we will never find the metadata from
+       both the request and the reply. We wait for 3 events (instead of 2)
+       to avoid false positives triggered by missing/dropped packets */
+    (flow->l4.tcp.http_asymmetric_stage == 2 &&
+     (flow->packet_direction_complete_counter[0] == 0 ||
+      flow->packet_direction_complete_counter[1] == 0))) {
     flow->extra_packets_func = NULL; /* We're good now */
 
     if(flow->initial_binary_bytes_len) ndpi_analyze_content_signature(ndpi_struct, flow);
diff --git a/tests/cfgs/default/result/http_asymmetric.pcapng.out b/tests/cfgs/default/result/http_asymmetric.pcapng.out
index 50c52b918..d65e7683a 100644
--- a/tests/cfgs/default/result/http_asymmetric.pcapng.out
+++ b/tests/cfgs/default/result/http_asymmetric.pcapng.out
@@ -1,6 +1,6 @@
-Guessed flow protos:	1
+Guessed flow protos:	0
 
-DPI Packets (TCP):	21	(10.50 pkts/flow)
+DPI Packets (TCP):	18	(9.00 pkts/flow)
 Confidence DPI              : 2 (flows)
 Num dissector calls: 26 (13.00 diss/flow)
 LRU cache ookla:      0/0/0 (insert/search/found)
@@ -26,4 +26,4 @@ Patricia protocols IPv6: 0/0 (search/found)
 HTTP	23	9961	2
 
 	1	TCP 192.168.1.146:80 -> 192.168.1.103:1044 [proto: 7/HTTP][IP: 0/Unknown][ClearText][Confidence: DPI][DPI packets: 11][cat: Web/5][13 pkts/8357 bytes -> 0 pkts/0 bytes][Goodput ratio: 91/0][5.11 sec][bytes ratio: 1.000 (Upload)][IAT c2s/s2c min/avg/max/stddev: 0/0 464/0 5000/0 1435/0][Pkt Len c2s/s2c min/avg/max/stddev: 60/0 643/0 1514/0 626/0][StatusCode: 404][Content-Type: text/html][Server: Apache/2.4.41 (Ubuntu)][Risk: ** HTTP Susp User-Agent **** Error Code **** Unidirectional Traffic **][Risk Score: 120][Risk Info: No client to server traffic / HTTP Error Code 404 / Empty or missing User-Agent][PLAIN TEXT (HTTP/1.1 200 OK)][Plen Bins: 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14,14,0,0,0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,57,0,0]
-	2	TCP 192.168.0.1:1044 -> 10.10.10.1:80 [proto: 7/HTTP][IP: 0/Unknown][ClearText][Confidence: DPI][DPI packets: 10][cat: Web/5][10 pkts/1604 bytes -> 0 pkts/0 bytes][Goodput ratio: 66/0][5.11 sec][Hostname/SNI: proxy.wiresharkfest.acropolis.local][bytes ratio: 1.000 (Upload)][IAT c2s/s2c min/avg/max/stddev: 0/0 567/0 4951/0 1550/0][Pkt Len c2s/s2c min/avg/max/stddev: 54/0 160/0 418/0 160/0][URL: proxy.wiresharkfest.acropolis.local/favicon.ico][User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101 Firefox/91.0][Risk: ** Unidirectional Traffic **][Risk Score: 10][Risk Info: No server to client traffic][PLAIN TEXT (GET / HTTP/1.1)][Plen Bins: 0,0,0,0,0,0,0,0,0,0,66,33,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+	2	TCP 192.168.0.1:1044 -> 10.10.10.1:80 [proto: 7/HTTP][IP: 0/Unknown][ClearText][Confidence: DPI][DPI packets: 7][cat: Web/5][10 pkts/1604 bytes -> 0 pkts/0 bytes][Goodput ratio: 66/0][5.11 sec][Hostname/SNI: proxy.wiresharkfest.acropolis.local][bytes ratio: 1.000 (Upload)][IAT c2s/s2c min/avg/max/stddev: 0/0 567/0 4951/0 1550/0][Pkt Len c2s/s2c min/avg/max/stddev: 54/0 160/0 418/0 160/0][URL: proxy.wiresharkfest.acropolis.local/favicon.ico][User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101 Firefox/91.0][Risk: ** Unidirectional Traffic **][Risk Score: 10][Risk Info: No server to client traffic][PLAIN TEXT (GET / HTTP/1.1)][Plen Bins: 0,0,0,0,0,0,0,0,0,0,66,33,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]