|
|
|
/* Copyright (C) 2007-2012 Open Information Security Foundation
|
|
|
|
*
|
|
|
|
* You can copy, redistribute or modify this Program under the terms of
|
|
|
|
* the GNU General Public License version 2 as published by the Free
|
|
|
|
* Software Foundation.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* version 2 along with this program; if not, write to the Free Software
|
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
|
|
|
|
* 02110-1301, USA.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/**
|
|
|
|
* \file
|
|
|
|
*
|
|
|
|
* \author Endace Technology Limited.
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
* \author Victor Julien <victor@inliniac.net>
|
|
|
|
*
|
|
|
|
* An API for profiling operations.
|
|
|
|
*
|
|
|
|
* Really just a wrapper around the existing perf counters.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "suricata-common.h"
|
|
|
|
#include "decode.h"
|
|
|
|
#include "detect.h"
|
|
|
|
#include "detect-engine-prefilter.h"
|
|
|
|
#include "conf.h"
|
|
|
|
#include "flow-worker.h"
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
#include "tm-threads.h"
|
|
|
|
|
|
|
|
#include "util-unittest.h"
|
|
|
|
#include "util-byte.h"
|
|
|
|
#include "util-profiling.h"
|
|
|
|
#include "util-profiling-locks.h"
|
|
|
|
|
|
|
|
#ifdef PROFILING
|
|
|
|
|
|
|
|
#ifndef MIN
|
|
|
|
#define MIN(a, b) (((a) < (b)) ? (a) : (b))
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#define DEFAULT_LOG_FILENAME "profile.log"
|
|
|
|
#define DEFAULT_LOG_MODE_APPEND "yes"
|
|
|
|
|
|
|
|
static pthread_mutex_t packet_profile_lock;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
static FILE *packet_profile_csv_fp = NULL;
|
|
|
|
|
|
|
|
extern int profiling_locks_enabled;
|
|
|
|
extern int profiling_locks_output_to_file;
|
|
|
|
extern char *profiling_locks_file_name;
|
|
|
|
extern const char *profiling_locks_file_mode;
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
typedef struct SCProfilePacketData_ {
|
|
|
|
uint64_t min;
|
|
|
|
uint64_t max;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
uint64_t tot;
|
|
|
|
uint64_t cnt;
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
uint64_t lock;
|
|
|
|
uint64_t ticks;
|
|
|
|
uint64_t contention;
|
|
|
|
|
|
|
|
uint64_t slock;
|
|
|
|
uint64_t sticks;
|
|
|
|
uint64_t scontention;
|
|
|
|
#endif
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
} SCProfilePacketData;
|
|
|
|
SCProfilePacketData packet_profile_data4[257]; /**< all proto's + tunnel */
|
|
|
|
SCProfilePacketData packet_profile_data6[257]; /**< all proto's + tunnel */
|
|
|
|
|
|
|
|
/* each module, each proto */
|
|
|
|
SCProfilePacketData packet_profile_tmm_data4[TMM_SIZE][257];
|
|
|
|
SCProfilePacketData packet_profile_tmm_data6[TMM_SIZE][257];
|
|
|
|
|
|
|
|
SCProfilePacketData packet_profile_app_data4[TMM_SIZE][257];
|
|
|
|
SCProfilePacketData packet_profile_app_data6[TMM_SIZE][257];
|
|
|
|
|
|
|
|
SCProfilePacketData packet_profile_app_pd_data4[257];
|
|
|
|
SCProfilePacketData packet_profile_app_pd_data6[257];
|
|
|
|
|
|
|
|
SCProfilePacketData packet_profile_detect_data4[PROF_DETECT_SIZE][257];
|
|
|
|
SCProfilePacketData packet_profile_detect_data6[PROF_DETECT_SIZE][257];
|
|
|
|
|
|
|
|
SCProfilePacketData packet_profile_log_data4[LOGGER_SIZE][256];
|
|
|
|
SCProfilePacketData packet_profile_log_data6[LOGGER_SIZE][256];
|
|
|
|
|
|
|
|
struct ProfileProtoRecords {
|
|
|
|
SCProfilePacketData records4[257];
|
|
|
|
SCProfilePacketData records6[257];
|
|
|
|
};
|
|
|
|
|
|
|
|
struct ProfileProtoRecords packet_profile_flowworker_data[PROFILE_FLOWWORKER_SIZE];
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
int profiling_packets_enabled = 0;
|
|
|
|
int profiling_output_to_file = 0;
|
|
|
|
|
|
|
|
static int profiling_packets_csv_enabled = 0;
|
|
|
|
static int profiling_packets_output_to_file = 0;
|
|
|
|
static char *profiling_file_name;
|
|
|
|
static char profiling_packets_file_name[PATH_MAX];
|
|
|
|
static char *profiling_csv_file_name;
|
|
|
|
static const char *profiling_packets_file_mode = "a";
|
|
|
|
|
|
|
|
static int rate = 1;
|
|
|
|
static SC_ATOMIC_DECLARE(uint64_t, samples);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Used as a check so we don't double enter a profiling run.
|
|
|
|
*/
|
|
|
|
thread_local int profiling_rules_entered = 0;
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
void SCProfilingDumpPacketStats(void);
|
|
|
|
const char * PacketProfileDetectIdToString(PacketProfileDetectId id);
|
|
|
|
const char * PacketProfileLoggertIdToString(LoggerId id);
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
static void PrintCSVHeader(void);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
static void FormatNumber(uint64_t num, char *str, size_t size)
|
|
|
|
{
|
|
|
|
if (num < 1000UL)
|
|
|
|
snprintf(str, size, "%"PRIu64, num);
|
|
|
|
else if (num < 1000000UL)
|
|
|
|
snprintf(str, size, "%3.1fk", (float)num/1000UL);
|
|
|
|
else if (num < 1000000000UL)
|
|
|
|
snprintf(str, size, "%3.1fm", (float)num/1000000UL);
|
|
|
|
else
|
|
|
|
snprintf(str, size, "%3.1fb", (float)num/1000000000UL);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* \brief Initialize profiling.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
SCProfilingInit(void)
|
|
|
|
{
|
|
|
|
ConfNode *conf;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
SC_ATOMIC_INIT(samples);
|
|
|
|
|
|
|
|
intmax_t rate_v = 0;
|
|
|
|
(void)ConfGetInt("profiling.sample-rate", &rate_v);
|
|
|
|
if (rate_v > 0 && rate_v < INT_MAX) {
|
|
|
|
rate = (int)rate_v;
|
|
|
|
if (rate != 1)
|
|
|
|
SCLogInfo("profiling runs for every %dth packet", rate);
|
|
|
|
else
|
|
|
|
SCLogInfo("profiling runs for every packet");
|
|
|
|
}
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
conf = ConfGetNode("profiling.packets");
|
|
|
|
if (conf != NULL) {
|
|
|
|
if (ConfNodeChildValueIsTrue(conf, "enabled")) {
|
|
|
|
profiling_packets_enabled = 1;
|
|
|
|
|
|
|
|
if (pthread_mutex_init(&packet_profile_lock, NULL) != 0) {
|
|
|
|
FatalError(SC_ERR_FATAL,
|
|
|
|
"Failed to initialize packet profiling mutex.");
|
|
|
|
}
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
memset(&packet_profile_data4, 0, sizeof(packet_profile_data4));
|
|
|
|
memset(&packet_profile_data6, 0, sizeof(packet_profile_data6));
|
|
|
|
memset(&packet_profile_tmm_data4, 0, sizeof(packet_profile_tmm_data4));
|
|
|
|
memset(&packet_profile_tmm_data6, 0, sizeof(packet_profile_tmm_data6));
|
|
|
|
memset(&packet_profile_app_data4, 0, sizeof(packet_profile_app_data4));
|
|
|
|
memset(&packet_profile_app_data6, 0, sizeof(packet_profile_app_data6));
|
|
|
|
memset(&packet_profile_app_pd_data4, 0, sizeof(packet_profile_app_pd_data4));
|
|
|
|
memset(&packet_profile_app_pd_data6, 0, sizeof(packet_profile_app_pd_data6));
|
|
|
|
memset(&packet_profile_detect_data4, 0, sizeof(packet_profile_detect_data4));
|
|
|
|
memset(&packet_profile_detect_data6, 0, sizeof(packet_profile_detect_data6));
|
|
|
|
memset(&packet_profile_log_data4, 0, sizeof(packet_profile_log_data4));
|
|
|
|
memset(&packet_profile_log_data6, 0, sizeof(packet_profile_log_data6));
|
|
|
|
memset(&packet_profile_flowworker_data, 0, sizeof(packet_profile_flowworker_data));
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
const char *filename = ConfNodeLookupChildValue(conf, "filename");
|
|
|
|
if (filename != NULL) {
|
|
|
|
const char *log_dir;
|
|
|
|
log_dir = ConfigGetLogDirectory();
|
|
|
|
|
|
|
|
snprintf(profiling_packets_file_name, sizeof(profiling_packets_file_name),
|
|
|
|
"%s/%s", log_dir, filename);
|
|
|
|
|
|
|
|
const char *v = ConfNodeLookupChildValue(conf, "append");
|
|
|
|
if (v == NULL || ConfValIsTrue(v)) {
|
|
|
|
profiling_packets_file_mode = "a";
|
|
|
|
} else {
|
|
|
|
profiling_packets_file_mode = "w";
|
|
|
|
}
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
profiling_packets_output_to_file = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
conf = ConfGetNode("profiling.packets.csv");
|
|
|
|
if (conf != NULL) {
|
|
|
|
if (ConfNodeChildValueIsTrue(conf, "enabled")) {
|
|
|
|
const char *filename = ConfNodeLookupChildValue(conf, "filename");
|
|
|
|
if (filename == NULL) {
|
|
|
|
filename = "packet_profile.csv";
|
|
|
|
}
|
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
const char *log_dir = ConfigGetLogDirectory();
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
profiling_csv_file_name = SCMalloc(PATH_MAX);
|
|
|
|
if (unlikely(profiling_csv_file_name == NULL)) {
|
|
|
|
FatalError(SC_ERR_FATAL, "out of memory");
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
snprintf(profiling_csv_file_name, PATH_MAX, "%s/%s", log_dir, filename);
|
|
|
|
|
|
|
|
packet_profile_csv_fp = fopen(profiling_csv_file_name, "w");
|
|
|
|
if (packet_profile_csv_fp == NULL) {
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
SCFree(profiling_csv_file_name);
|
|
|
|
profiling_csv_file_name = NULL;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
return;
|
|
|
|
}
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
|
|
|
|
PrintCSVHeader();
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
profiling_packets_csv_enabled = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
conf = ConfGetNode("profiling.locks");
|
|
|
|
if (conf != NULL) {
|
|
|
|
if (ConfNodeChildValueIsTrue(conf, "enabled")) {
|
|
|
|
#ifndef PROFILE_LOCKING
|
|
|
|
SCLogWarning(SC_WARN_PROFILE, "lock profiling not compiled in. Add --enable-profiling-locks to configure.");
|
|
|
|
#else
|
|
|
|
profiling_locks_enabled = 1;
|
|
|
|
|
|
|
|
LockRecordInitHash();
|
|
|
|
|
|
|
|
const char *filename = ConfNodeLookupChildValue(conf, "filename");
|
|
|
|
if (filename != NULL) {
|
|
|
|
const char *log_dir = ConfigGetLogDirectory();
|
|
|
|
|
|
|
|
profiling_locks_file_name = SCMalloc(PATH_MAX);
|
|
|
|
if (unlikely(profiling_locks_file_name == NULL)) {
|
|
|
|
FatalError(SC_ERR_FATAL, "can't duplicate file name");
|
|
|
|
}
|
|
|
|
|
|
|
|
snprintf(profiling_locks_file_name, PATH_MAX, "%s/%s", log_dir, filename);
|
|
|
|
|
|
|
|
const char *v = ConfNodeLookupChildValue(conf, "append");
|
|
|
|
if (v == NULL || ConfValIsTrue(v)) {
|
|
|
|
profiling_locks_file_mode = "a";
|
|
|
|
} else {
|
|
|
|
profiling_locks_file_mode = "w";
|
|
|
|
}
|
|
|
|
|
|
|
|
profiling_locks_output_to_file = 1;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* \brief Free resources used by profiling.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
SCProfilingDestroy(void)
|
|
|
|
{
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
if (profiling_packets_enabled) {
|
|
|
|
pthread_mutex_destroy(&packet_profile_lock);
|
|
|
|
}
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
if (profiling_packets_csv_enabled) {
|
|
|
|
if (packet_profile_csv_fp != NULL)
|
|
|
|
fclose(packet_profile_csv_fp);
|
|
|
|
packet_profile_csv_fp = NULL;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
|
|
|
|
if (profiling_csv_file_name != NULL)
|
|
|
|
SCFree(profiling_csv_file_name);
|
|
|
|
profiling_csv_file_name = NULL;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
if (profiling_file_name != NULL)
|
|
|
|
SCFree(profiling_file_name);
|
|
|
|
profiling_file_name = NULL;
|
|
|
|
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
LockRecordFreeHash();
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
SCProfilingDump(void)
|
|
|
|
{
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
SCProfilingDumpPacketStats();
|
|
|
|
SCLogPerf("Done dumping profiling data.");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void DumpFlowWorkerIP(FILE *fp, int ipv, uint64_t total)
|
|
|
|
{
|
|
|
|
char totalstr[256];
|
|
|
|
|
|
|
|
enum ProfileFlowWorkerId fwi;
|
|
|
|
for (fwi = 0; fwi < PROFILE_FLOWWORKER_SIZE; fwi++) {
|
|
|
|
struct ProfileProtoRecords *r = &packet_profile_flowworker_data[fwi];
|
|
|
|
for (int p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = ipv == 4 ? &r->records4[p] : &r->records6[p];
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-20s IPv%d %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %-6.2f\n",
|
|
|
|
ProfileFlowWorkerIdToString(fwi), ipv, p, pd->cnt,
|
|
|
|
pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void DumpFlowWorker(FILE *fp)
|
|
|
|
{
|
|
|
|
uint64_t total = 0;
|
|
|
|
|
|
|
|
enum ProfileFlowWorkerId fwi;
|
|
|
|
for (fwi = 0; fwi < PROFILE_FLOWWORKER_SIZE; fwi++) {
|
|
|
|
struct ProfileProtoRecords *r = &packet_profile_flowworker_data[fwi];
|
|
|
|
for (int p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &r->records4[p];
|
|
|
|
total += pd->tot;
|
|
|
|
pd = &r->records6[p];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(fp, "\n%-20s %-6s %-5s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"Flow Worker", "IP ver", "Proto", "cnt", "min", "max", "avg");
|
|
|
|
fprintf(fp, "%-20s %-6s %-5s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"--------------------", "------", "-----", "----------", "------------", "------------", "-----------");
|
|
|
|
DumpFlowWorkerIP(fp, 4, total);
|
|
|
|
DumpFlowWorkerIP(fp, 6, total);
|
|
|
|
fprintf(fp, "Note: %s includes app-layer for TCP\n",
|
|
|
|
ProfileFlowWorkerIdToString(PROFILE_FLOWWORKER_STREAM));
|
|
|
|
}
|
|
|
|
|
|
|
|
void SCProfilingDumpPacketStats(void)
|
|
|
|
{
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
int i;
|
|
|
|
FILE *fp;
|
|
|
|
char totalstr[256];
|
|
|
|
uint64_t total;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
if (profiling_packets_enabled == 0)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (profiling_packets_output_to_file == 1) {
|
|
|
|
fp = fopen(profiling_packets_file_name, profiling_packets_file_mode);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
if (fp == NULL) {
|
|
|
|
SCLogError(SC_ERR_FOPEN, "failed to open %s: %s",
|
|
|
|
profiling_packets_file_name, strerror(errno));
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
fp = stdout;
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(fp, "\n\nPacket profile dump:\n");
|
|
|
|
|
|
|
|
fprintf(fp, "\n%-6s %-5s %-12s %-12s %-12s %-12s %-12s %-3s\n",
|
|
|
|
"IP ver", "Proto", "cnt", "min", "max", "avg", "tot", "%%");
|
|
|
|
fprintf(fp, "%-6s %-5s %-12s %-12s %-12s %-12s %-12s %-3s\n",
|
|
|
|
"------", "-----", "----------", "------------", "------------", "-----------", "-----------", "---");
|
|
|
|
total = 0;
|
|
|
|
for (i = 0; i < 257; i++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_data4[i];
|
|
|
|
total += pd->tot;
|
|
|
|
pd = &packet_profile_data6[i];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
for (i = 0; i < 257; i++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_data4[i];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, " IPv4 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %6.2f\n", i, pd->cnt,
|
|
|
|
pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < 257; i++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_data6[i];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, " IPv6 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %6.2f\n", i, pd->cnt,
|
|
|
|
pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
fprintf(fp, "Note: Protocol 256 tracks pseudo/tunnel packets.\n");
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
fprintf(fp, "\nPer Thread module stats:\n");
|
|
|
|
|
|
|
|
fprintf(fp, "\n%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s %-3s",
|
|
|
|
"Thread Module", "IP ver", "Proto", "cnt", "min", "max", "avg", "tot", "%%");
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
fprintf(fp, " %-10s %-10s %-12s %-12s %-10s %-10s %-12s %-12s\n",
|
|
|
|
"locks", "ticks", "cont.", "cont.avg", "slocks", "sticks", "scont.", "scont.avg");
|
|
|
|
#else
|
|
|
|
fprintf(fp, "\n");
|
|
|
|
#endif
|
|
|
|
fprintf(fp, "%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s %-3s",
|
|
|
|
"------------------------", "------", "-----", "----------", "------------", "------------", "-----------", "-----------", "---");
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
fprintf(fp, " %-10s %-10s %-12s %-12s %-10s %-10s %-12s %-12s\n",
|
|
|
|
"--------", "--------", "----------", "-----------", "--------", "--------", "------------", "-----------");
|
|
|
|
#else
|
|
|
|
fprintf(fp, "\n");
|
|
|
|
#endif
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
int m;
|
|
|
|
total = 0;
|
|
|
|
for (m = 0; m < TMM_SIZE; m++) {
|
|
|
|
if (tmm_modules[m].flags & TM_FLAG_LOGAPI_TM)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_tmm_data4[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
|
|
|
|
pd = &packet_profile_tmm_data6[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
for (m = 0; m < TMM_SIZE; m++) {
|
|
|
|
if (tmm_modules[m].flags & TM_FLAG_LOGAPI_TM)
|
|
|
|
continue;
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_tmm_data4[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv4 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %6.2f",
|
|
|
|
TmModuleTmmIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
fprintf(fp, " %10.2f %12"PRIu64" %12"PRIu64" %10.2f %10.2f %12"PRIu64" %12"PRIu64" %10.2f\n",
|
|
|
|
(float)pd->lock/pd->cnt, (uint64_t)pd->ticks/pd->cnt, pd->contention, (float)pd->contention/pd->cnt, (float)pd->slock/pd->cnt, (uint64_t)pd->sticks/pd->cnt, pd->scontention, (float)pd->scontention/pd->cnt);
|
|
|
|
#else
|
|
|
|
fprintf(fp, "\n");
|
|
|
|
#endif
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (m = 0; m < TMM_SIZE; m++) {
|
|
|
|
if (tmm_modules[m].flags & TM_FLAG_LOGAPI_TM)
|
|
|
|
continue;
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_tmm_data6[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv6 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %6.2f\n",
|
|
|
|
TmModuleTmmIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
DumpFlowWorker(fp);
|
|
|
|
|
|
|
|
fprintf(fp, "\nPer App layer parser stats:\n");
|
|
|
|
|
|
|
|
fprintf(fp, "\n%-20s %-6s %-5s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"App Layer", "IP ver", "Proto", "cnt", "min", "max", "avg");
|
|
|
|
fprintf(fp, "%-20s %-6s %-5s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"--------------------", "------", "-----", "----------", "------------", "------------", "-----------");
|
|
|
|
|
|
|
|
total = 0;
|
|
|
|
for (m = 0; m < ALPROTO_MAX; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_app_data4[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
|
|
|
|
pd = &packet_profile_app_data6[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (m = 0; m < ALPROTO_MAX; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_app_data4[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-20s IPv4 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %-6.2f\n",
|
|
|
|
AppProtoToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (m = 0; m < ALPROTO_MAX; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_app_data6[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-20s IPv6 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %-6.2f\n",
|
|
|
|
AppProtoToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* proto detect output */
|
|
|
|
{
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_app_pd_data4[p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
fprintf(fp, "%-20s IPv4 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s\n",
|
|
|
|
"Proto detect", p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr);
|
|
|
|
}
|
|
|
|
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_app_pd_data6[p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
fprintf(fp, "%-20s IPv6 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s\n",
|
|
|
|
"Proto detect", p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
total = 0;
|
|
|
|
for (m = 0; m < PROF_DETECT_SIZE; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_detect_data4[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
|
|
|
|
pd = &packet_profile_detect_data6[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
fprintf(fp, "\n%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s %-3s",
|
|
|
|
"Log Thread Module", "IP ver", "Proto", "cnt", "min", "max", "avg", "tot", "%%");
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
fprintf(fp, " %-10s %-10s %-12s %-12s %-10s %-10s %-12s %-12s\n",
|
|
|
|
"locks", "ticks", "cont.", "cont.avg", "slocks", "sticks", "scont.", "scont.avg");
|
|
|
|
#else
|
|
|
|
fprintf(fp, "\n");
|
|
|
|
#endif
|
|
|
|
fprintf(fp, "%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s %-3s",
|
|
|
|
"------------------------", "------", "-----", "----------", "------------", "------------", "-----------", "-----------", "---");
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
fprintf(fp, " %-10s %-10s %-12s %-12s %-10s %-10s %-12s %-12s\n",
|
|
|
|
"--------", "--------", "----------", "-----------", "--------", "--------", "------------", "-----------");
|
|
|
|
#else
|
|
|
|
fprintf(fp, "\n");
|
|
|
|
#endif
|
|
|
|
total = 0;
|
|
|
|
for (m = 0; m < TMM_SIZE; m++) {
|
|
|
|
if (!(tmm_modules[m].flags & TM_FLAG_LOGAPI_TM))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_tmm_data4[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
|
|
|
|
pd = &packet_profile_tmm_data6[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (m = 0; m < TMM_SIZE; m++) {
|
|
|
|
if (!(tmm_modules[m].flags & TM_FLAG_LOGAPI_TM))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_tmm_data4[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv4 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %6.2f",
|
|
|
|
TmModuleTmmIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
fprintf(fp, " %10.2f %12"PRIu64" %12"PRIu64" %10.2f %10.2f %12"PRIu64" %12"PRIu64" %10.2f\n",
|
|
|
|
(float)pd->lock/pd->cnt, (uint64_t)pd->ticks/pd->cnt, pd->contention, (float)pd->contention/pd->cnt, (float)pd->slock/pd->cnt, (uint64_t)pd->sticks/pd->cnt, pd->scontention, (float)pd->scontention/pd->cnt);
|
|
|
|
#else
|
|
|
|
fprintf(fp, "\n");
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (m = 0; m < TMM_SIZE; m++) {
|
|
|
|
if (!(tmm_modules[m].flags & TM_FLAG_LOGAPI_TM))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_tmm_data6[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv6 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %6.2f\n",
|
|
|
|
TmModuleTmmIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(fp, "\nLogger/output stats:\n");
|
|
|
|
|
|
|
|
total = 0;
|
|
|
|
for (m = 0; m < LOGGER_SIZE; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 256; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_log_data4[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
pd = &packet_profile_log_data6[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(fp, "\n%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"Logger", "IP ver", "Proto", "cnt", "min", "max", "avg", "tot");
|
|
|
|
fprintf(fp, "%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"------------------------", "------", "-----", "----------", "------------", "------------", "-----------", "-----------");
|
|
|
|
for (m = 0; m < LOGGER_SIZE; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 256; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_log_data4[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv4 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %-6.2f\n",
|
|
|
|
PacketProfileLoggertIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (m = 0; m < LOGGER_SIZE; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 256; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_log_data6[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv6 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %-6.2f\n",
|
|
|
|
PacketProfileLoggertIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(fp, "\nGeneral detection engine stats:\n");
|
|
|
|
|
|
|
|
total = 0;
|
|
|
|
for (m = 0; m < PROF_DETECT_SIZE; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_detect_data4[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
pd = &packet_profile_detect_data6[m][p];
|
|
|
|
total += pd->tot;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(fp, "\n%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"Detection phase", "IP ver", "Proto", "cnt", "min", "max", "avg", "tot");
|
|
|
|
fprintf(fp, "%-24s %-6s %-5s %-12s %-12s %-12s %-12s %-12s\n",
|
|
|
|
"------------------------", "------", "-----", "----------", "------------", "------------", "-----------", "-----------");
|
|
|
|
for (m = 0; m < PROF_DETECT_SIZE; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_detect_data4[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv4 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %-6.2f\n",
|
|
|
|
PacketProfileDetectIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (m = 0; m < PROF_DETECT_SIZE; m++) {
|
|
|
|
int p;
|
|
|
|
for (p = 0; p < 257; p++) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_detect_data6[m][p];
|
|
|
|
|
|
|
|
if (pd->cnt == 0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
FormatNumber(pd->tot, totalstr, sizeof(totalstr));
|
|
|
|
double percent = (long double)pd->tot /
|
|
|
|
(long double)total * 100;
|
|
|
|
|
|
|
|
fprintf(fp, "%-24s IPv6 %3d %12"PRIu64" %12"PRIu64" %12"PRIu64" %12"PRIu64" %12s %-6.2f\n",
|
|
|
|
PacketProfileDetectIdToString(m), p, pd->cnt, pd->min, pd->max, (uint64_t)(pd->tot / pd->cnt), totalstr, percent);
|
|
|
|
}
|
|
|
|
}
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
fclose(fp);
|
|
|
|
}
|
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
static void PrintCSVHeader(void)
|
|
|
|
{
|
|
|
|
fprintf(packet_profile_csv_fp, "pcap_cnt,total,receive,decode,flowworker,");
|
|
|
|
fprintf(packet_profile_csv_fp, "threading,");
|
|
|
|
fprintf(packet_profile_csv_fp, "proto detect,");
|
|
|
|
|
|
|
|
for (enum ProfileFlowWorkerId fwi = 0; fwi < PROFILE_FLOWWORKER_SIZE; fwi++) {
|
|
|
|
fprintf(packet_profile_csv_fp, "%s,", ProfileFlowWorkerIdToString(fwi));
|
|
|
|
}
|
|
|
|
fprintf(packet_profile_csv_fp, "loggers,");
|
|
|
|
|
|
|
|
/* detect stages */
|
|
|
|
for (int i = 0; i < PROF_DETECT_SIZE; i++) {
|
|
|
|
fprintf(packet_profile_csv_fp, "%s,", PacketProfileDetectIdToString(i));
|
|
|
|
}
|
|
|
|
|
|
|
|
/* individual loggers */
|
|
|
|
for (LoggerId i = 0; i < LOGGER_SIZE; i++) {
|
|
|
|
fprintf(packet_profile_csv_fp, "%s,", PacketProfileLoggertIdToString(i));
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(packet_profile_csv_fp, "\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
void SCProfilingPrintPacketProfile(Packet *p)
|
|
|
|
{
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
if (profiling_packets_csv_enabled == 0 || p == NULL ||
|
|
|
|
packet_profile_csv_fp == NULL || p->profile == NULL) {
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
uint64_t tmm_total = 0;
|
|
|
|
uint64_t receive = 0;
|
|
|
|
uint64_t decode = 0;
|
|
|
|
|
|
|
|
/* total cost from acquisition to return to packetpool */
|
|
|
|
uint64_t delta = p->profile->ticks_end - p->profile->ticks_start;
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",%"PRIu64",",
|
|
|
|
p->pcap_cnt, delta);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
for (int i = 0; i < TMM_SIZE; i++) {
|
|
|
|
const PktProfilingTmmData *pdt = &p->profile->tmm[i];
|
|
|
|
uint64_t tmm_delta = pdt->ticks_end - pdt->ticks_start;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
if (tmm_modules[i].flags & TM_FLAG_RECEIVE_TM) {
|
|
|
|
if (tmm_delta) {
|
|
|
|
receive = tmm_delta;
|
|
|
|
}
|
|
|
|
continue;
|
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
} else if (tmm_modules[i].flags & TM_FLAG_DECODE_TM) {
|
|
|
|
if (tmm_delta) {
|
|
|
|
decode = tmm_delta;
|
|
|
|
}
|
|
|
|
continue;
|
|
|
|
}
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
tmm_total += tmm_delta;
|
|
|
|
}
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", receive);
|
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", decode);
|
|
|
|
PktProfilingTmmData *fw_pdt = &p->profile->tmm[TMM_FLOWWORKER];
|
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", fw_pdt->ticks_end - fw_pdt->ticks_start);
|
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", delta - tmm_total);
|
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
/* count ticks for app layer */
|
|
|
|
uint64_t app_total = 0;
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
for (AppProto i = 1; i < ALPROTO_FAILED; i++) {
|
|
|
|
const PktProfilingAppData *pdt = &p->profile->app[i];
|
|
|
|
|
|
|
|
if (p->proto == IPPROTO_TCP) {
|
|
|
|
app_total += pdt->ticks_spent;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", p->profile->proto_detect);
|
|
|
|
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
/* print flowworker steps */
|
|
|
|
for (enum ProfileFlowWorkerId fwi = 0; fwi < PROFILE_FLOWWORKER_SIZE; fwi++) {
|
|
|
|
const PktProfilingData *pd = &p->profile->flowworker[fwi];
|
|
|
|
uint64_t ticks_spent = pd->ticks_end - pd->ticks_start;
|
|
|
|
if (fwi == PROFILE_FLOWWORKER_STREAM) {
|
|
|
|
ticks_spent -= app_total;
|
|
|
|
} else if (fwi == PROFILE_FLOWWORKER_APPLAYERUDP && app_total) {
|
|
|
|
ticks_spent = app_total;
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", ticks_spent);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* count loggers cost and print as a single cost */
|
|
|
|
uint64_t loggers = 0;
|
|
|
|
for (LoggerId i = 0; i < LOGGER_SIZE; i++) {
|
|
|
|
const PktProfilingLoggerData *pd = &p->profile->logger[i];
|
|
|
|
loggers += pd->ticks_spent;
|
|
|
|
}
|
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", loggers);
|
|
|
|
|
|
|
|
/* detect steps */
|
|
|
|
for (int i = 0; i < PROF_DETECT_SIZE; i++) {
|
|
|
|
const PktProfilingDetectData *pdt = &p->profile->detect[i];
|
|
|
|
|
|
|
|
fprintf(packet_profile_csv_fp,"%"PRIu64",", pdt->ticks_spent);
|
|
|
|
}
|
profiling/csv: update output format
Update output to be:
pcap_cnt,total,receive,decode,flowworker,threading,proto detect,flow,
stream,app-layer,detect,tcp-prune,loggers,<detect stages>,<loggers>
For TCP, the app-layer cost is not part of stream anymore.
7 years ago
|
|
|
|
|
|
|
/* print individual loggers */
|
|
|
|
for (LoggerId i = 0; i < LOGGER_SIZE; i++) {
|
|
|
|
const PktProfilingLoggerData *pd = &p->profile->logger[i];
|
|
|
|
fprintf(packet_profile_csv_fp, "%"PRIu64",", pd->ticks_spent);
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(packet_profile_csv_fp,"\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketDetectRecord(PacketProfileDetectId id, uint8_t ipproto, PktProfilingDetectData *pdt, int ipver)
|
|
|
|
{
|
|
|
|
if (pdt == NULL) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
SCProfilePacketData *pd;
|
|
|
|
if (ipver == 4)
|
|
|
|
pd = &packet_profile_detect_data4[id][ipproto];
|
|
|
|
else
|
|
|
|
pd = &packet_profile_detect_data6[id][ipproto];
|
|
|
|
|
|
|
|
if (pd->min == 0 || pdt->ticks_spent < pd->min) {
|
|
|
|
pd->min = pdt->ticks_spent;
|
|
|
|
}
|
|
|
|
if (pd->max < pdt->ticks_spent) {
|
|
|
|
pd->max = pdt->ticks_spent;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += pdt->ticks_spent;
|
|
|
|
pd->cnt ++;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketDetectRecords(Packet *p)
|
|
|
|
{
|
|
|
|
PacketProfileDetectId i;
|
|
|
|
for (i = 0; i < PROF_DETECT_SIZE; i++) {
|
|
|
|
PktProfilingDetectData *pdt = &p->profile->detect[i];
|
|
|
|
|
|
|
|
if (pdt->ticks_spent > 0) {
|
|
|
|
if (PKT_IS_IPV4(p)) {
|
|
|
|
SCProfilingUpdatePacketDetectRecord(i, p->proto, pdt, 4);
|
|
|
|
} else {
|
|
|
|
SCProfilingUpdatePacketDetectRecord(i, p->proto, pdt, 6);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketAppPdRecord(uint8_t ipproto, uint32_t ticks_spent, int ipver)
|
|
|
|
{
|
|
|
|
SCProfilePacketData *pd;
|
|
|
|
if (ipver == 4)
|
|
|
|
pd = &packet_profile_app_pd_data4[ipproto];
|
|
|
|
else
|
|
|
|
pd = &packet_profile_app_pd_data6[ipproto];
|
|
|
|
|
|
|
|
if (pd->min == 0 || ticks_spent < pd->min) {
|
|
|
|
pd->min = ticks_spent;
|
|
|
|
}
|
|
|
|
if (pd->max < ticks_spent) {
|
|
|
|
pd->max = ticks_spent;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += ticks_spent;
|
|
|
|
pd->cnt ++;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketAppRecord(int alproto, uint8_t ipproto, PktProfilingAppData *pdt, int ipver)
|
|
|
|
{
|
|
|
|
if (pdt == NULL) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
SCProfilePacketData *pd;
|
|
|
|
if (ipver == 4)
|
|
|
|
pd = &packet_profile_app_data4[alproto][ipproto];
|
|
|
|
else
|
|
|
|
pd = &packet_profile_app_data6[alproto][ipproto];
|
|
|
|
|
|
|
|
if (pd->min == 0 || pdt->ticks_spent < pd->min) {
|
|
|
|
pd->min = pdt->ticks_spent;
|
|
|
|
}
|
|
|
|
if (pd->max < pdt->ticks_spent) {
|
|
|
|
pd->max = pdt->ticks_spent;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += pdt->ticks_spent;
|
|
|
|
pd->cnt ++;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketAppRecords(Packet *p)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0; i < ALPROTO_MAX; i++) {
|
|
|
|
PktProfilingAppData *pdt = &p->profile->app[i];
|
|
|
|
|
|
|
|
if (pdt->ticks_spent > 0) {
|
|
|
|
if (PKT_IS_IPV4(p)) {
|
|
|
|
SCProfilingUpdatePacketAppRecord(i, p->proto, pdt, 4);
|
|
|
|
} else {
|
|
|
|
SCProfilingUpdatePacketAppRecord(i, p->proto, pdt, 6);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (p->profile->proto_detect > 0) {
|
|
|
|
if (PKT_IS_IPV4(p)) {
|
|
|
|
SCProfilingUpdatePacketAppPdRecord(p->proto, p->profile->proto_detect, 4);
|
|
|
|
} else {
|
|
|
|
SCProfilingUpdatePacketAppPdRecord(p->proto, p->profile->proto_detect, 6);
|
|
|
|
}
|
|
|
|
}
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketTmmRecord(int module, uint8_t proto, PktProfilingTmmData *pdt, int ipver)
|
|
|
|
{
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
if (pdt == NULL) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
SCProfilePacketData *pd;
|
|
|
|
if (ipver == 4)
|
|
|
|
pd = &packet_profile_tmm_data4[module][proto];
|
|
|
|
else
|
|
|
|
pd = &packet_profile_tmm_data6[module][proto];
|
|
|
|
|
|
|
|
uint32_t delta = (uint32_t)pdt->ticks_end - pdt->ticks_start;
|
|
|
|
if (pd->min == 0 || delta < pd->min) {
|
|
|
|
pd->min = delta;
|
|
|
|
}
|
|
|
|
if (pd->max < delta) {
|
|
|
|
pd->max = delta;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += (uint64_t)delta;
|
|
|
|
pd->cnt ++;
|
|
|
|
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
pd->lock += pdt->mutex_lock_cnt;
|
|
|
|
pd->ticks += pdt->mutex_lock_wait_ticks;
|
|
|
|
pd->contention += pdt->mutex_lock_contention;
|
|
|
|
pd->slock += pdt->spin_lock_cnt;
|
|
|
|
pd->sticks += pdt->spin_lock_wait_ticks;
|
|
|
|
pd->scontention += pdt->spin_lock_contention;
|
|
|
|
#endif
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketTmmRecords(Packet *p)
|
|
|
|
{
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
int i;
|
|
|
|
for (i = 0; i < TMM_SIZE; i++) {
|
|
|
|
PktProfilingTmmData *pdt = &p->profile->tmm[i];
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
|
|
|
|
if (pdt->ticks_start == 0 || pdt->ticks_end == 0 || pdt->ticks_start > pdt->ticks_end) {
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (PKT_IS_IPV4(p)) {
|
|
|
|
SCProfilingUpdatePacketTmmRecord(i, p->proto, pdt, 4);
|
|
|
|
} else {
|
|
|
|
SCProfilingUpdatePacketTmmRecord(i, p->proto, pdt, 6);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void SCProfilingUpdatePacketGenericRecord(PktProfilingData *pdt,
|
|
|
|
SCProfilePacketData *pd)
|
|
|
|
{
|
|
|
|
if (pdt == NULL || pd == NULL) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
uint64_t delta = pdt->ticks_end - pdt->ticks_start;
|
|
|
|
if (pd->min == 0 || delta < pd->min) {
|
|
|
|
pd->min = delta;
|
|
|
|
}
|
|
|
|
if (pd->max < delta) {
|
|
|
|
pd->max = delta;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += delta;
|
|
|
|
pd->cnt ++;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketGenericRecords(Packet *p, PktProfilingData *pd,
|
|
|
|
struct ProfileProtoRecords *records, int size)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0; i < size; i++) {
|
|
|
|
PktProfilingData *pdt = &pd[i];
|
|
|
|
|
|
|
|
if (pdt->ticks_start == 0 || pdt->ticks_end == 0 || pdt->ticks_start > pdt->ticks_end) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct ProfileProtoRecords *r = &records[i];
|
|
|
|
SCProfilePacketData *store = NULL;
|
|
|
|
|
|
|
|
if (PKT_IS_IPV4(p)) {
|
|
|
|
store = &(r->records4[p->proto]);
|
|
|
|
} else {
|
|
|
|
store = &(r->records6[p->proto]);
|
|
|
|
}
|
|
|
|
|
|
|
|
SCProfilingUpdatePacketGenericRecord(pdt, store);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketLogRecord(LoggerId id,
|
|
|
|
uint8_t ipproto, PktProfilingLoggerData *pdt, int ipver)
|
|
|
|
{
|
|
|
|
if (pdt == NULL) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
SCProfilePacketData *pd;
|
|
|
|
if (ipver == 4)
|
|
|
|
pd = &packet_profile_log_data4[id][ipproto];
|
|
|
|
else
|
|
|
|
pd = &packet_profile_log_data6[id][ipproto];
|
|
|
|
|
|
|
|
if (pd->min == 0 || pdt->ticks_spent < pd->min) {
|
|
|
|
pd->min = pdt->ticks_spent;
|
|
|
|
}
|
|
|
|
if (pd->max < pdt->ticks_spent) {
|
|
|
|
pd->max = pdt->ticks_spent;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += pdt->ticks_spent;
|
|
|
|
pd->cnt++;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void SCProfilingUpdatePacketLogRecords(Packet *p)
|
|
|
|
{
|
|
|
|
for (LoggerId i = 0; i < LOGGER_SIZE; i++) {
|
|
|
|
PktProfilingLoggerData *pdt = &p->profile->logger[i];
|
|
|
|
|
|
|
|
if (pdt->ticks_spent > 0) {
|
|
|
|
if (PKT_IS_IPV4(p)) {
|
|
|
|
SCProfilingUpdatePacketLogRecord(i, p->proto, pdt, 4);
|
|
|
|
} else {
|
|
|
|
SCProfilingUpdatePacketLogRecord(i, p->proto, pdt, 6);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void SCProfilingAddPacket(Packet *p)
|
|
|
|
{
|
|
|
|
if (p == NULL || p->profile == NULL ||
|
|
|
|
p->profile->ticks_start == 0 || p->profile->ticks_end == 0 ||
|
|
|
|
p->profile->ticks_start > p->profile->ticks_end)
|
|
|
|
return;
|
|
|
|
|
|
|
|
pthread_mutex_lock(&packet_profile_lock);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
{
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
if (PKT_IS_IPV4(p)) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_data4[p->proto];
|
|
|
|
|
|
|
|
uint64_t delta = p->profile->ticks_end - p->profile->ticks_start;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
if (pd->min == 0 || delta < pd->min) {
|
|
|
|
pd->min = delta;
|
|
|
|
}
|
|
|
|
if (pd->max < delta) {
|
|
|
|
pd->max = delta;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += delta;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
pd->cnt ++;
|
|
|
|
|
|
|
|
if (IS_TUNNEL_PKT(p)) {
|
|
|
|
pd = &packet_profile_data4[256];
|
|
|
|
|
|
|
|
if (pd->min == 0 || delta < pd->min) {
|
|
|
|
pd->min = delta;
|
|
|
|
}
|
|
|
|
if (pd->max < delta) {
|
|
|
|
pd->max = delta;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += delta;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
pd->cnt ++;
|
|
|
|
}
|
|
|
|
|
|
|
|
SCProfilingUpdatePacketGenericRecords(p, p->profile->flowworker,
|
|
|
|
packet_profile_flowworker_data, PROFILE_FLOWWORKER_SIZE);
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
SCProfilingUpdatePacketTmmRecords(p);
|
|
|
|
SCProfilingUpdatePacketAppRecords(p);
|
|
|
|
SCProfilingUpdatePacketDetectRecords(p);
|
|
|
|
SCProfilingUpdatePacketLogRecords(p);
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
} else if (PKT_IS_IPV6(p)) {
|
|
|
|
SCProfilePacketData *pd = &packet_profile_data6[p->proto];
|
|
|
|
|
|
|
|
uint64_t delta = p->profile->ticks_end - p->profile->ticks_start;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
if (pd->min == 0 || delta < pd->min) {
|
|
|
|
pd->min = delta;
|
|
|
|
}
|
|
|
|
if (pd->max < delta) {
|
|
|
|
pd->max = delta;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += delta;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
pd->cnt ++;
|
|
|
|
|
|
|
|
if (IS_TUNNEL_PKT(p)) {
|
|
|
|
pd = &packet_profile_data6[256];
|
|
|
|
|
|
|
|
if (pd->min == 0 || delta < pd->min) {
|
|
|
|
pd->min = delta;
|
|
|
|
}
|
|
|
|
if (pd->max < delta) {
|
|
|
|
pd->max = delta;
|
|
|
|
}
|
|
|
|
|
|
|
|
pd->tot += delta;
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
pd->cnt ++;
|
|
|
|
}
|
|
|
|
|
|
|
|
SCProfilingUpdatePacketGenericRecords(p, p->profile->flowworker,
|
|
|
|
packet_profile_flowworker_data, PROFILE_FLOWWORKER_SIZE);
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
SCProfilingUpdatePacketTmmRecords(p);
|
|
|
|
SCProfilingUpdatePacketAppRecords(p);
|
|
|
|
SCProfilingUpdatePacketDetectRecords(p);
|
|
|
|
SCProfilingUpdatePacketLogRecords(p);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
|
|
|
|
if (profiling_packets_csv_enabled)
|
|
|
|
SCProfilingPrintPacketProfile(p);
|
|
|
|
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
pthread_mutex_unlock(&packet_profile_lock);
|
Add per packet profiling.
Per packet profiling uses tick based accounting. It has 2 outputs, a summary
and a csv file that contains per packet stats.
Stats per packet include:
1) total ticks spent
2) ticks spent per individual thread module
3) "threading overhead" which is simply calculated by subtracting (2) of (1).
A number of changes were made to integrate the new code in a clean way:
a number of generic enums are now placed in tm-threads-common.h so we can
include them from any part of the engine.
Code depends on --enable-profiling just like the rule profiling code.
New yaml parameters:
profiling:
# packet profiling
packets:
# Profiling can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: yes
filename: packet_stats.log
append: yes
# per packet csv output
csv:
# Output can be disabled here, but it will still have a
# performance impact if compiled in.
enabled: no
filename: packet_stats.csv
Example output of summary stats:
IP ver Proto cnt min max avg
------ ----- ------ ------ ---------- -------
IPv4 6 19436 11448 5404365 32993
IPv4 256 4 11511 49968 30575
Per Thread module stats:
Thread Module IP ver Proto cnt min max avg
------------------------ ------ ----- ------ ------ ---------- -------
TMM_DECODEPCAPFILE IPv4 6 19434 1242 47889 1770
TMM_DETECT IPv4 6 19436 1107 137241 1504
TMM_ALERTFASTLOG IPv4 6 19436 90 1323 155
TMM_ALERTUNIFIED2ALERT IPv4 6 19436 108 1359 138
TMM_ALERTDEBUGLOG IPv4 6 19436 90 1134 154
TMM_LOGHTTPLOG IPv4 6 19436 414 5392089 7944
TMM_STREAMTCP IPv4 6 19434 828 1299159 19438
The proto 256 is a counter for handling of pseudo/tunnel packets.
Example output of csv:
pcap_cnt,ipver,ipproto,total,TMM_DECODENFQ,TMM_VERDICTNFQ,TMM_RECEIVENFQ,TMM_RECEIVEPCAP,TMM_RECEIVEPCAPFILE,TMM_DECODEPCAP,TMM_DECODEPCAPFILE,TMM_RECEIVEPFRING,TMM_DECODEPFRING,TMM_DETECT,TMM_ALERTFASTLOG,TMM_ALERTFASTLOG4,TMM_ALERTFASTLOG6,TMM_ALERTUNIFIEDLOG,TMM_ALERTUNIFIEDALERT,TMM_ALERTUNIFIED2ALERT,TMM_ALERTPRELUDE,TMM_ALERTDEBUGLOG,TMM_ALERTSYSLOG,TMM_LOGDROPLOG,TMM_ALERTSYSLOG4,TMM_ALERTSYSLOG6,TMM_RESPONDREJECT,TMM_LOGHTTPLOG,TMM_LOGHTTPLOG4,TMM_LOGHTTPLOG6,TMM_PCAPLOG,TMM_STREAMTCP,TMM_DECODEIPFW,TMM_VERDICTIPFW,TMM_RECEIVEIPFW,TMM_RECEIVEERFFILE,TMM_DECODEERFFILE,TMM_RECEIVEERFDAG,TMM_DECODEERFDAG,threading
1,4,6,172008,0,0,0,0,0,0,47889,0,0,48582,1323,0,0,0,0,1359,0,1134,0,0,0,0,0,8028,0,0,0,49356,0,0,0,0,0,0,0,14337
First line of the file contains labels.
2 example gnuplot scripts added to plot the data.
14 years ago
|
|
|
}
|
|
|
|
|
|
|
|
PktProfiling *SCProfilePacketStart(void)
|
|
|
|
{
|
|
|
|
uint64_t sample = SC_ATOMIC_ADD(samples, 1);
|
|
|
|
if (sample % rate == 0)
|
|
|
|
return SCCalloc(1, sizeof(PktProfiling));
|
|
|
|
else
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* see if we want to profile rules for this packet */
|
|
|
|
int SCProfileRuleStart(Packet *p)
|
|
|
|
{
|
|
|
|
#ifdef PROFILE_LOCKING
|
|
|
|
if (p->profile != NULL) {
|
|
|
|
p->flags |= PKT_PROFILE;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
uint64_t sample = SC_ATOMIC_ADD(samples, 1);
|
|
|
|
if (sample % rate == 0) {
|
|
|
|
p->flags |= PKT_PROFILE;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
if (p->flags & PKT_PROFILE)
|
|
|
|
return 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
#define CASE_CODE(E) case E: return #E
|
|
|
|
|
|
|
|
/**
|
|
|
|
* \brief Maps the PacketProfileDetectId, to its string equivalent
|
|
|
|
*
|
|
|
|
* \param id PacketProfileDetectId id
|
|
|
|
*
|
|
|
|
* \retval string equivalent for the PacketProfileDetectId id
|
|
|
|
*/
|
|
|
|
const char * PacketProfileDetectIdToString(PacketProfileDetectId id)
|
|
|
|
{
|
|
|
|
switch (id) {
|
|
|
|
CASE_CODE (PROF_DETECT_SETUP);
|
|
|
|
CASE_CODE (PROF_DETECT_GETSGH);
|
|
|
|
CASE_CODE (PROF_DETECT_IPONLY);
|
|
|
|
CASE_CODE (PROF_DETECT_RULES);
|
|
|
|
CASE_CODE (PROF_DETECT_PF_PKT);
|
|
|
|
CASE_CODE (PROF_DETECT_PF_PAYLOAD);
|
|
|
|
CASE_CODE (PROF_DETECT_PF_TX);
|
|
|
|
CASE_CODE (PROF_DETECT_PF_SORT1);
|
|
|
|
CASE_CODE (PROF_DETECT_PF_SORT2);
|
|
|
|
CASE_CODE (PROF_DETECT_NONMPMLIST);
|
|
|
|
CASE_CODE (PROF_DETECT_TX);
|
|
|
|
CASE_CODE (PROF_DETECT_ALERT);
|
|
|
|
CASE_CODE (PROF_DETECT_TX_UPDATE);
|
|
|
|
CASE_CODE (PROF_DETECT_CLEANUP);
|
|
|
|
default:
|
|
|
|
return "UNKNOWN";
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* \brief Maps the LoggerId's to its string equivalent for profiling output.
|
|
|
|
*
|
|
|
|
* \param id LoggerId id
|
|
|
|
*
|
|
|
|
* \retval string equivalent for the LoggerId id
|
|
|
|
*/
|
|
|
|
const char * PacketProfileLoggertIdToString(LoggerId id)
|
|
|
|
{
|
|
|
|
switch (id) {
|
|
|
|
CASE_CODE (LOGGER_UNDEFINED);
|
|
|
|
CASE_CODE (LOGGER_ALERT_DEBUG);
|
|
|
|
CASE_CODE (LOGGER_ALERT_FAST);
|
|
|
|
CASE_CODE (LOGGER_ALERT_SYSLOG);
|
|
|
|
CASE_CODE (LOGGER_JSON_ALERT);
|
|
|
|
CASE_CODE (LOGGER_JSON_ANOMALY);
|
|
|
|
CASE_CODE (LOGGER_JSON_DROP);
|
|
|
|
CASE_CODE (LOGGER_JSON_SSH);
|
|
|
|
CASE_CODE (LOGGER_JSON_SMB);
|
|
|
|
CASE_CODE (LOGGER_JSON_NFS);
|
|
|
|
CASE_CODE(LOGGER_JSON_FRAME);
|
|
|
|
CASE_CODE (LOGGER_HTTP);
|
|
|
|
CASE_CODE(LOGGER_JSON_DNS);
|
|
|
|
CASE_CODE (LOGGER_JSON_DNP3_TS);
|
|
|
|
CASE_CODE (LOGGER_JSON_DNP3_TC);
|
|
|
|
CASE_CODE (LOGGER_JSON_HTTP);
|
|
|
|
CASE_CODE (LOGGER_JSON_DHCP);
|
|
|
|
CASE_CODE (LOGGER_JSON_KRB5);
|
|
|
|
CASE_CODE(LOGGER_JSON_IKE);
|
|
|
|
CASE_CODE(LOGGER_JSON_MODBUS);
|
|
|
|
CASE_CODE (LOGGER_JSON_FTP);
|
|
|
|
CASE_CODE (LOGGER_JSON_TFTP);
|
|
|
|
CASE_CODE (LOGGER_JSON_SMTP);
|
|
|
|
CASE_CODE (LOGGER_JSON_SNMP);
|
|
|
|
CASE_CODE (LOGGER_JSON_TLS);
|
|
|
|
CASE_CODE (LOGGER_JSON_SIP);
|
|
|
|
CASE_CODE (LOGGER_JSON_TEMPLATE_RUST);
|
add RFB parser
This commit adds support for the Remote Framebuffer Protocol (RFB) as
used, for example, by various VNC implementations. It targets the
official versions 3.3, 3.7 and 3.8 of the protocol and provides logging
for the RFB handshake communication for now. Logged events include
endpoint versions, details of the security (i.e. authentication)
exchange as well as metadata about the image transfer parameters.
Detection is enabled using keywords for:
- rfb.name: Session name as sticky buffer
- rfb.sectype: Security type, e.g. VNC-style challenge-response
- rfb.secresult: Result of the security exchange, e.g. OK, FAIL, ...
The latter could be used, for example, to detect brute-force attempts
on open VNC servers, while the name could be used to map unwanted VNC
sessions to the desktop owners or machines.
We also ship example EVE-JSON output and keyword docs as part of the
Sphinx source for Suricata's RTD documentation.
6 years ago
|
|
|
CASE_CODE (LOGGER_JSON_RFB);
|
|
|
|
CASE_CODE (LOGGER_JSON_MQTT);
|
|
|
|
CASE_CODE (LOGGER_JSON_TEMPLATE);
|
|
|
|
CASE_CODE (LOGGER_JSON_RDP);
|
|
|
|
CASE_CODE (LOGGER_JSON_DCERPC);
|
|
|
|
CASE_CODE (LOGGER_JSON_HTTP2);
|
|
|
|
CASE_CODE (LOGGER_TLS_STORE);
|
|
|
|
CASE_CODE (LOGGER_TLS);
|
|
|
|
CASE_CODE (LOGGER_FILE_STORE);
|
|
|
|
CASE_CODE (LOGGER_JSON_FILE);
|
|
|
|
CASE_CODE (LOGGER_TCP_DATA);
|
|
|
|
CASE_CODE (LOGGER_JSON_FLOW);
|
|
|
|
CASE_CODE (LOGGER_JSON_NETFLOW);
|
|
|
|
CASE_CODE (LOGGER_STATS);
|
|
|
|
CASE_CODE (LOGGER_JSON_STATS);
|
|
|
|
CASE_CODE (LOGGER_PCAP);
|
|
|
|
CASE_CODE (LOGGER_JSON_METADATA);
|
|
|
|
case LOGGER_SIZE:
|
|
|
|
return "UNKNOWN";
|
|
|
|
}
|
|
|
|
return "UNKNOWN";
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef UNITTESTS
|
|
|
|
|
|
|
|
static int
|
|
|
|
ProfilingGenericTicksTest01(void)
|
|
|
|
{
|
|
|
|
#define TEST_RUNS 1024
|
|
|
|
uint64_t ticks_start = 0;
|
|
|
|
uint64_t ticks_end = 0;
|
|
|
|
void *ptr[TEST_RUNS];
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
ptr[i] = SCMalloc(1024);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("malloc(1024) %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCFree(ptr[i]);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCFree(1024) %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
SCMutex m[TEST_RUNS];
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCMutexInit(&m[i], NULL);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCMutexInit() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCMutexLock(&m[i]);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCMutexLock() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCMutexUnlock(&m[i]);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCMutexUnlock() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCMutexDestroy(&m[i]);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCMutexDestroy() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
SCSpinlock s[TEST_RUNS];
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCSpinInit(&s[i], 0);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCSpinInit() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCSpinLock(&s[i]);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCSpinLock() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCSpinUnlock(&s[i]);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCSpinUnlock() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SCSpinDestroy(&s[i]);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SCSpinDestroy() %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
SC_ATOMIC_DECL_AND_INIT(unsigned int, test);
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
(void) SC_ATOMIC_ADD(test,1);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SC_ATOMIC_ADD %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
|
|
|
|
ticks_start = UtilCpuGetTicks();
|
|
|
|
for (i = 0; i < TEST_RUNS; i++) {
|
|
|
|
SC_ATOMIC_CAS(&test,i,i+1);
|
|
|
|
}
|
|
|
|
ticks_end = UtilCpuGetTicks();
|
|
|
|
printf("SC_ATOMIC_CAS %"PRIu64"\n", (ticks_end - ticks_start)/TEST_RUNS);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* UNITTESTS */
|
|
|
|
|
|
|
|
void
|
|
|
|
SCProfilingRegisterTests(void)
|
|
|
|
{
|
|
|
|
#ifdef UNITTESTS
|
|
|
|
UtRegisterTest("ProfilingGenericTicksTest01", ProfilingGenericTicksTest01);
|
|
|
|
#endif /* UNITTESTS */
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* PROFILING */
|