8a5385b9c9
Change-Id: Id4f987dcdacf06622d70263f4659a4400e30dc39 Reviewed-on: https://code.wireshark.org/review/26332 Petri-Dish: Dario Lombardo <lomato@gmail.com> Tested-by: Petri Dish Buildbot Reviewed-by: Anders Broman <a.broman58@gmail.com>
334 lines
12 KiB
Python
Executable file
334 lines
12 KiB
Python
Executable file
#!/usr/bin/env python
|
|
# -*- coding: utf-8 -*-
|
|
#
|
|
# Wireshark - Network traffic analyzer
|
|
# By Gerald Combs <gerald@wireshark.org>
|
|
# Copyright 1998 Gerald Combs
|
|
#
|
|
# SPDX-License-Identifier: GPL-2.0-or-later
|
|
#
|
|
'''\
|
|
Generate Sysdig event dissector sections from the sysdig sources.
|
|
|
|
Reads driver/event_table.c and driver/ppm_events_public.h and generates
|
|
corresponding dissection code in packet-sysdig-event.c. Updates are
|
|
performed in-place in the dissector code.
|
|
|
|
Requires an Internet connection. Assets are loaded from GitHub over HTTPS.
|
|
'''
|
|
|
|
import os
|
|
import os.path
|
|
import re
|
|
import urllib2
|
|
import sys
|
|
|
|
sysdig_repo_pfx = 'https://raw.githubusercontent.com/draios/sysdig/0.5.0/'
|
|
|
|
ppm_ev_pub = urllib2.urlopen(sysdig_repo_pfx + 'driver/ppm_events_public.h')
|
|
ppm_ev_pub_lines = ppm_ev_pub.readlines()
|
|
ppm_ev_pub.close()
|
|
|
|
ppme_re = re.compile('^\s+PPME_([A-Z0-9_]+_[EX])\s*=\s*([0-9]+)\s*,')
|
|
|
|
event_info_d = {}
|
|
|
|
def get_event_defines():
|
|
event_d = {}
|
|
for line in ppm_ev_pub_lines:
|
|
m = ppme_re.match(line)
|
|
if m:
|
|
event_d[int(m.group(2))] = m.group(1)
|
|
return event_d
|
|
|
|
ppm_ev_table = urllib2.urlopen(sysdig_repo_pfx + 'driver/event_table.c')
|
|
ppm_ev_table_lines = ppm_ev_table.readlines()
|
|
ppm_ev_table.close()
|
|
|
|
hf_d = {}
|
|
|
|
event_info_re = re.compile('^\s+/\*\s*PPME_.*\*\/\s*{\s*"([A-Za-z0-9_]+)"\s*,[^,]+,[^,]+,\s*([0-9]+)\s*[,{}]')
|
|
event_param_re = re.compile('{\s*"([A-Za-z0-9_]+)"\s*,\s*PT_([A-Z0-9_]+)\s*,\s*PF_([A-Z0-9_]+)\s*[,}]')
|
|
|
|
def get_event_names():
|
|
'''Return a contiguous list of event names. Names are lower case.'''
|
|
event_name_l = []
|
|
for line in ppm_ev_table_lines:
|
|
ei = event_info_re.match(line)
|
|
if ei:
|
|
event_name_l.append(ei.group(1))
|
|
return event_name_l
|
|
|
|
# PT_xxx to FT_xxx
|
|
pt_to_ft = {
|
|
'BYTEBUF': 'BYTES',
|
|
'CHARBUF': 'STRING',
|
|
'FD': 'INT64',
|
|
'FSPATH': 'STRING',
|
|
}
|
|
|
|
def get_event_params():
|
|
'''Return a list of dictionaries containing event names and parameter info.'''
|
|
event_param_l = []
|
|
event_num = 0
|
|
force_string_l = ['args', 'env']
|
|
for line in ppm_ev_table_lines:
|
|
ei = event_info_re.match(line)
|
|
ep = event_param_re.findall(line)
|
|
if ei and ep:
|
|
src_param_count = int(ei.group(2))
|
|
if len(ep) != src_param_count:
|
|
err_msg = '{}: found {} parameters. Expected {}. Params: {}'.format(
|
|
ei.group(1), len(ep), src_param_count, repr(ep))
|
|
raise NameError(err_msg)
|
|
for p in ep:
|
|
if p[0] in force_string_l:
|
|
param_type = 'STRING'
|
|
elif p[1] in pt_to_ft:
|
|
param_type = pt_to_ft[p[1]]
|
|
elif p[0] == 'flags' and p[1].startswith('INT') and 'HEX' in p[2]:
|
|
param_type = 'U' + p[1]
|
|
elif 'INT' in p[1]:
|
|
# Ints
|
|
param_type = p[1]
|
|
else:
|
|
# Fall back to bytes
|
|
param_type = 'BYTES'
|
|
|
|
if p[2] == 'NA':
|
|
if 'INT' in param_type:
|
|
param_format = 'DEC'
|
|
else:
|
|
param_format = 'NONE'
|
|
elif param_type == 'BYTES':
|
|
param_format = 'NONE'
|
|
else:
|
|
param_format = p[2]
|
|
param_d = {
|
|
'event_name': ei.group(1),
|
|
'event_num': event_num,
|
|
'param_name': p[0],
|
|
'param_type': param_type,
|
|
'param_format': param_format,
|
|
}
|
|
event_param_l.append(param_d)
|
|
if ei:
|
|
event_num += 1
|
|
return event_param_l
|
|
|
|
def param_to_hf_name(param):
|
|
return 'hf_param_{}_{}'.format(param['param_name'], param['param_type'].lower())
|
|
|
|
def param_to_value_string_name(param):
|
|
return '{}_{}_vals'.format(param['param_name'], param['param_type'].lower())
|
|
|
|
def get_param_desc(param):
|
|
# Try to coerce event names and parameters into human-friendly
|
|
# strings.
|
|
# XXX This could use some work.
|
|
|
|
# Specific descriptions. Event name + parameter name.
|
|
param_descs = {
|
|
'accept.queuepct': 'Accept queue per connection',
|
|
'execve.args': 'Program arguments',
|
|
'execve.comm': 'Command',
|
|
'execve.cwd': 'Current working directory',
|
|
}
|
|
# General descriptions. Event name only.
|
|
event_descs = {
|
|
'ioctl': 'I/O control',
|
|
}
|
|
|
|
event_name = param['event_name']
|
|
param_id = '{}.{}'.format(event_name, param['param_name'])
|
|
if param_id in param_descs:
|
|
param_desc = param_descs[param_id]
|
|
elif event_name in event_descs:
|
|
param_desc = '{}: {}'.format(event_descs[event_name], param['param_name'])
|
|
else:
|
|
param_desc = param['param_name']
|
|
return param_desc
|
|
|
|
def main():
|
|
# Event list
|
|
event_d = get_event_defines()
|
|
event_nums = event_d.keys()
|
|
event_nums.sort()
|
|
|
|
event_name_l = get_event_names()
|
|
event_param_l = get_event_params()
|
|
|
|
hf_d = {}
|
|
for param in event_param_l:
|
|
hf_name = param_to_hf_name(param)
|
|
hf_d[hf_name] = param
|
|
|
|
idx_id_to_name = { '': 'no' }
|
|
parameter_index_l = []
|
|
|
|
for en in range (0, len(event_nums)):
|
|
param_id = ''
|
|
param_l = []
|
|
event_var = event_d[en].lower()
|
|
for param in event_param_l:
|
|
if param['event_num'] == en:
|
|
hf_name = param_to_hf_name(param)
|
|
param_l.append(hf_name)
|
|
param_id += ':' + param['param_name'] + '_' + param['param_type']
|
|
|
|
ei_str = ''
|
|
if param_id not in idx_id_to_name:
|
|
idx_id_to_name[param_id] = event_var
|
|
ei_str = 'static const int *{}_indexes[] = {{ &{}, NULL }};'.format(
|
|
event_var,
|
|
', &'.join(param_l)
|
|
)
|
|
else:
|
|
ei_str = '#define {}_indexes {}_indexes'.format(event_var, idx_id_to_name[param_id])
|
|
|
|
parameter_index_l.append(ei_str)
|
|
|
|
dissector_path = os.path.join(os.path.dirname(__file__),
|
|
'..', 'epan', 'dissectors', 'packet-sysdig-event.c')
|
|
dissector_f = open(dissector_path, 'r')
|
|
dissector_lines = list(dissector_f)
|
|
dissector_f = open(dissector_path, 'w+')
|
|
|
|
# Strip out old content
|
|
strip_re_l = []
|
|
strip_re_l.append(re.compile('^static\s+int\s+hf_param_.*;'))
|
|
strip_re_l.append(re.compile('^#define\s+EVT_STR_[A-Z0-9_]+\s+"[A-Za-z0-9_]+"'))
|
|
strip_re_l.append(re.compile('^#define\s+EVT_[A-Z0-9_]+\s+[0-9]+'))
|
|
strip_re_l.append(re.compile('^\s*{\s*EVT_[A-Z0-9_]+\s*,\s*EVT_STR_[A-Z0-9_]+\s*}'))
|
|
strip_re_l.append(re.compile('^static\s+const\s+int\s+\*\s*[a-z0-9_]+_[ex]_indexes\[\]\s*=\s*\{\s*&hf_param_.*NULL\s*\}\s*;'))
|
|
strip_re_l.append(re.compile('^\s*#define\s+[a-z0-9_]+_[ex]_indexes\s+[a-z0-9_]+_indexes'))
|
|
strip_re_l.append(re.compile('^\s*\{\s*EVT_[A-Z0-9_]+_[EX]\s*,\s*[a-z0-9_]+_[ex]_indexes\s*}\s*,'))
|
|
strip_re_l.append(re.compile('^\s*{\s*&hf_param_.*},')) # Must all be on one line
|
|
|
|
for strip_re in strip_re_l:
|
|
dissector_lines = [l for l in dissector_lines if not strip_re.search(l)]
|
|
|
|
# Find our value strings
|
|
value_string_re = re.compile('static\s+const\s+value_string\s+([A-Za-z0-9_]+_vals)')
|
|
value_string_l = []
|
|
for line in dissector_lines:
|
|
vs = value_string_re.match(line)
|
|
if vs:
|
|
value_string_l.append(vs.group(1))
|
|
|
|
# Add in new content after comments.
|
|
|
|
header_fields_c = 'Header fields'
|
|
header_fields_re = re.compile('/\*\s+' + header_fields_c, flags = re.IGNORECASE)
|
|
header_fields_l = []
|
|
for hf_name in sorted(hf_d.keys()):
|
|
header_fields_l.append('static int {} = -1;'.format(hf_name))
|
|
|
|
event_names_c = 'Event names'
|
|
event_names_re = re.compile('/\*\s+' + event_names_c, flags = re.IGNORECASE)
|
|
event_names_l = []
|
|
event_str_l = list(set(event_name_l))
|
|
event_str_l.sort()
|
|
for evt_str in event_str_l:
|
|
event_names_l.append('#define EVT_STR_{0:24s} "{1:s}"'.format(evt_str.upper(), evt_str))
|
|
|
|
event_definitions_c = 'Event definitions'
|
|
event_definitions_re = re.compile('/\*\s+' + event_definitions_c, flags = re.IGNORECASE)
|
|
event_definitions_l = []
|
|
for evt in event_nums:
|
|
event_definitions_l.append('#define EVT_{0:24s} {1:3d}'.format(event_d[evt], evt))
|
|
|
|
value_strings_c = 'Value strings'
|
|
value_strings_re = re.compile('/\*\s+' + value_strings_c, flags = re.IGNORECASE)
|
|
value_strings_l = []
|
|
for evt in event_nums:
|
|
evt_num = 'EVT_{},'.format(event_d[evt])
|
|
evt_str = 'EVT_STR_' + event_name_l[evt].upper()
|
|
value_strings_l.append(' {{ {0:<32s} {1:s} }},'.format(evt_num, evt_str))
|
|
|
|
parameter_index_c = 'Parameter indexes'
|
|
parameter_index_re = re.compile('/\*\s+' + parameter_index_c, flags = re.IGNORECASE)
|
|
# parameter_index_l defined above.
|
|
|
|
event_tree_c = 'Event tree'
|
|
event_tree_re = re.compile('/\*\s+' + event_tree_c, flags = re.IGNORECASE)
|
|
event_tree_l = []
|
|
for evt in event_nums:
|
|
evt_num = 'EVT_{}'.format(event_d[evt])
|
|
evt_idx = '{}_indexes'.format(event_d[evt].lower())
|
|
event_tree_l.append(' {{ {}, {} }},'.format(evt_num, evt_idx))
|
|
|
|
header_field_reg_c = 'Header field registration'
|
|
header_field_reg_re = re.compile('/\*\s+' + header_field_reg_c, flags = re.IGNORECASE)
|
|
header_field_reg_l = []
|
|
for hf_name in sorted(hf_d.keys()):
|
|
param = hf_d[hf_name]
|
|
event_name = param['event_name']
|
|
param_desc = get_param_desc(param)
|
|
param_name = param['param_name']
|
|
param_type = param['param_type']
|
|
param_format = param['param_format']
|
|
fieldconvert = 'NULL'
|
|
vs_name = param_to_value_string_name(param)
|
|
if vs_name in value_string_l and 'INT' in param_type:
|
|
fieldconvert = 'VALS({})'.format(vs_name)
|
|
header_field_reg_l.append(' {{ &{}, {{ "{}", "sysdig.param.{}.{}", FT_{}, BASE_{}, {}, 0, NULL, HFILL }} }},'.format(
|
|
hf_name,
|
|
param_desc,
|
|
event_name,
|
|
param_name,
|
|
param_type,
|
|
param_format,
|
|
fieldconvert
|
|
))
|
|
|
|
for line in dissector_lines:
|
|
fill_comment = None
|
|
fill_l = []
|
|
|
|
if header_fields_re.match(line):
|
|
fill_comment = header_fields_c
|
|
fill_l = header_fields_l
|
|
elif event_names_re.match(line):
|
|
fill_comment = event_names_c
|
|
fill_l = event_names_l
|
|
elif event_definitions_re.match(line):
|
|
fill_comment = event_definitions_c
|
|
fill_l = event_definitions_l
|
|
elif value_strings_re.match(line):
|
|
fill_comment = value_strings_c
|
|
fill_l = value_strings_l
|
|
elif parameter_index_re.match(line):
|
|
fill_comment = parameter_index_c
|
|
fill_l = parameter_index_l
|
|
elif event_tree_re.match(line):
|
|
fill_comment = event_tree_c
|
|
fill_l = event_tree_l
|
|
elif header_field_reg_re.match(line):
|
|
fill_comment = header_field_reg_c
|
|
fill_l = header_field_reg_l
|
|
|
|
if fill_comment is not None:
|
|
# Write our comment followed by the content
|
|
print('Generating {}, {:d} lines'.format(fill_comment, len(fill_l)))
|
|
dissector_f.write('/* {}. Automatically generated by tools/{} */\n'.format(
|
|
fill_comment,
|
|
os.path.basename(__file__)
|
|
))
|
|
for line in fill_l:
|
|
dissector_f.write('{}\n'.format(line))
|
|
# Fill each section only once
|
|
del fill_l[:]
|
|
else:
|
|
# Existing content
|
|
dissector_f.write(line)
|
|
|
|
dissector_f.close()
|
|
|
|
#
|
|
# On with the show
|
|
#
|
|
|
|
if __name__ == "__main__":
|
|
sys.exit(main())
|