def get_args(self):
"""Try to get the args of the all XREFS of the function"""
# load the pattern
args_pattern = open(self.cfg["call_pattern"], "r").read()
args_pattern = args_pattern.replace("FILL_ADDR", str(hex(self.func_addr)))
#print Color.step("Searching for arguments of each XREFS to the function 0x{:X}".format(self.func_addr))
# match the pattern
args_matches = pygrap.match_graph(args_pattern, self.test_graph)
for match in args_matches.items()[0][1]:
self.args.append(match)
def get_xrefs(self):
"""Try to get the xrefs of the matched function"""
# generate our pattern for xrefs
xrefs_pattern = XREFS_PATTERN.replace("FILL_ADDR", str(hex(self.func_addr)))
#print Color.step("Searching for the XREFS to the function 0x{:X}".format(self.func_addr))
# match the pattern
call_matches = pygrap.match_graph(xrefs_pattern, self.test_graph)
# creation of the list of xrefs
if len(call_matches) == 0:
print Color.error("No XREFS to the function 0x{:X} was found".format(self.func_addr))
else:
for match in call_matches["xrefs"]:
self.xrefs.append(int(match["call"][0].info.address))
def decrypt_strings(pe, pe_baseaddr, pe_cfg):
r = parse_decrypt_function(pe, pe_baseaddr, pe_cfg)
if r is None:
print("ERROR: unable to find the decryption function")
sys.exit(1)
func_addr, xorkey_addr, cipher_addr, cipher_size = r
pattern_call_final = pattern_call.replace("FILL_ADDRESS", hex(func_addr))
matches = pygrap.match_graph(pattern_call_final, pe_cfg)
print("\nStrings:")
if "push_call_func" in matches:
for match in matches["push_call_func"]:
push_inst = match["PUSH"][0]
push_addr = push_inst.info.address
offset = pygrap.parse_first_immediate(push_inst.info.arg1)
if offset:
dec = decrypt_string(pe, pe_baseaddr, offset, xorkey_addr, cipher_addr, cipher_size)
print(hex(push_addr), hex(offset), dec)
def parse_decrypt_function(pe, pe_baseaddr, pe_cfg):
matches = pygrap.match_graph(pattern_function, pe_cfg)
if "qakbot_strings_parsing_func" in matches:
for match in matches["qakbot_strings_parsing_func"]:
func_addr = match["0_EP"][0].info.address
size1 = pygrap.parse_first_immediate(match["1_CIPHERSIZE1"][0].info.arg1)
cipher_addr = pygrap.parse_first_immediate(match["2_CIPHER"][0].info.arg1)
size2 = pygrap.parse_first_immediate(match["3_CIPHERSIZE2"][0].info.arg2)
xorkey_addr = pygrap.parse_first_address(match["5_XORKEY"][0].info.arg2)
if size1 != size2 + 1:
print("ERROR: size1 and size2 do not match")
return
print("Decryption function at:", hex(func_addr))
print("XOR key at:", hex(xorkey_addr))
print("XOR key:", pe.get_data(xorkey_addr - pe_baseaddr, 0x40).hex())
print("Cipher block at:", hex(cipher_addr))
print("Cipher block size:", hex(size1))
return func_addr, xorkey_addr, cipher_addr, size1
def get_func_ep_grap(self):
"""Try to get the entrypoint of the matched function"""
first_inst = self.match["first_inst"][0]
first_addr = first_inst.info.address
# Find the beginning of the function:
# It is a node with at least 2 fathers which address a fulfills: address - 0x7 <= a <= address
address_cond = "address >= " + str(hex(int(first_addr - 0x7))) + " and address <= " + str(hex(int(first_addr)))
ep_pattern = EP_PATTERN.replace("FILL_ADDR_COND", address_cond)
#print Color.step("Searching for the function entrypoint near 0x{:X}".format(first_addr))
# search for the EP
ep_matches = pygrap.match_graph(ep_pattern, self.test_graph)
if len(ep_matches) != 1 or len(ep_matches["ep_func"]) != 1:
print Color.error("Entrypoint not found or the function was not call for the address 0x{:X}".format(first_addr))
else:
self.func_addr = int(ep_matches["ep_func"][0]["ep"][0].info.address)
def __init__(self, cfg, test_graph):
"""Initialization
Arguments:
- cfg(Dict): Config for a given pattern
- test_graph(pygrap.graph_t): Grap representation of the binary
"""
self.cfg = cfg
self.test_graph = test_graph
self.matches = []
# search for patterns
matches = pygrap.match_graph(self.cfg["func_pattern"], self.test_graph).items()[0][1]
print Color.info("The graph {name} was found {size} time(s)".format(name=self.cfg["name"],
size=len(matches)))
# process matches
for match in matches:
try:
self.matches.append(Match(match, self.test_graph, self.cfg))
except:
continue
def main():
if len(sys.argv) <= 1:
print_usage()
sys.exit(1)
elif len(sys.argv) <= 2:
verbose = False
bin_path = sys.argv[1]
dot_path = sys.argv[1] + ".grapcfg"
else:
verbose = True
bin_path = sys.argv[2]
dot_path = sys.argv[2] + ".grapcfg"
if len(sys.argv) == 2 and (sys.argv[1] == "-h" or sys.argv[1] == "--help"):
print_usage()
sys.exit(1)
if bin_path[-8:] == ".grapcfg":
print "The argument should be a binary and not a .grapcfg file"
sys.exit(1)
# use_existing specifies wether an existing dot file should be used unchanged or overwritten
pygrap.disassemble_file(bin_path=bin_path,
dot_path=dot_path,
use_existing=True)
test_graph = pygrap.getGraphFromPath(dot_path)
if verbose:
print "Analyzing", bin_path
if not os.path.isfile(bin_path) or not os.path.isfile(dot_path):
print "Error: binary or dot file doesn't exist, exiting."
sys.exit(1)
pattern_decrypt = "backspace_decrypt_algos.grapp"
matches_decrypt = pygrap.match_graph(pattern_decrypt, test_graph)
if len(matches_decrypt) >= 2:
print "Error: two or more decryption algorithms matched, exiting."
sys.exit(1)
for algorithm_name in matches_decrypt:
if verbose:
print "Matched algorithm:", algorithm_name
first_match = matches_decrypt[algorithm_name][0]
first_group = first_match["A"]
first_instruction = first_group[0]
description = first_instruction.info.inst_str
address = first_instruction.info.address
if verbose:
print "Found decryption pattern at address", hex(int(address))
# Find the beginning of the function:
# It is a node with at least 5 fathers which address a fulfills: address - 30 <= a <= address
address_cond = "address >= " + str(hex(
int(address - 30))) + " and address <= " + str(hex(int(address)))
entrypoint_pattern = """
digraph decrypt_fun_begin{
ep [label="ep", cond="nfathers >= 5 and FILL_ADDR_COND", getid="ep"]
}
""".replace("FILL_ADDR_COND", address_cond)
if verbose:
print "Looking for entrypoint with pattern:"
print entrypoint_pattern
matches_entrypoint = pygrap.match_graph(entrypoint_pattern, test_graph)
if len(matches_entrypoint) != 1 or len(
matches_entrypoint["decrypt_fun_begin"]) != 1:
print "Error: Entrypoint not found, exiting"
sys.exit(1)
entrypoint = hex(
int(matches_entrypoint["decrypt_fun_begin"][0]["ep"]
[0].info.address))
if verbose:
print "Found decryption function at", entrypoint
push_call_pattern = """
digraph push_call_decrypt{
push [label="push", cond="opcode is push", repeat=2, getid=push]
call [label="call", cond="opcode is call"]
entrypoint [label="entrypoint", cond="address == FILL_ADDR"]
push -> call
call -> entrypoint [childnumber=2]
}
""".replace("FILL_ADDR", entrypoint)
if verbose:
print "Looking for calls to decrypt function with pattern:"
print push_call_pattern
matches_calls = pygrap.match_graph(push_call_pattern, test_graph)
if len(matches_calls) == 0:
print "error: No call found, exiting"
sys.exit(1)
if verbose:
print len(matches_calls["push_call_decrypt"]
), "calls to decrypt function found."
str_tuple = []
for m in matches_calls["push_call_decrypt"]:
# Work on matches with immediate arguments such as:
# PUSH (between 2 and 5) with hex arguments (for instance: 9, 0x12 or 0x4012a3)
# CALL entrypoint
if len(m["push"]
[-2].info.arg1) == 1 or "0x" in m["push"][-2].info.arg1:
str_tuple.append((int(m["push"][-2].info.arg1,
16), int(m["push"][-1].info.arg1, 16)))
decrypted_strings = decrypt_strings(algorithm_name, str_tuple,
bin_path)
print "Decrypted strings in", bin_path + ":"
for d in decrypted_strings:
print d
print ""
pygrap.graph_free(test_graph, True)
def _analyzing(self):
cfg = self.graph
# Clear the list
del self._analyzed_patterns[:]
#
# Control flow graph extraction
#
if not cfg.graph:
print "[I] Creation of the Control Flow Graph (can take few seconds)"
# Get the CFG of the binary
cfg.extract()
#
# Pattern searching
#
print "[I] Searching for patterns."
# Update "Test" patterns
MODULES["Test"]["Misc"] = get_test_misc()
# Group
for grp_name, grp in MODULES.iteritems():
#print "Group: " + grp_name
# Group->Type
for tp_name, tp in grp.iteritems():
#print "\tType: " + tp_name
if grp_name == "Test" and tp_name == "Misc":
patterns_path_list = []
for algo in tp:
for patterns in algo.get_patterns():
for pattern in patterns.get_patterns():
path = pattern.get_file()
#print "\t\tAdded patterns from " + path
print "Added patterns from " + path
patterns_path_list.append(path)
#print "\t\tMatching patterns against binary... this may take a few seconds"
print "Matching patterns against binary... this may take a few seconds"
matches = match_graph(patterns_path_list,
cfg.graph,
print_all_matches=True)
#print "\t\t", len(matches), "patterns found."
print len(matches), "patterns found."
for pattern_name in matches:
pattern = Pattern(name=pattern_name)
patterns_t = Patterns(patterns=[pattern],
name=pattern_name,
perform_analysis=False,
matches=matches[pattern_name])
pattern._matches = []
for c_match in matches[pattern_name]:
match = Match(c_match, pattern_name)
pattern._matches.append(match)
ana = PatternsAnalysis(patterns_t, None)
ana.filter_patterns()
self._analyzed_patterns.append(ana)
else:
for algo in tp:
# print algo
print "\t\tAlgorithm: %s" % algo.get_name()
# List of Patterns
for patterns in algo.get_patterns():
print "\t\t\tFunction: %s" % patterns.get_name()
# List of Pattern
for pattern in patterns.get_patterns():
print "\t\t\t\t[I] Searching for " + pattern.get_name(
)
pattern.parcourir(cfg.graph)
print "\t\t\t\t[I] %d %s pattern found" % (len(
pattern.get_matches()), pattern.get_name())
#
# Analyzing
#
if patterns._perform_analysis:
print "\t\t\t\t[I] Linking the patterns matches which are in the same area"
ana = PatternsAnalysis(patterns, algo)
print "\t\t\t\t[I] Filtering those patterns"
ana.filter_patterns()
# Add the analyzed patterns to the list
self._analyzed_patterns.append(ana)
def match(self, graph):
path_list = [p.get_file() for p in self._patterns]
self._matches = match_graph(path_list, graph)
