def testReadReferencedMap(self):
## Now we compare the mapped stream with the stream produced by SK:
FILENAME = './referencemap.00.zip'
mapfile = fif.FIFFile([FILENAME], None, False)
map_logfile_stream = mapfile.open_stream_by_name("logfile1.txt")
imagefile = fif.FIFFile(['ntfs1-gen2.00.zip'])
image_stream = imagefile.open_stream_by_name("./images/ntfs1-gen2.dd")
fs = sk.skfs(image_stream)
image_sk_file = fs.open('/RAW/logfile1.txt')
## Do big reads - sk is very slow with compressed ntfs files
BLOCKSIZE = 1024 * 1024
while 1:
test_data = map_logfile_stream.read(BLOCKSIZE)
if len(test_data) == 0: break
data = image_sk_file.read(BLOCKSIZE)
if data != test_data:
print len(data), len(test_data)
self.assertEqual(data, test_data)
image_sk_file.close()
fs.close()
image_stream.close()
map_logfile_stream.close()
mapfile.close()
imagefile.close()
def test_03_ZipVolume(self):
""" Tests the ZipVolume implementation """
print "Loading volume"
z = ZipVolume(None, 'r')
z.load_from(ResolverTests.filename)
## Test the stream implementation
fd = oracle.open(fully_qualified_name("properties", z.urn))
try:
data = z.zf.read("properties")
self.assertEqual(data, fd.read(1024))
finally:
oracle.cache_return(fd)
stream = oracle.open(fully_qualified_name("default", z.urn))
try:
fd = open("output.dd")
while 1:
data = stream.read(1024)
data2 = fd.read(1024)
if not data or not data2: break
self.assertEqual(data2, data)
import sk
stream.seek(0)
fs = sk.skfs(stream)
print fs.listdir('/')
finally:
oracle.cache_return(stream)
def test_03_ZipVolume(self):
""" Tests the ZipVolume implementation """
print "Loading volume"
z = ZipVolume(None, 'r')
z.load_from(ResolverTests.filename)
## Test the stream implementation
fd = oracle.open(fully_qualified_name("properties", z.urn))
try:
data = z.zf.read("properties")
self.assertEqual(data,fd.read(1024))
finally:
oracle.cache_return(fd)
stream = oracle.open(fully_qualified_name("default", z.urn))
try:
fd = open("output.dd")
while 1:
data = stream.read(1024)
data2 = fd.read(1024)
if not data or not data2: break
self.assertEqual(data2, data)
import sk
stream.seek(0)
fs = sk.skfs(stream)
print fs.listdir('/')
finally:
oracle.cache_return(stream)
def testReadAppendedMap(self):
## Now we compare the mapped stream with the stream produced by SK:
FILENAME = './ntfs1-gen2-mapinside.00.zip'
mapfile = fif.FIFFile([FILENAME], None, False)
map_logfile_stream = mapfile.open_stream_by_name("logfile1.txt")
image_stream = mapfile.open_stream_by_name("./images/ntfs1-gen2.dd")
fs = sk.skfs(image_stream)
image_sk_file = fs.open('/RAW/logfile1.txt')
## Do big reads - sk is very slow with compressed ntfs files
BLOCKSIZE = 1024*1024
while 1:
test_data = map_logfile_stream.read(BLOCKSIZE)
if len(test_data)==0: break
data = image_sk_file.read(BLOCKSIZE)
if data!=test_data:
print len(data), len(test_data)
self.assertEqual(data,test_data)
image_sk_file.close()
fs.close()
image_stream.close()
map_logfile_stream.close()
mapfile.close()
def scan(self, fd, scanners, type, mime, cookie, scores=None, **args):
if 'Filesystem' in type:
print "Will load %s" % fd.urn.value
fs = sk.skfs(fd)
for root, dirs, files in fs.walk('/', unalloc=True, inodes=True):
for d, dirname in dirs:
self.create_map(fd, fs, d, FlagFramework.sane_join(root[1], dirname))
for f, filename in files:
self.create_map(fd, fs, f, FlagFramework.sane_join(root[1], filename))
def test_time(fd):
fs = sk.skfs(fd)
f = fs.open('/Compressed/logfile1.txt')
count = 0
while 1:
data = f.read(1024 * 1024 * 30)
if len(data) == 0: break
count += len(data)
return count
def scan(self, fd, scanners, type, mime, cookie, scores=None, **args):
if 'Filesystem' in type:
print "Will load %s" % fd.urn.value
fs = sk.skfs(fd)
for root, dirs, files in fs.walk('/', unalloc=True, inodes=True):
for d, dirname in dirs:
self.create_map(fd, fs, d,
FlagFramework.sane_join(root[1], dirname))
for f, filename in files:
self.create_map(fd, fs, f,
FlagFramework.sane_join(root[1], filename))
def test_time(fd):
fs = sk.skfs(fd)
f = fs.open("/Compressed/logfile1.txt")
count = 0
while 1:
data = f.read(1024 * 1024 * 30)
if len(data) == 0:
break
count += len(data)
return count
def __init__(self, case, fd, inode):
File.__init__(self,case,fd,inode)
cache_key = "%s:%s" % (self.case, self.fd.inode)
try:
fs = SKCACHE.get(cache_key)
except KeyError:
fs = sk.skfs(self.fd)
SKCACHE.put(fs, key=cache_key)
inode = self.inode[self.inode.find('|K')+2:]
self.skfd = fs.open(inode=inode)
self.skfd.seek(0,2)
self.size = self.skfd.tell()
self.skfd.seek(0)
self.block_size = fs.block_size
def setUp(self):
shutil.copyfile('../samples/ntfs1-gen2.00.zip', './ntfs1-gen2.00.zip')
IMAGEFILENAME = 'ntfs1-gen2.00.zip'
fiffile = fif.FIFFile([IMAGEFILENAME])
image = fiffile.open_stream_by_name("./images/ntfs1-gen2.dd")
fs = sk.skfs(image)
f = fs.open('/RAW/logfile1.txt')
## We want to append to the last volume:
#fiffile.append_volume(FILENAME)
count = 0
mapfile = fif.FIFFile()
new_name = "%s.%02d.zip" % ("referencemap", count)
mapfile.create_new_volume(new_name)
## Create a new Map stream
new_stream = mapfile.create_stream_for_writing(
stream_type='aff2-storage:Map', target=image.getId())
new_stream.properties["aff2-storage:name"] = "logfile1.txt"
mapfile.properties["aff2-storage:containsImage"] = new_stream.getId()
mapfile.properties[
"aff2-storage:next_volume"] = "file://%s" % IMAGEFILENAME
count = 0
block_size = fs.block_size
## Build up the mapping function
for block in f.blocks():
new_stream.add_point(count * block_size, block * block_size, 0)
count += 1
new_stream.pack()
f.seek(0, 2)
new_stream.size = f.tell()
new_stream.save_map()
new_stream.close()
mapfile.close()
f.close()
fs.close()
image.close()
fiffile.close()
mapfile.close()
def setUp(self):
shutil.copyfile('../samples/ntfs1-gen2.00.zip', './ntfs1-gen2.00.zip')
IMAGEFILENAME = 'ntfs1-gen2.00.zip'
fiffile = fif.FIFFile([IMAGEFILENAME])
image = fiffile.open_stream_by_name("./images/ntfs1-gen2.dd")
fs = sk.skfs(image)
f = fs.open('/RAW/logfile1.txt')
## We want to append to the last volume:
#fiffile.append_volume(FILENAME)
count = 0
mapfile = fif.FIFFile()
new_name = "%s.%02d.zip" % ("referencemap", count)
mapfile.create_new_volume(new_name)
## Create a new Map stream
new_stream = mapfile.create_stream_for_writing(stream_type = 'aff2-storage:Map',
target = image.getId())
new_stream.properties["aff2-storage:name"] = "logfile1.txt"
mapfile.properties["aff2-storage:containsImage"] = new_stream.getId()
mapfile.properties["aff2-storage:next_volume"] = "file://%s" % IMAGEFILENAME
count = 0
block_size = fs.block_size
## Build up the mapping function
for block in f.blocks():
new_stream.add_point(count * block_size, block * block_size, 0)
count += 1
new_stream.pack()
f.seek(0,2)
new_stream.size = f.tell()
new_stream.save_map()
new_stream.close()
mapfile.close()
f.close()
fs.close()
image.close()
fiffile.close()
mapfile.close()
def scan(self, fd, scanners, type, mime, cookie, scores=None, **args):
if 'x86 boot sector' in type:
try:
parts = sk.mmls(fd)
except IOError, e:
print e
return
for part in parts:
## Make a unique and sensible name for this partition
name = "%s @ 0x%X" % (part[2], part[0])
## Add new maps for each partition
map = CacheManager.AFF4_MANAGER.create_cache_map(
fd.case, "%s/%s" % (fd.urn.parser.query, name))
map.write_from(fd.urn, SECTOR_SIZE * part[0],
SECTOR_SIZE * part[1])
map.close()
## Now we recursively scan each object
fsfd = FileSystem.DBFS(fd.case)
new_fd = fsfd.open(inode_id=map.inode_id)
try:
fs = sk.skfs(new_fd)
fs.close()
## Lets add a hint
Magic.set_magic(fd.case,
inode_id=map.inode_id,
mime="application/filesystem",
magic="Filesystem")
except:
pass
Scanner.scan_inode_distributed(fd.case, map.inode_id, scanners,
cookie)
def scan(self, fd, scanners, type, mime, cookie, scores=None, **args):
if 'x86 boot sector' in type:
try:
parts = sk.mmls(fd)
except IOError,e:
print e
return
for part in parts:
## Make a unique and sensible name for this partition
name = "%s @ 0x%X" % (part[2], part[0])
## Add new maps for each partition
map = CacheManager.AFF4_MANAGER.create_cache_map(
fd.case,
"%s/%s" % (fd.urn.parser.query, name))
map.write_from(fd.urn, SECTOR_SIZE * part[0],
SECTOR_SIZE * part[1])
map.close()
## Now we recursively scan each object
fsfd = FileSystem.DBFS(fd.case)
new_fd = fsfd.open(inode_id = map.inode_id)
try:
fs = sk.skfs(new_fd)
fs.close()
## Lets add a hint
Magic.set_magic(fd.case,
inode_id = map.inode_id,
mime = "application/filesystem",
magic = "Filesystem")
except: pass
Scanner.scan_inode_distributed(fd.case, map.inode_id,
scanners, cookie)
def setUp(self):
shutil.copyfile("../samples/ntfs1-gen2.00.zip", './ntfs1-gen2-mapinside.00.zip')
MAPFILENAME = './ntfs1-gen2-mapinside.00.zip'
fiffile = fif.FIFFile([MAPFILENAME], None, False)
image = fiffile.open_stream_by_name("./images/ntfs1-gen2.dd")
fs = sk.skfs(image)
f = fs.open('/RAW/logfile1.txt')
## We want to append to the last volume:
fiffile.append_volume(MAPFILENAME)
## Create a new Map stream
new_stream = fiffile.create_stream_for_writing(
stream_type = 'aff2-storage:Map',
target = image.getId())
new_stream.properties["aff2-storage:name"] = "logfile1.txt"
fiffile.properties["aff2-storage:containsImage"] = new_stream.getId()
global mappedStreamID
mappedStreamID = new_stream.getId()
count = 0
block_size = fs.block_size
## Build up the mapping function
for block in f.blocks():
new_stream.add_point(count * block_size, block * block_size, 0)
count += 1
new_stream.pack()
f.seek(0,2)
new_stream.size = f.tell()
new_stream.save_map()
new_stream.close()
f.close()
fs.close()
image.close()
fiffile.close()
def setUp(self):
shutil.copyfile("../samples/ntfs1-gen2.00.zip",
'./ntfs1-gen2-mapinside.00.zip')
MAPFILENAME = './ntfs1-gen2-mapinside.00.zip'
fiffile = fif.FIFFile([MAPFILENAME], None, False)
image = fiffile.open_stream_by_name("./images/ntfs1-gen2.dd")
fs = sk.skfs(image)
f = fs.open('/RAW/logfile1.txt')
## We want to append to the last volume:
fiffile.append_volume(MAPFILENAME)
## Create a new Map stream
new_stream = fiffile.create_stream_for_writing(
stream_type='aff2-storage:Map', target=image.getId())
new_stream.properties["aff2-storage:name"] = "logfile1.txt"
fiffile.properties["aff2-storage:containsImage"] = new_stream.getId()
global mappedStreamID
mappedStreamID = new_stream.getId()
count = 0
block_size = fs.block_size
## Build up the mapping function
for block in f.blocks():
new_stream.add_point(count * block_size, block * block_size, 0)
count += 1
new_stream.pack()
f.seek(0, 2)
new_stream.size = f.tell()
new_stream.save_map()
new_stream.close()
f.close()
fs.close()
image.close()
fiffile.close()
parser.add_option("-S",
"--subsys",
default=None,
help="Subsystem to use (e.g. EWF)")
(options, args) = parser.parse_args()
raid_map = mapper.load_map_file(options.map, options.period)
if options.print_map:
mapper.pretty_print(raid_map, options.period, options.number)
print mapper.calculate_map(raid_map, options.period, options.number)
sys.exit(0)
blocksize = mapper.parse_offsets(options.blocksize)
fds = []
for arg in args:
if arg != "None":
fds.append(mapper.open_image(arg, options.subsys))
else:
fds.append(
mapper.ParityDisk(
[mapper.open_image(arg) for arg in args if arg != 'None']))
fd = mapper.RaidReassembler(raid_map,
fds,
blocksize,
skip=mapper.parse_offsets(options.skip))
skfs = sk.skfs(fd, imgoff=128 * 1024 + 512 * 63)
print skfs.listdir("/")
#!/usr/bin/env python
# python implementation of dbtool using sk binding
# should produce the same results as dbtool (!)
import sys
import sk
import pyflag.DB as DB
img = open(sys.argv[1])
fs = sk.skfs(img)
mnt = "mnt"
ios = "Iios"
def runs(blocks):
# converts an ordered list e.g. [1,2,3,4,7,8,9] into a list of
# 'runs'; tuples of (start, length) e.g. [(1,4),(7,3)]
if len(blocks) == 0:
return
index = 0
start = None
length = 1
for i in blocks:
if start==None:
start = i
elif i==start+length:
length+=1
else:
yield index,start,length
index += 1
if "://" in options.output:
output = options.output
else:
output = "file://%s" % options.output
stream = oracle.open(fully_qualified_name(options.stream, VOLUME))
## Make an output volume FIXME - needs to be a utility
output_volume = ZipVolume(None, 'w')
oracle.set(output_volume.urn, AFF4_STORED, output)
output_volume.finish()
output_urn = output_volume.urn
oracle.cache_return(output_volume)
try:
fs = sk.skfs(stream)
block_size = fs.block_size
for root, dirs, files in fs.walk('/', unalloc=True, inodes=True):
for f, filename in files:
print root[1], filename
pathname = os.path.join(root[1], filename)
## The maps have urns based on the path names in the filesystem:
map_stream = Map(None, 'w')
map_stream.urn = fully_qualified_name(pathname, output_urn)
s = fs.stat(inode=str(f))
## Some of these properties should be stored
oracle.set(map_stream.urn, AFF4_SIZE, s.st_size)
oracle.set(map_stream.urn, AFF4_STORED, output_urn)
help = "print the map")
parser.add_option('-o','--output', default="output.dd",
help = "Name of the output file")
parser.add_option("-S", "--subsys",
default=None,
help="Subsystem to use (e.g. EWF)")
(options, args) = parser.parse_args()
raid_map = mapper.load_map_file(options.map, options.period)
if options.print_map:
mapper.pretty_print(raid_map, options.period, options.number)
print mapper.calculate_map(raid_map, options.period, options.number)
sys.exit(0)
blocksize = mapper.parse_offsets(options.blocksize)
fds=[]
for arg in args:
if arg != "None":
fds.append(mapper.open_image(arg, options.subsys))
else:
fds.append(mapper.ParityDisk([mapper.open_image(arg) for arg in args if arg != 'None']))
fd = mapper.RaidReassembler(raid_map, fds, blocksize, skip=mapper.parse_offsets(options.skip))
skfs = sk.skfs(fd, imgoff = 128 * 1024 + 512 * 63)
print skfs.listdir("/")
if "://" in options.output:
output = options.output
else:
output = "file://%s" % options.output
stream = oracle.open(fully_qualified_name(options.stream, VOLUME))
## Make an output volume FIXME - needs to be a utility
output_volume = ZipVolume(None, "w")
oracle.set(output_volume.urn, AFF4_STORED, output)
output_volume.finish()
output_urn = output_volume.urn
oracle.cache_return(output_volume)
try:
fs = sk.skfs(stream)
block_size = fs.block_size
for root, dirs, files in fs.walk("/", unalloc=True, inodes=True):
for f, filename in files:
print root[1], filename
pathname = os.path.join(root[1], filename)
## The maps have urns based on the path names in the filesystem:
map_stream = Map(None, "w")
map_stream.urn = fully_qualified_name(pathname, output_urn)
s = fs.stat(inode=str(f))
## Some of these properties should be stored
oracle.set(map_stream.urn, AFF4_SIZE, s.st_size)
oracle.set(map_stream.urn, AFF4_STORED, output_urn)
def load(self, mount_point, iosource_name, scanners = None, directory=None):
""" Loads the filesystem on mount point from iosource_name. If
scanners are specified - generate jobs for workers as soon as
the inodes are added. If directory is specified we only load
the specified directory.
"""
## Ensure that mount point is normalised:
mount_point = posixpath.normpath(mount_point)
DBFS.load(self, mount_point, iosource_name)
# open the skfs
iosrc = self.iosource
fs = sk.skfs(iosrc)
dbh_file=DB.DBO(self.case)
dbh_inode=DB.DBO(self.case)
dbh_block=DB.DBO(self.case)
if scanners:
scanner_string = ",".join(scanners)
pdbh = DB.DBO()
pdbh.mass_insert_start('jobs')
cookie = int(time.time())
dbh_file.cursor.ignore_warnings = True
dbh_inode.cursor.ignore_warnings = True
dbh_block.cursor.ignore_warnings = True
dbh_file.mass_insert_start("file")
#dbh_inode.mass_insert_start("inode")
dbh_block.mass_insert_start("block")
def insert_file(inode_id, inode, type, path, name):
path = path.decode("utf8","ignore")
name = name.decode("utf8","ignore")
inodestr = "I%s|K%s" % (iosource_name, inode)
pathstr = "%s%s/" % (mount_point, path)
if pathstr.startswith("//"):
pathstr = pathstr[1:]
if pathstr.endswith("//"):
pathstr = pathstr[:-1]
if inode.alloc == 0:
allocstr = "deleted"
type = type[:-1]+'-'
elif inode.alloc == 1:
allocstr = "alloc"
elif inode.alloc == 2:
allocstr = "realloc"
fields = {
"inode":inodestr,
"mode":type,
"status":allocstr,
"path":pathstr,
"name":name
}
if(inode_id):
fields['inode_id'] = inode_id
try:
fields["link"] = fs.readlink(inode=inode)
except IOError:
pass
# insert file entry
dbh_file.mass_insert(**fields)
def runs(blocks):
# converts an ordered list e.g. [1,2,3,4,7,8,9] into a list of
# 'runs'; tuples of (start, length) e.g. [(1,4),(7,3)]
if len(blocks) == 0:
return
index = 0
start = None
length = 1
for i in blocks:
if start==None:
start = i
elif i==start+length:
length+=1
else:
yield index,start,length
index += 1
start = i
length = 1
yield index,start,length
def insert_inode(inode):
""" Inserts inode into database and returns new inode_id and a
stat object for the newly inserted inode """
inode_id = None
# dont do anything for realloc inodes or those with an invalid
# inode number. inode_id 1 is the default (dummy) entry
#if inode.alloc == 2 or str(inode) == "0-0-0":
if str(inode) == "0-0-0":
return 1
inodestr = "I%s|K%s" % (iosource_name, inode)
if inode.alloc:
status = 'alloc'
else:
status = 'deleted'
args = dict(inode = inodestr,
status = status,
_fast = True)
try:
print "%r" % inode
if inode.__str__()=="22-0-0":
print "found it"
raise IOError("foo")
## If this fails we return the default deleted Inode
## because we dont know anything about this inode (we
## dont know its run list or attributes).
f = fs.open(inode=str(inode))
s = fs.fstat(f)
args.update(dict(
uid = s.st_uid,
gid = s.st_gid,
mode = s.st_mode,
links = s.st_nlink,
link = "",
size = s.st_size,
))
if s.st_mtime:
args['_mtime'] = "from_unixtime(%d)" % s.st_mtime
if s.st_atime:
args['_atime'] = "from_unixtime(%d)" % s.st_atime
if s.st_ctime:
args['_ctime'] = "from_unixtime(%d)" % s.st_ctime
#insert block runs
index = 0
for (index, start, count) in runs(f.blocks()):
dbh_block.mass_insert(
inode = inodestr,
index = index,
block = start,
count = count
)
#f.close()
except IOError,e:
pyflaglog.log(pyflaglog.WARNING, "Error creating inode: %s", e)
dbh_inode.insert( "inode", **args)
inode_id = dbh_inode.autoincrement()
## If needed schedule inode for scanning:
if scanners:
pdbh.mass_insert(
command = 'Scan',
arg1 = self.case,
arg2 = inodestr,
arg3= scanner_string,
cookie=cookie,
)
return inode_id
