Merge remote-tracking branch 'origin/topic/seth/faf-updates'

* origin/topic/seth/faf-updates: (27 commits)
  Undoing the FTP tests I updated earlier.
  Update the last two btest FAF tests.
  File analysis fixes and test updates.
  Fix a bug with getting analyzer tags.
  A few test updates.
  Some tests work now (at least they all don't fail anymore!)
  Forgot a file.
  Added protocol description functions that provide a super compressed log representation.
  Fix a bug where orig file information in http wasn't working right.
  Added mime types to http.log
  Clean up queued but unused file_over_new_connections event args.
  Add jar files to the default MHR lookups.
  Adding CAB files for MHR checking.
  Improve malware hash registry script.
  Fix a small issue with finding smtp entities.
  Added support for files to the notice framework.
  Make the custom libmagic database a git submodule.
  Add an is_orig parameter to file_over_new_connection event.
  Make magic for emitting application/msword mime type less strict.
  Disable more libmagic builtin checks that override the magic database.
  ...

Conflicts:
	doc/scripts/DocSourcesList.cmake
	scripts/base/init-bare.bro
	scripts/test-all-policy.bro
	testing/btest/Baseline/coverage.bare-load-baseline/canonified_loaded_scripts.log
	testing/btest/Baseline/coverage.default-load-baseline/canonified_loaded_scripts.log
This commit is contained in:
Robin Sommer 2013-07-29 14:21:52 -07:00
commit 984e9793db
196 changed files with 1548 additions and 5033 deletions

View file

@ -1,7 +1,7 @@
@load ./utils-commands
@load ./main
@load ./file-analysis
@load ./file-extract
@load ./utils
@load ./files
@load ./gridftp
@load-sigs ./dpd.sig
@load-sigs ./dpd.sig

View file

@ -1,48 +0,0 @@
@load ./main
@load base/utils/conn-ids
@load base/frameworks/file-analysis/main
module FTP;
export {
## Default file handle provider for FTP.
global get_file_handle: function(c: connection, is_orig: bool): string;
}
function get_handle_string(c: connection): string
{
return cat(Analyzer::ANALYZER_FTP_DATA, " ", c$start_time, " ", id_string(c$id));
}
function get_file_handle(c: connection, is_orig: bool): string
{
if ( [c$id$resp_h, c$id$resp_p] !in ftp_data_expected ) return "";
local info: FTP::Info = ftp_data_expected[c$id$resp_h, c$id$resp_p];
if ( info$passive )
# FTP client initiates data channel.
if ( is_orig )
# Don't care about FTP client data.
return "";
else
# Do care about FTP server data.
return get_handle_string(c);
else
# FTP server initiates dta channel.
if ( is_orig )
# Do care about FTP server data.
return get_handle_string(c);
else
# Don't care about FTP client data.
return "";
}
module GLOBAL;
event get_file_handle(tag: Analyzer::Tag, c: connection, is_orig: bool)
&priority=5
{
if ( tag != Analyzer::ANALYZER_FTP_DATA ) return;
set_file_handle(FTP::get_file_handle(c, is_orig));
}

View file

@ -1,90 +0,0 @@
##! File extraction support for FTP.
@load ./main
@load base/utils/files
module FTP;
export {
## Pattern of file mime types to extract from FTP transfers.
const extract_file_types = /NO_DEFAULT/ &redef;
## The on-disk prefix for files to be extracted from FTP-data transfers.
const extraction_prefix = "ftp-item" &redef;
}
redef record Info += {
## On disk file where it was extracted to.
extraction_file: string &log &optional;
## Indicates if the current command/response pair should attempt to
## extract the file if a file was transferred.
extract_file: bool &default=F;
};
function get_extraction_name(f: fa_file): string
{
local r = fmt("%s-%s.dat", extraction_prefix, f$id);
return r;
}
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "FTP_DATA" ) return;
if ( f?$mime_type && extract_file_types in f$mime_type )
{
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=get_extraction_name(f)]);
return;
}
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( [cid$resp_h, cid$resp_p] !in ftp_data_expected ) next;
local s = ftp_data_expected[cid$resp_h, cid$resp_p];
if ( ! s$extract_file ) next;
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=get_extraction_name(f)]);
return;
}
}
event file_state_remove(f: fa_file) &priority=4
{
if ( ! f?$source ) return;
if ( f$source != "FTP_DATA" ) return;
if ( ! f?$info ) return;
for ( filename in f$info$extracted_files )
{
local s: FTP::Info;
s$ts = network_time();
s$tags = set();
s$user = "<ftp-data>";
s$extraction_file = filename;
if ( f?$conns )
for ( cid in f$conns )
{
s$uid = f$conns[cid]$uid;
s$id = cid;
}
Log::write(FTP::LOG, s);
}
}
event log_ftp(rec: Info) &priority=-10
{
delete rec$extraction_file;
delete rec$extract_file;
}

View file

@ -0,0 +1,60 @@
@load ./main
@load ./utils
@load base/utils/conn-ids
@load base/frameworks/files
module FTP;
export {
redef record Info += {
## File unique ID.
fuid: string &optional &log;
};
## Default file handle provider for FTP.
global get_file_handle: function(c: connection, is_orig: bool): string;
## Describe the file being transferred.
global describe_file: function(f: fa_file): string;
}
function get_file_handle(c: connection, is_orig: bool): string
{
if ( [c$id$resp_h, c$id$resp_p] !in ftp_data_expected )
return "";
return cat(Analyzer::ANALYZER_FTP_DATA, c$start_time, c$id, is_orig);
}
function describe_file(f: fa_file): string
{
# This shouldn't be needed, but just in case...
if ( f$source != "FTP" )
return "";
for ( cid in f$conns )
{
if ( f$conns[cid]?$ftp )
return FTP::describe(f$conns[cid]$ftp);
}
return "";
}
event bro_init() &priority=5
{
Files::register_protocol(Analyzer::ANALYZER_FTP_DATA,
[$get_file_handle = FTP::get_file_handle,
$describe = FTP::describe_file]);
}
event file_over_new_connection(f: fa_file, c: connection, is_orig: bool) &priority=5
{
if ( [c$id$resp_h, c$id$resp_p] !in ftp_data_expected )
return;
local ftp = ftp_data_expected[c$id$resp_h, c$id$resp_p];
ftp$fuid = f$id;
if ( f?$mime_type )
ftp$mime_type = f$mime_type;
}

View file

@ -63,8 +63,6 @@ export {
reply_code: count &log &optional;
## Reply message from the server in response to the command.
reply_msg: string &log &optional;
## Arbitrary tags that may indicate a particular attribute of this command.
tags: set[string] &log;
## Expected FTP data channel.
data_channel: ExpectedDataChannel &log &optional;
@ -104,6 +102,8 @@ export {
global log_ftp: event(rec: Info);
}
@load ./utils
# Add the state tracking information variable to the connection record
redef record connection += {
ftp: Info &optional;
@ -171,37 +171,26 @@ function set_ftp_session(c: connection)
function ftp_message(s: Info)
{
# If it either has a tag associated with it (something detected)
# or it's a deliberately logged command.
if ( |s$tags| > 0 || (s?$cmdarg && s$cmdarg$cmd in logged_commands) )
s$ts=s$cmdarg$ts;
s$command=s$cmdarg$cmd;
s$arg = s$cmdarg$arg;
if ( s$cmdarg$cmd in file_cmds )
s$arg = build_url_ftp(s);
if ( s$arg == "" )
delete s$arg;
if ( s?$password &&
! s$capture_password &&
to_lower(s$user) !in guest_ids )
{
if ( s?$password &&
! s$capture_password &&
to_lower(s$user) !in guest_ids )
{
s$password = "<hidden>";
}
local arg = s$cmdarg$arg;
if ( s$cmdarg$cmd in file_cmds )
{
local comp_path = build_path_compressed(s$cwd, arg);
if ( comp_path[0] != "/" )
comp_path = cat("/", comp_path);
arg = fmt("ftp://%s%s", addr_to_uri(s$id$resp_h), comp_path);
}
s$ts=s$cmdarg$ts;
s$command=s$cmdarg$cmd;
if ( arg == "" )
delete s$arg;
else
s$arg=arg;
Log::write(FTP::LOG, s);
s$password = "<hidden>";
}
if ( s?$cmdarg && s$command in logged_commands)
Log::write(FTP::LOG, s);
# The MIME and file_size fields are specific to file transfer commands
# and may not be used in all commands so they need reset to "blank"
# values after logging.
@ -209,8 +198,6 @@ function ftp_message(s: Info)
delete s$file_size;
# Same with data channel.
delete s$data_channel;
# Tags are cleared everytime too.
s$tags = set();
}
function add_expected_data_channel(s: Info, chan: ExpectedDataChannel)
@ -218,8 +205,9 @@ function add_expected_data_channel(s: Info, chan: ExpectedDataChannel)
s$passive = chan$passive;
s$data_channel = chan;
ftp_data_expected[chan$resp_h, chan$resp_p] = s;
Analyzer::schedule_analyzer(chan$orig_h, chan$resp_h, chan$resp_p, Analyzer::ANALYZER_FTP_DATA,
5mins);
Analyzer::schedule_analyzer(chan$orig_h, chan$resp_h, chan$resp_p,
Analyzer::ANALYZER_FTP_DATA,
5mins);
}
event ftp_request(c: connection, command: string, arg: string) &priority=5

View file

@ -0,0 +1,47 @@
##! Utilities specific for FTP processing.
@load ./main
@load base/utils/addrs
module FTP;
export {
## Creates a URL from an :bro:type:`FTP::Info` record.
##
## rec: An :bro:type:`FTP::Info` record.
##
## Returns: A URL, not prefixed by "ftp://".
global build_url: function(rec: Info): string;
## Creates a URL from an :bro:type:`FTP::Info` record.
##
## rec: An :bro:type:`FTP::Info` record.
##
## Returns: A URL prefixed with "ftp://".
global build_url_ftp: function(rec: Info): string;
## Create an extremely shortened representation of a log line.
global describe: function(rec: Info): string;
}
function build_url(rec: Info): string
{
if ( !rec?$arg )
return "";
local comp_path = build_path_compressed(rec$cwd, rec$arg);
if ( comp_path[0] != "/" )
comp_path = cat("/", comp_path);
return fmt("%s%s", addr_to_uri(rec$id$resp_h), comp_path);
}
function build_url_ftp(rec: Info): string
{
return fmt("ftp://%s", build_url(rec));
}
function describe(rec: Info): string
{
return build_url_ftp(rec);
}

View file

@ -1,8 +1,6 @@
@load ./main
@load ./entities
@load ./utils
@load ./file-analysis
@load ./file-ident
@load ./file-hash
@load ./file-extract
@load ./files
@load-sigs ./dpd.sig

View file

@ -0,0 +1,109 @@
##! Analysis and logging for MIME entities found in HTTP sessions.
@load base/frameworks/files
@load base/utils/strings
@load base/utils/files
@load ./main
module HTTP;
export {
type Entity: record {
## Filename for the entity if discovered from a header.
filename: string &optional;
};
redef record Info += {
## An ordered vector of file unique IDs.
orig_fuids: vector of string &log &optional;
## An ordered vector of mime types.
orig_mime_types: vector of string &log &optional;
## An ordered vector of file unique IDs.
resp_fuids: vector of string &log &optional;
## An ordered vector of mime types.
resp_mime_types: vector of string &log &optional;
## The current entity.
current_entity: Entity &optional;
## Current number of MIME entities in the HTTP request message body.
orig_mime_depth: count &default=0;
## Current number of MIME entities in the HTTP response message body.
resp_mime_depth: count &default=0;
};
}
event http_begin_entity(c: connection, is_orig: bool) &priority=10
{
set_state(c, F, is_orig);
if ( is_orig )
++c$http$orig_mime_depth;
else
++c$http$resp_mime_depth;
c$http$current_entity = Entity();
}
event http_header(c: connection, is_orig: bool, name: string, value: string) &priority=3
{
if ( name == "CONTENT-DISPOSITION" &&
/[fF][iI][lL][eE][nN][aA][mM][eE]/ in value )
{
c$http$current_entity$filename = extract_filename_from_content_disposition(value);
}
else if ( name == "CONTENT-TYPE" &&
/[nN][aA][mM][eE][:blank:]*=/ in value )
{
c$http$current_entity$filename = extract_filename_from_content_disposition(value);
}
}
event file_over_new_connection(f: fa_file, c: connection, is_orig: bool) &priority=5
{
if ( f$source == "HTTP" && c?$http )
{
if ( c$http?$current_entity && c$http$current_entity?$filename )
f$info$filename = c$http$current_entity$filename;
if ( f$is_orig )
{
if ( ! c$http?$orig_mime_types )
c$http$orig_fuids = string_vec(f$id);
else
c$http$orig_fuids[|c$http$orig_fuids|] = f$id;
if ( f?$mime_type )
{
if ( ! c$http?$orig_mime_types )
c$http$orig_mime_types = string_vec(f$mime_type);
else
c$http$orig_mime_types[|c$http$orig_mime_types|] = f$mime_type;
}
}
else
{
if ( ! c$http?$resp_mime_types )
c$http$resp_fuids = string_vec(f$id);
else
c$http$resp_fuids[|c$http$resp_fuids|] = f$id;
if ( f?$mime_type )
{
if ( ! c$http?$resp_mime_types )
c$http$resp_mime_types = string_vec(f$mime_type);
else
c$http$resp_mime_types[|c$http$resp_mime_types|] = f$mime_type;
}
}
}
}
event http_end_entity(c: connection, is_orig: bool) &priority=5
{
if ( c?$http && c$http?$current_entity )
delete c$http$current_entity;
}

View file

@ -1,54 +0,0 @@
@load ./main
@load ./utils
@load base/utils/conn-ids
@load base/frameworks/file-analysis/main
module HTTP;
export {
redef record HTTP::Info += {
## Number of MIME entities in the HTTP request message body so far.
request_mime_level: count &default=0;
## Number of MIME entities in the HTTP response message body so far.
response_mime_level: count &default=0;
};
## Default file handle provider for HTTP.
global get_file_handle: function(c: connection, is_orig: bool): string;
}
event http_begin_entity(c: connection, is_orig: bool) &priority=5
{
if ( ! c?$http )
return;
if ( is_orig )
++c$http$request_mime_level;
else
++c$http$response_mime_level;
}
function get_file_handle(c: connection, is_orig: bool): string
{
if ( ! c?$http ) return "";
local mime_level: count =
is_orig ? c$http$request_mime_level : c$http$response_mime_level;
local mime_level_str: string = mime_level > 1 ? cat(mime_level) : "";
if ( c$http$range_request )
return cat(Analyzer::ANALYZER_HTTP, " ", is_orig, " ", c$id$orig_h, " ",
build_url(c$http));
return cat(Analyzer::ANALYZER_HTTP, " ", c$start_time, " ", is_orig, " ",
c$http$trans_depth, mime_level_str, " ", id_string(c$id));
}
module GLOBAL;
event get_file_handle(tag: Analyzer::Tag, c: connection, is_orig: bool)
&priority=5
{
if ( tag != Analyzer::ANALYZER_HTTP ) return;
set_file_handle(HTTP::get_file_handle(c, is_orig));
}

View file

@ -1,100 +0,0 @@
##! Extracts the items from HTTP traffic, one per file. At this time only
##! the message body from the server can be extracted with this script.
@load ./main
@load ./file-analysis
module HTTP;
export {
## Pattern of file mime types to extract from HTTP response entity bodies.
const extract_file_types = /NO_DEFAULT/ &redef;
## The on-disk prefix for files to be extracted from HTTP entity bodies.
const extraction_prefix = "http-item" &redef;
redef record Info += {
## On-disk location where files in request body were extracted.
extracted_request_files: vector of string &log &optional;
## On-disk location where files in response body were extracted.
extracted_response_files: vector of string &log &optional;
## Indicates if the response body is to be extracted or not. Must be
## set before or by the first :bro:see:`file_new` for the file content.
extract_file: bool &default=F;
};
}
function get_extraction_name(f: fa_file): string
{
local r = fmt("%s-%s.dat", extraction_prefix, f$id);
return r;
}
function add_extraction_file(c: connection, is_orig: bool, fn: string)
{
if ( is_orig )
{
if ( ! c$http?$extracted_request_files )
c$http$extracted_request_files = vector();
c$http$extracted_request_files[|c$http$extracted_request_files|] = fn;
}
else
{
if ( ! c$http?$extracted_response_files )
c$http$extracted_response_files = vector();
c$http$extracted_response_files[|c$http$extracted_response_files|] = fn;
}
}
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "HTTP" ) return;
if ( ! f?$conns ) return;
local fname: string;
local c: connection;
if ( f?$mime_type && extract_file_types in f$mime_type )
{
fname = get_extraction_name(f);
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=fname]);
for ( cid in f$conns )
{
c = f$conns[cid];
if ( ! c?$http ) next;
add_extraction_file(c, f$is_orig, fname);
}
return;
}
local extracting: bool = F;
for ( cid in f$conns )
{
c = f$conns[cid];
if ( ! c?$http ) next;
if ( ! c$http$extract_file ) next;
fname = get_extraction_name(f);
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=fname]);
extracting = T;
break;
}
if ( extracting )
for ( cid in f$conns )
{
c = f$conns[cid];
if ( ! c?$http ) next;
add_extraction_file(c, f$is_orig, fname);
}
}

View file

@ -1,68 +0,0 @@
##! Calculate hashes for HTTP body transfers.
@load ./main
@load ./file-analysis
module HTTP;
export {
redef record Info += {
## MD5 sum for a file transferred over HTTP calculated from the
## response body.
md5: string &log &optional;
## This value can be set per-transfer to determine per request
## if a file should have an MD5 sum generated. It must be
## set to T at the time of or before the first chunk of body data.
calc_md5: bool &default=F;
};
## Generate MD5 sums for these filetypes.
const generate_md5 = /application\/x-dosexec/ # Windows and DOS executables
| /application\/x-executable/ # *NIX executable binary
&redef;
}
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "HTTP" ) return;
if ( f?$mime_type && generate_md5 in f$mime_type )
{
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_MD5]);
return;
}
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( ! c?$http ) next;
if ( ! c$http$calc_md5 ) next;
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_MD5]);
return;
}
}
event file_state_remove(f: fa_file) &priority=4
{
if ( ! f?$source ) return;
if ( f$source != "HTTP" ) return;
if ( ! f?$conns ) return;
if ( ! f?$info ) return;
if ( ! f$info?$md5 ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( ! c?$http ) next;
c$http$md5 = f$info$md5;
}
}

View file

@ -1,105 +0,0 @@
##! Identification of file types in HTTP response bodies with file content sniffing.
@load base/frameworks/notice
@load ./main
@load ./utils
@load ./file-analysis
module HTTP;
export {
redef enum Notice::Type += {
## Indicates when the file extension doesn't seem to match the file
## contents.
Incorrect_File_Type,
};
redef record Info += {
## Mime type of response body identified by content sniffing.
mime_type: string &log &optional;
};
## Mapping between mime type strings (without character set) and
## regular expressions for URLs.
## The :bro:enum:`HTTP::Incorrect_File_Type` notice is generated if the
## pattern doesn't match the mime type that was discovered.
const mime_types_extensions: table[string] of pattern = {
["application/x-dosexec"] = /\.([eE][xX][eE]|[dD][lL][lL])/,
} &redef;
## A pattern for filtering out :bro:enum:`HTTP::Incorrect_File_Type` urls
## that are not noteworthy before a notice is created. Each
## pattern added should match the complete URL (the matched URLs include
## "http://" at the beginning).
const ignored_incorrect_file_type_urls = /^$/ &redef;
}
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "HTTP" ) return;
if ( ! f?$mime_type ) return;
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( ! c?$http ) next;
c$http$mime_type = f$mime_type;
local mime_str: string = c$http$mime_type;
if ( mime_str !in mime_types_extensions ) next;
if ( ! c$http?$uri ) next;
if ( mime_types_extensions[mime_str] in c$http$uri ) next;
local url = build_url_http(c$http);
if ( url == ignored_incorrect_file_type_urls ) next;
local message = fmt("%s %s %s", mime_str, c$http$method, url);
NOTICE([$note=Incorrect_File_Type,
$msg=message,
$conn=c]);
}
}
event file_over_new_connection(f: fa_file, c: connection) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "HTTP" ) return;
if ( ! f?$mime_type ) return;
if ( ! c?$http ) return;
# Spread the mime around (e.g. for partial content, file_type event only
# happens once for the first connection, but if there's subsequent
# connections to transfer the same file, they'll be lacking the mime_type
# field if we don't do this).
c$http$mime_type = f$mime_type;
}
# Tracks byte-range request / partial content response mime types, indexed
# by [connection, uri] pairs. This is needed because a person can pipeline
# byte-range requests over multiple connections to the same uri. Without
# the tracking, only the first request in the pipeline for each connection
# would get a mime_type field assigned to it (by the FileAnalysis policy hooks).
global partial_types: table[conn_id, string] of string &read_expire=5mins;
# Priority 4 so that it runs before the handler that will write to http.log.
event http_message_done(c: connection, is_orig: bool, stat: http_message_stat)
&priority=4
{
if ( ! c$http$range_request ) return;
if ( ! c$http?$uri ) return;
if ( c$http?$mime_type )
{
partial_types[c$id, c$http$uri] = c$http$mime_type;
return;
}
if ( [c$id, c$http$uri] in partial_types )
c$http$mime_type = partial_types[c$id, c$http$uri];
}

View file

@ -0,0 +1,56 @@
@load ./main
@load ./entities
@load ./utils
@load base/utils/conn-ids
@load base/frameworks/files
module HTTP;
export {
## Default file handle provider for HTTP.
global get_file_handle: function(c: connection, is_orig: bool): string;
## Default file describer for HTTP.
global describe_file: function(f: fa_file): string;
}
function get_file_handle(c: connection, is_orig: bool): string
{
if ( ! c?$http )
return "";
if ( c$http$range_request && ! is_orig )
{
# Any multipart responses from the server are pieces of same file
# that correspond to range requests, so don't use mime depth to
# identify the file.
return cat(Analyzer::ANALYZER_HTTP, is_orig, c$id$orig_h, build_url(c$http));
}
else
{
local mime_depth = is_orig ? c$http$orig_mime_depth : c$http$resp_mime_depth;
return cat(Analyzer::ANALYZER_HTTP, c$start_time, is_orig,
c$http$trans_depth, mime_depth, id_string(c$id));
}
}
function describe_file(f: fa_file): string
{
# This shouldn't be needed, but just in case...
if ( f$source != "HTTP" )
return "";
for ( cid in f$conns )
{
if ( f$conns[cid]?$http )
return build_url_http(f$conns[cid]$http);
}
return "";
}
event bro_init() &priority=5
{
Files::register_protocol(Analyzer::ANALYZER_HTTP,
[$get_file_handle = HTTP::get_file_handle,
$describe = HTTP::describe_file]);
}

View file

@ -1,5 +1,5 @@
##! Implements base functionality for HTTP analysis. The logging model is
##! to log request/response pairs and all relevant metadata together in
##! Implements base functionality for HTTP analysis. The logging model is
##! to log request/response pairs and all relevant metadata together in
##! a single record.
@load base/utils/numbers
@ -15,10 +15,10 @@ export {
## Placeholder.
EMPTY
};
## This setting changes if passwords used in Basic-Auth are captured or not.
const default_capture_password = F &redef;
type Info: record {
## Timestamp for when the request happened.
ts: time &log;
@ -26,7 +26,7 @@ export {
uid: string &log;
## The connection's 4-tuple of endpoint addresses/ports.
id: conn_id &log;
## Represents the pipelined depth into the connection of this
## Represents the pipelined depth into the connection of this
## request/response transaction.
trans_depth: count &log;
## Verb used in the HTTP request (GET, POST, HEAD, etc.).
@ -60,24 +60,24 @@ export {
## A set of indicators of various attributes discovered and
## related to a particular request/response pair.
tags: set[Tags] &log;
## Username if basic-auth is performed for the request.
username: string &log &optional;
## Password if basic-auth is performed for the request.
password: string &log &optional;
## Determines if the password will be captured for this request.
capture_password: bool &default=default_capture_password;
## All of the headers that may indicate if the request was proxied.
proxied: set[string] &log &optional;
## Indicates if this request can assume 206 partial content in
## response.
range_request: bool &default=F;
range_request: bool &default=F;
};
## Structure to maintain state for an HTTP connection with multiple
## Structure to maintain state for an HTTP connection with multiple
## requests and responses.
type State: record {
## Pending requests.
@ -87,7 +87,7 @@ export {
## Current response in the pending queue.
current_response: count &default=0;
};
## A list of HTTP headers typically used to indicate proxied requests.
const proxy_headers: set[string] = {
"FORWARDED",
@ -100,8 +100,8 @@ export {
} &redef;
## A list of HTTP methods. Other methods will generate a weird. Note
## that the HTTP analyzer will only accept methods consisting solely
## of letters ``[A-Za-z]``.
## that the HTTP analyzer will only accept methods consisting solely
## of letters ``[A-Za-z]``.
const http_methods: set[string] = {
"GET", "POST", "HEAD", "OPTIONS",
"PUT", "DELETE", "TRACE", "CONNECT",
@ -111,8 +111,8 @@ export {
"POLL", "REPORT", "SUBSCRIBE", "BMOVE",
"SEARCH"
} &redef;
## Event that can be handled to access the HTTP record as it is sent on
## Event that can be handled to access the HTTP record as it is sent on
## to the logging framework.
global log_http: event(rec: Info);
}
@ -147,12 +147,12 @@ function new_http_session(c: connection): Info
tmp$ts=network_time();
tmp$uid=c$uid;
tmp$id=c$id;
# $current_request is set prior to the Info record creation so we
# $current_request is set prior to the Info record creation so we
# can use the value directly here.
tmp$trans_depth = c$http_state$current_request;
return tmp;
}
function set_state(c: connection, request: bool, is_orig: bool)
{
if ( ! c?$http_state )
@ -160,19 +160,19 @@ function set_state(c: connection, request: bool, is_orig: bool)
local s: State;
c$http_state = s;
}
# These deal with new requests and responses.
if ( request || c$http_state$current_request !in c$http_state$pending )
c$http_state$pending[c$http_state$current_request] = new_http_session(c);
if ( ! is_orig && c$http_state$current_response !in c$http_state$pending )
c$http_state$pending[c$http_state$current_response] = new_http_session(c);
if ( is_orig )
c$http = c$http_state$pending[c$http_state$current_request];
else
c$http = c$http_state$pending[c$http_state$current_response];
}
event http_request(c: connection, method: string, original_URI: string,
unescaped_URI: string, version: string) &priority=5
{
@ -181,17 +181,17 @@ event http_request(c: connection, method: string, original_URI: string,
local s: State;
c$http_state = s;
}
++c$http_state$current_request;
set_state(c, T, T);
c$http$method = method;
c$http$uri = unescaped_URI;
if ( method !in http_methods )
event conn_weird("unknown_HTTP_method", c, method);
}
event http_reply(c: connection, version: string, code: count, reason: string) &priority=5
{
if ( ! c?$http_state )
@ -199,7 +199,7 @@ event http_reply(c: connection, version: string, code: count, reason: string) &p
local s: State;
c$http_state = s;
}
# If the last response was an informational 1xx, we're still expecting
# the real response to the request, so don't create a new Info record yet.
if ( c$http_state$current_response !in c$http_state$pending ||
@ -207,7 +207,7 @@ event http_reply(c: connection, version: string, code: count, reason: string) &p
! code_in_range(c$http_state$pending[c$http_state$current_response]$status_code, 100, 199)) )
++c$http_state$current_response;
set_state(c, F, F);
c$http$status_code = code;
c$http$status_msg = reason;
if ( code_in_range(code, 100, 199) )
@ -216,33 +216,33 @@ event http_reply(c: connection, version: string, code: count, reason: string) &p
c$http$info_msg = reason;
}
}
event http_header(c: connection, is_orig: bool, name: string, value: string) &priority=5
{
set_state(c, F, is_orig);
if ( is_orig ) # client headers
{
if ( name == "REFERER" )
c$http$referrer = value;
else if ( name == "HOST" )
# The split is done to remove the occasional port value that shows up here.
c$http$host = split1(value, /:/)[1];
else if ( name == "RANGE" )
c$http$range_request = T;
else if ( name == "USER-AGENT" )
c$http$user_agent = value;
else if ( name in proxy_headers )
{
if ( ! c$http?$proxied )
c$http$proxied = set();
add c$http$proxied[fmt("%s -> %s", name, value)];
}
else if ( name == "AUTHORIZATION" )
{
if ( /^[bB][aA][sS][iI][cC] / in value )
@ -264,25 +264,19 @@ event http_header(c: connection, is_orig: bool, name: string, value: string) &pr
}
}
}
else # server headers
{
if ( name == "CONTENT-DISPOSITION" &&
/[fF][iI][lL][eE][nN][aA][mM][eE]/ in value )
c$http$filename = extract_filename_from_content_disposition(value);
}
}
event http_message_done(c: connection, is_orig: bool, stat: http_message_stat) &priority = 5
{
set_state(c, F, is_orig);
if ( is_orig )
c$http$request_body_len = stat$body_length;
else
c$http$response_body_len = stat$body_length;
}
event http_message_done(c: connection, is_orig: bool, stat: http_message_stat) &priority = -5
{
# The reply body is done so we're ready to log.
@ -311,4 +305,4 @@ event connection_state_remove(c: connection) &priority=-5
}
}
}

View file

@ -32,6 +32,9 @@ export {
##
## Returns: A URL prefixed with "http://".
global build_url_http: function(rec: Info): string;
## Create an extremely shortened representation of a log line.
global describe: function(rec: Info): string;
}
@ -62,3 +65,8 @@ function build_url_http(rec: Info): string
{
return fmt("http://%s", build_url(rec));
}
function describe(rec: Info): string
{
return build_url_http(rec);
}

View file

@ -1,5 +1,5 @@
@load ./main
@load ./dcc-send
@load ./file-analysis
@load ./files
@load-sigs ./dpd.sig

View file

@ -2,7 +2,7 @@
##!
##! There is a major problem with this script in the cluster context because
##! we might see A send B a message that a DCC connection is to be expected,
##! but that connection will actually be between B and C which could be
##! but that connection will actually be between B and C which could be
##! analyzed on a different worker.
##!
@ -15,12 +15,6 @@
module IRC;
export {
## Pattern of file mime types to extract from IRC DCC file transfers.
const extract_file_types = /NO_DEFAULT/ &redef;
## On-disk prefix for files to be extracted from IRC DCC file transfers.
const extraction_prefix = "irc-dcc-item" &redef;
redef record Info += {
## DCC filename requested.
dcc_file_name: string &log &optional;
@ -28,101 +22,10 @@ export {
dcc_file_size: count &log &optional;
## Sniffed mime type of the file.
dcc_mime_type: string &log &optional;
## The file handle for the file to be extracted
extraction_file: string &log &optional;
## A boolean to indicate if the current file transfer should be extracted.
extract_file: bool &default=F;
};
}
global dcc_expected_transfers: table[addr, port] of Info &read_expire=5mins;
function set_dcc_mime(f: fa_file)
{
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( [cid$resp_h, cid$resp_p] !in dcc_expected_transfers ) next;
local s = dcc_expected_transfers[cid$resp_h, cid$resp_p];
s$dcc_mime_type = f$mime_type;
}
}
function set_dcc_extraction_file(f: fa_file, filename: string)
{
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( [cid$resp_h, cid$resp_p] !in dcc_expected_transfers ) next;
local s = dcc_expected_transfers[cid$resp_h, cid$resp_p];
s$extraction_file = filename;
}
}
function get_extraction_name(f: fa_file): string
{
local r = fmt("%s-%s.dat", extraction_prefix, f$id);
return r;
}
# this handler sets the IRC::Info mime type
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "IRC_DATA" ) return;
if ( ! f?$mime_type ) return;
set_dcc_mime(f);
}
# this handler check if file extraction is desired
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "IRC_DATA" ) return;
local fname: string;
if ( f?$mime_type && extract_file_types in f$mime_type )
{
fname = get_extraction_name(f);
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=fname]);
set_dcc_extraction_file(f, fname);
return;
}
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( [cid$resp_h, cid$resp_p] !in dcc_expected_transfers ) next;
local s = dcc_expected_transfers[cid$resp_h, cid$resp_p];
if ( ! s$extract_file ) next;
fname = get_extraction_name(f);
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=fname]);
s$extraction_file = fname;
return;
}
}
global dcc_expected_transfers: table[addr, port] of Info &synchronized &read_expire=5mins;
function log_dcc(f: fa_file)
{
@ -141,24 +44,21 @@ function log_dcc(f: fa_file)
Log::write(IRC::LOG, irc);
irc$command = tmp;
# Delete these values in case another DCC transfer
# Delete these values in case another DCC transfer
# happens during the IRC session.
delete irc$extract_file;
delete irc$extraction_file;
delete irc$dcc_file_name;
delete irc$dcc_file_size;
delete irc$dcc_mime_type;
delete dcc_expected_transfers[cid$resp_h, cid$resp_p];
return;
}
}
event file_new(f: fa_file) &priority=-5
{
if ( ! f?$source ) return;
if ( f$source != "IRC_DATA" ) return;
log_dcc(f);
if ( f$source == "IRC_DATA" )
log_dcc(f);
}
event irc_dcc_message(c: connection, is_orig: bool,

View file

@ -1,25 +0,0 @@
@load ./dcc-send.bro
@load base/utils/conn-ids
@load base/frameworks/file-analysis/main
module IRC;
export {
## Default file handle provider for IRC.
global get_file_handle: function(c: connection, is_orig: bool): string;
}
function get_file_handle(c: connection, is_orig: bool): string
{
if ( is_orig ) return "";
return cat(Analyzer::ANALYZER_IRC_DATA, " ", c$start_time, " ", id_string(c$id));
}
module GLOBAL;
event get_file_handle(tag: Analyzer::Tag, c: connection, is_orig: bool)
&priority=5
{
if ( tag != Analyzer::ANALYZER_IRC_DATA ) return;
set_file_handle(IRC::get_file_handle(c, is_orig));
}

View file

@ -0,0 +1,39 @@
@load ./dcc-send
@load base/utils/conn-ids
@load base/frameworks/files
module IRC;
export {
redef record Info += {
## File unique ID.
fuid: string &log &optional;
};
## Default file handle provider for IRC.
global get_file_handle: function(c: connection, is_orig: bool): string;
}
function get_file_handle(c: connection, is_orig: bool): string
{
return cat(Analyzer::ANALYZER_IRC_DATA, c$start_time, c$id, is_orig);
}
event bro_init() &priority=5
{
Files::register_protocol(Analyzer::ANALYZER_IRC_DATA,
[$get_file_handle = IRC::get_file_handle]);
}
event file_over_new_connection(f: fa_file, c: connection, is_orig: bool) &priority=5
{
if ( [c$id$resp_h, c$id$resp_p] !in dcc_expected_transfers )
return;
local irc = dcc_expected_transfers[c$id$resp_h, c$id$resp_p];
irc$fuid = f$id;
if ( irc?$dcc_file_name )
f$info$filename = irc$dcc_file_name;
if ( f?$mime_type )
irc$dcc_mime_type = f$mime_type;
}

View file

@ -1,6 +1,5 @@
@load ./main
@load ./entities
@load ./entities-excerpt
@load ./file-analysis
@load ./files
@load-sigs ./dpd.sig

View file

@ -1,37 +0,0 @@
##! This script is for optionally adding a body excerpt to the SMTP
##! entities log.
@load ./entities
module SMTP;
export {
redef record SMTP::EntityInfo += {
## The entity body excerpt.
excerpt: string &log &default="";
};
## This is the default value for how much of the entity body should be
## included for all MIME entities. The lesser of this value and
## :bro:see:`default_file_bof_buffer_size` will be used.
const default_entity_excerpt_len = 0 &redef;
}
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "SMTP" ) return;
if ( ! f?$bof_buffer ) return;
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( ! c?$smtp ) next;
if ( default_entity_excerpt_len > 0 )
c$smtp$current_entity$excerpt =
f$bof_buffer[0:default_entity_excerpt_len];
}
}

View file

@ -1,5 +1,6 @@
##! Analysis and logging for MIME entities found in SMTP sessions.
@load base/frameworks/files
@load base/utils/strings
@load base/utils/files
@load ./main
@ -7,217 +8,55 @@
module SMTP;
export {
redef enum Log::ID += { ENTITIES_LOG };
type EntityInfo: record {
## This is the timestamp of when the MIME content transfer began.
ts: time &log;
uid: string &log;
id: conn_id &log;
## A count to represent the depth of this message transaction in a
## single connection where multiple messages were transferred.
trans_depth: count &log;
## The filename seen in the Content-Disposition header.
filename: string &log &optional;
## Track how many bytes of the MIME encoded file have been seen.
content_len: count &log &default=0;
## The mime type of the entity discovered through magic bytes identification.
mime_type: string &log &optional;
## The calculated MD5 sum for the MIME entity.
md5: string &log &optional;
## Optionally calculate the file's MD5 sum. Must be set prior to the
## first data chunk being see in an event.
calc_md5: bool &default=F;
## Optionally write the file to disk. Must be set prior to first
## data chunk being seen in an event.
extract_file: bool &default=F;
## Store the file handle here for the file currently being extracted.
extraction_file: string &log &optional;
type Entity: record {
## Filename for the entity if discovered from a header.
filename: string &optional;
};
redef record Info += {
## The in-progress entity information.
current_entity: EntityInfo &optional;
## The current entity being seen.
entity: Entity &optional;
};
redef record State += {
## Track the number of MIME encoded files transferred during a session.
mime_level: count &default=0;
## Track the number of MIME encoded files transferred
## during a session.
mime_depth: count &default=0;
};
## Generate MD5 sums for these filetypes.
const generate_md5 = /application\/x-dosexec/ # Windows and DOS executables
| /application\/x-executable/ # *NIX executable binary
&redef;
## Pattern of file mime types to extract from MIME bodies.
const extract_file_types = /NO_DEFAULT/ &redef;
## The on-disk prefix for files to be extracted from MIME entity bodies.
const extraction_prefix = "smtp-entity" &redef;
## If set, never generate MD5s. This is mainly for testing purposes to create
## reproducable output in the case that the decision whether to create
## checksums depends on environment specifics.
const never_calc_md5 = F &redef;
global log_mime: event(rec: EntityInfo);
}
event bro_init() &priority=5
{
Log::create_stream(SMTP::ENTITIES_LOG, [$columns=EntityInfo, $ev=log_mime]);
}
function set_session(c: connection, new_entity: bool)
{
if ( ! c$smtp?$current_entity || new_entity )
{
local info: EntityInfo;
info$ts=network_time();
info$uid=c$uid;
info$id=c$id;
info$trans_depth=c$smtp$trans_depth;
c$smtp$current_entity = info;
++c$smtp_state$mime_level;
}
}
function get_extraction_name(f: fa_file): string
{
local r = fmt("%s-%s.dat", extraction_prefix, f$id);
return r;
}
event mime_begin_entity(c: connection) &priority=10
{
if ( ! c?$smtp ) return;
set_session(c, T);
c$smtp$entity = Entity();
++c$smtp_state$mime_depth;
}
event file_new(f: fa_file) &priority=5
event file_over_new_connection(f: fa_file, c: connection, is_orig: bool) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "SMTP" ) return;
if ( ! f?$conns ) return;
local fname: string;
local extracting: bool = F;
for ( cid in f$conns )
if ( f$source == "SMTP" && c?$smtp )
{
local c: connection = f$conns[cid];
if ( ! c?$smtp ) next;
if ( ! c$smtp?$current_entity ) next;
if ( c$smtp$current_entity$extract_file )
{
if ( ! extracting )
{
fname = get_extraction_name(f);
FileAnalysis::add_analyzer(f,
[$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=fname]);
extracting = T;
}
c$smtp$current_entity$extraction_file = fname;
}
if ( c$smtp$current_entity$calc_md5 )
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_MD5]);
if ( c$smtp?$entity && c$smtp$entity?$filename )
f$info$filename = c$smtp$entity$filename;
f$info$depth = c$smtp_state$mime_depth;
}
}
function check_extract_by_type(f: fa_file)
event mime_one_header(c: connection, h: mime_header_rec) &priority=5
{
if ( extract_file_types !in f$mime_type ) return;
if ( f?$info && FileAnalysis::ANALYZER_EXTRACT in f$info$analyzers )
if ( ! c?$smtp )
return;
local fname: string = get_extraction_name(f);
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_EXTRACT,
$extract_filename=fname]);
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( ! c?$smtp ) next;
c$smtp$current_entity$extraction_file = fname;
}
}
function check_md5_by_type(f: fa_file)
{
if ( never_calc_md5 ) return;
if ( generate_md5 !in f$mime_type ) return;
FileAnalysis::add_analyzer(f, [$tag=FileAnalysis::ANALYZER_MD5]);
}
event file_new(f: fa_file) &priority=5
{
if ( ! f?$source ) return;
if ( f$source != "SMTP" ) return;
if ( ! f?$mime_type ) return;
if ( f?$conns )
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( ! c?$smtp ) next;
if ( ! c$smtp?$current_entity ) next;
c$smtp$current_entity$mime_type = f$mime_type;
}
check_extract_by_type(f);
check_md5_by_type(f);
}
event file_state_remove(f: fa_file) &priority=4
{
if ( ! f?$source ) return;
if ( f$source != "SMTP" ) return;
if ( ! f?$conns ) return;
for ( cid in f$conns )
{
local c: connection = f$conns[cid];
if ( ! c?$smtp ) next;
if ( ! c$smtp?$current_entity ) next;
# Only log if there was some content.
if ( f$seen_bytes == 0 ) next;
if ( f?$info && f$info?$md5 )
c$smtp$current_entity$md5 = f$info$md5;
c$smtp$current_entity$content_len = f$seen_bytes;
Log::write(SMTP::ENTITIES_LOG, c$smtp$current_entity);
delete c$smtp$current_entity;
return;
}
}
event mime_one_header(c: connection, h: mime_header_rec)
{
if ( ! c?$smtp ) return;
if ( h$name == "CONTENT-DISPOSITION" &&
/[fF][iI][lL][eE][nN][aA][mM][eE]/ in h$value )
c$smtp$current_entity$filename = extract_filename_from_content_disposition(h$value);
c$smtp$entity$filename = extract_filename_from_content_disposition(h$value);
if ( h$name == "CONTENT-TYPE" &&
/[nN][aA][mM][eE][:blank:]*=/ in h$value )
c$smtp$current_entity$filename = extract_filename_from_content_disposition(h$value);
c$smtp$entity$filename = extract_filename_from_content_disposition(h$value);
}
event mime_end_entity(c: connection) &priority=5
{
if ( c?$smtp && c$smtp?$entity )
delete c$smtp$entity;
}

View file

@ -1,27 +0,0 @@
@load ./main
@load ./entities
@load base/utils/conn-ids
@load base/frameworks/file-analysis/main
module SMTP;
export {
## Default file handle provider for SMTP.
global get_file_handle: function(c: connection, is_orig: bool): string;
}
function get_file_handle(c: connection, is_orig: bool): string
{
if ( ! c?$smtp ) return "";
return cat(Analyzer::ANALYZER_SMTP, " ", c$start_time, " ", c$smtp$trans_depth, " ",
c$smtp_state$mime_level);
}
module GLOBAL;
event get_file_handle(tag: Analyzer::Tag, c: connection, is_orig: bool)
&priority=5
{
if ( tag != Analyzer::ANALYZER_SMTP ) return;
set_file_handle(SMTP::get_file_handle(c, is_orig));
}

View file

@ -0,0 +1,53 @@
@load ./main
@load ./entities
@load base/utils/conn-ids
@load base/frameworks/files
module SMTP;
export {
redef record Info += {
## An ordered vector of file unique IDs seen attached to
## the message.
fuids: vector of string &log &default=string_vec();
};
## Default file handle provider for SMTP.
global get_file_handle: function(c: connection, is_orig: bool): string;
## Default file describer for SMTP.
global describe_file: function(f: fa_file): string;
}
function get_file_handle(c: connection, is_orig: bool): string
{
return cat(Analyzer::ANALYZER_SMTP, c$start_time, c$smtp$trans_depth,
c$smtp_state$mime_depth);
}
function describe_file(f: fa_file): string
{
# This shouldn't be needed, but just in case...
if ( f$source != "SMTP" )
return "";
for ( cid in f$conns )
{
local c = f$conns[cid];
return SMTP::describe(c$smtp);
}
return "";
}
event bro_init() &priority=5
{
Files::register_protocol(Analyzer::ANALYZER_SMTP,
[$get_file_handle = SMTP::get_file_handle,
$describe = SMTP::describe_file]);
}
event file_over_new_connection(f: fa_file, c: connection, is_orig: bool) &priority=5
{
if ( c?$smtp )
c$smtp$fuids[|c$smtp$fuids|] = f$id;
}

View file

@ -72,7 +72,10 @@ export {
## ALL_HOSTS - always capture the entire path.
## NO_HOSTS - never capture the path.
const mail_path_capture = ALL_HOSTS &redef;
## Create an extremely shortened representation of a log line.
global describe: function(rec: Info): string;
global log_smtp: event(rec: Info);
}
@ -271,3 +274,29 @@ event connection_state_remove(c: connection) &priority=-5
if ( c?$smtp )
smtp_message(c);
}
function describe(rec: Info): string
{
if ( rec?$mailfrom && rec?$rcptto )
{
local one_to = "";
for ( to in rec$rcptto )
{
one_to = to;
break;
}
local abbrev_subject = "";
if ( rec?$subject )
{
if ( |rec$subject| > 20 )
{
abbrev_subject = rec$subject[0:20] + "...";
}
}
return fmt("%s -> %s%s%s", rec$mailfrom, one_to,
(|rec$rcptto|>1 ? fmt(" (plus %d others)", |rec$rcptto|-1) : ""),
(abbrev_subject != "" ? fmt(": %s", abbrev_subject) : ""));
}
return "";
}