zeek/scripts/base/frameworks/file-analysis/main.bro

##! An interface for driving the analysis of files, possibly independent of
##! any network protocol over which they're transported.

@load base/bif/file_analysis.bif
@load base/frameworks/logging

module FileAnalysis;

export {
	redef enum Log::ID += {
		## Logging stream for file analysis.
		LOG
	};

	## A structure which represents a desired type of file analysis.
	type AnalyzerArgs: record {
		## The type of analysis.
		tag: FileAnalysis::Tag;

		## The local filename to which to write an extracted file.  Must be
		## set when *tag* is :bro:see:`FileAnalysis::ANALYZER_EXTRACT`.
		extract_filename: string &optional;

		## An event which will be generated for all new file contents,
		## chunk-wise.  Used when *tag* is
		## :bro:see:`FileAnalysis::ANALYZER_DATA_EVENT`.
		chunk_event: event(f: fa_file, data: string, off: count) &optional;

		## An event which will be generated for all new file contents,
		## stream-wise.  Used when *tag* is
		## :bro:see:`FileAnalysis::ANALYZER_DATA_EVENT`.
		stream_event: event(f: fa_file, data: string) &optional;
	} &redef;

	## Contains all metadata related to the analysis of a given file.
	## For the most part, fields here are derived from ones of the same name
	## in :bro:see:`fa_file`.
	type Info: record {
		## An identifier associated with a single file.
		id: string &log;

		## Identifier associated with a container file from which this one was
		## extracted as part of the file analysis.
		parent_id: string &log &optional;

		## An identification of the source of the file data.  E.g. it may be
		## a network protocol over which it was transferred, or a local file
		## path which was read, or some other input source.
		source: string &log &optional;

		## If the source of this file is is a network connection, this field
		## may be set to indicate the directionality.
		is_orig: bool &log &optional;

		## The time at which the last activity for the file was seen.
		last_active: time &log;

		## Number of bytes provided to the file analysis engine for the file.
		seen_bytes: count &log &default=0;

		## Total number of bytes that are supposed to comprise the full file.
		total_bytes: count &log &optional;

		## The number of bytes in the file stream that were completely missed
		## during the process of analysis e.g. due to dropped packets.
		missing_bytes: count &log &default=0;

		## The number of not all-in-sequence bytes in the file stream that
		## were delivered to file analyzers due to reassembly buffer overflow.
		overflow_bytes: count &log &default=0;

		## The amount of time between receiving new data for this file that
		## the analysis engine will wait before giving up on it.
		timeout_interval: interval &log &optional;

		## The number of bytes at the beginning of a file to save for later
		## inspection in *bof_buffer* field.
		bof_buffer_size: count &log &optional;

		## A mime type provided by libmagic against the *bof_buffer*, or
		## in the cases where no buffering of the beginning of file occurs,
		## an initial guess of the mime type based on the first data seen.
		mime_type: string &log &optional;

		## Whether the file analysis timed out at least once for the file.
		timedout: bool &log &default=F;

		## Connection UIDS over which the file was transferred.
		conn_uids: set[string] &log;

		## A set of analysis types done during the file analysis.
		analyzers: set[FileAnalysis::Tag];

		## Local filenames of extracted files.
		extracted_files: set[string] &log;

		## An MD5 digest of the file contents.
		md5: string &log &optional;

		## A SHA1 digest of the file contents.
		sha1: string &log &optional;

		## A SHA256 digest of the file contents.
		sha256: string &log &optional;
	} &redef;

	## A table that can be used to disable file analysis completely for
	## any files transferred over given network protocol analyzers.
	const disable: table[Analyzer::Tag] of bool = table() &redef;

	## Event that can be handled to access the Info record as it is sent on
	## to the logging framework.
	global log_file_analysis: event(rec: Info);

	## The salt concatenated to unique file handle strings generated by
	## :bro:see:`get_file_handle` before hashing them in to a file id
	## (the *id* field of :bro:see:`fa_file`).
	## Provided to help mitigate the possiblility of manipulating parts of
	## network connections that factor in to the file handle in order to
	## generate two handles that would hash to the same file id.
	const salt = "I recommend changing this." &redef;

	## Sets the *timeout_interval* field of :bro:see:`fa_file`, which is
	## used to determine the length of inactivity that is allowed for a file
	## before internal state related to it is cleaned up.  When used within a
	## :bro:see:`file_timeout` handler, the analysis will delay timing out
	## again for the period specified by *t*.
	##
	## f: the file.
	##
	## t: the amount of time the file can remain inactive before discarding.
	##
	## Returns: true if the timeout interval was set, or false if analysis
	##          for the *id* isn't currently active.
	global set_timeout_interval: function(f: fa_file, t: interval): bool;

	## Adds an analyzer to the analysis of a given file.
	##
	## f: the file.
	##
	## args: the analyzer type to add along with any arguments it takes.
	##
	## Returns: true if the analyzer will be added, or false if analysis
	##          for the *id* isn't currently active or the *args*
	##          were invalid for the analyzer type.
	global add_analyzer: function(f: fa_file, args: AnalyzerArgs): bool;

	## Removes an analyzer from the analysis of a given file.
	##
	## f: the file.
	##
	## args: the analyzer (type and args) to remove.
	##
	## Returns: true if the analyzer will be removed, or false if analysis
	##          for the *id* isn't currently active.
	global remove_analyzer: function(f: fa_file, args: AnalyzerArgs): bool;

	## Stops/ignores any further analysis of a given file.
	##
	## f: the file.
	##
	## Returns: true if analysis for the given file will be ignored for the
	##          rest of it's contents, or false if analysis for the *id*
	##          isn't currently active.
	global stop: function(f: fa_file): bool;
}

redef record fa_file += {
	info: Info &optional;
};

function set_info(f: fa_file)
	{
	if ( ! f?$info )
		{
		local tmp: Info;
		f$info = tmp;
		}

	f$info$id = f$id;
	if ( f?$parent_id ) f$info$parent_id = f$parent_id;
	if ( f?$source ) f$info$source = f$source;
	if ( f?$is_orig ) f$info$is_orig = f$is_orig;
	f$info$last_active = f$last_active;
	f$info$seen_bytes = f$seen_bytes;
	if ( f?$total_bytes ) f$info$total_bytes = f$total_bytes;
	f$info$missing_bytes = f$missing_bytes;
	f$info$overflow_bytes = f$overflow_bytes;
	f$info$timeout_interval = f$timeout_interval;
	f$info$bof_buffer_size = f$bof_buffer_size;
	if ( f?$mime_type ) f$info$mime_type = f$mime_type;
	if ( f?$conns )
		for ( cid in f$conns )
			add f$info$conn_uids[f$conns[cid]$uid];
	}

function set_timeout_interval(f: fa_file, t: interval): bool
	{
	return __set_timeout_interval(f$id, t);
	}

function add_analyzer(f: fa_file, args: AnalyzerArgs): bool
	{
	if ( ! __add_analyzer(f$id, args) ) return F;

	set_info(f);
	add f$info$analyzers[args$tag];

	if ( args$tag == FileAnalysis::ANALYZER_EXTRACT )
		add f$info$extracted_files[args$extract_filename];

	return T;
	}

function remove_analyzer(f: fa_file, args: AnalyzerArgs): bool
	{
	return __remove_analyzer(f$id, args);
	}

function stop(f: fa_file): bool
	{
	return __stop(f$id);
	}

event bro_init() &priority=5
	{
	Log::create_stream(FileAnalysis::LOG,
	                   [$columns=Info, $ev=log_file_analysis]);
	}

event file_timeout(f: fa_file) &priority=5
	{
	set_info(f);
	f$info$timedout = T;
	}

event file_hash(f: fa_file, kind: string, hash: string) &priority=5
	{
	set_info(f);
	switch ( kind ) {
	case "md5":
		f$info$md5 = hash;
		break;
	case "sha1":
		f$info$sha1 = hash;
		break;
	case "sha256":
		f$info$sha256 = hash;
		break;
	}
	}

event file_state_remove(f: fa_file) &priority=5
	{
	set_info(f);
	}

event file_state_remove(f: fa_file) &priority=-5
	{
	Log::write(FileAnalysis::LOG, f$info);
	}