You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
suricata/src/app-layer-detect-proto.c

3635 lines
124 KiB
C

App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/* Copyright (C) 2007-2014 Open Information Security Foundation
*
* You can copy, redistribute or modify this Program under the terms of
* the GNU General Public License version 2 as published by the Free
* Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
* 02110-1301, USA.
*/
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/**
* \file
*
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
* \author Victor Julien <victor@inliniac.net>
* \author Anoop Saldanha <anoopsaldanha@gmail.com>
*/
#include "suricata-common.h"
#include "debug.h"
#include "decode.h"
#include "threads.h"
#include "threadvars.h"
#include "tm-threads.h"
#include "detect.h"
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
#include "detect-engine-port.h"
#include "detect-parse.h"
#include "detect-engine.h"
#include "detect-content.h"
#include "detect-engine-mpm.h"
#include "detect-engine-state.h"
#include "util-print.h"
#include "util-pool.h"
#include "util-unittest.h"
#include "util-unittest-helper.h"
#include "flow.h"
#include "flow-util.h"
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
#include "flow-private.h"
#include "stream-tcp-private.h"
#include "stream-tcp-reassemble.h"
#include "stream-tcp.h"
#include "stream.h"
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
#include "app-layer.h"
#include "app-layer-protos.h"
#include "app-layer-parser.h"
#include "app-layer-detect-proto.h"
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
#include "conf.h"
#include "util-memcmp.h"
#include "util-spm.h"
#include "util-cuda.h"
#include "util-debug.h"
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
#include "runmodes.h"
typedef struct AppLayerProtoDetectProbingParserElement_ {
AppProto alproto;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/* \todo don't really need it. See if you can get rid of it */
uint16_t port;
/* \todo calculate at runtime and get rid of this var */
uint32_t alproto_mask;
/* \todo check if we can reduce the bottom 2 vars to uint16_t */
/* the min length of data that has to be supplied to invoke the parser */
uint32_t min_depth;
/* the max length of data after which this parser won't be invoked */
uint32_t max_depth;
/* the probing parser function */
ProbingParserFPtr ProbingParser;
struct AppLayerProtoDetectProbingParserElement_ *next;
} AppLayerProtoDetectProbingParserElement;
typedef struct AppLayerProtoDetectProbingParserPort_ {
/* the port no for which probing parser(s) are invoked */
uint16_t port;
uint32_t toserver_alproto_mask;
uint32_t toclient_alproto_mask;
/* the max depth for all the probing parsers registered for this port */
uint16_t toserver_max_depth;
uint16_t toclient_max_depth;
AppLayerProtoDetectProbingParserElement *toserver;
AppLayerProtoDetectProbingParserElement *toclient;
struct AppLayerProtoDetectProbingParserPort_ *next;
} AppLayerProtoDetectProbingParserPort;
typedef struct AppLayerProtoDetectProbingParser_ {
uint16_t ip_proto;
AppLayerProtoDetectProbingParserPort *port;
struct AppLayerProtoDetectProbingParser_ *next;
} AppLayerProtoDetectProbingParser;
typedef struct AppLayerProtoDetectPMSignature_ {
AppProto alproto;
/* \todo Change this into a non-pointer */
DetectContentData *cd;
struct AppLayerProtoDetectPMSignature_ *next;
} AppLayerProtoDetectPMSignature;
typedef struct AppLayerProtoDetectPMCtx_ {
uint16_t max_len;
uint16_t min_len;
MpmCtx mpm_ctx;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/** Mapping between pattern id and signature. As each signature has a
* unique pattern with a unique id, we can lookup the signature by
* the pattern id. */
AppLayerProtoDetectPMSignature **map;
AppLayerProtoDetectPMSignature *head;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/* \todo we don't need this except at setup time. Get rid of it. */
PatIntId max_pat_id;
} AppLayerProtoDetectPMCtx;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
typedef struct AppLayerProtoDetectCtxIpproto_ {
/* 0 - toserver, 1 - toclient */
AppLayerProtoDetectPMCtx ctx_pm[2];
} AppLayerProtoDetectCtxIpproto;
/**
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
* \brief The app layer protocol detection context.
*/
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
typedef struct AppLayerProtoDetectCtx_ {
/* Context per ip_proto.
* \todo Modify ctx_ipp to hold for only tcp and udp. The rest can be
* implemented if needed. Waste of space otherwise. */
AppLayerProtoDetectCtxIpproto ctx_ipp[FLOW_PROTO_DEFAULT];
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParser *ctx_pp;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/* Indicates the protocols that have registered themselves
* for protocol detection. This table is independent of the
* ipproto. */
char *alproto_names[ALPROTO_MAX];
} AppLayerProtoDetectCtx;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/**
* \brief The app layer protocol detection thread context.
*/
typedef struct AppLayerProtoDetectThreadCtx_ {
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
PatternMatcherQueue pmq;
/* The value 2 is for direction(0 - toserver, 1 - toclient). */
MpmThreadCtx mpm_tctx[FLOW_PROTO_DEFAULT][2];
} AppLayerProtoDetectThreadCtx;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/* The global app layer proto detection context. */
AppLayerProtoDetectCtx alpd_ctx;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/***** Static Internal Calls: Protocol Retrieval *****/
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static uint16_t AppLayerProtoDetectPMMatchSignature(AppLayerProtoDetectPMSignature *s,
uint8_t *buf, uint16_t buflen,
uint8_t ipproto)
{
SCEnter();
uint16_t proto = ALPROTO_UNKNOWN;
uint8_t *found = NULL;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (s->cd->offset > buflen) {
SCLogDebug("s->co->offset (%"PRIu16") > buflen (%"PRIu16")",
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
s->cd->offset, buflen);
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (s->cd->depth > buflen) {
SCLogDebug("s->co->depth (%"PRIu16") > buflen (%"PRIu16")",
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
s->cd->depth, buflen);
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint8_t *sbuf = buf + s->cd->offset;
uint16_t sbuflen = s->cd->depth - s->cd->offset;
SCLogDebug("s->co->offset (%"PRIu16") s->cd->depth (%"PRIu16")",
s->cd->offset, s->cd->depth);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (s->cd->flags & DETECT_CONTENT_NOCASE)
found = SpmNocaseSearch(sbuf, sbuflen, s->cd->content, s->cd->content_len);
else
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
found = SpmSearch(sbuf, sbuflen, s->cd->content, s->cd->content_len);
if (found != NULL)
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
proto = s->alproto;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
SCReturnInt(proto);
16 years ago
}
static uint16_t AppLayerProtoDetectPMGetProto(AppLayerProtoDetectThreadCtx *tctx,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
Flow *f,
uint8_t *buf, uint16_t buflen,
uint8_t direction,
uint8_t ipproto,
AppProto *pm_results)
{
SCEnter();
pm_results[0] = ALPROTO_UNKNOWN;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPMCtx *pm_ctx;
MpmThreadCtx *mpm_tctx;
uint16_t pm_matches = 0;
uint8_t cnt;
uint16_t searchlen;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (direction & STREAM_TOSERVER) {
pm_ctx = &alpd_ctx.ctx_ipp[FlowGetProtoMapping(ipproto)].ctx_pm[0];
mpm_tctx = &tctx->mpm_tctx[FlowGetProtoMapping(ipproto)][0];
} else {
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
pm_ctx = &alpd_ctx.ctx_ipp[FlowGetProtoMapping(ipproto)].ctx_pm[1];
mpm_tctx = &tctx->mpm_tctx[FlowGetProtoMapping(ipproto)][1];
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (pm_ctx->mpm_ctx.pattern_cnt == 0)
goto end;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
searchlen = buflen;
if (searchlen > pm_ctx->max_len)
searchlen = pm_ctx->max_len;
uint32_t search_cnt = 0;
/* do the mpm search */
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
search_cnt = mpm_table[pm_ctx->mpm_ctx.mpm_type].Search(&pm_ctx->mpm_ctx,
mpm_tctx,
&tctx->pmq,
buf, searchlen);
if (search_cnt == 0)
goto end;
/* alproto bit field */
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint8_t pm_results_bf[(ALPROTO_MAX / 8) + 1];
memset(pm_results_bf, 0, sizeof(pm_results_bf));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
for (cnt = 0; cnt < search_cnt; cnt++) {
AppLayerProtoDetectPMSignature *s = pm_ctx->map[tctx->pmq.pattern_id_array[cnt]];
while (s != NULL) {
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint16_t proto = AppLayerProtoDetectPMMatchSignature(s, buf, searchlen, ipproto);
if (proto != ALPROTO_UNKNOWN &&
!(pm_results_bf[proto / 8] & (1 << (proto % 8))) )
{
pm_results[pm_matches++] = proto;
pm_results_bf[proto / 8] |= 1 << (proto % 8);
}
s = s->next;
}
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
PmqReset(&tctx->pmq);
if (buflen >= pm_ctx->max_len)
FLOW_SET_PM_DONE(f, direction);
SCReturnUInt(pm_matches);
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static AppLayerProtoDetectProbingParserPort *AppLayerProtoDetectGetProbingParsers(AppLayerProtoDetectProbingParser *pp,
uint16_t ip_proto,
uint16_t port)
{
AppLayerProtoDetectProbingParserPort *pp_port = NULL;
while (pp != NULL) {
if (pp->ip_proto == ip_proto)
16 years ago
break;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
pp = pp->next;
}
if (pp == NULL)
goto end;
pp_port = pp->port;
while (pp_port != NULL) {
if (pp_port->port == port || pp_port->port == 0) {
break;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
}
pp_port = pp_port->next;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
SCReturnPtr(pp_port, "AppLayerProtoDetectProbingParserPort *");
}
/**
* \brief Call the probing parser if it exists for this src or dst port.
*/
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static AppProto AppLayerProtoDetectPPGetProto(Flow *f,
uint8_t *buf, uint32_t buflen,
uint8_t ipproto, uint8_t direction)
{
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParserPort *pp_port = NULL;
AppLayerProtoDetectProbingParserElement *pe = NULL;
AppProto alproto = ALPROTO_UNKNOWN;
uint32_t *alproto_masks;
if (direction & STREAM_TOSERVER) {
pp_port = AppLayerProtoDetectGetProbingParsers(alpd_ctx.ctx_pp, ipproto, f->dp);
alproto_masks = &f->probing_parser_toserver_alproto_masks;
if (pp_port == NULL) {
SCLogDebug("toserver-No probing parser registered for port %"PRIu16,
f->dp);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
FLOW_SET_PP_DONE(f, direction);
goto end;
}
pe = pp_port->toserver;
} else {
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
pp_port = AppLayerProtoDetectGetProbingParsers(alpd_ctx.ctx_pp, ipproto, f->sp);
alproto_masks = &f->probing_parser_toclient_alproto_masks;
if (pp_port == NULL) {
SCLogDebug("toclient-No probing parser registered for port %"PRIu16,
f->sp);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
FLOW_SET_PP_DONE(f, direction);
goto end;
}
pe = pp_port->toclient;
}
while (pe != NULL) {
if ((buflen < pe->min_depth) ||
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
(alproto_masks[0] & pe->alproto_mask)) {
pe = pe->next;
continue;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
alproto = pe->ProbingParser(buf, buflen, NULL);
if (alproto != ALPROTO_UNKNOWN && alproto != ALPROTO_FAILED)
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
goto end;
if (alproto == ALPROTO_FAILED ||
(pe->max_depth != 0 && buflen > pe->max_depth)) {
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
alproto_masks[0] |= pe->alproto_mask;
}
pe = pe->next;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (direction & STREAM_TOSERVER) {
if (alproto_masks[0] == pp_port->toserver_alproto_mask) {
FLOW_SET_PP_DONE(f, direction);
goto end;
}
} else {
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alproto_masks[0] == pp_port->toclient_alproto_mask) {
FLOW_SET_PP_DONE(f, direction);
goto end;
}
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
SCReturnCT(alproto, "AppProto");
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/***** Static Internal Calls: PP registration *****/
static void AppLayerProtoDetectPPGetIpprotos(AppProto alproto,
uint8_t *ipprotos)
{
SCEnter();
AppLayerProtoDetectProbingParser *pp;
AppLayerProtoDetectProbingParserPort *pp_port;
AppLayerProtoDetectProbingParserElement *pp_pe;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
for (pp = alpd_ctx.ctx_pp; pp != NULL; pp = pp->next) {
for (pp_port = pp->port; pp_port != NULL; pp_port = pp_port->next) {
for (pp_pe = pp_port->toserver; pp_pe != NULL; pp_pe = pp_pe->next) {
if (alproto == pp_pe->alproto)
ipprotos[pp->ip_proto / 8] |= 1 << (pp->ip_proto % 8);
}
for (pp_pe = pp_port->toclient; pp_pe != NULL; pp_pe = pp_pe->next) {
if (alproto == pp_pe->alproto)
ipprotos[pp->ip_proto / 8] |= 1 << (pp->ip_proto % 8);
}
}
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturn;
}
static uint32_t AppLayerProtoDetectProbingParserGetMask(AppProto alproto)
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (!(alproto > ALPROTO_UNKNOWN && alproto < ALPROTO_FAILED)) {
SCLogError(SC_ERR_ALPARSER, "Unknown protocol detected - %"PRIu16,
alproto);
exit(EXIT_FAILURE);
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturnUInt(1 << alproto);
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline AppLayerProtoDetectProbingParserElement *AllocAppLayerProtoDetectProbingParserElement(void)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParserElement *p = SCMalloc(sizeof(AppLayerProtoDetectProbingParserElement));
if (unlikely(p == NULL)) {
exit(EXIT_FAILURE);
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(p, 0, sizeof(AppLayerProtoDetectProbingParserElement));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturnPtr(p, "AppLayerProtoDetectProbingParserElement");
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline void DeAllocAppLayerProtoDetectProbingParserElement(AppLayerProtoDetectProbingParserElement *p)
{
SCEnter();
SCFree(p);
SCReturn;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline AppLayerProtoDetectProbingParserPort *AllocAppLayerProtoDetectProbingParserPort(void)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParserPort *p = SCMalloc(sizeof(AppLayerProtoDetectProbingParserPort));
if (unlikely(p == NULL)) {
exit(EXIT_FAILURE);
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(p, 0, sizeof(AppLayerProtoDetectProbingParserPort));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturnPtr(p, "AppLayerProtoDetectProbingParserPort");
}
static inline void DeAllocAppLayerProtoDetectProbingParserPort(AppLayerProtoDetectProbingParserPort *p)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParserElement *e;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
e = p->toserver;
while (e != NULL) {
AppLayerProtoDetectProbingParserElement *e_next = e->next;
DeAllocAppLayerProtoDetectProbingParserElement(e);
e = e_next;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
e = p->toclient;
while (e != NULL) {
AppLayerProtoDetectProbingParserElement *e_next = e->next;
DeAllocAppLayerProtoDetectProbingParserElement(e);
e = e_next;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCFree(p);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturn;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline AppLayerProtoDetectProbingParser *AllocAppLayerProtoDetectProbingParser(void)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParser *p = SCMalloc(sizeof(AppLayerProtoDetectProbingParser));
if (unlikely(p == NULL)) {
exit(EXIT_FAILURE);
}
memset(p, 0, sizeof(AppLayerProtoDetectProbingParser));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturnPtr(p, "AppLayerProtoDetectProbingParser");
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline void DeAllocAppLayerProtoDetectProbingParser(AppLayerProtoDetectProbingParser *p)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParserPort *pt = p->port;
while (pt != NULL) {
AppLayerProtoDetectProbingParserPort *pt_next = pt->next;
DeAllocAppLayerProtoDetectProbingParserPort(pt);
pt = pt_next;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCFree(p);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturn;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static AppLayerProtoDetectProbingParserElement *
AppLayerProtoDetectCreateAppLayerProtoDetectProbingParserElement(AppProto alproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint16_t port,
uint16_t min_depth,
uint16_t max_depth,
uint16_t (*AppLayerProtoDetectProbingParser)
(uint8_t *input, uint32_t input_len, uint32_t *offset))
{
AppLayerProtoDetectProbingParserElement *pe = AllocAppLayerProtoDetectProbingParserElement();
pe->alproto = alproto;
pe->port = port;
pe->alproto_mask = AppLayerProtoDetectProbingParserGetMask(alproto);
pe->min_depth = min_depth;
pe->max_depth = max_depth;
pe->ProbingParser = AppLayerProtoDetectProbingParser;
pe->next = NULL;
if (max_depth != 0 && min_depth >= max_depth) {
SCLogError(SC_ERR_ALPARSER, "Invalid arguments sent to "
"register the probing parser. min_depth >= max_depth");
goto error;
}
if (alproto <= ALPROTO_UNKNOWN || alproto >= ALPROTO_MAX) {
SCLogError(SC_ERR_ALPARSER, "Invalid arguments sent to register "
"the probing parser. Invalid alproto - %d", alproto);
goto error;
}
if (AppLayerProtoDetectProbingParser == NULL) {
SCLogError(SC_ERR_ALPARSER, "Invalid arguments sent to "
"register the probing parser. Probing parser func NULL");
goto error;
}
SCReturnPtr(pe, "AppLayerProtoDetectProbingParserElement");
error:
DeAllocAppLayerProtoDetectProbingParserElement(pe);
SCReturnPtr(NULL, "AppLayerProtoDetectProbingParserElement");
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static AppLayerProtoDetectProbingParserElement *
DuplicateAppLayerProtoDetectProbingParserElement(AppLayerProtoDetectProbingParserElement *pe)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParserElement *new_pe = AllocAppLayerProtoDetectProbingParserElement();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
new_pe->alproto = pe->alproto;
new_pe->port = pe->port;
new_pe->alproto_mask = pe->alproto_mask;
new_pe->min_depth = pe->min_depth;
new_pe->max_depth = pe->max_depth;
new_pe->ProbingParser = pe->ProbingParser;
new_pe->next = NULL;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturnPtr(new_pe, "AppLayerProtoDetectProbingParserElement");
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
void AppLayerProtoDetectPrintProbingParsers(AppLayerProtoDetectProbingParser *pp)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectProbingParserPort *pp_port = NULL;
AppLayerProtoDetectProbingParserElement *pp_pe = NULL;
printf("\n");
for ( ; pp != NULL; pp = pp->next) {
/* print ip protocol */
if (pp->ip_proto == IPPROTO_TCP)
printf("IPProto: TCP\n");
else if (pp->ip_proto == IPPROTO_UDP)
printf("IPProto: UDP\n");
else
printf("IPProto: %"PRIu16"\n", pp->ip_proto);
pp_port = pp->port;
for ( ; pp_port != NULL; pp_port = pp_port->next) {
if (pp_port->toserver == NULL)
goto AppLayerProtoDetectPrintProbingParsers_jump_toclient;
printf(" Port: %"PRIu16 "\n", pp_port->port);
printf(" To_Server: (max-depth: %"PRIu16 ", "
"mask - %"PRIu32")\n",
pp_port->toserver_max_depth,
pp_port->toserver_alproto_mask);
pp_pe = pp_port->toserver;
for ( ; pp_pe != NULL; pp_pe = pp_pe->next) {
if (pp_pe->alproto == ALPROTO_HTTP)
printf(" alproto: ALPROTO_HTTP\n");
else if (pp_pe->alproto == ALPROTO_FTP)
printf(" alproto: ALPROTO_FTP\n");
else if (pp_pe->alproto == ALPROTO_SMTP)
printf(" alproto: ALPROTO_SMTP\n");
else if (pp_pe->alproto == ALPROTO_TLS)
printf(" alproto: ALPROTO_TLS\n");
else if (pp_pe->alproto == ALPROTO_SSH)
printf(" alproto: ALPROTO_SSH\n");
else if (pp_pe->alproto == ALPROTO_IMAP)
printf(" alproto: ALPROTO_IMAP\n");
else if (pp_pe->alproto == ALPROTO_MSN)
printf(" alproto: ALPROTO_MSN\n");
else if (pp_pe->alproto == ALPROTO_JABBER)
printf(" alproto: ALPROTO_JABBER\n");
else if (pp_pe->alproto == ALPROTO_SMB)
printf(" alproto: ALPROTO_SMB\n");
else if (pp_pe->alproto == ALPROTO_SMB2)
printf(" alproto: ALPROTO_SMB2\n");
else if (pp_pe->alproto == ALPROTO_DCERPC)
printf(" alproto: ALPROTO_DCERPC\n");
else if (pp_pe->alproto == ALPROTO_IRC)
printf(" alproto: ALPROTO_IRC\n");
else
printf("impossible\n");
printf(" port: %"PRIu16 "\n", pp_pe->port);
printf(" mask: %"PRIu32 "\n", pp_pe->alproto_mask);
printf(" min_depth: %"PRIu32 "\n", pp_pe->min_depth);
printf(" max_depth: %"PRIu32 "\n", pp_pe->max_depth);
printf("\n");
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPrintProbingParsers_jump_toclient:
if (pp_port->toclient == NULL) {
continue;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
printf(" To_Client: (max-depth: %"PRIu16 ", "
"mask - %"PRIu32")\n",
pp_port->toclient_max_depth,
pp_port->toclient_alproto_mask);
pp_pe = pp_port->toclient;
for ( ; pp_pe != NULL; pp_pe = pp_pe->next) {
if (pp_pe->alproto == ALPROTO_HTTP)
printf(" alproto: ALPROTO_HTTP\n");
else if (pp_pe->alproto == ALPROTO_FTP)
printf(" alproto: ALPROTO_FTP\n");
else if (pp_pe->alproto == ALPROTO_SMTP)
printf(" alproto: ALPROTO_SMTP\n");
else if (pp_pe->alproto == ALPROTO_TLS)
printf(" alproto: ALPROTO_TLS\n");
else if (pp_pe->alproto == ALPROTO_SSH)
printf(" alproto: ALPROTO_SSH\n");
else if (pp_pe->alproto == ALPROTO_IMAP)
printf(" alproto: ALPROTO_IMAP\n");
else if (pp_pe->alproto == ALPROTO_MSN)
printf(" alproto: ALPROTO_MSN\n");
else if (pp_pe->alproto == ALPROTO_JABBER)
printf(" alproto: ALPROTO_JABBER\n");
else if (pp_pe->alproto == ALPROTO_SMB)
printf(" alproto: ALPROTO_SMB\n");
else if (pp_pe->alproto == ALPROTO_SMB2)
printf(" alproto: ALPROTO_SMB2\n");
else if (pp_pe->alproto == ALPROTO_DCERPC)
printf(" alproto: ALPROTO_DCERPC\n");
else if (pp_pe->alproto == ALPROTO_IRC)
printf(" alproto: ALPROTO_IRC\n");
else
printf("impossible\n");
printf(" port: %"PRIu16 "\n", pp_pe->port);
printf(" mask: %"PRIu32 "\n", pp_pe->alproto_mask);
printf(" min_depth: %"PRIu32 "\n", pp_pe->min_depth);
printf(" max_depth: %"PRIu32 "\n", pp_pe->max_depth);
printf("\n");
}
}
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCReturn;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline void AppendAppLayerProtoDetectProbingParserElement(AppLayerProtoDetectProbingParserElement **head_pe,
AppLayerProtoDetectProbingParserElement *new_pe)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (*head_pe == NULL) {
*head_pe = new_pe;
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if ((*head_pe)->port == 0) {
if (new_pe->port != 0) {
new_pe->next = *head_pe;
*head_pe = new_pe;
} else {
AppLayerProtoDetectProbingParserElement *temp_pe = *head_pe;
while (temp_pe->next != NULL)
temp_pe = temp_pe->next;
temp_pe->next = new_pe;
}
} else {
AppLayerProtoDetectProbingParserElement *temp_pe = *head_pe;
if (new_pe->port == 0) {
while (temp_pe->next != NULL)
temp_pe = temp_pe->next;
temp_pe->next = new_pe;
} else {
while (temp_pe->next != NULL && temp_pe->next->port != 0)
temp_pe = temp_pe->next;
new_pe->next = temp_pe->next;
temp_pe->next = new_pe;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
}
}
end:
SCReturn;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline void AppendAppLayerProtoDetectProbingParser(AppLayerProtoDetectProbingParser **head_pp,
AppLayerProtoDetectProbingParser *new_pp)
{
SCEnter();
if (*head_pp == NULL) {
*head_pp = new_pp;
goto end;
}
AppLayerProtoDetectProbingParser *temp_pp = *head_pp;
while (temp_pp->next != NULL)
temp_pp = temp_pp->next;
temp_pp->next = new_pp;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
SCReturn;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static inline void AppendAppLayerProtoDetectProbingParserPort(AppLayerProtoDetectProbingParserPort **head_port,
AppLayerProtoDetectProbingParserPort *new_port)
{
SCEnter();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (*head_port == NULL) {
*head_port = new_port;
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if ((*head_port)->port == 0) {
new_port->next = *head_port;
*head_port = new_port;
} else {
AppLayerProtoDetectProbingParserPort *temp_port = *head_port;
while (temp_port->next != NULL && temp_port->next->port != 0) {
temp_port = temp_port->next;
}
new_port->next = temp_port->next;
temp_port->next = new_port;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
SCReturn;
}
static inline void AppLayerProtoDetectInsertNewProbingParser(AppLayerProtoDetectProbingParser **pp,
uint16_t ip_proto,
uint16_t port,
AppProto alproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint16_t min_depth, uint16_t max_depth,
uint8_t direction,
ProbingParserFPtr ProbingParser)
{
SCEnter();
/* get the top level ipproto pp */
AppLayerProtoDetectProbingParser *curr_pp = *pp;
while (curr_pp != NULL) {
if (curr_pp->ip_proto == ip_proto)
break;
curr_pp = curr_pp->next;
}
if (curr_pp == NULL) {
AppLayerProtoDetectProbingParser *new_pp = AllocAppLayerProtoDetectProbingParser();
new_pp->ip_proto = ip_proto;
AppendAppLayerProtoDetectProbingParser(pp, new_pp);
curr_pp = new_pp;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/* get the top level port pp */
AppLayerProtoDetectProbingParserPort *curr_port = curr_pp->port;
while (curr_port != NULL) {
if (curr_port->port == port)
break;
curr_port = curr_port->next;
}
if (curr_port == NULL) {
AppLayerProtoDetectProbingParserPort *new_port = AllocAppLayerProtoDetectProbingParserPort();
new_port->port = port;
AppendAppLayerProtoDetectProbingParserPort(&curr_pp->port, new_port);
curr_port = new_port;
if (direction & STREAM_TOSERVER) {
curr_port->toserver_max_depth = max_depth;
} else {
curr_port->toclient_max_depth = max_depth;
}
AppLayerProtoDetectProbingParserPort *zero_port;
zero_port = curr_pp->port;
while (zero_port != NULL && zero_port->port != 0) {
zero_port = zero_port->next;
}
if (zero_port != NULL) {
AppLayerProtoDetectProbingParserElement *zero_pe;
zero_pe = zero_port->toserver;
for ( ; zero_pe != NULL; zero_pe = zero_pe->next) {
if (curr_port->toserver == NULL)
curr_port->toserver_max_depth = zero_pe->max_depth;
if (zero_pe->max_depth == 0)
curr_port->toserver_max_depth = zero_pe->max_depth;
if (curr_port->toserver_max_depth != 0 &&
curr_port->toserver_max_depth < zero_pe->max_depth) {
curr_port->toserver_max_depth = zero_pe->max_depth;
}
AppLayerProtoDetectProbingParserElement *dup_pe =
DuplicateAppLayerProtoDetectProbingParserElement(zero_pe);
AppendAppLayerProtoDetectProbingParserElement(&curr_port->toserver, dup_pe);
curr_port->toserver_alproto_mask |= dup_pe->alproto_mask;
}
zero_pe = zero_port->toclient;
for ( ; zero_pe != NULL; zero_pe = zero_pe->next) {
if (curr_port->toclient == NULL)
curr_port->toclient_max_depth = zero_pe->max_depth;
if (zero_pe->max_depth == 0)
curr_port->toclient_max_depth = zero_pe->max_depth;
if (curr_port->toclient_max_depth != 0 &&
curr_port->toclient_max_depth < zero_pe->max_depth) {
curr_port->toclient_max_depth = zero_pe->max_depth;
}
AppLayerProtoDetectProbingParserElement *dup_pe =
DuplicateAppLayerProtoDetectProbingParserElement(zero_pe);
AppendAppLayerProtoDetectProbingParserElement(&curr_port->toclient, dup_pe);
curr_port->toclient_alproto_mask |= dup_pe->alproto_mask;
}
} /* if (zero_port != NULL) */
} /* if (curr_port == NULL) */
/* insert the pe_pp */
AppLayerProtoDetectProbingParserElement *curr_pe;
if (direction & STREAM_TOSERVER)
curr_pe = curr_port->toserver;
else
curr_pe = curr_port->toclient;
while (curr_pe != NULL) {
if (curr_pe->alproto == alproto) {
SCLogError(SC_ERR_ALPARSER, "Duplicate pp registered - "
"ip_proto - %"PRIu16" Port - %"PRIu16" "
"App Protocol - NULL, App Protocol(ID) - "
"%"PRIu16" min_depth - %"PRIu16" "
"max_dept - %"PRIu16".",
ip_proto, port, alproto,
min_depth, max_depth);
goto error;
}
curr_pe = curr_pe->next;
}
/* Get a new parser element */
AppLayerProtoDetectProbingParserElement *new_pe =
AppLayerProtoDetectCreateAppLayerProtoDetectProbingParserElement(alproto,
curr_port->port,
min_depth, max_depth,
ProbingParser);
if (new_pe == NULL)
goto error;
curr_pe = new_pe;
AppLayerProtoDetectProbingParserElement **head_pe;
if (direction & STREAM_TOSERVER) {
if (curr_port->toserver == NULL)
curr_port->toserver_max_depth = new_pe->max_depth;
if (new_pe->max_depth == 0)
curr_port->toserver_max_depth = new_pe->max_depth;
if (curr_port->toserver_max_depth != 0 &&
curr_port->toserver_max_depth < new_pe->max_depth) {
curr_port->toserver_max_depth = new_pe->max_depth;
}
curr_port->toserver_alproto_mask |= new_pe->alproto_mask;
head_pe = &curr_port->toserver;
} else {
if (curr_port->toclient == NULL)
curr_port->toclient_max_depth = new_pe->max_depth;
if (new_pe->max_depth == 0)
curr_port->toclient_max_depth = new_pe->max_depth;
if (curr_port->toclient_max_depth != 0 &&
curr_port->toclient_max_depth < new_pe->max_depth) {
curr_port->toclient_max_depth = new_pe->max_depth;
}
curr_port->toclient_alproto_mask |= new_pe->alproto_mask;
head_pe = &curr_port->toclient;
}
AppendAppLayerProtoDetectProbingParserElement(head_pe, new_pe);
if (curr_port->port == 0) {
AppLayerProtoDetectProbingParserPort *temp_port = curr_pp->port;
while (temp_port != NULL && temp_port->port != 0) {
if (direction & STREAM_TOSERVER) {
if (temp_port->toserver == NULL)
temp_port->toserver_max_depth = curr_pe->max_depth;
if (curr_pe->max_depth == 0)
temp_port->toserver_max_depth = curr_pe->max_depth;
if (temp_port->toserver_max_depth != 0 &&
temp_port->toserver_max_depth < curr_pe->max_depth) {
temp_port->toserver_max_depth = curr_pe->max_depth;
}
AppendAppLayerProtoDetectProbingParserElement(&temp_port->toserver,
DuplicateAppLayerProtoDetectProbingParserElement(curr_pe));
temp_port->toserver_alproto_mask |= curr_pe->alproto_mask;
} else {
if (temp_port->toclient == NULL)
temp_port->toclient_max_depth = curr_pe->max_depth;
if (curr_pe->max_depth == 0)
temp_port->toclient_max_depth = curr_pe->max_depth;
if (temp_port->toclient_max_depth != 0 &&
temp_port->toclient_max_depth < curr_pe->max_depth) {
temp_port->toclient_max_depth = curr_pe->max_depth;
}
AppendAppLayerProtoDetectProbingParserElement(&temp_port->toclient,
DuplicateAppLayerProtoDetectProbingParserElement(curr_pe));
temp_port->toclient_alproto_mask |= curr_pe->alproto_mask;
}
temp_port = temp_port->next;
} /* while */
} /* if */
error:
SCReturn;
}
/***** Static Internal Calls: PM registration *****/
static void AppLayerProtoDetectPMGetIpprotos(AppProto alproto,
uint8_t *ipprotos)
{
SCEnter();
AppLayerProtoDetectPMSignature *s;
int pat_id, max_pat_id;
int i, j;
uint8_t ipproto;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
ipproto = FlowGetReverseProtoMapping(i);
for (j = 0; j < 2; j++) {
AppLayerProtoDetectPMCtx *pm_ctx = &alpd_ctx.ctx_ipp[i].ctx_pm[j];
max_pat_id = pm_ctx->max_pat_id;
for (pat_id = 0; pat_id < max_pat_id; pat_id++) {
s = pm_ctx->map[pat_id];
while (s != NULL) {
if (s->alproto == alproto)
ipprotos[ipproto / 8] |= 1 << (ipproto % 8);
s = s->next;
}
}
}
}
SCReturn;
}
static int AppLayerProtoDetectPMSetContentIDs(AppLayerProtoDetectPMCtx *ctx)
{
SCEnter();
typedef struct TempContainer_ {
PatIntId id;
uint16_t content_len;
uint8_t *content;
} TempContainer;
AppLayerProtoDetectPMSignature *s = NULL;
uint32_t struct_total_size = 0;
uint32_t content_total_size = 0;
/* array hash buffer */
uint8_t *ahb = NULL;
uint8_t *content = NULL;
uint8_t content_len = 0;
PatIntId max_id = 0;
TempContainer *struct_offset = NULL;
uint8_t *content_offset = NULL;
TempContainer *dup = NULL;
int ret = 0;
if (ctx->head == NULL)
goto end;
for (s = ctx->head; s != NULL; s = s->next) {
struct_total_size += sizeof(TempContainer);
content_total_size += s->cd->content_len;
}
ahb = SCMalloc(sizeof(uint8_t) * (struct_total_size + content_total_size));
if (unlikely(ahb == NULL))
goto error;
struct_offset = (TempContainer *)ahb;
content_offset = ahb + struct_total_size;
for (s = ctx->head; s != NULL; s = s->next) {
dup = (TempContainer *)ahb;
content = s->cd->content;
content_len = s->cd->content_len;
for (; dup != struct_offset; dup++) {
if (dup->content_len != content_len ||
SCMemcmp(dup->content, content, dup->content_len) != 0)
{
continue;
}
break;
}
if (dup != struct_offset) {
s->cd->id = dup->id;
continue;
}
struct_offset->content_len = content_len;
struct_offset->content = content_offset;
content_offset += content_len;
memcpy(struct_offset->content, content, content_len);
struct_offset->id = max_id++;
s->cd->id = struct_offset->id;
struct_offset++;
}
ctx->max_pat_id = max_id;
goto end;
error:
ret = -1;
end:
if (ahb != NULL)
SCFree(ahb);
SCReturnInt(ret);
}
static int AppLayerProtoDetectPMMapSignatures(AppLayerProtoDetectPMCtx *ctx)
{
SCEnter();
int ret = 0;
PatIntId max_pat_id = 0, tmp_pat_id;
AppLayerProtoDetectPMSignature *s, *next_s;
int mpm_ret;
max_pat_id = ctx->max_pat_id;
ctx->map = SCMalloc((max_pat_id) * sizeof(ctx->map));
if (ctx->map == NULL)
goto error;
memset(ctx->map, 0, (max_pat_id) * sizeof(ctx->map));
/* add an array indexed by pattern id to look up the sig */
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
for (s = ctx->head; s != NULL;) {
next_s = s->next;
s->next = ctx->map[s->cd->id];
ctx->map[s->cd->id] = s;
s = next_s;
}
ctx->head = NULL;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
for (tmp_pat_id = 0; tmp_pat_id < max_pat_id; tmp_pat_id++) {
s = NULL;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
for (s = ctx->map[tmp_pat_id]; s != NULL; s = s->next) {
if (s->cd->flags & DETECT_CONTENT_NOCASE) {
break;
}
}
/* if s != NULL now, it's CI. If NULL, CS */
if (s != NULL) {
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
mpm_ret = MpmAddPatternCI(&ctx->mpm_ctx,
s->cd->content, s->cd->content_len,
0, 0,
tmp_pat_id, 0, 0);
if (mpm_ret < 0)
goto error;
} else {
s = ctx->map[tmp_pat_id];
if (s == NULL)
goto error;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
mpm_ret = MpmAddPatternCS(&ctx->mpm_ctx,
s->cd->content, s->cd->content_len,
0, 0,
tmp_pat_id, 0, 0);
if (mpm_ret < 0)
goto error;
}
}
goto end;
error:
ret = -1;
end:
SCReturnInt(ret);
}
static int AppLayerProtoDetectPMPrepareMpm(AppLayerProtoDetectPMCtx *ctx)
{
SCEnter();
int ret = 0;
MpmCtx *mpm_ctx = &ctx->mpm_ctx;
if (mpm_table[mpm_ctx->mpm_type].Prepare(mpm_ctx) < 0)
goto error;
goto end;
error:
ret = -1;
end:
SCReturnInt(ret);
}
static void AppLayerProtoDetectPMFreeSignature(AppLayerProtoDetectPMSignature *sig)
{
SCEnter();
if (sig == NULL)
SCReturn;
if (sig->cd)
DetectContentFree(sig->cd);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCFree(sig);
SCReturn;
}
static int AppLayerProtoDetectPMAddSignature(AppLayerProtoDetectPMCtx *ctx, DetectContentData *cd,
AppProto alproto)
{
SCEnter();
int ret = 0;
AppLayerProtoDetectPMSignature *s = SCMalloc(sizeof(*s));
if (unlikely(s == NULL))
goto error;
memset(s, 0, sizeof(*s));
s->alproto = alproto;
s->cd = cd;
/* prepend to the list */
s->next = ctx->head;
ctx->head = s;
goto end;
error:
ret = -1;
end:
SCReturnInt(ret);
}
static int AppLayerProtoDetectPMRegisterPattern(uint8_t ipproto, AppProto alproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
char *pattern,
uint16_t depth, uint16_t offset,
uint8_t direction,
uint8_t is_cs)
{
SCEnter();
AppLayerProtoDetectCtxIpproto *ctx_ipp = &alpd_ctx.ctx_ipp[FlowGetProtoMapping(ipproto)];
AppLayerProtoDetectPMCtx *ctx_pm = NULL;
DetectContentData *cd;
int ret = 0;
cd = DetectContentParseEncloseQuotes(pattern);
if (cd == NULL)
goto error;
cd->depth = depth;
cd->offset = offset;
if (!is_cs)
cd->flags |= DETECT_CONTENT_NOCASE;
if (depth < cd->content_len)
goto error;
if (direction & STREAM_TOSERVER)
ctx_pm = (AppLayerProtoDetectPMCtx *)&ctx_ipp->ctx_pm[0];
else
ctx_pm = (AppLayerProtoDetectPMCtx *)&ctx_ipp->ctx_pm[1];
if (depth > ctx_pm->max_len)
ctx_pm->max_len = depth;
if (depth < ctx_pm->min_len)
ctx_pm->min_len = depth;
/* Finally turn it into a signature and add to the ctx. */
AppLayerProtoDetectPMAddSignature(ctx_pm, cd, alproto);
goto end;
error:
ret = -1;
end:
SCReturnInt(ret);
}
/***** Protocol Retrieval *****/
AppProto AppLayerProtoDetectGetProto(void *tctx,
Flow *f,
uint8_t *buf, uint32_t buflen,
uint8_t ipproto, uint8_t direction)
{
SCEnter();
AppProto alproto = ALPROTO_UNKNOWN;
uint16_t pm_results[ALPROTO_MAX];
uint16_t pm_matches;
if (!FLOW_IS_PM_DONE(f, direction)) {
pm_matches = AppLayerProtoDetectPMGetProto(tctx,
f,
buf, buflen,
direction,
ipproto,
pm_results);
if (pm_matches > 0) {
alproto = pm_results[0];
goto end;
}
}
if (!FLOW_IS_PP_DONE(f, direction))
alproto = AppLayerProtoDetectPPGetProto(f, buf, buflen, ipproto, direction);
end:
SCReturnCT(alproto, "AppProto");
}
static void AppLayerProtoDetectFreeProbingParsers(AppLayerProtoDetectProbingParser *pp)
{
SCEnter();
AppLayerProtoDetectProbingParser *tmp_pp = NULL;
if (pp == NULL)
goto end;
while (pp != NULL) {
tmp_pp = pp->next;
DeAllocAppLayerProtoDetectProbingParser(pp);
pp = tmp_pp;
}
end:
SCReturn;
}
/***** State Preparation *****/
int AppLayerProtoDetectPrepareState(void)
{
SCEnter();
AppLayerProtoDetectPMCtx *ctx_pm;
int i, j;
int ret = 0;
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
for (j = 0; j < 2; j++) {
ctx_pm = &alpd_ctx.ctx_ipp[i].ctx_pm[j];
if (AppLayerProtoDetectPMSetContentIDs(ctx_pm) < 0)
goto error;
if (ctx_pm->max_pat_id == 0)
continue;
if (AppLayerProtoDetectPMMapSignatures(ctx_pm) < 0)
goto error;
if (AppLayerProtoDetectPMPrepareMpm(ctx_pm) < 0)
goto error;
}
}
goto end;
error:
ret = -1;
end:
SCReturnInt(ret);
}
/***** PP registration *****/
void AppLayerProtoDetectPPRegister(uint8_t ipproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
char *portstr,
AppProto alproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint16_t min_depth, uint16_t max_depth,
uint8_t direction,
ProbingParserFPtr ProbingParser)
{
SCEnter();
DetectPort *head = NULL;
DetectPortParse(&head, portstr);
DetectPort *temp_dp = head;
while (temp_dp != NULL) {
uint32_t port = temp_dp->port;
if (port == 0 && temp_dp->port2 != 0)
port++;
for ( ; port <= temp_dp->port2; port++) {
AppLayerProtoDetectInsertNewProbingParser(&alpd_ctx.ctx_pp,
ipproto,
port,
alproto,
min_depth, max_depth,
direction,
ProbingParser);
}
temp_dp = temp_dp->next;
}
DetectPortCleanupList(head);
SCReturn;
}
void AppLayerProtoDetectPPParseConfPorts(const char *ipproto_name,
uint8_t ipproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
const char *alproto_name,
AppProto alproto,
uint16_t min_depth, uint16_t max_depth,
ProbingParserFPtr ProbingParser)
{
SCEnter();
char param[100];
int r;
ConfNode *node;
ConfNode *port_node = NULL;
r = snprintf(param, sizeof(param), "%s%s%s", "app-layer.protocols.",
alproto_name, ".detection-ports");
if (r < 0) {
SCLogError(SC_ERR_FATAL, "snprintf failure.");
exit(EXIT_FAILURE);
} else if (r > (int)sizeof(param)) {
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
exit(EXIT_FAILURE);
}
node = ConfGetNode(param);
if (node == NULL) {
SCLogDebug("Entry for %s not found.", param);
r = snprintf(param, sizeof(param), "%s%s%s%s%s", "app-layer.protocols.",
alproto_name, ".", ipproto_name, ".detection-ports");
if (r < 0) {
SCLogError(SC_ERR_FATAL, "snprintf failure.");
exit(EXIT_FAILURE);
} else if (r > (int)sizeof(param)) {
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
exit(EXIT_FAILURE);
}
node = ConfGetNode(param);
if (node == NULL)
goto end;
}
port_node = ConfNodeLookupChild(node, "toserver");
if (port_node != NULL && port_node->val != NULL) {
AppLayerProtoDetectPPRegister(ipproto,
port_node->val,
alproto,
min_depth, max_depth,
STREAM_TOSERVER,
ProbingParser);
}
port_node = ConfNodeLookupChild(node, "toclient");
if (port_node != NULL && port_node->val != NULL) {
AppLayerProtoDetectPPRegister(ipproto,
port_node->val,
alproto,
min_depth, max_depth,
STREAM_TOCLIENT,
ProbingParser);
}
end:
SCReturn;
}
/***** PM registration *****/
int AppLayerProtoDetectPMRegisterPatternCS(uint8_t ipproto, AppProto alproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
char *pattern,
uint16_t depth, uint16_t offset,
uint8_t direction)
{
SCEnter();
SCReturnInt(AppLayerProtoDetectPMRegisterPattern(ipproto, alproto,
pattern,
depth, offset,
direction,
1 /* case-sensitive */));
}
int AppLayerProtoDetectPMRegisterPatternCI(uint8_t ipproto, AppProto alproto,
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
char *pattern,
uint16_t depth, uint16_t offset,
uint8_t direction)
{
SCEnter();
SCReturnInt(AppLayerProtoDetectPMRegisterPattern(ipproto, alproto,
pattern,
depth, offset,
direction,
0 /* !case-sensitive */));
}
/***** Setup/General Registration *****/
int AppLayerProtoDetectSetup(void)
{
SCEnter();
int i, j;
memset(&alpd_ctx, 0, sizeof(alpd_ctx));
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
for (j = 0; j < 2; j++) {
MpmInitCtx(&alpd_ctx.ctx_ipp[i].ctx_pm[j].mpm_ctx, MPM_AC);
}
}
SCReturnInt(0);
}
/**
* \todo incomplete. Need more work.
*/
int AppLayerProtoDetectDeSetup(void)
{
SCEnter();
int ipproto_map = 0;
int dir = 0;
PatIntId id = 0;
AppLayerProtoDetectPMCtx *pm_ctx = NULL;
AppLayerProtoDetectPMSignature *sig = NULL, *next_sig = NULL;
for (ipproto_map = 0; ipproto_map < FLOW_PROTO_DEFAULT; ipproto_map++) {
for (dir = 0; dir < 2; dir++) {
pm_ctx = &alpd_ctx.ctx_ipp[ipproto_map].ctx_pm[dir];
mpm_table[pm_ctx->mpm_ctx.mpm_type].DestroyCtx(pm_ctx->mpm_ctx.ctx);
for (id = 0; id < pm_ctx->max_pat_id; id++) {
sig = pm_ctx->map[id];
while (sig != NULL) {
next_sig = sig->next;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPMFreeSignature(sig);
sig = next_sig;
}
}
}
}
AppLayerProtoDetectFreeProbingParsers(alpd_ctx.ctx_pp);
SCReturnInt(0);
}
void AppLayerProtoDetectRegisterProtocol(AppProto alproto, char *alproto_name)
{
SCEnter();
if (alpd_ctx.alproto_names[alproto] != NULL)
goto end;
alpd_ctx.alproto_names[alproto] = alproto_name;
goto end;
end:
SCReturn;
}
int AppLayerProtoDetectConfProtoDetectionEnabled(const char *ipproto,
const char *alproto)
{
SCEnter();
BUG_ON(ipproto == NULL || alproto == NULL);
int enabled = 1;
char param[100];
ConfNode *node;
int r;
if (RunmodeIsUnittests())
goto enabled;
r = snprintf(param, sizeof(param), "%s%s%s", "app-layer.protocols.",
alproto, ".enabled");
if (r < 0) {
SCLogError(SC_ERR_FATAL, "snprintf failure.");
exit(EXIT_FAILURE);
} else if (r > (int)sizeof(param)) {
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
exit(EXIT_FAILURE);
}
node = ConfGetNode(param);
if (node == NULL) {
SCLogDebug("Entry for %s not found.", param);
r = snprintf(param, sizeof(param), "%s%s%s%s%s", "app-layer.protocols.",
alproto, ".", ipproto, ".enabled");
if (r < 0) {
SCLogError(SC_ERR_FATAL, "snprintf failure.");
exit(EXIT_FAILURE);
} else if (r > (int)sizeof(param)) {
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
exit(EXIT_FAILURE);
}
node = ConfGetNode(param);
if (node == NULL) {
SCLogDebug("Entry for %s not found.", param);
goto enabled;
}
}
if (strcasecmp(node->val, "yes") == 0) {
goto enabled;
} else if (strcasecmp(node->val, "no") == 0) {
goto disabled;
} else if (strcasecmp(node->val, "detection-only") == 0) {
goto enabled;
} else {
SCLogError(SC_ERR_FATAL, "Invalid value found for %s.", param);
exit(EXIT_FAILURE);
}
disabled:
enabled = 0;
enabled:
SCReturnInt(enabled);
}
void *AppLayerProtoDetectGetCtxThread(void)
{
SCEnter();
AppLayerProtoDetectThreadCtx *alpd_tctx = NULL;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
MpmCtx *mpm_ctx;
MpmThreadCtx *mpm_tctx;
int i, j;
PatIntId max_pat_id = 0;
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
for (j = 0; j < 2; j++) {
if (max_pat_id == 0) {
max_pat_id = alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id;
} else if (alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id &&
max_pat_id < alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id)
{
max_pat_id = alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
}
}
alpd_tctx = SCMalloc(sizeof(*alpd_tctx));
if (alpd_tctx == NULL)
goto error;
memset(alpd_tctx, 0, sizeof(*alpd_tctx));
/* Get the max pat id for all the mpm ctxs. */
if (PmqSetup(&alpd_tctx->pmq, max_pat_id) < 0)
goto error;
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
for (j = 0; j < 2; j++) {
mpm_ctx = &alpd_ctx.ctx_ipp[i].ctx_pm[j].mpm_ctx;
mpm_tctx = &alpd_tctx->mpm_tctx[i][j];
mpm_table[mpm_ctx->mpm_type].InitThreadCtx(mpm_ctx, mpm_tctx, 0);
}
}
goto end;
error:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
alpd_tctx = NULL;
end:
SCReturnPtr(alpd_tctx, "void *");
}
void AppLayerProtoDetectDestroyCtxThread(void *tctx)
{
SCEnter();
AppLayerProtoDetectThreadCtx *alpd_tctx = (AppLayerProtoDetectThreadCtx *)tctx;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
MpmCtx *mpm_ctx;
MpmThreadCtx *mpm_tctx;
int ipproto_map, dir;
for (ipproto_map = 0; ipproto_map < FLOW_PROTO_DEFAULT; ipproto_map++) {
for (dir = 0; dir < 2; dir++) {
mpm_ctx = &alpd_ctx.ctx_ipp[ipproto_map].ctx_pm[dir].mpm_ctx;
mpm_tctx = &alpd_tctx->mpm_tctx[ipproto_map][dir];
mpm_table[mpm_ctx->mpm_type].DestroyThreadCtx(mpm_ctx, mpm_tctx);
}
}
PmqFree(&alpd_tctx->pmq);
SCFree(alpd_tctx);
SCReturn;
}
/***** Utility *****/
void AppLayerProtoDetectSupportedIpprotos(AppProto alproto, uint8_t *ipprotos)
{
SCEnter();
AppLayerProtoDetectPMGetIpprotos(alproto, ipprotos);
AppLayerProtoDetectPPGetIpprotos(alproto, ipprotos);
SCReturn;
}
AppProto AppLayerProtoDetectGetProtoByName(char *alproto_name)
{
SCEnter();
AppProto a = ALPROTO_UNKNOWN;
for (a = 0; a < ALPROTO_MAX; a++) {
if (alpd_ctx.alproto_names[a] != NULL &&
strlen(alpd_ctx.alproto_names[a]) == strlen(alproto_name) &&
(SCMemcmp(alpd_ctx.alproto_names[a], alproto_name, strlen(alproto_name)) == 0))
{
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
}
end:
SCReturnCT(a, "AppProto");
}
char *AppLayerProtoDetectGetProtoName(AppProto alproto)
{
return alpd_ctx.alproto_names[alproto];
}
void AppLayerProtoDetectSupportedAppProtocols(AppProto *alprotos)
{
SCEnter();
memset(alprotos, 0, ALPROTO_MAX * sizeof(AppProto));
int alproto;
for (alproto = 0; alproto != ALPROTO_MAX; alproto++) {
if (alpd_ctx.alproto_names[alproto] != NULL)
alprotos[alproto] = 1;
}
SCReturn;
}
/***** Unittests *****/
#ifdef UNITTESTS
static AppLayerProtoDetectCtx alpd_ctx_ut;
void AppLayerProtoDetectUnittestCtxBackup(void)
{
SCEnter();
alpd_ctx_ut = alpd_ctx;
memset(&alpd_ctx, 0, sizeof(alpd_ctx));
SCReturn;
}
void AppLayerProtoDetectUnittestCtxRestore(void)
{
SCEnter();
alpd_ctx = alpd_ctx_ut;
memset(&alpd_ctx_ut, 0, sizeof(alpd_ctx_ut));
SCReturn;
}
int AppLayerProtoDetectTest01(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
char *buf;
int r = 0;
buf = "HTTP";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
buf = "GET";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOSERVER);
AppLayerProtoDetectPrepareState();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 1) {
printf("Failure - "
"alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 1\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
printf("Failure - "
"alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1\n");
goto end;
}
r = 1;
end:
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
int AppLayerProtoDetectTest02(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
char *buf;
int r = 0;
buf = "HTTP";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
buf = "ftp";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
AppLayerProtoDetectPrepareState();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
goto end;
}
r = 1;
end:
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
int AppLayerProtoDetectTest03(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "HTTP/1.1 200 OK\r\nServer: Apache/1.0\r\n\r\n";
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
memset(pm_results, 0, sizeof(pm_results));
buf = "HTTP";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
buf = "220 ";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
goto end;
}
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
printf("cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
goto end;
}
r = 1;
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
int AppLayerProtoDetectTest04(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "HTTP/1.1 200 OK\r\nServer: Apache/1.0\r\n\r\n";
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
memset(pm_results, 0, sizeof(pm_results));
buf = "200 ";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 13, 0, STREAM_TOCLIENT);
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_HTTP\n");
goto end;
}
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
printf("cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
goto end;
}
r = 1;
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
int AppLayerProtoDetectTest05(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "HTTP/1.1 200 OK\r\nServer: Apache/1.0\r\n\r\n<HTML><BODY>Blahblah</BODY></HTML>";
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
memset(pm_results, 0, sizeof(pm_results));
buf = "HTTP";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
buf = "220 ";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
printf("cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int AppLayerProtoDetectTest06(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "220 Welcome to the OISF FTP server\r\n";
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
buf = "HTTP";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
buf = "220 ";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_FTP) {
printf("cnt != 1 && pm_results[0] != AlPROTO_FTP\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int AppLayerProtoDetectTest07(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "220 Welcome to the OISF HTTP/FTP server\r\n";
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
memset(pm_results, 0, sizeof(pm_results));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
buf = "HTTP";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_HTTP\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 0) {
printf("cnt != 0\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int AppLayerProtoDetectTest08(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = {
0x00, 0x00, 0x00, 0x85, 0xff, 0x53, 0x4d, 0x42,
0x72, 0x00, 0x00, 0x00, 0x00, 0x18, 0x53, 0xc8,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0xfe,
0x00, 0x00, 0x00, 0x00, 0x00, 0x62, 0x00, 0x02,
0x50, 0x43, 0x20, 0x4e, 0x45, 0x54, 0x57, 0x4f,
0x52, 0x4b, 0x20, 0x50, 0x52, 0x4f, 0x47, 0x52,
0x41, 0x4d, 0x20, 0x31, 0x2e, 0x30, 0x00, 0x02,
0x4c, 0x41, 0x4e, 0x4d, 0x41, 0x4e, 0x31, 0x2e,
0x30, 0x00, 0x02, 0x57, 0x69, 0x6e, 0x64, 0x6f,
0x77, 0x73, 0x20, 0x66, 0x6f, 0x72, 0x20, 0x57,
0x6f, 0x72, 0x6b, 0x67, 0x72, 0x6f, 0x75, 0x70,
0x73, 0x20, 0x33, 0x2e, 0x31, 0x61, 0x00, 0x02,
0x4c, 0x4d, 0x31, 0x2e, 0x32, 0x58, 0x30, 0x30,
0x32, 0x00, 0x02, 0x4c, 0x41, 0x4e, 0x4d, 0x41,
0x4e, 0x32, 0x2e, 0x31, 0x00, 0x02, 0x4e, 0x54,
0x20, 0x4c, 0x4d, 0x20, 0x30, 0x2e, 0x31, 0x32,
0x00
};
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
memset(pm_results, 0, sizeof(pm_results));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
buf = "|ff|SMB";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_SMB, buf, 8, 4, STREAM_TOCLIENT);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_SMB) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_SMB\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_SMB) {
printf("cnt != 1 && pm_results[0] != AlPROTO_SMB\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int AppLayerProtoDetectTest09(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = {
0x00, 0x00, 0x00, 0x66, 0xfe, 0x53, 0x4d, 0x42,
0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x24, 0x00, 0x01, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x02, 0x02
};
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
16 years ago
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(pm_results, 0, sizeof(pm_results));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
buf = "|fe|SMB";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_SMB2, buf, 8, 4, STREAM_TOCLIENT);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_SMB2) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_SMB2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_SMB2) {
printf("cnt != 1 && pm_results[0] != AlPROTO_SMB2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int AppLayerProtoDetectTest10(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = {
0x05, 0x00, 0x0b, 0x03, 0x10, 0x00, 0x00, 0x00,
0x48, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0xd0, 0x16, 0xd0, 0x16, 0x00, 0x00, 0x00, 0x00,
0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00,
0xb8, 0x4a, 0x9f, 0x4d, 0x1c, 0x7d, 0xcf, 0x11,
0x86, 0x1e, 0x00, 0x20, 0xaf, 0x6e, 0x7c, 0x57,
0x00, 0x00, 0x00, 0x00, 0x04, 0x5d, 0x88, 0x8a,
0xeb, 0x1c, 0xc9, 0x11, 0x9f, 0xe8, 0x08, 0x00,
0x2b, 0x10, 0x48, 0x60, 0x02, 0x00, 0x00, 0x00
};
char *buf;
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
memset(pm_results, 0, sizeof(pm_results));
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
buf = "|05 00|";
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_DCERPC, buf, 4, 0, STREAM_TOCLIENT);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_DCERPC) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_DCERPC\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_DCERPC) {
printf("cnt != 1 && pm_results[0] != AlPROTO_DCERPC\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/**
* \test Why we still get http for connect... obviously because
* we also match on the reply, duh
*/
int AppLayerProtoDetectTest11(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "CONNECT www.ssllabs.com:443 HTTP/1.0\r\n";
uint8_t l7data_resp[] = "HTTP/1.1 405 Method Not Allowed\r\n";
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
memset(pm_results, 0, sizeof(pm_results));
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "GET", 3, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "PUT", 3, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "POST", 4, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "TRACE", 5, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "OPTIONS", 7, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "CONNECT", 7, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOCLIENT);
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 7) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 7\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[1]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[2]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[3]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[4]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[5]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[6]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP)
{
printf("failure 1\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(pm_results, 0, sizeof(pm_results));
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOSERVER,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
printf("l7data - cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(pm_results, 0, sizeof(pm_results));
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data_resp, sizeof(l7data_resp),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
printf("l7data_resp - cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/**
* \test AlpProtoSignature test
*/
int AppLayerProtoDetectTest12(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = 0;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].head == NULL ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL)
{
printf("failure 1\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPrepareState();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 1) {
printf("failure 2\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].head != NULL ||
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map == NULL)
{
printf("failure 3\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP) {
printf("failure 4\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->cd->id != 0) {
printf("failure 5\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->next != NULL) {
printf("failure 6\n");
goto end;
}
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
/**
* \test What about if we add some sigs only for udp but call for tcp?
* It should not detect any proto
*/
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int AppLayerProtoDetectTest13(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "CONNECT www.ssllabs.com:443 HTTP/1.0\r\n";
uint8_t l7data_resp[] = "HTTP/1.1 405 Method Not Allowed\r\n";
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
uint32_t cnt;
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "GET", 3, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "PUT", 3, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "POST", 4, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "TRACE", 5, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "OPTIONS", 7, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "CONNECT", 7, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOCLIENT);
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[1]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[2]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[3]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[4]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[5]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[6]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP)
{
printf("failure 1\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(pm_results, 0, sizeof(pm_results));
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOSERVER,
IPPROTO_TCP,
pm_results);
if (cnt != 0) {
printf("l7data - cnt != 0\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(pm_results, 0, sizeof(pm_results));
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data_resp, sizeof(l7data_resp),
STREAM_TOCLIENT,
IPPROTO_TCP,
pm_results);
if (cnt != 0) {
printf("l7data_resp - cnt != 0\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
/**
* \test What about if we add some sigs only for udp calling it for UDP?
* It should detect ALPROTO_HTTP (over udp). This is just a check
* to ensure that TCP/UDP differences work correctly.
*/
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int AppLayerProtoDetectTest14(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
uint8_t l7data[] = "CONNECT www.ssllabs.com:443 HTTP/1.0\r\n";
uint8_t l7data_resp[] = "HTTP/1.1 405 Method Not Allowed\r\n";
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = 0;
Flow f;
uint16_t pm_results[ALPROTO_MAX];
void *alpd_tctx;
uint32_t cnt;
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "GET", 3, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "PUT", 3, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "POST", 4, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "TRACE", 5, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "OPTIONS", 7, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "CONNECT", 7, 0, STREAM_TOSERVER);
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOCLIENT);
AppLayerProtoDetectPrepareState();
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
* it sets internal structures which depends on the above function. */
alpd_tctx = AppLayerProtoDetectGetCtxThread();
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7\n");
goto end;
}
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1) {
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[1]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[2]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[3]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[4]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[5]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[6]->alproto != ALPROTO_HTTP ||
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP)
{
printf("failure 1\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(pm_results, 0, sizeof(pm_results));
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data, sizeof(l7data),
STREAM_TOSERVER,
IPPROTO_UDP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
printf("l7data - cnt != 0\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
memset(pm_results, 0, sizeof(pm_results));
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
&f,
l7data_resp, sizeof(l7data_resp),
STREAM_TOCLIENT,
IPPROTO_UDP,
pm_results);
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
printf("l7data_resp - cnt != 0\n");
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
r = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alpd_tctx != NULL)
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return r;
}
typedef struct AppLayerProtoDetectPPTestDataElement_ {
char *alproto_name;
AppProto alproto;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
uint16_t port;
uint32_t alproto_mask;
uint32_t min_depth;
uint32_t max_depth;
} AppLayerProtoDetectPPTestDataElement;
typedef struct AppLayerProtoDetectPPTestDataPort_ {
uint16_t port;
uint32_t toserver_alproto_mask;
uint32_t toclient_alproto_mask;
uint16_t toserver_max_depth;
uint16_t toclient_max_depth;
AppLayerProtoDetectPPTestDataElement *toserver_element;
AppLayerProtoDetectPPTestDataElement *toclient_element;
int ts_no_of_element;
int tc_no_of_element;
} AppLayerProtoDetectPPTestDataPort;
typedef struct AppLayerProtoDetectPPTestDataIPProto_ {
uint16_t ip_proto;
AppLayerProtoDetectPPTestDataPort *port;
int no_of_port;
} AppLayerProtoDetectPPTestDataIPProto;
static int AppLayerProtoDetectPPTestData(AppLayerProtoDetectProbingParser *pp,
AppLayerProtoDetectPPTestDataIPProto *ip_proto,
int no_of_ip_proto)
{
int result = 0;
int i = -1, j = -1 , k = -1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
#ifdef DEBUG
int dir = 0;
#endif
for (i = 0; i < no_of_ip_proto; i++, pp = pp->next) {
if (pp->ip_proto != ip_proto[i].ip_proto)
goto end;
AppLayerProtoDetectProbingParserPort *pp_port = pp->port;
for (k = 0; k < ip_proto[i].no_of_port; k++, pp_port = pp_port->next) {
if (pp_port->port != ip_proto[i].port[k].port)
goto end;
if (pp_port->toserver_alproto_mask != ip_proto[i].port[k].toserver_alproto_mask)
goto end;
if (pp_port->toclient_alproto_mask != ip_proto[i].port[k].toclient_alproto_mask)
goto end;
if (pp_port->toserver_max_depth != ip_proto[i].port[k].toserver_max_depth)
goto end;
if (pp_port->toclient_max_depth != ip_proto[i].port[k].toclient_max_depth)
goto end;
AppLayerProtoDetectProbingParserElement *pp_element = pp_port->toserver;
#ifdef DEBUG
dir = 0;
#endif
for (j = 0 ; j < ip_proto[i].port[k].ts_no_of_element;
j++, pp_element = pp_element->next) {
if (pp_element->alproto != ip_proto[i].port[k].toserver_element[j].alproto) {
goto end;
}
if (pp_element->port != ip_proto[i].port[k].toserver_element[j].port) {
goto end;
}
if (pp_element->alproto_mask != ip_proto[i].port[k].toserver_element[j].alproto_mask) {
goto end;
}
if (pp_element->min_depth != ip_proto[i].port[k].toserver_element[j].min_depth) {
goto end;
}
if (pp_element->max_depth != ip_proto[i].port[k].toserver_element[j].max_depth) {
goto end;
}
} /* for */
if (pp_element != NULL)
goto end;
pp_element = pp_port->toclient;
#ifdef DEBUG
dir = 1;
#endif
for (j = 0 ; j < ip_proto[i].port[k].tc_no_of_element; j++, pp_element = pp_element->next) {
if (pp_element->alproto != ip_proto[i].port[k].toclient_element[j].alproto) {
goto end;
}
if (pp_element->port != ip_proto[i].port[k].toclient_element[j].port) {
goto end;
}
if (pp_element->alproto_mask != ip_proto[i].port[k].toclient_element[j].alproto_mask) {
goto end;
}
if (pp_element->min_depth != ip_proto[i].port[k].toclient_element[j].min_depth) {
goto end;
}
if (pp_element->max_depth != ip_proto[i].port[k].toclient_element[j].max_depth) {
goto end;
}
} /* for */
if (pp_element != NULL)
goto end;
}
if (pp_port != NULL)
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
if (pp != NULL)
goto end;
result = 1;
end:
#ifdef DEBUG
printf("i = %d, k = %d, j = %d(%s)\n", i, k, j, (dir == 0) ? "ts" : "tc");
#endif
return result;
}
static uint16_t ProbingParserDummyForTesting(uint8_t *input,
uint32_t input_len,
uint32_t *offset)
{
return 0;
}
static int AppLayerProtoDetectTest15(void)
{
AppLayerProtoDetectUnittestCtxBackup();
AppLayerProtoDetectSetup();
int result = 0;
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"80",
ALPROTO_HTTP,
5, 8,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"80",
ALPROTO_SMB,
5, 6,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"80",
ALPROTO_FTP,
7, 10,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"81",
ALPROTO_DCERPC,
9, 10,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"81",
ALPROTO_FTP,
7, 15,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"0",
ALPROTO_SMTP,
12, 0,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"0",
ALPROTO_TLS,
12, 18,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"85",
ALPROTO_DCERPC,
9, 10,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"85",
ALPROTO_FTP,
7, 15,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
result = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
AppLayerProtoDetectPPRegister(IPPROTO_UDP,
"85",
ALPROTO_IMAP,
12, 23,
STREAM_TOSERVER,
ProbingParserDummyForTesting);
/* toclient */
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"0",
ALPROTO_JABBER,
12, 23,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"0",
ALPROTO_IRC,
12, 14,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"85",
ALPROTO_DCERPC,
9, 10,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"81",
ALPROTO_FTP,
7, 15,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"0",
ALPROTO_TLS,
12, 18,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"80",
ALPROTO_HTTP,
5, 8,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"81",
ALPROTO_DCERPC,
9, 10,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"90",
ALPROTO_FTP,
7, 15,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"80",
ALPROTO_SMB,
5, 6,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_UDP,
"85",
ALPROTO_IMAP,
12, 23,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"0",
ALPROTO_SMTP,
12, 17,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
"80",
ALPROTO_FTP,
7, 10,
STREAM_TOCLIENT,
ProbingParserDummyForTesting);
AppLayerProtoDetectPPTestDataElement element_ts_80[] = {
{ "http", ALPROTO_HTTP, 80, 1 << ALPROTO_HTTP, 5, 8 },
{ "smb", ALPROTO_SMB, 80, 1 << ALPROTO_SMB, 5, 6 },
{ "ftp", ALPROTO_FTP, 80, 1 << ALPROTO_FTP, 7, 10 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
};
AppLayerProtoDetectPPTestDataElement element_tc_80[] = {
{ "http", ALPROTO_HTTP, 80, 1 << ALPROTO_HTTP, 5, 8 },
{ "smb", ALPROTO_SMB, 80, 1 << ALPROTO_SMB, 5, 6 },
{ "ftp", ALPROTO_FTP, 80, 1 << ALPROTO_FTP, 7, 10 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
};
AppLayerProtoDetectPPTestDataElement element_ts_81[] = {
{ "dcerpc", ALPROTO_DCERPC, 81, 1 << ALPROTO_DCERPC, 9, 10 },
{ "ftp", ALPROTO_FTP, 81, 1 << ALPROTO_FTP, 7, 15 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
};
AppLayerProtoDetectPPTestDataElement element_tc_81[] = {
{ "ftp", ALPROTO_FTP, 81, 1 << ALPROTO_FTP, 7, 15 },
{ "dcerpc", ALPROTO_DCERPC, 81, 1 << ALPROTO_DCERPC, 9, 10 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
};
AppLayerProtoDetectPPTestDataElement element_ts_85[] = {
{ "dcerpc", ALPROTO_DCERPC, 85, 1 << ALPROTO_DCERPC, 9, 10 },
{ "ftp", ALPROTO_FTP, 85, 1 << ALPROTO_FTP, 7, 15 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
};
AppLayerProtoDetectPPTestDataElement element_tc_85[] = {
{ "dcerpc", ALPROTO_DCERPC, 85, 1 << ALPROTO_DCERPC, 9, 10 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
};
AppLayerProtoDetectPPTestDataElement element_ts_90[] = {
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
};
AppLayerProtoDetectPPTestDataElement element_tc_90[] = {
{ "ftp", ALPROTO_FTP, 90, 1 << ALPROTO_FTP, 7, 15 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
};
AppLayerProtoDetectPPTestDataElement element_ts_0[] = {
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
};
AppLayerProtoDetectPPTestDataElement element_tc_0[] = {
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
};
AppLayerProtoDetectPPTestDataElement element_ts_85_udp[] = {
{ "imap", ALPROTO_IMAP, 85, 1 << ALPROTO_IMAP, 12, 23 },
};
AppLayerProtoDetectPPTestDataElement element_tc_85_udp[] = {
{ "imap", ALPROTO_IMAP, 85, 1 << ALPROTO_IMAP, 12, 23 },
};
AppLayerProtoDetectPPTestDataPort ports_tcp[] = {
{ 80,
((1 << ALPROTO_HTTP) | (1 << ALPROTO_SMB) | (1 << ALPROTO_FTP) |
(1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
((1 << ALPROTO_HTTP) | (1 << ALPROTO_SMB) | (1 << ALPROTO_FTP) |
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
0, 23,
element_ts_80, element_tc_80,
sizeof(element_ts_80) / sizeof(AppLayerProtoDetectPPTestDataElement),
sizeof(element_tc_80) / sizeof(AppLayerProtoDetectPPTestDataElement),
},
{ 81,
((1 << ALPROTO_DCERPC) | (1 << ALPROTO_FTP) |
(1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
((1 << ALPROTO_FTP) | (1 << ALPROTO_DCERPC) |
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
0, 23,
element_ts_81, element_tc_81,
sizeof(element_ts_81) / sizeof(AppLayerProtoDetectPPTestDataElement),
sizeof(element_tc_81) / sizeof(AppLayerProtoDetectPPTestDataElement),
},
{ 85,
((1 << ALPROTO_DCERPC) | (1 << ALPROTO_FTP) |
(1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
((1 << ALPROTO_DCERPC) |
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
0, 23,
element_ts_85, element_tc_85,
sizeof(element_ts_85) / sizeof(AppLayerProtoDetectPPTestDataElement),
sizeof(element_tc_85) / sizeof(AppLayerProtoDetectPPTestDataElement)
},
{ 90,
((1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
((1 << ALPROTO_FTP) |
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
0, 23,
element_ts_90, element_tc_90,
sizeof(element_ts_90) / sizeof(AppLayerProtoDetectPPTestDataElement),
sizeof(element_tc_90) / sizeof(AppLayerProtoDetectPPTestDataElement)
},
{ 0,
((1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
((1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
0, 23,
element_ts_0, element_tc_0,
sizeof(element_ts_0) / sizeof(AppLayerProtoDetectPPTestDataElement),
sizeof(element_tc_0) / sizeof(AppLayerProtoDetectPPTestDataElement)
}
};
AppLayerProtoDetectPPTestDataPort ports_udp[] = {
{ 85,
(1 << ALPROTO_IMAP),
(1 << ALPROTO_IMAP),
23, 23,
element_ts_85_udp, element_tc_85_udp,
sizeof(element_ts_85_udp) / sizeof(AppLayerProtoDetectPPTestDataElement),
sizeof(element_tc_85_udp) / sizeof(AppLayerProtoDetectPPTestDataElement),
},
};
AppLayerProtoDetectPPTestDataIPProto ip_proto[] = {
{ IPPROTO_TCP,
ports_tcp,
sizeof(ports_tcp) / sizeof(AppLayerProtoDetectPPTestDataPort),
},
{ IPPROTO_UDP,
ports_udp,
sizeof(ports_udp) / sizeof(AppLayerProtoDetectPPTestDataPort),
},
};
if (AppLayerProtoDetectPPTestData(alpd_ctx.ctx_pp, ip_proto,
sizeof(ip_proto) / sizeof(AppLayerProtoDetectPPTestDataIPProto)) == 0) {
goto end;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
result = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
AppLayerProtoDetectDeSetup();
AppLayerProtoDetectUnittestCtxRestore();
return result;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
/** \test test if the engine detect the proto and match with it */
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static int AppLayerProtoDetectTest16(void)
{
int result = 0;
Flow *f = NULL;
HtpState *http_state = NULL;
uint8_t http_buf1[] = "POST /one HTTP/1.0\r\n"
"User-Agent: Mozilla/1.0\r\n"
"Cookie: hellocatch\r\n\r\n";
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
TcpSession ssn;
Packet *p = NULL;
Signature *s = NULL;
ThreadVars tv;
DetectEngineThreadCtx *det_ctx = NULL;
DetectEngineCtx *de_ctx = NULL;
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
memset(&tv, 0, sizeof(ThreadVars));
memset(&ssn, 0, sizeof(TcpSession));
p = UTHBuildPacket(NULL, 0, IPPROTO_TCP);
if (p == NULL) {
printf("packet setup failed: ");
goto end;
}
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
if (f == NULL) {
printf("flow setup failed: ");
goto end;
}
f->protoctx = &ssn;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
f->proto = IPPROTO_TCP;
p->flow = f;
p->flowflags |= FLOW_PKT_TOSERVER;
p->flowflags |= FLOW_PKT_ESTABLISHED;
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
f->alproto = ALPROTO_HTTP;
StreamTcpInitConfig(TRUE);
de_ctx = DetectEngineCtxInit();
if (de_ctx == NULL) {
goto end;
}
de_ctx->mpm_matcher = MPM_B2G;
de_ctx->flags |= DE_QUIET;
s = de_ctx->sig_list = SigInit(de_ctx, "alert http any any -> any any "
"(msg:\"Test content option\"; "
"sid:1;)");
if (s == NULL) {
goto end;
}
SigGroupBuild(de_ctx);
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
SCMutexLock(&f->m);
int r = AppLayerParserParse(alp_tctx, f, ALPROTO_HTTP, STREAM_TOSERVER, http_buf1, http_buf1_len);
if (r != 0) {
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
SCMutexUnlock(&f->m);
goto end;
}
SCMutexUnlock(&f->m);
http_state = f->alstate;
if (http_state == NULL) {
printf("no http state: ");
goto end;
}
/* do detect */
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
if (!PacketAlertCheck(p, 1)) {
printf("sig 1 didn't alert, but it should: ");
goto end;
}
result = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alp_tctx != NULL)
AppLayerParserThreadCtxFree(alp_tctx);
if (det_ctx != NULL)
DetectEngineThreadCtxDeinit(&tv, det_ctx);
if (de_ctx != NULL)
SigGroupCleanup(de_ctx);
if (de_ctx != NULL)
DetectEngineCtxFree(de_ctx);
StreamTcpFreeConfig(TRUE);
UTHFreePackets(&p, 1);
UTHFreeFlow(f);
return result;
}
/** \test test if the engine detect the proto on a non standar port
* and match with it */
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static int AppLayerProtoDetectTest17(void)
{
int result = 0;
Flow *f = NULL;
HtpState *http_state = NULL;
uint8_t http_buf1[] = "POST /one HTTP/1.0\r\n"
"User-Agent: Mozilla/1.0\r\n"
"Cookie: hellocatch\r\n\r\n";
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
TcpSession ssn;
Packet *p = NULL;
Signature *s = NULL;
ThreadVars tv;
DetectEngineThreadCtx *det_ctx = NULL;
DetectEngineCtx *de_ctx = NULL;
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
memset(&tv, 0, sizeof(ThreadVars));
memset(&ssn, 0, sizeof(TcpSession));
p = UTHBuildPacketSrcDstPorts(http_buf1, http_buf1_len, IPPROTO_TCP, 12345, 88);
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
if (f == NULL)
goto end;
f->protoctx = &ssn;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
f->proto = IPPROTO_TCP;
p->flow = f;
p->flowflags |= FLOW_PKT_TOSERVER;
p->flowflags |= FLOW_PKT_ESTABLISHED;
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
f->alproto = ALPROTO_HTTP;
StreamTcpInitConfig(TRUE);
de_ctx = DetectEngineCtxInit();
if (de_ctx == NULL) {
goto end;
}
de_ctx->mpm_matcher = MPM_B2G;
de_ctx->flags |= DE_QUIET;
s = de_ctx->sig_list = SigInit(de_ctx, "alert http any !80 -> any any "
"(msg:\"http over non standar port\"; "
"sid:1;)");
if (s == NULL) {
goto end;
}
SigGroupBuild(de_ctx);
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
SCMutexLock(&f->m);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = AppLayerParserParse(alp_tctx, f, ALPROTO_HTTP, STREAM_TOSERVER, http_buf1, http_buf1_len);
if (r != 0) {
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
SCMutexUnlock(&f->m);
goto end;
}
SCMutexUnlock(&f->m);
http_state = f->alstate;
if (http_state == NULL) {
printf("no http state: ");
goto end;
}
/* do detect */
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
if (!PacketAlertCheck(p, 1)) {
printf("sig 1 didn't alert, but it should: ");
goto end;
}
result = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alp_tctx != NULL)
AppLayerParserThreadCtxFree(alp_tctx);
if (det_ctx != NULL)
DetectEngineThreadCtxDeinit(&tv, det_ctx);
if (de_ctx != NULL)
SigGroupCleanup(de_ctx);
if (de_ctx != NULL)
DetectEngineCtxFree(de_ctx);
StreamTcpFreeConfig(TRUE);
UTHFreePackets(&p, 1);
UTHFreeFlow(f);
return result;
}
/** \test test if the engine detect the proto and doesn't match
* because the sig expects another proto (ex ftp)*/
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static int AppLayerProtoDetectTest18(void)
{
int result = 0;
Flow *f = NULL;
HtpState *http_state = NULL;
uint8_t http_buf1[] = "POST /one HTTP/1.0\r\n"
"User-Agent: Mozilla/1.0\r\n"
"Cookie: hellocatch\r\n\r\n";
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
TcpSession ssn;
Packet *p = NULL;
Signature *s = NULL;
ThreadVars tv;
DetectEngineThreadCtx *det_ctx = NULL;
DetectEngineCtx *de_ctx = NULL;
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
memset(&tv, 0, sizeof(ThreadVars));
memset(&ssn, 0, sizeof(TcpSession));
p = UTHBuildPacket(http_buf1, http_buf1_len, IPPROTO_TCP);
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
if (f == NULL)
goto end;
f->protoctx = &ssn;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
f->proto = IPPROTO_TCP;
p->flow = f;
p->flowflags |= FLOW_PKT_TOSERVER;
p->flowflags |= FLOW_PKT_ESTABLISHED;
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
f->alproto = ALPROTO_HTTP;
StreamTcpInitConfig(TRUE);
de_ctx = DetectEngineCtxInit();
if (de_ctx == NULL) {
goto end;
}
de_ctx->mpm_matcher = MPM_B2G;
de_ctx->flags |= DE_QUIET;
s = de_ctx->sig_list = SigInit(de_ctx, "alert ftp any any -> any any "
"(msg:\"Test content option\"; "
"sid:1;)");
if (s == NULL) {
goto end;
}
SigGroupBuild(de_ctx);
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
SCMutexLock(&f->m);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = AppLayerParserParse(alp_tctx, f, ALPROTO_HTTP, STREAM_TOSERVER, http_buf1, http_buf1_len);
if (r != 0) {
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
SCMutexUnlock(&f->m);
goto end;
}
SCMutexUnlock(&f->m);
http_state = f->alstate;
if (http_state == NULL) {
printf("no http state: ");
goto end;
}
/* do detect */
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
if (PacketAlertCheck(p, 1)) {
printf("sig 1 alerted, but it should not (it's not ftp): ");
goto end;
}
result = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alp_tctx != NULL)
AppLayerParserThreadCtxFree(alp_tctx);
if (det_ctx != NULL)
DetectEngineThreadCtxDeinit(&tv, det_ctx);
if (de_ctx != NULL)
SigGroupCleanup(de_ctx);
if (de_ctx != NULL)
DetectEngineCtxFree(de_ctx);
StreamTcpFreeConfig(TRUE);
UTHFreePackets(&p, 1);
UTHFreeFlow(f);
return result;
}
/** \test test if the engine detect the proto and doesn't match
* because the packet has another proto (ex ftp) */
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static int AppLayerProtoDetectTest19(void)
{
int result = 0;
Flow *f = NULL;
uint8_t http_buf1[] = "MPUT one\r\n";
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
TcpSession ssn;
Packet *p = NULL;
Signature *s = NULL;
ThreadVars tv;
DetectEngineThreadCtx *det_ctx = NULL;
DetectEngineCtx *de_ctx = NULL;
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
memset(&tv, 0, sizeof(ThreadVars));
memset(&ssn, 0, sizeof(TcpSession));
p = UTHBuildPacketSrcDstPorts(http_buf1, http_buf1_len, IPPROTO_TCP, 12345, 88);
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
if (f == NULL)
goto end;
f->protoctx = &ssn;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
f->proto = IPPROTO_TCP;
p->flow = f;
p->flowflags |= FLOW_PKT_TOSERVER;
p->flowflags |= FLOW_PKT_ESTABLISHED;
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
f->alproto = ALPROTO_FTP;
StreamTcpInitConfig(TRUE);
de_ctx = DetectEngineCtxInit();
if (de_ctx == NULL) {
goto end;
}
de_ctx->mpm_matcher = MPM_B2G;
de_ctx->flags |= DE_QUIET;
s = de_ctx->sig_list = SigInit(de_ctx, "alert http any !80 -> any any "
"(msg:\"http over non standar port\"; "
"sid:1;)");
if (s == NULL) {
goto end;
}
SigGroupBuild(de_ctx);
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
SCMutexLock(&f->m);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = AppLayerParserParse(alp_tctx, f, ALPROTO_FTP, STREAM_TOSERVER, http_buf1, http_buf1_len);
if (r != 0) {
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
SCMutexUnlock(&f->m);
goto end;
}
SCMutexUnlock(&f->m);
/* do detect */
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
if (PacketAlertCheck(p, 1)) {
printf("sig 1 alerted, but it should not (it's ftp): ");
goto end;
}
result = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alp_tctx != NULL)
AppLayerParserThreadCtxFree(alp_tctx);
if (det_ctx != NULL)
DetectEngineThreadCtxDeinit(&tv, det_ctx);
if (de_ctx != NULL)
SigGroupCleanup(de_ctx);
if (de_ctx != NULL)
DetectEngineCtxFree(de_ctx);
StreamTcpFreeConfig(TRUE);
UTHFreePackets(&p, 1);
UTHFreeFlow(f);
return result;
}
/** \test test if the engine detect the proto and match with it
* and also against a content option */
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
static int AppLayerProtoDetectTest20(void)
{
int result = 0;
Flow *f = NULL;
uint8_t http_buf1[] = "POST /one HTTP/1.0\r\n"
"User-Agent: Mozilla/1.0\r\n"
"Cookie: hellocatch\r\n\r\n";
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
TcpSession ssn;
Packet *p = NULL;
Signature *s = NULL;
ThreadVars tv;
DetectEngineThreadCtx *det_ctx = NULL;
DetectEngineCtx *de_ctx = NULL;
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
memset(&tv, 0, sizeof(ThreadVars));
memset(&ssn, 0, sizeof(TcpSession));
p = UTHBuildPacket(http_buf1, http_buf1_len, IPPROTO_TCP);
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
if (f == NULL)
goto end;
f->protoctx = &ssn;
p->flow = f;
p->flowflags |= FLOW_PKT_TOSERVER;
p->flowflags |= FLOW_PKT_ESTABLISHED;
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
f->alproto = ALPROTO_HTTP;
f->proto = IPPROTO_TCP;
p->flags |= PKT_STREAM_ADD;
p->flags |= PKT_STREAM_EOF;
de_ctx = DetectEngineCtxInit();
if (de_ctx == NULL) {
goto end;
}
StreamTcpInitConfig(TRUE);
StreamMsg *stream_msg = StreamMsgGetFromPool();
if (stream_msg == NULL) {
printf("no stream_msg: ");
goto end;
}
memcpy(stream_msg->data.data, http_buf1, http_buf1_len);
stream_msg->data.data_len = http_buf1_len;
ssn.toserver_smsg_head = stream_msg;
ssn.toserver_smsg_tail = stream_msg;
de_ctx->mpm_matcher = MPM_B2G;
de_ctx->flags |= DE_QUIET;
s = de_ctx->sig_list = SigInit(de_ctx, "alert http any any -> any any "
"(msg:\"Test content option\"; "
"content:\"one\"; sid:1;)");
if (s == NULL) {
goto end;
}
SigGroupBuild(de_ctx);
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
SCMutexLock(&f->m);
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
int r = AppLayerParserParse(alp_tctx, f, ALPROTO_HTTP, STREAM_TOSERVER, http_buf1, http_buf1_len);
if (r != 0) {
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
SCMutexUnlock(&f->m);
goto end;
}
SCMutexUnlock(&f->m);
/* do detect */
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
if (!PacketAlertCheck(p, 1)) {
printf("sig 1 didn't alert, but it should: ");
goto end;
}
result = 1;
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
end:
if (alp_tctx != NULL)
AppLayerParserThreadCtxFree(alp_tctx);
if (det_ctx != NULL)
DetectEngineThreadCtxDeinit(&tv, det_ctx);
if (de_ctx != NULL)
SigGroupCleanup(de_ctx);
if (de_ctx != NULL)
DetectEngineCtxFree(de_ctx);
StreamTcpFreeConfig(TRUE);
UTHFreePackets(&p, 1);
UTHFreeFlow(f);
return result;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
void AppLayerProtoDetectUnittestsRegister(void)
{
SCEnter();
UtRegisterTest("AppLayerProtoDetectTest01", AppLayerProtoDetectTest01, 1);
UtRegisterTest("AppLayerProtoDetectTest02", AppLayerProtoDetectTest02, 1);
UtRegisterTest("AppLayerProtoDetectTest03", AppLayerProtoDetectTest03, 1);
UtRegisterTest("AppLayerProtoDetectTest04", AppLayerProtoDetectTest04, 1);
UtRegisterTest("AppLayerProtoDetectTest05", AppLayerProtoDetectTest05, 1);
UtRegisterTest("AppLayerProtoDetectTest06", AppLayerProtoDetectTest06, 1);
UtRegisterTest("AppLayerProtoDetectTest07", AppLayerProtoDetectTest07, 1);
UtRegisterTest("AppLayerProtoDetectTest08", AppLayerProtoDetectTest08, 1);
UtRegisterTest("AppLayerProtoDetectTest09", AppLayerProtoDetectTest09, 1);
UtRegisterTest("AppLayerProtoDetectTest10", AppLayerProtoDetectTest10, 1);
UtRegisterTest("AppLayerProtoDetectTest11", AppLayerProtoDetectTest11, 1);
UtRegisterTest("AppLayerProtoDetectTest12", AppLayerProtoDetectTest12, 1);
UtRegisterTest("AppLayerProtoDetectTest13", AppLayerProtoDetectTest13, 1);
UtRegisterTest("AppLayerProtoDetectTest14", AppLayerProtoDetectTest14, 1);
UtRegisterTest("AppLayerProtoDetectTest15", AppLayerProtoDetectTest15, 1);
UtRegisterTest("AppLayerProtoDetectTest16", AppLayerProtoDetectTest16, 1);
UtRegisterTest("AppLayerProtoDetectTest17", AppLayerProtoDetectTest17, 1);
UtRegisterTest("AppLayerProtoDetectTest18", AppLayerProtoDetectTest18, 1);
UtRegisterTest("AppLayerProtoDetectTest19", AppLayerProtoDetectTest19, 1);
UtRegisterTest("AppLayerProtoDetectTest20", AppLayerProtoDetectTest20, 1);
SCReturn;
}
App layer API rewritten. The main files in question are: app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch]. Things addressed in this commit: - Brings out a proper separation between protocol detection phase and the parser phase. - The dns app layer now is registered such that we don't use "dnstcp" and "dnsudp" in the rules. A user who previously wrote a rule like this - "alert dnstcp....." or "alert dnsudp....." would now have to use, alert dns (ipproto:tcp;) or alert udp (app-layer-protocol:dns;) or alert ip (ipproto:udp; app-layer-protocol:dns;) The same rules extend to other another such protocol, dcerpc. - The app layer parser api now takes in the ipproto while registering callbacks. - The app inspection/detection engine also takes an ipproto. - All app layer parser functions now take direction as STREAM_TOSERVER or STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the functions. - FlowInitialize() and FlowRecycle() now resets proto to 0. This is needed by unittests, which would try to clean the flow, and that would call the api, AppLayerParserCleanupParserState(), which would try to clean the app state, but the app layer now needs an ipproto to figure out which api to internally call to clean the state, and if the ipproto is 0, it would return without trying to clean the state. - A lot of unittests are now updated where if they are using a flow and they need to use the app layer, we would set a flow ipproto. - The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
#endif /* UNITTESTS */