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-parser.c

1884 lines
59 KiB
C

/* Copyright (C) 2007-2020 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.
*/
/**
* \file
*
* \author Victor Julien <victor@inliniac.net>
*
* Generic App-layer parsing functions.
*/
#include "suricata-common.h"
#include "debug.h"
#include "util-unittest.h"
#include "decode.h"
#include "threads.h"
#include "util-print.h"
#include "util-pool.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 "detect-engine-state.h"
#include "detect-engine-port.h"
#include "stream-tcp.h"
#include "stream-tcp-private.h"
#include "stream.h"
#include "stream-tcp-reassemble.h"
#include "app-layer.h"
#include "app-layer-protos.h"
#include "app-layer-parser.h"
#include "app-layer-smb.h"
#include "app-layer-htp.h"
#include "app-layer-ftp.h"
#include "app-layer-ssl.h"
#include "app-layer-ssh.h"
#include "app-layer-smtp.h"
#include "app-layer-modbus.h"
#include "app-layer-enip.h"
#include "app-layer-dnp3.h"
#include "app-layer-nfs-tcp.h"
#include "app-layer-nfs-udp.h"
#include "app-layer-ntp.h"
#include "app-layer-tftp.h"
#include "app-layer-ike.h"
#include "app-layer-krb5.h"
#include "app-layer-snmp.h"
#include "app-layer-sip.h"
#include "app-layer-rfb.h"
#include "app-layer-mqtt.h"
#include "app-layer-template.h"
#include "app-layer-template-rust.h"
#include "app-layer-rdp.h"
#include "app-layer-http2.h"
#include "conf.h"
#include "util-spm.h"
#include "util-debug.h"
#include "decode-events.h"
#include "util-unittest-helper.h"
#include "util-validate.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"
#include "rust.h"
struct AppLayerParserThreadCtx_ {
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 *alproto_local_storage[FLOW_PROTO_MAX][ALPROTO_MAX];
};
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 App layer protocol parser context.
*/
typedef struct AppLayerParserProtoCtx_
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
{
/* 0 - to_server, 1 - to_client. */
AppLayerParserFPtr Parser[2];
bool logger;
uint32_t logger_bits; /**< registered loggers for this 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
void *(*StateAlloc)(void *, 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
void (*StateFree)(void *);
void (*StateTransactionFree)(void *, uint64_t);
void *(*LocalStorageAlloc)(void);
void (*LocalStorageFree)(void *);
void (*Truncate)(void *, uint8_t);
FileContainer *(*StateGetFiles)(void *, uint8_t);
AppLayerDecoderEvents *(*StateGetEvents)(void *);
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 (*StateGetProgress)(void *alstate, uint8_t direction);
uint64_t (*StateGetTxCnt)(void *alstate);
void *(*StateGetTx)(void *alstate, uint64_t tx_id);
AppLayerGetTxIteratorFunc StateGetTxIterator;
int complete_ts;
int complete_tc;
int (*StateGetEventInfoById)(int event_id, const char **event_name,
AppLayerEventType *event_type);
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 (*StateGetEventInfo)(const char *event_name,
int *event_id, AppLayerEventType *event_type);
DetectEngineState *(*GetTxDetectState)(void *tx);
int (*SetTxDetectState)(void *tx, DetectEngineState *);
AppLayerTxData *(*GetTxData)(void *tx);
bool (*ApplyTxConfig)(void *state, void *tx, int mode, AppLayerTxConfig);
void (*SetStreamDepthFlag)(void *tx, uint8_t flags);
/* each app-layer has its own value */
uint32_t stream_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
/* Indicates the direction the parser is ready to see the data
* the first time for a flow. Values accepted -
* STREAM_TOSERVER, STREAM_TOCLIENT */
uint8_t first_data_dir;
/* Option flags such as supporting gaps or not. */
uint32_t option_flags;
/* coccinelle: AppLayerParserProtoCtx:option_flags:APP_LAYER_PARSER_OPT_ */
uint32_t internal_flags;
/* coccinelle: AppLayerParserProtoCtx:internal_flags:APP_LAYER_PARSER_INT_ */
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 UNITTESTS
void (*RegisterUnittests)(void);
#endif
} AppLayerParserProtoCtx;
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 AppLayerParserCtx_ {
AppLayerParserProtoCtx ctxs[FLOW_PROTO_MAX][ALPROTO_MAX];
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
} AppLayerParserCtx;
struct AppLayerParserState_ {
/* coccinelle: AppLayerParserState:flags:APP_LAYER_PARSER_ */
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 flags;
/* Indicates the current transaction that is being inspected.
* We have a var per direction. */
uint64_t inspect_id[2];
/* Indicates the current transaction being logged. Unlike inspect_id,
* we don't need a var per direction since we don't log a transaction
* unless we have the entire transaction. */
uint64_t log_id;
uint64_t min_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
/* Used to store decoder events. */
AppLayerDecoderEvents *decoder_events;
};
#ifdef UNITTESTS
void UTHAppLayerParserStateGetIds(void *ptr, uint64_t *i1, uint64_t *i2, uint64_t *log, uint64_t *min)
{
struct AppLayerParserState_ *s = ptr;
*i1 = s->inspect_id[0];
*i2 = s->inspect_id[1];
*log = s->log_id;
*min = s->min_id;
}
#endif
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 global version of the parser context.
* Post 2.0 let's look at changing this to move it out to app-layer.c. */
static AppLayerParserCtx alp_ctx;
int AppLayerParserProtoIsRegistered(uint8_t ipproto, AppProto alproto)
{
uint8_t ipproto_map = FlowGetProtoMapping(ipproto);
return (alp_ctx.ctxs[ipproto_map][alproto].StateAlloc != NULL) ? 1 : 0;
}
AppLayerParserState *AppLayerParserStateAlloc(void)
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();
AppLayerParserState *pstate = (AppLayerParserState *)SCMalloc(sizeof(*pstate));
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 (pstate == NULL)
goto end;
memset(pstate, 0, sizeof(*pstate));
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(pstate, "AppLayerParserState");
}
void AppLayerParserStateFree(AppLayerParserState *pstate)
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();
if (pstate->decoder_events != NULL)
AppLayerDecoderEventsFreeEvents(&pstate->decoder_events);
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(pstate);
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
int AppLayerParserSetup(void)
{
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();
memset(&alp_ctx, 0, sizeof(alp_ctx));
SCReturnInt(0);
}
void AppLayerParserPostStreamSetup(void)
{
AppProto alproto = 0;
int flow_proto = 0;
/* lets set a default value for stream_depth */
for (flow_proto = 0; flow_proto < FLOW_PROTO_DEFAULT; flow_proto++) {
for (alproto = 0; alproto < ALPROTO_MAX; alproto++) {
if (!(alp_ctx.ctxs[flow_proto][alproto].internal_flags &
APP_LAYER_PARSER_INT_STREAM_DEPTH_SET)) {
alp_ctx.ctxs[flow_proto][alproto].stream_depth =
stream_config.reassembly_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
int AppLayerParserDeSetup(void)
{
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();
FTPParserCleanup();
SMTPParserCleanup();
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
SCReturnInt(0);
}
AppLayerParserThreadCtx *AppLayerParserThreadCtxAlloc(void)
{
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();
AppProto alproto = 0;
int flow_proto = 0;
AppLayerParserThreadCtx *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
tctx = SCMalloc(sizeof(*tctx));
if (tctx == NULL)
goto end;
memset(tctx, 0, sizeof(*tctx));
for (flow_proto = 0; flow_proto < FLOW_PROTO_DEFAULT; flow_proto++) {
for (alproto = 0; alproto < ALPROTO_MAX; alproto++) {
uint8_t ipproto = FlowGetReverseProtoMapping(flow_proto);
tctx->alproto_local_storage[flow_proto][alproto] =
AppLayerParserGetProtocolParserLocalStorage(ipproto, 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:
SCReturnPtr(tctx, "void *");
}
void AppLayerParserThreadCtxFree(AppLayerParserThreadCtx *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
SCEnter();
AppProto alproto = 0;
int flow_proto = 0;
for (flow_proto = 0; flow_proto < FLOW_PROTO_DEFAULT; flow_proto++) {
for (alproto = 0; alproto < ALPROTO_MAX; alproto++) {
uint8_t ipproto = FlowGetReverseProtoMapping(flow_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
AppLayerParserDestroyProtocolParserLocalStorage(ipproto, alproto,
tctx->alproto_local_storage[flow_proto][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
}
}
SCFree(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
SCReturn;
}
/** \brief check if a parser is enabled in the config
* Returns enabled always if: were running unittests
*/
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 AppLayerParserConfParserEnabled(const char *ipproto,
const char *alproto_name)
{
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();
int enabled = 1;
char param[100];
ConfNode *node;
int 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
if (RunmodeIsUnittests())
goto enabled;
r = snprintf(param, sizeof(param), "%s%s%s", "app-layer.protocols.",
alproto_name, ".enabled");
if (r < 0) {
FatalError(SC_ERR_FATAL, "snprintf 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
} else if (r > (int)sizeof(param)) {
FatalError(SC_ERR_FATAL, "buffer not big enough to write param.");
}
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
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, ".enabled");
if (r < 0) {
FatalError(SC_ERR_FATAL, "snprintf 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
} else if (r > (int)sizeof(param)) {
FatalError(SC_ERR_FATAL, "buffer not big enough to write param.");
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
}
node = ConfGetNode(param);
if (node == NULL) {
SCLogDebug("Entry for %s not found.", param);
goto enabled;
}
}
if (ConfValIsTrue(node->val)) {
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 enabled;
} else if (ConfValIsFalse(node->val)) {
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 disabled;
} else if (strcasecmp(node->val, "detection-only") == 0) {
goto disabled;
} 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
SCLogError(SC_ERR_FATAL, "Invalid value found for %s.", param);
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
disabled:
enabled = 0;
enabled:
SCReturnInt(enabled);
}
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
/***** Parser related registration *****/
int AppLayerParserRegisterParser(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
uint8_t direction,
AppLayerParserFPtr Parser)
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
Parser[(direction & STREAM_TOSERVER) ? 0 : 1] = Parser;
SCReturnInt(0);
}
void AppLayerParserRegisterParserAcceptableDataDirection(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
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
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].first_data_dir |=
(direction & (STREAM_TOSERVER | STREAM_TOCLIENT));
SCReturn;
}
void AppLayerParserRegisterOptionFlags(uint8_t ipproto, AppProto alproto,
uint32_t flags)
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].option_flags |= flags;
SCReturn;
}
uint32_t AppLayerParserGetOptionFlags(uint8_t protomap, AppProto alproto)
{
SCEnter();
SCReturnUInt(alp_ctx.ctxs[protomap][alproto].option_flags);
}
void AppLayerParserRegisterStateFuncs(uint8_t ipproto, AppProto alproto,
void *(*StateAlloc)(void *, AppProto), void (*StateFree)(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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateAlloc =
StateAlloc;
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateFree =
StateFree;
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;
}
void AppLayerParserRegisterLocalStorageFunc(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
void *(*LocalStorageAlloc)(void),
void (*LocalStorageFree)(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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].LocalStorageAlloc =
LocalStorageAlloc;
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].LocalStorageFree =
LocalStorageFree;
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;
}
void AppLayerParserRegisterGetFilesFunc(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
FileContainer *(*StateGetFiles)(void *, uint8_t))
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateGetFiles =
StateGetFiles;
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;
}
void AppLayerParserRegisterGetEventsFunc(uint8_t ipproto, AppProto alproto,
AppLayerDecoderEvents *(*StateGetEvents)(void *))
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateGetEvents =
StateGetEvents;
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;
}
void AppLayerParserRegisterLoggerBits(uint8_t ipproto, AppProto alproto, LoggerId bits)
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].logger_bits = bits;
SCReturn;
}
void AppLayerParserRegisterLogger(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
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].logger = true;
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;
}
void AppLayerParserRegisterTruncateFunc(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
void (*Truncate)(void *, uint8_t))
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].Truncate = Truncate;
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;
}
void AppLayerParserRegisterGetStateProgressFunc(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
int (*StateGetProgress)(void *alstate, uint8_t direction))
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetProgress = StateGetProgress;
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;
}
void AppLayerParserRegisterTxFreeFunc(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
void (*StateTransactionFree)(void *, uint64_t))
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateTransactionFree = StateTransactionFree;
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;
}
void AppLayerParserRegisterGetTxCnt(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
uint64_t (*StateGetTxCnt)(void *alstate))
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetTxCnt = StateGetTxCnt;
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;
}
void AppLayerParserRegisterGetTx(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
void *(StateGetTx)(void *alstate, uint64_t tx_id))
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetTx = StateGetTx;
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;
}
void AppLayerParserRegisterGetTxIterator(uint8_t ipproto, AppProto alproto,
AppLayerGetTxIteratorFunc Func)
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateGetTxIterator = Func;
SCReturn;
}
void AppLayerParserRegisterStateProgressCompletionStatus(
AppProto alproto, const int ts, const int tc)
{
BUG_ON(ts == 0);
BUG_ON(tc == 0);
BUG_ON(!AppProtoIsValid(alproto));
BUG_ON(alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_ts != 0 &&
alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_ts != ts);
BUG_ON(alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_tc != 0 &&
alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_tc != tc);
alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_ts = ts;
alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_tc = tc;
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 AppLayerParserRegisterGetEventInfoById(uint8_t ipproto, AppProto alproto,
int (*StateGetEventInfoById)(int event_id, const char **event_name,
AppLayerEventType *event_type))
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetEventInfoById = StateGetEventInfoById;
SCReturn;
}
void AppLayerParserRegisterGetEventInfo(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
int (*StateGetEventInfo)(const char *event_name, int *event_id,
AppLayerEventType *event_type))
{
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
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetEventInfo = StateGetEventInfo;
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;
}
void AppLayerParserRegisterDetectStateFuncs(uint8_t ipproto, AppProto alproto,
DetectEngineState *(*GetTxDetectState)(void *tx),
int (*SetTxDetectState)(void *tx, DetectEngineState *))
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].GetTxDetectState = GetTxDetectState;
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].SetTxDetectState = SetTxDetectState;
SCReturn;
}
void AppLayerParserRegisterTxDataFunc(uint8_t ipproto, AppProto alproto,
AppLayerTxData *(*GetTxData)(void *tx))
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].GetTxData = GetTxData;
SCReturn;
}
void AppLayerParserRegisterApplyTxConfigFunc(uint8_t ipproto, AppProto alproto,
bool (*ApplyTxConfig)(void *state, void *tx, int mode, AppLayerTxConfig))
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].ApplyTxConfig = ApplyTxConfig;
SCReturn;
}
void AppLayerParserRegisterSetStreamDepthFlag(uint8_t ipproto, AppProto alproto,
void (*SetStreamDepthFlag)(void *tx, uint8_t flags))
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].SetStreamDepthFlag = SetStreamDepthFlag;
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
/***** Get and transaction functions *****/
void *AppLayerParserGetProtocolParserLocalStorage(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
{
SCEnter();
void * r = 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 (alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
LocalStorageAlloc != NULL)
{
r = alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
LocalStorageAlloc();
}
SCReturnPtr(r, "void *");
}
void AppLayerParserDestroyProtocolParserLocalStorage(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
void *local_data)
{
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 (alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
LocalStorageFree != NULL)
{
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
LocalStorageFree(local_data);
}
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;
}
/** \brief default tx iterator
*
* Used if the app layer parser doesn't register its own iterator.
* Simply walks the tx_id space until it finds a tx. Uses 'state' to
* keep track of where it left off.
*
* \retval txptr or NULL if no more txs in list
*/
static AppLayerGetTxIterTuple AppLayerDefaultGetTxIterator(
const uint8_t ipproto, const AppProto alproto,
void *alstate, uint64_t min_tx_id, uint64_t max_tx_id,
AppLayerGetTxIterState *state)
{
uint64_t ustate = *(uint64_t *)state;
uint64_t tx_id = MAX(min_tx_id, ustate);
for ( ; tx_id < max_tx_id; tx_id++) {
void *tx_ptr = AppLayerParserGetTx(ipproto, alproto, alstate, tx_id);
if (tx_ptr != NULL) {
ustate = tx_id + 1;
*state = *(AppLayerGetTxIterState *)&ustate;
AppLayerGetTxIterTuple tuple = {
.tx_ptr = tx_ptr,
.tx_id = tx_id,
.has_next = (tx_id + 1 < max_tx_id),
};
SCLogDebug("tuple: %p/%"PRIu64"/%s", tuple.tx_ptr, tuple.tx_id,
tuple.has_next ? "true" : "false");
return tuple;
}
}
AppLayerGetTxIterTuple no_tuple = { NULL, 0, false };
return no_tuple;
}
AppLayerGetTxIteratorFunc AppLayerGetTxIterator(const uint8_t ipproto,
const AppProto alproto)
{
AppLayerGetTxIteratorFunc Func =
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateGetTxIterator;
return Func ? Func : AppLayerDefaultGetTxIterator;
}
uint64_t AppLayerParserGetTransactionLogId(AppLayerParserState *pstate)
{
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();
SCReturnCT((pstate == NULL) ? 0 : pstate->log_id, "uint64_t");
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 AppLayerParserSetTransactionLogId(AppLayerParserState *pstate, uint64_t tx_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
{
SCEnter();
if (pstate != NULL)
pstate->log_id = tx_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
SCReturn;
}
uint64_t AppLayerParserGetTransactionInspectId(AppLayerParserState *pstate, 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
SCEnter();
if (pstate == NULL)
SCReturnCT(0ULL, "uint64_t");
SCReturnCT(pstate->inspect_id[direction & STREAM_TOSERVER ? 0 : 1], "uint64_t");
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 uint64_t GetTxDetectFlags(AppLayerTxData *txd, const uint8_t dir)
{
uint64_t detect_flags =
(dir & STREAM_TOSERVER) ? txd->detect_flags_ts : txd->detect_flags_tc;
return detect_flags;
}
static inline void SetTxDetectFlags(AppLayerTxData *txd, const uint8_t dir, const uint64_t detect_flags)
{
if (dir & STREAM_TOSERVER) {
txd->detect_flags_ts = detect_flags;
} else {
txd->detect_flags_tc = detect_flags;
}
}
static inline uint32_t GetTxLogged(AppLayerTxData *txd)
{
return txd->logged.flags;
}
void AppLayerParserSetTransactionInspectId(const Flow *f, AppLayerParserState *pstate,
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
void *alstate, const uint8_t flags,
bool tag_txs_as_inspected)
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();
const int direction = (flags & STREAM_TOSERVER) ? 0 : 1;
const uint64_t total_txs = AppLayerParserGetTxCnt(f, alstate);
uint64_t idx = AppLayerParserGetTransactionInspectId(pstate, flags);
const int state_done_progress = AppLayerParserGetStateProgressCompletionStatus(f->alproto, flags);
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
const uint8_t ipproto = f->proto;
const AppProto alproto = f->alproto;
AppLayerGetTxIteratorFunc IterFunc = AppLayerGetTxIterator(ipproto, alproto);
AppLayerGetTxIterState state;
memset(&state, 0, sizeof(state));
SCLogDebug("called: %s, tag_txs_as_inspected %s",direction==0?"toserver":"toclient",
tag_txs_as_inspected?"true":"false");
/* mark all txs as inspected if the applayer progress is
* at the 'end state'. */
while (1) {
AppLayerGetTxIterTuple ires = IterFunc(ipproto, alproto, alstate, idx, total_txs, &state);
if (ires.tx_ptr == NULL)
break;
void *tx = ires.tx_ptr;
idx = ires.tx_id;
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
int state_progress = AppLayerParserGetStateProgress(ipproto, alproto, tx, flags);
if (state_progress < state_done_progress)
break;
AppLayerTxData *txd = AppLayerParserGetTxData(ipproto, alproto, tx);
if (txd && tag_txs_as_inspected) {
uint64_t detect_flags = GetTxDetectFlags(txd, flags);
if ((detect_flags & APP_LAYER_TX_INSPECTED_FLAG) == 0) {
detect_flags |= APP_LAYER_TX_INSPECTED_FLAG;
SetTxDetectFlags(txd, flags, detect_flags);
SCLogDebug("%p/%"PRIu64" in-order tx is done for direction %s. Flag %016"PRIx64,
tx, idx, flags & STREAM_TOSERVER ? "toserver" : "toclient", detect_flags);
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
}
}
idx++;
if (!ires.has_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
break;
}
pstate->inspect_id[direction] = idx;
SCLogDebug("inspect_id now %"PRIu64, pstate->inspect_id[direction]);
/* if necessary we flag all txs that are complete as 'inspected'
* also move inspect_id forward. */
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
if (tag_txs_as_inspected) {
/* continue at idx */
while (1) {
AppLayerGetTxIterTuple ires = IterFunc(ipproto, alproto, alstate, idx, total_txs, &state);
if (ires.tx_ptr == NULL)
break;
void *tx = ires.tx_ptr;
/* if we got a higher id than the minimum we requested, we
* skipped a bunch of 'null-txs'. Lets see if we can up the
* inspect tracker */
if (ires.tx_id > idx && pstate->inspect_id[direction] == idx) {
pstate->inspect_id[direction] = ires.tx_id;
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
}
idx = ires.tx_id;
const int state_progress = AppLayerParserGetStateProgress(ipproto, alproto, tx, flags);
if (state_progress < state_done_progress)
break;
/* txd can be NULL for HTTP sessions where the user data alloc failed */
AppLayerTxData *txd = AppLayerParserGetTxData(ipproto, alproto, tx);
if (likely(txd)) {
uint64_t detect_flags = GetTxDetectFlags(txd, flags);
if ((detect_flags & APP_LAYER_TX_INSPECTED_FLAG) == 0) {
detect_flags |= APP_LAYER_TX_INSPECTED_FLAG;
SetTxDetectFlags(txd, flags, detect_flags);
SCLogDebug("%p/%"PRIu64" out of order tx is done for direction %s. Flag %016"PRIx64,
tx, idx, flags & STREAM_TOSERVER ? "toserver" : "toclient", detect_flags);
SCLogDebug("%p/%"PRIu64" out of order tx. Update inspect_id? %"PRIu64,
tx, idx, pstate->inspect_id[direction]);
if (pstate->inspect_id[direction]+1 == idx)
pstate->inspect_id[direction] = idx;
}
} else {
if (pstate->inspect_id[direction]+1 == idx)
pstate->inspect_id[direction] = idx;
}
if (!ires.has_next)
break;
idx++;
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 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
SCReturn;
}
AppLayerDecoderEvents *AppLayerParserGetDecoderEvents(AppLayerParserState *pstate)
{
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();
SCReturnPtr(pstate->decoder_events,
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
"AppLayerDecoderEvents *");
}
void AppLayerParserSetDecoderEvents(AppLayerParserState *pstate, AppLayerDecoderEvents *devents)
{
pstate->decoder_events = devents;
}
AppLayerDecoderEvents *AppLayerParserGetEventsByTx(uint8_t ipproto, AppProto alproto,
void *tx)
{
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
AppLayerDecoderEvents *ptr = 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 (alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetEvents != NULL)
{
ptr = alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetEvents(tx);
}
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(ptr, "AppLayerDecoderEvents *");
}
FileContainer *AppLayerParserGetFiles(const Flow *f, const uint8_t direction)
{
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
FileContainer *ptr = NULL;
if (alp_ctx.ctxs[f->protomap][f->alproto].StateGetFiles != 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
{
ptr = alp_ctx.ctxs[f->protomap][f->alproto].
StateGetFiles(f->alstate, 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
SCReturnPtr(ptr, "FileContainer *");
}
#define IS_DISRUPTED(flags) ((flags) & (STREAM_DEPTH | STREAM_GAP))
extern int g_detect_disabled;
extern bool g_file_logger_enabled;
/**
* \brief remove obsolete (inspected and logged) transactions
*/
void AppLayerParserTransactionsCleanup(Flow *f)
{
SCEnter();
DEBUG_ASSERT_FLOW_LOCKED(f);
AppLayerParserProtoCtx *p = &alp_ctx.ctxs[f->protomap][f->alproto];
if (unlikely(p->StateTransactionFree == NULL))
SCReturn;
const bool has_tx_detect_flags = !g_detect_disabled;
const uint8_t ipproto = f->proto;
const AppProto alproto = f->alproto;
void * const alstate = f->alstate;
AppLayerParserState * const alparser = f->alparser;
if (alstate == NULL || alparser == NULL)
SCReturn;
const uint64_t min = alparser->min_id;
const uint64_t total_txs = AppLayerParserGetTxCnt(f, alstate);
const LoggerId logger_expectation = AppLayerParserProtocolGetLoggerBits(ipproto, alproto);
const int tx_end_state_ts = AppLayerParserGetStateProgressCompletionStatus(alproto, STREAM_TOSERVER);
const int tx_end_state_tc = AppLayerParserGetStateProgressCompletionStatus(alproto, STREAM_TOCLIENT);
const uint8_t ts_disrupt_flags = FlowGetDisruptionFlags(f, STREAM_TOSERVER);
const uint8_t tc_disrupt_flags = FlowGetDisruptionFlags(f, STREAM_TOCLIENT);
AppLayerGetTxIteratorFunc IterFunc = AppLayerGetTxIterator(ipproto, alproto);
AppLayerGetTxIterState state;
memset(&state, 0, sizeof(state));
uint64_t i = min;
uint64_t new_min = min;
SCLogDebug("start min %"PRIu64, min);
bool skipped = false;
const bool is_unidir =
AppLayerParserGetOptionFlags(f->protomap, f->alproto) & APP_LAYER_PARSER_OPT_UNIDIR_TXS;
while (1) {
AppLayerGetTxIterTuple ires = IterFunc(ipproto, alproto, alstate, i, total_txs, &state);
if (ires.tx_ptr == NULL)
break;
bool tx_skipped = false;
void *tx = ires.tx_ptr;
i = ires.tx_id; // actual tx id for the tx the IterFunc returned
SCLogDebug("%p/%"PRIu64" checking", tx, i);
const int tx_progress_tc =
AppLayerParserGetStateProgress(ipproto, alproto, tx, tc_disrupt_flags);
if (tx_progress_tc < tx_end_state_tc) {
SCLogDebug("%p/%"PRIu64" skipping: tc parser not done", tx, i);
skipped = true;
goto next;
}
const int tx_progress_ts =
AppLayerParserGetStateProgress(ipproto, alproto, tx, ts_disrupt_flags);
if (tx_progress_ts < tx_end_state_ts) {
SCLogDebug("%p/%"PRIu64" skipping: ts parser not done", tx, i);
skipped = true;
goto next;
}
AppLayerTxData *txd = AppLayerParserGetTxData(ipproto, alproto, tx);
bool inspected = false;
if (txd && has_tx_detect_flags) {
if (!IS_DISRUPTED(ts_disrupt_flags) && f->sgh_toserver != NULL) {
uint64_t detect_flags_ts = GetTxDetectFlags(txd, STREAM_TOSERVER);
if (!(detect_flags_ts & APP_LAYER_TX_INSPECTED_FLAG)) {
SCLogDebug("%p/%"PRIu64" skipping: TS inspect not done: ts:%"PRIx64,
tx, i, detect_flags_ts);
tx_skipped = true;
} else {
inspected = true;
}
}
if (!IS_DISRUPTED(tc_disrupt_flags) && f->sgh_toclient != NULL) {
uint64_t detect_flags_tc = GetTxDetectFlags(txd, STREAM_TOCLIENT);
if (!(detect_flags_tc & APP_LAYER_TX_INSPECTED_FLAG)) {
SCLogDebug("%p/%"PRIu64" skipping: TC inspect not done: tc:%"PRIx64,
tx, i, detect_flags_tc);
tx_skipped = true;
} else {
inspected = true;
}
}
}
// If not a unidirectional transaction both sides are required to have
// been inspected.
if (!is_unidir && tx_skipped) {
SCLogDebug("%p/%" PRIu64 " !is_unidir && tx_skipped", tx, i);
skipped = true;
goto next;
}
// If this is a unidirectional transaction require only one side to be
// inspected, which the inspected flag tells us. This is also guarded
// with skip to limit this check to transactions that actually had the
// tx inspected flag checked.
if (is_unidir && tx_skipped && !inspected) {
SCLogDebug("%p/%" PRIu64 " is_unidir && tx_skipped && !inspected", tx, i);
skipped = true;
goto next;
}
if (txd && logger_expectation != 0) {
LoggerId tx_logged = GetTxLogged(txd);
if (tx_logged != logger_expectation) {
SCLogDebug("%p/%"PRIu64" skipping: logging not done: want:%"PRIx32", have:%"PRIx32,
tx, i, logger_expectation, tx_logged);
skipped = true;
goto next;
}
}
/* if file logging is enabled, we keep a tx active while some of the files aren't
* logged yet. */
if (txd && txd->files_opened && g_file_logger_enabled) {
if (txd->files_opened != txd->files_logged) {
skipped = true;
goto next;
}
}
/* if we are here, the tx can be freed. */
p->StateTransactionFree(alstate, i);
SCLogDebug("%p/%"PRIu64" freed", tx, i);
/* if we didn't skip any tx so far, up the minimum */
SCLogDebug("skipped? %s i %"PRIu64", new_min %"PRIu64, skipped ? "true" : "false", i, new_min);
if (!skipped)
new_min = i + 1;
SCLogDebug("final i %"PRIu64", new_min %"PRIu64, i, new_min);
next:
if (!ires.has_next) {
/* this was the last tx. See if we skipped any. If not
* we removed all and can update the minimum to the max
* id. */
SCLogDebug("no next: cur tx i %"PRIu64", total %"PRIu64, i, total_txs);
if (!skipped) {
new_min = total_txs;
SCLogDebug("no next: cur tx i %"PRIu64", total %"PRIu64": "
"new_min updated to %"PRIu64, i, total_txs, new_min);
}
break;
}
i++;
}
/* see if we need to bring all trackers up to date. */
SCLogDebug("update f->alparser->min_id? %"PRIu64" vs %"PRIu64, new_min, alparser->min_id);
if (new_min > alparser->min_id) {
const uint64_t next_id = new_min;
alparser->min_id = next_id;
alparser->inspect_id[0] = MAX(alparser->inspect_id[0], next_id);
alparser->inspect_id[1] = MAX(alparser->inspect_id[1], next_id);
alparser->log_id = MAX(alparser->log_id, next_id);
SCLogDebug("updated f->alparser->min_id %"PRIu64, alparser->min_id);
}
SCReturn;
}
static inline int StateGetProgressCompletionStatus(const AppProto alproto, const uint8_t flags)
{
if (flags & STREAM_TOSERVER) {
return alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_ts;
} else if (flags & STREAM_TOCLIENT) {
return alp_ctx.ctxs[FLOW_PROTO_DEFAULT][alproto].complete_tc;
} else {
DEBUG_VALIDATE_BUG_ON(1);
return 0;
}
}
/**
* \brief get the progress value for a tx/protocol
*
* If the stream is disrupted, we return the 'completion' value.
*/
int AppLayerParserGetStateProgress(uint8_t ipproto, AppProto alproto,
void *alstate, uint8_t flags)
{
SCEnter();
int r;
if (unlikely(IS_DISRUPTED(flags))) {
r = StateGetProgressCompletionStatus(alproto, flags);
} else {
r = alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
StateGetProgress(alstate, flags);
}
SCReturnInt(r);
}
uint64_t AppLayerParserGetTxCnt(const Flow *f, void *alstate)
{
SCEnter();
uint64_t r = alp_ctx.ctxs[f->protomap][f->alproto].StateGetTxCnt(alstate);
SCReturnCT(r, "uint64_t");
}
void *AppLayerParserGetTx(uint8_t ipproto, AppProto alproto, void *alstate, uint64_t tx_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
SCEnter();
void *r = alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateGetTx(alstate, tx_id);
SCReturnPtr(r, "void *");
}
int AppLayerParserGetStateProgressCompletionStatus(AppProto alproto,
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
SCEnter();
int r = StateGetProgressCompletionStatus(alproto, direction);
SCReturnInt(r);
}
int AppLayerParserGetEventInfo(uint8_t ipproto, AppProto alproto, const char *event_name,
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 *event_id, AppLayerEventType *event_type)
{
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();
const int ipproto_map = FlowGetProtoMapping(ipproto);
12 years ago
int r = (alp_ctx.ctxs[ipproto_map][alproto].StateGetEventInfo == NULL) ?
-1 : alp_ctx.ctxs[ipproto_map][alproto].StateGetEventInfo(event_name, event_id, event_type);
SCReturnInt(r);
}
int AppLayerParserGetEventInfoById(uint8_t ipproto, AppProto alproto, int event_id,
const char **event_name, AppLayerEventType *event_type)
{
SCEnter();
const int ipproto_map = FlowGetProtoMapping(ipproto);
*event_name = (const char *)NULL;
int r = (alp_ctx.ctxs[ipproto_map][alproto].StateGetEventInfoById == NULL) ?
-1 : alp_ctx.ctxs[ipproto_map][alproto].StateGetEventInfoById(event_id, event_name, event_type);
SCReturnInt(r);
}
uint8_t AppLayerParserGetFirstDataDir(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
SCEnter();
uint8_t r = alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].first_data_dir;
SCReturnCT(r, "uint8_t");
}
uint64_t AppLayerParserGetTransactionActive(const Flow *f,
AppLayerParserState *pstate, 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
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
uint64_t active_id;
uint64_t log_id = pstate->log_id;
uint64_t inspect_id = pstate->inspect_id[direction & STREAM_TOSERVER ? 0 : 1];
if (alp_ctx.ctxs[f->protomap][f->alproto].logger == true) {
active_id = MIN(log_id, inspect_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
} else {
active_id = inspect_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
SCReturnCT(active_id, "uint64_t");
}
int AppLayerParserSupportsFiles(uint8_t ipproto, AppProto alproto)
{
// Custom case for only signature-only protocol so far
if (alproto == ALPROTO_HTTP) {
return AppLayerParserSupportsFiles(ipproto, ALPROTO_HTTP1) ||
AppLayerParserSupportsFiles(ipproto, ALPROTO_HTTP2);
}
if (alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].StateGetFiles != NULL)
return TRUE;
return FALSE;
}
DetectEngineState *AppLayerParserGetTxDetectState(uint8_t ipproto, AppProto alproto, void *tx)
{
SCEnter();
DetectEngineState *s = alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].GetTxDetectState(tx);
SCReturnPtr(s, "DetectEngineState");
}
int AppLayerParserSetTxDetectState(const Flow *f,
void *tx, DetectEngineState *s)
{
SCEnter();
int r = alp_ctx.ctxs[f->protomap][f->alproto].SetTxDetectState(tx, s);
SCReturnInt(r);
}
AppLayerTxData *AppLayerParserGetTxData(uint8_t ipproto, AppProto alproto, void *tx)
{
SCEnter();
AppLayerTxData *d = alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].GetTxData(tx);
SCReturnPtr(d, "AppLayerTxData");
}
void AppLayerParserApplyTxConfig(uint8_t ipproto, AppProto alproto,
void *state, void *tx, enum ConfigAction mode, AppLayerTxConfig config)
{
SCEnter();
const int ipproto_map = FlowGetProtoMapping(ipproto);
if (alp_ctx.ctxs[ipproto_map][alproto].ApplyTxConfig) {
alp_ctx.ctxs[ipproto_map][alproto].ApplyTxConfig(state, tx, mode, config);
}
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
/***** General *****/
static inline void SetEOFFlags(AppLayerParserState *pstate, const uint8_t flags)
{
if ((flags & (STREAM_EOF|STREAM_TOSERVER)) == (STREAM_EOF|STREAM_TOSERVER)) {
SCLogDebug("setting APP_LAYER_PARSER_EOF_TS");
AppLayerParserStateSetFlag(pstate, APP_LAYER_PARSER_EOF_TS);
} else if ((flags & (STREAM_EOF|STREAM_TOCLIENT)) == (STREAM_EOF|STREAM_TOCLIENT)) {
SCLogDebug("setting APP_LAYER_PARSER_EOF_TC");
AppLayerParserStateSetFlag(pstate, APP_LAYER_PARSER_EOF_TC);
}
}
/** \retval int -1 in case of unrecoverable error. App-layer tracking stops for this flow.
* \retval int 0 ok: we did not update app_progress
* \retval int 1 ok: we updated app_progress */
int AppLayerParserParse(ThreadVars *tv, AppLayerParserThreadCtx *alp_tctx, Flow *f, AppProto alproto,
uint8_t flags, const uint8_t *input, uint32_t input_len)
{
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();
#ifdef DEBUG_VALIDATION
BUG_ON(f->protomap != FlowGetProtoMapping(f->proto));
#endif
AppLayerParserState *pstate = f->alparser;
AppLayerParserProtoCtx *p = &alp_ctx.ctxs[f->protomap][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
void *alstate = NULL;
uint64_t p_tx_cnt = 0;
uint32_t consumed = input_len;
const int direction = (flags & STREAM_TOSERVER) ? 0 : 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
/* we don't have the parser registered for this protocol */
if (p->StateAlloc == 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 (flags & STREAM_GAP) {
if (!(p->option_flags & APP_LAYER_PARSER_OPT_ACCEPT_GAPS)) {
SCLogDebug("app-layer parser does not accept gaps");
if (f->alstate != NULL && !FlowChangeProto(f)) {
AppLayerParserStreamTruncated(f->proto, alproto, f->alstate,
flags);
}
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
/* Get the parser state (if any) */
if (pstate == NULL) {
f->alparser = pstate = AppLayerParserStateAlloc();
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 (pstate == NULL)
goto error;
}
SetEOFFlags(pstate, flags);
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
alstate = f->alstate;
if (alstate == NULL || FlowChangeProto(f)) {
f->alstate = alstate = p->StateAlloc(alstate, f->alproto_orig);
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 (alstate == NULL)
goto error;
SCLogDebug("alloced new app layer state %p (name %s)",
alstate, AppLayerGetProtoName(f->alproto));
} else {
SCLogDebug("using existing app layer state %p (name %s))",
alstate, AppLayerGetProtoName(f->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
p_tx_cnt = AppLayerParserGetTxCnt(f, f->alstate);
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
/* invoke the recursive parser, but only on data. We may get empty msgs on EOF */
if (input_len > 0 || (flags & STREAM_EOF)) {
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
/* invoke the parser */
AppLayerResult res = p->Parser[direction](f, alstate, pstate,
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
input, input_len,
alp_tctx->alproto_local_storage[f->protomap][alproto],
flags);
if (res.status < 0) {
goto error;
} else if (res.status > 0) {
DEBUG_VALIDATE_BUG_ON(res.consumed > input_len);
DEBUG_VALIDATE_BUG_ON(res.needed + res.consumed < input_len);
DEBUG_VALIDATE_BUG_ON(res.needed == 0);
/* incomplete is only supported for TCP */
DEBUG_VALIDATE_BUG_ON(f->proto != IPPROTO_TCP);
/* put protocol in error state on improper use of the
* return codes. */
if (res.consumed > input_len || res.needed + res.consumed < input_len) {
goto error;
}
if (f->proto == IPPROTO_TCP && f->protoctx != NULL) {
TcpSession *ssn = f->protoctx;
SCLogDebug("direction %d/%s", direction,
(flags & STREAM_TOSERVER) ? "toserver" : "toclient");
if (direction == 0) {
/* parser told us how much data it needs on top of what it
* consumed. So we need tell stream engine how much we need
* before the next call */
ssn->client.data_required = res.needed;
SCLogDebug("setting data_required %u", ssn->client.data_required);
} else {
/* parser told us how much data it needs on top of what it
* consumed. So we need tell stream engine how much we need
* before the next call */
ssn->server.data_required = res.needed;
SCLogDebug("setting data_required %u", ssn->server.data_required);
}
}
consumed = res.consumed;
}
}
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
/* set the packets to no inspection and reassembly if required */
if (pstate->flags & APP_LAYER_PARSER_NO_INSPECTION) {
AppLayerParserSetEOF(pstate);
FlowSetNoPayloadInspectionFlag(f);
if (f->proto == IPPROTO_TCP) {
StreamTcpDisableAppLayer(f);
/* Set the no reassembly flag for both the stream in this TcpSession */
if (pstate->flags & APP_LAYER_PARSER_NO_REASSEMBLY) {
/* Used only if it's TCP */
TcpSession *ssn = f->protoctx;
if (ssn != NULL) {
StreamTcpSetSessionNoReassemblyFlag(ssn, 0);
StreamTcpSetSessionNoReassemblyFlag(ssn, 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
}
/* Set the bypass flag for both the stream in this TcpSession */
if (pstate->flags & APP_LAYER_PARSER_BYPASS_READY) {
/* Used only if it's TCP */
TcpSession *ssn = f->protoctx;
if (ssn != NULL) {
StreamTcpSetSessionBypassFlag(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
}
}
/* In cases like HeartBleed for TLS we need to inspect AppLayer but not Payload */
if (!(f->flags & FLOW_NOPAYLOAD_INSPECTION) && pstate->flags & APP_LAYER_PARSER_NO_INSPECTION_PAYLOAD) {
FlowSetNoPayloadInspectionFlag(f);
/* Set the no reassembly flag for both the stream in this TcpSession */
if (f->proto == IPPROTO_TCP) {
/* Used only if it's TCP */
TcpSession *ssn = f->protoctx;
if (ssn != NULL) {
StreamTcpSetDisableRawReassemblyFlag(ssn, 0);
StreamTcpSetDisableRawReassemblyFlag(ssn, 1);
}
}
}
/* get the diff in tx cnt for stats keeping */
uint64_t cur_tx_cnt = AppLayerParserGetTxCnt(f, f->alstate);
if (cur_tx_cnt > p_tx_cnt && tv) {
AppLayerIncTxCounter(tv, f, cur_tx_cnt - p_tx_cnt);
}
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
/* stream truncated, inform app layer */
if (flags & STREAM_DEPTH)
AppLayerParserStreamTruncated(f->proto, alproto, alstate, flags);
end:
/* update app progress */
if (consumed != input_len && f->proto == IPPROTO_TCP && f->protoctx != NULL) {
TcpSession *ssn = f->protoctx;
StreamTcpUpdateAppLayerProgress(ssn, direction, consumed);
SCReturnInt(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
SCReturnInt(0);
error:
/* Set the no app layer inspection flag for both
* the stream in this Flow */
if (f->proto == IPPROTO_TCP) {
StreamTcpDisableAppLayer(f);
}
AppLayerParserSetEOF(pstate);
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
SCReturnInt(-1);
}
void AppLayerParserSetEOF(AppLayerParserState *pstate)
{
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 (pstate == NULL)
goto end;
SCLogDebug("setting APP_LAYER_PARSER_EOF_TC and APP_LAYER_PARSER_EOF_TS");
AppLayerParserStateSetFlag(pstate, (APP_LAYER_PARSER_EOF_TS|APP_LAYER_PARSER_EOF_TC));
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;
}
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
/* return true if there are app parser decoder events. These are
* only the ones that are set during protocol detection. */
bool AppLayerParserHasDecoderEvents(AppLayerParserState *pstate)
{
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();
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
if (pstate == NULL)
return false;
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
const AppLayerDecoderEvents *decoder_events = AppLayerParserGetDecoderEvents(pstate);
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 (decoder_events && decoder_events->cnt)
detect: rewrite of the detect engine Use per tx detect_flags to track prefilter. Detect flags are used for 2 things: 1. marking tx as fully inspected 2. tracking already run prefilter (incl mpm) engines This supercedes the MpmIDs API for directionless tracking of the prefilter engines. When we have no SGH we have to flag the txs that are 'complete' as inspected as well. Special handling for the stream engine: If a rule mixes TX inspection and STREAM inspection, we can encounter the case where the rule is evaluated against multiple transactions during a single inspection run. As the stream data is exactly the same for each of those runs, it's wasteful to rerun inspection of the stream portion of the rule. This patch enables caching of the stream 'inspect engine' result in the local 'RuleMatchCandidateTx' array. This is valid only during the live of a single inspection run. Remove stateful inspection from 'mask' (SignatureMask). The mask wasn't used in most cases for those rules anyway, as there we rely on the prefilter. Add a alproto check to catch the remaining cases. When building the active non-mpm/non-prefilter list check not just the mask, but also the alproto. This especially helps stateful rules with negated mpm. Simplify AppLayerParserHasDecoderEvents usage in detection to only return true if protocol detection events are set. Other detection is done in inspect engines. Move rule group lookup and handling into it's own function. Handle 'post lookup' tasks immediately, instead of after the first detect run. The tasks were independent of the initial detection. Many cleanups and much refactoring.
8 years ago
return true;
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 we have reached here, we don't have events */
return false;
}
/** \brief simple way to globally test if a alproto is registered
* and fully enabled in the configuration.
*/
int AppLayerParserIsEnabled(AppProto alproto)
{
for (int i = 0; i < FLOW_PROTO_APPLAYER_MAX; i++) {
if (alp_ctx.ctxs[i][alproto].StateGetProgress != NULL) {
return 1;
}
}
return 0;
}
int AppLayerParserProtocolIsTxEventAware(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
SCEnter();
int ipproto_map = FlowGetProtoMapping(ipproto);
int r = (alp_ctx.ctxs[ipproto_map][alproto].StateGetEvents == NULL) ? 0 : 1;
12 years ago
SCReturnInt(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 AppLayerParserProtocolHasLogger(uint8_t ipproto, AppProto alproto)
{
SCEnter();
int ipproto_map = FlowGetProtoMapping(ipproto);
int r = (alp_ctx.ctxs[ipproto_map][alproto].logger == false) ? 0 : 1;
SCReturnInt(r);
}
LoggerId AppLayerParserProtocolGetLoggerBits(uint8_t ipproto, AppProto alproto)
{
SCEnter();
const int ipproto_map = FlowGetProtoMapping(ipproto);
LoggerId r = alp_ctx.ctxs[ipproto_map][alproto].logger_bits;
SCReturnUInt(r);
}
void AppLayerParserTriggerRawStreamReassembly(Flow *f, int 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
SCEnter();
SCLogDebug("f %p tcp %p direction %d", f, f ? f->protoctx : NULL, 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
if (f != NULL && f->protoctx != NULL)
StreamTcpReassembleTriggerRawReassembly(f->protoctx, 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
SCReturn;
}
void AppLayerParserSetStreamDepth(uint8_t ipproto, AppProto alproto, uint32_t stream_depth)
{
SCEnter();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].stream_depth = stream_depth;
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].internal_flags |=
APP_LAYER_PARSER_INT_STREAM_DEPTH_SET;
SCReturn;
}
uint32_t AppLayerParserGetStreamDepth(const Flow *f)
{
SCReturnInt(alp_ctx.ctxs[f->protomap][f->alproto].stream_depth);
}
void AppLayerParserSetStreamDepthFlag(uint8_t ipproto, AppProto alproto, void *state, uint64_t tx_id, uint8_t flags)
{
SCEnter();
void *tx = NULL;
if (state != NULL) {
if ((tx = AppLayerParserGetTx(ipproto, alproto, state, tx_id)) != NULL) {
if (alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].SetStreamDepthFlag != NULL) {
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].SetStreamDepthFlag(tx, flags);
}
}
}
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
/***** Cleanup *****/
void AppLayerParserStateProtoCleanup(
uint8_t protomap, AppProto alproto, void *alstate, AppLayerParserState *pstate)
{
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();
AppLayerParserProtoCtx *ctx = &alp_ctx.ctxs[protomap][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
if (ctx->StateFree != NULL && alstate != NULL)
ctx->StateFree(alstate);
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
/* free the app layer parser api state */
if (pstate != NULL)
AppLayerParserStateFree(pstate);
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;
}
void AppLayerParserStateCleanup(const Flow *f, void *alstate, AppLayerParserState *pstate)
{
AppLayerParserStateProtoCleanup(f->protomap, f->alproto, alstate, pstate);
}
static void ValidateParserProtoDump(AppProto alproto, uint8_t ipproto)
{
uint8_t map = FlowGetProtoMapping(ipproto);
const AppLayerParserProtoCtx *ctx = &alp_ctx.ctxs[map][alproto];
printf("ERROR: incomplete app-layer registration\n");
printf("AppLayer protocol %s ipproto %u\n", AppProtoToString(alproto), ipproto);
printf("- option flags %"PRIx32"\n", ctx->option_flags);
printf("- first_data_dir %"PRIx8"\n", ctx->first_data_dir);
printf("Mandatory:\n");
printf("- Parser[0] %p Parser[1] %p\n", ctx->Parser[0], ctx->Parser[1]);
printf("- StateAlloc %p StateFree %p\n", ctx->StateAlloc, ctx->StateFree);
printf("- StateGetTx %p StateGetTxCnt %p StateTransactionFree %p\n",
ctx->StateGetTx, ctx->StateGetTxCnt, ctx->StateTransactionFree);
printf("- GetTxData %p\n", ctx->GetTxData);
printf("- StateGetProgress %p\n", ctx->StateGetProgress);
printf("- GetTxDetectState %p SetTxDetectState %p\n", ctx->GetTxDetectState, ctx->SetTxDetectState);
printf("Optional:\n");
printf("- LocalStorageAlloc %p LocalStorageFree %p\n", ctx->LocalStorageAlloc, ctx->LocalStorageFree);
printf("- StateGetEvents %p StateGetEventInfo %p StateGetEventInfoById %p\n", ctx->StateGetEvents, ctx->StateGetEventInfo,
ctx->StateGetEventInfoById);
}
#define BOTH_SET(a, b) ((a) != NULL && (b) != NULL)
#define BOTH_SET_OR_BOTH_UNSET(a, b) (((a) == NULL && (b) == NULL) || ((a) != NULL && (b) != NULL))
#define THREE_SET_OR_THREE_UNSET(a, b, c) (((a) == NULL && (b) == NULL && (c) == NULL) || ((a) != NULL && (b) != NULL && (c) != NULL))
#define THREE_SET(a, b, c) ((a) != NULL && (b) != NULL && (c) != NULL)
static void ValidateParserProto(AppProto alproto, uint8_t ipproto)
{
uint8_t map = FlowGetProtoMapping(ipproto);
const AppLayerParserProtoCtx *ctx = &alp_ctx.ctxs[map][alproto];
if (ctx->Parser[0] == NULL && ctx->Parser[1] == NULL)
return;
if (!(BOTH_SET(ctx->Parser[0], ctx->Parser[1]))) {
goto bad;
}
if (!(BOTH_SET(ctx->StateFree, ctx->StateAlloc))) {
goto bad;
}
if (!(THREE_SET(ctx->StateGetTx, ctx->StateGetTxCnt, ctx->StateTransactionFree))) {
goto bad;
}
if (ctx->StateGetProgress == NULL) {
goto bad;
}
/* local storage is optional, but needs both set if used */
if (!(BOTH_SET_OR_BOTH_UNSET(ctx->LocalStorageAlloc, ctx->LocalStorageFree))) {
goto bad;
}
if (!(BOTH_SET(ctx->GetTxDetectState, ctx->SetTxDetectState))) {
goto bad;
}
if (ctx->GetTxData == NULL) {
goto bad;
}
return;
bad:
ValidateParserProtoDump(alproto, ipproto);
exit(EXIT_FAILURE);
}
#undef BOTH_SET
#undef BOTH_SET_OR_BOTH_UNSET
#undef THREE_SET_OR_THREE_UNSET
#undef THREE_SET
static void ValidateParser(AppProto alproto)
{
ValidateParserProto(alproto, IPPROTO_TCP);
ValidateParserProto(alproto, IPPROTO_UDP);
}
static void ValidateParsers(void)
{
AppProto p = 0;
for ( ; p < ALPROTO_MAX; p++) {
ValidateParser(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
void AppLayerParserRegisterProtocolParsers(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
RegisterHTPParsers();
RegisterSSLParsers();
rs_dcerpc_register_parser();
rs_dcerpc_udp_register_parser();
RegisterSMBParsers();
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
RegisterFTPParsers();
RegisterSSHParsers();
RegisterSMTPParsers();
rs_dns_udp_register_parser();
rs_dns_tcp_register_parser();
RegisterModbusParsers();
RegisterENIPUDPParsers();
RegisterENIPTCPParsers();
RegisterDNP3Parsers();
RegisterNFSTCPParsers();
RegisterNFSUDPParsers();
RegisterNTPParsers();
RegisterTFTPParsers();
RegisterIKEParsers();
RegisterKRB5Parsers();
rs_dhcp_register_parser();
RegisterSNMPParsers();
RegisterSIPParsers();
RegisterTemplateRustParsers();
RegisterRFBParsers();
RegisterMQTTParsers();
RegisterTemplateParsers();
RegisterRdpParsers();
RegisterHTTP2Parsers();
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
/** IMAP */
AppLayerProtoDetectRegisterProtocol(ALPROTO_IMAP, "imap");
if (AppLayerProtoDetectConfProtoDetectionEnabled("tcp", "imap")) {
if (AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_IMAP,
"1|20|capability", 12, 0, STREAM_TOSERVER) < 0)
{
SCLogInfo("imap proto registration 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
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
} else {
SCLogInfo("Protocol detection and parser disabled for %s protocol.",
"imap");
}
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
ValidateParsers();
return;
}
/* coccinelle: AppLayerParserStateSetFlag():2,2:APP_LAYER_PARSER_ */
void AppLayerParserStateSetFlag(AppLayerParserState *pstate, uint8_t flag)
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();
pstate->flags |= flag;
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;
}
/* coccinelle: AppLayerParserStateIssetFlag():2,2:APP_LAYER_PARSER_ */
int AppLayerParserStateIssetFlag(AppLayerParserState *pstate, uint8_t flag)
{
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();
SCReturnInt(pstate->flags & flag);
}
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 AppLayerParserStreamTruncated(uint8_t ipproto, AppProto alproto, void *alstate,
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 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
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 (alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].Truncate != NULL)
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].Truncate(alstate, direction);
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
#ifdef DEBUG
void AppLayerParserStatePrintDetails(AppLayerParserState *pstate)
{
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 (pstate == NULL)
SCReturn;
AppLayerParserState *p = pstate;
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
SCLogDebug("AppLayerParser parser state information for parser state p(%p). "
"p->inspect_id[0](%"PRIu64"), "
"p->inspect_id[1](%"PRIu64"), "
"p->log_id(%"PRIu64"), "
"p->decoder_events(%p).",
pstate, p->inspect_id[0], p->inspect_id[1], p->log_id,
p->decoder_events);
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
#endif
/***** Unittests *****/
#ifdef UNITTESTS
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 AppLayerParserCtx alp_ctx_backup_unittest;
typedef struct TestState_ {
uint8_t test;
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
} TestState;
/**
* \brief Test parser function to test the memory deallocation of app layer
* parser of occurence of an error.
*/
static AppLayerResult TestProtocolParser(Flow *f, void *test_state, AppLayerParserState *pstate,
const uint8_t *input, uint32_t input_len,
void *local_data, const uint8_t flags)
{
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();
SCReturnStruct(APP_LAYER_ERROR);
}
/** \brief Function to allocates the Test protocol state memory
*/
static void *TestProtocolStateAlloc(void *orig_state, AppProto proto_orig)
{
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();
void *s = SCMalloc(sizeof(TestState));
if (unlikely(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
goto end;
memset(s, 0, sizeof(TestState));
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(s, "TestState");
}
/** \brief Function to free the Test Protocol state memory
*/
static void TestProtocolStateFree(void *s)
{
SCFree(s);
}
static uint64_t TestGetTxCnt(void *state)
{
/* single tx */
return 1;
}
static void TestStateTransactionFree(void *state, uint64_t tx_id)
{
/* do nothing */
}
static void *TestGetTx(void *state, uint64_t tx_id)
{
TestState *test_state = (TestState *)state;
return test_state;
}
void AppLayerParserRegisterProtocolUnittests(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
void (*RegisterUnittests)(void))
{
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();
alp_ctx.ctxs[FlowGetProtoMapping(ipproto)][alproto].
RegisterUnittests = RegisterUnittests;
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
void AppLayerParserBackupParserTable(void)
{
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();
alp_ctx_backup_unittest = alp_ctx;
memset(&alp_ctx, 0, sizeof(alp_ctx));
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
void AppLayerParserRestoreParserTable(void)
{
SCEnter();
alp_ctx = alp_ctx_backup_unittest;
memset(&alp_ctx_backup_unittest, 0, sizeof(alp_ctx_backup_unittest));
SCReturn;
}
/**
* \test Test the deallocation of app layer parser memory on occurance of
* error in the parsing process.
*/
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 AppLayerParserTest01(void)
{
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
AppLayerParserBackupParserTable();
Flow *f = NULL;
uint8_t testbuf[] = { 0x11 };
uint32_t testlen = sizeof(testbuf);
TcpSession ssn;
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
memset(&ssn, 0, sizeof(ssn));
/* Register the Test protocol state and parser functions */
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
AppLayerParserRegisterParser(IPPROTO_TCP, ALPROTO_TEST, STREAM_TOSERVER,
TestProtocolParser);
AppLayerParserRegisterStateFuncs(IPPROTO_TCP, ALPROTO_TEST,
TestProtocolStateAlloc, TestProtocolStateFree);
AppLayerParserRegisterTxFreeFunc(IPPROTO_TCP, ALPROTO_TEST, TestStateTransactionFree);
AppLayerParserRegisterGetTx(IPPROTO_TCP, ALPROTO_TEST, TestGetTx);
AppLayerParserRegisterGetTxCnt(IPPROTO_TCP, ALPROTO_TEST, TestGetTxCnt);
f = UTHBuildFlow(AF_INET, "1.2.3.4", "4.3.2.1", 20, 40);
FAIL_IF_NULL(f);
f->protoctx = &ssn;
f->alproto = ALPROTO_TEST;
f->proto = IPPROTO_TCP;
StreamTcpInitConfig(true);
int r = AppLayerParserParse(NULL, alp_tctx, f, ALPROTO_TEST,
STREAM_TOSERVER | STREAM_EOF, testbuf,
testlen);
FAIL_IF(r != -1);
FAIL_IF(!(ssn.flags & STREAMTCP_FLAG_APP_LAYER_DISABLED));
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
AppLayerParserRestoreParserTable();
StreamTcpFreeConfig(true);
UTHFreeFlow(f);
PASS;
}
/**
* \test Test the deallocation of app layer parser memory on occurance of
* error in the parsing process for UDP.
*/
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 AppLayerParserTest02(void)
{
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
AppLayerParserBackupParserTable();
Flow *f = NULL;
uint8_t testbuf[] = { 0x11 };
uint32_t testlen = sizeof(testbuf);
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
/* Register the Test protocol state and parser functions */
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
AppLayerParserRegisterParser(IPPROTO_UDP, ALPROTO_TEST, STREAM_TOSERVER,
TestProtocolParser);
AppLayerParserRegisterStateFuncs(IPPROTO_UDP, ALPROTO_TEST,
TestProtocolStateAlloc, TestProtocolStateFree);
AppLayerParserRegisterTxFreeFunc(IPPROTO_UDP, ALPROTO_TEST, TestStateTransactionFree);
AppLayerParserRegisterGetTx(IPPROTO_UDP, ALPROTO_TEST, TestGetTx);
AppLayerParserRegisterGetTxCnt(IPPROTO_UDP, ALPROTO_TEST, TestGetTxCnt);
f = UTHBuildFlow(AF_INET, "1.2.3.4", "4.3.2.1", 20, 40);
FAIL_IF_NULL(f);
f->alproto = ALPROTO_TEST;
f->proto = IPPROTO_UDP;
f->protomap = FlowGetProtoMapping(f->proto);
StreamTcpInitConfig(true);
int r = AppLayerParserParse(NULL, alp_tctx, f, ALPROTO_TEST,
STREAM_TOSERVER | STREAM_EOF, testbuf,
testlen);
FAIL_IF(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
AppLayerParserRestoreParserTable();
StreamTcpFreeConfig(true);
UTHFreeFlow(f);
PASS;
}
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 AppLayerParserRegisterUnittests(void)
{
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
int ip;
AppProto alproto;
AppLayerParserProtoCtx *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
for (ip = 0; ip < FLOW_PROTO_DEFAULT; ip++) {
for (alproto = 0; alproto < ALPROTO_MAX; alproto++) {
ctx = &alp_ctx.ctxs[ip][alproto];
if (ctx->RegisterUnittests == NULL)
continue;
ctx->RegisterUnittests();
}
}
UtRegisterTest("AppLayerParserTest01", AppLayerParserTest01);
UtRegisterTest("AppLayerParserTest02", AppLayerParserTest02);
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
#endif