App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/* Copyright (C) 2007-2014 Open Information Security Foundation
|
|
|
|
|
*
|
|
|
|
|
* You can copy, redistribute or modify this Program under the terms of
|
|
|
|
|
* the GNU General Public License version 2 as published by the Free
|
|
|
|
|
* Software Foundation.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
|
* version 2 along with this program; if not, write to the Free Software
|
|
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
|
|
|
|
|
* 02110-1301, USA.
|
|
|
|
|
*/
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/**
|
|
|
|
|
* \file
|
|
|
|
|
*
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
* \author Victor Julien <victor@inliniac.net>
|
|
|
|
|
* \author Anoop Saldanha <anoopsaldanha@gmail.com>
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include "suricata-common.h"
|
|
|
|
|
#include "debug.h"
|
|
|
|
|
#include "decode.h"
|
|
|
|
|
#include "threads.h"
|
|
|
|
|
#include "threadvars.h"
|
|
|
|
|
#include "tm-threads.h"
|
|
|
|
|
|
|
|
|
|
#include "detect.h"
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#include "detect-engine-port.h"
|
|
|
|
|
#include "detect-parse.h"
|
|
|
|
|
#include "detect-engine.h"
|
|
|
|
|
#include "detect-content.h"
|
|
|
|
|
#include "detect-engine-mpm.h"
|
|
|
|
|
#include "detect-engine-state.h"
|
|
|
|
|
|
|
|
|
|
#include "util-print.h"
|
|
|
|
|
#include "util-pool.h"
|
|
|
|
|
#include "util-unittest.h"
|
|
|
|
|
#include "util-unittest-helper.h"
|
|
|
|
|
|
|
|
|
|
#include "flow.h"
|
|
|
|
|
#include "flow-util.h"
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#include "flow-private.h"
|
|
|
|
|
|
|
|
|
|
#include "stream-tcp-private.h"
|
|
|
|
|
#include "stream-tcp-reassemble.h"
|
|
|
|
|
#include "stream-tcp.h"
|
|
|
|
|
#include "stream.h"
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#include "app-layer.h"
|
|
|
|
|
#include "app-layer-protos.h"
|
|
|
|
|
#include "app-layer-parser.h"
|
|
|
|
|
#include "app-layer-detect-proto.h"
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#include "conf.h"
|
|
|
|
|
#include "util-memcmp.h"
|
|
|
|
|
#include "util-spm.h"
|
|
|
|
|
#include "util-cuda.h"
|
|
|
|
|
#include "util-debug.h"
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#include "runmodes.h"
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectProbingParserElement_ {
|
|
|
|
|
AppProto alproto;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/* \todo don't really need it. See if you can get rid of it */
|
|
|
|
|
uint16_t port;
|
|
|
|
|
/* \todo calculate at runtime and get rid of this var */
|
|
|
|
|
uint32_t alproto_mask;
|
|
|
|
|
/* \todo check if we can reduce the bottom 2 vars to uint16_t */
|
|
|
|
|
/* the min length of data that has to be supplied to invoke the parser */
|
|
|
|
|
uint32_t min_depth;
|
|
|
|
|
/* the max length of data after which this parser won't be invoked */
|
|
|
|
|
uint32_t max_depth;
|
|
|
|
|
|
|
|
|
|
/* the to_server probing parser function */
|
|
|
|
|
ProbingParserFPtr ProbingParserTs;
|
|
|
|
|
|
|
|
|
|
/* the to_client probing parser function */
|
|
|
|
|
ProbingParserFPtr ProbingParserTc;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
struct AppLayerProtoDetectProbingParserElement_ *next;
|
|
|
|
|
} AppLayerProtoDetectProbingParserElement;
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectProbingParserPort_ {
|
|
|
|
|
/* the port no for which probing parser(s) are invoked */
|
|
|
|
|
uint16_t port;
|
|
|
|
|
|
|
|
|
|
uint32_t alproto_mask;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/* the max depth for all the probing parsers registered for this port */
|
|
|
|
|
uint16_t dp_max_depth;
|
|
|
|
|
uint16_t sp_max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *dp;
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *sp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
struct AppLayerProtoDetectProbingParserPort_ *next;
|
|
|
|
|
} AppLayerProtoDetectProbingParserPort;
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectProbingParser_ {
|
|
|
|
|
uint8_t ipproto;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectProbingParserPort *port;
|
|
|
|
|
|
|
|
|
|
struct AppLayerProtoDetectProbingParser_ *next;
|
|
|
|
|
} AppLayerProtoDetectProbingParser;
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectPMSignature_ {
|
|
|
|
|
AppProto alproto;
|
|
|
|
|
SigIntId 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
|
|
|
/* \todo Change this into a non-pointer */
|
|
|
|
|
DetectContentData *cd;
|
|
|
|
|
struct AppLayerProtoDetectPMSignature_ *next;
|
|
|
|
|
} AppLayerProtoDetectPMSignature;
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectPMCtx_ {
|
|
|
|
|
uint16_t max_len;
|
|
|
|
|
uint16_t min_len;
|
|
|
|
|
MpmCtx mpm_ctx;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/** Mapping between pattern id and signature. As each signature has a
|
|
|
|
|
* unique pattern with a unique id, we can lookup the signature by
|
|
|
|
|
* the pattern id. */
|
|
|
|
|
AppLayerProtoDetectPMSignature **map;
|
|
|
|
|
AppLayerProtoDetectPMSignature *head;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/* \todo we don't need this except at setup time. Get rid of it. */
|
|
|
|
|
PatIntId max_pat_id;
|
|
|
|
|
SigIntId max_sig_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
|
|
|
} AppLayerProtoDetectPMCtx;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
typedef struct AppLayerProtoDetectCtxIpproto_ {
|
|
|
|
|
/* 0 - toserver, 1 - toclient */
|
|
|
|
|
AppLayerProtoDetectPMCtx ctx_pm[2];
|
|
|
|
|
} AppLayerProtoDetectCtxIpproto;
|
|
|
|
|
|
|
|
|
|
/**
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
* \brief The app layer protocol detection context.
|
|
|
|
|
*/
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
typedef struct AppLayerProtoDetectCtx_ {
|
|
|
|
|
/* Context per ip_proto.
|
|
|
|
|
* \todo Modify ctx_ipp to hold for only tcp and udp. The rest can be
|
|
|
|
|
* implemented if needed. Waste of space otherwise. */
|
|
|
|
|
AppLayerProtoDetectCtxIpproto ctx_ipp[FLOW_PROTO_DEFAULT];
|
|
|
|
|
|
|
|
|
|
/* Global SPM thread context prototype. */
|
|
|
|
|
SpmGlobalThreadCtx *spm_global_thread_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
|
|
|
AppLayerProtoDetectProbingParser *ctx_pp;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/* Indicates the protocols that have registered themselves
|
|
|
|
|
* for protocol detection. This table is independent of the
|
|
|
|
|
* ipproto. */
|
|
|
|
|
const char *alproto_names[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
|
|
|
} AppLayerProtoDetectCtx;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/**
|
|
|
|
|
* \brief The app layer protocol detection thread context.
|
|
|
|
|
*/
|
|
|
|
|
struct AppLayerProtoDetectThreadCtx_ {
|
|
|
|
|
PrefilterRuleStore pmq;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/* The value 2 is for direction(0 - toserver, 1 - toclient). */
|
|
|
|
|
MpmThreadCtx mpm_tctx[FLOW_PROTO_DEFAULT][2];
|
|
|
|
|
SpmThreadCtx *spm_thread_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
|
|
|
/* The global app layer proto detection context. */
|
|
|
|
|
static AppLayerProtoDetectCtx alpd_ctx;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/***** Static Internal Calls: Protocol Retrieval *****/
|
|
|
|
|
|
|
|
|
|
/** \internal
|
|
|
|
|
* \brief Handle SPM search for Signature */
|
|
|
|
|
static AppProto AppLayerProtoDetectPMMatchSignature(const AppLayerProtoDetectPMSignature *s,
|
|
|
|
|
AppLayerProtoDetectThreadCtx *tctx,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint8_t *buf, uint16_t buflen,
|
|
|
|
|
uint8_t ipproto)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
AppProto proto = ALPROTO_UNKNOWN;
|
|
|
|
|
uint8_t *found = NULL;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (s->cd->offset > buflen) {
|
|
|
|
|
SCLogDebug("s->co->offset (%"PRIu16") > buflen (%"PRIu16")",
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
s->cd->offset, buflen);
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (s->cd->depth > buflen) {
|
|
|
|
|
SCLogDebug("s->co->depth (%"PRIu16") > buflen (%"PRIu16")",
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
s->cd->depth, buflen);
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint8_t *sbuf = buf + s->cd->offset;
|
|
|
|
|
uint16_t sbuflen = s->cd->depth - s->cd->offset;
|
|
|
|
|
SCLogDebug("s->co->offset (%"PRIu16") s->cd->depth (%"PRIu16")",
|
|
|
|
|
s->cd->offset, s->cd->depth);
|
|
|
|
|
|
|
|
|
|
found = SpmScan(s->cd->spm_ctx, tctx->spm_thread_ctx, sbuf, sbuflen);
|
|
|
|
|
if (found != NULL)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
proto = s->alproto;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
SCReturnUInt(proto);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** \internal
|
|
|
|
|
* \brief Run Pattern Sigs against buffer
|
|
|
|
|
* \param pm_results[out] AppProto array of size ALPROTO_MAX */
|
|
|
|
|
static AppProto AppLayerProtoDetectPMGetProto(AppLayerProtoDetectThreadCtx *tctx,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
Flow *f,
|
|
|
|
|
uint8_t *buf, uint16_t buflen,
|
|
|
|
|
uint8_t direction,
|
|
|
|
|
uint8_t ipproto,
|
|
|
|
|
AppProto *pm_results)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
pm_results[0] = ALPROTO_UNKNOWN;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPMCtx *pm_ctx;
|
|
|
|
|
MpmThreadCtx *mpm_tctx;
|
|
|
|
|
uint16_t pm_matches = 0;
|
|
|
|
|
uint8_t cnt;
|
|
|
|
|
uint16_t searchlen;
|
|
|
|
|
|
|
|
|
|
if (f->protomap >= FLOW_PROTO_DEFAULT)
|
|
|
|
|
return ALPROTO_UNKNOWN;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (direction & STREAM_TOSERVER) {
|
|
|
|
|
pm_ctx = &alpd_ctx.ctx_ipp[f->protomap].ctx_pm[0];
|
|
|
|
|
mpm_tctx = &tctx->mpm_tctx[f->protomap][0];
|
|
|
|
|
} else {
|
|
|
|
|
pm_ctx = &alpd_ctx.ctx_ipp[f->protomap].ctx_pm[1];
|
|
|
|
|
mpm_tctx = &tctx->mpm_tctx[f->protomap][1];
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (pm_ctx->mpm_ctx.pattern_cnt == 0)
|
|
|
|
|
goto end;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
searchlen = buflen;
|
|
|
|
|
if (searchlen > pm_ctx->max_len)
|
|
|
|
|
searchlen = pm_ctx->max_len;
|
|
|
|
|
|
|
|
|
|
uint32_t search_cnt = 0;
|
|
|
|
|
|
|
|
|
|
/* do the mpm search */
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
search_cnt = mpm_table[pm_ctx->mpm_ctx.mpm_type].Search(&pm_ctx->mpm_ctx,
|
|
|
|
|
mpm_tctx,
|
|
|
|
|
&tctx->pmq,
|
|
|
|
|
buf, searchlen);
|
|
|
|
|
if (search_cnt == 0)
|
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
/* alproto bit field */
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint8_t pm_results_bf[(ALPROTO_MAX / 8) + 1];
|
|
|
|
|
memset(pm_results_bf, 0, sizeof(pm_results_bf));
|
|
|
|
|
|
|
|
|
|
/* loop through unique pattern id's. Can't use search_cnt here,
|
|
|
|
|
* as that contains all matches, tctx->pmq.pattern_id_array_cnt
|
|
|
|
|
* contains only *unique* matches. */
|
|
|
|
|
for (cnt = 0; cnt < tctx->pmq.rule_id_array_cnt; cnt++) {
|
|
|
|
|
const AppLayerProtoDetectPMSignature *s = pm_ctx->map[tctx->pmq.rule_id_array[cnt]];
|
|
|
|
|
while (s != NULL) {
|
|
|
|
|
AppProto proto = AppLayerProtoDetectPMMatchSignature(s,
|
|
|
|
|
tctx, buf, searchlen, ipproto);
|
|
|
|
|
|
|
|
|
|
/* store each unique proto once */
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (proto != ALPROTO_UNKNOWN &&
|
|
|
|
|
!(pm_results_bf[proto / 8] & (1 << (proto % 8))) )
|
|
|
|
|
{
|
|
|
|
|
pm_results[pm_matches++] = proto;
|
|
|
|
|
pm_results_bf[proto / 8] |= 1 << (proto % 8);
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
s = s->next;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
PmqReset(&tctx->pmq);
|
|
|
|
|
if (buflen >= pm_ctx->max_len)
|
|
|
|
|
FLOW_SET_PM_DONE(f, direction);
|
|
|
|
|
SCReturnUInt(pm_matches);
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static AppLayerProtoDetectProbingParserPort *AppLayerProtoDetectGetProbingParsers(AppLayerProtoDetectProbingParser *pp,
|
|
|
|
|
uint8_t ipproto,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t port)
|
|
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *pp_port = NULL;
|
|
|
|
|
|
|
|
|
|
while (pp != NULL) {
|
|
|
|
|
if (pp->ipproto == ipproto)
|
|
|
|
|
break;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
pp = pp->next;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (pp == NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
pp_port = pp->port;
|
|
|
|
|
while (pp_port != NULL) {
|
|
|
|
|
if (pp_port->port == port || pp_port->port == 0) {
|
|
|
|
|
break;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
pp_port = pp_port->next;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
SCReturnPtr(pp_port, "AppLayerProtoDetectProbingParserPort *");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* \brief Call the probing parser if it exists for this flow.
|
|
|
|
|
*
|
|
|
|
|
* First we check the flow's dp as it's most likely to match. If that didn't
|
|
|
|
|
* lead to a PP, we try the sp.
|
|
|
|
|
*
|
|
|
|
|
*/
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static AppProto AppLayerProtoDetectPPGetProto(Flow *f,
|
|
|
|
|
uint8_t *buf, uint32_t buflen,
|
|
|
|
|
uint8_t ipproto, uint8_t direction)
|
|
|
|
|
{
|
|
|
|
|
const AppLayerProtoDetectProbingParserPort *pp_port_dp = NULL;
|
|
|
|
|
const AppLayerProtoDetectProbingParserPort *pp_port_sp = NULL;
|
|
|
|
|
const AppLayerProtoDetectProbingParserElement *pe = NULL;
|
|
|
|
|
const AppLayerProtoDetectProbingParserElement *pe1 = NULL;
|
|
|
|
|
const AppLayerProtoDetectProbingParserElement *pe2 = 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
|
|
|
AppProto alproto = ALPROTO_UNKNOWN;
|
|
|
|
|
uint32_t *alproto_masks;
|
|
|
|
|
uint32_t mask = 0;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
const uint16_t dp = f->protodetect_dp ? f->protodetect_dp : f->dp;
|
|
|
|
|
const uint16_t sp = f->sp;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (direction & STREAM_TOSERVER) {
|
|
|
|
|
/* first try the destination port */
|
|
|
|
|
pp_port_dp = AppLayerProtoDetectGetProbingParsers(alpd_ctx.ctx_pp, ipproto, dp);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
alproto_masks = &f->probing_parser_toserver_alproto_masks;
|
|
|
|
|
if (pp_port_dp != NULL) {
|
|
|
|
|
SCLogDebug("toserver - Probing parser found for destination port %"PRIu16, dp);
|
|
|
|
|
|
|
|
|
|
/* found based on destination port, so use dp registration */
|
|
|
|
|
pe1 = pp_port_dp->dp;
|
|
|
|
|
} else {
|
|
|
|
|
SCLogDebug("toserver - No probing parser registered for dest port %"PRIu16, dp);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
pp_port_sp = AppLayerProtoDetectGetProbingParsers(alpd_ctx.ctx_pp, ipproto, sp);
|
|
|
|
|
if (pp_port_sp != NULL) {
|
|
|
|
|
SCLogDebug("toserver - Probing parser found for source port %"PRIu16, sp);
|
|
|
|
|
|
|
|
|
|
/* found based on source port, so use sp registration */
|
|
|
|
|
pe2 = pp_port_sp->sp;
|
|
|
|
|
} else {
|
|
|
|
|
SCLogDebug("toserver - No probing parser registered for source port %"PRIu16, sp);
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
/* first try the destination port */
|
|
|
|
|
pp_port_dp = AppLayerProtoDetectGetProbingParsers(alpd_ctx.ctx_pp, ipproto, dp);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
alproto_masks = &f->probing_parser_toclient_alproto_masks;
|
|
|
|
|
if (pp_port_dp != NULL) {
|
|
|
|
|
SCLogDebug("toclient - Probing parser found for destination port %"PRIu16, dp);
|
|
|
|
|
|
|
|
|
|
/* found based on destination port, so use dp registration */
|
|
|
|
|
pe1 = pp_port_dp->dp;
|
|
|
|
|
} else {
|
|
|
|
|
SCLogDebug("toclient - No probing parser registered for dest port %"PRIu16, dp);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
pp_port_sp = AppLayerProtoDetectGetProbingParsers(alpd_ctx.ctx_pp, ipproto, sp);
|
|
|
|
|
if (pp_port_sp != NULL) {
|
|
|
|
|
SCLogDebug("toclient - Probing parser found for source port %"PRIu16, sp);
|
|
|
|
|
|
|
|
|
|
pe2 = pp_port_sp->sp;
|
|
|
|
|
} else {
|
|
|
|
|
SCLogDebug("toclient - No probing parser registered for source port %"PRIu16, sp);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (pe1 == NULL && pe2 == NULL) {
|
|
|
|
|
SCLogDebug("%s - No probing parsers found for either port",
|
|
|
|
|
(direction & STREAM_TOSERVER) ? "toserver":"toclient");
|
|
|
|
|
FLOW_SET_PP_DONE(f, direction);
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* run the parser(s) */
|
|
|
|
|
pe = pe1;
|
|
|
|
|
while (pe != NULL) {
|
|
|
|
|
if ((buflen < pe->min_depth) ||
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
(alproto_masks[0] & pe->alproto_mask)) {
|
|
|
|
|
pe = pe->next;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (direction & STREAM_TOSERVER && pe->ProbingParserTs != NULL) {
|
|
|
|
|
alproto = pe->ProbingParserTs(buf, buflen, NULL);
|
|
|
|
|
} else if (pe->ProbingParserTc != NULL) {
|
|
|
|
|
alproto = pe->ProbingParserTc(buf, buflen, NULL);
|
|
|
|
|
}
|
|
|
|
|
if (alproto != ALPROTO_UNKNOWN && alproto != ALPROTO_FAILED)
|
|
|
|
|
goto end;
|
|
|
|
|
if (alproto == ALPROTO_FAILED ||
|
|
|
|
|
(pe->max_depth != 0 && buflen > pe->max_depth)) {
|
|
|
|
|
alproto_masks[0] |= pe->alproto_mask;
|
|
|
|
|
}
|
|
|
|
|
pe = pe->next;
|
|
|
|
|
}
|
|
|
|
|
pe = pe2;
|
|
|
|
|
while (pe != NULL) {
|
|
|
|
|
if ((buflen < pe->min_depth) ||
|
|
|
|
|
(alproto_masks[0] & pe->alproto_mask)) {
|
|
|
|
|
pe = pe->next;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (direction & STREAM_TOSERVER && pe->ProbingParserTs != NULL) {
|
|
|
|
|
alproto = pe->ProbingParserTs(buf, buflen, NULL);
|
|
|
|
|
} else if (pe->ProbingParserTc != NULL) {
|
|
|
|
|
alproto = pe->ProbingParserTc(buf, buflen, NULL);
|
|
|
|
|
}
|
|
|
|
|
if (alproto != ALPROTO_UNKNOWN && alproto != ALPROTO_FAILED)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto end;
|
|
|
|
|
if (alproto == ALPROTO_FAILED ||
|
|
|
|
|
(pe->max_depth != 0 && buflen > pe->max_depth)) {
|
|
|
|
|
alproto_masks[0] |= pe->alproto_mask;
|
|
|
|
|
}
|
|
|
|
|
pe = pe->next;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* get the mask we need for this direction */
|
|
|
|
|
if (pp_port_dp && pp_port_sp)
|
|
|
|
|
mask = pp_port_dp->alproto_mask|pp_port_sp->alproto_mask;
|
|
|
|
|
else if (pp_port_dp)
|
|
|
|
|
mask = pp_port_dp->alproto_mask;
|
|
|
|
|
else if (pp_port_sp)
|
|
|
|
|
mask = pp_port_sp->alproto_mask;
|
|
|
|
|
|
|
|
|
|
if (alproto_masks[0] == mask) {
|
|
|
|
|
FLOW_SET_PP_DONE(f, direction);
|
|
|
|
|
SCLogDebug("%s, mask is now %08x, needed %08x, so done",
|
|
|
|
|
(direction & STREAM_TOSERVER) ? "toserver":"toclient", alproto_masks[0], mask);
|
|
|
|
|
} else {
|
|
|
|
|
SCLogDebug("%s, mask is now %08x, need %08x",
|
|
|
|
|
(direction & STREAM_TOSERVER) ? "toserver":"toclient", alproto_masks[0], mask);
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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:
|
|
|
|
|
SCLogDebug("%s, mask is now %08x",
|
|
|
|
|
(direction & STREAM_TOSERVER) ? "toserver":"toclient", alproto_masks[0]);
|
|
|
|
|
SCReturnUInt(alproto);
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/***** Static Internal Calls: PP registration *****/
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectPPGetIpprotos(AppProto alproto,
|
|
|
|
|
uint8_t *ipprotos)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
const AppLayerProtoDetectProbingParser *pp;
|
|
|
|
|
const AppLayerProtoDetectProbingParserPort *pp_port;
|
|
|
|
|
const AppLayerProtoDetectProbingParserElement *pp_pe;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
for (pp = alpd_ctx.ctx_pp; pp != NULL; pp = pp->next) {
|
|
|
|
|
for (pp_port = pp->port; pp_port != NULL; pp_port = pp_port->next) {
|
|
|
|
|
for (pp_pe = pp_port->dp; pp_pe != NULL; pp_pe = pp_pe->next) {
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alproto == pp_pe->alproto)
|
|
|
|
|
ipprotos[pp->ipproto / 8] |= 1 << (pp->ipproto % 8);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
for (pp_pe = pp_port->sp; pp_pe != NULL; pp_pe = pp_pe->next) {
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alproto == pp_pe->alproto)
|
|
|
|
|
ipprotos[pp->ipproto / 8] |= 1 << (pp->ipproto % 8);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static uint32_t AppLayerProtoDetectProbingParserGetMask(AppProto alproto)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (!(alproto > ALPROTO_UNKNOWN && alproto < ALPROTO_FAILED)) {
|
|
|
|
|
SCLogError(SC_ERR_ALPARSER, "Unknown protocol detected - %"PRIu16,
|
|
|
|
|
alproto);
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturnUInt(1 << alproto);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static AppLayerProtoDetectProbingParserElement *AppLayerProtoDetectProbingParserElementAlloc(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
|
|
|
AppLayerProtoDetectProbingParserElement *p = SCMalloc(sizeof(AppLayerProtoDetectProbingParserElement));
|
|
|
|
|
if (unlikely(p == NULL)) {
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(p, 0, sizeof(AppLayerProtoDetectProbingParserElement));
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturnPtr(p, "AppLayerProtoDetectProbingParserElement");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectProbingParserElementFree(AppLayerProtoDetectProbingParserElement *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
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
SCFree(p);
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static AppLayerProtoDetectProbingParserPort *AppLayerProtoDetectProbingParserPortAlloc(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
|
|
|
AppLayerProtoDetectProbingParserPort *p = SCMalloc(sizeof(AppLayerProtoDetectProbingParserPort));
|
|
|
|
|
if (unlikely(p == NULL)) {
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(p, 0, sizeof(AppLayerProtoDetectProbingParserPort));
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturnPtr(p, "AppLayerProtoDetectProbingParserPort");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectProbingParserPortFree(AppLayerProtoDetectProbingParserPort *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
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectProbingParserElement *e;
|
|
|
|
|
|
|
|
|
|
e = p->dp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
while (e != NULL) {
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *e_next = e->next;
|
|
|
|
|
AppLayerProtoDetectProbingParserElementFree(e);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
e = e_next;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
e = p->sp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
while (e != NULL) {
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *e_next = e->next;
|
|
|
|
|
AppLayerProtoDetectProbingParserElementFree(e);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
e = e_next;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCFree(p);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static AppLayerProtoDetectProbingParser *AppLayerProtoDetectProbingParserAlloc(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
|
|
|
AppLayerProtoDetectProbingParser *p = SCMalloc(sizeof(AppLayerProtoDetectProbingParser));
|
|
|
|
|
if (unlikely(p == NULL)) {
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
|
|
|
|
memset(p, 0, sizeof(AppLayerProtoDetectProbingParser));
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturnPtr(p, "AppLayerProtoDetectProbingParser");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectProbingParserFree(AppLayerProtoDetectProbingParser *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
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectProbingParserPort *pt = p->port;
|
|
|
|
|
while (pt != NULL) {
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *pt_next = pt->next;
|
|
|
|
|
AppLayerProtoDetectProbingParserPortFree(pt);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
pt = pt_next;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCFree(p);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static AppLayerProtoDetectProbingParserElement *
|
|
|
|
|
AppLayerProtoDetectProbingParserElementCreate(AppProto alproto,
|
|
|
|
|
uint16_t port,
|
|
|
|
|
uint16_t min_depth,
|
|
|
|
|
uint16_t max_depth)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *pe = AppLayerProtoDetectProbingParserElementAlloc();
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
pe->alproto = alproto;
|
|
|
|
|
pe->port = port;
|
|
|
|
|
pe->alproto_mask = AppLayerProtoDetectProbingParserGetMask(alproto);
|
|
|
|
|
pe->min_depth = min_depth;
|
|
|
|
|
pe->max_depth = max_depth;
|
|
|
|
|
pe->next = NULL;
|
|
|
|
|
|
|
|
|
|
if (max_depth != 0 && min_depth >= max_depth) {
|
|
|
|
|
SCLogError(SC_ERR_ALPARSER, "Invalid arguments sent to "
|
|
|
|
|
"register the probing parser. min_depth >= max_depth");
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
if (alproto <= ALPROTO_UNKNOWN || alproto >= ALPROTO_MAX) {
|
|
|
|
|
SCLogError(SC_ERR_ALPARSER, "Invalid arguments sent to register "
|
|
|
|
|
"the probing parser. Invalid alproto - %d", alproto);
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SCReturnPtr(pe, "AppLayerProtoDetectProbingParserElement");
|
|
|
|
|
error:
|
|
|
|
|
AppLayerProtoDetectProbingParserElementFree(pe);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturnPtr(NULL, "AppLayerProtoDetectProbingParserElement");
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static AppLayerProtoDetectProbingParserElement *
|
|
|
|
|
AppLayerProtoDetectProbingParserElementDuplicate(AppLayerProtoDetectProbingParserElement *pe)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *new_pe = AppLayerProtoDetectProbingParserElementAlloc();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
new_pe->alproto = pe->alproto;
|
|
|
|
|
new_pe->port = pe->port;
|
|
|
|
|
new_pe->alproto_mask = pe->alproto_mask;
|
|
|
|
|
new_pe->min_depth = pe->min_depth;
|
|
|
|
|
new_pe->max_depth = pe->max_depth;
|
|
|
|
|
new_pe->ProbingParserTs = pe->ProbingParserTs;
|
|
|
|
|
new_pe->ProbingParserTc = pe->ProbingParserTc;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
new_pe->next = NULL;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturnPtr(new_pe, "AppLayerProtoDetectProbingParserElement");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef DEBUG
|
|
|
|
|
static void AppLayerProtoDetectPrintProbingParsers(AppLayerProtoDetectProbingParser *pp)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectProbingParserPort *pp_port = NULL;
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *pp_pe = NULL;
|
|
|
|
|
|
|
|
|
|
printf("\nProtocol Detection Configuration\n");
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
for ( ; pp != NULL; pp = pp->next) {
|
|
|
|
|
/* print ip protocol */
|
|
|
|
|
if (pp->ipproto == IPPROTO_TCP)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
printf("IPProto: TCP\n");
|
|
|
|
|
else if (pp->ipproto == IPPROTO_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
|
|
|
printf("IPProto: UDP\n");
|
|
|
|
|
else
|
|
|
|
|
printf("IPProto: %"PRIu8"\n", pp->ipproto);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
pp_port = pp->port;
|
|
|
|
|
for ( ; pp_port != NULL; pp_port = pp_port->next) {
|
|
|
|
|
if (pp_port->dp != NULL) {
|
|
|
|
|
printf(" Port: %"PRIu16 "\n", pp_port->port);
|
|
|
|
|
|
|
|
|
|
printf(" Destination port: (max-depth: %"PRIu16 ", "
|
|
|
|
|
"mask - %"PRIu32")\n",
|
|
|
|
|
pp_port->dp_max_depth,
|
|
|
|
|
pp_port->alproto_mask);
|
|
|
|
|
pp_pe = pp_port->dp;
|
|
|
|
|
for ( ; pp_pe != NULL; pp_pe = pp_pe->next) {
|
|
|
|
|
|
|
|
|
|
if (pp_pe->alproto == ALPROTO_HTTP)
|
|
|
|
|
printf(" alproto: ALPROTO_HTTP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_FTP)
|
|
|
|
|
printf(" alproto: ALPROTO_FTP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SMTP)
|
|
|
|
|
printf(" alproto: ALPROTO_SMTP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_TLS)
|
|
|
|
|
printf(" alproto: ALPROTO_TLS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SSH)
|
|
|
|
|
printf(" alproto: ALPROTO_SSH\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_IMAP)
|
|
|
|
|
printf(" alproto: ALPROTO_IMAP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_MSN)
|
|
|
|
|
printf(" alproto: ALPROTO_MSN\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_JABBER)
|
|
|
|
|
printf(" alproto: ALPROTO_JABBER\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SMB)
|
|
|
|
|
printf(" alproto: ALPROTO_SMB\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SMB2)
|
|
|
|
|
printf(" alproto: ALPROTO_SMB2\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_DCERPC)
|
|
|
|
|
printf(" alproto: ALPROTO_DCERPC\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_IRC)
|
|
|
|
|
printf(" alproto: ALPROTO_IRC\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_DNS)
|
|
|
|
|
printf(" alproto: ALPROTO_DNS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_MODBUS)
|
|
|
|
|
printf(" alproto: ALPROTO_MODBUS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_ENIP)
|
|
|
|
|
printf(" alproto: ALPROTO_ENIP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_NFS)
|
|
|
|
|
printf(" alproto: ALPROTO_NFS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_TEMPLATE)
|
|
|
|
|
printf(" alproto: ALPROTO_TEMPLATE\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_DNP3)
|
|
|
|
|
printf(" alproto: ALPROTO_DNP3\n");
|
|
|
|
|
else
|
|
|
|
|
printf("impossible\n");
|
|
|
|
|
|
|
|
|
|
printf(" port: %"PRIu16 "\n", pp_pe->port);
|
|
|
|
|
printf(" mask: %"PRIu32 "\n", pp_pe->alproto_mask);
|
|
|
|
|
printf(" min_depth: %"PRIu32 "\n", pp_pe->min_depth);
|
|
|
|
|
printf(" max_depth: %"PRIu32 "\n", pp_pe->max_depth);
|
|
|
|
|
|
|
|
|
|
printf("\n");
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (pp_port->sp == 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
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
printf(" Source port: (max-depth: %"PRIu16 ", "
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
"mask - %"PRIu32")\n",
|
|
|
|
|
pp_port->sp_max_depth,
|
|
|
|
|
pp_port->alproto_mask);
|
|
|
|
|
pp_pe = pp_port->sp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
for ( ; pp_pe != NULL; pp_pe = pp_pe->next) {
|
|
|
|
|
|
|
|
|
|
if (pp_pe->alproto == ALPROTO_HTTP)
|
|
|
|
|
printf(" alproto: ALPROTO_HTTP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_FTP)
|
|
|
|
|
printf(" alproto: ALPROTO_FTP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SMTP)
|
|
|
|
|
printf(" alproto: ALPROTO_SMTP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_TLS)
|
|
|
|
|
printf(" alproto: ALPROTO_TLS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SSH)
|
|
|
|
|
printf(" alproto: ALPROTO_SSH\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_IMAP)
|
|
|
|
|
printf(" alproto: ALPROTO_IMAP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_MSN)
|
|
|
|
|
printf(" alproto: ALPROTO_MSN\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_JABBER)
|
|
|
|
|
printf(" alproto: ALPROTO_JABBER\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SMB)
|
|
|
|
|
printf(" alproto: ALPROTO_SMB\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_SMB2)
|
|
|
|
|
printf(" alproto: ALPROTO_SMB2\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_DCERPC)
|
|
|
|
|
printf(" alproto: ALPROTO_DCERPC\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_IRC)
|
|
|
|
|
printf(" alproto: ALPROTO_IRC\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_DNS)
|
|
|
|
|
printf(" alproto: ALPROTO_DNS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_MODBUS)
|
|
|
|
|
printf(" alproto: ALPROTO_MODBUS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_ENIP)
|
|
|
|
|
printf(" alproto: ALPROTO_ENIP\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_NFS)
|
|
|
|
|
printf(" alproto: ALPROTO_NFS\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_TEMPLATE)
|
|
|
|
|
printf(" alproto: ALPROTO_TEMPLATE\n");
|
|
|
|
|
else if (pp_pe->alproto == ALPROTO_DNP3)
|
|
|
|
|
printf(" alproto: ALPROTO_DNP3\n");
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
else
|
|
|
|
|
printf("impossible\n");
|
|
|
|
|
|
|
|
|
|
printf(" port: %"PRIu16 "\n", pp_pe->port);
|
|
|
|
|
printf(" mask: %"PRIu32 "\n", pp_pe->alproto_mask);
|
|
|
|
|
printf(" min_depth: %"PRIu32 "\n", pp_pe->min_depth);
|
|
|
|
|
printf(" max_depth: %"PRIu32 "\n", pp_pe->max_depth);
|
|
|
|
|
|
|
|
|
|
printf("\n");
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectProbingParserElementAppend(AppLayerProtoDetectProbingParserElement **head_pe,
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *new_pe)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (*head_pe == NULL) {
|
|
|
|
|
*head_pe = new_pe;
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if ((*head_pe)->port == 0) {
|
|
|
|
|
if (new_pe->port != 0) {
|
|
|
|
|
new_pe->next = *head_pe;
|
|
|
|
|
*head_pe = new_pe;
|
|
|
|
|
} else {
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *temp_pe = *head_pe;
|
|
|
|
|
while (temp_pe->next != NULL)
|
|
|
|
|
temp_pe = temp_pe->next;
|
|
|
|
|
temp_pe->next = new_pe;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *temp_pe = *head_pe;
|
|
|
|
|
if (new_pe->port == 0) {
|
|
|
|
|
while (temp_pe->next != NULL)
|
|
|
|
|
temp_pe = temp_pe->next;
|
|
|
|
|
temp_pe->next = new_pe;
|
|
|
|
|
} else {
|
|
|
|
|
while (temp_pe->next != NULL && temp_pe->next->port != 0)
|
|
|
|
|
temp_pe = temp_pe->next;
|
|
|
|
|
new_pe->next = temp_pe->next;
|
|
|
|
|
temp_pe->next = new_pe;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectProbingParserAppend(AppLayerProtoDetectProbingParser **head_pp,
|
|
|
|
|
AppLayerProtoDetectProbingParser *new_pp)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
if (*head_pp == NULL) {
|
|
|
|
|
*head_pp = new_pp;
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParser *temp_pp = *head_pp;
|
|
|
|
|
while (temp_pp->next != NULL)
|
|
|
|
|
temp_pp = temp_pp->next;
|
|
|
|
|
temp_pp->next = new_pp;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectProbingParserPortAppend(AppLayerProtoDetectProbingParserPort **head_port,
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *new_port)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (*head_port == NULL) {
|
|
|
|
|
*head_port = new_port;
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if ((*head_port)->port == 0) {
|
|
|
|
|
new_port->next = *head_port;
|
|
|
|
|
*head_port = new_port;
|
|
|
|
|
} else {
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *temp_port = *head_port;
|
|
|
|
|
while (temp_port->next != NULL && temp_port->next->port != 0) {
|
|
|
|
|
temp_port = temp_port->next;
|
|
|
|
|
}
|
|
|
|
|
new_port->next = temp_port->next;
|
|
|
|
|
temp_port->next = new_port;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectInsertNewProbingParser(AppLayerProtoDetectProbingParser **pp,
|
|
|
|
|
uint8_t ipproto,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t port,
|
|
|
|
|
AppProto alproto,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t min_depth, uint16_t max_depth,
|
|
|
|
|
uint8_t direction,
|
|
|
|
|
ProbingParserFPtr ProbingParser1,
|
|
|
|
|
ProbingParserFPtr ProbingParser2)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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();
|
|
|
|
|
|
|
|
|
|
/* get the top level ipproto pp */
|
|
|
|
|
AppLayerProtoDetectProbingParser *curr_pp = *pp;
|
|
|
|
|
while (curr_pp != NULL) {
|
|
|
|
|
if (curr_pp->ipproto == ipproto)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
break;
|
|
|
|
|
curr_pp = curr_pp->next;
|
|
|
|
|
}
|
|
|
|
|
if (curr_pp == NULL) {
|
|
|
|
|
AppLayerProtoDetectProbingParser *new_pp = AppLayerProtoDetectProbingParserAlloc();
|
|
|
|
|
new_pp->ipproto = ipproto;
|
|
|
|
|
AppLayerProtoDetectProbingParserAppend(pp, new_pp);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
curr_pp = new_pp;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/* get the top level port pp */
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *curr_port = curr_pp->port;
|
|
|
|
|
while (curr_port != NULL) {
|
|
|
|
|
if (curr_port->port == port)
|
|
|
|
|
break;
|
|
|
|
|
curr_port = curr_port->next;
|
|
|
|
|
}
|
|
|
|
|
if (curr_port == NULL) {
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *new_port = AppLayerProtoDetectProbingParserPortAlloc();
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
new_port->port = port;
|
|
|
|
|
AppLayerProtoDetectProbingParserPortAppend(&curr_pp->port, new_port);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
curr_port = new_port;
|
|
|
|
|
if (direction & STREAM_TOSERVER) {
|
|
|
|
|
curr_port->dp_max_depth = max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
} else {
|
|
|
|
|
curr_port->sp_max_depth = max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *zero_port;
|
|
|
|
|
|
|
|
|
|
zero_port = curr_pp->port;
|
|
|
|
|
while (zero_port != NULL && zero_port->port != 0) {
|
|
|
|
|
zero_port = zero_port->next;
|
|
|
|
|
}
|
|
|
|
|
if (zero_port != NULL) {
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *zero_pe;
|
|
|
|
|
|
|
|
|
|
zero_pe = zero_port->dp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
for ( ; zero_pe != NULL; zero_pe = zero_pe->next) {
|
|
|
|
|
if (curr_port->dp == NULL)
|
|
|
|
|
curr_port->dp_max_depth = zero_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (zero_pe->max_depth == 0)
|
|
|
|
|
curr_port->dp_max_depth = zero_pe->max_depth;
|
|
|
|
|
if (curr_port->dp_max_depth != 0 &&
|
|
|
|
|
curr_port->dp_max_depth < zero_pe->max_depth) {
|
|
|
|
|
curr_port->dp_max_depth = zero_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *dup_pe =
|
|
|
|
|
AppLayerProtoDetectProbingParserElementDuplicate(zero_pe);
|
|
|
|
|
AppLayerProtoDetectProbingParserElementAppend(&curr_port->dp, dup_pe);
|
|
|
|
|
curr_port->alproto_mask |= dup_pe->alproto_mask;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
zero_pe = zero_port->sp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
for ( ; zero_pe != NULL; zero_pe = zero_pe->next) {
|
|
|
|
|
if (curr_port->sp == NULL)
|
|
|
|
|
curr_port->sp_max_depth = zero_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (zero_pe->max_depth == 0)
|
|
|
|
|
curr_port->sp_max_depth = zero_pe->max_depth;
|
|
|
|
|
if (curr_port->sp_max_depth != 0 &&
|
|
|
|
|
curr_port->sp_max_depth < zero_pe->max_depth) {
|
|
|
|
|
curr_port->sp_max_depth = zero_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *dup_pe =
|
|
|
|
|
AppLayerProtoDetectProbingParserElementDuplicate(zero_pe);
|
|
|
|
|
AppLayerProtoDetectProbingParserElementAppend(&curr_port->sp, dup_pe);
|
|
|
|
|
curr_port->alproto_mask |= dup_pe->alproto_mask;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (zero_port != NULL) */
|
|
|
|
|
} /* if (curr_port == NULL) */
|
|
|
|
|
|
|
|
|
|
/* insert the pe_pp */
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *curr_pe;
|
|
|
|
|
if (direction & STREAM_TOSERVER)
|
|
|
|
|
curr_pe = curr_port->dp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
else
|
|
|
|
|
curr_pe = curr_port->sp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
while (curr_pe != NULL) {
|
|
|
|
|
if (curr_pe->alproto == alproto) {
|
|
|
|
|
SCLogError(SC_ERR_ALPARSER, "Duplicate pp registered - "
|
|
|
|
|
"ipproto - %"PRIu8" Port - %"PRIu16" "
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
"App Protocol - NULL, App Protocol(ID) - "
|
|
|
|
|
"%"PRIu16" min_depth - %"PRIu16" "
|
|
|
|
|
"max_dept - %"PRIu16".",
|
|
|
|
|
ipproto, port, 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
|
|
|
min_depth, max_depth);
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
curr_pe = curr_pe->next;
|
|
|
|
|
}
|
|
|
|
|
/* Get a new parser element */
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *new_pe =
|
|
|
|
|
AppLayerProtoDetectProbingParserElementCreate(alproto,
|
|
|
|
|
curr_port->port,
|
|
|
|
|
min_depth, max_depth);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (new_pe == NULL)
|
|
|
|
|
goto error;
|
|
|
|
|
curr_pe = new_pe;
|
|
|
|
|
AppLayerProtoDetectProbingParserElement **head_pe;
|
|
|
|
|
if (direction & STREAM_TOSERVER) {
|
|
|
|
|
curr_pe->ProbingParserTs = ProbingParser1;
|
|
|
|
|
curr_pe->ProbingParserTc = ProbingParser2;
|
|
|
|
|
if (curr_port->dp == NULL)
|
|
|
|
|
curr_port->dp_max_depth = new_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (new_pe->max_depth == 0)
|
|
|
|
|
curr_port->dp_max_depth = new_pe->max_depth;
|
|
|
|
|
if (curr_port->dp_max_depth != 0 &&
|
|
|
|
|
curr_port->dp_max_depth < new_pe->max_depth) {
|
|
|
|
|
curr_port->dp_max_depth = new_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
curr_port->alproto_mask |= new_pe->alproto_mask;
|
|
|
|
|
head_pe = &curr_port->dp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
} else {
|
|
|
|
|
curr_pe->ProbingParserTs = ProbingParser2;
|
|
|
|
|
curr_pe->ProbingParserTc = ProbingParser1;
|
|
|
|
|
if (curr_port->sp == NULL)
|
|
|
|
|
curr_port->sp_max_depth = new_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (new_pe->max_depth == 0)
|
|
|
|
|
curr_port->sp_max_depth = new_pe->max_depth;
|
|
|
|
|
if (curr_port->sp_max_depth != 0 &&
|
|
|
|
|
curr_port->sp_max_depth < new_pe->max_depth) {
|
|
|
|
|
curr_port->sp_max_depth = new_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
curr_port->alproto_mask |= new_pe->alproto_mask;
|
|
|
|
|
head_pe = &curr_port->sp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
AppLayerProtoDetectProbingParserElementAppend(head_pe, new_pe);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
if (curr_port->port == 0) {
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *temp_port = curr_pp->port;
|
|
|
|
|
while (temp_port != NULL && temp_port->port != 0) {
|
|
|
|
|
if (direction & STREAM_TOSERVER) {
|
|
|
|
|
if (temp_port->dp == NULL)
|
|
|
|
|
temp_port->dp_max_depth = curr_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (curr_pe->max_depth == 0)
|
|
|
|
|
temp_port->dp_max_depth = curr_pe->max_depth;
|
|
|
|
|
if (temp_port->dp_max_depth != 0 &&
|
|
|
|
|
temp_port->dp_max_depth < curr_pe->max_depth) {
|
|
|
|
|
temp_port->dp_max_depth = curr_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
AppLayerProtoDetectProbingParserElementAppend(&temp_port->dp,
|
|
|
|
|
AppLayerProtoDetectProbingParserElementDuplicate(curr_pe));
|
|
|
|
|
temp_port->alproto_mask |= curr_pe->alproto_mask;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (temp_port->sp == NULL)
|
|
|
|
|
temp_port->sp_max_depth = curr_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (curr_pe->max_depth == 0)
|
|
|
|
|
temp_port->sp_max_depth = curr_pe->max_depth;
|
|
|
|
|
if (temp_port->sp_max_depth != 0 &&
|
|
|
|
|
temp_port->sp_max_depth < curr_pe->max_depth) {
|
|
|
|
|
temp_port->sp_max_depth = curr_pe->max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
AppLayerProtoDetectProbingParserElementAppend(&temp_port->sp,
|
|
|
|
|
AppLayerProtoDetectProbingParserElementDuplicate(curr_pe));
|
|
|
|
|
temp_port->alproto_mask |= curr_pe->alproto_mask;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
temp_port = temp_port->next;
|
|
|
|
|
} /* while */
|
|
|
|
|
} /* if */
|
|
|
|
|
|
|
|
|
|
error:
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** Static Internal Calls: PM registration *****/
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectPMGetIpprotos(AppProto alproto,
|
|
|
|
|
uint8_t *ipprotos)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
const AppLayerProtoDetectPMSignature *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
|
|
|
int i, j;
|
|
|
|
|
uint8_t ipproto;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
|
|
|
|
|
ipproto = FlowGetReverseProtoMapping(i);
|
|
|
|
|
for (j = 0; j < 2; j++) {
|
|
|
|
|
AppLayerProtoDetectPMCtx *pm_ctx = &alpd_ctx.ctx_ipp[i].ctx_pm[j];
|
|
|
|
|
|
|
|
|
|
SigIntId x;
|
|
|
|
|
for (x = 0; x < pm_ctx->max_sig_id;x++) {
|
|
|
|
|
s = pm_ctx->map[x];
|
|
|
|
|
if (s->alproto == alproto)
|
|
|
|
|
ipprotos[ipproto / 8] |= 1 << (ipproto % 8);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectPMSetContentIDs(AppLayerProtoDetectPMCtx *ctx)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
typedef struct TempContainer_ {
|
|
|
|
|
PatIntId id;
|
|
|
|
|
uint16_t content_len;
|
|
|
|
|
uint8_t *content;
|
|
|
|
|
} TempContainer;
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPMSignature *s = NULL;
|
|
|
|
|
uint32_t struct_total_size = 0;
|
|
|
|
|
uint32_t content_total_size = 0;
|
|
|
|
|
/* array hash buffer */
|
|
|
|
|
uint8_t *ahb = NULL;
|
|
|
|
|
uint8_t *content = NULL;
|
|
|
|
|
uint8_t content_len = 0;
|
|
|
|
|
PatIntId max_id = 0;
|
|
|
|
|
TempContainer *struct_offset = NULL;
|
|
|
|
|
uint8_t *content_offset = NULL;
|
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
|
|
if (ctx->head == NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
for (s = ctx->head; s != NULL; s = s->next) {
|
|
|
|
|
struct_total_size += sizeof(TempContainer);
|
|
|
|
|
content_total_size += s->cd->content_len;
|
|
|
|
|
ctx->max_sig_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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ahb = SCMalloc(sizeof(uint8_t) * (struct_total_size + content_total_size));
|
|
|
|
|
if (unlikely(ahb == NULL))
|
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
|
|
struct_offset = (TempContainer *)ahb;
|
|
|
|
|
content_offset = ahb + struct_total_size;
|
|
|
|
|
for (s = ctx->head; s != NULL; s = s->next) {
|
|
|
|
|
TempContainer *tcdup = (TempContainer *)ahb;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
content = s->cd->content;
|
|
|
|
|
content_len = s->cd->content_len;
|
|
|
|
|
|
|
|
|
|
for (; tcdup != struct_offset; tcdup++) {
|
|
|
|
|
if (tcdup->content_len != content_len ||
|
|
|
|
|
SCMemcmp(tcdup->content, content, tcdup->content_len) != 0)
|
|
|
|
|
{
|
|
|
|
|
continue;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (tcdup != struct_offset) {
|
|
|
|
|
s->cd->id = tcdup->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
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
struct_offset->content_len = content_len;
|
|
|
|
|
struct_offset->content = content_offset;
|
|
|
|
|
content_offset += content_len;
|
|
|
|
|
memcpy(struct_offset->content, content, content_len);
|
|
|
|
|
struct_offset->id = max_id++;
|
|
|
|
|
s->cd->id = struct_offset->id;
|
|
|
|
|
|
|
|
|
|
struct_offset++;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ctx->max_pat_id = max_id;
|
|
|
|
|
|
|
|
|
|
goto end;
|
|
|
|
|
error:
|
|
|
|
|
ret = -1;
|
|
|
|
|
end:
|
|
|
|
|
if (ahb != NULL)
|
|
|
|
|
SCFree(ahb);
|
|
|
|
|
SCReturnInt(ret);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectPMMapSignatures(AppLayerProtoDetectPMCtx *ctx)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
int ret = 0;
|
|
|
|
|
AppLayerProtoDetectPMSignature *s, *next_s;
|
|
|
|
|
int mpm_ret;
|
|
|
|
|
SigIntId id = 0;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
ctx->map = SCMalloc(ctx->max_sig_id * sizeof(AppLayerProtoDetectPMSignature *));
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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->map == NULL)
|
|
|
|
|
goto error;
|
|
|
|
|
memset(ctx->map, 0, ctx->max_sig_id * sizeof(AppLayerProtoDetectPMSignature *));
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
/* add an array indexed by rule id to look up the sig */
|
|
|
|
|
for (s = ctx->head; 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
|
|
|
next_s = s->next;
|
|
|
|
|
s->id = id++;
|
|
|
|
|
SCLogDebug("s->id %u", s->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
|
|
|
|
|
|
|
|
if (s->cd->flags & DETECT_CONTENT_NOCASE) {
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
mpm_ret = MpmAddPatternCI(&ctx->mpm_ctx,
|
|
|
|
|
s->cd->content, s->cd->content_len,
|
|
|
|
|
0, 0, s->cd->id, s->id, 0);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (mpm_ret < 0)
|
|
|
|
|
goto error;
|
|
|
|
|
} else {
|
|
|
|
|
mpm_ret = MpmAddPatternCS(&ctx->mpm_ctx,
|
|
|
|
|
s->cd->content, s->cd->content_len,
|
|
|
|
|
0, 0, s->cd->id, s->id, 0);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (mpm_ret < 0)
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ctx->map[s->id] = s;
|
|
|
|
|
s->next = NULL;
|
|
|
|
|
s = next_s;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
ctx->head = NULL;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
goto end;
|
|
|
|
|
error:
|
|
|
|
|
ret = -1;
|
|
|
|
|
end:
|
|
|
|
|
SCReturnInt(ret);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectPMPrepareMpm(AppLayerProtoDetectPMCtx *ctx)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
int ret = 0;
|
|
|
|
|
MpmCtx *mpm_ctx = &ctx->mpm_ctx;
|
|
|
|
|
|
|
|
|
|
if (mpm_table[mpm_ctx->mpm_type].Prepare(mpm_ctx) < 0)
|
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
|
|
goto end;
|
|
|
|
|
error:
|
|
|
|
|
ret = -1;
|
|
|
|
|
end:
|
|
|
|
|
SCReturnInt(ret);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectPMFreeSignature(AppLayerProtoDetectPMSignature *sig)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
if (sig == NULL)
|
|
|
|
|
SCReturn;
|
|
|
|
|
if (sig->cd)
|
|
|
|
|
DetectContentFree(sig->cd);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
SCFree(sig);
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectPMAddSignature(AppLayerProtoDetectPMCtx *ctx, DetectContentData *cd,
|
|
|
|
|
AppProto alproto)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
int ret = 0;
|
|
|
|
|
AppLayerProtoDetectPMSignature *s = SCMalloc(sizeof(*s));
|
|
|
|
|
if (unlikely(s == NULL))
|
|
|
|
|
goto error;
|
|
|
|
|
memset(s, 0, sizeof(*s));
|
|
|
|
|
|
|
|
|
|
s->alproto = alproto;
|
|
|
|
|
s->cd = cd;
|
|
|
|
|
|
|
|
|
|
/* prepend to the list */
|
|
|
|
|
s->next = ctx->head;
|
|
|
|
|
ctx->head = s;
|
|
|
|
|
|
|
|
|
|
goto end;
|
|
|
|
|
error:
|
|
|
|
|
ret = -1;
|
|
|
|
|
end:
|
|
|
|
|
SCReturnInt(ret);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectPMRegisterPattern(uint8_t ipproto, AppProto alproto,
|
|
|
|
|
const char *pattern,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t depth, uint16_t offset,
|
|
|
|
|
uint8_t direction,
|
|
|
|
|
uint8_t is_cs)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectCtxIpproto *ctx_ipp = &alpd_ctx.ctx_ipp[FlowGetProtoMapping(ipproto)];
|
|
|
|
|
AppLayerProtoDetectPMCtx *ctx_pm = NULL;
|
|
|
|
|
DetectContentData *cd;
|
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
|
|
cd = DetectContentParseEncloseQuotes(alpd_ctx.spm_global_thread_ctx,
|
|
|
|
|
pattern);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (cd == NULL)
|
|
|
|
|
goto error;
|
|
|
|
|
cd->depth = depth;
|
|
|
|
|
cd->offset = offset;
|
|
|
|
|
if (!is_cs) {
|
|
|
|
|
/* Rebuild as nocase */
|
|
|
|
|
SpmDestroyCtx(cd->spm_ctx);
|
|
|
|
|
cd->spm_ctx = SpmInitCtx(cd->content, cd->content_len, 1,
|
|
|
|
|
alpd_ctx.spm_global_thread_ctx);
|
|
|
|
|
if (cd->spm_ctx == NULL) {
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
cd->flags |= DETECT_CONTENT_NOCASE;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (depth < cd->content_len)
|
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
|
|
if (direction & STREAM_TOSERVER)
|
|
|
|
|
ctx_pm = (AppLayerProtoDetectPMCtx *)&ctx_ipp->ctx_pm[0];
|
|
|
|
|
else
|
|
|
|
|
ctx_pm = (AppLayerProtoDetectPMCtx *)&ctx_ipp->ctx_pm[1];
|
|
|
|
|
|
|
|
|
|
if (depth > ctx_pm->max_len)
|
|
|
|
|
ctx_pm->max_len = depth;
|
|
|
|
|
if (depth < ctx_pm->min_len)
|
|
|
|
|
ctx_pm->min_len = depth;
|
|
|
|
|
|
|
|
|
|
/* Finally turn it into a signature and add to the ctx. */
|
|
|
|
|
AppLayerProtoDetectPMAddSignature(ctx_pm, cd, alproto);
|
|
|
|
|
|
|
|
|
|
goto end;
|
|
|
|
|
error:
|
|
|
|
|
ret = -1;
|
|
|
|
|
end:
|
|
|
|
|
SCReturnInt(ret);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** Protocol Retrieval *****/
|
|
|
|
|
|
|
|
|
|
AppProto AppLayerProtoDetectGetProto(AppLayerProtoDetectThreadCtx *tctx,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
Flow *f,
|
|
|
|
|
uint8_t *buf, uint32_t buflen,
|
|
|
|
|
uint8_t ipproto, uint8_t direction)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
AppProto alproto = ALPROTO_UNKNOWN;
|
|
|
|
|
AppProto pm_results[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
|
|
|
uint16_t pm_matches;
|
|
|
|
|
|
|
|
|
|
if (!FLOW_IS_PM_DONE(f, direction)) {
|
|
|
|
|
pm_matches = AppLayerProtoDetectPMGetProto(tctx, f,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
buf, buflen,
|
|
|
|
|
direction,
|
|
|
|
|
ipproto,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (pm_matches > 0) {
|
|
|
|
|
alproto = pm_results[0];
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!FLOW_IS_PP_DONE(f, direction))
|
|
|
|
|
alproto = AppLayerProtoDetectPPGetProto(f, buf, buflen, ipproto, direction);
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
SCReturnUInt(alproto);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void AppLayerProtoDetectFreeProbingParsers(AppLayerProtoDetectProbingParser *pp)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParser *tmp_pp = NULL;
|
|
|
|
|
|
|
|
|
|
if (pp == NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
while (pp != NULL) {
|
|
|
|
|
tmp_pp = pp->next;
|
|
|
|
|
AppLayerProtoDetectProbingParserFree(pp);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
pp = tmp_pp;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** State Preparation *****/
|
|
|
|
|
|
|
|
|
|
int AppLayerProtoDetectPrepareState(void)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPMCtx *ctx_pm;
|
|
|
|
|
int i, j;
|
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
|
|
|
|
|
for (j = 0; j < 2; j++) {
|
|
|
|
|
ctx_pm = &alpd_ctx.ctx_ipp[i].ctx_pm[j];
|
|
|
|
|
|
|
|
|
|
if (AppLayerProtoDetectPMSetContentIDs(ctx_pm) < 0)
|
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
|
|
if (ctx_pm->max_sig_id == 0)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (AppLayerProtoDetectPMMapSignatures(ctx_pm) < 0)
|
|
|
|
|
goto error;
|
|
|
|
|
if (AppLayerProtoDetectPMPrepareMpm(ctx_pm) < 0)
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef DEBUG
|
|
|
|
|
if (SCLogDebugEnabled()) {
|
|
|
|
|
AppLayerProtoDetectPrintProbingParsers(alpd_ctx.ctx_pp);
|
|
|
|
|
}
|
|
|
|
|
#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
|
|
|
goto end;
|
|
|
|
|
error:
|
|
|
|
|
ret = -1;
|
|
|
|
|
end:
|
|
|
|
|
SCReturnInt(ret);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** PP registration *****/
|
|
|
|
|
|
|
|
|
|
/** \brief register parser at a port
|
|
|
|
|
*
|
|
|
|
|
* \param direction STREAM_TOSERVER or STREAM_TOCLIENT for dp or sp
|
|
|
|
|
*/
|
|
|
|
|
void AppLayerProtoDetectPPRegister(uint8_t ipproto,
|
|
|
|
|
const char *portstr,
|
|
|
|
|
AppProto alproto,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t min_depth, uint16_t max_depth,
|
|
|
|
|
uint8_t direction,
|
|
|
|
|
ProbingParserFPtr ProbingParser1,
|
|
|
|
|
ProbingParserFPtr ProbingParser2)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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();
|
|
|
|
|
|
|
|
|
|
DetectPort *head = NULL;
|
|
|
|
|
DetectPortParse(NULL,&head, portstr);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
DetectPort *temp_dp = head;
|
|
|
|
|
while (temp_dp != NULL) {
|
|
|
|
|
uint32_t port = temp_dp->port;
|
|
|
|
|
if (port == 0 && temp_dp->port2 != 0)
|
|
|
|
|
port++;
|
|
|
|
|
for ( ; port <= temp_dp->port2; port++) {
|
|
|
|
|
AppLayerProtoDetectInsertNewProbingParser(&alpd_ctx.ctx_pp,
|
|
|
|
|
ipproto,
|
|
|
|
|
port,
|
|
|
|
|
alproto,
|
|
|
|
|
min_depth, max_depth,
|
|
|
|
|
direction,
|
|
|
|
|
ProbingParser1,
|
|
|
|
|
ProbingParser2);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
temp_dp = temp_dp->next;
|
|
|
|
|
}
|
|
|
|
|
DetectPortCleanupList(head);
|
|
|
|
|
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int AppLayerProtoDetectPPParseConfPorts(const char *ipproto_name,
|
|
|
|
|
uint8_t ipproto,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
const char *alproto_name,
|
|
|
|
|
AppProto alproto,
|
|
|
|
|
uint16_t min_depth, uint16_t max_depth,
|
|
|
|
|
ProbingParserFPtr ProbingParserTs,
|
|
|
|
|
ProbingParserFPtr ProbingParserTc)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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();
|
|
|
|
|
|
|
|
|
|
char param[100];
|
|
|
|
|
int r;
|
|
|
|
|
ConfNode *node;
|
|
|
|
|
ConfNode *port_node = NULL;
|
|
|
|
|
int config = 0;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
r = snprintf(param, sizeof(param), "%s%s%s", "app-layer.protocols.",
|
|
|
|
|
alproto_name, ".detection-ports");
|
|
|
|
|
if (r < 0) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "snprintf failure.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
} else if (r > (int)sizeof(param)) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
|
|
|
|
node = ConfGetNode(param);
|
|
|
|
|
if (node == NULL) {
|
|
|
|
|
SCLogDebug("Entry for %s not found.", param);
|
|
|
|
|
r = snprintf(param, sizeof(param), "%s%s%s%s%s", "app-layer.protocols.",
|
|
|
|
|
alproto_name, ".", ipproto_name, ".detection-ports");
|
|
|
|
|
if (r < 0) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "snprintf failure.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
} else if (r > (int)sizeof(param)) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
|
|
|
|
node = ConfGetNode(param);
|
|
|
|
|
if (node == NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* detect by destination port of the flow (e.g. port 53 for DNS) */
|
|
|
|
|
port_node = ConfNodeLookupChild(node, "dp");
|
|
|
|
|
if (port_node == NULL)
|
|
|
|
|
port_node = ConfNodeLookupChild(node, "toserver");
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (port_node != NULL && port_node->val != NULL) {
|
|
|
|
|
AppLayerProtoDetectPPRegister(ipproto,
|
|
|
|
|
port_node->val,
|
|
|
|
|
alproto,
|
|
|
|
|
min_depth, max_depth,
|
|
|
|
|
STREAM_TOSERVER, /* to indicate dp */
|
|
|
|
|
ProbingParserTs, ProbingParserTc);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* detect by source port of flow */
|
|
|
|
|
port_node = ConfNodeLookupChild(node, "sp");
|
|
|
|
|
if (port_node == NULL)
|
|
|
|
|
port_node = ConfNodeLookupChild(node, "toclient");
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
if (port_node != NULL && port_node->val != NULL) {
|
|
|
|
|
AppLayerProtoDetectPPRegister(ipproto,
|
|
|
|
|
port_node->val,
|
|
|
|
|
alproto,
|
|
|
|
|
min_depth, max_depth,
|
|
|
|
|
STREAM_TOCLIENT, /* to indicate sp */
|
|
|
|
|
ProbingParserTc, ProbingParserTs);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
config = 1;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
SCReturnInt(config);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** PM registration *****/
|
|
|
|
|
|
|
|
|
|
int AppLayerProtoDetectPMRegisterPatternCS(uint8_t ipproto, AppProto alproto,
|
|
|
|
|
const char *pattern,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t depth, uint16_t offset,
|
|
|
|
|
uint8_t direction)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
int r = 0;
|
|
|
|
|
r = AppLayerProtoDetectPMRegisterPattern(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
|
|
|
pattern,
|
|
|
|
|
depth, offset,
|
|
|
|
|
direction,
|
|
|
|
|
1 /* case-sensitive */);
|
|
|
|
|
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 AppLayerProtoDetectPMRegisterPatternCI(uint8_t ipproto, AppProto alproto,
|
|
|
|
|
const char *pattern,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t depth, uint16_t offset,
|
|
|
|
|
uint8_t direction)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
int r = 0;
|
|
|
|
|
r = AppLayerProtoDetectPMRegisterPattern(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
|
|
|
pattern,
|
|
|
|
|
depth, offset,
|
|
|
|
|
direction,
|
|
|
|
|
0 /* !case-sensitive */);
|
|
|
|
|
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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** Setup/General Registration *****/
|
|
|
|
|
|
|
|
|
|
int AppLayerProtoDetectSetup(void)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
int i, j;
|
|
|
|
|
|
|
|
|
|
memset(&alpd_ctx, 0, sizeof(alpd_ctx));
|
|
|
|
|
|
|
|
|
|
uint16_t spm_matcher = SinglePatternMatchDefaultMatcher();
|
|
|
|
|
uint16_t mpm_matcher = PatternMatchDefaultMatcher();
|
|
|
|
|
|
|
|
|
|
#ifdef __SC_CUDA_SUPPORT__
|
|
|
|
|
/* CUDA won't work here, so fall back to AC */
|
|
|
|
|
if (mpm_matcher == MPM_AC_CUDA)
|
|
|
|
|
mpm_matcher = mpm_default_matcher;
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
alpd_ctx.spm_global_thread_ctx = SpmInitGlobalThreadCtx(spm_matcher);
|
|
|
|
|
if (alpd_ctx.spm_global_thread_ctx == NULL) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "Unable to alloc SpmGlobalThreadCtx.");
|
|
|
|
|
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
|
|
|
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
|
|
|
|
|
for (j = 0; j < 2; j++) {
|
|
|
|
|
MpmInitCtx(&alpd_ctx.ctx_ipp[i].ctx_pm[j].mpm_ctx, mpm_matcher);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* \todo incomplete. Need more work.
|
|
|
|
|
*/
|
|
|
|
|
int AppLayerProtoDetectDeSetup(void)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
int ipproto_map = 0;
|
|
|
|
|
int dir = 0;
|
|
|
|
|
PatIntId id = 0;
|
|
|
|
|
AppLayerProtoDetectPMCtx *pm_ctx = NULL;
|
|
|
|
|
AppLayerProtoDetectPMSignature *sig = NULL;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (ipproto_map = 0; ipproto_map < FLOW_PROTO_DEFAULT; ipproto_map++) {
|
|
|
|
|
for (dir = 0; dir < 2; dir++) {
|
|
|
|
|
pm_ctx = &alpd_ctx.ctx_ipp[ipproto_map].ctx_pm[dir];
|
|
|
|
|
mpm_table[pm_ctx->mpm_ctx.mpm_type].DestroyCtx(&pm_ctx->mpm_ctx);
|
|
|
|
|
for (id = 0; id < pm_ctx->max_sig_id; 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
|
|
|
sig = pm_ctx->map[id];
|
|
|
|
|
AppLayerProtoDetectPMFreeSignature(sig);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SpmDestroyGlobalThreadCtx(alpd_ctx.spm_global_thread_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
|
|
|
AppLayerProtoDetectFreeProbingParsers(alpd_ctx.ctx_pp);
|
|
|
|
|
|
|
|
|
|
SCReturnInt(0);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void AppLayerProtoDetectRegisterProtocol(AppProto alproto, 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();
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.alproto_names[alproto] != NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
alpd_ctx.alproto_names[alproto] = alproto_name;
|
|
|
|
|
|
|
|
|
|
goto end;
|
|
|
|
|
end:
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** \brief request applayer to wrap up this protocol and rerun protocol
|
|
|
|
|
* detection.
|
|
|
|
|
*
|
|
|
|
|
* When this is called, the old session is reset unconditionally. A
|
|
|
|
|
* 'detect/log' flush packet is generated for both direction before
|
|
|
|
|
* the reset, so allow for final detection and logging.
|
|
|
|
|
*
|
|
|
|
|
* \param f flow to act on
|
|
|
|
|
* \param dp destination port to use in protocol detection. Set to 443
|
|
|
|
|
* for start tls, set to the HTTP uri port for CONNECT and
|
|
|
|
|
* set to 0 to not use it.
|
|
|
|
|
* \param expect_proto expected protocol. AppLayer event will be set if
|
|
|
|
|
* detected protocol differs from this.
|
|
|
|
|
*/
|
|
|
|
|
void AppLayerRequestProtocolChange(Flow *f, uint16_t dp, AppProto expect_proto)
|
|
|
|
|
{
|
|
|
|
|
FlowSetChangeProtoFlag(f);
|
|
|
|
|
f->protodetect_dp = dp;
|
|
|
|
|
f->alproto_expect = expect_proto;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** \brief request applayer to wrap up this protocol and rerun protocol
|
|
|
|
|
* detection with expectation of TLS. Used by STARTTLS.
|
|
|
|
|
*
|
|
|
|
|
* Sets detection port to 443 to make port based TLS detection work for
|
|
|
|
|
* SMTP, FTP etc as well.
|
|
|
|
|
*
|
|
|
|
|
* \param f flow to act on
|
|
|
|
|
*/
|
|
|
|
|
void AppLayerRequestProtocolTLSUpgrade(Flow *f)
|
|
|
|
|
{
|
|
|
|
|
AppLayerRequestProtocolChange(f, 443, ALPROTO_TLS);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void AppLayerProtoDetectReset(Flow *f)
|
|
|
|
|
{
|
|
|
|
|
FlowUnsetChangeProtoFlag(f);
|
|
|
|
|
FLOW_RESET_PM_DONE(f, STREAM_TOSERVER);
|
|
|
|
|
FLOW_RESET_PM_DONE(f, STREAM_TOCLIENT);
|
|
|
|
|
FLOW_RESET_PP_DONE(f, STREAM_TOSERVER);
|
|
|
|
|
FLOW_RESET_PP_DONE(f, STREAM_TOCLIENT);
|
|
|
|
|
f->probing_parser_toserver_alproto_masks = 0;
|
|
|
|
|
f->probing_parser_toclient_alproto_masks = 0;
|
|
|
|
|
|
|
|
|
|
AppLayerParserStateCleanup(f, f->alstate, f->alparser);
|
|
|
|
|
f->alstate = NULL;
|
|
|
|
|
f->alparser = NULL;
|
|
|
|
|
f->alproto = ALPROTO_UNKNOWN;
|
|
|
|
|
f->alproto_ts = ALPROTO_UNKNOWN;
|
|
|
|
|
f->alproto_tc = ALPROTO_UNKNOWN;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int AppLayerProtoDetectConfProtoDetectionEnabled(const char *ipproto,
|
|
|
|
|
const char *alproto)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
BUG_ON(ipproto == NULL || alproto == NULL);
|
|
|
|
|
|
|
|
|
|
int enabled = 1;
|
|
|
|
|
char param[100];
|
|
|
|
|
ConfNode *node;
|
|
|
|
|
int r;
|
|
|
|
|
|
|
|
|
|
#ifdef AFLFUZZ_APPLAYER
|
|
|
|
|
goto enabled;
|
|
|
|
|
#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
|
|
|
if (RunmodeIsUnittests())
|
|
|
|
|
goto enabled;
|
|
|
|
|
|
|
|
|
|
r = snprintf(param, sizeof(param), "%s%s%s", "app-layer.protocols.",
|
|
|
|
|
alproto, ".enabled");
|
|
|
|
|
if (r < 0) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "snprintf failure.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
} else if (r > (int)sizeof(param)) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
node = ConfGetNode(param);
|
|
|
|
|
if (node == NULL) {
|
|
|
|
|
SCLogDebug("Entry for %s not found.", param);
|
|
|
|
|
r = snprintf(param, sizeof(param), "%s%s%s%s%s", "app-layer.protocols.",
|
|
|
|
|
alproto, ".", ipproto, ".enabled");
|
|
|
|
|
if (r < 0) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "snprintf failure.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
} else if (r > (int)sizeof(param)) {
|
|
|
|
|
SCLogError(SC_ERR_FATAL, "buffer not big enough to write param.");
|
|
|
|
|
exit(EXIT_FAILURE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
node = ConfGetNode(param);
|
|
|
|
|
if (node == NULL) {
|
|
|
|
|
SCLogDebug("Entry for %s not found.", param);
|
|
|
|
|
goto enabled;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (node->val) {
|
|
|
|
|
if (ConfValIsTrue(node->val)) {
|
|
|
|
|
goto enabled;
|
|
|
|
|
} else if (ConfValIsFalse(node->val)) {
|
|
|
|
|
goto disabled;
|
|
|
|
|
} else if (strcasecmp(node->val, "detection-only") == 0) {
|
|
|
|
|
goto 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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Invalid or null value. */
|
|
|
|
|
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);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectThreadCtx *AppLayerProtoDetectGetCtxThread(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();
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx = NULL;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
MpmCtx *mpm_ctx;
|
|
|
|
|
MpmThreadCtx *mpm_tctx;
|
|
|
|
|
int i, j;
|
|
|
|
|
PatIntId max_pat_id = 0;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
|
|
|
|
|
for (j = 0; j < 2; j++) {
|
|
|
|
|
if (max_pat_id == 0) {
|
|
|
|
|
max_pat_id = alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id;
|
|
|
|
|
|
|
|
|
|
} else if (alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id &&
|
|
|
|
|
max_pat_id < alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id)
|
|
|
|
|
{
|
|
|
|
|
max_pat_id = alpd_ctx.ctx_ipp[i].ctx_pm[j].max_pat_id;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
alpd_tctx = SCMalloc(sizeof(*alpd_tctx));
|
|
|
|
|
if (alpd_tctx == NULL)
|
|
|
|
|
goto error;
|
|
|
|
|
memset(alpd_tctx, 0, sizeof(*alpd_tctx));
|
|
|
|
|
|
|
|
|
|
/* Get the max pat id for all the mpm ctxs. */
|
|
|
|
|
if (PmqSetup(&alpd_tctx->pmq) < 0)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto error;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < FLOW_PROTO_DEFAULT; i++) {
|
|
|
|
|
for (j = 0; j < 2; j++) {
|
|
|
|
|
mpm_ctx = &alpd_ctx.ctx_ipp[i].ctx_pm[j].mpm_ctx;
|
|
|
|
|
mpm_tctx = &alpd_tctx->mpm_tctx[i][j];
|
|
|
|
|
mpm_table[mpm_ctx->mpm_type].InitThreadCtx(mpm_ctx, mpm_tctx);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
alpd_tctx->spm_thread_ctx = SpmMakeThreadCtx(alpd_ctx.spm_global_thread_ctx);
|
|
|
|
|
if (alpd_tctx->spm_thread_ctx == NULL) {
|
|
|
|
|
goto error;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto end;
|
|
|
|
|
error:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
alpd_tctx = NULL;
|
|
|
|
|
end:
|
|
|
|
|
SCReturnPtr(alpd_tctx, "AppLayerProtoDetectThreadCtx");
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void AppLayerProtoDetectDestroyCtxThread(AppLayerProtoDetectThreadCtx *alpd_tctx)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
MpmCtx *mpm_ctx;
|
|
|
|
|
MpmThreadCtx *mpm_tctx;
|
|
|
|
|
int ipproto_map, dir;
|
|
|
|
|
|
|
|
|
|
for (ipproto_map = 0; ipproto_map < FLOW_PROTO_DEFAULT; ipproto_map++) {
|
|
|
|
|
for (dir = 0; dir < 2; dir++) {
|
|
|
|
|
mpm_ctx = &alpd_ctx.ctx_ipp[ipproto_map].ctx_pm[dir].mpm_ctx;
|
|
|
|
|
mpm_tctx = &alpd_tctx->mpm_tctx[ipproto_map][dir];
|
|
|
|
|
mpm_table[mpm_ctx->mpm_type].DestroyThreadCtx(mpm_ctx, mpm_tctx);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
PmqFree(&alpd_tctx->pmq);
|
|
|
|
|
if (alpd_tctx->spm_thread_ctx != NULL) {
|
|
|
|
|
SpmDestroyThreadCtx(alpd_tctx->spm_thread_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
|
|
|
SCFree(alpd_tctx);
|
|
|
|
|
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** Utility *****/
|
|
|
|
|
|
|
|
|
|
void AppLayerProtoDetectSupportedIpprotos(AppProto alproto, uint8_t *ipprotos)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPMGetIpprotos(alproto, ipprotos);
|
|
|
|
|
AppLayerProtoDetectPPGetIpprotos(alproto, ipprotos);
|
|
|
|
|
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
AppProto AppLayerProtoDetectGetProtoByName(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();
|
|
|
|
|
|
|
|
|
|
AppProto a;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 (a = 0; a < ALPROTO_MAX; a++) {
|
|
|
|
|
if (alpd_ctx.alproto_names[a] != NULL &&
|
|
|
|
|
strlen(alpd_ctx.alproto_names[a]) == strlen(alproto_name) &&
|
|
|
|
|
(SCMemcmp(alpd_ctx.alproto_names[a], alproto_name, strlen(alproto_name)) == 0))
|
|
|
|
|
{
|
|
|
|
|
SCReturnCT(a, "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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SCReturnCT(ALPROTO_UNKNOWN, "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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
const char *AppLayerProtoDetectGetProtoName(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
|
|
|
{
|
|
|
|
|
return alpd_ctx.alproto_names[alproto];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void AppLayerProtoDetectSupportedAppProtocols(AppProto *alprotos)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
memset(alprotos, 0, ALPROTO_MAX * sizeof(AppProto));
|
|
|
|
|
|
|
|
|
|
int alproto;
|
|
|
|
|
|
|
|
|
|
for (alproto = 0; alproto != ALPROTO_MAX; alproto++) {
|
|
|
|
|
if (alpd_ctx.alproto_names[alproto] != NULL)
|
|
|
|
|
alprotos[alproto] = 1;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/***** Unittests *****/
|
|
|
|
|
|
|
|
|
|
#ifdef UNITTESTS
|
|
|
|
|
|
|
|
|
|
static AppLayerProtoDetectCtx alpd_ctx_ut;
|
|
|
|
|
|
|
|
|
|
void AppLayerProtoDetectUnittestCtxBackup(void)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
alpd_ctx_ut = alpd_ctx;
|
|
|
|
|
memset(&alpd_ctx, 0, sizeof(alpd_ctx));
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void AppLayerProtoDetectUnittestCtxRestore(void)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
alpd_ctx = alpd_ctx_ut;
|
|
|
|
|
memset(&alpd_ctx_ut, 0, sizeof(alpd_ctx_ut));
|
|
|
|
|
SCReturn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest01(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
|
|
|
|
|
buf = "HTTP";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
buf = "GET";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOSERVER);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 1) {
|
|
|
|
|
printf("Failure - "
|
|
|
|
|
"alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("Failure - "
|
|
|
|
|
"alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
r = 1;
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest02(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
int r = 0;
|
|
|
|
|
|
|
|
|
|
const char *buf = "HTTP";
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
buf = "ftp";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
r = 1;
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest03(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "HTTP/1.1 200 OK\r\nServer: Apache/1.0\r\n\r\n";
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
|
|
|
|
buf = "HTTP";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
buf = "220 ";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
|
|
|
|
|
printf("cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
r = 1;
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest04(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "HTTP/1.1 200 OK\r\nServer: Apache/1.0\r\n\r\n";
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
|
|
|
|
buf = "200 ";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 13, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
|
|
|
|
|
printf("cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
r = 1;
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest05(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "HTTP/1.1 200 OK\r\nServer: Apache/1.0\r\n\r\n<HTML><BODY>Blahblah</BODY></HTML>";
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
|
|
|
|
buf = "HTTP";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
buf = "220 ";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
|
|
|
|
|
printf("cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest06(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "220 Welcome to the OISF FTP server\r\n";
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
|
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
buf = "HTTP";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
buf = "220 ";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_FTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_FTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_FTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1]->alproto != ALPROTO_HTTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[1].alproto != ALPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_FTP) {
|
|
|
|
|
printf("cnt != 1 && pm_results[0] != AlPROTO_FTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest07(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "220 Welcome to the OISF HTTP/FTP server\r\n";
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
buf = "HTTP";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 0) {
|
|
|
|
|
printf("cnt != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest08(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = {
|
|
|
|
|
0x00, 0x00, 0x00, 0x85, 0xff, 0x53, 0x4d, 0x42,
|
|
|
|
|
0x72, 0x00, 0x00, 0x00, 0x00, 0x18, 0x53, 0xc8,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0xfe,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x62, 0x00, 0x02,
|
|
|
|
|
0x50, 0x43, 0x20, 0x4e, 0x45, 0x54, 0x57, 0x4f,
|
|
|
|
|
0x52, 0x4b, 0x20, 0x50, 0x52, 0x4f, 0x47, 0x52,
|
|
|
|
|
0x41, 0x4d, 0x20, 0x31, 0x2e, 0x30, 0x00, 0x02,
|
|
|
|
|
0x4c, 0x41, 0x4e, 0x4d, 0x41, 0x4e, 0x31, 0x2e,
|
|
|
|
|
0x30, 0x00, 0x02, 0x57, 0x69, 0x6e, 0x64, 0x6f,
|
|
|
|
|
0x77, 0x73, 0x20, 0x66, 0x6f, 0x72, 0x20, 0x57,
|
|
|
|
|
0x6f, 0x72, 0x6b, 0x67, 0x72, 0x6f, 0x75, 0x70,
|
|
|
|
|
0x73, 0x20, 0x33, 0x2e, 0x31, 0x61, 0x00, 0x02,
|
|
|
|
|
0x4c, 0x4d, 0x31, 0x2e, 0x32, 0x58, 0x30, 0x30,
|
|
|
|
|
0x32, 0x00, 0x02, 0x4c, 0x41, 0x4e, 0x4d, 0x41,
|
|
|
|
|
0x4e, 0x32, 0x2e, 0x31, 0x00, 0x02, 0x4e, 0x54,
|
|
|
|
|
0x20, 0x4c, 0x4d, 0x20, 0x30, 0x2e, 0x31, 0x32,
|
|
|
|
|
0x00
|
|
|
|
|
};
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
buf = "|ff|SMB";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_SMB, buf, 8, 4, STREAM_TOCLIENT);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_SMB) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_SMB\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_SMB) {
|
|
|
|
|
printf("cnt != 1 && pm_results[0] != AlPROTO_SMB\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest09(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = {
|
|
|
|
|
0x00, 0x00, 0x00, 0x66, 0xfe, 0x53, 0x4d, 0x42,
|
|
|
|
|
0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x24, 0x00, 0x01, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x00, 0x02, 0x02
|
|
|
|
|
};
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
|
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
buf = "|fe|SMB";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_SMB2, buf, 8, 4, STREAM_TOCLIENT);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_SMB2) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_SMB2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_SMB2) {
|
|
|
|
|
printf("cnt != 1 && pm_results[0] != AlPROTO_SMB2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest10(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = {
|
|
|
|
|
0x05, 0x00, 0x0b, 0x03, 0x10, 0x00, 0x00, 0x00,
|
|
|
|
|
0x48, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0xd0, 0x16, 0xd0, 0x16, 0x00, 0x00, 0x00, 0x00,
|
|
|
|
|
0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00,
|
|
|
|
|
0xb8, 0x4a, 0x9f, 0x4d, 0x1c, 0x7d, 0xcf, 0x11,
|
|
|
|
|
0x86, 0x1e, 0x00, 0x20, 0xaf, 0x6e, 0x7c, 0x57,
|
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x04, 0x5d, 0x88, 0x8a,
|
|
|
|
|
0xeb, 0x1c, 0xc9, 0x11, 0x9f, 0xe8, 0x08, 0x00,
|
|
|
|
|
0x2b, 0x10, 0x48, 0x60, 0x02, 0x00, 0x00, 0x00
|
|
|
|
|
};
|
|
|
|
|
const char *buf;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
buf = "|05 00|";
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_DCERPC, buf, 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_DCERPC) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0].alproto != ALPROTO_DCERPC\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_DCERPC) {
|
|
|
|
|
printf("cnt != 1 && pm_results[0] != AlPROTO_DCERPC\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/**
|
|
|
|
|
* \test Why we still get http for connect... obviously because
|
|
|
|
|
* we also match on the reply, duh
|
|
|
|
|
*/
|
|
|
|
|
static int AppLayerProtoDetectTest11(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "CONNECT www.ssllabs.com:443 HTTP/1.0\r\n";
|
|
|
|
|
uint8_t l7data_resp[] = "HTTP/1.1 405 Method Not Allowed\r\n";
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "GET", 3, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "PUT", 3, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "POST", 4, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "TRACE", 5, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "OPTIONS", 7, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "CONNECT", 7, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 7) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 7\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].max_pat_id != 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map == NULL) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map != NULL\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[1]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[2]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[3]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[4]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[5]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[6]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP)
|
|
|
|
|
{
|
|
|
|
|
printf("failure 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
uint32_t cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
|
|
|
|
|
printf("l7data - cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data_resp, sizeof(l7data_resp),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
|
|
|
|
|
printf("l7data_resp - cnt != 1 && pm_results[0] != AlPROTO_HTTP\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
/**
|
|
|
|
|
* \test AlpProtoSignature test
|
|
|
|
|
*/
|
|
|
|
|
static int AppLayerProtoDetectTest12(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_TCP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].head == NULL ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map != NULL)
|
|
|
|
|
{
|
|
|
|
|
printf("failure 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].max_pat_id != 1) {
|
|
|
|
|
printf("failure 2\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].head != NULL ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map == NULL)
|
|
|
|
|
{
|
|
|
|
|
printf("failure 3\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP) {
|
|
|
|
|
printf("failure 4\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->cd->id != 0) {
|
|
|
|
|
printf("failure 5\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_TCP].ctx_pm[0].map[0]->next != NULL) {
|
|
|
|
|
printf("failure 6\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
r = 1;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* \test What about if we add some sigs only for udp but call for tcp?
|
|
|
|
|
* It should not detect any proto
|
|
|
|
|
*/
|
|
|
|
|
static int AppLayerProtoDetectTest13(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "CONNECT www.ssllabs.com:443 HTTP/1.0\r\n";
|
|
|
|
|
uint8_t l7data_resp[] = "HTTP/1.1 405 Method Not Allowed\r\n";
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt;
|
|
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_TCP);
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "GET", 3, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "PUT", 3, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "POST", 4, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "TRACE", 5, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "OPTIONS", 7, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "CONNECT", 7, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[1]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[2]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[3]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[4]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[5]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[6]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP)
|
|
|
|
|
{
|
|
|
|
|
printf("failure 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 0) {
|
|
|
|
|
printf("l7data - cnt != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data_resp, sizeof(l7data_resp),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_TCP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 0) {
|
|
|
|
|
printf("l7data_resp - cnt != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
|
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* \test What about if we add some sigs only for udp calling it for UDP?
|
|
|
|
|
* It should detect ALPROTO_HTTP (over udp). This is just a check
|
|
|
|
|
* to ensure that TCP/UDP differences work correctly.
|
|
|
|
|
*/
|
|
|
|
|
static int AppLayerProtoDetectTest14(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
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
uint8_t l7data[] = "CONNECT www.ssllabs.com:443 HTTP/1.0\r\n";
|
|
|
|
|
uint8_t l7data_resp[] = "HTTP/1.1 405 Method Not Allowed\r\n";
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
int r = 0;
|
|
|
|
|
Flow f;
|
|
|
|
|
AppProto pm_results[ALPROTO_MAX];
|
|
|
|
|
AppLayerProtoDetectThreadCtx *alpd_tctx;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint32_t cnt;
|
|
|
|
|
|
|
|
|
|
memset(&f, 0x00, sizeof(f));
|
|
|
|
|
f.protomap = FlowGetProtoMapping(IPPROTO_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
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "GET", 3, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "PUT", 3, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "POST", 4, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "TRACE", 5, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "OPTIONS", 7, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "CONNECT", 7, 0, STREAM_TOSERVER);
|
|
|
|
|
AppLayerProtoDetectPMRegisterPatternCS(IPPROTO_UDP, ALPROTO_HTTP, "HTTP", 4, 0, STREAM_TOCLIENT);
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPrepareState();
|
|
|
|
|
/* AppLayerProtoDetectGetCtxThread() should be called post AppLayerProtoDetectPrepareState(), since
|
|
|
|
|
* it sets internal structures which depends on the above function. */
|
|
|
|
|
alpd_tctx = AppLayerProtoDetectGetCtxThread();
|
|
|
|
|
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].max_pat_id != 7\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1) {
|
|
|
|
|
printf("alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].max_pat_id != 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[0]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[1]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[2]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[3]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[4]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[5]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[0].map[6]->alproto != ALPROTO_HTTP ||
|
|
|
|
|
alpd_ctx.ctx_ipp[FLOW_PROTO_UDP].ctx_pm[1].map[0]->alproto != ALPROTO_HTTP)
|
|
|
|
|
{
|
|
|
|
|
printf("failure 1\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data, sizeof(l7data),
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
IPPROTO_UDP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
|
|
|
|
|
printf("l7data - cnt != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
memset(pm_results, 0, sizeof(pm_results));
|
|
|
|
|
cnt = AppLayerProtoDetectPMGetProto(alpd_tctx,
|
|
|
|
|
&f,
|
|
|
|
|
l7data_resp, sizeof(l7data_resp),
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
IPPROTO_UDP,
|
|
|
|
|
pm_results);
|
|
|
|
|
if (cnt != 1 && pm_results[0] != ALPROTO_HTTP) {
|
|
|
|
|
printf("l7data_resp - cnt != 0\n");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
r = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alpd_tctx != NULL)
|
|
|
|
|
AppLayerProtoDetectDestroyCtxThread(alpd_tctx);
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return r;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectPPTestDataElement_ {
|
|
|
|
|
const char *alproto_name;
|
|
|
|
|
AppProto alproto;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
uint16_t port;
|
|
|
|
|
uint32_t alproto_mask;
|
|
|
|
|
uint32_t min_depth;
|
|
|
|
|
uint32_t max_depth;
|
|
|
|
|
} AppLayerProtoDetectPPTestDataElement;
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectPPTestDataPort_ {
|
|
|
|
|
uint16_t port;
|
|
|
|
|
uint32_t alproto_mask;
|
|
|
|
|
uint16_t dp_max_depth;
|
|
|
|
|
uint16_t sp_max_depth;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement *toserver_element;
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement *toclient_element;
|
|
|
|
|
int ts_no_of_element;
|
|
|
|
|
int tc_no_of_element;
|
|
|
|
|
} AppLayerProtoDetectPPTestDataPort;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
typedef struct AppLayerProtoDetectPPTestDataIPProto_ {
|
|
|
|
|
uint8_t ipproto;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataPort *port;
|
|
|
|
|
int no_of_port;
|
|
|
|
|
} AppLayerProtoDetectPPTestDataIPProto;
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectPPTestData(AppLayerProtoDetectProbingParser *pp,
|
|
|
|
|
AppLayerProtoDetectPPTestDataIPProto *ip_proto,
|
|
|
|
|
int no_of_ip_proto)
|
|
|
|
|
{
|
|
|
|
|
int result = 0;
|
|
|
|
|
int i = -1, j = -1 , k = -1;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#ifdef DEBUG
|
|
|
|
|
int dir = 0;
|
|
|
|
|
#endif
|
|
|
|
|
for (i = 0; i < no_of_ip_proto; i++, pp = pp->next) {
|
|
|
|
|
if (pp->ipproto != ip_proto[i].ipproto)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParserPort *pp_port = pp->port;
|
|
|
|
|
for (k = 0; k < ip_proto[i].no_of_port; k++, pp_port = pp_port->next) {
|
|
|
|
|
if (pp_port->port != ip_proto[i].port[k].port)
|
|
|
|
|
goto end;
|
|
|
|
|
if (pp_port->alproto_mask != ip_proto[i].port[k].alproto_mask)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto end;
|
|
|
|
|
if (pp_port->alproto_mask != ip_proto[i].port[k].alproto_mask)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto end;
|
|
|
|
|
if (pp_port->dp_max_depth != ip_proto[i].port[k].dp_max_depth)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto end;
|
|
|
|
|
if (pp_port->sp_max_depth != ip_proto[i].port[k].sp_max_depth)
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectProbingParserElement *pp_element = pp_port->dp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#ifdef DEBUG
|
|
|
|
|
dir = 0;
|
|
|
|
|
#endif
|
|
|
|
|
for (j = 0 ; j < ip_proto[i].port[k].ts_no_of_element;
|
|
|
|
|
j++, pp_element = pp_element->next) {
|
|
|
|
|
|
|
|
|
|
if (pp_element->alproto != ip_proto[i].port[k].toserver_element[j].alproto) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->port != ip_proto[i].port[k].toserver_element[j].port) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->alproto_mask != ip_proto[i].port[k].toserver_element[j].alproto_mask) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->min_depth != ip_proto[i].port[k].toserver_element[j].min_depth) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->max_depth != ip_proto[i].port[k].toserver_element[j].max_depth) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
} /* for */
|
|
|
|
|
if (pp_element != NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
pp_element = pp_port->sp;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
#ifdef DEBUG
|
|
|
|
|
dir = 1;
|
|
|
|
|
#endif
|
|
|
|
|
for (j = 0 ; j < ip_proto[i].port[k].tc_no_of_element; j++, pp_element = pp_element->next) {
|
|
|
|
|
if (pp_element->alproto != ip_proto[i].port[k].toclient_element[j].alproto) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->port != ip_proto[i].port[k].toclient_element[j].port) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->alproto_mask != ip_proto[i].port[k].toclient_element[j].alproto_mask) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->min_depth != ip_proto[i].port[k].toclient_element[j].min_depth) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_element->max_depth != ip_proto[i].port[k].toclient_element[j].max_depth) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
} /* for */
|
|
|
|
|
if (pp_element != NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
if (pp_port != NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
if (pp != NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
|
|
result = 1;
|
|
|
|
|
end:
|
|
|
|
|
#ifdef DEBUG
|
|
|
|
|
printf("i = %d, k = %d, j = %d(%s)\n", i, k, j, (dir == 0) ? "ts" : "tc");
|
|
|
|
|
#endif
|
|
|
|
|
return result;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static uint16_t ProbingParserDummyForTesting(uint8_t *input,
|
|
|
|
|
uint32_t input_len,
|
|
|
|
|
uint32_t *offset)
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int AppLayerProtoDetectTest15(void)
|
|
|
|
|
{
|
|
|
|
|
AppLayerProtoDetectUnittestCtxBackup();
|
|
|
|
|
AppLayerProtoDetectSetup();
|
|
|
|
|
|
|
|
|
|
int result = 0;
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"80",
|
|
|
|
|
ALPROTO_HTTP,
|
|
|
|
|
5, 8,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"80",
|
|
|
|
|
ALPROTO_SMB,
|
|
|
|
|
5, 6,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"80",
|
|
|
|
|
ALPROTO_FTP,
|
|
|
|
|
7, 10,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"81",
|
|
|
|
|
ALPROTO_DCERPC,
|
|
|
|
|
9, 10,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"81",
|
|
|
|
|
ALPROTO_FTP,
|
|
|
|
|
7, 15,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"0",
|
|
|
|
|
ALPROTO_SMTP,
|
|
|
|
|
12, 0,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"0",
|
|
|
|
|
ALPROTO_TLS,
|
|
|
|
|
12, 18,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"85",
|
|
|
|
|
ALPROTO_DCERPC,
|
|
|
|
|
9, 10,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"85",
|
|
|
|
|
ALPROTO_FTP,
|
|
|
|
|
7, 15,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
result = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_UDP,
|
|
|
|
|
"85",
|
|
|
|
|
ALPROTO_IMAP,
|
|
|
|
|
12, 23,
|
|
|
|
|
STREAM_TOSERVER,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
|
|
|
|
|
/* toclient */
|
|
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"0",
|
|
|
|
|
ALPROTO_JABBER,
|
|
|
|
|
12, 23,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"0",
|
|
|
|
|
ALPROTO_IRC,
|
|
|
|
|
12, 14,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"85",
|
|
|
|
|
ALPROTO_DCERPC,
|
|
|
|
|
9, 10,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"81",
|
|
|
|
|
ALPROTO_FTP,
|
|
|
|
|
7, 15,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"0",
|
|
|
|
|
ALPROTO_TLS,
|
|
|
|
|
12, 18,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"80",
|
|
|
|
|
ALPROTO_HTTP,
|
|
|
|
|
5, 8,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"81",
|
|
|
|
|
ALPROTO_DCERPC,
|
|
|
|
|
9, 10,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"90",
|
|
|
|
|
ALPROTO_FTP,
|
|
|
|
|
7, 15,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"80",
|
|
|
|
|
ALPROTO_SMB,
|
|
|
|
|
5, 6,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_UDP,
|
|
|
|
|
"85",
|
|
|
|
|
ALPROTO_IMAP,
|
|
|
|
|
12, 23,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"0",
|
|
|
|
|
ALPROTO_SMTP,
|
|
|
|
|
12, 17,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
AppLayerProtoDetectPPRegister(IPPROTO_TCP,
|
|
|
|
|
"80",
|
|
|
|
|
ALPROTO_FTP,
|
|
|
|
|
7, 10,
|
|
|
|
|
STREAM_TOCLIENT,
|
|
|
|
|
ProbingParserDummyForTesting, 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
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_ts_80[] = {
|
|
|
|
|
{ "http", ALPROTO_HTTP, 80, 1 << ALPROTO_HTTP, 5, 8 },
|
|
|
|
|
{ "smb", ALPROTO_SMB, 80, 1 << ALPROTO_SMB, 5, 6 },
|
|
|
|
|
{ "ftp", ALPROTO_FTP, 80, 1 << ALPROTO_FTP, 7, 10 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
};
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_tc_80[] = {
|
|
|
|
|
{ "http", ALPROTO_HTTP, 80, 1 << ALPROTO_HTTP, 5, 8 },
|
|
|
|
|
{ "smb", ALPROTO_SMB, 80, 1 << ALPROTO_SMB, 5, 6 },
|
|
|
|
|
{ "ftp", ALPROTO_FTP, 80, 1 << ALPROTO_FTP, 7, 10 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_ts_81[] = {
|
|
|
|
|
{ "dcerpc", ALPROTO_DCERPC, 81, 1 << ALPROTO_DCERPC, 9, 10 },
|
|
|
|
|
{ "ftp", ALPROTO_FTP, 81, 1 << ALPROTO_FTP, 7, 15 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
};
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_tc_81[] = {
|
|
|
|
|
{ "ftp", ALPROTO_FTP, 81, 1 << ALPROTO_FTP, 7, 15 },
|
|
|
|
|
{ "dcerpc", ALPROTO_DCERPC, 81, 1 << ALPROTO_DCERPC, 9, 10 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_ts_85[] = {
|
|
|
|
|
{ "dcerpc", ALPROTO_DCERPC, 85, 1 << ALPROTO_DCERPC, 9, 10 },
|
|
|
|
|
{ "ftp", ALPROTO_FTP, 85, 1 << ALPROTO_FTP, 7, 15 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
};
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_tc_85[] = {
|
|
|
|
|
{ "dcerpc", ALPROTO_DCERPC, 85, 1 << ALPROTO_DCERPC, 9, 10 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_ts_90[] = {
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
};
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_tc_90[] = {
|
|
|
|
|
{ "ftp", ALPROTO_FTP, 90, 1 << ALPROTO_FTP, 7, 15 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_ts_0[] = {
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 0 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 25 },
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
};
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_tc_0[] = {
|
|
|
|
|
{ "jabber", ALPROTO_JABBER, 0, 1 << ALPROTO_JABBER, 12, 23 },
|
|
|
|
|
{ "irc", ALPROTO_IRC, 0, 1 << ALPROTO_IRC, 12, 14 },
|
|
|
|
|
{ "tls", ALPROTO_TLS, 0, 1 << ALPROTO_TLS, 12, 18 },
|
|
|
|
|
{ "smtp", ALPROTO_SMTP, 0, 1 << ALPROTO_SMTP, 12, 17 }
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_ts_85_udp[] = {
|
|
|
|
|
{ "imap", ALPROTO_IMAP, 85, 1 << ALPROTO_IMAP, 12, 23 },
|
|
|
|
|
};
|
|
|
|
|
AppLayerProtoDetectPPTestDataElement element_tc_85_udp[] = {
|
|
|
|
|
{ "imap", ALPROTO_IMAP, 85, 1 << ALPROTO_IMAP, 12, 23 },
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataPort ports_tcp[] = {
|
|
|
|
|
{ 80,
|
|
|
|
|
((1 << ALPROTO_HTTP) | (1 << ALPROTO_SMB) | (1 << ALPROTO_FTP) |
|
|
|
|
|
(1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
|
|
|
|
|
((1 << ALPROTO_HTTP) | (1 << ALPROTO_SMB) | (1 << ALPROTO_FTP) |
|
|
|
|
|
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
|
|
|
|
|
23,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
element_ts_80, element_tc_80,
|
|
|
|
|
sizeof(element_ts_80) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
sizeof(element_tc_80) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
},
|
|
|
|
|
{ 81,
|
|
|
|
|
((1 << ALPROTO_DCERPC) | (1 << ALPROTO_FTP) |
|
|
|
|
|
(1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
|
|
|
|
|
((1 << ALPROTO_FTP) | (1 << ALPROTO_DCERPC) |
|
|
|
|
|
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
|
|
|
|
|
23,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
element_ts_81, element_tc_81,
|
|
|
|
|
sizeof(element_ts_81) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
sizeof(element_tc_81) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
},
|
|
|
|
|
{ 85,
|
|
|
|
|
((1 << ALPROTO_DCERPC) | (1 << ALPROTO_FTP) |
|
|
|
|
|
(1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
|
|
|
|
|
((1 << ALPROTO_DCERPC) |
|
|
|
|
|
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
|
|
|
|
|
23,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
element_ts_85, element_tc_85,
|
|
|
|
|
sizeof(element_ts_85) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
sizeof(element_tc_85) / sizeof(AppLayerProtoDetectPPTestDataElement)
|
|
|
|
|
},
|
|
|
|
|
{ 90,
|
|
|
|
|
((1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
|
|
|
|
|
((1 << ALPROTO_FTP) |
|
|
|
|
|
(1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
|
|
|
|
|
23,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
element_ts_90, element_tc_90,
|
|
|
|
|
sizeof(element_ts_90) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
sizeof(element_tc_90) / sizeof(AppLayerProtoDetectPPTestDataElement)
|
|
|
|
|
},
|
|
|
|
|
{ 0,
|
|
|
|
|
((1 << ALPROTO_SMTP) | (1 << ALPROTO_TLS) | (1 << ALPROTO_IRC) | (1 << ALPROTO_JABBER)),
|
|
|
|
|
((1 << ALPROTO_JABBER) | (1 << ALPROTO_IRC) | (1 << ALPROTO_TLS) | (1 << ALPROTO_SMTP)),
|
|
|
|
|
23,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
element_ts_0, element_tc_0,
|
|
|
|
|
sizeof(element_ts_0) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
sizeof(element_tc_0) / sizeof(AppLayerProtoDetectPPTestDataElement)
|
|
|
|
|
}
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataPort ports_udp[] = {
|
|
|
|
|
{ 85,
|
|
|
|
|
(1 << ALPROTO_IMAP),
|
|
|
|
|
(1 << ALPROTO_IMAP),
|
|
|
|
|
23,
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
element_ts_85_udp, element_tc_85_udp,
|
|
|
|
|
sizeof(element_ts_85_udp) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
sizeof(element_tc_85_udp) / sizeof(AppLayerProtoDetectPPTestDataElement),
|
|
|
|
|
},
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
AppLayerProtoDetectPPTestDataIPProto ip_proto[] = {
|
|
|
|
|
{ IPPROTO_TCP,
|
|
|
|
|
ports_tcp,
|
|
|
|
|
sizeof(ports_tcp) / sizeof(AppLayerProtoDetectPPTestDataPort),
|
|
|
|
|
},
|
|
|
|
|
{ IPPROTO_UDP,
|
|
|
|
|
ports_udp,
|
|
|
|
|
sizeof(ports_udp) / sizeof(AppLayerProtoDetectPPTestDataPort),
|
|
|
|
|
},
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if (AppLayerProtoDetectPPTestData(alpd_ctx.ctx_pp, ip_proto,
|
|
|
|
|
sizeof(ip_proto) / sizeof(AppLayerProtoDetectPPTestDataIPProto)) == 0) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
result = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
AppLayerProtoDetectDeSetup();
|
|
|
|
|
AppLayerProtoDetectUnittestCtxRestore();
|
|
|
|
|
return result;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
|
|
|
|
|
/** \test test if the engine detect the proto and match with it */
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static int AppLayerProtoDetectTest16(void)
|
|
|
|
|
{
|
|
|
|
|
int result = 0;
|
|
|
|
|
Flow *f = NULL;
|
|
|
|
|
HtpState *http_state = NULL;
|
|
|
|
|
uint8_t http_buf1[] = "POST /one HTTP/1.0\r\n"
|
|
|
|
|
"User-Agent: Mozilla/1.0\r\n"
|
|
|
|
|
"Cookie: hellocatch\r\n\r\n";
|
|
|
|
|
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
|
|
|
|
|
TcpSession ssn;
|
|
|
|
|
Packet *p = NULL;
|
|
|
|
|
Signature *s = NULL;
|
|
|
|
|
ThreadVars tv;
|
|
|
|
|
DetectEngineThreadCtx *det_ctx = NULL;
|
|
|
|
|
DetectEngineCtx *de_ctx = NULL;
|
|
|
|
|
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
|
|
|
|
|
|
|
|
|
|
memset(&tv, 0, sizeof(ThreadVars));
|
|
|
|
|
memset(&ssn, 0, sizeof(TcpSession));
|
|
|
|
|
|
|
|
|
|
p = UTHBuildPacket(NULL, 0, IPPROTO_TCP);
|
|
|
|
|
if (p == NULL) {
|
|
|
|
|
printf("packet setup failed: ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
|
|
|
|
|
if (f == NULL) {
|
|
|
|
|
printf("flow setup failed: ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
f->protoctx = &ssn;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
f->proto = IPPROTO_TCP;
|
|
|
|
|
p->flow = f;
|
|
|
|
|
|
|
|
|
|
p->flowflags |= FLOW_PKT_TOSERVER;
|
|
|
|
|
p->flowflags |= FLOW_PKT_ESTABLISHED;
|
|
|
|
|
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
|
|
|
|
|
|
|
|
|
|
f->alproto = ALPROTO_HTTP;
|
|
|
|
|
|
|
|
|
|
StreamTcpInitConfig(TRUE);
|
|
|
|
|
|
|
|
|
|
de_ctx = DetectEngineCtxInit();
|
|
|
|
|
if (de_ctx == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
de_ctx->flags |= DE_QUIET;
|
|
|
|
|
|
|
|
|
|
s = de_ctx->sig_list = SigInit(de_ctx, "alert http any any -> any any "
|
|
|
|
|
"(msg:\"Test content option\"; "
|
|
|
|
|
"sid:1;)");
|
|
|
|
|
if (s == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SigGroupBuild(de_ctx);
|
|
|
|
|
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
|
|
|
|
|
|
|
|
|
|
FLOWLOCK_WRLOCK(f);
|
|
|
|
|
int r = AppLayerParserParse(NULL, alp_tctx, f, ALPROTO_HTTP,
|
|
|
|
|
STREAM_TOSERVER, http_buf1, http_buf1_len);
|
|
|
|
|
if (r != 0) {
|
|
|
|
|
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
|
|
|
|
|
http_state = f->alstate;
|
|
|
|
|
if (http_state == NULL) {
|
|
|
|
|
printf("no http state: ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* do detect */
|
|
|
|
|
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
|
|
|
|
|
|
|
|
|
|
if (!PacketAlertCheck(p, 1)) {
|
|
|
|
|
printf("sig 1 didn't alert, but it should: ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
result = 1;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alp_tctx != NULL)
|
|
|
|
|
AppLayerParserThreadCtxFree(alp_tctx);
|
|
|
|
|
if (det_ctx != NULL)
|
|
|
|
|
DetectEngineThreadCtxDeinit(&tv, det_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
SigGroupCleanup(de_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
DetectEngineCtxFree(de_ctx);
|
|
|
|
|
|
|
|
|
|
StreamTcpFreeConfig(TRUE);
|
|
|
|
|
|
|
|
|
|
UTHFreePackets(&p, 1);
|
|
|
|
|
UTHFreeFlow(f);
|
|
|
|
|
return result;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** \test test if the engine detect the proto on a non standar port
|
|
|
|
|
* and match with it */
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static int AppLayerProtoDetectTest17(void)
|
|
|
|
|
{
|
|
|
|
|
int result = 0;
|
|
|
|
|
Flow *f = NULL;
|
|
|
|
|
HtpState *http_state = NULL;
|
|
|
|
|
uint8_t http_buf1[] = "POST /one HTTP/1.0\r\n"
|
|
|
|
|
"User-Agent: Mozilla/1.0\r\n"
|
|
|
|
|
"Cookie: hellocatch\r\n\r\n";
|
|
|
|
|
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
|
|
|
|
|
TcpSession ssn;
|
|
|
|
|
Packet *p = NULL;
|
|
|
|
|
Signature *s = NULL;
|
|
|
|
|
ThreadVars tv;
|
|
|
|
|
DetectEngineThreadCtx *det_ctx = NULL;
|
|
|
|
|
DetectEngineCtx *de_ctx = NULL;
|
|
|
|
|
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
|
|
|
|
|
|
|
|
|
|
memset(&tv, 0, sizeof(ThreadVars));
|
|
|
|
|
memset(&ssn, 0, sizeof(TcpSession));
|
|
|
|
|
|
|
|
|
|
p = UTHBuildPacketSrcDstPorts(http_buf1, http_buf1_len, IPPROTO_TCP, 12345, 88);
|
|
|
|
|
|
|
|
|
|
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
|
|
|
|
|
if (f == NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
f->protoctx = &ssn;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
f->proto = IPPROTO_TCP;
|
|
|
|
|
p->flow = f;
|
|
|
|
|
p->flowflags |= FLOW_PKT_TOSERVER;
|
|
|
|
|
p->flowflags |= FLOW_PKT_ESTABLISHED;
|
|
|
|
|
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
|
|
|
|
|
f->alproto = ALPROTO_HTTP;
|
|
|
|
|
|
|
|
|
|
StreamTcpInitConfig(TRUE);
|
|
|
|
|
|
|
|
|
|
de_ctx = DetectEngineCtxInit();
|
|
|
|
|
if (de_ctx == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
de_ctx->flags |= DE_QUIET;
|
|
|
|
|
|
|
|
|
|
s = de_ctx->sig_list = SigInit(de_ctx, "alert http any !80 -> any any "
|
|
|
|
|
"(msg:\"http over non standar port\"; "
|
|
|
|
|
"sid:1;)");
|
|
|
|
|
if (s == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SigGroupBuild(de_ctx);
|
|
|
|
|
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
|
|
|
|
|
|
|
|
|
|
FLOWLOCK_WRLOCK(f);
|
|
|
|
|
int r = AppLayerParserParse(NULL, alp_tctx, f, ALPROTO_HTTP,
|
|
|
|
|
STREAM_TOSERVER, http_buf1, http_buf1_len);
|
|
|
|
|
if (r != 0) {
|
|
|
|
|
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
|
|
|
|
|
http_state = f->alstate;
|
|
|
|
|
if (http_state == NULL) {
|
|
|
|
|
printf("no http state: ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* do detect */
|
|
|
|
|
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
|
|
|
|
|
|
|
|
|
|
if (!PacketAlertCheck(p, 1)) {
|
|
|
|
|
printf("sig 1 didn't alert, but it should: ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
result = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alp_tctx != NULL)
|
|
|
|
|
AppLayerParserThreadCtxFree(alp_tctx);
|
|
|
|
|
if (det_ctx != NULL)
|
|
|
|
|
DetectEngineThreadCtxDeinit(&tv, det_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
SigGroupCleanup(de_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
DetectEngineCtxFree(de_ctx);
|
|
|
|
|
|
|
|
|
|
StreamTcpFreeConfig(TRUE);
|
|
|
|
|
|
|
|
|
|
UTHFreePackets(&p, 1);
|
|
|
|
|
UTHFreeFlow(f);
|
|
|
|
|
return result;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** \test test if the engine detect the proto and doesn't match
|
|
|
|
|
* because the sig expects another proto (ex ftp)*/
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static int AppLayerProtoDetectTest18(void)
|
|
|
|
|
{
|
|
|
|
|
int result = 0;
|
|
|
|
|
Flow *f = NULL;
|
|
|
|
|
HtpState *http_state = NULL;
|
|
|
|
|
uint8_t http_buf1[] = "POST /one HTTP/1.0\r\n"
|
|
|
|
|
"User-Agent: Mozilla/1.0\r\n"
|
|
|
|
|
"Cookie: hellocatch\r\n\r\n";
|
|
|
|
|
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
|
|
|
|
|
TcpSession ssn;
|
|
|
|
|
Packet *p = NULL;
|
|
|
|
|
Signature *s = NULL;
|
|
|
|
|
ThreadVars tv;
|
|
|
|
|
DetectEngineThreadCtx *det_ctx = NULL;
|
|
|
|
|
DetectEngineCtx *de_ctx = NULL;
|
|
|
|
|
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
|
|
|
|
|
|
|
|
|
|
memset(&tv, 0, sizeof(ThreadVars));
|
|
|
|
|
memset(&ssn, 0, sizeof(TcpSession));
|
|
|
|
|
|
|
|
|
|
p = UTHBuildPacket(http_buf1, http_buf1_len, IPPROTO_TCP);
|
|
|
|
|
|
|
|
|
|
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
|
|
|
|
|
if (f == NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
f->protoctx = &ssn;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
f->proto = IPPROTO_TCP;
|
|
|
|
|
p->flow = f;
|
|
|
|
|
p->flowflags |= FLOW_PKT_TOSERVER;
|
|
|
|
|
p->flowflags |= FLOW_PKT_ESTABLISHED;
|
|
|
|
|
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
|
|
|
|
|
f->alproto = ALPROTO_HTTP;
|
|
|
|
|
|
|
|
|
|
StreamTcpInitConfig(TRUE);
|
|
|
|
|
|
|
|
|
|
de_ctx = DetectEngineCtxInit();
|
|
|
|
|
if (de_ctx == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
de_ctx->flags |= DE_QUIET;
|
|
|
|
|
|
|
|
|
|
s = de_ctx->sig_list = SigInit(de_ctx, "alert ftp any any -> any any "
|
|
|
|
|
"(msg:\"Test content option\"; "
|
|
|
|
|
"sid:1;)");
|
|
|
|
|
if (s == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SigGroupBuild(de_ctx);
|
|
|
|
|
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
|
|
|
|
|
|
|
|
|
|
FLOWLOCK_WRLOCK(f);
|
|
|
|
|
int r = AppLayerParserParse(NULL, alp_tctx, f, ALPROTO_HTTP,
|
|
|
|
|
STREAM_TOSERVER, http_buf1, http_buf1_len);
|
|
|
|
|
if (r != 0) {
|
|
|
|
|
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
|
|
|
|
|
http_state = f->alstate;
|
|
|
|
|
if (http_state == NULL) {
|
|
|
|
|
printf("no http state: ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* do detect */
|
|
|
|
|
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
|
|
|
|
|
|
|
|
|
|
if (PacketAlertCheck(p, 1)) {
|
|
|
|
|
printf("sig 1 alerted, but it should not (it's not ftp): ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
result = 1;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alp_tctx != NULL)
|
|
|
|
|
AppLayerParserThreadCtxFree(alp_tctx);
|
|
|
|
|
if (det_ctx != NULL)
|
|
|
|
|
DetectEngineThreadCtxDeinit(&tv, det_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
SigGroupCleanup(de_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
DetectEngineCtxFree(de_ctx);
|
|
|
|
|
|
|
|
|
|
StreamTcpFreeConfig(TRUE);
|
|
|
|
|
|
|
|
|
|
UTHFreePackets(&p, 1);
|
|
|
|
|
UTHFreeFlow(f);
|
|
|
|
|
return result;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** \test test if the engine detect the proto and doesn't match
|
|
|
|
|
* because the packet has another proto (ex ftp) */
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
static int AppLayerProtoDetectTest19(void)
|
|
|
|
|
{
|
|
|
|
|
int result = 0;
|
|
|
|
|
Flow *f = NULL;
|
|
|
|
|
uint8_t http_buf1[] = "MPUT one\r\n";
|
|
|
|
|
uint32_t http_buf1_len = sizeof(http_buf1) - 1;
|
|
|
|
|
TcpSession ssn;
|
|
|
|
|
Packet *p = NULL;
|
|
|
|
|
Signature *s = NULL;
|
|
|
|
|
ThreadVars tv;
|
|
|
|
|
DetectEngineThreadCtx *det_ctx = NULL;
|
|
|
|
|
DetectEngineCtx *de_ctx = NULL;
|
|
|
|
|
AppLayerParserThreadCtx *alp_tctx = AppLayerParserThreadCtxAlloc();
|
|
|
|
|
|
|
|
|
|
memset(&tv, 0, sizeof(ThreadVars));
|
|
|
|
|
memset(&ssn, 0, sizeof(TcpSession));
|
|
|
|
|
|
|
|
|
|
p = UTHBuildPacketSrcDstPorts(http_buf1, http_buf1_len, IPPROTO_TCP, 12345, 88);
|
|
|
|
|
|
|
|
|
|
f = UTHBuildFlow(AF_INET, "1.1.1.1", "2.2.2.2", 1024, 80);
|
|
|
|
|
if (f == NULL)
|
|
|
|
|
goto end;
|
|
|
|
|
f->protoctx = &ssn;
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
f->proto = IPPROTO_TCP;
|
|
|
|
|
p->flow = f;
|
|
|
|
|
p->flowflags |= FLOW_PKT_TOSERVER;
|
|
|
|
|
p->flowflags |= FLOW_PKT_ESTABLISHED;
|
|
|
|
|
p->flags |= PKT_HAS_FLOW|PKT_STREAM_EST;
|
|
|
|
|
f->alproto = ALPROTO_FTP;
|
|
|
|
|
|
|
|
|
|
StreamTcpInitConfig(TRUE);
|
|
|
|
|
|
|
|
|
|
de_ctx = DetectEngineCtxInit();
|
|
|
|
|
if (de_ctx == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
de_ctx->flags |= DE_QUIET;
|
|
|
|
|
|
|
|
|
|
s = de_ctx->sig_list = SigInit(de_ctx, "alert http any !80 -> any any "
|
|
|
|
|
"(msg:\"http over non standar port\"; "
|
|
|
|
|
"sid:1;)");
|
|
|
|
|
if (s == NULL) {
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
SigGroupBuild(de_ctx);
|
|
|
|
|
DetectEngineThreadCtxInit(&tv, (void *)de_ctx, (void *)&det_ctx);
|
|
|
|
|
|
|
|
|
|
FLOWLOCK_WRLOCK(f);
|
|
|
|
|
int r = AppLayerParserParse(NULL, alp_tctx, f, ALPROTO_FTP,
|
|
|
|
|
STREAM_TOSERVER, http_buf1, http_buf1_len);
|
|
|
|
|
if (r != 0) {
|
|
|
|
|
printf("toserver chunk 1 returned %" PRId32 ", expected 0: ", r);
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
FLOWLOCK_UNLOCK(f);
|
|
|
|
|
|
|
|
|
|
/* do detect */
|
|
|
|
|
SigMatchSignatures(&tv, de_ctx, det_ctx, p);
|
|
|
|
|
|
|
|
|
|
if (PacketAlertCheck(p, 1)) {
|
|
|
|
|
printf("sig 1 alerted, but it should not (it's ftp): ");
|
|
|
|
|
goto end;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
result = 1;
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
end:
|
|
|
|
|
if (alp_tctx != NULL)
|
|
|
|
|
AppLayerParserThreadCtxFree(alp_tctx);
|
|
|
|
|
if (det_ctx != NULL)
|
|
|
|
|
DetectEngineThreadCtxDeinit(&tv, det_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
SigGroupCleanup(de_ctx);
|
|
|
|
|
if (de_ctx != NULL)
|
|
|
|
|
DetectEngineCtxFree(de_ctx);
|
|
|
|
|
|
|
|
|
|
StreamTcpFreeConfig(TRUE);
|
|
|
|
|
UTHFreePackets(&p, 1);
|
|
|
|
|
UTHFreeFlow(f);
|
|
|
|
|
return result;
|
|
|
|
|
}
|
|
|
|
|
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app layer, we would set a flow ipproto.
- The "app-layer" section in the yaml conf has also been updated as well.
12 years ago
|
|
|
void AppLayerProtoDetectUnittestsRegister(void)
|
|
|
|
|
{
|
|
|
|
|
SCEnter();
|
|
|
|
|
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest01", AppLayerProtoDetectTest01);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest02", AppLayerProtoDetectTest02);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest03", AppLayerProtoDetectTest03);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest04", AppLayerProtoDetectTest04);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest05", AppLayerProtoDetectTest05);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest06", AppLayerProtoDetectTest06);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest07", AppLayerProtoDetectTest07);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest08", AppLayerProtoDetectTest08);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest09", AppLayerProtoDetectTest09);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest10", AppLayerProtoDetectTest10);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest11", AppLayerProtoDetectTest11);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest12", AppLayerProtoDetectTest12);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest13", AppLayerProtoDetectTest13);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest14", AppLayerProtoDetectTest14);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest15", AppLayerProtoDetectTest15);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest16", AppLayerProtoDetectTest16);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest17", AppLayerProtoDetectTest17);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest18", AppLayerProtoDetectTest18);
|
|
|
|
|
UtRegisterTest("AppLayerProtoDetectTest19", AppLayerProtoDetectTest19);
|
App layer API rewritten. The main files in question are:
app-layer.[ch], app-layer-detect-proto.[ch] and app-layer-parser.[ch].
Things addressed in this commit:
- Brings out a proper separation between protocol detection phase and the
parser phase.
- The dns app layer now is registered such that we don't use "dnstcp" and
"dnsudp" in the rules. A user who previously wrote a rule like this -
"alert dnstcp....." or
"alert dnsudp....."
would now have to use,
alert dns (ipproto:tcp;) or
alert udp (app-layer-protocol:dns;) or
alert ip (ipproto:udp; app-layer-protocol:dns;)
The same rules extend to other another such protocol, dcerpc.
- The app layer parser api now takes in the ipproto while registering
callbacks.
- The app inspection/detection engine also takes an ipproto.
- All app layer parser functions now take direction as STREAM_TOSERVER or
STREAM_TOCLIENT, as opposed to 0 or 1, which was taken by some of the
functions.
- FlowInitialize() and FlowRecycle() now resets proto to 0. This is
needed by unittests, which would try to clean the flow, and that would
call the api, AppLayerParserCleanupParserState(), which would try to
clean the app state, but the app layer now needs an ipproto to figure
out which api to internally call to clean the state, and if the ipproto
is 0, it would return without trying to clean the state.
- A lot of unittests are now updated where if they are using a flow and
they need to use the app 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 */
|