Tradeoff between key comparison efficiency (i.e. no `memcmp`) and key
length.
At least in the ipv4 cases, we have no more different entries with the
same key.
* Normalization of host_server_name
The ndpi_hostname_sni_set() function replaces all non-printable
characters with the "?" character and removing whitespace characters
at the end of the line.
* Added conditional hostname normalization.
When switching to (D)TLS dissector from the STUN one, we need to clear
any flow risks set from the latter (because we don't have anymore
`NDPI_PROTOCOL_STUN` in the classification results)
Keep looking for RTP packets but remove the monitoring concept.
We will re-introduce a more general concept of "flow in monitoring
state" later.
The function was disabled by default.
Some configuration knobs will be provided when/if #2190 is merged.
Keep demultiplexing STUN/RTP/RTCP packets after DTLS ones.
We might end up processing the session a little longer, because we will
process the STUN/RTP/RTCP packets after the DTLS handshake.
1) Public API/headers in `src/include/` [as it has always been]
2) Private API/headers in `src/lib/`
Try to keep the "ndpi_" prefix only for the public functions
Try to have a faster classification, on first packet; use standard extra
dissection data path for sub-classification, metadata extraction and
monitoring.
STUN caches:
* use the proper confidence value
* lookup into the caches only once per flow, after having found a proper
STUN classification
Add identification of Telegram VoIP calls.
Regardless of the name, the removed trace doesn't contain meaningful
Hangout traffic.
Remove last piece of sub-classifiction based only on ip addresses.
For a lot of protocols, reduce the number of packets after which the
protocols dissector gives up.
The values are quite arbitary, tring to not impact on classification
Remove `FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION` define from
`fuzz/Makefile.am`; it is already included by the main configure script
(when fuzzing).
Add a knob to force disabling of AESNI optimizations: this way we can
fuzz also no-aesni crypto code.
Move CRC32 algorithm into the library.
Add some fake traces to extend fuzzing coverage. Note that these traces
are hand-made (via scapy/curl) and must not be used as "proof" that the
dissectors are really able to identify this kind of traffic.
Some small updates to some dissectors:
CSGO: remove a wrong rule (never triggered, BTW). Any UDP packet starting
with "VS01" will be classified as STEAM (see steam.c around line 111).
Googling it, it seems right so.
XBOX: XBOX only analyses UDP flows while HTTP only TCP ones; therefore
that condition is false.
RTP, STUN: removed useless "break"s
Zattoo: `flow->zattoo_stage` is never set to any values greater or equal
to 5, so these checks are never true.
PPStream: `flow->l4.udp.ppstream_stage` is never read. Delete it.
TeamSpeak: we check for `flow->packet_counter == 3` just above, so the
following check `flow->packet_counter >= 3` is always false.
All dissector callbacks should not be exported by the library; make static
some other local functions.
The callback logic in `ndpiReader` has never been used.
With internal libgcrypt, `gcry_control()` should always return no
errors.
We can check `categories` length at compilation time.
The application may enable only some protocols.
Disabling a protocol means:
*) don't register/use the protocol dissector code (if any)
*) disable classification by-port for such a protocol
*) disable string matchings for domains/certificates involving this protocol
*) disable subprotocol registration (if any)
This feature can be tested with `ndpiReader -B list_of_protocols_to_disable`.
Custom protocols are always enabled.
Technically speaking, this commit doesn't introduce any API/ABI
incompatibility. However, calling `ndpi_set_protocol_detection_bitmask2()`
is now mandatory, just after having called `ndpi_init_detection_module()`.
Most of the diffs (and all the diffs in `/src/lib/protocols/`) are due to
the removing of some function parameters.
Fix the low level macro `NDPI_LOG`. This issue hasn't been detected
sooner simply because almost all the code uses only the helpers `NDPI_LOG_*`
Simplest solution, keeping the existing cache data structure
TLS certificate cache is used for DTLS traffic, too.
Note that Ookla cache already works with ipv6 flows.
TODO:
* make the key/hashing more robust (extending the key size?)
* update bittorrent cache too. That task is quite difficult because
ntopng uses a public function (`ndpi_guess_undetected_protocol()`)
intrinsically ipv4 only...
LRU callbacks have been added in 460ff3c7a, but they have never been
used and they have never been extended to the other LRU caches.
`ndpi_search_tcp_or_udp()` basically returns the classification by port/ip
of the flow; calling it from the dissector is useless.
The same for TOR detection: ips are checked in the generic code
0 as size value disable the cache.
The diffs in unit tests are due to the fact that some lookups are
performed before the first insert: before this change these lookups
weren't counted because the cache was not yet initialized, now they are.
Basically:
* "classification by-ip" (i.e. `flow->guessed_protocol_id_by_ip` is
NEVER returned in the protocol stack (i.e.
`flow->detected_protocol_stack[]`);
* if the application is interested into such information, it can access
`ndpi_protocol->protocol_by_ip` itself.
There are mainly 4 points in the code that set the "classification
by-ip" in the protocol stack: the generic `ndpi_set_detected_protocol()`/
`ndpi_detection_giveup()` functions and the HTTP/STUN dissectors.
In the unit tests output, a print about `ndpi_protocol->protocol_by_ip`
has been added for each flow: the huge diff of this commit is mainly due
to that.
Strictly speaking, this change is NOT an API/ABI breakage, but there are
important differences in the classification results. For examples:
* TLS flows without the initial handshake (or without a matching
SNI/certificate) are simply classified as `TLS`;
* similar for HTTP or QUIC flows;
* DNS flows without a matching request domain are simply classified as
`DNS`; we don't have `DNS/Google` anymore just because the server is
8.8.8.8 (that was an outrageous behaviour...);
* flows previusoly classified only "by-ip" are now classified as
`NDPI_PROTOCOL_UNKNOWN`.
See #1425 for other examples of why adding the "classification by-ip" in
the protocol stack is a bad idea.
Please, note that IPV6 is not supported :( (long standing issue in nDPI) i.e.
`ndpi_protocol->protocol_by_ip` wil be always `NDPI_PROTOCOL_UNKNOWN` for
IPv6 flows.
Define `NDPI_CONFIDENCE_MATCH_BY_IP` has been removed.
Close#1687
The field `flow->guessed_host_protocol_id` is set at the beginning of
the flow analysis and it represents the "classification by ip" of the flow
itself.
This field should never be changed. Dissectors which want to provide an
"hint" about the classification, should update `flow->guessed_protocol_id`
instead. Such "hint" is useless if the dissector set the "extra-dissection"
data-path.
Rename such field to `guessed_protocol_id_by_ip` to better describe its
role.
Preliminary work necessary for #1687
Add detection over TCP and fix detection over IPv6.
Rename some variables since Stun dissector is no more "udp-centric".
Stun dissector should always classified the flow as `STUN` or
`STUN/Something`.
Don't touch `flow->guessed_host_protocol_id` field, which should be
always be related to "ip-classification" only.