| Commit message (Collapse) | Author | Age |
... | |
| |
|
| |
|
|
|
|
|
|
| |
"Invalid DNS Header"-risk should be set only if the flow has been
already classified as DNS. Otherwise, almost any non-DNS flows on port 53
will end up having the `NDPI_MALFORMED_PACKET` risk set, which is a little
bit confusing for non DNS traffic
|
|
|
|
|
|
|
|
|
|
|
|
| |
Based on the paper: "OpenVPN is Open to VPN Fingerprinting"
See: https://www.usenix.org/conference/usenixsecurity22/presentation/xue-diwen
Basic idea:
* the distribution of the first byte of the messages (i.e. the distribution
of the op-codes) is quite unique
* this fingerprint might be still detectable even if the OpenVPN packets are
somehow fully encrypted/obfuscated
The heuristic is disabled by default.
|
| |
|
| |
|
|
|
|
| |
We can do definitely better, but this change is a big improvements
respect the current broken code
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Allow sub-classification of OpenVPN/Wireguard flows using their server IP.
That is useful to detect the specific VPN application/app used.
At the moment, the supported protocols are: Mullvad, NordVPN, ProtonVPN.
This feature is configurable.
|
| |
|
| |
|
| |
|
|
|
|
|
| |
(#2541)
Signed-off-by: Toni Uhlig <matzeton@googlemail.com>
|
| |
|
|
|
|
|
|
|
| |
On extra-dissection data-path we only need to look for the hash (the
flow is already classified as Bittorrent).
As a nice side-effect, the confidence is now always with the right
value.
|
| |
|
| |
|
|
|
| |
The Train Real Time Data Protocol (TRDP) is a UDP/TCP-based communication protocol designed for IP networks in trains, enabling data exchange between devices such as door controls and air conditioning systems. It is standardized by the IEC under IEC 61375-2-3 and is not related to the Remote Desktop Protocol (RDP).
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
See also #2523
---------
Co-authored-by: Nardi Ivan <nardi.ivan@gmail.com>
|
|
|
| |
ISO/IEC 14908-4 defines how to tunnel Control Network Protocol (CNP) over IP networks. It encapsulates protocols like EIA-709, EIA-600, and CNP, making it a versatile solution for building automation and control systems.
|
| |
|
| |
|
|
|
|
|
| |
The `suffix_id` is simply an incremental index (see
`ndpi_load_domain_suffixes`), so its value might changes every time we
update the public suffix list.
|
| |
|
|
|
|
| |
If the flow is classified (via DPI) after the first packet, we should
use this information as FPC
|
| |
|
|
|
|
| |
Add printing of fpc_dns statistics and add a general cconfiguration option.
Rework the code to be more generic and ready to handle other logics.
|
|
|
|
|
|
|
|
|
| |
Use DNS information to get a better First Packet Classification.
See: #2322
---------
Co-authored-by: Nardi Ivan <nardi.ivan@gmail.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Reformatted shell scripts according to [ShellCheck](https://github.com/koalaman/shellcheck/).
I. Most common changes:
1. https://github.com/koalaman/shellcheck/wiki/SC2086
`$var` → `"$var"`
Note: this isn't always necessary and I've been careful not to substitute where it wasn't necessary in meaning.
2. https://github.com/koalaman/shellcheck/wiki/SC2006
`` `command` `` → `$(command)`
3. https://github.com/koalaman/shellcheck/wiki/SC2004
`$(( $a + $b ))` → `$(( a + b ))`
4. https://github.com/koalaman/shellcheck/wiki/SC2164
`cd "$dir"` → `cd "$dir" || exit 1`
5. https://github.com/koalaman/shellcheck/wiki/SC2166
`[ check1 -o check2 ]` → `[ check1 ] || [ check2 ]`
6. https://github.com/koalaman/shellcheck/wiki/SC2002
`cat "${file}" | wc -c` → `< "${file}" wc -c`
Note: this looks a bit uglier but works faster.
II. Some special changes:
1. In file `utils/common.sh`:
https://github.com/koalaman/shellcheck/wiki/SC2112
This script is interpreted by `sh`, not by `bash`, but uses the keyword `function`.
So I replaced `#!/usr/bin/env sh` to `#!/usr/bin/env bash`.
2. After that I thought of replacing all shebangs to `#!/usr/bin/env bash` for consistency and cross-platform compatibility, especially since most of the files already use bash.
3. But in cases when it was `#!/bin/sh -e` or `#!/bin/bash -eu` another problem appears:
https://github.com/koalaman/shellcheck/wiki/SC2096
So I decided to make all shebangs look uniform:
```
#!/usr/bin/env bash
set -e (or set -eu) (if needed)
```
4. In file `tests/ossfuzz.sh`:
https://github.com/koalaman/shellcheck/wiki/SC2162
`read i` → `read -r i`
Note: I think that there is no need in special treatment for backslashes, but I could be wrong.
5. In file `tests/do.sh.in`:
https://github.com/koalaman/shellcheck/wiki/SC2035
`ls *.*cap*` → `ls -- *.*cap*`
6. In file `utils/verify_dist_tarball.sh`:
https://github.com/koalaman/shellcheck/wiki/SC2268
`[ "x${TARBALL}" = x ]` → `[ -z "${TARBALL}" ]`
7. In file `utils/check_symbols.sh`:
https://github.com/koalaman/shellcheck/wiki/SC2221
`'[ndpi_utils.o]'|'[ndpi_memory.o]'|'[roaring.o]')` → `'[ndpi_utils.o]'|'[ndpi_memory.o]')`
8. In file `autogen.sh`:
https://github.com/koalaman/shellcheck/wiki/SC2145
`echo "./configure $@"` → `echo "./configure $*"`
https://github.com/koalaman/shellcheck/wiki/SC2068
`./configure $@` → `./configure "$@"`
III. `LIST6_MERGED` and `LIST_MERGED6`
There were typos with this variables in files `utils/aws_ip_addresses_download.sh`, `utils/aws_ip_addresses_download.sh` and `utils/microsoft_ip_addresses_download.sh` where variable `LIST6_MERGED` was defined, but `LIST_MERGED6` was removed by `rm`.
I changed all `LIST_MERGED6` to `LIST6_MERGED`.
Not all changes are absolutely necessary, but some may save you from future bugs.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix detection of WebDAV and Gnutella (over HTTP)
Fix detection of z3950
Add two fuzzers to test `ndpi_memmem()` and `ndpi_strnstr()`
Remove some dead code:
* RTP: the same exact check is performed at the very beginning of the
function
* MQTT: use a better helper to exclude the protocol
* Colletd: `ndpi_hostname_sni_set()` never fails
Update pl7m code (fix a Use-of-uninitialized-value error)
|
|
|
| |
See: #2484
|
|
|
|
|
| |
Let's start with some basic helpers and with FPC based on flow addresses.
See: #2322
|
|
|
|
|
|
| |
To avoid the following error with some old libpcap versions:
```
ERROR: could not open pcap file: invalid file capture length 524288, bigger than maximum of 262144
```
|
|
|
| |
See #2484
|
| |
|
|
|
|
|
|
| |
Remove some code never triggered
AFP: the removed check is included in the following one
MQTT: fix flags extraction
|
| |
|
|
|
| |
Signed-off-by: Toni Uhlig <matzeton@googlemail.com>
|
| |
|
|
|
|
|
| |
Since 070a0908b we are able to detect P2P calls directly from the packet
content, without any correlation among flows
|