This currently supports automatic decapsulation of GTP-U packets on
UDP port 2152.
The GTPv1 headers for such tunnels can be inspected by handling the
"gtpv1_g_pdu_packet" event, which has a parameter of type "gtpv1_hdr".
Analyzer and test cases are derived from submissions by Carsten Langer.
Addresses #690.
This option indicates that the Teredo analyzer should wait until
it sees both sides of a connection using a valid Teredo encapsulation
before issuing a protocol_confirmation. Previous behavior confirmed
on the first instance of a valid encapsulation, which could result
in more false positives (and e.g. bogus entries in known-services.log).
Addresses #890.
Authentication Headers with a Payload Len field set to zero would cause
a crash due to invalid memory allocation because the previous code
assumed Payload Len would always be great enough to contain all
mandatory fields of the header. This changes it so the length of
the header is explicitly checked before attempting to extract fields
located past the minimum length (8 bytes) of an Authentication Header.
Crashes due to this are only possible when handling script-layer events
ipv6_ext_headers, new_packet, esp_packet, or teredo_*. Or also when
implementing one of the discarder_check_* family of functions.
Otherwise, Bro correctly parses past such a header.
Otherwise functions could be called with a mismatching argument list
and cause a crash at run-time. The incorrect function type is now
reported at parse-time.
If the metrics break interval happened to occur between first and second
worker starting up and getting connected to the cluster, the test
would fail because the second worker didn't get a chance to connect and
send data. The test now waits for the cluster setup to complete
before workers send metrics data.
- Some baselines for tests in "leaks" group were outdated.
- Changed a few of the cluster/communication tests to terminate
more explicitly instead of relying on btest-bg-wait to kill
processes. This makes the tests finish faster in the success case
and makes the reason for failing clearer in the that case.
Turns out the finish methods weren't called correctly, caused by a
mess up with method names which all sounded too similar and the wrong
one ended up being called. I've reworked this by changing the
thread/writer/reader interfaces, which actually also simplifies them
by getting rid of the requirement for writer backends to call their
parent methods (i.e., less opportunity for errors).
This commit also includes the following (because I noticed the problem
above when working on some of these):
- The ASCII log writer now includes "#start <timestamp>" and
"#end <timestamp> lines in the each file. The latter supersedes
Bernhard's "EOF" patch.
This required a number of tests updates. The standard canonifier
removes the timestamps, but some tests compare files directly,
which doesn't work if they aren't printing out the same
timestamps (like the comm tests).
- The above required yet another change to the writer API to
network_time to methods.
- Renamed ASCII logger "header" options to "meta".
- Fixes#763 "Escape # when first character in log file line".
All btests pass for me on Linux FC15. Will try MacOS next.
Also removed RefExpr::Eval(Val*) method since it was never called
(Clang emitted warning about this hiding overloaded virtual function
UnaryExpr::Eval(Frame*)) and doesn't appear to be necessary even if it
was called to avoid the default vector handling of UnaryExpr::Eval
(as the comment suggests as the intention).
This adds a new parameter of type "icmp6_nd_options" to the ICMPv6
neighbor discovery events (icmp_redirect, icmp_router_solicitation,
icmp_router_advertisement, icmp_neighbor_solicitation,
icmp_neighbor_advertisement) which includes data extracted from
all neighbor discovery options (RFC 4861) that are present in the
ICMPv6 message.
Teredo bubble packets (IPv6 w/ No Next Header and zero Payload Length)
with data extending past the inner IPv6 header (the outer IPv4 header's
Total Length and UDP header's Length indicate this) now only raises
a "Teredo_payload_len_mismatch" weird instead of causing a
ProtocolViolation().
This also fixes a crash in NetSessions::ParseIPPacket() that occurred
when the packet length didn't match the payload length field.
If some expression in an event handler body causes an
InterpreterException internally, then the rest of that body doesn't
get executed, but also the bodies of any other handlers were not
executed.
The Tunnel::yielding_teredo_decapsulation (on by default) makes it so
the Teredo analyzer doesn't attempt to decapsulate payloads when
there's already a sibling analyzer that thinks it's parsing the right
protocol. Sometimes, UDP payloads just happen to look like they are
validly Teredo-encapsulated and doing further analysis on the
decapsulated packet can quickly turn into a weird; this change helps
reduce such weirds.
These are called "teredo_packet", "teredo_authentication",
"teredo_origin_indication", and "teredo_bubble" and may be raised
on a per-packet basis depending on Teredo encapsulation method.
The ICMP/ICMPv6 analyzers function correctly when full packets have
not been captured, but everything up to and including the ICMP header
is there (e.g. the functions that inspect ICMP error message context
correctly check the caplen to see if more info can be extracted).
The "Should have been caught earlier already." comment may have referred
to NetSessions::CheckHeaderTrunc, which works as intended to catch cases
where the ICMP header is not there in full, but then the assert was
still not correctly formulated for that...
Also changed the ICMP checksum calculation to not occur when the full
packet has not been captured, which seems consistent with what the UDP
analysis does.
Without this change, flow labeling of connections over IPv6 are
only available in the per-packet types of events (e.g. new_packet)
in which header fields can be inspected, but now minimal tracking
of the most recent flow label is done internally and that's available
per-connection for all events that use connection record arguments.
Specifically, this adds a "flow_label" field to the "endpoint" record
type, which is used for both the "orig" and "resp" fields of
"connection" records. The new "connection_flow_label_changed" event
also allows tracking of changes in flow labels: it's raised each time
one direction of the connection starts using a different label.
* origin/topic/robin/dataseries:
Moving trace for rotation test into traces directory.
Fixing a rotation race condition at termination.
Portability fixes.
Extending DS docs with some examples.
Updating doc.
Fixing pack_scale and time-as-int.
Adding format specifier to DS spec to print out double as %.6f.
DataSeries updates and fixes.
DataSeries tuning.
Tweaking DataSeries support.
Extending log post-processor call to include the name of the writer.
Removing an unnecessary const cast.
DataSeries TODO list with open issues/questions.
Starting DataSeries HowTo.
Additional test output canonification for ds2txt's timestamps.
In threads, an internal error now immediately aborts.
DataSeries cleanup.
Working on DataSeries support.
Merging in DataSeries support from topic/gilbert/logging.
Fixing threads' DoFinish() method.
- Unserializing files that were previously kicked out of the open-file
cache would cause them to be fopen'd with the original access
permissions which is usually 'w' and causes truncation. They
are now opened in 'a' mode. (addresses #780)
- Add 'max_files_in_cache' script option to manually set the maximum
amount of opened files to keep cached. Mainly this just helped
to create a simple test case for the above change.
- Remove unused NO_HAVE_SETRLIMIT preprocessor switch.
- On systems that don't enforce a limit on number of files opened for
the process, raise default max size of open-file cache from
32 to 512.
conn.log now sets a field indicating all the parent tunnel UIDs over
which a connection operated and cross reference the UIDs found in
the tunnel.log.
Also some renaming of tunnel related types at the scripting layer.
UDP tunnel support removed for now, to be re-added in specific
analyzers later, but IP-in-IP is now decapsulated recursively
so nested tunnels can be seen and the inner packets get sent
through the IP fragment reassembler if necessary.
- The 'icmp_conn' record now contains an 'hlim' field since hop limit
in the IP header is an interesting field for at least these ND
messages.
- Changed 'icmp_router_advertisement' event parameters.
'router_lifetime' is now an interval. Fix 'reachable_time' and
'retrans_timer' using wrong internal Val type for intervals.
Made more of the known router advertisement flags available through
boolean parameters.
- Changed 'icmp_neighbor_advertisement' event parameters to add
more of the known boolean flags.
* origin/topic/icmp6:
Fixes for IPv6 truncation and ICMP/ICMP6 analysis.
Change ICMPv6 checksum calculation to use IP_Hdr wrapper.
Update IPv6 atomic fragment unit test to filter output of ICMPv6.
Add more data to icmp events
More code cleanup
Add more icmpv6 events, and general code cleanup
Fix compile failure after merge from master
Significant edit pass over ICMPv6 code.
Porting Matti's branch to git.
Closes#808.
- Add more guards against trying to analyze captured packets with a
truncated IPv6 static header or extension header chain.
- Add back in the ICMP payload tracking for ICMP "connections".
- Fix 'icmp_context' record construction. Some field assignments
were mismatched for ICMP and ICMP6. Source and destination
addresses were set incorrectly for context packets that don't
contain a full IP header. Some fields for ICMP6 weren't filled out.
- Changed ICMP Time Exceeded packets to raise the 'icmp_time_exceeded'
event instead of 'icmp_error_message'.
- Add unit tests for truncation and the main types of ICMP/ICMP6
that have specific events.
- Documentation clarifications.
So that src/dst addresses used in pseudo-header are correct when
there's certain extension headers (routing/destination).
Add ICMP/ICMPv6 checksum unit tests.
