Wednesday, April 6, 2011

2.9.0.5 is available for download!

Now available for download from the link here, 2.9.0.5 brings many improvements to Snort in terms of bug fixes.  Below is a cut and paste from the Changelog.


2011-03-23 Steven Sturges <ssturges@sourcefire.com>
  * src/build.h:
      Increment Snort build number to 134
  * src/: decode.h, encode.c:
  * src/dynamic-plugins/sf_engine/: sf_snort_packet.h:
  * src/preprocessors/: spp_sfportscan.c, spp_frag3.c:
  * src/output-plugins/: spo_alert_fast.c:
  * src/preprocessors/Stream5/: stream5_common.c:
      Updated portscan to set protocol correctly in raw packet for
      IPv6 and changed the encoder to recognize portscan packets as pseudo
 packets so that the checksum isn't calculated
  * src/: sfdaq.c, util.c:
      Improve handling of DAQ failure codes when Snort is shutting down.
  * src/preprocessors/spp_perfmonitor.c:
      Update perfmonitor to create now files prior to dropping privs

2011-03-16 Ryan Jordan <ryan.jordan@sourcefire.com>
Snort 2.9.0.5
  * src/build.h:
      Increment Snort build number to 132
  * src/snort.c:
  * src/preprocessors/: normalize.c, perf-base.c, perf-base.h,
    Stream5/snort_stream5_tcp.c:
      TCP timestamp options are only NOPed by the Normalization preprocessor
      if Stream5 has seen a full 3-way handshake, and timestamps weren't
      negotiated.

      The IPS mode reassembly policy has been refactored to do stream
      normalization within the first policy.

      Packets injected by the normalization preprocessor are now counted
      in the packet statistics.
  * doc/snort_manual.tex:
  * src/: parser.c, parser.h:
  * src/preprocessors/: spp_frag3.c, Stream5/snort_stream5_session.c:
      Added a "config vlan_agnostic" setting that globally disables Stream's
      use of vlan tag in session tracking.
  * src/: snort.c, preprocessors/normalize.c,
    preprocessors/spp_normalize.c, preprocessors/spp_normalize.h,
    preprocessors/perf-base.c, preprocessors/perf-base.h:
  * doc/: README.normalize, snort_manual.pdf, snort_manual.tex:
      Fixed the normalization preprocessor to call its post-initialization
      config functions during a policy reload.

      Packets can no longer be trimmed below the minimum ethernet frame
      length. Trimming is now configurable with the "normalize_ip4: trim;"
      option. TOS clearing is now configurable with "normalize_ip4: tos;".

      The "normalize_ip4: trim" option is automatically disabled if the
      DAQ can't inject packets. If the DAQ tries and fails to inject
      a given packet, the wire packet is not blocked.

      Updated documentation regarding these changes.
  * src/detection-plugins/sp_cvs.c:
      Fixed a false positive in the CVS detection plugin. It was incorrectly
      parsing CVS entries that had a '+' in between the 3rd and 4th slashes.
  * src/preprocessors/HttpInspect/: client/hi_client.c,
    server/hi_server.c:
      Changed a pointer comparison to a size check for code readability.
      Belated thanks to Dwane Atkins and Parker Crook for reporting a
      related issue that was fixed in Snort 2.9.0.4 build 111.

      Moved the zlib initialization such that gzipped responses are still
      inspected if the zipped data starts after the first Stream-reassembled
      packet is inspected.
  * src/decode.c:
      Fixed an issue with decoding too many IP layers in a single packet. The
      Teredo proto bit was not unset after hitting the limit on IP layers.
      Thanks to Dwane Atkins for reporting this issue.

      IPv6 fragmented packets are no longer inspected unless they have an
      offset of zero and the next layer is UDP. This behavior is consistent
      with IPv4 decoding.
      Thanks to Martin Schütte for reporting an issue where fragged ICMPv6
      packets were being inspected.

      The decoder no longer attempts to decode Teredo packets inside of
      IPv4 fragments, instead waiting for the reassembled packet.
  * src/encode.c:
      Fixed a problem where encoded packets had their lengths calculated
      incorrectly. This caused the active response feature to generate
      incorrect RST packets if the original packet had a VLAN tag.
  * preproc_rules/preprocessor.rules:
      Updated references to rule 125:1:1
  * src/preprocessors/spp_perfmonitor.c:
      Perfmonitor files are now created after Snort changes uid/gid.
  * src/dynamic-plugins/sf_preproc_example/sf_dynamic_preproc_lib.c:
      Fixed the size formatting of an error message argument when
      compiling with --enable-rzb-saac.
      Thanks to Cleber S. Brandão for reporting this issue.
  * etc/snort.conf:
      Updated the default snort.conf with max compress and decompress
      depths to enable unlimited decompression of gzipped HTTP responses.
  * snort.8:
      Fixed the man page's URL regarding the location of Snort rules.
      Thanks to Michael Scheidell for reporting an out-of-date man page section.
  * doc/README.http_inspect, doc/snort_manual.tex,
    src/preprocessors/snort_httpinspect.c:
      HTTP Inspect's "unlimited_decompress" option now requires that
      "compress_depth" and "decompress_depth" are set to their max values.
  * src/: fpcreate.c, dynamic-plugins/sf_dynamic_define.h,
    dynamic-plugins/sf_dynamic_engine.h,
    preprocessors/Stream5/snort_stream5_tcp.c:
      Fixed an error that prevented compiling with --disable-dynamicplugin.
      Thanks to Jason Wallace for reporting this issue.
  * src/dynamic-preprocessors/ftptelnet/: snort_ftptelnet.c,
    snort_ftptelnet.h, spp_ftptelnet.c:
      Changed the names of ProcessGlobalConf() and PrintGlobalConf() inside
      the ftp_telnet preprocessor to avoid a naming conflict with similar
      functions in HTTP Inspect.
      Thanks to Bruce Corwin for reporting this issue.
  * src/preprocessors/: perf.c, perf-base.c, perf-base.h, perf-flow.c,
    perf-flow.h:
      Fixed comparisons between signed and unsigned int, which lead to
      a faulty length check.
      Thanks to Cihan Ayyildiz and Jason Wallace for helping us debug this
      issue.
 Please upgrade your version of Snort!

No comments:

Post a Comment