|
Nmap Development
mailing list archives
Nmap 5.10BETA1 released
From: Fyodor <fyodor () insecure org>
Date: Mon, 23 Nov 2009 00:33:07 -0800
Hi folks. I'm happy to announce our first post-5.00 release! It
contains the results of a full five months of work, since the 5.00
release candidate was branched off in June. Good work, everybody!
Changes include:
o 14 new NSE scripts, bringing the total to 72!
o A brand new (much faster and more efficient) --traceroute system
o The Zenmap host filter (Ctrl-L) for drilling down to the exact
systems you want to see in a big scan based on the criteria you
specify. For example, type "apache" to see all the hosts running a
version of the Apache web server.
o UDP protocol-specific payloads make UDP scanning and host discovery
much more effective.
o And more than 100 other significant changes!
With such a massive release, I wouldn't be surprised to find some bugs
lurking about. Please give this some serious testing, as I'd like to
turn it into a stable release in December. If you find any problems
or have questions, see the instructions at
http://nmap.org/book/man-bugs.html. If the problems are serious, I'll
build and release 5.10BETA2.
You can download Nmap 5.10BETA1 from the normal location:
http://nmap.org/download.html
Now here is the giant list of changes in this release (I've tried to
put the most important ones near the top):
Nmap 5.10BETA1 [2009-11-23]
o Added 14 new NSE scripts for a grand total of 72! You can learn
about them all at http://nmap.org/nsedoc/. Here are the new ones:
o smb-psexec implements remote process execution similar to the
Sysinternals' psexec tool (or Metasploit's psexec "exploit"),
allowing a user to run a series of programs on a remote machine
and read the output. This is great for gathering information about
servers, running the same tool on a range of system, or even
installing a backdoor on a collection of computers. See
http://nmap.org/nsedoc/scripts/smb-psexec.html [Ron]
o dhcp-discover sends out DHCP probes on UDP/67 and displays all
interesting results (or, with verbosity, all results).
Optionally, multiple probes can be sent and the MAC address can be
randomized in an attempt to exhaust the DHCP server's address pool
and potentially create a denial of service condition. See
http://nmap.org/nsedoc/scripts/dhcp-discover.html. [Ron]
o http-enum enumerates URLs used by popular web applications and
servers and reports which ones exist on a target web server. See
http://nmap.org/nsedoc/scripts/http-enum.html. [Ron, Andrew Orr,
Rob Nicholls]
o ssl-cert retrieves and prints a target server's SSL
certificate. See
http://nmap.org/nsedoc/scripts/ssl-cert.html. [David]
o x11-access checks whether access to an X11 server is allowed (as
with "xhost +" for example). See
http://nmap.org/nsedoc/scripts/x11-access.html. [jlanthea]
o db2-info enhances DB2 database instance detection. It provides
detection when version probes fail, but will default to the
version detection probe value if that is more precise. It also
detects the server platform and database instance name. The DB2
version detection port ranges were broadened to 50000-50025 and
60000-60025 as well. See
http://nmap.org/nsedoc/scripts/db2-info.html. [Tom]
o smbv2-enabled checks if the smbv2 protocol is enabled on target
servers. SMBv2 has already suffered from at least one major
security vulnerability. See
http://nmap.org/nsedoc/scripts/smbv2-enabled.html. [Ron]
o http-favicon obtains the favicon file (/favicon.ico or whatever is
specified by the HTML link tag) and tries to identify its source
(such as a certain web application) using a database lookup. See
http://nmap.org/nsedoc/scripts/http-favicon.html. [Vladz]
o http-date obtains the Date: header field value from an HTTP server
then displays it along with how much it differs from local
time. See http://nmap.org/nsedoc/scripts/http-date.html. [David]
o http-userdir-enum attempts to enumerate users on a system by
trying URLs with common usernames in the Apache mod_userdir format
(e.g. http://target-server.com/~john). See
http://nmap.org/nsedoc/scripts/http-userdir-enum.html. [Jah]
o pjl-ready-message allows viewing and setting the status message on
printers which support the Printer Job Language (many HP printers
do). See http://nmap.org/nsedoc/scripts/pjl-ready-message.html.
[Aaron Leininger]
o http-headers performs a GET request for the root folder ("/") of a
web server and displays the HTTP headers returned. See
http://nmap.org/nsedoc/scripts/http-headers.html. [Ron]
o http-malware-host is designed to discover hosts that are serving
malware (perhaps because they were compromised), but so far it
only checks for one specific attack. See
http://nmap.org/nsedoc/scripts/http-malware-host.html. [Ron]
o smb-enum-groups displays a list of groups on the remote system
along with their membership (like enum.exe -G). See
http://nmap.org/nsedoc/scripts/smb-enum-users.html [Ron]
o Nmap's --traceroute has been rewritten for better performance.
Probes are sent in parallel to individual hosts, not just across all
hosts as before. Trace consolidation is more sophisticated, allowing
common traces to be identified sooner and fewer probes to be sent.
The older traceroute could be very slow (taking minutes per target)
if the target did not respond to the trace probes, and this new
traceroute avoids that. In a trace of 110 hosts in a /24 over the
Internet, the number of probes sent dropped 50% from 1565 to 743,
and the time taken dropped 92% from 95 seconds to 7.6
seconds. Traceroute now uses an ICMP echo request probe if no
working probes against the target were discovered during
scanning. [David]
o [Zenmap] After performing or loading a scan, you can now filter
results to just the hosts you are interested in by pressing Ctrl+L
(or the "Filter Hosts" button) to open the host filtering interface.
This makes it easy to select just Linux hosts, or those running a
certain version of Apache, or whatever interests you. You can easily
modify the filter or remove it to see the whole scan again. See
http://nmap.org/book/zenmap-filter.html. [Josh Marlow]
o For some UDP ports, Nmap will now send a protocol-specific payload
that is more likely to get a response than an empty packet is. This
improves the effectiveness of probes to those ports for host
discovery, and also makes an open port more likely to be classified
open rather than open|filtered. The ports and payloads are defined
in payload.cc. The ports that have a payload are 7 (echo),
53 (domain), 111 (rpcbind), 123 (ntp), 137 (netbios-ns), 161 (snmp),
177 (xdmcp), 500 (isakmp), 520 (route), 1645 and 1812 (radius),
2049 (nfs), 5353 (zeroconf), and 10080 (amanda). [David]
o Integrated 1,349 fingerprints (and 81 corrections) submitted by Nmap
users! They resulted in 342 new fingerprints (a 17% increase),
including Google's Android Linux system for smart phones, Mac OS X
10.6 (Snow Leopard), the Chumby, and a slew number of printers, broadband
routers, and other devices (40 new vendors). See
http://seclists.org/nmap-dev/2009/q4/416 [David]
o [NSE] For all the services which are commonly tunneled over SSL
(pop3, http, imap, irc, smtp, etc.), we audited the scripts to
ensure they can support that tunneling. The com.tryssl function
was added for easy SSL detection. See
http://nmap.org/nsedoc/lib/comm.html [Joao]
o Nmap now prefers to display the hostname supplied by the user instead
of the reverse-DNS name in most places. If a reverse DNS record
exists, and it differs from the user-supplied name, it is printed
like this:
Nmap scan report for www.google.com (74.125.53.103)
rDNS record for 74.125.53.103: pw-in-f103.1e100.net
And in XML it looks like:
<hostnames>
<hostname name="openbsd.org" type="user"/>
<hostname name="cvs.openbsd.org" type="PTR"/>
</hostnames>
Host latency is now printed more often. See
http://seclists.org/nmap-dev/2009/q4/199 for a summary of other
output changes. [David]
o Ndiff now shows changes in script (NSE) output for each target
host (in both text output format and XML). [David]
o We now print output for down hosts, even when doing scanning beyond
just a ping scan. This always prints to XML and grepable output,
and is printed to normal and interactive output in verbose mode. The
format for printing a down host has changed slightly: "Nmap scan
report for 1.1.1.1 [host down]" [David]
o [NSE] Default socket parallelism has been doubled from 10 to 20,
which doubles speed in some situations. See
http://seclists.org/nmap-dev/2009/q3/161. [Patrick]
o Version detection's maximum socket concurrency has been increased
from 10-20 based on timing level to 20-40. This can dramatically
speed up version detection when there are many open ports in a host
group being scanned. [Fyodor]
o The Nmap source tarball (and RPMs) now included man page
translations (16 languages so far). Nmap always installs the English
man page, and installs the translations by default. If you only want
some of the translations, set the LINGUAS environmental variable to
the language codes you are interested in (e.g. "es de"). You can
specify the configure option --disable-nls or set LINGUAS to the
empty string to avoid installation of any man page translations. The
RPM always installs them. [David]
o [NSE] Added a function for scripts to format their output in a
consistent way. See
http://nmap.org/nsedoc/lib/stdnse.html#format_output. [Ron]
o [NSE] Now supports worker threads so that a single script can
perform multiple network operations concurrently. This patch also
includes condition variables for synchronization. See
http://nmap.org/nsedoc/lib/stdnse.html#new_thread,
http://nmap.org/nsedoc/lib/nmap.html#condvar, and
http://seclists.org/nmap-dev/2009/q4/294.
o Fixed a problem in which the Nmap installer wrongly reported that
the Microsoft Visual C++ 2008 Redistributable Package (vcredist.exe)
failed to install. We had to update a registry key--see
http://seclists.org/nmap-dev/2009/q3/164. [Jah]
o Added support for connecting to nameservers over IPv6. IPv6 addresses
can be used in /etc/resolv.conf or with the --dns-servers option. The
parallel reverse DNS resolver still only support IPv4 addresses, but
it can look them up over IPv6. [Ankur Nandwani]
o Zenmap now includes ports in the services view whenever Nmap found
them "interesting," whatever their state. Previously they were only
included if the state was "open", "filtered", or "open|filtered",
which led to confusing behavior when a closed port showed up in the
Services column but clicking on the service showed no ports in the
display. [David]
o [Ncat] Now has configure-time ASCII art just like Nmap does:
. .
\`-"'"-'/
} 6 6 {
==. Y ,==
/^^^\ .
/ \ ) Ncat: A modern interpretation of classic Netcat
( )-( )/
-""---""--- /
/ Ncat \_/
( ____
\_.=|____E
o [NSE] Added HTTP pipelining support to the HTTP library and and to
the http-enum, http-userdir-enum, and sql-injection.nse
scripts. Pipelining can increase speed dramatically for scripts
which make many requests.
o [NSE] The HTTP library now caches responses from http.get or
http.head so that resources aren't requested multiple times during
the same Nmap run even if several scripts request them. See
http://seclists.org/nmap-dev/2009/q3/733. [Patrick]
o [Ncat, Ndiff] The exit codes of these programs now reflect whether
they succeeded. For Ncat, 0 means the connection was successful, 1
indicates a network error, and 2 indicates any other error. For
Ndiff, 0 means the scans were equal, 1 means they were different,
and 2 indicates a runtime error. [David]
o [Ncat] In verbose mode, Ncat now prints the number of bytes read and
written after the client connection is terminated. Ncat also now
prints elapsed time. For example, "Ncat finished: 16 bytes sent, 566
bytes received in 8.05 seconds." [Venkat]
o [NSE] telnet-brute.nse now uses the unpw database instead of a
hard coded list. [Ron]
o [NSE] ssl-cert.nse now supports TLS negotiation against SMTP ports
that support it. [Tom Sellers, David]
o [NSE] Scripts that are listed by name with the --script option now
have their verbosity level automatically increased by one. Many
will print negative results ("no infection found") at a higher
verbosity level. The idea is that if you ask for a script
specifically, you are more interested in such results.
[David, Patrick]
o Upgraded our Winpcap installer to use the new WinPcap version 4.1.1.
A bug which could prevent proper uninstallation of previous versions
was fixed at the same time. Later we made it set some registry keys
for compatibility with the official Winpcap project installer (see
http://seclists.org/nmap-dev/2009/q4/237). [Rob Nicholls]
o [Ncat] Ncat now prints a message like "Connection refused." by
default when a socket error occurs. This used to require -v, but
printing no message at all could make a failed connection look like
success in a case like
ncat remote < short-file
o Zenmap no longer displays down hosts in the GUI. [Josh]
o The Ndiff man page was dramatically improved with examples and
sample output. See http://nmap.org/ndiff/man.html.
[David]
o [NSE] At debug level 2 or higher (-d2), Nmap now prints all active
scripts (running & waiting) and a backtrace whenever a key is
pressed. This can be quite helpful in debugging deadlocks and other
script/NSE problems. [Patrick]
o Nmap now allows you to specify --data-length 0, and that is now the
documented way to disable the new UDP protocol-specific probe
payload feature. [David]
o Fixed compilation of our libdnet on Debian GNU/kFreeBSD (patch from
Petr Salinger).
o Our Windows packages are now built on Windows 7, though they are
32-bit binaries and should continue to work on Win2K and later.
o Fixed a bug that could cause an infinite loop ("Unable to find
listening socket in get_rpc_results") in RPC scan. The loop would
happen when scanning a port that sent no responses, and there was at
least one other port to scan. Thanks to Lionel Cons for reporting
the problem. [David]
o [NSE] The dns-zone-transfer and whois script argument table syntax has been
improved so you don't need curly braces.
o [NSE] smb-enum-shares.nse now checks whether or not a share is
writable by attempting to write a file (and deleting it if it's
successful). Significantly cleaned up the code, as well. [Ron]
o The nselib/data directory is now installed. It was not installed
before because of an error in the Makefile. The scripts that would
not have worked after installation because they were missing data
files are http-enum.nse, http-favicon.nse, http-iis-webdav-vuln.nse,
http-userdir-enum.nse, smb-pwdump.nse, pop3-brute.nse,
smb-brute.nse, and snmp-brute.nse. [David]
o Upgraded the included libpcap to 1.0.0. [David]
o Optimize MAC address prefix lookup by using an std::map rather than
a custom hash table. This increases performance and code simplicity
at the cost of some extra memory consumption. In one test, this
reduced the time of a single target ARP ping scan from 0.59 seconds
to 0.13. [David]
o Added -Pn and -sn as aliases for -PN and -sP, respectively. They
will eventually become the recommended and documented way to disable
host discovery (ping scanning) and port scanning. They are more
consistent and also match the existing -n option for disabling
reverse DNS resolution. [David]
o Fixed an error in the handling of exclude groups that used IPv4
ranges. Si Stransky reported the problem and provided a number of
useful test cases in http://seclists.org/nmap-dev/2009/q4/276. The
error caused various assertion failures along the lines of
TargetGroup.cc:465: int
TargetGroup::get_next_host(sockaddr_storage*, size_t*):
Assertion `ipsleft > 1' failed.
[David]
o [NSE] Improved the authentication used by the smb-* scripts. Instead of
looking in a bunch of places (registry, command-line, etc) for the
usernames/passwords, a table is kept. This lets us store any number
of accounts for later use, and remove them if they stop working. This
also fixes a bug where typing in a password incorrectly would lock
out an account (since it wouldn't stop trying the account in question).
[Ron]
o Removed IP ID matching in packet headers returned in ICMP errors.
This was already the case for some operating systems that are known
to mangle the IDs of sent IP packets. Requiring such a match could
occasionally cause valid replies to be ignored. See
http://seclists.org/nmap-dev/2009/q2/580 for an example of host
order affecting scan results due to this phenomenon. [David]
o [NSE] The HTTP library now handles chunked transfer decoding more
robustly. See http://seclists.org/nmap-dev/2009/q3/13 [David]
o [NSE] Unexpected error messages from scripts now include the target
host and port number. [David]
o [NSE] Fixed many libraries which were inappropriately using global
variables, meaning that multiple scripts running concurrently could
overwrite each others values. NSE now automatically checks for this
problem at runtime, and we have a static code checker
(check_globals) available as well. See this whole thread
http://seclists.org/nmap-dev/2009/q3/70. [Patrick]
o Added some additional matching rules to keep a reply to a SYN probe
from matching an ACK probe to the same port, or vice versa, in ping
scans that include both scan types. Such a mismatch could cause an
ineffective timing ping or traceroute probe to be selected. [David]
o [Zenmap] There is a new command-line option, --confdir, which sets
the per-user configuration directory. Its value defaults to
$HOME/.zenmap. This was suggested by Jesse McCoppin. [David]
o Open bpf devices in read/write mode, not read-only, in libdnet on
BSD. This is to work around a bug in Mac OS X 10.6 that causes
incoming traffic to become invisible. [David]
o "make install" now removes from the Nmap script directory some
scripts which only existed in previous versions of Nmap but weren't
deleted during upgrades. [David]
o [NSE] Added the reconnect_ssl method for sockets. We sometimes need
to reconnect a socket with SSL because the initial communication on
the socket is done without SSL. See this thread for more details:
http://seclists.org/nmap-dev/2009/q4/3 [Patrick, Tom Sellers]
o [Zenmap] Fixed a crash that could occur when entering certain
characters in the target entry (those whose UTF-8 encoding contains
a byte that counts as whitespace in the Windows locale):
File "zenmapGUI\ScanNotebook.pyo", line 184, in _target_entry_changed
File "zenmapCore\NmapOptions.pyo", line 719, in render_string
UnicodeDecodeError: 'utf8' codec can't decode byte 0xc3 in position 1:
unexpected end of data
For more details on this curious problem, see
http://seclists.org/nmap-dev/2009/q4/82 [David]
o [NSE] There is a new function, nmap.bind, to set the source address
of a socket. [David]
o [Nsock] Made it a fatal error instead of silent memory corruption
when an attempt is made to use a file descriptor whose number is not
less than FD_SETSIZE. This applies only on non-Windows platforms
where FD_SETSIZE is a limit on the value of file descriptors as well
as a limit on the number of descriptors in the set. The error will
look like
nsock_core.c:186: Attempt to FD_SET fd 1024, which is not less
than FD_SETSIZE (1024). Try using a lower parallelism.
Thanks to Brandon Enright for discovering the problem and much help
debugging it, and to Jay Fink for submitting an initial patch. [David]
o [Ncat] Fixed proxy connections in connect mode on Windows. Because
the dup function does not work on Windows, an assertion failure
would be raised reading
(fh >= 0 && (unsigned)fd < (unsigned)_nhandle)
[David]
o [Ncat] Fixed the combination of --max-conns and --exec on Windows.
The count of connected clients was not decreased when the program
spawned by --exec finished. With --max-conns 5, for example, no more
connections would be allowed after the fifth, even if some of the
earlier ones had ended. Jon Greaves reported the problem and Venkat
contributed a patch.
o [Ncat] The code that manages the count of connected clients has been
made robust with respect to signals. The code was contributed by
Solar Designer.
o The files read by the -iL (input from file) and --excludefile
options now support comments that start with # and go to the end of
the line. [Tom Sellers]
o [Zenmap] On Windows, Zenmap no longer uses the cmd.exe shell to run
Nmap sub-processes. This means that canceling a scan will kill the
Nmap process as it does on other platforms (previously it would just
kill the shell). It also means that that scanning will work as a
user whose name contains characters like '&' that are significant to
the shell. Mike Crawford and Nick Marsh reported bugs related to
this. [David]
o [NSE] All scripts (except for those in "version" or "demo"
categories) are now classified in either the "safe" or "intrusive"
categories, based on how likely they are to cause problems when run
against other machines on the network. Those classifications already
existed, but weren't used consistently. [Fyodor]
o Added a check for a SMBv2 vulnerability (CVE-2009-3103) to
smb-check-vulns. Due to its nature (it performs a DoS, then checks
if the system is still online), the script isn't run by default and
requires a special script-arg to work. See
http://nmap.org/nsedoc/scripts/smb-check-vulns.html. [Ron]
o Fixed an integer overflow in uptime calculation which could occur
when a target with a low TCP timestamp clock frequency uses large
timestamp values, such that a naive uptime calculation shows a boot
time before the epoch. Also fixed a printf format specifier mismatch
that was revealed by the bug. Toby Simmons reported the problem and
helped with the fix. [David]
o [NSE] The HTTP library now supports HTTP cookies. [Joao Correa]
o Fixed a compile error on NetBSD. It was
tcpip.cc:2948: error: pointer of type 'void *' used in arithmetic
Thanks to Jay Fink for reporting the problem and submitting a patch.
o [Zenmap] If you have any hosts or services selected, they will
remain selected after aggregating another scan or running a filter
(as long as they are still up and visible). Previously the selection
was lost whenever the scan inventory was changed. This is
particularly important due to the new host filter system. [David]
o [Zenmap] New translation: Russian (contributed by Alexander Khodyrev).
Updated translations: French and German.
o Nmap now generates IP addresses without duplicates (until you cycle
through all the allowed IPs) thanks to a new collision-free 32-bit
number generator in nbase_rnd.c. See
http://seclists.org/nmap-dev/2009/q3/695 [Brandon]
o There is a new OS detection pseudo-test, SCAN.DC, which records how
the network distance in SCAN.DS was calculated. Its value can be "L"
for localhost, "D" for a direct connection, "I" for an ICMP TTL
calculation, and "T" for a traceroute hop count. This is mainly for
the benefit of OS integration, when it is sometimes important to
distinguish between DS=1%DC=I (probably the result of forged TTLs)
and DS=1%DC=D (a true one-hop connection.) [David]
o Canonicalized the list of OS detection device types to a smaller set
with descriptions: http://nmap.org/svn/docs/device-types.txt.
[David, Fyodor, Doug]
o [Ncat] The --idle-timeout option now exits when *both* stdin and the
socket have been idle for the given time. Previously it would exit
when *either* of them had been idle, meaning that the program would
quit contrary to your expectation when downloading a large file
without sending anything, for example. [David]
o [Ncat] Ncat now always prefixes its own output messages with "Ncat: "
or "NCAT DEBUG: " to make it clear that they are not coming from the
remote host. This only matters when output goes to a terminal, where
the standard output and standard error streams are mixed. [David]
o Nmap's Nbase library now has a new hexdump() function which produces
output similar to Wireshark. nmap_hexdump() is a wrapper which
prints the output using Nmap's log_write facility. The old hdump()
and lamont_dump() functions have been removed. [Luis]
o Added explicit casts to (int)(unsigned char) for arguments to ctype function
calls in nmap, ncat and nbase. Thanks to Solar Designer for pointing out
the need and fix for this. [Josh]
o Ncat now supports wildcard SSL certificates. The wildcard character
(*) can be in commonname field or in DNS field of Subject
Alternative Name(SAN) Extension of SSL certificate. Matching Rules:
-'*' should be only on the leftmost component of FQDN.(*.example.com
but not www.*.com or www.example*.com).
-The leftmost component should contain only '*' and it should be
followed by '.'(*.example.com but not *w.example.com or
w*.example.com).
-There should be at least three components in FQDN.(*.exmaple.com but
not *.com or *.com.).[venkat]
o Nmap now handles the case when a primary network interface (venet0)
does not have an address assigned but its aliases do (venet0:1
etc.). This could result in the error messages
Failed to find device venet0 which was referenced in /proc/net/route
Failed to lookup subnet/netmask for device (venet0): venet0: no IPv4 address assigned
This was observed under OpenVZ. [Dmitry Levin]
o [Ncat] The --ssl-cert, --ssl-key, and --ssl-trustfile options now
automatically turn on SSL mode. Previously they were ignored if
--ssl was not also used. [David]
o [Nsock] Now Nsock supports pure TLSv1 and SSLv3 servers in addition
to the (already supported and far more common) SSLv2 and SSLv23
servers. Ncat currently never uses SSLv2 for security reasons, so
it is unaffected by this change.
o [Ncat] Implemented basic SCTP client functionality (server already
exists). Only the default SCTP stream is used. This is also called
TCP compatible mode. While it allows Ncat to be used for manually
probing open SCTP ports, more complicated services making use of
multiple streams or depending on specific message boundaries cannot
be talked to successfully. [Daniel Roethlisberger]
o [Ncat] Implemented SSL over SCTP in both client (connect) and server
(listen) modes. [Daniel Roethlisberger]
o Nmap now filters received ARP packets based on their target address
address field, not the destination address in the enclosing ethernet
frame. Some operating systems, including Windows 7 and Solaris 10,
are known to at least sometimes send their ARP replies to the
broadcast address and Nmap wouldn't notice them. The symptom of this
was that root scans wouldn't work ("Host seems down") but non-root
scans would work. Thanks to Mike Calmus and Vijay Sankar for
reporting the problem, and Marcus Haebler for suggesting the
fix. [David]
o The -fno-strict-aliasing option is now used unconditionally when
using GCC. It was already this way, in effect, because a test
against the GCC version number was reversed: <= 4 rather than >= 4.
Solar Designer reported the problem.
o Nmap now prints a warning instead of a fatal error when the hardware
address of an interface can't be found. This is the case for
FireWire interfaces, which have a hardware address format not
supported by libdnet. Thanks to Julian Berdych for the bug report.
[David]
o Zenmap's UI performance has improved significantly thanks to
optimization of the update_ui() function. In particular, this speeds
up the new host filter system. [Josh]
o Add a service probe for DNS-based service discovery (DNS-SD). See
http://seclists.org/nmap-dev/2009/q3/0610.html. [David]
o Made RPC grinding work from service detection again by changing the
looked-for service name from "rpc" to "rpcbind", the name it has in
nmap-service-probes. Also removed some dead code. [David]
o Fixed a log_write call and a pfatal call to use a syntax which is
safer from format strings bugs. This allows Nmap to build with the
gcc -Wformat -Werror=format-security options. [Guillaume Rousse,
Dmitry Levin]
o A bug in Nsock was fixed: On systems where a non-blocking connect
could succeed immediately, connections that were requested to be
tunneled through SSL would actually be plain text. This could be
verified with an Ncat client and server running on localhost. This
was observed to happen with localhost connections on FreeBSD 7.2.
Non-localhost connections were likely not affected. The bug was
reported by Daniel Roethlisberger. [David]
o Ncat proxy now hides the proxy's response ("HTTP/1.0 200 OK" or
whatever it may be). Before, if you retrieved a file through a
proxy, it would have the "HTTP/1.0 200 OK" stuck to the top of
it. For this Ncat uses blocking sockets until the proxy negotiation
is done and once it is successful, Nsock takes over for rest of the
connection.[Venkat]
o [NSE] socket garbage collection was rewritten for better performance
and to ensure that socket slots are immediately available to others
after a socket is closed. See
http://seclists.org/nmap-dev/2009/q2/0624.html. [Patrick]
o [NSE] Fixed a rare but possible segfault which could occur if the
nsock binding attempted to push values on the stack of a thread
which had already ended due to an error, and if that internal Lua
stack was already completely full. This bug is very hard to
reproduce with a SEGFAULT but is usually visible when Lua assertion
checks are turned on. A socket handler routine must be called AFTER
a thread has ended in error. [Patrick]
o [Ncat] Fixed an error that would cause Ncat to use 100% CPU in
broker mode after a client disconnected or a read error happened.
[Kris, David]
o [NSE] --script-args may now have whitespace in unquoted strings (but
surrounding whitespace is ignored). For example,
--script-args 'greeting = This is a greeting' Becomes:
{ ["greeting"] = "This is a greeting" } [Patrick]
o [Ncat] Using --send-only in conjunction with the plain listen or
broker modes now behaves as it should: nothing will be read from the
network end. Ncat previously read and discarded any data
received. [Kris]
o [Nsock] Added a socket_count abstraction that counts the number of
read or write events pending on a socket, for the purpose of
maintaining an fd_set. The bit is set in the fd_set whenever the
count is positive, and cleared when it is zero. The reason for doing
this was that write bits were not being properly cleared when using
Ncat with SSL in connect mode, such that a client send would cause
Ncat to use 100% CPU until it received something from the
server. See the thread at
http://seclists.org/nmap-dev/2009/q2/0413.html. This change will
also make it easier to use a different back end than select in the
future. [David]
o [Nsock] Added compilation dependency generation (makefile.dep)
[David]
o [Ncat] The --broker option now automatically implies --listen. [David]
o Fixed a logic error in getinterfaces_siocgifconf. The check for
increasing the capacity of the list of interfaces was off by
one. This caused a crash on initialization for systems with more
than 16 network interfaces. [David]
o Added Apache JServe protocol version detection probe and signatures
and some some other nmap-service-probes patches. [Tom Sellers]
o Fixed two memory leaks in ncat_posix.c and a bug where an open file was not
being closed in libdnet-stripped/src/intf.c [Josh Marlow]
o [Zenmap] Added profile editor support for the Nmap SCTP options:
-PY, -sY and -sZ. [Josh Marlow]
o Fixed a bug in --data-length parsing which in some cases could
result in useless buffer allocations and unpredictable payload
lengths. See http://seclists.org/nmap-dev/2009/q2/0763.html [Luis]
o The configure script now allows cross-compiling by assuming that
libpcap is recent enough to use rather than trying to compile and
run a test program. Libpcap will always be recent enough when Nmap's
included copy is used. [Mike Frysinger]
o Updated the IANA assignment IP list for random IP (-iR)
generation. The Mac OS prefix file was updated as
well. [Kris, Fyodor]
o [Zenmap] Fix a bug which could cause a crash in the (very rare) case
where Nmap would produce port tags in XML output without a state
attribute. [David]
o Added a convenience top-level BSDmakefile which automatically
redirects BSD make to GNU make on BSD systems. The Nmap Makefile
relies on numerous GNU Make extensions. [Daniel Roethlisberger]
Enjoy the new release!
-Fyodor
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/
 By Date 
By Thread
Current thread:
- Nmap 5.10BETA1 released Fyodor (Nov 23)
|
Intro
