Age | Commit message (Collapse) | Author | Files | Lines |
|
clang gives the following warning about function meas_type():
proto_80211_mac_hdr.c:1704:1: warning: control may reach end of non-void function [-Wreturn-type]
}
^
Even though this is a false positive (since we check the entire range of
an u8 in the switch/case), fix it by turning the case 13 ... 255 into
default.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Both sparse and clang warn about the initializers overriding previous
initialization of the packet_types array. Since every access of the
packet_types array checks the value for NULL (the default value, since the
array is static) and prints a "?" if it isNULL, we don't need the prior
initialization with "?".
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Currently we ignore the return value of the write() calls to
/proc/sys/net/core/{r,w}mem_{default,max} in set_system_socket_mem().
Better check the return value and notify the user about it.
This also fixes a clang compiler warning about a variable explicitely
assigned to itself.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
If ctx.device_in is NULL after option parsing, it is always set to
"any", which is before this check. Thus, it serves no purpose and can be
removed.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Don't reset the variable CC in the Makefile, but set it to gcc by
default (can be overriden when e.g. compiling with `make CC=clang').
This way we can also have non-verbose compiles with compilers other than
gcc. Furthermore make sure that the CC variable is also respected for
LD.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Try to convert the directory entry name to an unsigned integer with
strtoul() instead of using strspn() to determine if a proc entry is a
PID. If it is a valid PID (i.e. strtoul returned a value != 0), we can
directly use it to pass into walk_process() and there set
flow_entry->proc_num.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Instead of testing three skip conditions for every entry, make use of
the short-circuit evaluation of the boolean OR operator to only test as
few conditions as necessary.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Create a make variable $(COMPILER) which is set to either `gcc' or
`clang' depending on the compiler used. Use this variable to filter out
command line options not supported by clang (for now only
-fno-delete-null-pointer-checks).
Compiler detection snippet extracted from the Linux kernel Makefile.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Make sure we also use clang for the compile tests in the configure
script.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
When running the configure script with `CC=clang ./configure' clang
fails all the compile test programs due to the main() function not
having return type int. Change it, so we can use clang as a compiler in
the configure script.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Move the copyright/bug report/license string to an own constant and use
it for the output of help() and version() to avoid duplication and
prevent the strings from getting out of sync. This makes the text
section of bpfc.o slightly smaller:
before:
text data bss dec hex filename
3594 0 0 3594 e0a bpfc/bpfc.o
after:
text data bss dec hex filename
3313 0 0 3313 cf1 bpfc/bpfc.o
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Move the copyright/bug report/license string to an own constant and use
it for the output of help() and version() to avoid duplication and
prevent the strings from getting out of sync. This makes the text
section of flowtop.o slightly smaller:
before:
text data bss dec hex filename
15601 4 48 15653 3d25 flowtop/flowtop.o
after:
text data bss dec hex filename
15228 4 48 15280 3bb0 flowtop/flowtop.o
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Move the copyright/bug report/license string to an own constant and use
it for the output of help() and version() to avoid duplication and
prevent the strings from getting out of sync. This makes the text
section of astraceroute.o slightly smaller:
before:
text data bss dec hex filename
18481 0 4 18485 4835 astraceroute/astraceroute.o
after:
text data bss dec hex filename
18201 0 4 18205 471d astraceroute/astraceroute.o
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Move the copyright/bug report/license string to an own constant and use
it for the output of help() and version() to avoid duplication and
prevent the strings from getting out of sync. This makes the text
section of ifpps.o slightly smaller:
before:
text data bss dec hex filename
22146 0 1456 23602 5c32 ifpps/ifpps.o
after:
text data bss dec hex filename
21807 0 1456 23263 5adf ifpps/ifpps.o
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Move the copyright/bug report/license string to an own constant and use
it for the output of help() and version() to avoid duplication and
prevent the strings from getting out of sync. This makes the text
section of netsniff-ng.o slightly smaller:
before:
text data bss dec hex filename
26998 8 68 27074 69c2 netsniff-ng/netsniff-ng.o
after:
text data bss dec hex filename
26582 8 68 26658 6822 netsniff-ng/netsniff-ng.o
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Move the copyright/bug report/license string to an own constant and use
it for the output of help() and version() to avoid duplication and
prevent the strings from getting out of sync. This makes the text
section of trafgen.o slightly smaller:
before:
text data bss dec hex filename
18512 0 104 18616 48b8 trafgen/trafgen.o
after:
text data bss dec hex filename
18240 0 104 18344 47a8 trafgen/trafgen.o
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
The number of CPUs is stored in ctx.cpus which is unsigned int, so use
unsigned int consistently when using CPU number. Negative CPU numbers
wont occur anyhow.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
clang doesn't understand all our (GCC-spcific) compiler flags. We need
to work around it by filtering them out properly in case clang is used.
For now, just disable clang on travis, until I figured this out locally.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
If travis compiles with clang, the environment variable CC is set. Since
the netsniff-ng build system doesn't currently pick it up unless
explicitly specified as a make variable, pass it on as such.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
In order to make sure all tools (except curvetun) are built by Travis
CI, install the libraries they depend upon.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
ANSI C doesn't allow empty parameter list, thus use void where
appropriate. This fixes the corresponding sparse warning.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
This allows any libnl github repo to link to travis-ci which
will result in new commits and pull requests to be automatically
built with both gcc and clang. Any build failures including the
addition of warnings are reported to the author and committer.
Following Thomas Graf's suggestion:
https://plus.google.com/103961622207897185587/posts/GaVC6jwcMgp
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Use zmalloc()'ed buffer or statically sized array to replace usage of
variable sized arrays. Found by the sparse static checker.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Binary constants are not C99, but a GCC extension. Moreover, Mohan
reports compilation errors resulting from these constructs when using a
PowerPC cross-compiler. Thus, replace them by the corresponding
hexadecimal constants.
Reported-by: Mohan Kannekanti <mohan.kannekanti@gmail.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Commit 76486ece10c6 ("trafgen: add option to not adjust system socket
mem during testrun") introduced the new command line option
-A/--no-sock-mem for trafgen but didn't add it to the zsh completion. Do
it now.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Pushed last commit a bit too fast. The help text described that
the single letter option for "--no-sock-mem" were "-m", which is
wrong the right option is "-A" (only help txt, correct in the code
and manual).
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
|
|
The default behavior of trafgen is to boost the systems default
socket memory limits during a testrun. This behaviour does not
always result in improved performance, because this will stress
the kernels SLAB allocators unnecessary.
Introducing an option "--no-sock-mem" that instruct trafgen
to not perform any socket memory adjustments.
Fixes: #130
Acked-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
|
|
Fixes sparse warnings like the following in some dissectors:
proto_arp.c:158:17: warning: symbol 'arp_ops' was not declared. Should it be static?
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Update oui.conf using oui-update.py
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Fixes the following sparse warnings:
geoip.c:76:30: warning: Using plain integer as NULL pointer
geoip.c:78:30: warning: Using plain integer as NULL pointer
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Instead of defining POLLRDNORM in built_in.h if it isn't provided by
system/kernel headers, define it in ring.h where its only user is. This
way we make sure that <poll.h> is included before checking #ifndef
POLLRDNORM. This fixes the following sparse warning:
/usr/include/bits/poll.h:32:10: warning: preprocessor token POLLRDNORM redefined
built_in.h:378:10: this was the original definition
Also remove the defines for POLLWRNORM and POLLRDHUP which are not used
anywhere in the code.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Similar to the 'make C=1' for the Linux kernel, wire up the sparse
command for our build system in a similar fashion. Use 'make C=1 [tool]'
to check the source files using sparse. Of course, the sparse tool needs
to be installed and available in your $PATH for this to work.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
dissector_eth.h is only used in dissectors which need to manipulate the
eth_lay2 and eth_lay3 hash tables. Remove the include from all
dissectors that don't access them.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
time for real)
Commit 0fab564a98d1 ("netsniff-ng: Properly wrap usage of all tpacket v3
structs") took care of protecting _some_ tpacket v3 structures with
compile error when building with !HAVE_TPACKET3 (reported by Mike
Reeves):
> CC ring_rx.c
> ring_rx.c: In function 'setup_rx_ring_layout':
> ring_rx.c:124: warning: implicit declaration of function 'set_sockopt_tpacket_v3'
> ring_rx.c: In function 'sock_rx_net_stats':
> ring_rx.c:194: error: field 'k3' has incomplete type
> make: *** [netsniff-ng/ring_rx.o] Error 1
Many thanks to Mike for helping me sort out these problems.
Reported-by: Mike Reeves <luke@geekempire.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Commit edca6174b09 ("dissector: Restore paket type if capturing from
nlmon device") changed the signature of show_frame_hdr(). The call to
this function was not updated in the !HAVE_TPACKET3 part of netsniff-ng
introduced in commit 97e6f994785c ("netsniff-ng: Restore tpacket v2
capturing"), causing a compile error. Fix this by providing the correct
parameters to show_frame_hdr() also in this case.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Mike Reeves reports the following compilation error if tpacket v3 is not
available:
> CC ring_rx.c
> ring_rx.c: In function 'alloc_rx_ring_frames':
> ring_rx.c:143: error: 'struct ring' has no member named 'layout3'
> ring_rx.c:144: error: 'struct ring' has no member named 'layout3'
> ring_rx.c: In function 'sock_rx_net_stats':
> ring_rx.c:172: error: field 'k3' has incomplete type
> make: *** [netsniff-ng/ring_rx.o] Error 1
The layout3 member of struct ring is only available for HAVE_TPACKET3.
Thus, wrap all access to it into inline functions defined depending on
wheter HAVE_TPACKET3 is defined.
Reported-by: Mike Reeves <luke@geekempire.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Instead of having #ifdef HAVE_TPACKET3 spread all over the code,
encapsulate the functionality depending on it inside inline functions:
the existing is_tpacket_v3() introduced in commit 5bc19d0b84d0
("netsniff-ng: Only use TPACKET_V3 if HAVE_TPACKET3 is defined") and the
newly introduced get_ring_layout_size() to get the ring layout size
depending on the tpacket version available and the version actually in
use.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
TPACKET_V3 is not defined if tpacket v3 is not available, thus make its
use conditional on HAVE_TPACKET3. Wrap the check for TPACKET_V3 in
ring_rx in an inline function which always returns false if
HAVE_TPACKET3 is not defined.
Reported-by: Mike Reeves <luke@geekempire.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
It's supposed to be LINKTYPE_NETLINK, not AF_NETLINK as otherwise
the pkt_type fixup cannot be done correctly.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Some older systems (e.g. RHEL 6) don't have tpacket v3 available, but
only tpacket v2. However, since commit d8cdc6a ("ring: netsniff-ng:
migrate capture only to TPACKET_V3") we solely rely on tpacket v3 for
capturing packets.
This patch restores the possibility to capture using tpacket v2. For now
this is just a fallback if the configure script doesn't detect tpacket
v3 (and thus HAVE_TPACKET3 isn't set). Thus, on most modern systems this
shouldn't change anything and they will continue using tpacket v3.
For now this fix contains quite a bit of ugly #ifdefery which should be
cleaned up in the future.
Fixes #76
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Allow to disable hardware time stamping using the command line switch
(-N/--no-hwtimestamp). This might be useful in situations where hardware
time stamps are skewed somehow.
Reference: #129
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Github shows 33 commits for me:
https://github.com/netsniff-ng/netsniff-ng/graphs/contributors
$ git shortlog | grep -i jon
Jon Schipp (29):
jonschipp (1):
Not sure what the difference is here, but this one will make 31 based
on git shortlog (I hope that's not cheating) :) Otherwise I can wait.
And I believe I have a few commits from the old history (is this still
up somewhere?). I think it used to be on the cryptoism site.
Signed-off-by: Jon Schipp <jschipp@illinois.edu>
Acked-by: Daniel Borkmann <borkmann@iogearbox.net>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Consistently vse the __packed macro defined in built_in.h instead of
__attribute__((packed)) in all dissectors (and the entire code base for
that matter).
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Specify the listen-addr for the CLI in /etc/netsniff-ng/mausezahn.conf
and remove the message, "mz: device not given, will use eth0" since
we're listening by IP, rather than interface now, and this shows continues
to show.
$ grep listen-addr /etc/netsniff-ng/mausezahn.conf
listen-addr = 127.0.0.1
$ mausezahn -V -x
mz: device eth0 got assigned 10.0.2.15 and is a possible candidate.
mz: device lo got assigned 127.0.0.1 (loopback)
Opening config file /etc/netsniff-ng/mausezahn.conf...
...
Mausezahn accepts incoming Telnet connections on 127.0.0.1:25542.
Signed-off-by: Jon Schipp <jonschipp@gmail.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
This documents two more configuration file options that can be used.
One, automops, is in the development phase (not compiled by default),
and the other one, management-only, is usable. It prevents a user from
using a network interface to pass data traffic with the intention that
the interface is used for machine managment (remote control).
$ cat /etc/netsniff-ng/mausezahn.conf
user = mzadmin
password = better-password
management-only = eth0
automops = /path/to/an/automops-file.xml
Signed-off-by: Jon Schipp <jonschipp@gmail.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Specify the port for the CLI in /etc/netsniff-ng/mausezahn.conf like so:
$ grep port /etc/netsniff-ng/mausezahn.conf
port = 65000
$ mausezahn -V -x -l 127.0.0.1
Mausezahn 0.40 - (C) 2007-2010 by Herbert Haas - http://www.perihel.at/sec/mz/
Use at your own risk and responsibility!
-- Verbose mode --
...
Opening config file /etc/netsniff-ng/mausezahn.conf...
...
Mausezahn accepts incoming Telnet connections on 127.0.0.1:65000.
Signed-off-by: Jon Schipp <jonschipp@gmail.com>
[tklauser: Fixed whitespace issues for patch to apply cleanly]
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Signed-off-by: Jon Schipp <jonschipp@gmail.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
|
|
Tobias pointed out and fixed in commit 9ef1dd15551f ("mausezahn:
Increase buffer size for listen address string") that the buffer isn't
log enough to hold the null terminator character. The declaration of the
buffer in cli.h was missed though.
Signed-off-by: Jon Schipp <jonschipp@gmail.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch> [commit msg edited]
|