Update from Jason (Xin) Li to reflect changes to the FreeBSD
SIOCGIFDESCR implementation - it now doesn't return an error if the
buffer is too short, it sets the buffer pointer to NULL. No FreeBSD
release has SIOCGIFDESCR, so this doesn't break on any release.
The loop, trying to increase the buffer size until it's big enough,
works only on FreeBSD, as that's the only OS where you get told what
length to use; OpenBSD clamps the description length at IFDESCRSIZE, so
we just use that.
Update to work with FreeBSD flavor of SIOCGIFDESCR as well as the
OpenBSD flavor. Also, a small tweak from me to use "ifdescr" as a
pointer to the string buffer in both cases ("&ifdescr" and "ifdescr",
although they have different types, point to the same location), and to
put "s" back inside #ifdef SIOCGIFDESCR/#endif, as it's not used if
SIOCGIFDESCR isn't defined.
This patch adds support for our NICs when run in a specialized capture mode.
It is diffed against the current master.
The Myricom Sniffer10G software uses Myri-10G programmable Network Interface
Cards (NICs), a firmware extension, a specialized driver and a user-level
library (libsnf) to enable sustained capture of 10-Gigabit Ethernet traffic.
Small-packet coalescing and an efficient zero-copy path to host memory allow
Sniffer10G to capture streams at line rate for all Ethernet packet sizes.
Optionally, libpcap can be used concurrently by multiple processes on a single
NIC port to partition the incoming traffic across processes. While the Linux
kernel enables this through multiple receive queues, the difference is that the
myri_snf driver cooperates with libsnf to set up multiple queues that are each
independently accessible through user-space.
More information here: http://www.myri.com/scs/SNF/doc
Signed-off-by: Guy Harris <guy@alum.mit.edu>
From NetBSD; to quote the checkin comment:
Fix pcap_lookupnet(): reset ifr before SIOCGIFNETMASK. Without
it we get back a bogus netmask.
Presumably some stuff left over in ifr from the previous ioctl confuses
the next ioctl.
several files:
date: 2006/02/27 15:53:24; author: drochner; state: Exp;
avoid shadowing globals, for WARNS=2
date: 2006/02/27 15:55:30; author: drochner; state: Exp;
minor constification, good for WARNS=3 now
date: 2006/02/27 15:57:17; author: drochner; state: Exp;
NetBSD adaption:
...
-const pcap_strerror() for consistency
gencode.c:
date: 2006/04/26 09:24:33; author: tron; state: Exp;
Add missing "const" keywords to match declarations in "pcap.h".
date: 2006/10/15 19:27:21; author: christos; state: Exp;
add a volatile variable to prevent vfork/longjmp clobbering.
optimize.c:
date: 2006/05/17 17:48:36; author: drochner; state: Exp;
Make the optimizer use unsigned numbers as the kernel does.
While it is not agreed on that purely unsigned arithmetics is nice,
different behaviour of optimized and unoptimized code is less desirable.
pcap-bpf.c:
date: 2006/02/27 15:51:38; author: drochner; state: Exp;
pull in from NetBSD's libpcap: use cloning bpf device on NetBSD
Have the configure script check for paths.h, so that we can include it
only if we have it, and use the cloning BPF device only if we're on
NetBSD *and* _PATH_BPF is defined (hopefully this will keep us from
using it on versions of NetBSD that don't have a cloning BPF device; if,
in the future, other OSes with BPF get cloning BPF devices, we can make
this work for them as well).
captured on, without trying to open it - if "enN" is an AirPort device,
opening "wltN" is like opening "enN", except that the device is put into
monitor mode, which, for many devices, causes the device to deassociate
from the network with which it's associated.
the list for the interface, and try to open the interface only if we
don't already have an interface, so if you have multiple addresses for
an interface, you don't open the interface for each of them. (See
Wireshark bug 1151:
http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=1151
for the reason to do this; pcap_findalldevs() spent a *lot* of time and
energy opening en0 over and over again for all 64K-128K(!) addresses.)
lacking it, so use it in some other cases, instead of duplicating its
functionality.
Check, however, whether it succeeds, and fail appropriately when it
fails.
Also, when other mallocs fails, free up any memory allocated for data
structures, so it doesn't leak.
to the list of interfaces, so we don't, for example, include the
loopback device in the list on Solaris - SIOCGIFCONF/SIOCGLIFCONF will
find the device, as it has an IP address, but it's not a DLPI device so
you can't open it for capturing.
argument to "pcap_open_live()" a "const" pointer.
Constify some additional device name arguments, and update the man page
to reflect some arguments that were already consts.
each source file, only the headers that file needs, and all the headers
it needs in order to compile on various platforms and not to get any
avoidable compiler warnings on those platforms (as well as any
incomplete structure definitions needed to avoid those warnings).
That also means that <pcap.h> doesn't include <pcap-stdinc.h> on UNIX;
we don't want it to include <pcap-stdinc.h>, at least on UNIX, as doing
so
1) would mean we'd have to install that, so that programs can
build with libpcap
and
2) would mean that programs including <pcap.h> would drag in a
bunch of header files that they don't need.
Put a newline at the end of "inet.c" - the Sun C compiler doesn't like
it if the last line doesn't end with a newline.
source files, rather than having a pile of #ifdefs in "inet.c". Add
code to the configure script to determine which implementation to use on
the platform for which libpcap is being built.
Add a "pcap_findalldevs()" implementation for Solaris 8 and later that
handles IPv6 addresses.
"struct rtentry" (in <net/if.h> on some systems, e.g. Digital UNIX 4.0,
there are prototypes for kernel functions that include "struct mbuf *"
and "struct rtentry *" arguments, and they are included even if you're
just user-mode code).
field, and make a PCAP_IF_LOOPBACK flag be the first flag bit in that
field, specifying whether the interface is a loopback interface; this
allows us to add more flags without changing the layout of the
structure.
"getifaddrs()"), after processing the list returned by SIOCGIFCONF, scan
"/proc/net/dev" for interface names, and add to the list of interfaces
entries for those interfaces, with no associated addresses (if the
interfaces were already added, with addresses, from the list returned by
SIOCGIFCONF, they won't get added again).
Clean up the error handling a bit.
whether we have "freeifaddrs()" (we don't check whether we have
"getifaddrs()", and if we have "getifaddrs()" but not "freeifaddrs()",
we're stuck with leaking memory).
Give the "any" device an instance number of INT_MAX, so it shows up
after all other non-loopback devices.
"getifaddrs()" sometimes appears to supply a destination address even
for non-point-to-point interfaces (it did so on a FreeBSD 4.1 system);
don't use the broadcast address it supplies if an interface isn't a
broadcast interface, and don't use the destination address it supplies
if an interface isn't a point-to-point interface.
If we had an error constructing the list of interfaces, don't attempt to
add the "any" device to the list.
Patrick McHardy
Gianluca Varenni