-
BTW, I checked in a change to the top-of-tree Git version to fix the problems with autoconf 2.64 and later.
2010-01-02 22:25:06 UTC in The libpcap project
-
Also, are CAN devices like other network devices, so that they might show up in, for example, /proc/net/dev - or so that you can bind a PF_PACKET socket to them, and capture on that socket and see CAN packets?
If the latter is true, would the existing Linux PF_PACKET socket code in libpcap be able to capture on CAN devices?.
2010-01-01 03:02:21 UTC in The libpcap project
-
It's still a pain, but that's nothing we can fix. At least I got a damn e-mail when I added this; hopefully it didn't just do so because I'm the one who submitted it....
("Rejected" indeed. Why does SourceForge make angry ticks fire out of my nipples? :-))
2009-12-31 02:25:06 UTC in The tcpdump project
-
*NOTHING* about it is convenient.
(This is a test, to see whether I've finally managed to figure out how to get it to send me e-mail, directly, when a new tracker item is added. Wish me luck....)
2009-12-31 02:22:46 UTC in The tcpdump project
-
We already require in other places, including pcap-linux.c, and we don't check for its availability anywhere. Other tests in the configure script just assume it's available; I changed the SocketCAN test not to bother checking for either, and checked it in.
Is there any way to check what canN/vcanN devices exist, so that we can enumerate them in...
2009-12-31 01:19:47 UTC in The libpcap project
-
Fixed in the top of the Git tree.
2009-12-30 21:01:32 UTC in The libpcap project
-
pcap-can-linux.c itself includes before ; should the check for include rather than checking for and including it if present?
Also, the fourth argument to AC_CHECK_HEADER() should just be the additional headers required, such as , not the header you're checking for (AC_CHECK_HEADER() generates the include...
2009-12-30 20:59:33 UTC in The libpcap project
-
I've checked the most recent version of the patch in, with the first 4 bytes of the header being in network byte order.
I still have a question - the configure script checks for a header, but I don't see any version.h header in the Linux 2.6.29 kernel's "include/linux/can" directory, and pcap-can-linux.c only includes and . Should the...
2009-12-29 09:31:30 UTC in The libpcap project
-
1. The buffer into which you read from the socket should be large enough for the largest possible packet you can get from the socket.
2. Move the data up only if it either makes code that processes the packet simpler or if it makes the packet look like a CAN frame on the wire - as I remember from the discussion on tcpdump-workers, the packet you get from the socket is suffiiciently...
2009-12-07 19:36:19 UTC in The libpcap project
-
No, it's due to fixing OS X bug 5729372; the behavior you're seeing in Snow Leopard is what you're supposed to be seeing on all systems that use BPF for packet capturing, including OS X and {Free,Net,Open,DragonFly}BSD. (Well, not on AIX, but AIX's BPF is weird.) FreeBSD had that bug a while ago, but it was fixed; OS X fixed it in 10.6. The bug is in BPF, so it affects *all* network adapters;...
2009-12-07 09:33:55 UTC in The libpcap project
