On Thu, Aug 23, 2001 at 02:46:58PM -0700, Guy Harris wrote:
> (We could, I guess, just set the
> buffer size to the snapshot length on 2.2 and later kernels, which might
> save some memory if you use a small snapshot, but that's just a
> refinement.)
I've checked that in.
Currently:
if you build libpcap on a system whose build environment
supports PF_PACKET sockets, it falls back on SOCK_PACKET sockets
if the system on which you're running it doesn't support
PF_PACKET sockets;
if you build libpcap on a system whose build environment
supports MSG_TRUNC, it assumes MSG_TRUNC won't work on 2.0[.x]
kernels (which is, in fact, the case), and should work even on
those systems.
Is there any reason to care whether a libpcap built on a system with
PF_PACKET socket support and MSG_TRUNC support (i.e., 2.2 or later) will
work on a 2.0[.x] system? If not, we should probably have libpcap use
only PF_PACKET sockets, and assume MSG_TRUNC works, if built on a
platform that supports PF_PACKET sockets, and use only SOCK_PACKET
sockets, and assume MSG_TRUNC doesn't work, if built on a platform that
doesn't support PF_PACKET sockets. (As far as I know, you can't get
SOCK_PACKET support in a 2.2 or later kernel without also getting
PF_PACKET support.)
-
This is the TCPDUMP workers list. It is archived at
http://www.tcpdump.org/lists/workers/index.html
To unsubscribe use mailto:[EMAIL PROTECTED]?body=unsubscribe
