Joel L. Breazeale on Mon, 26 May 2003 00:13:05 -0400 |
Michael, I used --no-select and found better results for 127.0.0.1 and 24.163.210.43. Side-by-side results follow: $ sudo ./tcptraceroute 127.0.0.1 Selected device lo0, address 127.0.0.1, port 49517 for outgoing packets Tracing the path to 127.0.0.1 on TCP port 80 (http), 30 hops max 1 * * * ^C $ sudo ./tcptraceroute --no-select 127.0.0.1 Selected device lo0, address 127.0.0.1, port 49518 for outgoing packets Tracing the path to 127.0.0.1 on TCP port 80 (http), 30 hops max 1 localhost (127.0.0.1) [open] 0.333 ms 0.244 ms 0.186 ms $ sudo ./tcptraceroute 24.163.210.43 Selected device en0, address 192.168.1.200, port 49519 for outgoing packets Tracing the path to 24.163.210.43 on TCP port 80 (http), 30 hops max 1 * * * ^C $ sudo ./tcptraceroute --no-select 24.163.210.43 Selected device en0, address 192.168.1.200, port 49520 for outgoing packets Tracing the path to 24.163.210.43 on TCP port 80 (http), 30 hops max 1 cpe-24-163-210-43.mn.rr.com (24.163.210.43) [open] 1.354 ms 0.999 ms 0.952 ms $ Here's the non-local case too w/ and w/o --no-select: $ sudo ./tcptraceroute 66.135.192.87 Selected device en0, address 192.168.1.200, port 49530 for outgoing packets Tracing the path to 66.135.192.87 on TCP port 80 (http), 30 hops max 1 * * * ^C $ $ sudo ./tcptraceroute --no-select 66.135.192.87 Selected device en0, address 192.168.1.200, port 49529 for outgoing packets Tracing the path to 66.135.192.87 on TCP port 80 (http), 30 hops max 1 * * * ^C Here's all the data you wanted me to gather: $ sudo ./tcptraceroute -dv debug: tcptraceroute 1.5beta1 debug: Compiled with libpcap 0.6, libnet 1.0.2a (API 100) tcptraceroute 1.5beta1 Copyright (c) 2001, 2002, 2003 Michael C. Toren <mct@toren.net> Updates are available from http://michael.toren.net/code/tcptraceroute/ $ uname -a Darwin mogwai.local. 6.6 Darwin Kernel Version 6.6: Thu May 1 21:48:54 PDT 2003; root:xnu/xnu-344.34.obj~1/RELEASE_PPC Power Macintosh powerpc $ gcc -E -dM - </dev/null #define __BIG_ENDIAN__ 1 #define __HAVE_BUILTIN_SETJMP__ 1 #define __SIZE_TYPE__ long unsigned int #define __GNUC_PATCHLEVEL__ 0 #define __NATURAL_ALIGNMENT__ 1 #define __APPLE_CC__ 1175 #define __USER_LABEL_PREFIX__ #define __STDC_HOSTED__ 1 #define __DYNAMIC__ 1 #define __WCHAR_TYPE__ int #define __ppc__ 1 #define __APPLE__ 1 #define __MACH__ 1 #define __WINT_TYPE__ unsigned int #define __GNUC__ 3 #define __GNUC_MINOR__ 1 #define __STDC__ 1 #define __PTRDIFF_TYPE__ int #define __REGISTER_PREFIX__ #define __POWERPC__ 1 #define __NO_INLINE__ 1 #define __VERSION__ "3.1 20020420 (prerelease)" $ Regards, Joel Breazeale > > I tried tcptraceroute 1.5 beta1 on MacOS X 10.2.6 with the December > > 2002 Developer Tools and had some problems. I also had problems with > > v1.4 via Fink, so it may be there's something else going on here. > > > > I tried 127.0.0.1 (local loopback), 24.163.210.43 (my network interface > > IP address), and a network address (66.135.192.87) and got the same re- > > sult every time. Here the are (I quit them before 30 iterations): > > Hi Joel, > > This is typical behavior on some BSD-like systems that don't support > select(2) on BPF file descriptors. As a workaround, you can try running > tcptraceroute with the --no-select command line argument, which will force > it to check for new data without first consulting select(2) to see weather > or not it's available. If that works for you, can you please send me the > output of "tcptraceroute -dv", "uname -a" and, "gcc -E -dM - </dev/null" on > your system? Armed with that data, it may be possible to write an autoconf > check such that --no-select is the default behavior on MacOS X in the > future. > > Thanks! > -mct > _______________________________________________ > tcptraceroute-dev mailing list > tcptraceroute-dev@netisland.net > https://lists.netisland.net/mailman/listinfo/tcptraceroute-dev > _______________________________________________ tcptraceroute-dev mailing list tcptraceroute-dev@netisland.net https://lists.netisland.net/mailman/listinfo/tcptraceroute-dev
|
|