site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com Now time to debug the panic in ipfw_chk for unknown opcode. Thanks Giuliano _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... After a day of poking into the FreeBSD sources and printf-ing here and there, I might have found the solution. Well, at least now ipfw fwd works and the system does not seem to suffer... The file to change is bsd/netinet/ip_output.c diff -u ip_output.c ip_output.c.xnu-1228.7.58 --- ip_output.c 2008-10-16 12:42:21.000000000 +0200 +++ ip_output.c.xnu-1228.7.58 2008-10-16 12:48:24.000000000 +0200 @@ -1121,7 +1121,7 @@ dst->sin_addr.s_addr) break; } - if (ia_fw) { //the arg was ia, presumably wrong as it would brake fwd + if (ia) { /* tell ip_input "dont filter" */ struct m_tag *fwd_tag; struct ip_fwd_tag *ipfwd_tag; this makes the source more consistent with the original FreeBSD one. Might have been originated by the changed name of the ia variable in the Darwin source, change that was not brought forward in the next block of code. Reported to Apple but someone in the list might know if this has any ill side-effects. This email sent to site_archiver@lists.apple.com