On Wed, 31 Aug 2011, Iain Hibbert wrote: > Module Name: src > Committed By: plunky > Date: Wed Aug 31 18:31:04 UTC 2011 > > Modified Files: > src/sys/coda: coda_subr.c > src/sys/compat/ndis: subr_ntoskrnl.c > src/sys/dev/pci: twa.c > src/sys/dev/raidframe: rf_aselect.c rf_cvscan.c rf_decluster.c > rf_reconmap.c > src/sys/fs/ntfs: ntfs_subr.c > src/sys/kern: kern_drvctl.c kern_malloc.c subr_autoconf.c uipc_mbuf.c > uipc_socket2.c > src/sys/miscfs/fifofs: fifo_vnops.c > src/sys/netatalk: ddp_input.c > src/sys/netinet: ip_input.c ip_mroute.c tcp_input.c tcp_timer.c > src/sys/netinet6: icmp6.c in6_pcb.c ip6_mroute.c mld6.c udp6_output.c > src/sys/netipsec: ipsec_output.c xform_esp.c > src/sys/netiso: tp_inet.c > src/sys/netsmb: smb_trantcp.c > > Log Message: > NULL does not need a cast
reviewing these, I note that ip_output() is a vararg function, and I changed some instances where a NULL was cast before being passed as vararg to that However, there are many instances elsewhere where NULL is passed to ip_output() without a cast, and presumably these have caused no problems. For example, from my change below.. --- src/sys/netinet/ip_input.c:1.295 Tue May 3 17:44:31 2011 +++ src/sys/netinet/ip_input.c Wed Aug 31 18:31:03 2011 @@ -1405,7 +1405,7 @@ error = ip_output(m, NULL, &ipforward_rt, (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), - (struct ip_moptions *)NULL, (struct socket *)NULL); + NULL, NULL); if (error) IP_STATINC(IP_STAT_CANTFORWARD); and so, as NULL is already defined in NetBSD as ((void *)0) I am inclined to leave these few changes as-is. iain