Stuart Henderson <s...@spacehopper.org> writes: > On 2017/01/20 11:40, Marc Peters wrote: >> Hi, >> >> the attached patch fixes the interface index for dhcpv6 answers on >> OpenBSD. It is based on the patch from >> https://marc.info/?l=openbsd-misc&m=144067760416819&w=2 >> >> Without the patch, every request will create a log entry with no route >> to host: >> Jan 20 11:05:50 infra1-DG dhcpd: send_packet6: No route to host >> Jan 20 11:05:50 infra1-DG dhcpd: dhcpv6: send_packet6() sent -1 of 133 bytes
I hit this bug some months/years ago but forgot to fix it. :-/ >> Comments, oks? > >> Index: patches/patch-common_socket_c >> =================================================================== >> RCS file: patches/patch-common_socket_c >> diff -N patches/patch-common_socket_c >> --- /dev/null 1 Jan 1970 00:00:00 -0000 >> +++ patches/patch-common_socket_c 20 Jan 2017 09:39:19 -0000 >> @@ -0,0 +1,21 @@ >> +--- common/socket.c.orig Thu Jan 19 15:55:02 2017 >> ++++ common/socket.c Thu Jan 19 15:58:34 2017 >> +@@ -793,9 +793,18 @@ >> + memcpy(&dst, to, sizeof(dst)); >> + m.msg_name = &dst; >> + m.msg_namelen = sizeof(dst); >> ++ >> ++ /* >> ++ * For OpenBSD, needing interface index. >> ++ * The preprocessor test is added . . . >> ++ */ >> ++ #if defined(__OpenBSD__) >> ++ dst.sin6_scope_id = ifindex = if_nametoindex(interface->name); >> ++ #else /* ! defined(__OpenBSD__) */ >> + ifindex = if_nametoindex(interface->name); >> + if (no_global_v6_socket) >> + dst.sin6_scope_id = ifindex; >> ++ #endif /* ! defined(__OpenBSD__) */ >> + >> + /* >> + * Set the data buffer we're sending. (Using this wacky > > Here is the upstream code with more context. > > 789 /* > 790 * Set the target address we're sending to. > 791 * Enforce the scope ID for bogus BSDs. > 792 */ > 793 memcpy(&dst, to, sizeof(dst)); > 794 m.msg_name = &dst; > 795 m.msg_namelen = sizeof(dst); > 796 ifindex = if_nametoindex(interface->name); > 797 if (no_global_v6_socket) > 798 dst.sin6_scope_id = ifindex; > > So, setting the scope ID is exactly what they are already trying to do. > I don't really want to spend much time understanding code which they > have marked as "XXX: this is gross. we need to go back and overhaul the > API for socket handling". But it looks like they'd be expecting > no_global_v6_socket to be set on a BSD which needs the scope id. Whether no_global_v6_socket is true doesn't seem to depend on the OS: it appears true when running dhclient, false when running dhcpd/dhcrelay; see if_register_linklocal6(). This looks more like a lack of testing, as it appears safe to set sin6_scope_id everywhere. > Better if someone who already knows the code looks at it - can you > bring this up on dhcp-users or somewhere similar? I don't volunteer to move this to dhcp-users... yet. Maybe the actual fix is in our base system. It *looks like* our kernel has indeed a bug, it should use the ifindex in the cmsg. > (Also preprocessor directives should start in column 1 so they're > obvious..but this won't be an acceptable fix upstream anyway). Indeed. -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
