March 2014 - Darwin-kernel - lists.apple.com

EADDRNOTAVAIL when resetting IPv6 multicast interface
by Saúl Ibarra Corretgé 26 Mar '14

26 Mar '14

Hi, (Disclaimer: it's my first post to the list. I found no previous reports of this problem, apologies in advance if I managed to miss them.) While working on a platform abstraction library I've ran into what looks like a bug in the XNU kernel when setting the IPv6 multicast interface to the default value using the IPV6_MULTICAST_IF option with setsockopt. >From the manual page: "IPV6_MULTICAST_IF u_int * Get or set the interface from which multicast packets will be sent. For hosts with multiple interfaces, each multicast transmission is sent from the primary network interface. The interface is specified as its index as provided by if_nametoindex(3). A value of zero specifies the default interface." However, when passing 0 in order to reset to the default value, EADDRNOTAVAIL happens. The attached file (test-multicast-if.c) demonstrates the problem. I tested it on OSX 10.6.8 and 10.9.3 with the same result. I had a look at the XNU source for version 2422.1.72 and made a small (untested) patch (see ipv6-multicast-if.diff) which I think would solve the issue. The same problem manifests when trying to join a multicast group using IPV6_JOIN_GROUP when the ipv6mr_interface field of struct ipv6_mreq is set to 0. Kind regards, -- Saúl Ibarra Corretgé bettercallsaghul.com #include <assert.h> #include <stdlib.h> #include <stdio.h> #include <sys/socket.h> #include <netinet/in.h> #include <net/if.h> #define IFACE "lo0" int set_multicast_if(int fd, int index) { int opt; int optlen; opt = index; optlen = sizeof(opt); return setsockopt(fd, IPPROTO_IPV6, IPV6_MULTICAST_IF, &opt, optlen); } int get_multicast_if(int fd) { int opt; int optlen; optlen = sizeof(opt); if (getsockopt(fd, IPPROTO_IPV6, IPV6_MULTICAST_IF, &opt, &optlen) == 0) return opt; return -1; } int main(int argc, const char *argv[]) { int index; int fd; int r; fd = socket(AF_INET6, SOCK_DGRAM, 0); if (fd == -1) { perror("creating socket"); abort(); } /* get default index value */ r = get_multicast_if(fd); if (r < 0) { perror("getting default interface"); abort(); } assert(r == 0); index = if_nametoindex(IFACE); assert(index > 0); /* set new interface index value */ r = set_multicast_if(fd, index); if (r < 0) { perror("setting interface"); abort(); } /* get current index value */ r = get_multicast_if(fd); if (r < 0) { perror("getting interface"); abort(); } assert(r == index); /* set index value back to default */ r = set_multicast_if(fd, 0); if (r < 0) { perror("setting interface"); abort(); } /* get current index value */ r = get_multicast_if(fd); if (r < 0) { perror("getting interface"); abort(); } assert(r == 0); return 0; } --- bsd/netinet6/in6_mcast.c.orig 2014-03-26 21:48:29.847976899 +0100 +++ bsd/netinet6/in6_mcast.c 2014-03-26 21:52:04.823969044 +0100 @@ -2616,7 +2616,7 @@ ifp = ifindex2ifnet[ifindex]; ifnet_head_done(); - if (ifp == NULL || (ifp->if_flags & IFF_MULTICAST) == 0) + if (ifp != NULL && (ifp->if_flags & IFF_MULTICAST) == 0) return (EADDRNOTAVAIL); imo = in6p_findmoptions(inp); _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Two-machine GDB debugging results in frequent hangs
by John Kraft 12 Mar '14

12 Mar '14

Yes, I’ve had little luck single-stepping in kernel code as well (though to be honest I haven’t really tried it for years). At this point, I do almost all of my debugging using fwkpfv. -John _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Two-machine GDB debugging results in frequent hangs
by Joe Lowe 12 Mar '14

12 Mar '14

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (mailto:Darwin-kernel@lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/joe.lowe%40pismotechn… This email sent to mailto:joe.lowe@pismotechnic.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Two-machine GDB debugging results in frequent hangs
by Peter Gibson 12 Mar '14

12 Mar '14

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Problem with NKE iff_event_func
by Peter Gibson 12 Mar '14

12 Mar '14

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Interpreting a kernel backtrace
by Andrew Keller 07 Mar '14

07 Mar '14

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Interpreting a kernel backtrace
by Shantonu Sen 07 Mar '14

07 Mar '14

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Interpreting a kernel backtrace
by Andrew Keller 07 Mar '14

07 Mar '14

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Lock-free list atomics
by Derek Kumar 07 Mar '14

07 Mar '14

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Interpreting a kernel backtrace
by Shantonu Sen 07 Mar '14

07 Mar '14

You'd have to be more specific about the backtraces, the OS, the architecture, and what tools you are using to symbolicate Sent from my iPhone > On Mar 6, 2014, at 5:44 PM, Andrew Keller <andrew(a)kellerfarm.com> wrote: > > Hi all, > > When analyzing the backtrace in a kernel panic log, is it normal for the symbolized addresses to point to weird locations, such as open and close brackets, and unused functions? > > Thanks, > - Andrew Keller > > > _______________________________________________ > Do not post admin requests to the list. They will be ignored. > Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) > Help/Unsubscribe/Update your Subscription: > https://lists.apple.com/mailman/options/darwin-kernel/ssen%40apple.com > > This email sent to ssen(a)apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0