Date: Mon, 07 Apr 2014 11:24:54 +0200 From: Julien Charbon <jcharbon@verisign.com> To: freebsd-net@freebsd.org Subject: Re: panic in -HEAD multicast code Message-ID: <53426EE6.5020409@verisign.com> In-Reply-To: <CAJ-Vmo=q3dweaedFK=3nVP=YQc59PHKzCv-fPCm04yRLw%2BhxYw@mail.gmail.com> References: <CAJ-Vmo=q3dweaedFK=3nVP=YQc59PHKzCv-fPCm04yRLw%2BhxYw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Adrian, On 07/04/14 04:58, Adrian Chadd wrote: > I'm seeing a panic in the multicast code path. I reproduce this by > trying to browse network sharing on VLC. > > The panic: > > http://people.freebsd.org/~adrian/ath/core.txt.0 > > Any ideas? We believe this issue is due to a race condition in multicast code. We found it tracking down another unrelated bug: http://www.freebsd.org/cgi/query-pr.cgi?pr=185043 See provided details on this race condition in PR description section starting with: "With even this patch in place, we have further found a subsequent race condition that...". In short, it looks like a race between a thread setting/allocating a new multicast address on an interface and calling: inp_setmoptions() -> inp_join_group() -> inp_join_group_locked() -> in_getmulti() and a thread detaching this interface, purging all associated multicast addresses and calling: if_detach() -> if_detach_internal() -> if_purgemaddrs() -> if_delmulti_locked() -> if_freemulti() However, we did not find a way to fix it without unwanted side effects and we asked for comments/ideas. -- Julien
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53426EE6.5020409>