Date: Tue, 30 Jun 2015 21:10:39 +0300 From: Guy Yur <guyyur@gmail.com> To: Arjan van der Velde <vandervelde.ag@gmail.com> Cc: freebsd-arm <freebsd-arm@freebsd.org> Subject: Re: STABLE "pfctl -s states" core dumps on Beaglebone Black (was -s rules) Message-ID: <CAC67Hz8vchTUgoMcsubSp7bdvWhu%2BKfZvQqsTz1Y65f1J6eA1Q@mail.gmail.com> In-Reply-To: <55762A9C.7080504@gmail.com> References: <CD502960-38E8-4336-867A-9E292ECBC438@gmail.com> <55762A9C.7080504@gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, On Tue, Jun 9, 2015 at 2:51 AM, Arjan van der Velde <vandervelde.ag@gmail.com> wrote: > Ok, I meant to say "pfctl -s states" here. Sorry for the confusion. > > -- Arjan > > > On 06/08/2015 16:54, Arjan van der Velde wrote: >> >> Hello, >> >> I'm running stable on my BBB and I noticed that pfctl core dumps when >> doing "pfctl -s rules". I am on: >> >> FreeBSD beaglebone 10.1-STABLE FreeBSD 10.1-STABLE #1 r283308 >> >> gdb doesn't seem to give me any useful information. >> >> dmesg shows: pid 4192 (pfctl), uid 0: exited on signal 10 (core dumped) >> >> and gdb: >> (snip) >> >> >> Is this a known issue? How would I go about analyzing the cause of this? The cause for the core dump is unaligned access when trying to copy struct pf_addr in print_host(). The compiler expects the struct to be aligned on 4 bytes (it contains a union of a u_int32_t array) but it is not since it is passed from inside pfsync_state struct which size is not divisible by 4. On the BeagleBone Black strict alignment fault checking is enabled. Discussion on the issue: https://lists.freebsd.org/pipermail/freebsd-arm/2014-January/007237.html I filed bug report 185617 with changes that will break KBI: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=185617 >> >> Thanks. >> >> -- Arjan Regards, -- Guy
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAC67Hz8vchTUgoMcsubSp7bdvWhu%2BKfZvQqsTz1Y65f1J6eA1Q>