Date: Thu, 1 Jun 2006 15:01:13 +0400 From: dawnshade <dawnshade@mail.ru> To: freebsd-stable@freebsd.org Subject: Re: periodical crash 6.1-REL Message-ID: <200606011501.14348.dawnshade@mail.ru> In-Reply-To: <20060601112405.Q27942@fledge.watson.org> References: <200605311022.00463.dawnshade@mail.ru> <20060601112405.Q27942@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday 01 June 2006 14:26, Robert Watson wrote: > > I have periodically crashes 6.1-RELEASE running on Intel MB S875WP1 > > > > in messages i see: May 31 06:14:50 mail savecore: reboot after panic: > > page fault > > > > i installed debug kernel and got a core, but can't understand what > > happens. results of backtrace : > > There's also an existing PR for this, 97095. =A0This problem is believed > fixed in 7-CURRENT, but it requires fairly significant changes in the > socket/netinet code, which are scheduled for merging in about a month and= a > half. =A0I'll investigate whether there are any easy work-arounds that ca= n be > applied in the mean time. =A0It is made a little tricky because the > ip_ctloutput() path is shared by varius netinet consumers, and the only w= ay > to prevent a change in the so->so_pcb pointer in 6.x is to hold the pcbin= fo > lock, which requires you (ip_ctloutput()) to have more information about > its calling context than it currently does. =A0So this will definitely be > fixed in another month or so, but I'll see if I can find something in the > short term that will do the trick with a bit less disruption. Thanks for reply. I my case it reproduced with Exim (~5-6K emails/day) running on 6.1-RELEASE. Thanks again, waiting for MFC.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200606011501.14348.dawnshade>