Date: Wed, 21 May 2008 13:52:18 -0700 From: "Jack Vogel" <jfvogel@gmail.com> To: "Neil Hoggarth" <neil@hoggarth.me.uk> Cc: freebsd-net@freebsd.org Subject: Re: [em][panic] recursed on non-recursive mutex em0 Message-ID: <2a41acea0805211352s37e6fd40maaaac90e51912947@mail.gmail.com> In-Reply-To: <2a41acea0805211326v74baa017tc03fcd329a764ef5@mail.gmail.com> References: <alpine.BSF.1.10.0805211932250.1133@neilhoggarth-2.dsl.easynet.co.uk> <2a41acea0805211326v74baa017tc03fcd329a764ef5@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, May 21, 2008 at 1:26 PM, Jack Vogel <jfvogel@gmail.com> wrote: > On Wed, May 21, 2008 at 12:11 PM, Neil Hoggarth <neil@hoggarth.me.uk> wrote: >> Hi Folks, >> >> I opened PR kern/122928 last month, describing my problems with Intel >> PRO/1000 MT adaptor on 7-STABLE, with v6.7.3 of the em driver: every >> so often the machine would get into a state where it would repeatedly >> watchdog timeout the em0 interface, and the interface would stop >> receiving packets. >> >> A few weeks ago I cvsupped a newer version of the 7-STABLE source >> tree, after v6.9.0 of the em driver was MFCed. Since I started running >> on the newer build watchdog timeouts appear to be followed by a system >> panic rather than the freezing up of the interface. I have no hard >> evidence but it "feels like" the panics are happening under the same >> sort of circumstance that was previously triggering my watchdog >> problem (basically network activity combined with CPU load). A similar >> observation was made by Yani Karydis on the freebsd-stable mailing >> list earlier in the month: >> >> http://lists.freebsd.org/pipermail/freebsd-stable/2008-May/042311.html >> >> I've now got a serial console set up, and I'm running a kernel with >> DDB support compiled in. I managed to get the information below out of >> the most recent panic. Is anyone able to use this to debug the >> situation? What else might I try to gather useful information? > > This is a helpful start, I'll stare at the code and see... sounds like some > rx lock design issue. OPPS, I think I found the problem, look for a delta on if_em.c shortly. Let me know if this fixes your problem. Jack
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?2a41acea0805211352s37e6fd40maaaac90e51912947>