Date: Sun, 31 Jan 2016 14:27:26 +0000 From: Tom Vijlbrief <tvijlbrief@gmail.com> To: bob prohaska <fbsd@www.zefox.net>, freebsd-arm@freebsd.org Subject: Re: Files needed to analyze kernel panic Message-ID: <CAOQrpVd4%2BAsuoC%2Bwcdi7W=nAwDX8=NLeeyM9qzPeuT01Sr%2BR1g@mail.gmail.com> In-Reply-To: <20160131012448.GB75928@www.zefox.net> References: <20160131012448.GB75928@www.zefox.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Op zo 31 jan. 2016 om 02:27 schreef bob prohaska <fbsd@www.zefox.net>: > Hi all, > > It's possible to reliably crash an RPI2 running 11-CURRENT using stress2. > In most cases the crash occurs during the syscall tests and then usually > the opening > line is of the form "translation fault L2" [sometimes L1]. > > What information is useful to collect for purposes of debugging by others > off-host? > So far I'm aware of backtrace, kernel and modules, vmcore, core.txt and > recently > kernel.debug (three places). One update/build/test/crash cycle per day is > possible. > > I can indeed reproduce the crash on my RPI2 without too much trouble with just the syscall tests. Steps to reproduce: download http://people.freebsd.org/~pho/stress/src/stress2.tgz and extract it. do a "make -i" do build the binary tests. Create a scall.cfg file with the following contents: ################ . ./default.cfg export RUNTIME=15s export TESTPROGS="testcases/syscall/syscall" export syscallKILL=1 export syscallNODELAY=1 export syscallLOAD=100 ################ run: ./run.sh scall.cfg and it will crash after 5 to 15 minutes with a translation fault panic.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAOQrpVd4%2BAsuoC%2Bwcdi7W=nAwDX8=NLeeyM9qzPeuT01Sr%2BR1g>