Date: Mon, 30 Aug 2004 11:19:56 -0700 (PDT) From: Doug White <dwhite@gumbysoft.com> To: Garance A Drosihn <drosih@rpi.edu> Cc: current@freebsd.org Subject: Re: 5.3-RELEASE TODO - make/kqueue Message-ID: <20040830111532.H85743@carver.gumbysoft.com> In-Reply-To: <p0611042cbd58649cefba@[128.113.24.47]> References: <200408271337.i7RDbXgu052801@pooker.samsco.org> <20040829174521.H69068@carver.gumbysoft.com> <p0611042cbd58649cefba@[128.113.24.47]>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 30 Aug 2004, Garance A Drosihn wrote: > At 5:45 PM -0700 8/29/04, Doug White wrote: > >On Sat, 28 Aug 2004, Garance A Drosihn wrote: > > > >> I do still get the "*** Signal 6"s, even though I am now running > > > with v1.76 of src/sys/kern/kern_lock.c. ... > > > >If you're sure you've updated, and have tried rebuilding make > >to eliminate a corrupted binary, then you might have hardware > >problems. > > This seems much too repeatable to be hardware. The more times I > repeat my testing, the more consistent the problem seems. (but > I'll spare you the details until I narrow it down even more...). > > I am sure I have rebuilt make several times, because I am switching > between "make WITH_KQUEUE" and make without kqueue, and I do > complete recompiles each time I switch. These signal-6's are not > coming up in "normal operation" for me. It is only when I have > been stress-testing changes to the make command. What counts as a 'stress-test'? I can attempt to reproduce it here. So far my machines build world OK, but I don't have WITH_KQUEUE set yet. It doesn't sound like the kqueue option follows the problem, though.... > Please note that I did not mean to make a big deal about these > signal 6's. Well you shouldn't be getting them, and it'd be nice to fix it now while we have a known broken case. -- Doug White | FreeBSD: The Power to Serve dwhite@gumbysoft.com | www.FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040830111532.H85743>