Date: Mon, 23 Aug 2004 11:38:50 -0700 From: Julian Elischer <julian@elischer.org> To: Mike Jakubik <mikej@rogers.com> Cc: freebsd-current@freebsd.org Subject: Re: SMP deadlocks Message-ID: <412A39BA.8030003@elischer.org> In-Reply-To: <1599.66.11.183.182.1093285772.squirrel@66.11.183.182> References: <1569.66.11.183.182.1093284444.squirrel@66.11.183.182> <412A3590.4030003@elischer.org> <1599.66.11.183.182.1093285772.squirrel@66.11.183.182>
next in thread | previous in thread | raw e-mail | index | archive | help
Mike Jakubik wrote: >Julian Elischer said: > > >>have you the published IPI patch? >>have you preemption defined? >> > >No, i have not tried the patch yet, everything seems fine when debugging >is in the kernel. Weather preemption is defined im not sure, where is this >set? The machine is a supermicro server 6023P-8R, dual xeon, 2GB ram. I >have tried with and without hyperthreading, doesnt seem to make a >difference. > > ok, here's the patch.. http://people.freebsd.org/~dwhite/smp_rv_mtx.patch if you don't know how to turn on pre-emption you probably don't have it on... :) (it's disabled for now in i386/include/param.h from memory) > >>Mike Jakubik wrote: >> >> >>>I am also experiencing SMP related deadlocks when doing parallel compiles >>>and make -j#. I am running RELENG_5, cvsuped a few hours ago. The >>>interesting thing is that i can only reproduce the system freeze when i >>>do >>>not have any debugging enabled in the kernel. With the GENERIC debugger >>>stuff in the kernel, it does not freeze. This is a real show stopper for >>>5.3. >>> >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?412A39BA.8030003>