Date: Mon, 30 Sep 1996 07:44:58 +0800 From: Peter Wemm <peter@spinner.dialix.com> To: erich@uruk.org Cc: Steve Passe <smp@csn.net>, freebsd-smp@freebsd.org Subject: Re: Tyan S1562D Tomcat2 Message-ID: <199609292344.HAA04870@spinner.DIALix.COM> In-Reply-To: Your message of "Sun, 29 Sep 1996 16:24:11 MST." <199609292324.QAA28911@uruk.org>
next in thread | previous in thread | raw e-mail | index | archive | help
erich@uruk.org wrote: > > Steve Passe <smp@csn.net> writes: > > I have 64 meg and never see it. I suspect your getting hit by this > > problem. Unfortunately I don't have a solution. Anyone have a suggestion > > for "proving" this theory? How about commenting out the invalidate-tlb calls and see how long it lasts? :-) That's about the size of what's happening to the second cpu. > I have 64MB and see something like this problem quite a bit once I turn > on the second CPU (well, if I'm doing something big like a GCC compile). > I'm not claiming it is exactly the same problem. > > Interestingly enough, it always seems particular problems are caused by > "ln" (it gets memory core dumps very often) and GCC's "enquire" simply > locks the machine solid. Hmm. I have some old 386 motherboards that lock solid on enquire, I seem to recall it was dismissed as "FERR->INT 13 linkage foulup".. Perhaps enquire is triggering a floating point IRQ-13 on the second cpu which is currently not being handled? The second cpu does not get any IRQ's at all at present, and if IRQ-13 is being generated directly within the cpu then that might explain it. > -- > Erich Stefan Boleyn \_ E-mail (preferred): <erich@uruk.org Cheers, -Peter
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199609292344.HAA04870>