Date: Fri, 16 Jul 2004 00:47:28 -0500 From: Jon Noack <noackjr@alumni.rice.edu> To: Robert Watson <rwatson@freebsd.org> Cc: Eric Anderson <anderson@centtech.com> Subject: Re: ACPI degrading.. ? Message-ID: <40F76BF0.1000108@alumni.rice.edu> In-Reply-To: <40F752A1.8020705@alumni.rice.edu> References: <Pine.NEB.3.96L.1040715234852.50858B-100000@fledge.watson.org> <40F752A1.8020705@alumni.rice.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On 07/15/04 22:59, Jon Noack wrote: > On 07/15/04 22:49, Robert Watson wrote: >> On Thu, 15 Jul 2004, Jon Noack wrote: >>> Come to think of it, all of my hard lockups were when Firefox was >>> running. I think (some of) the issues native preemption uncovered >>> are related to threading. In any case, I commented out "#define >>> PREEMPTION" in src/sys/<ARCH>/include/param.h and rebuilt my kernel >>> so it would stay up for more than 30 minutes. >> >> I've also drawn the same conclusion -- to trigger the hangs, I run MySQL >> with a threaded benchmark. Otherwise, things seem fairly stable on my >> test boxes. Could you try running with PREEMPTION and using libthr >> instead of libkse and see if you get the same result? > > I'm rebuilding my kernel with PREEMPTION and will add the following > "global" values to /etc/libmap.conf before I reboot: > libpthread.so.1 libthr.so.1 > libpthread.so libthr.so > > I'll let you know how things pan out -- shouldn't be too long. Worked great with Firefox until I started XMMS. Everything was fine when I started playing an mp3, but it locked up hard when I switched to another desktop (had a Firefox window on the desktop I was switching to; it painted the Firefox widgets but hadn't repainted the webpage I was viewing before the lockup). Looks like it's thread library independent too. Jon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40F76BF0.1000108>