Date: Tue, 18 Mar 2003 00:13:20 -0700 From: Soren Harward <soren@byu.edu> To: Andrew Gallatin <gallatin@cs.duke.edu> Cc: freebsd-alpha@FreeBSD.ORG Subject: Re: AS2100 SMP clock skew -- fixed or not? Message-ID: <20030318071319.GA592@gandalf.tmmc.dyndns.org> In-Reply-To: <15990.8815.810592.704436@grasshopper.cs.duke.edu> References: <20030317191924.GA84539@gandalf.tmmc.dyndns.org> <20030317202115.A35066@freebie.xs4all.nl> <15990.8815.810592.704436@grasshopper.cs.duke.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon 17 Mar 2003 at 14:30:55, Andrew Gallatin said: > Search the arhives. I think there may be a hack you can use in there > somewhere. This is what I ended up doing: In alpha_clock_interrupt() in /usr/src/sys/alpha/alpha/interrupt.c I changed the "#ifdef SMP" lines to "#ifdef SMP_BROKEN_AS2100" (which of course isn't #define'd). I re-compiled and now "sleep 5" actually sleeps for 5 seconds instead of 10. I haven't put this through really rigorous testing yet (the machine's been up for all of a half hour now -- I'm going to let it run overnight without ntpd to see if I still get some minor clock skew). Ideally, we could look to see if DEC_2100_A500 (this particular machine type) is defined in the config to see if we need to tweak this setting, however, GENERIC comes with DEC_2100_A500 #define'd by default, along with #define's for every other machine, so we'd probably need some other flag. Of course, if there's still some clock skew, we'd just need to fix the bug instead :) -- Soren Harward soren@byu.edu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030318071319.GA592>