Date: Fri, 26 Oct 2001 15:28:35 -0700 From: Mike Smith <msmith@freebsd.org> To: Matthew Dillon <dillon@apollo.backplane.com> Cc: Poul-Henning Kamp <phk@critter.freebsd.dk>, John Baldwin <jhb@FreeBSD.ORG>, arch@FreeBSD.ORG, Peter Wemm <peter@wemm.org>, Bakul Shah <bakul@bitblocks.com> Subject: Re: 64 bit times revisited.. Message-ID: <200110262228.f9QMSZv04542@mass.dis.org> In-Reply-To: Message from Matthew Dillon <dillon@apollo.backplane.com> of "Fri, 26 Oct 2001 15:09:48 PDT." <200110262209.f9QM9m739133@apollo.backplane.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> > :> Before this gets misinterpreted, the 'ticks' I am talking about is > :> not the kernel timer interrupt ticks... it's the high resolution cpu > :> or 825x ticks we get. e.g. frequency dependant on the timer we use. > : > :Matt, that is the mess Linux is fighting with. We have had a superior > :solution for years by now which even allows us to change timekeeping > :hardware on the fly as we find more suitable timebases. > > I don't consider our solution to be superior, I consider it to be a > huge mess. It's a huge hack to deal with i386-specific time counter > issues and, frankly, it doesn't even do that good a job at it. Actually, this isn't true at all. It's a fairly neat solution to the requirement that we have largely MI timekeeping. > We've > been plagued by backwards-time notifications and weird things happening > for YEARS now. This is a combination if implementation issues and flat-out broken hardware. > It is far too sensitive to environmental conditions > like laptops going into sleep mode and such. One unbelievably large > mess. Again, these are implementation issues, and shouldn't be confused with the basic design which is actually quite sound. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200110262228.f9QMSZv04542>