Date: Mon, 11 Aug 97 19:21:26 -0700 From: "Mike Burgett" <mburgett@cmnsens.zoom.com> To: "hardware@freebsd.org" <hardware@FreeBSD.ORG>, "Randy Berndt" <rberndt@nething.com> Subject: Re: K6-200 Has anyone successfully done a 'make world' ? Message-ID: <199708120221.TAA06191@dragon.cmnsens.zoom.com>
next in thread | raw e-mail | index | archive | help
On Mon, 11 Aug 1997 20:56:42 -0500, Randy Berndt wrote: >>Hmmmm. What started off as an annoyance, is getting *really* interesting. I >>wonder if 2_2_RELENG is somehow incompatible with the chip? AMD claims >>compatibility with FreeBSD 2.0 on their compatability list, but I wonder if >>that included rebuilding all the sources? Seems to run fine when it's just >>perking along doing normal stuff'n quick builds... > >Pardon me for being REALLY dense, but .... Isn't the K6 supposed to be >compatible with the Pentium >> CHIP << ??? Not a program that operates on >that chip??? Sure it is, and AMD publishes a list of software that they've tested to try and prove that point. >If Freebsd uses pentium instructions, how can it become incompatible? Well, therein lies the rub. It uses AMD's implementation of the pentium instructions. Possibly there's bugs, hardware flaws, or even a subtle difference in things that happen atomically and things that don't that aren't specified (in the published instruction set) that could be causing the problem. Looking through the kernel sources, there's reasons for indentifiying different compatible chips, sometimes running on specific hardware, that needs things done or not done in the kernel, or just during init. I've seend differences in failure times, in running a 'make world' depending on what flags I gave the npx in my kernel config file, as well as things like the AUTO_EOI? options... So far, I haven't found a combination that gets me beyond about 1.5 hours of a 'make world'.... --Mike
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199708120221.TAA06191>