Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!nntp.coast.net!news-res.gsl.net!news.gsl.net!news.mathworks.com!newsfeed.internetmci.com!pier2.bayarea.net!baygate.bayarea.net!thorpej From: thorpej@baygate.bayarea.net (Jason R. Thorpe) Newsgroups: comp.unix.bsd.freebsd.misc,comp.unix.bsd.netbsd.misc Subject: Re: Curious about *BSD History Date: 26 Jun 1996 19:16:32 GMT Organization: George's NetBSD answer man Lines: 45 Message-ID: <4qs2ag$bg0@pier2.bayarea.net> References: <4k1nue$lm8@orb.direct.ca> <317006C4.77C0450E@lambert.org> <4qkegg$9as@pier2.bayarea.net> <31D0A2C9.72741EA8@lambert.org> NNTP-Posting-Host: baygate.bayarea.net Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:22367 comp.unix.bsd.netbsd.misc:3886 In article <31D0A2C9.72741EA8@lambert.org>, Terry Lambert <terry@lambert.org> wrote: >Doesn't it really want a "machine independent target range >restricted DMA soloution"? The Amiga, for instance, does >not have ISA, but has the DMA targetting problem for "fast RAM". That's not really the approach we've been taking. We've been focusing more on doing it as the bus-level. I.e. make the bus code machine-independent (done for ISA, EISA, and PCI, and partially for TurboChannel, under NetBSD), and make the DMA mapping functions bus-dependent. I think, in general, this approach is more feasible. >Yes, the generic soloution is the best. The 21066 had the ISA >bridged off the PCI instead of the other way around (like in >most PC's), and so the problem was inverted. The real issue >is generalized DMA routines that say "this will work" to the >driver coder, and then a generalized subsystem that hooks to >that interface so the Amiga and PC ISA and similar systems with >problems (ie: EISA systems with HiNT chipsets, like the old NiCE >"EISA" motherboards) can share common management code. To my knowledge, all Alpha systems with an ISA bus have the ISA bus behind some sort of bridge. The AlphaStation 600 is the same way, as are the AlphaStation 200 and 400. (If you havne't seen a 600, they're a _really_ impressive tangle of bridges :-) In fact, I've always been under the impression that even PCs indeed had the ISA behind a PCI->ISA bridge (Intel SIO), but it was the BIOS that frobbed things to make the ISA look like it is directly accessible. I could be wrong... However, I would think that if the PCI were behind the ISA, then you'd have all of the standard ISA address line lossage on the PCI, too. >I have to admit that I've been considering writing software >to make the MMU hardware pretend to be PC MMU hardware so I >can more easily track the FreeBSD VM changes on the PPC (I >known, I know, "evil incarnate"). Ick! Why is there i386 MMU-specific stuff (or assumptions) in the FreeBSD VM code? All of that should live completely in pmap! -- Jason R. Thorpe <thorpej@bayarea.net>