From owner-freebsd-stable Fri Jul 26 14:21:50 1996 Return-Path: owner-stable Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id OAA27733 for stable-outgoing; Fri, 26 Jul 1996 14:21:50 -0700 (PDT) Received: from Sisyphos (Sisyphos.MI.Uni-Koeln.DE [134.95.212.10]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id OAA27720 for ; Fri, 26 Jul 1996 14:21:45 -0700 (PDT) Received: from x14.mi.uni-koeln.de (annexr2-42.slip.Uni-Koeln.DE) by Sisyphos with SMTP id AA03202 (5.67b/IDA-1.5 for ); Fri, 26 Jul 1996 23:21:05 +0200 Received: (from se@localhost) by x14.mi.uni-koeln.de (8.7.5/8.6.9) id XAA00503; Fri, 26 Jul 1996 23:19:48 +0200 (MET DST) Date: Fri, 26 Jul 1996 23:19:48 +0200 (MET DST) Message-Id: <199607262119.XAA00503@x14.mi.uni-koeln.de> From: Stefan Esser To: swhite@gov.za Cc: stable@freebsd.org Subject: Re: PCI bus... or why I don't like Acer In-Reply-To: <31927694E70@jedi.wcape.gov.za.> References: <31927694E70@jedi.wcape.gov.za.> Sender: owner-stable@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Sean White writes: > Hiya, all! > > I've downloaded the 2.1.5-R boot floppy and briefly checked the Acer > machine that gave me so many problems before with 2.1.0-R ... Hi Stefan! > > The problem detecting the PCI bus is still there. Now, I'm quite happy that > it's not FreeBSD's fault at all, more the rather strange PCI setup in this > particular machine. My fix for 2.1.0-R was to compile a custom kernel which > reversed the PCI config mode probes (mode 2 before mode 1) and this has > proven to work quite successfully. I've also tested the modified kernel on > a couple of other machines including IBM and Unisys with no problems > either. Compaq Prolinea model 590 machines are still a problem but I > couldn't be bothered with those things anyway. Don't have one, don't want > one. Hmmm, they should work, too ... If anybody has problems with the PCI probe of an unpatched 2.1.5 kernel on his system, please let me know and send verbose boot messages! > My question is, would anything be broken or disadvantaged if this change > was incorporated into the FreeBSD release, obviously in a later version? > Would the "team" be prepared to consider such a change? I'll happily post a > diff for pcibus.c here if anyone is interested. Well, there are several problems with reversing the order of the probes. Did you test the boot floppy from the latest 2.2-SNAP on those machines ? (No need to do an actuall install, just test whether the boot floppy PCI probe is successful ...) The PCI probe in 2.2 is quite different from that in 2.1 and 2.1.5. It might be possible to select a mode from the boot config menu in a later version. But this is getting less important, since all currently sold systems use config mechanism 1 only. > My thanks again go to Stefan Esser for the help given in getting 2.1.0-R to > run on this machine. Well, my thanks go to those people (including you :) who actually give some feedback ... Regards, STefan