Date: Thu, 5 Dec 1996 13:23:30 +0100 From: se@freebsd.org (Stefan Esser) To: Terje.N.Marthinussen@cc.uit.no (Terje Normann Marthinussen) Cc: Freebsd-current@freebsd.org Subject: Re: pci-pci bridge on HP netserver Message-ID: <Mutt.19961205132330.se@x14.mi.uni-koeln.de> In-Reply-To: <199612050053.BAA17484@slibo.cc.uit.no>; from Terje Normann Marthinussen on Dec 5, 1996 01:53:22 %2B0100 References: <199612050053.BAA17484@slibo.cc.uit.no>
next in thread | previous in thread | raw e-mail | index | archive | help
On Dec 5, Terje.N.Marthinussen@cc.uit.no (Terje Normann Marthinussen) wrote: > We have a HP Netserver 5/133 LS4 > > It has 5 PCI slots on two pci busses. > The default FreeBSD kernel is not able to find the two onboard Adaptec > aic7870 PCI controllers that is on the second pci bus. > > It will find the first PCI bus, but not the second where the SCSI controllers > are. > > >From boot output: > chip0 <generic PCI bridge (vendor=8086 device=1225 subclass=0)> rev 2 on pci0:0 > > All what seems to be needed is to bounce 'pciroots' like it is done with > the Orion chipset (I now boot it by just hardcoding the right value in > pci.c). > > I could write this patch myself, however, I don't have, or know where > to get the information needed (is it enough to just bounce pciroots > with '1' or is something more general needed?). I'm not even sure about NO! Hardcoding to 1 does not work! Think about any PCI card with an integrated PCI to PCI bridge in bus 0 (an AH3940 or a 4channel Ethernet card, for example). The chips behind that bridge will be on bus 1, the second PCI bus on the motherboard will be 2 now ... > what this chipset is called. Please send me a verbose boot message log (from your patched kernel, which is able to see the SCSI chips). I will try to find out whcih Intel chip has ID 1225. There will most probably be two CPU to PCI bridge chips, and one will have the secondary bus register set to 1. If you have a card with a PCI to PCI bridge, you may try to put that into bus 0, and watch the secondary bus register of the mainboard chip become 2 ... (Look into the 0x40 to 0x5f config register area if you want to find the bus number registers yourself. Or send me a dump of those registers, and I'll try to identify the correct one. It could even be 0x48, as is used in case of the Orion. If you are running -current, there is an ioctl() interface into PCI config space. Create /dev/pci (defined in MAKEDEV in -current) and use the pciconf program (as root) to read out individual registers: # pciconf -l pci0:0:0: class=0x000000 card=0x00000000 chip=0x04838086 rev=0x04 hdr=0xff pci0:1:0: class=0x000000 card=0x00000000 chip=0x00011000 rev=0x01 hdr=0xff pci0:2:0: class=0x000000 card=0x00000000 chip=0x04848086 rev=0x03 hdr=0xff pci0:4:0: class=0x010000 card=0x00000000 chip=0x000f1000 rev=0x01 hdr=0xff pci0:5:0: class=0x000100 card=0x00000000 chip=0x88c15333 rev=0x00 hdr=0xff pci0:6:0: class=0x020000 card=0x00000000 chip=0x802910ec rev=0x00 hdr=0xff# pciconf -r pci0:0:0 0x48 0x00000000 Seems there is onthing interesting in reg. 0x48 of my Saturn II, but the output might be different on your system ... :) Regards, STefan > The only chips inside with text or numbers that looks like PCI chipsets > are S82374SB and S82375SB. From what I can see from information on the > INTEL web server, these have to do with the EISA slots. > > Anyone know what "device=1225" is named? Where to find information about it? > Want to put support for it into current? :) > > Terje Marthinussen > terjem@cc.uit.no > > > > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Mutt.19961205132330.se>