Date: Tue, 14 May 1996 09:19:39 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@GndRsh.aac.dev.com> To: dave@persprog.com (David Alderman) Cc: ewv@boom.bsdi.com, freebsd-hardware@freebsd.org, hsu@clinet.fi Subject: Re: Motherboards Message-ID: <199605141619.JAA04268@GndRsh.aac.dev.com> In-Reply-To: <270B9E0758A@novell.persprog.com> from David Alderman at "May 13, 96 10:08:00 am"
next in thread | previous in thread | raw e-mail | index | archive | help
> > > > > > ASUS doesn't seem to understand what a PCI-PCI bridge is and why > > > someone might want to use one. Their BIOS only detects and initializes > > > > > >I have at least 6 ASUS based routers which use SMC EtherPower 2 and ZNYX > > >4-port ethernet boards. One router has got 10 ports, all occupied. The > > >motherboard is the SIS chipset one, P90. I have got Triton II boards also, > > >but I'm not absolutely sure that I have tested those with multiple ports in > > >use yet (but they are correctly probed and both ports on EP2s are usable > > >alone). > > > > It must only be certain models then... the ones we've had trouble with > > were the P54NP4's (Neptune EISA/PCI). It isn't a BSD/OS problem, > > these wouldn't even run the DOS diagnostics that come with the > > EP2's. The I_LINE register in the second device is left programmed > > with apparently random values after POST (I've seen 0xfe, 0x20, > > and 0xff). > > > > They've probably corrected the problems in their newer boards (the > > Neptune is at least a couple of years old). > > > > - -Eric > > > Eric, there was a generic problem with Neptune chipsets not > supporting bridging. This may have been fixed in a later revision of > the Neptune chip. We have one machine in house using an ASUS board > which will not support two PCI bus-mastering devices. There was a work around for the Neptune chip set to support 2 or more bus mastering PCI devices, Intel released it to board designers way after most of them had been in full production for some time, and the fixed involved adding several pals to the board. (It replaced the PCI bus arbitier built into the chip set with an ad hoc one build out of Pal's.) > Note that this does not in any way invalidate waht you have stated > about the PIN-A etc support on the Asus Neptune boards. Even if the > problem with the chipset was corrected, it is possible that Asus has > not changed the design of the board to accomodate the newer chips. I > am nearly certain that what you described is a problem with the old > Neptune board I have in house. I don't have any doubts that the old ASUS Neptune boards are broken with respect to both PCI-PCI bridges and multiple bus mastering devices. I am unclear if the INT-A, B, C and D signals are fully routed to all slots. > Maybe Mr. Grimes would care to comment on the bridging support issue > (hint, hint!). I already commented once, but most of the brokeness in the Neptune based boards is due to the fact that the ink on the PCI 2.0 spec was not dry at the time this chipset was designed. No PCI-PCI bridge chips had completed design at the time Neptune was being done, pretty hard to make sure it would really work when you don't have a design to test it with. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199605141619.JAA04268>