From owner-freebsd-hackers Wed Nov 15 06:06:46 1995 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id GAA00220 for hackers-outgoing; Wed, 15 Nov 1995 06:06:46 -0800 Received: from skynet.ctr.columbia.edu (skynet.ctr.columbia.edu [128.59.64.70]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id GAA00190 for ; Wed, 15 Nov 1995 06:06:37 -0800 Received: (from wpaul@localhost) by skynet.ctr.columbia.edu (8.6.11/8.6.9) id JAA04385; Wed, 15 Nov 1995 09:06:30 -0500 From: A boy and his worm gear Message-Id: <199511151406.JAA04385@skynet.ctr.columbia.edu> Subject: Re: getting close - 1104-snap ed0 not working To: archive@cps.cmich.edu (Mail Archive) Date: Wed, 15 Nov 1995 09:06:29 -0500 (EST) Cc: hackers@freebsd.org In-Reply-To: from "Mail Archive" at Nov 14, 95 10:53:12 pm X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 4094 Sender: owner-hackers@freebsd.org Precedence: bulk Of all the gin joints in all the towns in all the world, Mail Archive had to walk into mine and say: > On Tue, 14 Nov 1995, David Greenman wrote: > > > > > There have been no changes to the driver code in the past 2 months. Can you > > be more specific about "it doesn't work"? > > > it will probe the card find the card but refuse to initialize the network > with an ifconfig. All I get from the ifconfig is the ability to ping my > own IP number but not my router. I can throw a linux boot disk in and > setup network with the same values and am able to NFS mount the remote > filesystem where the installation lives. I have also tried both > 10base5/10base2/10baseT on all these cards and "NONE" of them will work > under freebsd. Yet all of them are probed and will run on linux. I do not > know how much farther I can go with this. But the only factor I have > found is that all the motherboards are Intel with Neptune chipsets > P5-90's. I then got brave this afternoon and took the same hardware MINUS > motherboard (replaced it with an ASUS PCI/I-P55TP4XE board and guess what > all the ISA cards work fine now. I have a feeling that this a problem of > the Neptune chipset. > > This is all the insight I have.... > > Matthew S. Bailey > mbailey@cps.cmich.edu Hmm. Here's another datapoint for you (which may not help, since you said Linux worked, and I'm pretty sure this trouble would affect just about any OS). We got some PCI boxes from Intel a while back (part of a grant) and I remember having difficulty with installing ethernet adapters in these things until I went into the BIOS and changed a few things. These had AMI BIOSes. One of the screens (I think it was a plug & pray setup, though it didn't really identify it as such) allowed you to configure the size and location of a shared memory window (I set mine to D0000. 32K) as well as IRQ assignments. By default, the shared memory segment was disabled. I had both an Intel Etherexpress16 (ix0) and a 3Com 3c503 (ed0) and both gave me trouble until I enabled that damn shared memory zone and set the cards' shared memory configuration to match. The Intel Etherexpress card would probe, but when I tried to ifconfig it, I would get timeout errors from the kernel. The address would be installed, but the device wouldn't work. With the 3c503, the card would be detected but the shared memory test would fail. As it happened, Windows NT also had trouble on the same hardware until I set the BIOS right, but NT was a bit more tight-lipped about the problem: it gave me no errors or warnings whatsoeve -- the cards simply wouldn't work. Once I set the shared memory segment correctly however, both Windows NT and FreeBSD worked fine with either card. Both the Intel Etherexpress16 and the 3c503 require a shared memory segment to work. And I think the 3c503 is the only card for which the ed driver performs a memory test (I could be wrong: it's the first thing in the morning and my brain is still fuzzy), so you may not notice the trouble with the SMC/WD or Novell cards. The ed driver uses programmed I/O with the Novell NE2000 though, so I'm not sure why this one would fail. (Murphy's Law applies though, so you never know.) I've used the ed driver with Novell NE2000 clones, SMC and 3Com adapters on a fairly wide range of hardware (386s, 486s, Pentiums, ISA/PCI, EISA...) including my machines at home, and I've had very good results with it once I got the hardware set up right. :) -Bill -- ============================================================================= -Bill Paul (212) 854-6020 | System Manager Work: wpaul@ctr.columbia.edu | Center for Telecommunications Research Home: wpaul@skynet.ctr.columbia.edu | Columbia University, New York City ============================================================================= License error: The license for this .sig file has expired. You must obtain a new license key before any more witty phrases will appear in this space. =============================================================================