Date: Tue, 7 Jul 1998 16:39:04 -0400 (EDT) From: Bill Paul <wpaul@skynet.ctr.columbia.edu> To: lile@stdio.com (Larry S. Lile) Cc: hackers@FreeBSD.ORG Subject: Re: Object library formats Message-ID: <199807072039.QAA18807@skynet.ctr.columbia.edu> In-Reply-To: <Pine.SUN.3.91.980707103029.18086C-100000@heathers2.stdio.com> from "Larry S. Lile" at Jul 7, 98 10:45:51 am
next in thread | previous in thread | raw e-mail | index | archive | help
Of all the gin joints in all the towns in all the world, Larry S. Lile had to walk into mine and say: [snippage] > > I would avoid trying to accomodate foreign object modules. It sets a > > bad precedent. Repeat after me: "You want source code. You want source > > code. These aren't the droids you're looking for." > > I really like having the source, but sometimes that is just not possible. Nothing is impossible. > If I can make their object code work in a driver then I will use it, > if not then I will push them for a native object or the source. I > understand their point though, they have spent a lot of money to make > good token ring cards and dont want to give away their technology > but they are making an effort to provide third party accessability. Unfortunately, this doesn't wash simply because there are many other companies in the world that manufacture hardware like this, and they don't have any problems at all releasing programming info. This includes companies that make their own silicon, not just their own boards. Remember: we're not talking about incredibly sensitive industry secrets here. We're not asking for their chip masks or the source to their own drivers. What we are asking for is decent documentation that others can use to write their own device drivers, without being smothered by an NDA. If they think their business will be hurt by providing such information, then they're wrong. And no, it won't cost tons of money of force them to reorganize their infrastructures or any such nonsense: all they have to do is exactly what they're doing now, only without puckering up their sphyncters so tightly. [chop] > Unfortunately, I don't know where to find > > the PCI vendor and product IDs for their cards: all the 'datasheets' > > on their web server are basically just marketing hype. I found one > > place with patches for the Linux driver that claim to add Olicom > > support, but in typical Linux fashion, the patches are broken: the > > patch for the header file is incomplete and lists only the Olicom > > vendor ID, whereas the patch for the C code refers to product IDs for > > the OC-2183/2185, OC-2326 and 2326 NICs. These are nowhere to be found. > > > > I asked Olicom about getting more useful datasheets via their > > 'feedback' section on the web server but haven't heard anything back > > yet. > > They were happy to help me, but it did take several days. July 2 is > not a good day to start asking technical questions, oops. Also their > development staff is in Denmark. Actually, I think I found what I needed. It turns out the patches were to add support for newer Olicom cards: the existing version of the Linux tlan driver at Caldera already had the initial support. I found the product IDs for the relevant cards in the source, and got the Olicom vendor ID from the netbsd pcidevs file. As a bonus, I also found some device IDs for some additional Compaq NICs. Of course, it would be nice if I actually had an Olicom card to test with. > > May I ask where you got the token ring device driver kit? Do they have > > it available for download anywhere or did you have to speak to someone > > in particular at Olicom to get it? Do you know if they have kits for > > the ethernet NICs too? > > The tokenring kit is at: > http://www.olicom.com > software > TOKEN-RING:Token-Ring PowerMACH Works The file they mention is a self-extracting .EXE, which is kinda useless for me since I have a SPARC IPX on my desk. :( > I have not seen an ethernet ddk, but that does not mean there isn't > one. You should call them (Corporate office) and ask about it. I > will warn you that you will get passed around and finally end up > most likely talking to the head marketing rep. in your area. The > lady I ended up talking to however was very nice and got me the > ddk (when it was not on the webserver/or the link was broken?) via > e-mail. I called the NJ office (the closest one) and left a message earlier today. Haven't heard anything yet. > > PS: The ThunderLAN chip appears to support both ethernet and token rin > > PHYs; if this is a PCI token-ring NIC thay you have, it may have > > this chip on it, in which case I would visit the Texas Instruments > > web server and obtain the programmer's guide from there. > > Why reinvent the wheel? Because your existing wheel is sealed in a decidedly non-wheel-shaped block of epoxy with part of an axle sticking out. > Besides I already looked at the two olicom > cards I have and couldn't find the chips on TI's website :( I found > some very similar numbers but no matches. (Of course I was only looking > for token ring stuff) I would try contacting TI directly then. The fact that they don't have the documentation on their web server doesn't mean it doesn't exist at all. -Bill -- ============================================================================= -Bill Paul (212) 854-6020 | System Manager, Master of Unix-Fu Work: wpaul@ctr.columbia.edu | Center for Telecommunications Research Home: wpaul@skynet.ctr.columbia.edu | Columbia University, New York City ============================================================================= "It is not I who am crazy; it is I who am mad!" - Ren Hoek, "Space Madness" ============================================================================= To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199807072039.QAA18807>