Date: Tue, 9 May 95 20:07:42 MDT From: terry@cs.weber.edu (Terry Lambert) To: gibbs@estienne.CS.Berkeley.EDU (Justin T. Gibbs) Cc: rgrimes@gndrsh.aac.dev.com, brian@MediaCity.Com, freebsd-hackers@FreeBSD.org Subject: Re: A question of downloading device drivers Message-ID: <9505100207.AA20621@cs.weber.edu> In-Reply-To: <199505100030.RAA11375@estienne.cs.berkeley.edu> from "Justin T. Gibbs" at May 9, 95 05:30:49 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> >This was a conscious choice for the AIC7770 support. The POST on the > > A conscious choice on the part of Adaptec or the developer? The developer. Note I did *not* say "a bad choice"... I render no judgment other than to note that the static inclusion of that code in binary form puts kernels distributed with it under obligation to the GPL as long as it remains GPL'ed code. For the CDROM distribution, this isn't a problem, but FTP code could be. > >A proof, I offer the fact the the darn things boot and run DOS without > >any special software being necessary. > > Offer the technical specs for the interface to the default microcode > and you offer an alternative (asuming we don't have VM86). Until then, > there is no choice to be made. The board must be running microcode that > we know the interfaces too in order for the kernel driver to talk to the > board. The only way we can guarantee that now is by downloading our own > from the get-go. Actually, that should be "assuming no one bothers to port NetBSD's VM86". I find the idea of importing NetBSD kernel code less abhorent than importing GPL'ed code into the kernel, with all the baggage that entails. I'm sure that someone with an AIC7770 based card finds booting less abhorrent than arguing about distribution. 8-). On the other hand, Adaptec is willing to document *not requiring a non-disclosure agreement* the default interface. It's in their programming DOC for the card alone. Failing that, the INT 13 and INT 21 BIOS is available for disassembly if you have the card installed. It just happens to be a terrible piece of single threaded crap that I would personally prefer wasn't supported (but it *is* an option, even if an unpleasent one, if the VM86 code isn't imported). Actually, if the VM86 code was going to be imported, I could point you to an FTP site that has the specifications necessary to write NetWare protected mode SCSI drivers -- which turns out to be the spec required to be able to use them. All you'd have to do would be write the glue and NLM loader routines (famous "all's ya gotta do") and you'd have a driver for every card that has a BIOS interface and which can run NetWare (you can pretty much read this as "all of them"). If you work on the VM86 now, I might be done with the NLM loader code for the ODI drivers in time for you to use that (it's mostly there already in the GCC object file tools, latest release, but I'd prefer to just use the binary documentation from the FTP site for it, since I want to put the code in the kernel: it'll have to be there for the disk drivers to be loaded by the VM86 mode disk driver). Terry Lambert terry@cs.weber.edu --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9505100207.AA20621>