Date: Wed, 07 Oct 1998 11:21:35 -0700 From: Mike Smith <mike@smith.net.au> To: "Matthew N. Dodd" <winter@jurai.net> Cc: Mike Smith <mike@smith.net.au>, Marc Slemko <marcs@znep.com>, hardware@FreeBSD.ORG, wpaul@ctr.columbia.edu Subject: Re: Lite-On PNIC (fwd) Message-ID: <199810071821.LAA01419@dingo.cdrom.com> In-Reply-To: Your message of "Wed, 07 Oct 1998 14:02:22 EDT." <Pine.BSF.4.02.9810071353260.16487-100000@sasami.jurai.net>
next in thread | previous in thread | raw e-mail | index | archive | help
> > [Added Bill Paul to the CC:] > > On Wed, 7 Oct 1998, Mike Smith wrote: > > Bill Paul has more or less inadvertently rewritten the 'de' driver > > while writing support for the Winbond 100Mbps chips. We might want to > > go with something that's slightly less of an ifdef tangle (have you > > tried looking at the 'de' sources?), especially if the maintainer > > continues to be too busy to maintain it. > > > > It would be interesting to test these changes though. > > You're proposing to provide Bill with examples of -all- the different > varients of the Tulip cards and enough hard drugs to keep in in the state > of mind required to do the work? I was actually planning on just feeding the non-Digital-based cards to Bill, as most of them are cut-downs. We're certainly more than happy to facilitate providing hardware to developers, especially ones with track records like Bill's. 8) > The PNIC changes appear to be straight forward (though they could probably > use some un-grunging) and in line with existing code to support other > Tulip varients. Though I would speak to Bill about this first, I'm > betting the best course would be to apply some flavor of this patch and > move on to better things. I'm pretty sure that more duplicated code would > be a step backwards. It almost certainly would be. The problem is that the 'de' driver is (meant to be) maintained by Matt Thomas, who has been extremely quiet of late (due to a move and other things). We've supplied Matt with hardware in the past (eg. a Macronix card a month or so back), but his responses have been kinda patchy, and as you've seen the net result is an almost unmaintainable mess. > (I would really fear a Bill-written Tulip driver; his other code is really > clean and it would be interesting to see his solution to the Tulip mess.) Having spent some time headfirst in the 'de' dungheap, I'd be inclined to say that it really needs to be a pile of well-specified function vectors with a large heap of access and personality macros. There's just far too much "if tulip_chipid == foo" logic in there, not to mention more #ifdefs than anyone wants to see. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199810071821.LAA01419>