From owner-freebsd-hackers Thu Jun 1 05:45:06 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id FAA15649 for hackers-outgoing; Thu, 1 Jun 1995 05:45:06 -0700 Received: from Root.COM (implode.Root.COM [198.145.90.1]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id FAA15643 for ; Thu, 1 Jun 1995 05:45:02 -0700 Received: from corbin.Root.COM (corbin.Root.COM [198.145.90.18]) by Root.COM (8.6.8/8.6.5) with ESMTP id FAA04489; Thu, 1 Jun 1995 05:47:46 -0700 Received: from localhost (localhost [127.0.0.1]) by corbin.Root.COM (8.6.11/8.6.5) with SMTP id FAA02515; Thu, 1 Jun 1995 05:44:56 -0700 Message-Id: <199506011244.FAA02515@corbin.Root.COM> To: Peter da Silva cc: bde@zeta.org.au (Bruce Evans), freebsd-hackers@FreeBSD.ORG, henrich@crh.cl.msu.edu Subject: Re: Problem with 2.0.5-Alpha and SMC Elite Ultra In-reply-to: Your message of "Thu, 01 Jun 95 05:39:03 CDT." <199506011039.FAA04243@bonkers.taronga.com> From: David Greenman Reply-To: davidg@Root.COM Date: Thu, 01 Jun 1995 05:44:51 -0700 Sender: hackers-owner@FreeBSD.ORG Precedence: bulk >> >Wait a minute, how come the change? I would expect that the vast vast majority >> >of cards out there are using soft settings, why not make the hard-settings >> >folks have to go into -c and leave it for soft settings as the normal case >> >(Which is the more common case afterall!) > >> Because the old behaviour was a bug. Configuring the irq to anything other >> than "?" (-1 in userconfig) says that you really want that irq no matter >> what the soft settings say. You get soft settings for the normal case by >> saying what you mean ("irq ?") in the configuration. > >What he's saying is, I think, that the default boot kernel should use IRQ ?. It can't - more than half of the if_ed supported cards don't have a soft config capability, so changing it to ? would break them. ...besides, this would still create a mystery for cards that do have soft config but are currently configured for the hard setting. I like the new method better, but I do see a need for better diagnostics at probe time to tell the user that the irq is bogus. If it weren't for this being very difficult to get right for all cases, it would have already been done. My general sentiment is that this is the GENERIC kernel and that it's not a big deal for people to set this in userconfig (-c) - it will be made automatically permanent via 'dset' during the first normal bootup. ...or people can build a custom kernel (they should anyway). -DG