From owner-freebsd-mobile Sun Oct 26 17:13:50 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id RAA03345 for mobile-outgoing; Sun, 26 Oct 1997 17:13:50 -0800 (PST) (envelope-from owner-freebsd-mobile) Received: from ns.mt.sri.com (SRI-56K-FR.mt.net [206.127.65.42]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id RAA03340 for ; Sun, 26 Oct 1997 17:13:45 -0800 (PST) (envelope-from nate@rocky.mt.sri.com) Received: from rocky.mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.7/8.8.7) with ESMTP id SAA08850 for ; Sun, 26 Oct 1997 18:13:37 -0700 (MST) Received: (from nate@localhost) by rocky.mt.sri.com (8.7.5/8.7.3) id SAA28015; Sun, 26 Oct 1997 18:13:37 -0700 (MST) Date: Sun, 26 Oct 1997 18:13:37 -0700 (MST) Message-Id: <199710270113.SAA28015@rocky.mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: mobile@freebsd.org Subject: Real hackes run -current X-Mailer: VM 6.29 under 19.15 XEmacs Lucid Sender: owner-freebsd-mobile@freebsd.org X-Loop: FreeBSD.org Precedence: bulk To quote a famous person (hi PHK!). In any case, I made *lots* of changes this weekend, although most of them aren't *major* functionality additions (more below). Hopefully it'll pave the way to see some more funcationality additions, though. But, before that happens, I need people to test out the code. I don't anticipate many problems, but there's always someone whose computers quits working. You *MUST* re-config your kernel in order to get things working, since there are new files, and the options have changed (new filenames are generated.) What I did this last week (in no particular order): 1) Re-named a bunch of things. This of course won't make *any* difference in functionality, but it will make things easier to follow. For example the old code had pccard_drv and pccard_dev, and sometimes it was easy to get the two confused. Other misc. cleanups went in as well to make the code hopefully easier to maintain. 2) Only try to suspend/resume 'active' cards. In the past it attempted to suspend/resume any card that was ever active in the system. It does this no more, and only does operations on 'currently' active cards. 3) IRQ's are now allocated from the bottom up. I had changed it to be more like Win95, but at least one very popular laptop had problems doing it that way (the ThinkPad 560). Until we can figure out why, I reverted to the old code. 4) Gave up trying to do suspend/resume. Basically, everything I tried to get things working on my laptop at resume time *except* pretending the card was inserted from scratch. More works needs to be done here, but the code that's in place should work on at least *some* laptops (it works with mine). Pay attention to changes here. 5) pccard_beep is now a new file which contains the 'beeping' code. I liked bringing the code out like was done in PAO, but I didn't like how it was done, and where things were placed. I think this is cleaner, and affects things less. 6) Linker sets are used to register PCCARD drivers. This was taken directly from PAO, and was a *great* idea, and one that I should have used a long time ago. But, fishing season was on. :) :) 7) The aic6360 driver was updated with PCCARD support. So, the Adaptec SlimSCSI is now supported (at least the one that is distributed with my Sony CD-ROM works, other may need tweaks to /etc/pccard.conf). My todo list is: 1) Work on the suspend/resume stuff until we can get things working reliably on *all* of the systems. If suspend/resume doesn't work, I want to know about it so we can make it work. If it works, I want to know about it as well! 2) Bring in additional drivers from PAO. The current list of supported drivers is pretty weak, and I'd like to 'upgrade' our current drives (if necessary), and bring in 'safe' changes from PAO that don't affect desktop stability/functionality. 3) Look at the PAO code to see how to support PCI/CardBus machines. I now understand what's done there, and I think with some slight changes it can be brought into a maintainable state in FreeBSD. I'd like some help/feedback on this though, since I know *nothing* about PCI or CardBus. Once that is done, FreeBSD won't be as easy to setup as PAO is now, but we should have almost all of the functionality (modulo the desktop destabilizing patches). They have done lots of things to the user-land pccard code, but in the opinions of the FreeBSD developers we need to completely re-think and re-write the user-land code, so bringing in more junk into a pretty hairy piece of code will only make things more difficult in the long run. I welcome dissenting opinions though. None of the above is really too hard to do, or involves much brain power, it just takes time. If anyone has some and is willing to do the work, send me some email, or better yet reply to the list and we can get on it. Who knows how long my current 'spurt of energy' will last, so take advantage of it while you can. *grin* Nate