Oh boy. It has been forever since I have written. I did some cool stuff in the last month, but I don't think I am going to get to any of it in this entry. I am in a kind of wierd position.
I have an airport basestation, and moved to a new apartment over break. Since I have not had DSL installed yet I decided to use it to dial into my universities modem bank. As it turns out the thing only works with a couple of premade modem scripts Apple provides, none of which were suitable. I sat down, and hacked the image around and figured out how to write my own. It works and I am happy.
Normally I would pass this kind of info on to other people (like the guy who wrote airport configurator), but I am in a really awkward spot. For various reasons I am being sent some NDA materials from apple. I have not signed the forms yet, but I have every intention of doing so. I have not seen anything I should not have, and the stuff that I am getting involved with has nothing to do with the airport stuff, but it could be messy. To top it off Apple is flying me out for an on campus interview later this month.
Not quite sure what I should do with the info I figured out. I sent an email to my primary contact at Apple to see what he thinks.
Well, it is really busy around here, end of the semester and all.
I started a port of FreeBSDs new random driver to Darwin. Currently it builds, runs, and panics. I think that is because I just converted the mutex calls, and there are apparently some semantic differences. Particularly FreeBSD supports recursive mutexes, and it appears that mach does not. I have a way to fix it, I just need some time, hopefully tomorrow or over the weekend.
So much work, so little sleep
Well, I am in Philadelphia for Thanksgiving. I was surprised I when I saw this story about ZDnet. Or more exactly I was surprised that it mentioned the /dev/random port I had done. Too bad, since that code will not be integrated, and my rewrite is a while off still, but it certainly brightened my evening a little.
My ufs work on darwin is coming along nicely, it can read and write files, though it panics when I write directories. I decided to install Darwin Intel finally so that I could have a remote debugging machine. It is currently getting past the first stage booter, and the framebuffer appears to be unhappy, I kind of expected that though...
I wiped out a drive on that machine to install Darwin, and wanted to transfer some file from it to another box first. Boy was I shocked that the md5's did not match, then I determined the Mac OS X's md5 is broken (it reports incorrect, though self-consistent sums), so I made a working copy, and submitted a bug report.
So now after two notable detours I am almost back to working on ufs. Maybe I will try to get my tulip working while I am at it, for right now I am just going to to use an EtherExpress.
It has been a good week:
I suppose that is about everything. The zip drive lets me have a nice test ground to fiddle with new fs code. It also lets me find new and amusing ways to panic the kernel (hot removal seems to have some issues...)
Still trying to catch up with everything that happened while I was away at BSDcon. I am almost caught up which means I can get back to working on useful things. Here are things that I intend to get some work on in the near future, in no particular order:
Not sure about the last one, it is one of those things I always claim I am going to do, but never seem to get around to. Some of things on that list are definately not going to happen in the near future. Hopefully while I am doing work on some of them other people will do the some of them (well, if someone paid for everything to move my systems to rackmount I might be able to make the time to do everything else on my list ;-)
Got back from BSDCon on saturday. I would have updated from there, but I had some issues openning up a reasonably secure connection.
Did some work on Darwin, fixed a cosmetic bug in the Ethernet driver for my laptop, ported some optimizations from FreeBSD for /dev/null and /dev/zero, went back and forth some on /dev/random, and got myself into the position of porting softupdates ;-)
I met a bunch of cool people there, and generally just had a great time. I am still a little out of it (stupid time difference), otherwise I would probably write more.
Went to the SIG-BSD meeting last night. We have a lot of technicly competent people this year. We decided we should consolidate our work and try to do a major project. A couple of ideas were suggested. In the short term people are going to work on a custom FreeBSD install for rescue situations that could be used to make a bootable buisness card, and working on Jon Chen's cardbus code. Seems like a good time to get involved with that, since Jon told us last night it no longer panicked on insertion of a 16bit card 8-)
Longer term I am certain people are interested in working on the Open Packages project, it just seems a little to early to get involved with it.
I said I would be more then willing to help out on anything, but that my time was mostly consumed by Darwin. I suppose I have a funny definition of priorities, since I am primarily working on Darwin, but all I have been doing is working on Linux code recently. I think that is because my main machine is gone. Incidetnally, I love Apple's Repair Status page. So much better then waiting on the phone for an hour for the same info. When it comes home I will get my priorities back in line.
Thanx to Andrew Morton who figured out some more info about what is going on with the 3c556B. Gave the info to Jon and Dave, with an appropriate patch to the BSD driver. The card mostly resets correctly now, and probably is just fine for the casual user, but it is pretty clear we are not doing things quite right. I think we should have both the Linux and FreeBSD drivers done RSN.
Spent a good portion of this weekend fiddling with the 3c556B code, seemed like the reasonable thing to do since my Powerbook is not in working order.
The upside is that I now have tracked down the reset bug, and I ported the patch to 2.2.17, and 2.4.0-test8, and sent it to the vortex mailing list for people to tear apart. Hopefully it should make it into major release streams fairly soon.
I have to concede when my name was used in the same sentence as Donald Becker's on a post to the vortex mailing list I was a little taken aback 8-)
Jeff Carr from LinuxPPC emailed me and offered me a LinuxPPC CD. Having lurked on their mailing lists for a while (2 years or so) I happen to think most people involved with it really cool. Too bad I just don't really have the time for it right now.
Now I should go to bed, too bad I won't...
New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.
Keep up with the latest Advogato features by reading the Advogato status blog.
If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!