Phones, Slips and Virtualization

My cell phone had been trying to die on me for the past week or two, so over the weekend, I finally got around to getting a new phone. Since I had no real reason to stick with Verizon anymore, I looked around a bit and decided that T-Mobile had the best plans for me. I got the Samsung T809 which supports oodles of things including bluetooth, GPRS and EDGE. Although T-Mobile's EDGE network isn't huge now, they are in the process of expanding it. But, since my phone now supports data service, I decided I would also look into getting my laptop to connect to it and use it for dial-up. I have to say that setting up a dial-up connection is just as crappy as it was a few years ago… I'm pretty tempted to sit down and write some decent dial-up support for NetworkManager just to make it nicer to use my phone. Especially as dialup is (ironically) becoming more prevalent again as the wireless providers start to push their data networks. Overall, though, I'm pretty happy with my new provider and phone.

In less happy news, despite everyone's best efforts, FC5 test3 has had to slip, which I'm a bit bummed about. Unfortunately, shipping a non-final compiler really isn't an option. So, rebuild is now under way and after a little bit of a hiccup yesterday evening, it seems to be moving along quite nicely. We should be on track for test3 to be able to freeze for real on Monday. What effect (if any) this will have on the final schedule remains to be seen. I'm leaning towards saying we're better off just sticking with the schedule as it is and not push back the final by a week. I can probably be convinced either way, though.

One plus side of the slip is that it's given some time to get a few more things done. I'm been spending some time working on cleaning things up for Xen. First on this list has been getting the support for virtualization hardware (Intel's VT and AMD's SVM) working — the HVM infrastructure was merged into upstream Xen and so it was mostly pulling things down and getting them tested. Thus far, it's looking like they basically work. Also high on the list has been x86_64 support. I'm glad to say that we finally this evening got a kernel package compiled and working for an x86_64 dom0. I spent a little bit of time afterwards poking and all of the installation stuff I've done for x86 is working just fine for x86_64 as well. I'd screenshot it, but it's not all that exciting 😉 This won't make it into rawhide tomorrow as some merging needs to be done, but it should hopefully be there for Friday.

Tomorrow, I'm going to sit down and spend some quality time with some patches that are being contributed for iSCSI support in anaconda. At this point, I've reviewed them a few times so they should be to the point where I can commit them. It's still not to the point most people will want, though, as we don't yet have support for iSCSI root in mkinitrd. The path to get there is pretty clear at this point and hopefully I'll make some good progress on it as well as iSCSI could be very nice to use within a Xen guest for storage — then you really aren't tied to your host and migration becomes much more reasonable. I also need to try to give Kadischi a try again and start trying to help them in cleaning up some of their interactions with the base system.

So much to do, so little time…