If you happen to run out of bpfx devices, Joerg Sonnenberger has an easy solution: MAKEDEV more.
Thomas Spanjaard committed his new version (from FreeBSD, avoiding GEOM) of the ata architecture (nata). It can be compiled with this patch, though it’s not ready for use yet.
BSDStats.org sent out a summary of reporting hosts; there’s about 2,000 FreeBSD users and almost nobody else on other architectures. There’s only one DragonFly user reporting in North America, which I assume is me.
Peter Avalos has a port of OpenBSD’s dhclient program available for testing. The OpenBSD version runs with reduced privileges, compared to the current DragonFly dhclient.
Also, Victor Balada Diaz has his newest version of the jail code, allowing multiple IPs within a jail and also IPv6.
The newest DragonFly developer with commit access: Thomas Spanjaard. He’s currently working on an upgrade of the disk system.
Matthew Dillon has added some rough notes on how to port FreeBSD drivers over to DragonFly.
Matthew Dillon has removed the global VM page hash table, and replaced it with a per-VM-object red-black tree. According to his post, memory usage is reduced, with no impact on performance.
Matthew Dillon posted an in-depth explanation of how he plans to both simplify and speed up filesystem access, starting after the start of the new year.
I posted about it before, but Simon ‘corecode’ Schubert supplied more details: there’s going to be a DragonFly hackathon at the 23rd Chaos Communication Congress in Berlin at the end of the year. Visit the IRC channel #dragonflybsd on EFNet if you’d like to coordinate rooms/board/meeting up.
This could be quite good; the recent NetBSD hackathon generated a lot of results.
Dmitri Nikulin happened to post an interesting idea: complete virtualization of DragonFly. (Last paragraph of message.) It’s similar to Inferno.
BSDNews has an interesting link to a Inquirer article benchmarking the new Intel processors. While it does wander into excessive acronyms, it’s interesting that the benchmarking is done using a variety of BSDs.
YONETANI Tomokazu has a patch to update to the latest version of ACPI. Please test, if you’re running bleeding edge code, and especially if you have a laptop.
Are you going to the 23C3 (23rd Chaos Communication Congress) conference at the end of 2006? If so, there’s an informal DragonFly Hackathon planned, as a number of DragonFly developers will be there. There’s a list of potential tasks on the wiki.
Matthew Dillon reported that DragonFly Preview code (version 1.7) have been synchronized with the bleeding-edge code, as it’s been stable. Also, the 1.8 release is definitely scheduled for January, at which point he plans to have “at least a basic userland kernel binary”.
‘nega’ reports a DragonFly 1.4 has gone for most of a year without issue; good news for an operating system undergoing heavy surgery.
A fellow named Trismegistos is interested in creating an Italian BSD community; if you’re interested, contact him at tr1sm3g1st0s@gmail.com.
Something less obvious about the open source model (and, of course, the DragonFly project) is the relatively egalitarian playing field for anyone who wants to contribute. The worst thing that can happen is a rude email. Via Slashdot comes the story of how a simple menu was too long and still
wrong.
Mike Tancsa is doing some throughput testing on different versions of FreeBSD, Linux, and DragonFly. DragonFly does relatively well for a system in the middle of a dramatic change.
The new site look has been extended to wiki.dragonflybsd.org, the CVSWeb output, and the mail archive. Also, usage of the bug tracker has increased lately, with a significant reduction in the number of outstanding bugs. If you do happen to have any pending bugs reports on the tracker (which includes posts to the bugs@ mailing list), please update or close them.