A running description of activity related to DragonFly BSD.
Recently, Sascha Wildner committed a huge number of changes to the various games, bringing them in line with what’s on NetBSD and style(9). This was all put together by Ulrich Spoerlein.
I draw attention to this not because it changed anything with the games in a functional sense, but because it’s huge (450 files changed, 31450 insertions(+), 29998 deletions(-)) and because it came out of nowhere. It’s always nice to have new surprise contributions arrive.
There’s a recent libc vulnerability that appears to be present in every BSD and Linux flavor. (Nearly every? There’s a lot…) Antonio Huete Jimenez committed the fix, with instructions on how to just rebuild libc for thatupdate.
Jan Lentfer noticed a lot of errors with his vr(4) card under load. Matthew Dillon suggested some reasons/fixes, and then made a commit that may fix it. Please test if you have an older Rhine card.
As previously noted, some of Matthew Dillon’s work was potentially destabilizing. The “danger period” (I saw no issues reported) is over.
Matthew Dillon is working on moving more of DragonFly out from under the Giant Lock. This may mean some instability this week if you’re following the bleeding-edge. He’s already posted a warning and an explanation (with numbers!) of work already completed.
Matthew Dillon has refactored the lwkt_token code, for an unspecified speed improvement. He’s been doing a lot of MP-lock cleanup recently…
If you have previously tried 64-bit DragonFly on a system with more than 3G of RAM and it failed to boot, the problem is fixed.
Sascha Wildner has added -Werror to the kernel build process. Warnings will now result in an error that stops the kernel from building. If you’re a developer, this will force you to create warning-free code when doing kernel development. If you’re a user, this will result in a cleaner, more stable kernel.
If you’re running DragonFly 2.5, Matthew Dillon has changed thread and process structures, meaning that a full rebuild of kernel and modules is necessary on the next system update.
Matthew Dillon has moved the Big Giant Lock off of a whole bunch of syscalls. This should make a noticeable difference in a multiprocessing context, though I don’t have measured results to point at. (hint, hint…)
Thanks to Michael Neumann, it’s now possible to remove a drive from a Hammer volume. It’s experimental, so all the standard warnings apply.
This can’t be done on a root volume, for hopefully obvious reasons.
Did you know you a Hammer volume can span multiple disks? And that you can add extra disks later on? There’s no RAID-like features – it’s just a straight multiple-disk volume, but it works. The Hammer command to do it is now “hammer volume-add“
A number of recent changes will be important to you if you develop on DragonFly:
Matthew Dillon has made version 4 of Hammer the default; the upgrade is a relatively painless ‘hammer upgrade’ command. This new version cuts out a chunk of the disk syncs needed, speeding up Hammer disk operations.
Alexander Polakov has imported OpenBSD’s hotplugd(8). It monitors for hotplug-style events, like disk additions and removals, and executes corresponding scripts to handles those events.
It’s now possible to boot a vkernel using an NFS share as the root. Now, you can have a networked virtual system!
Aggelos Economopoulos has committed Jan Lentfer’s update of BIND to 9.5.2-P1. It fixes CVE-2009-4022, though that bug never affected DragonFly by default.
SSH, on DragonFly, now defaults to allowing root logins, but does not allow plaintext password logins. This is on new installs only, so any existing installations won’t be affected, even after upgrades. Plaintext passwords are under constant brute-force attack for some years now, so this is probably safer.
Sascha Wildner has added pkgin to the base DragonFly system. It’s still present as a pkgsrc package, so it’s manageable and upgradeable with the normal pkg_* tools. See prior discussion here for the history.