A running description of activity related to DragonFly BSD.
I just removed old pkgsrc binary packages for DragonFly 2.6/2.7 from avalon, so if somehow you are running a version of DragonFly that old, and still using binary packages, you’ll want to upgrade. I’m pretty confident that describes nobody.
Also, I have plans for coordinating the next pkgsrc release of 2012Q1, due April 6th, with the probably next minor upgrade of DragonFly, 3.0.3. I wrote out my plans already, so go read. (plus followup)
John Marino has changed the default search path for ldconfig; it no longer looks along /usr/lib/gcc* since that’s already included via rpath. The end result: you will need to do ‘make upgrade’ after your next buildworld build/installworld.
I’ve seen a few people complain about poor video performance in DragonFly, in Xorg. If you see a bunch of “contigmalloc_map: failed …” errors in your dmesg, your video card needs more contiguous memory allocated. Set vm.dma_reserved to 32M in /boot/loader.conf and you should be set. If that doesn’t work, try 64M.
A torrent for DragonFly 3.0.2, found via Google search. Which ISO or img files does it include? I don’t know. Which architectures? I don’t know. Is it legit? I don’t know. Click at your own risk, just like any other link.
DragonFly 3.0.2 is out, and you can update (see /usr/src/UPDATING) an existing install or download a new one. This release turns off I/O APIC when booting in a VM because it caused issues for some users.
Student applications for Google Summer of Code (and DragonFly) can now be submitted, until April 6th. Now’s your chance!
Matthew Dillon has posted a link to the errata for the AMD CPU bug that he found. Venkatesh Srinivas has followed with a test case for the bug.
Matthew Dillon also pointed out there’s a workaround to fix it, with no performance impact, it’s only found on revision 10h CPUs (not Bulldozer), and it’s extremely hard to duplicate. Why draw such a heavy line under that? The news of this bug rippled out through various news sites and was almost universally misreported, in a way that made it look bad for AMD without actually realistically quantifying the problem. Remember, it took 6 months just to find it – and he was looking for it!
This report from yours truly is using pkgsrc-current, so it reflects some of what will show up in pkgsrc-2012Q1. John Marino has already fixed some of the “top breakage” items, so the numbers should be even better for the next one…
If you’re running bleeding-edge DragonFly (meaning version 3.1), you will need to do a full buildworld on your next update. ‘make quickworld’ will appear to succeed but the kernel won’t work.
If you’re running DragonFly 3.0.x, this does not affect you.
Konrad Neuwirth is running Apache inside a jail, and getting some weird errors. Obviously I don’t know the fix, but Chris Turner knows what the settings need to be.
We have pkgsrc binaries still around for DragonFly 2.6/2.7. As I posted, I’d like to get rid of them. Would that inconvenience anyone?
We don’t have a set expiration policy. We probably should.
Carsten Mattner wrote out his notes on EFI booting on a Mac. This gets you closer to booting DragonFly on there, but I don’t think it is completely working yet.
Update: Carsten Mattner has a better summation than what I wrote.
If you do, acpi_hp could use some testing. Sascha Wildner just brought some improvements in for that module. I’ve seen discounted HP laptops show up in various places, recently.
Here’s the page, with a convenient mentor application note at the bottom. That’s the next step, so if you were thinking of mentoring, now is the time.
John Marino’s updated DragonFly’s version of GCC 4.4 to 4.4.7, apparently the final version of GCC 4.4. What’s next? I imagine GCC 4.6 at some point. It’s always a fun (maybe bikeshed-ish) conversation on which compiler to install, and which to have in base.
For the curious, I recently sent a bulk build report for pkgsrc-2011Q4 to the lists. Other than ruby-193 (which is fixed in pkgsrc HEAD thanks to John Marino), we’re looking pretty good! I’m curious if KDE or Gnome could actually get installed via binary; that’s sort of an ultimate goal due to the number of packages involved.
Speaking of Ruby, the default in pkgsrc may change soon, along with some of the involved Rails packages.
If you’re trying DragonFly 3 in a virtual machine, you may have noticed some issues in booting in (for instance) Qemu. Sepherosa Ziehau committed a change that sets the sysctl hw.ioapic_enable to 0 in virtual environments. It can always be turned back on, but the recent MSI/MSI-X improvements seem to cause trouble in some virtual environment. You can also set that tunable at boot to get an initial install going.
(I haven’t had trouble in Virtualbox or VMWare, so you may or may not need this.)
As several people have told me, there’s benchmarks of DragonFly 3.0 vs. 2.10, available on Phoronix. CPU performance shows a significant improvement, in tests that actually test it. (I’d think a file compression test would be disk-limited, for instance.) Disk performance isn’t as great, but that may be in part because Hammer no longer will starve reading to benefit writing; that makes benchmarks look worse but improves real-world interactivity. I’m sure there’s more quibbling to do, since it’s lies, damn lies benchmarks.