This is a sort of nice non-report report, cause EFI booting just works fine, as you’d hope/expect.
I always like hearing about small systems, in this case potential single-purpose DragonFly firewalls.
The cdce(4) driver has been ported to DragonFly from FreeBSD, by Markus Pfeiffer. It’s for networking over USB, whether it’s USB on both ends or Ethernet on one.
If you are using gpt(8) to format a disk, Matthew Dillon’s added a “init” option. It’s similar to ‘fdisk -Ib’, though don’t ask me how to use it because I have always been bad at manual disk formatting.
For your Monday entertainment: the boot log from DragonFly on a system with 11 sockets, 10 cores per socket, for 110 CPUs. Plus 8 TB of RAM.
(Skip past the control codes at the start)
The default kernel config for DragonFly has changed: Sascha Wildner has added the acpi, gpio_acpi, gpio_intel, smbus and smbacp devices. If you are using a custom kernel, you’ll probably want to add these. If you aren’t using a custom kernel – you should have no negative effect.
Sepherosa Ziehau’s updated the em(4) driver for Intel network cards, with the 7.6.2 release directly from Intel. For once, Intel news not related to CPU hardware bugs.
The title really says it all – if you have a Coffee Lake series Intel chipset, your video is accelerated on DragonFly.
DragonFly now has support for the Adaptec 1420. “Now” means since last month, cause I am working through my link backlog.
One side effect of Meltdown/Spectre are CPU microcode (firmware) updates. For future needs: sysutils/devcpu-data is the port that has the updates for Intel, and cpucontrol(8) is the program you run on DragonFly to add them.
I haven’t used this myself, yet, so I can’t tell you how necessary an immediate update could be – but you will probably want to use it soon.
Update: Newer CPUs might require this sizing change.
Update update: a better explanation of applying microcode updates. There’s new ones out, too. (via)
There’s been a number of commits lately from Sascha Wildner with mentions of “Tianocore EDK II” a term that’s new to me. Googling for it tells me it’s a development environment for UEFI programs, useful for anyone booting on modern hardware.
Peeter Must has added evdev support in DragonFly. It’s a ‘generic input event interface’, meaning at least at first it’s for keyboards and mice. It requires a kernel rebuild with ‘device evdev’ and ‘options EVDEV_SUPPORT’ included.
I’m far enough backlogged that Sepherosa Ziehau’s igb(4) update is already in, but as a side effect, a PC Engines apu2b4 is a good DragonFly machine.
However, if you have em(4), here’s your chance to help test.
If you are running DragonFly-current, you can get your floppy drive running again. This is actually hard to test; floppy drives are becoming an endangered species.
The ix(4) driver in DragonFly has been updated to match a new vendor release, and the faith(4)/faithd(4) driver is gone.
If you’re booting DragonFly in UEFI mode, and you have unsupported video (i.e. NVIDIA), there’s the scfb driver for xorg. It doesn’t support NVIDIA chipsets either, but it gives more options than the generic vesa driver. It appears to be present in all the BSDs to some extent.
I just wasted an hour trying to figure out why xorg had strange output but no errors on this laptop, and it’s because I had i915_load=”YES” in /boot/loader.conf instead of i915_load=”YES” in /etc/rc.conf. I’m almost nearly sure I’ve mentioned that before, but if not: here you go.
(though if you never plan to run X, you can put it in loader.conf and everything will just work.)
(Title updated for a more correct sentence)
The Areca driver, arcmsr(4), has been updated to version 1.40.00.00. This comes right from the company, too, which is very nice of them.
Sepherosa Ziehau has an update for the Realtek re(4) network driver. Try it if you have the hardware, whether older or newer.
If for some reason smartmontools seems to think your disks aren’t SMART-capable, force smartd to use SCSI ioctls. has a snippet to use in your smartd.conf for just that.