If you are using virtio drivers, there’s no longer a need for ‘device virtio_pci’ in your kernel config. It’s autoloaded as a dependency. If you run a custom kernel, remember to take it out. You’ll want to do that now if you’re on 5.1, or later at the next version upgrade if you are on 5.0.
The default options on the math/py-numpy port slowed it down. Francois Tigeot noticed, and committed a change that takes advantage of all processors. Read his note to users@ for details.
This isn’t really a dramatic event, but Rimvydas Jasinskas has added support for DWARF-4 line number tables in binutils 2.27. I am linking it to remind everyone that a little bit of Tolkien, in the form of elves (elfs?) and dwarfs (dwarves?) lives in your computer. We need a ORC standard. Oh. Hobbit? Hobbit.
Tomohiro Kusumi has brought in exFAT support to DragonFly from FreeBSD. Useful for cross-platform drives when FAT32 isn’t enough, and NTFS brings its own problems.
Rimvydas Jasinskas created a loader.conf(5) hint that keeps various nata(4) devices from attaching during boot. This is super useful if it’s a device that screws up your boot process. and I think it’s also great if you get irritated having something in your dmesg every time about the device you never use, like a CDROM.
The ls(1) command has a lot of options (Look at the man page synopsis!). So much so, that the most recent option added in DragonFly is “_”. That’s to show nanotime as part of the -l output. This will be most useful when you have multiple files being created within the same second of time, and you want to see which came first.
I’m still catching up with the pre-2018, pre-Meltdown commits, so here’s one: Changing the staged packet count in DragonFly changes forwarding rate, for the better – up to a point. There’s probably some specific ratio in the change that makes sense, though I don’t know it.
The commands rcp(1), rlogin(1), rlogind(1), rsh(1) and rshd(1) have been removed from DragonFly. There’s a net/bsdrcmds port if you still need them… though I imagine/hope ssh is filling the void for everyone.
Matthew Dillon posted a summary of IBRS and IBPB support in DragonFly, and some numbers showing its impact in various configurations of options and CPUs.
It’s turned off and on by the sysctl machdep.isolated_user_pmap – and defaults to on for Intel CPUs. Buildworld tests show about a 4-5% performance hit, but that’s only one form of activity, measured, so there will surely be other effects.
Note that Spectre is not mitigated by this commit series, and as I understand it, cannot be realistically fixed in software.
Update: Matthew Dillon posted a summary to users@.
Update 2: He told us so.