DragonFly 4.8 is officially released! Download from your nearest mirror, where it should appear in the next 24 hours. If you’re upgrading your existing install, you can use the generic instructions in the release notes or in my users@ email; whichever you click first. Don’t forget to ‘pkg upgrade’!
15 Replies to “DragonFly 4.8 released!”
Comments are closed.
Question: the release notes refer to “MFC” a few times. What does MFC mean?
paraphrasing, it means to take a fix on the current branch and copy it to one of the official release branches.
So is “MFC” another way to say “backporting”?
It seems MFC is an old term, it was used perfectly fine 10+ years ago, it basically means:
[MFC] = [M]erged [F]rom [C]urrent, Source: Michael W Lucas Book (2008)
as in, from current to stable.
It might be a FreeBSD slang.
As a side note, hey, 4.8! The hype, anybody feeling the hype? Don’t know much about DFLY but who does if there is hype! This is for the hoardes of filesystem/storage fans out there! Since it’s Hammer1 version7, V7 is the thing that gets MFCed! Hype! Phoronix will go crazy soon. Where is the benchmark?!
Always great to see a new release, though I’m surprised the mention of Ryzen support got buried at the very end of the release notes. That seems pretty significant to me.
@Guest
Re: Ryzen.
Why is support an x64 chip so surprising?
We’re hitting what looks like a cpu bug on Ryzen. Not sure if the FMA3 fix (which my 0511 ASUS BIOS doesn’t have yet) will fix it (it isn’t specific to FMA). I have to wait for at least that fix to hit the BIOS and then re-test. Validating the issue may take a while.
We’re also not parsing the cpu topology correctly for Ryzen, which is a minor issue.
Other than that there are a few other things we will be tweaking. For example AMD CPU’s have traditionally had issues with invariant TSC support so in the past we force use of the HPET for the cpu clock. But it looks like the Ryzen supports it properly so we will be able to use the TSC instead of the HPET for the cpu clock on Ryzen. And probably a few other things will come out of the woodwork.
Its an interesting cpu. I’d like AMD the Mobo vendors to move a little faster fixing the hardware issues, though.
-Matt
@Matthew
I’ve always wonder but what’s your “day job”?
Are you supporting yourself from DragonflyBSD donations or are you employeed by a company and who?
I’m more entrepreneurial, I’ve done a number of startups and small ventures over the years and will probably be doing another one soon. I have a long-running relationship with a small engineering firm, and do occasional consulting for a water district. When I’m not doing other things I spend my time on DragonFlyBSD.
The Donations are used to fund other developers. Not me. The donation pile is relatively small so we let it build up and then its available for things. The bigger problem is our developers are flung all over the world so buying stuff and shipping it through customs is non-trivial.
-Matt
Matt Dillon:
Since it sounds like you live a fairly busy life and only work on DragonflyBSD, as you state “when [your] not doing other things” – what motivates you to continue development of DragonflyBSD?
I ask because it’s not like your employer to funding you to do so.
Nor is there many people even using DragonflyBSD.
I don’t mean to sound like I’m unappreciative of your work. I certainly am.
But I’m just genuinely curious to know what keeps you going working on DragonflyBSD, since it isn’t like you’re getting rich/fame/recognition for it.
@Matthew
The community really appreciates the Dragonfly BSD team’s efforts. Your team is doing significant R&D work on numerous fronts, specially H2.
Keep up the great work. Thank you.
Well, first, I love BSD. And Linux. And all of Open Source. But especially BSD because I did some kernel programming at U.C. Berkeley in the mid 80’s on the original CSRG BSD systems running on DEC VAXen. I literally grew up on BSD from well before Linux even existed.
What I will say is that open-source in general has never employed me directly, but much of my success over the years… in at least one startup (BEST Internet) and probably half a dozen contracting and engineering jobs, I attribute to being able to leverage my expertise and use in particular of BSD as the underlying OS in the hardware used in those contracts.
So while Open Source may not directly pay the bills, it is certainly responsible for a great deal of my personal success and I take great pride in being to bring that expertise to other participants in open-source development who I associate with, that they then use to leverage their own interests and jobs.
It might be a bit difficult to describe why I like it so much… but for me, and for many kernel developers like me, the thrill is in doing something that suddenly boosts performance for X or Y to unheard of levels, enabling capability on commodity hardware that rivals that of commercial companies. I like the idea and the simple fact that open-source development forces commercial companies to improve their own offerings… much less ‘crap’ is sold these days because of open-source. Open-source forces a minimum level of capability and service on even the largest companies like Google or Apple or Microsoft. We keep them honest (or at least a lot more honest than they would be otherwise), we keep them on their toes, and our work forces prices down for thousands of products across the board because most commercial companies don’t have the expertise to throw on enough *real* value-add to the baseline to be able to charge a premium. In the entire world, only Apple and Google have been able to break free of these base line constraints and one can always feel open-source clawing them back down to earth even now :-). The world is a better place because of this effect.
-Matt
Thanks Matt.
Love what you do. I would love to see Dragonfly take over the world.
This comment section could be a separate article! Interview with MD.
“…the thrill is in doing something that suddenly boosts performance for X or Y to unheard of levels, enabling capability on commodity hardware that rivals that of commercial companies…”
In a nutshell that is what drives all the good engineers I’ve worked with. Curiousity, the need to learn, to find out what makes things tick.
The feedback loop is what makes all of this interesting: DragonFly does something that boosts performance, other BSDs look to see if it can be adapted, the whole community benefits.
Now to scrounge up some more hardware