The Fiasco Team is pleased to announce Fiasco version 1.0, the first official release of the Fiasco L4 microkernel developed at the University of Technology, Dresden.Fiasco currently supports Intel and AMD IA-32 processors (i486 and higher) and implements both version 2 and version X.0 of the L4 interface, as well as several other features such as I/O flexpages and small address spaces. Also included in this release are JDB, the powerful Fiasco kernel debugger, and Fiasco-UX, a user-mode port of Fiasco to the Linux system call interface.
Fiasco 1.0 Microkernel Released
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Isn’t the Quark kernel from MorphOS based on L4?
too bad that Liedke is not with us 🙁
Its been a while since I last looked at l4-fiasco.. at that time it was nice c++ code. I wonder how many operating systems is running on top of it?
Does any know if there is a BSD on top on L4 ?
Did you know that “Fiasco” means failure in Italian? 😉
I hope this accelerates Hurd moving over from outdated, bloated mach. Die mach, die!
I hope a lot of development gets done on this microkernel, from what I read in their papers it’s a very nice system with support for hard realtime, non-blocking data structures and much more.
Good job to everyone who worked on that project!
this IS an unfortunate name..
not very confidence building
whats next “disaster c runtime library”?
Fiasco 1.0 implements the L4 X.0 spec, just as Hazelnut does. However, I had heard that clans and chiefs got dropped from the X.0 spec. What security does Fiasco 1.0 use?
Well…. Fiasco is also failure in Spanish…. so it really is a bad name IMHO. Unless they intended on alienating most latin developers/users out there from the start 🙂
Fiasco is also a failure in English.
What are you people smoking? Sure, “fiasco” means failure in many languages, but it means exactly “total failure” in English as well:
http://www.m-w.com/
The etymology of the word (origin) is from Italian and English got it through French – but I’ve got news for you – a ton of words in standard English comes from other languages. It is a STANDARD word in English, and for anyone to say: “it means xxx in _italian_” merely shows you don’t know English well enough.
As a matter of fact, I use the word “fiasco” pretty often. It is not that uncommon a word. Bottom line: no need to say what it means in italian – IT MEANS THE EXACT SAME THING IN ENGLISH. It is as much of an English word as “manage” even though “manage” also comes from Italian.
They may be smoking something, but a rant like that leads me to believe you’re sniffing something.
fiasco > noun (event)
Meaning: A sudden and violent collapse.
Broader: collapse
Synonyms: debacle
Of course everyone knows what it means. It’s fun, isn’t it? Think about it… Developing a kernel and naming it Fiasco. ROFL! Has everyone lost their sense of humor?
Since when did *everything* become about marketability? If marketoids ran the world, all the cool, fun, silly, *meaningful* names that prevade all fields (science, engineering, you name it) would be turned into bland, MTV-friendly drivel…
Fiasco, in portuguese, is failure too!!!!!
Michael Vinícius de Oliveira
~ BlueEyedOS.com Webmaster ~
obs:- new website design still in this week.
I hope this accelerates Hurd moving over from outdated, bloated mach. Die mach, die!
The Hurd isn’t moving anyhere, considering it’s almost a decade and a half old and still can’t warm boot, or support filesystems larger than 1GB.
Also, the primary advantage of L4 over Mach (context-switch free IPC) is fairly x86 specific, since it relies on the use of segmented memory. Segmented memory isn’t available on x86-64, for example.
Meanwhile, a Mach based kernel, OS X’s XNU, continues to grow in popularity…
I’m afraid L4 isn’t going anywhere…
Just today there was some mroe talk on the L4 HURD List. I do admit things are slow but I look forward to L4/HURd being full feature. If you read some of the things that HURD has, it is a very powerful system with translators and all. Also the console layer rewrite seems very impressive and shows the commitment to doing things the right way
Looks like Marcus Brinkmann, a fairly major HURD hacker, and soem others have been in discussion about how to implenment HURD on top of L4. From what I read it seemed its been off list with osme posts stating the goal for more feedback. Some of the discussion was more about the purposed driver implementation while the other was about thread IDs.
Ive lookled at Fiasco and it seemed great. The only problem back then was it was using the 2.0 API and so it is nice to see taht they are on X.0 (THough I think the spec is all the way to X.2) The advantage of L4Ka is that they are keeping more up to date with the Spec.
IIRC, “fiasco” stems from Latin “fio, fieri, factus (sum)” meaning “to become”.
Didn’t GM have a car in Brazil back in the 70’s, named something in portugese equivalent to ‘Doesn’t Run’? I remember reading a website somewhere that had a long list of these marketing mixups.
Found it. http://www.industryweek.com/Columns/ASP/columns.asp?ColumnId=400
The Hurd isn’t moving anyhere, considering it’s almost a decade and a half old and still can’t warm boot, or support filesystems larger than 1GB.[i]
*cough* Bullshit. I’m running GNU/Linux right now, but I’ll mount the partition…..
root@lionel:/# df -h /hurd
Filesystem Size Used Avail Use% Mounted on
/dev/hda3 1.4G 154M 1.1G 12% /hurd
And, doing my math, coding didn’t begin on the hurd in its current incarnation until 1994, so that’s nine years. Not a decade and a half. I’m not quite sure what you’re referring to when you claim that it can’t warm boot.
[i]Also, the primary advantage of L4 over Mach (context-switch free IPC) is fairly x86 specific, since it relies on the use of segmented memory. Segmented memory isn’t available on x86-64, for example.[i]
And you haven’t been watching what they’re saying about what the L4 port is there for. Or a possible QNX port. They’re trying to make it microkernel-independent. Besides, L4 is so small (it’s not really a microkernel, it’s more of a nanokernel), the assumption is that you just write a new kernel for each architecture with the same external interfaces. Then all you have to worry about with porting is things like big-endian versus little-endian…things that are pretty well taken care of today.
[i]Meanwhile, a Mach based kernel, OS X’s XNU, continues to grow in popularity…
XNU is exactly the wrong way to do things.
1. It’s an older version of Mach than gnumach or the hurd extensions on top of L4.
2. The BSD subsystem shares address space with Mach, which kills modularity. If you’re running XNU, you have to use the BSD layer. You couldn’t, for example, run a modified Linux kernel for your Unix services.
3. There are still APSL problems.
4. I’d venture to say that hardware support for gnumach (and the newer oskit-mach) is wider than XNU’s on the pc side of things. I don’t have a machine where the hurd has refused to run, however, I only have two pc’s where xnu will run, and those aren’t the ones I use very often.
5. The only feature that XNU has over a conventional unix is Mach interfaces. Well, and crappy performance. XNU *might* gain an advantage if the GNUStep people implement network stuff over MachIPC where it exists, but it’ll probably be slower than native unix communication. Besides, NetBSD has already implemented Mach IPC within the kernel on PPC….it might show up on other platforms. I like NetBSD better than XNU, anyway.
I’m afraid L4 isn’t going anywhere…
I don’t think that you can reach that conclusion yet.
In german it sounds much more catastrophic (Katastrophe)because of the “k” atonation. That’s why Kafka shortly denotes his desperate characters simply Mister K.
Posts by:
Linguists 13 (~59%)
Engineers 8 (~36%)
Statisticians (~4%)
Also, the primary advantage of L4 over Mach (context-switch free IPC) is fairly x86 specific, since it relies on the use of segmented memory. Segmented memory isn’t available on x86-64, for example.
I suggest you take a look at Mungi ( http://mungi.org ). Source code available.
It uses a L4 microkernel on (64-bit) MIPS R4x00 and Alpha 21×64 platforms.
Students at UNSW use a custom built MIPS R4600/4700 based development machines for their work. I suspect some DEC Alpha’s and older SGI’s may be used as well.
http://www.cse.unsw.edu.au/~cs9242/asyst/
Naturally, the real information for this debate is in the research papers: http://www.cse.unsw.edu.au/~disy/papers/index.html
I’ll tell you a name which is a real Fiasaco – “Hurd”. Hurd just sounds like “Turd” to me. Why did the GNU people give it that name? Mach suggests speed, technology etc. “Hurd”. hmm…turd.
Where do these numbers come from??? Did I get put on the engineer list???
The current translators in Hurd, last I checked (a while ago) don’t support filesystems more than 2Gb, not 1Gb.