“We have now reached the Release Candidate stage of the FreeBSD 6.2 release cycle. A few significant problems had been discovered during the initial BETA testing and those issues should now be fixed. RC1 is the first of two planned Release Candidate builds. If no more significant problems are reported 6.2-RELEASE builds will be done after RC2.”
Another Job well done guys, looking forward to final.
For the announcement you can read this.
http://archive.netbsd.se/?ml=freebsd-announce&a=2006-11&m=2562174
I’m still waiting the USB get fixed.
Well, they are replacing the entire USB stack – but, it won’t be included in it (changes are too extensive) – it will debut in 7.0. Still, you can easily install it on 6.x as well.
Details here:
http://www.turbocat.net/~hselasky/usb4bsd/
Edited 2006-11-18 08:13
I’m still waiting the USB get fixed.
There are so many fun replies to this sort of post:
-I didn’t realize it was broken
-Send diffs!
-Oh, you didn’t open a PR?
…every step that brings us closer to FreeBSD 6.2 also brings us closer to DesktopBSD 1.6
I’ve installed the RC1 on my test machine where I last had 5.3 on it. Wow… FreeBSD simply magickally does what other OS products (some Linusi and all the memory garbage from MICROS~1) don’t seem to be able to: It runs faster the higher the version number gets. How do they do that? It’s impressing. I can even afford using a P133 with 128 MB RAM as a file and / or web server. For other “new” products (e. g. “Vista”) I would have to buy new x86 stuff (I refuse to do so), but FreeBSD makes older hardware beeing worth the money spent on it in the past. And the ports collection grows. Still, all the standard tools you know from your commercial UNIX systems are provided. It’s just amazing how good an OS can be. (I use FreeBSD right next to Solaris and IRIX.) Power, simplicity, portability, compatibility, fantastic documentation, standard conformity, features of the future today – I can’t be happier, but I surely will be with the next release. FreeBSD is one of the greatest x86 OSes I used at home and at work. It truly can compete with commercail UNIXes.
Great work, FreeBSD developers! Thank you very much!
Edited 2006-11-18 06:30
I’ve had no real speed issue with the 5 series, but I did find out how fast my laptop could be since I started using 4.8 on it. Before it ran Windows 2000 and I had to script and schedule maintenance to keep it from slowing down after a few months.
“I’ve had no real speed issue with the 5 series, […]”
Maybe I had sone halluzinations.
“[…] but I did find out how fast my laptop could be since I started using 4.8 on it.”
At work I use an old Toshiba T1600/40 with FreeBSD 4.1 on it – with cu(1) as a serial maintenance terminal. Shame on me, why don’t I use my museum DEC VT100? (Answer: It’s too big to place it in the filing cabinet.)
As far as I know, DragonflyBSD uses the 4.x branck and did further development on it.
Oh, legendary 4.x branch.
It just flies!
Especially on smp (irony)
> Wow… FreeBSD simply magickally does what other OS products (some Linusi and all the memory garbage from MICROS~1) don’t seem to be able to: It runs faster the higher the version number gets.
Hmm. I just tested the stable versions both PC-BSD and Kubuntu on my 1GHz/256MB laptop, and got much better results from the latter.
Opening more than 10 tabs in PC-BSD’s konqueror resulted in furious trashing, behaviour which I haven’t met in Kubuntu yet.
Therefore, I don’t think what you’re saying (except for the MICROS~1 part) is true.
Opening more than 10 tabs in PC-BSD’s konqueror resulted in furious trashing, behaviour which I haven’t met in Kubuntu yet.
You are comparing KDE 3.5.3 with KDE 3.5.5, of course you got better results with newer version. You should try out PC-BSD latest beta release instead: http://www.pcbsd.com/1.3/PCBSD-1.3-BETA2-test1.iso
KDE crashing on your computer got nothing to do with FreeBSD stability and better handling of your hardware.
Edited 2006-11-18 09:31
“KDE crashing on your computer got nothing to do with FreeBSD stability and better handling of your hardware.”
Show us some benchmarks, plz ?
You are comparing KDE 3.5.3 with KDE 3.5.5, of course you got better results with newer version. You should try out PC-BSD latest beta release instead:
This probably has more to do with the kerneln than KDE versions. FreeBSD is still using the old 4BSD scheduler inherited from 4.3BSD while Linux has a new and modern one since 2.6.
This probably has more to do with the kerneln than KDE versions. FreeBSD is still using the old 4BSD scheduler inherited from 4.3BSD while Linux has a new and modern one since 2.6.
PC-BSD kernel is built with ULE scheduler as default.
SCHED_ULE has been all but abondoned. There’s no one maintaining it. While there are the occasional workloads where it is better than 4BSD, it should not be used as the default scheduler unless you test it again 4BSD on your hardware and your workloads.
Most of the advances in SCHED_ULE have been ported to SCHED_4BSD.
There’s a new scheduler under development with lots of developers working on it: SCHED_CORE. It is in -CURRENT, but not enabled (I don’t think, don’t have a -CURRENT system). There’s talk of it becoming available in 7.0, with a look to possibly make it the default after that.
Can those who modded me down please explain why? There are three resons for modding a post down: ads, personal attacks and off topic. Which rule did I violate?
My post is factual correct. It’s no secret that the FreeBSD devs couldn't get the ULE scheduler working in time. That’s why 5.3 was reverted to 4BSD.
I would imagine you got marked down because you stated a factual inaccuracy. This is what you said:
“This probably has more to do with the kerneln than KDE versions. FreeBSD is still using the old 4BSD scheduler inherited from 4.3BSD while Linux has a new and modern one since 2.6.”
This is a blatant lie. Infact, FreeBSD is not “still using the old 4BSD scheduler inherited from 4.3BSD.” There have been major and significant improvements to the scheduler since then. Don’t believe me what you wrote is a blatant lie, try this out:
“This probably has more to do with the kernel than KDE versions. RedHat is still using the same Linux kernel built in 1991 while …”
I could also dig through the archives and explicitly state all the changes since 4.3BSD, but I don’t necesarrily think this is necessary.
BTW, don’t get me wrong. It is okay to have well-informed opinions. It is also okay to state facts that people might not like. So for example, you could easily have stated that you think the problem with speed has a lot more to do with how the scheduler handles things rather than KDE versions. Better yet, you could have presented a technical discussion with why exactly FreeBSD is not able to handle multitasking on a desktop well. However, you chose to state a factual inaccuracy.
Edited 2006-11-20 22:33
“Blatant lie”? Is FreeBSD using something else than 4BSD now? No, it isn’t! Of course it has been improved since the fork, just like any part of the system, but it’s still the same design. That’s why they’ve been working to replace it since 5.X.
Since you obviously have no idea what your’e talking about I suggest you read this PDF by FreeBSD developer Jeff Roberson about the schduler:
http://www.chesapeake.net/~jroberson/ULE.pdf
It’s a bit old but many parts are still relevant.
Wow, don’t get so emotional. I was just explaining to you why someone else might mark you down. Inferring I have no idea what I am talking about is wrong and I take insult to your comment.
In actuality, I do know quite a lot of what is happening in the FreeBSD world. I also know there are quite a lot of improvements that are being made to FreeBSD, including the scheduler. I also know the only operating system that is not undergoing fundamental improvements is a dead operating system. The fact that so much effort is being put in changing Linux, FreeBSD, and Solaris is a very good sign for all of them. Finally, I know BSD runs very well as a desktop system and a lot of recent “speed improvements” Linux users are experiencing have relatively little to do with kernel changes, but with major improvements in Gnome and KDE. These improvements will be enjoyed by everyone, even including us poor people using the same old 4BSD scheduler from 1991.
Anyways, going back to the original topic, I don’t know what you might have meant or what you might have not meant, but what you wrote was blatantly wrong.
Again, let me repeat what you said, focusing on the important part:
“FreeBSD is still using the old 4BSD scheduler inherited from 4.3BSD while Linux has a new and modern one since 2.6.”
What is FreeBSD using according to you? A modernized scheduler based on the original 4BSD scheduler? Nope. A scheduler that imported some of the ideas from the ULE scheduler? Nope. A scheduler that has been changed for SMP workloads? Nope. A scheduler that has seen quite a few improvements in the 5.x series? Nope. A scheduler that has seen incremental improvements over 15 years of existence? Nope. By your account, it is still using the old 4BSD scheduler inherited from 4.3 BSD.
As I said, I am not saying sched_4BSD is perfect, nor that it does not contribute to any problems with FreeBSD. However, that is not what you wrote. And I hate to beat a dead horse, but you wrote what you wrote and what you wrote was wrong. I can’t say anything else to convince you about it.
Finally, I did not mark you down, I am just trying to explain to you why someone might have.
Edited 2006-11-21 00:35
Ok, I’m sorry. Actually I tried to edit my post after i submited it but I wasn’t able to.
I just got angry because I felt people where modding me down because I critizised FreeBSD.
“Hmm. I just tested the stable versions both PC-BSD and Kubuntu on my 1GHz/256MB laptop, and got much better results from the latter.”
I don’t use KDE or Gnome, I prefer WindowMaker (because I don’t need a desktop, I rather need a fast and well-formed window manager). Surely the use of KDE isn’t that good with “only” 256 MB RAM. I don’t know if Gnome would be better / faster.
“Opening more than 10 tabs in PC-BSD’s konqueror resulted in furious trashing, behaviour which I haven’t met in Kubuntu yet.”
I found that if you set FreeBSD on load, it may slow down, but does not halt like some Linusi do. If you want to test it, just run “while(1) mozilla &”.
Maybe you have different results because you can consult more / different hardware. My warehouse of x86 stuff isn’t that big, I could build myself a house of Suns and SGIs instead.
“I’ve installed the RC1 on my test machine where I last had 5.3 on it. Wow… FreeBSD simply magickally does what other OS products (some Linusi and all the memory garbage from MICROS~1) don’t seem to be able to: It runs faster the higher the version number gets.”
It’s no wonder, 5.3 weren’t partculary fast.
Edited 2006-11-18 08:50
Hi,
I get my new Macbook next week and I know that 6.1 doesn’t boot up on the macbook, because either of acpi (if you enable it) or some wired error about s-ata.
So has somebody given a try booting 6.2 RC1 on a Macbook?
About the only disappointment, Intel 3945ABG support isn’t in there yet
hopefully WPI + WPA_Supplicant support will merge into the tree soon; got knows is stupid to think that there is only one issue holding me back from re-adopting it 
Yup, I had to switch to Ubuntu because I moved and had to get a new laptop… I miss FreeBSD so much
Hopefully soon…
I’ve since had a look on the OpenSolaris website; it appears there is an issue with the NDIS implementation which stops one from using the NDIS drivers with OpenSolaris – hopeing that Sun get this corrected in a timely manner, this will be merged back into FreeBSD’s version of NDIS.
With that being said, if OpenSolaris get their act together and release one in a timely manner, it will win be over going to FreeBSD – which is ashame give that I have a soft sport for FreeBSD and the ports system (which is lacking in OpenSolaris).
You mean Nexenta GNU/OpenSolaris?
Nexenta is now more of a hybrid GNU and Solaris operating system. It now contains a GNU userland but optionally can have Solaris userland tools as well.
I might just buy a Nvidia card and switch to it from SUSE once the stable version is released.
We’ll see.
Sorry, I meant the Solaris Express; I’d prefer using the complete Solaris system rather than a franktenstein fusion of GNU and Solaris, compiled by many different compilers – they reason for using Solaris is the degree of consistancy, and using GNU user space, you’ve thrown that benefit out he window.