Revlin OS

http://p-quinn.com/

I have talked to one of the devs of this project and he admitted that they would prefer to use FreeBSD as the base, but are using Linux due to the larger(?) driver base. Does anyone have any info(comparisons, updates, etc.) showing/explaining the differences in driver support/availability for FreeBSD vs Linux?

Just so everyone knows, I brought this question to them because of their use of open web standards(licensed BSD/X11 equivalents) as the primary technology for the desktop gui. Their use of BSD as a base would allow for a much farther range of adoption of companies, organizations and individuals.

Please any help would be nice. Thank you.
 
I'll answer this.
Licensing is the reason for the difference.
Popularity is the reason for the current choice.
 
Linux supports a lot more "desktop" hardware such as audio cards, TV tuners, USB gimmicks, et cetera. IMHO those support in most cases comes from heavy kernel patching and isn't fully stable in most cases.
 
Linux will always support desktop computers much better than FreeBSD because of popularity of distros like ubundu among non specialists and amount of developers working on the kernel and it's modules.
 
I definitely have to agree that drivers in linux tend to be a bit more unstable from what I have seen. I guess it becomes an issue of quality over quantity.

What about the fact that FreeBSD supports NDIS natively in the kernel? How does that help in comparison to linux? Does anyone know? Thanks.
 
I had to recently dabble with linux to use Intel VTune to test some code. I find the drivers system a right mess, and udev is satan!

FreeBSD may have less drivers but I believe that to be better than too have millions scattered around the place of poor quality. Otherwise we might have to bring in a blacklist... Hah!

Suffice to say, just buy Thinkpad laptops and you have a very high chance of your hardware being supported (>= Linux). (I had trouble with a T43's Radeon a while back but since it's a Radeon, they were designed to be awkward)
 
Hello

My name is Patrick Quinn, im head of development on the RevLin OS project, I just wanted to let you guys know that we are currently doing some tests runa with the BSD kernel, seeing how things match up and what technologies are available that we are using in out linux based kernel. I am personally very interested in getting grand central dispatch to the platform.

I would like to extend an invitation to any developer who might be interested in helping out with the kernel and subsystem to see if BSD is a viable option,
Like i mentioned my preference would be BSD with a few linux compat layers but due to A lack of man power and B lack of desktop hardware drivers it seems hard to justify.

If a way around this can be found then it will be BSD that we use as the base of PopCorn.

Thanks for your time guys

Patrick Quinn
 
revlin said:
I would like to extend an invitation to any developer who might be interested in helping out with the kernel and subsystem to see if BSD is a viable option, Like i mentioned my preference would be BSD with a few linux compat layers but due to A lack of man power and B lack of desktop hardware drivers it seems hard to justify.

We already have a thread for that, just ask, many FreeBSD developers read these forums, we also can evaluate many questions.
 
We are moving toward BSD but the only reservation i would have is boot time. low a boot time is a very key feature. We are aiming for a boot time similar to intel's Moblin across various platforms. How achievable would quick boot be with FreeBSD?
 
Boot speeds

The aim is 10 seconds or below. It just not seen to be acceptable to have a "next gen" operating system that takes longer than this.
 
You can easily get a boottime of 30-35 sec on a relatively old machine. Hardware probing takes most of it, so I guess if you kept it to the bare minimum, you might get what you want. You can check "amateur" projects such as MenuetOS or its fork KolibriOS for examples of the minimalism we're talking about.

But it all depends on your needs and the needs of the users, and since RevlinOS is a desktop OS (as opposed to a highly secialized system made for engineers or scientists), this is going to be hard. Desktop users usually have many, diverse needs, and are rarely computer literate, so the needed customization will be much harder if not impossible to achieve, in my opinion at least.
 
Thank you for your input, i will do some experimentations with BSD and see how quickly we can get it to boot on normal hardware (1gb ram, 1.6ghz core 2 duo) and then get back to you guys with the results. As far as i can tell kolibriOS is micro kernel based?

Thanks for your time
 
They both have monolithic kernels.
What I meant is that by being amateur projects and supporting rather limited types of hardware, they can avoid very long startup time. That and being made as small as possible to fit on a floppy and being written in assembly. But once they implement some of these (especially the hardware-related entries), I dare to speculate they'll take a little more time to boot.
 
Can any one enlighten me if NetBSD would be a better option than FreeBSD?

So one over at the Haiku forums mentioned it as an option.
 
@revlin

About Your question regarding FreeBSD boot speed, on Lenovo X300 (dual core 1.2GHz) the 8.0-RELEASE boots on about 25 seconds (timer started after BIOS post), 19 seconds take hardware detection, all the USB ports and such takes time, after it was detected, its failry fast running all the startup scripts.

One of the targets of FreeBSD 9.0 (will be released about end of 2011/early 2012) is asonchronous device detection at boot, so You may follow freebsd-current mailing lists to see when (and if) it will be implemented.
 
Hmm that really seems like a reasonable time to me, is that with any optimization at all? im fairly sure that with the right amount of work that could be paired down quit significantly. :)
 
revlin said:
Hmm that really seems like a reasonable time to me, is that with any optimization at all? im fairly sure that with the right amount of work that could be paired down quit significantly. :)

Its stock 8.0-RELEASE-p2 without ANY other optimizations, but default FreeBSD 'base system' CFLAGS are -O2 -fno-strict-aliasing -pipe, so not a lot room for optimization here.

I would definitely remove floppy controllers from the kernel, they often slow down the boot process by about 1-2 secs, You can also check what sleep values are put in startup scripts and tight them up a little.

Removing exotic devices as FireWire can also speed that a bit (left in modules so no functionality is cut off).

As for other BSDs, NetBSD booted very fast for me, but I do not have any numbers through.
 
Ok we will keep our options open. Ill see if NetBSD is a worthwhile venture. Also i personally just got blasted on the Haiku forums. I was accused of waisting your time. So if any of you feel this way i am sorry. I am just looking at all available avenues for the project..
 
revlin said:
Ok we will keep our options open. Ill see if NetBSD is a worthwhile venture.
You could also try to get the newest snapshot of current work for NetBSD 6, NetBSD 5.x still has rather old sound/audio architecture comparing to oss@freebsd/oss4, but I am not sue if NetBSD already implemented that.

revlin said:
Also i personally just got blasted on the Haiku forums. I was accused of waisting your time.
Interesting, any link to that thread?

revlin said:
So if any of you feel this way i am sorry. I am just looking at all available avenues for the project..
Cant speak for others but I do not have anything against our little chit chat.
 
its the second last post at the bottom ( i go by sabianadmin there ) Here
I am happy to hear that. i hoped you didn't take offense to me asking a few questions. your all being a huge help :) .
 
Back
Top