Kernel desgine(probbaly a stupid question?)

The place to bring up any design issues, or post your own creations

Moderator: Moderator Team

Locked
frog-o
Posts: 10
Joined: Fri Dec 23, 2005 6:58 pm

Kernel desgine(probbaly a stupid question?)

Post by frog-o » Fri Jan 27, 2006 8:38 pm

If some kind devolpor has the time can you please tell me,

Has anyone every gave it a thought to try to change an exiting linux kenel to use windows and linux drives?

Would this be posibble?

if so ,would this not help bring the rose project online quicker sine you are not wirte the whole os?

Would it not allow you to foucus more on wine?

Would you be able to use one of the source code for the ros installer from another linux distrubution such as onebase, mandrake etc?

Note:
I not much of a devoloper and know nothing about write drivers or an os or , where to begin ether, so this might sound stupid but i can't help woundering,

oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm » Sat Jan 28, 2006 5:24 am

Yes this is slightly of topic of this place but I will answer it.

Yes developers have thought of it. ndiswrapper. Captive. And Windows 3.11 printer drivers under Wine.

Internals of the Linux kernel are a lot differnet to Windows. To use Windows Drivers under Linux you need a wrapper of some form.

Sad part is that Captive has been harmed by Reactos crash.

This could be expanded with time. ndiswrapper is the most like what you are talking about. Ndis allows to use most Windows Network Card Drivers Under Linux.

This really does not help the reactos project much. This has more overhead.

frog-o
Posts: 10
Joined: Fri Dec 23, 2005 6:58 pm

Post by frog-o » Sat Jan 28, 2006 4:56 pm

Thanks for you reply, I did make it sound of topic, sorry my fault :( .
But it is not of topic when you know what i was thinking. What I was geting at
and thouse suggesting is that we leen reactos more toward this method(writing wrappers driver for windows under linux).

This has many bennifit if i not mistaken wich I list bellow. Please feel free to corect me. I'm not an expert on this type of thing after all.

-No more rewrite soucecode for the install program for the os (or any other on nesssary code, for linux would have it).

There are many linux distribution that have an install program and as far as i see most do they even have all the bell and wisel we ever need, such as reformating partion and creating new ones mandrake).

-all window driver wrappers for linux in one spot ,encourging devolpment.

I don't know about you but i hate look all over the web to find program that would use my windows drive a centeral source is needed. we sould combine the NdisWrapper and Captive drives in one spot makeing it easer to find.

-get reactos svn going sooner

since you not rewrite the os, but the driver, you have a stable os to use as soon as you choose a linux distribution to use and copy it to svn. I would suggest onebase linux if we go this way if possible. I would think this would be a matter of weeks not months wich it would take other wise.

-Probbaly more wich i can't explain or think of.

If nothing else we could create what i suggest above as a fork of reactos and have it like reacos is to wine write now, Wroking closely together to form the perfect windows compatible os.

oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm » Mon Jan 30, 2006 3:28 am

Project of Wrappers for Linux really are not a Reactos Project.

Yes the exist. Yes it would be nice to see a central place. At this time it does not really fit with Reactos.

Reason Reactos is trying to have a Windows Style Kernel. Advanages no wraping to loss speed.

Wine project is basicly the Linux/Freebsd Side. What advantages would Reactos have over its relitive Wine on Linux. None.

Infomation Reactos Creates but running Windows Drivers inside a real kernel makes driver wrapper makers life better. More complete infomation.

More likely for a Central Wrapper point be linked to Wine than Reactos. This still might not fit.

The Audit will just take time.

frog-o
Posts: 10
Joined: Fri Dec 23, 2005 6:58 pm

Post by frog-o » Mon Jan 30, 2006 8:26 pm

Project of Wrappers for Linux really are not a Reactos Project

Yes the exist. Yes it would be nice to see a central place. At this time it does not really fit with Reactos.
Yes it dose not fit totally with Reactos that the reson i suggested a fork.
Reason Reactos is trying to have a Windows Style Kernel. Advanages no wraping to loss speed.
Personnaly i think the lose in speed would be very minimal becuase you only use drivers that linux dose not support for such devices as winmodem.
Wine project is basicly the Linux/Freebsd Side. Reactos have over its relitive Wine on Linux. None.
I would disagree with this statment. It may be true in the futter but not now. I also think you said it in reverse (wich i have a bad habit of do my self all the time. :) )

Just to make sure i understand you ment to say this right,

What advantages would wine have over reactos. None.

To me Reactos is hilly alpha where as linux is very stable, this make working on wine easyer if you devolop wine in the os enviorment you are working on, wich is very often the case. People that use linux like build linux in linux for example. I think pepole working in reactos would want a stable reactos to work on (I know i do.) an thous the reson for the fork.

In other word
I think that my way would encourage devolopment (on reactos brother, the fork i want implemented, if pepole don't mind working on that sort of thing) and thouse reactos itself(sine it would be like reactos is to wine right now.
Infomation Reactos Creates but running Windows Drivers inside a real kernel makes driver wrapper makers life better. More complete infomation.
This might be the case but is it not like comparing reactos with wine and refuseing to create a reactos port.

what is the big deference in these 2 types of wrapper you are still rewrite the hole wrapper for window drivers i would think they be very simaler.
More likely for a Central Wrapper point be linked to Wine than Reactos. This still might not fit.
Again a reson for a fork.


P.S. I sure this is getting frustrating for you as it is me and I again thankyou for you time put in . It is very much like.(i can't spell apprecated)

Do you think we could create a pole to see who would be interest in working on such a fork or would this be a dumb idea. If theres no one even interest in this type of devolopment then there no point to this conversation. I think the title of this tread throwing thing off right now.

oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm » Tue Jan 31, 2006 1:42 am

Reactos will stablise it will just take time. Wine is Stablised it just took 10 years. Alpha to Beta Moving towards Stable. Reactos should not take that long. Some of Reactos Stable status is dependent on Wine.

My statements are long term not short.

If reactos is running a Linux kernel. Its really just Wine in that case.

Wrappers cost between 1 to 10 percent performace on the Windows Drivers runing under Linux. Reactos will not loss this 1 to 10 percent with windows drivers.
The NTFS driver under Reactos is lossing speed because its a Wrapper the other way linux->windows.

Really we enough projects. Forking Reactos only splits resources.

Wine Mailling list is most likely a better place for this.

One of the key parts to run Drivers is ntoskrnl.exe the kernel. Wine is developing this as well. Difference this kernel is built to run on top of BSD and Linux not directly to hardware. Really suits wrapper development better than Reactos less speed loss.

Now depending on what space you want the driver to run in kernel or user. You might be better off following Ndiswrapper. Making wrappers that support enough of the ntoskrnl in the linux kernel to run the driver. Now this does not fit with Wine or Reactos. Ntoskrnl interface provided directly inside a linux kernel has almost nothing in common with a Ntoskrnl buit to run on real hardware. There are more chances that the Wine Ntoskrnl and running inside the linux kernel might have common code.

It just seams like spliting effort to me forking where you get code to use makes sence. Forking and getting no code makes no sence. I don't know what Wine's long term plan is with Driver compad.

middings
Posts: 1008
Joined: Tue May 07, 2013 9:18 pm
Location: California, USA

Re: Kernel desgine(probbaly a stupid question?)

Post by middings » Wed Jan 21, 2015 10:17 am

"The main goal of the ReactOS® project is to provide an operating system which is binary compatible with Windows. This will allow your Windows® applications and drivers to run as they would on your Windows system. Additionally, the look and feel of the Windows operating system is used, such that people accustomed to the familiar user interface of Windows® would find using ReactOS straightforward. The ultimate goal of ReactOS® is to allow you to use it as alternative to Windows® without the need to change software you are used to."
--from "ReactOS Project" on the ReactOS.org Home Page
The ReactOS Project participants support this goal, Wine Project participants support their project's goals.

The interesting questions frog-o asks are answered in the ReactOS User FAQ.

The answers to frog-o's other questions are "No" and "Not possible". Those questions have been asked before. If there was an easier way to achieve the ReactOS Project's goals the developers would choose it.

EmuandCo
Developer
Posts: 4334
Joined: Sun Nov 28, 2004 7:52 pm
Location: Germany, Bavaria, Steinfeld
Contact:

Re: Kernel desgine(probbaly a stupid question?)

Post by EmuandCo » Wed Jan 21, 2015 11:38 am

Ca you explain me, why you answer to a post which is dead since 2006??
Image
ReactOS is still in alpha stage, meaning it is not feature-complete and is recommended only for evaluation and testing purposes.

Locked

Who is online

Users browsing this forum: No registered users and 3 guests