Hi
There is a possibility to do demand paging in kernel space address ?
thanks
----- Original Message -----
From: "Breno" <[email protected]>
Date: Sat, 15 Mar 2003 13:29:52 -0300
To: "Kernel List" <[email protected]>
Subject: Demand paging - Kernel
> There is a possibility to do demand paging in kernel space address ?
Why? I would like to know your reasons :-)
--
______________________________________________
http://www.linuxmail.org/
Now with e-mail forwarding for only US$5.95/yr
Powered by Outblaze
Hi Filipe
There is no reason , just i want know if possible.
thanks
----- Original Message -----
From: "Felipe Alfaro Solana" <[email protected]>
To: "Breno" <[email protected]>; "Kernel List"
<[email protected]>
Sent: Saturday, March 15, 2003 11:11 PM
Subject: Re: Demand paging - Kernel
> ----- Original Message -----
> From: "Breno" <[email protected]>
> Date: Sat, 15 Mar 2003 13:29:52 -0300
> To: "Kernel List" <[email protected]>
> Subject: Demand paging - Kernel
>
> > There is a possibility to do demand paging in kernel space address ?
>
> Why? I would like to know your reasons :-)
> --
> ______________________________________________
> http://www.linuxmail.org/
> Now with e-mail forwarding for only US$5.95/yr
>
> Powered by Outblaze
Uttered "Breno" <[email protected]>, spoke thus:
> There is a possibility to do demand paging in kernel space address ?
No. The entire kernel, and all of its data structures, are resident in
memory all of the time. Kernel demand paging is not possible, not
necessary and not implemented.
Tommy Reynolds <[email protected]> said:
> Uttered "Breno" <[email protected]>, spoke thus:
> > There is a possibility to do demand paging in kernel space address ?
> No. The entire kernel, and all of its data structures, are resident in
> memory all of the time.
In the current Linux kernel, that is.
> Kernel demand paging is not possible,
It is certainly possible (you could mark areas that don't contain currently
used stuff for pageout). It is extremely hairy to do right (just see the
mess with module loading/unloading, which is some of the same stuff, very
tamed).
> not
> necessary and not implemented.
The cost of doing it right in a monolitic kernel would probably outweigh
the gains manyfold, and require massive redesign for Linux. In microkernels
it is a lot easier to do (but their performance sucks baby elephants
through straws, so they are moot :-)
--
Dr. Horst H. von Brand User #22616 counter.li.org
Departamento de Informatica Fono: +56 32 654431
Universidad Tecnica Federico Santa Maria +56 32 654239
Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513
Uttered Horst von Brand <[email protected]>, spoke thus:
> Tommy Reynolds <[email protected]> said:
> > Uttered "Breno" <[email protected]>, spoke thus:
> > > There is a possibility to do demand paging in kernel space address ?
> > No. The entire kernel, and all of its data structures, are resident in
> > memory all of the time.
> The cost of doing it right in a monolitic kernel would probably outweigh
> the gains manyfold, and require massive redesign for Linux. In microkernels
> it is a lot easier to do (but their performance sucks baby elephants
> through straws, so they are moot :-)
Dr. von Brand,
Thanks for following up on this. You are right, of course: with the
source code and enough patience anything could be crafted. I answered
in the sense of "is this currently implemented", and it's not ;-)