If I don't active the page mechanism

Fan Yang lljyangfan at gmail.com
Thu Nov 1 06:22:47 EDT 2012


2012/11/1 Mulyadi Santosa <mulyadi.santosa at gmail.com>

> Hi..
>
> On Wed, Oct 31, 2012 at 9:30 PM, Fan Yang <lljyangfan at gmail.com> wrote:
> > Hi all:
> >     I have a problem,  if i don't active the page mechanism just use the
> > segment mechanism at the x86 cpu, how can I manage the Multi-tasking?
>
> multitasking and memory management are not directly related IMHO, but
> you're right that we need process address space switching.
>
> I think what you're going to do is already done by Linux kernel port
> to non MMU  architectures. Can't recall which ones, so try to check
> https://en.wikipedia.org/wiki/Memory_management_unit for further
> examples
>
>
> -
> regards,
>
> Mulyadi Santosa
> Freelance Linux trainer and consultant
>
> blog: the-hydra.blogspot.com
> training: mulyaditraining.blogspot.com


Hi Mulyadi Santosa
    Very honored to see your reply.Yes, the page mechanism can protect
different task to interfere each other by mapping different task to
different physical memory, and this work is already done by kernel.

    In order to solve the prev problem I look through the intel manual
http://download.intel.com/design/processor/manuals/253668.pdf ,and in
chapter 3.1 I see this sentence "Segmentation provides a mechanism of
isolating individual code, data, and stack modules so that multiple
programs (or tasks) can run on the same processor without interfering with
one another." then this manual tell me that "There is no mode bit to
disable segmentation. The use of paging, however, is optional." that is to
say the segment mechanism must be support in the OS, but the page mechanism
is optional, but at the manual I can't find any example to explain how to
use the single segment mechanism to protect the different tasks to
interfere with one another ,so I have a guess you just see.
    BTW, I try the 2.6.32.27 kernel version but nothing changed.

Thanks
Fan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20121101/f55124ce/attachment.html 


More information about the Kernelnewbies mailing list