Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: More Microsoft annoyances ...part 2

Author: Robert Hyatt

Date: 07:44:52 08/14/01

Go up one level in this thread


On August 14, 2001 at 07:26:22, Adam Oellermann wrote:

>On August 14, 2001 at 04:11:58, Jeremiah Penery wrote:
>
>>On August 14, 2001 at 00:44:43, Slater Wold wrote:
>>
>>>Oh, and like the MS clip said, there are programs that when run under Windows 98
>>>or ME will USE a 2nd CPU.  But they are not going through the OS.  CAD and even
>>>the deep programs will use a 2nd CPU under Windows 98 and ME.
>>
>>This is wrong.  Win9x/Me do not even acknowledge that a second processor exists.
>> Since in this case the program can't know more than the OS, it is impossible
>>that a program can use both CPUs under such an operating system.  Really, it
>>just does not work.
>
>Before I start, I don't know either way, having never tried an SMP box on a
>silly operating system. However, the notion that software is limited to what is
>exposed by the OS is surely incorrect.

Actually it isn't.  Programs run in a "user mode" on the cpu which prevents
them from doing certain operations that would tend to make security impossible
or would tend to affect other programs in a negative way.  IE a program can't
directly read/write a disk drive, it has to go through the O/S file management
stuff to enforce file permissions.  A program can't store anywhere in memory
it wants or it could/would corrupt other programs.

Of course a user can (in unix) run as "super-user" and get the processor into a
privileged mode and do the above things.  But with the very likely affect that
the machine will crash and burn at some point.  A user-mode program gets whacked
if it does something illegal.  A kernel-mode program brings the whole machine
down if it does something illegal.




> It seems to me that low-level software
>can choose to use OS services or bypass them. Slater may be talking about
>software simply bypassing OS services in this instance. It may not be easy, but
>it should surely be *possible*.


Bypassing the O/S _must_ be impossible.  Otherwise it offers no protection/
security between the programs that are running concurrently.  As a result, it
would be unreliable and unstable.  The antithesis of what an operating system
is supposed to be.


>
>- Adam



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.