Re: What exactly is Protected Mode?

From: Beth (BethStone21_at_hotmail.NOSPICEDHAM.com)
Date: 05/02/04 

Date: Sun, 2 May 2004 07:20:42 +0100

Herman Dullink wrote:
> Important!!! Regardless of any state or mode, most units in a CPU
will
> operate
> the same way all the time. Even if the CPU is in Real Mode, the
segmentation
> unit will check every memory access against the values in the
descriptor
> cache.
> In normal operation, the descriptor cache is loaded with values that
makes
> the CPU operate in a 8086 compatible way (however, not 100%).

Evenbit, are you reading here too?

The above and most of Herman's post is making the point I was trying
to make but in an infinitely better and clearer way...I wasn't talking
about "V86 mode", actually...but, as Herman suggests, that the CPU is
actually always operating "the same way all the time" and "real mode"
is just a "special case" where all the various values of segments and
other "protection" things are delibrately set to behave in (as close
as possible to) the original 8086 behaviour...perhaps I confused
things by saying it as "always in protected mode" (I meant "always
employing the protection mechanisms", as Herman phrases infinitely
better as "Even if the CPU is in Real Mode, the segmentation unit will
check every memory access against the values in the descriptor cache."
by not mixing up his jargon, as I did...technically, what I said about
"always in protected mode" is, of course, total nonsense and makes no
sense...but, Hopefully, people can decipher the terrible way I phrased
it to understand that I basically meant what Herman is saying here
without mixing up his jargon at all ;)...

"V86 mode" would actually be a further "special case" where it _does_
leave the "protections" partially on but changes the segmentation
stuff to behave as "real mode", so that it is operating a little as
"real mode" and a little as "protected mode" to bridge the
"compatibility gap" to make things like the "DOS virtual machine" in
Windows actually work (namely, that old 8086 software can run like old
8086 software but that the machine has not completely reverted back to
"real mode" so that some "protections" are enabled...such as allowing
the OS to still "intercept" things and "redirect" them so that the
"DOS emulation" and Windows can happily run side-by-side without
clashing into each other all the time ;)...

Beth :)

Relevant Pages

  • Re: Help with error code
    ... bad capacitors next to the CPU socket. ... holds the CPU heatsink in place, ... Not all computers have overheat protection. ... That will help confirm either an overheating ...
    (microsoft.public.windowsxp.general)
  • Re: Help with error code
    ... off with a 2 tone error that I had thought was perhaps bad RAM, I ran ram test and it was ok. ... bad capacitors next to the CPU socket. ... holds the CPU heatsink in place, ... Not all computers have overheat protection. ...
    (microsoft.public.windowsxp.general)
  • Re: AMD processor overheating protection on Asus motherboards
    ... Just because ASUS doesn't say anything about it doesn't mean it's not there. ... Overheating Protection), that has been available on their AMD CPU ...
    (alt.comp.periphs.mainboard.asus)
  • Re: Protecting Encryption Algorithms
    ... I suppose that a system with specialized hardware ... > sealed chip doesn't have to encrypt its data to/from its internal memory. ... sending the unencrypted code to the CPU. ... > The truth of this statement is dependent on what level of protection you ...
    (sci.crypt)
  • Re: [RFC PATCH 2/4] Rename lock_cpu_hotplug to get_online_cpus
    ... the refcount semantics in these operations. ... protection against cpus coming up/ going down in the code below, ... since the cpu-hotplug operation doesn't affect the cpu_present_map. ... Online operations must check that a cpu is present ...
    (Linux-Kernel)