30 May
2000
30 May
'00
5:01 a.m.
What I'm
saying is that the way is *should* work is:
pointer fault on attempt to execute faulted pointer to
DLL routine
segment fault on attempt to execute code in
unmapped
DLL routine's
segment
page fault on attempt to execute code in unmapped DLL
routine's page
And in a really proper OS you'd also get a page fault if
the page map
containing the PageMapTableEntries weren't
currently mapped in (and
yes you can page out page maps).
And before you point out that segments went away from Win32 let
me say that too is another fatal flaw with the Windows family of
OS's.
Is it just me -- or does this sound like Multics segments? 
30 May
30 May
10:44 p.m.
To the best of my knowledge, win32, including win95/98 still suffer from
64k segments. In particular, the video subsystems are still 16bit,
non-reentrant code which means programmers still have to chop thier
graphics into 64k chunks and feed those to the GUI. A lot of people were
really disapointed with this when Microsoft first revealed it prior to
95's debut.
There used to be an excellent IBM whitepaper comparing Win95, OS/2 warp,
and Windows NT but I have been unable to find it alive anywhere.
Douglas Quebbeman wrote:
Dang- busted again. :-)
-doug q
What I'm saying is that the way is *should* work
is:
pointer fault on attempt to execute faulted pointer to
DLL routine
segment fault on attempt to execute code in
unmapped
DLL routine's
segment
page fault on attempt to execute code in unmapped DLL
routine's page
And in a really proper OS you'd also get a page fault if
the page map
containing the PageMapTableEntries weren't
currently mapped in (and
yes you can page out page maps).
And before you point out that segments went away from Win32 let
me say that too is another fatal flaw with the Windows family of
OS's.
Is it just me -- or does this sound like Multics segments?
8971
days inactive
8972
days old
1 comments
2 participants
participants (2)
-
dhquebbeman＠theestopinalgroup.com -
Technoid＠cheta.net