From: muta...@gmail.com   
      
   On Sunday, July 18, 2021 at 1:55:18 PM UTC+10, muta...@gmail.com wrote:   
      
   > > > > Since I can only get 512 MiB for 16-bit programs anyway,   
   > > > > I may as well have the mandatory 6 selectors or whatever   
   > > > > to address the first 3.5 GiB of memory for 32-bit programs   
   > > > > and the rest of the selectors used for 16-bit programs in   
   > > > > the other 512 MiB of memory.   
      
   And this will solve another problem.   
      
   A few weeks ago I was talking to someone who was using   
   "debug" to assemble 8086 assembler to do some simple   
   INT 21H calls and I was trying to figure out how to give him   
   the equivalent in PDOS/386.   
      
   I was thinking of reserving the region above 1 MiB for one   
   single 32-bit .com program to reside. So 32-bit "debug.exe"   
   would take you to that region. But the 32-bit INT 21H is   
   expecting full edx or whatever to be provided for values,   
   so the entered instructions will need to change. Although   
   come to think of it, I can zero all the registers before   
   executing the .com residing at x'0010 0100'.   
      
   Come to think of it, I should be able to execute multiple   
   32-bit .com programs at arbitrary locations so long as it is   
   on a 64k boundary.   
      
   "debug" would generate x'66' prefixes in the above situation   
   if only 16-bit instructions were entered. But I'll make it part   
   of "the rules" that the high 16-bits of 32-bit registers will be   
   zeroed upon entry to a .com program. Maybe a .exe too?   
      
   Anyway, my original idea was that if I support 16-bit 8086   
   executables in the 512 MiB region I can provide a nice   
   environment for people running "debug", because the   
   selectors will all be set appropriately to give the equivalent   
   of what people are used to.   
      
   BFN. Paul.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)