From: nowhere@never.at   
      
   On 29/12/2023 07:13, Paul Edwards wrote:   
   > I recently "found out" (ie is it true?) that   
   > there is apparently a BIOS call to go from   
   > PM16 to RM16 on an 80286.   
   >   
   > It originally did that via the convoluted   
   > and slow keyboard thing. In principle I have   
   > no problem with that, and I also have no   
   > problem with replacing the BIOS ROM with an   
   > updated version that uses a more efficient   
   > (and documented in section 9.6.2 of the   
   > "80286 and 80287 Programmer's Reference Manual")   
   > triple fault mechanism:   
   >   
   > https://bitsavers.org/components/intel/80286/210498-005_80286_   
   nd_80287_Programmers_Reference_Manual_1987.pdf   
   >   
   > I also don't mind the BIOS ROM being updated   
   > with the LOADALL instruction which has been   
   > belatedly documented (I believe that's the   
   > fastest solution).   
      
   I'd forget about LOADALL it's opcode has a new meaning now.   
   and return from protected mode needed a forced crash on 286.   
      
   > So speed doesn't bother me.   
   >   
   > But it occurs to me. If there is a BIOS interrupt   
   > to go from PM16 to RM16, why not have one to go   
   > from PM32 to RM16 too? As that is what I need in   
   > PDOS/386. I rolled my own. I wasn't aware that   
   > there might be a BIOS function to do that.   
   >   
   > So maybe in PDOS/286 at least I can use the   
   > official BIOS routines, assuming they exist   
   > and I can find them and I'm not misunderstanding   
   > something.   
      
   why bother with BIOS calls (which may not be present at all)?   
   my "all-mode switches"   
   [from any to any of RM16,PM16,PM32,LM64,CM16,CM32]   
   are just a few bytes large and share only one system-stack,   
   all whats required is proper setup of GDT and IDT entries up front.   
   __   
   wolfgang   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)