🚀 go-pugleaf

In article <roelof.8@parst1.UUCP>, roelof@parst1.UUCP (Roelof van Suilichem) writes:
 > In article <1992Mar30.171014.3133@exu.ericsson.se> exuhag@exu.ericsson.se (James Hague) writes:
 > 
 > >Raymond Chen writes:
 > >>
 > >>Many people (including me, even) suggest the standard "jmp FFFF:0000"
 > >>method.
 > >>
 > >>But this is extremely dangerous if the user is running disk cacheing
 > >>software, since it bypasses *everything*!  Caches don't get flushed,
 > >>and you run the risk of trashing your hard drive!
 > 
 > As far as I know, they just precede the jmp with a 1 second delay so that
 > the cache can flush.  This is still dangerous.
 > 
 > Roelof

Another solution (and one used by configuration menus like Autocon and ConEd)
is to execute a DOS shell right before boot, to call the cache's own flush
function.  If you don't do any more disk access after the shell returns,
you're pretty safe to boot.

The disadvantage is that no two cache's use the same commands to flush.
One solution is to let the user provide a command line for flushing.  You'll
then need to explain the user that it's in his own interest that this is
done.

A typical command line is for example: MYCACHE /FLUSH

-- 
Regards from Robert Schmidt, PC C/asm is my game, fast graphics is my 
pain  -  Watch for my VGA/SB demo!
     e-mail to robert@solan.unit.no
 snail-mail to Stud.techn. Robert Schmidt, stud.post 170,
               Norwegian Institute of Technology, Trondheim, NORWAY