Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!mailrus!wuarchive!usc!ucsd!pacbell.com!pacbell!att!mcdchg!laidbak!obdient!vpnet!cgordon
From: cgordon@vpnet.chi.il.us (Gordon Hlavenka)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Using Int 19h?
Keywords: warm boot cold boot
Message-ID: <2697a1f3-20d1.4comp.sys.ibm.pc-1@vpnet.chi.il.us>
Date: 8 Jul 90 21:55:04 GMT
References: <90184.155514J0S@psuvm.psu.edu> <1990Jul5.151534.3823@sj.ate.slb.co
Lines: 24


INT19 will work if you have not messed with interrupts.  In other words, NO
TSRs, NO *.SYS drivers, etc.

I was never able to get INT19 to work.  I heard (somewhere) that INT19 does
not rebuild the interrupt vector table.  I tried booting vanilla DOS and
then running a COM file which consisted of the two bytes 'CD 19' (INT 19),
and it _works_!  It's just not real useful.  (Some copy protection schemes
call INT19 on a protection failure: the machine just keeps rebooting)

The jump to the 8086 boot address is usually effective, just set the flag at
0000:0472 to 0 for cold and 1234H for warm boots.

But even that gets into trouble when you are running protected mode
applications.  I use 386MAX, and have been unable to get a software reboot
running.  MAX has that address remapped, so I need to unmap it and shut down
MAX before I do the jump.  Qualitas was somewhat helpful, but eventually I
was told I was on my own.

Anybody else?

----------------------------------------------------------
Gordon S. Hlavenka                 cgordon@vpnet.chi.il.us
Disclaimer: He's lying