Date: Sun, 3 Mar 1996 10:02:01 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Martynas Kunigelis <martynas DOT kunigelis AT vm DOT ktu DOT lt>
Cc: DJGPP mailing list <djgpp AT delorie DOT com>
Subject: Re: HW interrupts
In-Reply-To: <199603011224.HAA08125@delorie.com>
Message-Id: <Pine.SUN.3.91.960303095759.18857M-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Fri, 1 Mar 1996, Martynas Kunigelis wrote:

>   Shame on me, but read the FAQ section on hw int handling only after posting
> the above zip to DJ. And one thing confused me, so here's my question:
>   if I hook the protected mode herdware interrupt and *don't* chain, what will
> happen if the interrupt happens while CPU is in real mode (e.g. reading from
> HD) ? I thought if I hook the PM int 9 and don't chain, the keyboard is mine
> *forever*. So how would it be?

It *is* yours forever.  Under DPMI, the hw interrupts are always passed 
to PM handler first, and only if unhandled, the RM handler is called.  
That means that if the interrupt happens while the processor is in real 
mode, it switches to protected mode, calls the PM handler, then switches 
back to real mode and returns to whatever it was doing.  For what I know, 
the FAQ says this explicitly.  If it isn't clear enough, please tell me 
and I'll change the wording.