Date: Sun, 31 Dec 1995 11:18:35 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Larry Blanchard <larryb AT comtch DOT iea DOT com>
Cc: djgpp AT delorie DOT com
Subject: Re: interrupts - again


On 27 Dec 1995, Larry Blanchard wrote:

> For example, the FAQ says "jump to the previous handler".  The info data
> says "chain to the next handler".  Which is it?

Both.  ``Chaining'' is accomplished by ``jumping'' to the previous 
handler.

>                                              Does this mean the
> next/previous real mode or protected mode handler or either/both?

If you install a protected-mode handler, you chain to the previous PM 
handler; real-mode handler should chain to RM handler.

> In PCTIME13 I more or less understand the real mode handler, but don't
> see why the protected mode handler doesn't chain every time (thus getting
> the clock off).

PCTIME13 supports boosting the timer tick frequency, while BIOS routines 
only know about the timer ticking 18.2 times a second.

>                  Does the EOI prevent the chaining?  Incidentally, I'm
> surprised that the wrapper doesn't do the EOI but as I said, I'm still a
> little confused.

EOI tells the PIC (programmable interrupt controller) that the interrupt 
has been handled, so other interrupts can now be serviced.  EOI should be 
issued by the last handler in the chain, or else you risk getting another 
timer interrupt while this one has not been processed yet.  That's why 
the wrapper doesn't do it: it has no means of knowing whether it's the 
last handler in the chain.