gossamer axe wrote:


> For number 2 I had no f-ing clue, and should I on a hardware exam?

IRQ=4, COM 2&4 use IRQ3
I never understood the history of 1&3=4, and 2&4=3 ?
Surely when they first added COM ports they would have taken the first
available IRQ, and that would have been 3 ?

IRQ numbers were assigned by Intel. IBM for some odd reason, screwed up on their use of some of them for unknown reasons (I seem to remember that there was an NMI IRQ was assigned to something else by IBM).

Back when I ran a BBS I had one machine, two modems...mouse was on COM1, and
the BBS modem was on COM2...I had to put the 2nd modem on COM3, but would
share an IRQ with the mouse, so had to get the tweezers (if my nails weren't
long enough) to change the jumpers to put it to IRQ 9 (the only free one
back then) and it worked fine.  But, this was back when modems actually HAD
jumpers or dipswitches.  I remembered the IRQs because Com1 is an even IRQ
and COM2 is an odd IRQ.  I believe IRQ 1 and 2 are reserved, so 3 & 4 were
and easy way to remember it.  9 was usually open, and I seem to remember 7
also, but again, I am probably wrong about that.  It's been a while since
fiddling with IRQs.

In the early PC-DOS days, IRQ's 2, 5, 9, and 10 were often unused and available for other things, quite often for serial ports. IRQ7 was traditionally used for the parallel (PIO, or printer) port.

Eventually, IRQ5 got abducted by sound cards, with IRQ's 2 and 9 as options. IRQ's 2 and 10 were common options for bus mouse cards.

The poor new kids on the block, network cards, especially Ethernet cards very often complained about an IRQ already being used during the configuration process. IRQ 5 was very often a default for NIC's in the workplace because PC's with sound cards were rare. Also a common practice was to turn off one serial (COM) port and reclaim its IRQ for something else.

Ahh, those were the good old days. #-o

--
   Best Regards,
      ~DJA.


--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to