NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/comp/sys/intel/2230

Internet Message Format | 1992-11-11 | 1.8 KB

Xref: sparky comp.sys.intel:2230 comp.sys.ibm.pc.hardware:29565
Newsgroups: comp.sys.intel,comp.sys.ibm.pc.hardware
Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!zaphod.mps.ohio-state.edu!sample.eng.ohio-state.edu!purdue!yuma!csn!stortek!stortek!aarond
From: [e-mail] (Aaron Dailey)
Subject: 8259 PIC programming questions
Message-ID: <[e-mail]>
Sender: [e-mail]
Nntp-Posting-Host: xibm.stortek.com
Organization: StorageTek
Date: Wed, 11 Nov 1992 19:06:40 GMT
Lines: 26

I'm programming what is essentially a PC with several peripherals
connected via 8259 programmable interrupt controller to a 386SX
running in protected mode.  There two 8259s, in the normal AT cascaded
arrangement.

In the interrupt handler for these devices, I send a non specific end
of interrupt (0x20) to the interrupt controller as the last thing I
do.  I've looked at some source code, which sends a specific end of
interrupt command to the PIC before exiting the ISR.  Other code I've
seen uses nonspecific EOIs.  I've looked at the IBM Tech Ref for the
AT, which documents that the BIOS initiailizes the 8259 to not special
fully nested mode, which indicates to me that I should send it a
specific EOI command.  Which do I need?

Also, I've looked at some source code which, while handling IRQ9, on
the slave PIC, only sends an EOI to the master.  The Intel data sheets
for the 8259 clearly state that EOIs must be sent to the master and
slave PICs for an interrupt on a slave PIC.  What's the deal?  I
understand that IRQ9 was IRQ2 on the master PIC on PC (8088) class
machines, but doesn't the slave PIC need to be given an EOI?

--
Aaron Dailey Internet: [e-mail]
StorageTek Corporation   Voice: [phone], FAX: [phone]
Mail:  MS0242, 2270 South 88th Street, Louisville, C0 80028