teg@cci632.UUCP (Thomas Gauger) (01/03/91)
I need an immediate reply to a troublesome interrupt question. Today is Thursday 1/3/91, quickness is of the essence (sp?). I have a Real-Time Devices AD1000 A/D conversion board. I am tring to intercept an external interrupt using the EXTRI input line. This line is tied to IRQ5 via a jumper on the board. That is the hardware configuration I seem to have no problem there. Interrupts are received as IRQ5 by software. The software for the interrupt handler is broken down into 3 parts: an initialization routine, an interrupt handler, and a clean up routine (for termination of the process handling the interrupts). The initialization routine sets up the interrupt vector using (I am using the MS Quick C compiler) dos_getvect() and dos_setvect(). I do the dos_getvect() to save the original IRQ5 handler. I then read the OCW1 control word of the 8259A Intel Interrupt control chip, to get the current interrupt mask, and then enable IRQ5. Interesting at this point, IRQ5 was already enabled. I save the current mask for restoration in the clean up routine. Interrupts are recieved at this point (when generated). The interrupt handler reads some ports off the AD1000 board and then issues a SPECIFIC EOI (end of interrupt) command to the 8259A chip. The specific interrupt is of course IRQ5. I also reenable the IRQ5 interrupt in OCW1 (I don't know if this is necessary or not, but it does not seem to hurt). End of interrupt handler. Please note that I don't currently chain to the original interrupt handler, maybe this is a problem. The clean up routine, which is called at process termination, resets the IRQ5 mask in OCW1 to its original value and restores the original interrupt vector with a dos_setvect(). Now for my problem, according to all of the documentation that I have, IRQ5 (I am using an Everex IBM PC/XT compatible) is used for LPT2 interrupts. LPT2 is not installed. Therefore I should have no problem, but when I exit the program that uses the IRQ5, the floppy drive A: (the box has no B: drive) no longer functions. I always get a time out error resulting in drive not ready. A reboot corrects the problem, but this is not acceptable. I did see in some of my documentation that IRQ5 was used on the original PC is the floppy disk interrupt. Is there a conflict here? Should I use a different interrupt line? I can not use IRQ7, I have another board from another customer (not currently installed) that will be using IRQ7 with software that they will be providing. Can anyone tell me what the problem is? Please E-Mail all responses ASAP, I need the info quickly. Thanks in advance. -------------------------------------------------- Thomas E. Gauger The TEGGER Rochester's self proclaimed #1 Bill's fan Section C5, Row 32, Seats 109 & 110 Buffalo: 38 Opponents: 14
dubner@hpspkla.spk.hp.com (Joe Dubner) (01/07/91)
> The initialization routine sets up the interrupt vector using (I am using > the MS Quick C compiler) dos_getvect() and dos_setvect(). I do the > dos_getvect() to save the original IRQ5 handler. I then read the OCW1 control I don't know a lot about the interrupt controller chip in the PC, and know even less about MS Quick C, but I'd be willing to bet good money that dos_getvect() works on software interrupts, not hardware IRQ lines. When you do a dos_getvect() on "5", you're saving the INT 05h vector, which is used to sense the PrtSc key. Lots of luck. ... Joe