Page 1 of 1

Does interrupts totally disappear with CLI?

Posted: Mon Jan 26, 2004 3:38 am
by Sam
Or are they "suspended" until next "sti" and will occur immedietly after???

Re:Does interrupts totally disappear with CLI?

Posted: Mon Jan 26, 2004 5:08 am
by Solar
Deleted: I stand corrected.

Re:Does interrupts totally disappear with CLI?

Posted: Mon Jan 26, 2004 5:13 am
by Pype.Clicker
actually, they do not "disappear" ... The CPU will ignore any external IRQ request until you STI. When you STI, it will acknowledge the IRQ request and receive interrupts that were pending at the Programmable Interrupt Controller.

Btw, even when IF is cleared, you can still issue interrupts with INT nn instruction or receive exceptions.

Oh, for completeness, CLI doesn't mask "Non-maskable Interrupts", but unless you activate "issue NMI when memory checksum fails" and have memory chips with ECC, i dunno how you could even get an NMI ...

Re:Does interrupts totally disappear with CLI?

Posted: Mon Jan 26, 2004 6:59 am
by Candy
Pype.Clicker wrote: Oh, for completeness, CLI doesn't mask "Non-maskable Interrupts", but unless you activate "issue NMI when memory checksum fails" and have memory chips with ECC, i dunno how you could even get an NMI ...
Well, you don't need chips with ECC :D just shorting some pins of the ISA bus gives an NMI :-[

Re:Does interrupts totally disappear with CLI?

Posted: Mon Jan 26, 2004 7:41 am
by Pype.Clicker
shorting pins ?? you mean cutting them, or wire-strap them or something ??

Re:Does interrupts totally disappear with CLI?

Posted: Mon Jan 26, 2004 9:41 am
by Candy
Pype.Clicker wrote: shorting pins ?? you mean cutting them, or wire-strap them or something ??
well, you take a wire, connect them and the system says: Memory parity error. System halted.

At least, on a 8088 we tried it on :). We tried to get some pins connected normally so we could make our own 8-bit ISA cards... didn't work out as we had hoped :)