Hi all, Does anybody know what does the system do when a division by zero occurs? We are running RTX 51 ful kernel on Philips' 80c51, 80c552 and 80c660 controllers. Thanks, Laurent.
A similar question came up here a few months ago because a particular chip returns a strange value (all bits set) from the math accelerator when a division by zero occurs. I believe the compiler returns a value of 0, but of course the real result is undefined. After some discussion, we determined that the math accelerator returns all bits set because this was fewer internal gates than returning 0's. And, division by 0 is undefined anyway, so... Don't rely on the results of a division by 0. Jon
Geez! Isn't there any exception/software interrupt thrown that we could catch? Laurent.
Geez! Isn't there any exception/software interrupt thrown that we could catch? Not for 8051 derivatives. I'm not sure about the 166 series, but you will need to go with a platform more complex than the 8051 if you want hardware-based exceptions.
Geez! Isn't there any exception/software interrupt thrown that we could catch? Nope. You seem to believe that the '51 is a microprocessor, it is not, it is a microcontroller. Erik
I don't believe the C166 has a hardware trap for this but division by zero causes the overflow bit to be set. I know the HC16 provides a hardware trap, I guess that makes it a microprocessor in Erik's eyes. -Walt
Walt, Yeah... In my opinion, the use of the word microcontroller has been muddied a bit to include anything that includes some built-in "peripherals." For instance, Motorola/Freescale dub the 68340 a microcontroller, but I think most people would have put that one squarely in the processor category. I think it just comes down to the 8051 being an architecture designed for simplicity.
I don't believe the C166 has a hardware trap for this but division by zero causes the overflow bit to be set. same for the '51 I know the HC16 provides a hardware trap, I guess that makes it a microprocessor in Erik's eyes. Nope, if a controller get a few gadgets, that's nice, if it does not, do not *** about it. Erik
"In my opinion, the use of the word microcontroller has been muddied..." I don't know that there was ever a hard-and-fast distinction between "controller" and "processor." These days, I believe that Intel refer to the '386 as a "microcontroller" - that's just their view of market segmentation, rather than any specific architectural or functional distinction!