??? 03/29/10 23:35 Read: times |
#174612 - Soft interrupts Responding to: ???'s previous message |
Erik Malund said:
A little known fact is that what we consider triggering interrupts actually does not. HUH??? as an example timer 1 does not trigger an interrupt, it just set TF1, Then when service is possible TF1 triggers the interrupt. Same principle for the rest of the bunch.
Erik Some processors/architecures supports "soft" interrupts by writing to interrupt bits, similarly to if the hardware had set them. This can sometimes be (ab)used to create critical sections that are trigged by the normal code and protected from interrupts of low or medium level, while still allowing critical high-level interrupts to be serviced. Some processors may show interrupt bits as R/W but with the little addendum that a write may only clear the bits, so you can't fake an interrupt but can "disarm" a pending interrupt for a device before enabling interrupts for it. |
Topic | Author | Date |
Timer Problem Assembly Code | 01/01/70 00:00 | |
Interrupt priorities... | 01/01/70 00:00 | |
Also... | 01/01/70 00:00 | |
The missing words | 01/01/70 00:00 | |
You sure? | 01/01/70 00:00 | |
many ways to skin a cat | 01/01/70 00:00 | |
Actually very relevant | 01/01/70 00:00 | |
My impression is | 01/01/70 00:00 | |
Interesting, but missing the point surely... | 01/01/70 00:00 | |
Interrupts are 'saved' | 01/01/70 00:00 | |
Soft interrupts | 01/01/70 00:00 | |
in a '51 forum | 01/01/70 00:00 | |
Still missing the point... | 01/01/70 00:00 | |
the missing reti's have already been mentioned | 01/01/70 00:00 | |
Apart from missing reti...![]() | 01/01/70 00:00 | |
Interupt priorities | 01/01/70 00:00 |