Should enabling a previously flagged interrupt force it to be handled?

[Lucas Parzych]

Currently, if an already flagged, but previously disabled interrupt is suddenly enabled, the interrupt handler will not be invoked.

Do we want this to be the behavior, or should pending interrupts be handled when the appropriate bits go high in the IER's?

[William Roberts]

Right now we check transitions between clock cycles between new and old IFR registers. So if something is flagged into the IFR and it're re-enabled it won't be handled (as you point out). I think we only want to move the IFR register bits into the old IFR when interrupts are enabled globally and the IER is set for that interrupt as well. I wonder if Amos can get a test around this.