Re: [avr-gcc-list] Clarification for interrupt specific changes inoutput_movhi() in revision 136238

[Anatoly Sokolov]

Hi.


> There is one issue I would like to get clarified. In prologue, conditional 
> check for is_OS_main is happening but not the same in epilogue.
>
> expand_prologue() {
> ...
> +       else if (TARGET_NO_INTERRUPTS
> +        || cfun->machine->is_signal
> +        || cfun->machine->is_OS_main)
> + {
   insn =
     emit_insn (gen_movhi_sp_r_>>>irq_off<<< (stack_pointer_rtx,
            frame_pointer_rtx));

In time when prologue of the "main" function  or interrupt handler declared 
with  "signal" attribute  is executed the interrupt is disabled and need not 
to save-disable-restore "I" flag when new value is written in SP.

> expand_epilogue() {
> ...
> +       else if (TARGET_NO_INTERRUPTS
> +        || cfun->machine->is_signal)
> + {
> ...
   emit_insn (gen_movhi_sp_r_>>>irq_off<<< (stack_pointer_rtx,
          frame_pointer_rtx));
In time when prologue of the interrupt handler declared with  "signal" 
attribute  is executed the interrupt is disabled and need not to 
save-disable-restore "I" flag when new value is written in SP.

On the other hand in "main" function interrupt may be enabled and writing to 
SP soul be de done in safe way.


Anatoly.