[ECOS] question about isr function signature (profiling)

[Andrew Lunn]

On Fri, Mar 26, 2004 at 05:30:25PM -0500, Yuxin Jiang wrote:
> 
> For ARM7TDMI core, I want to define a ISR service routine, as following:
> 
> // ISR.
> static cyg_uint32
> timer_isr(CYG_ADDRWORD vector, CYG_ADDRWORD data, HAL_SavedRegisters *regs)
> {
> 
>     HAL_INTERRUPT_ACKNOWLEDGE (CYGNUM_HAL_INTERRUPT_PROFILE_TIMER);
>     __profile_hit(regs->pc);
> 
>     return CYG_ISR_HANDLED;
> }
> 
> 
> My question is: is the signature for the timer_isr correct?

It looks correct.
 
> To be specific, is it correct to declare that there is a third parameter for
> a isr, and this third parameter is a pointer to the HAL_SavedRegisters
> structure, and the structure has correct (updated) context registers value
> at the moment the interrupt occurs?
> 
> As you can see, I am trying to do profiling. But I am not sure I actually
> get the correct pc value passed into the __profile_hit() function call.

For ARM the PC will be for the next instruction to be executed, not
the interrupted instruction. It could also be wrong when the
instruction just executed was a branch.
            
      Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss