[En-Nut-Discussion] a strange "trace" code in usartavr.c

[Matthias Ringwald]

Hi Michael

> My question is:
>      if I move "trace 4" to the end of the function, will it change  
> the semantics of this function ?
Nope. It will fix a bug. :) It really should be at the end, I guess  
that was my mistake.
I'vll updated the CVS as soon as I can access it. :)

The loop is only added based on a compilte time flag which is set to  
false as default
and if the tracer is not enabled, it also doesn't matter.

Thanks for pointing out the mistake,

matthias

..


> #ifdef NUTTRACER
>        TRACE_ADD_ITEM 
> (TRACE_TAG_INTERRUPT_EXIT,TRACE_INT_UART_RXCOMPL); /* trace 4 */
> #endif
>
> #ifdef UART_READMULTIBYTE
>    } while ( inb(UCSRnA) & _BV(RXC) ); // byte in buffer?
>
>    // Eventually post event to wake thread
>    if (postEvent)
>        NutEventPostFromIrq(&rbf->rbf_que);
> #endif
>
> }
>
> ---------------------------------------------------------------------- 
> ----------------------
>