sergio masci wrote:
> On Thu, 30 Jul 2009, Dave Tweed wrote:
> > There is one ambiguity -- when the indices are equal, is the FIFO
> > completely full or completely empty? This means that you can't
> > actually completely fill the FIFO in a simple implementation.
> 
> This means that the fifo buffer can only ever store a max of 
> ('buffer_length' - 1) items (in other words one item becomes a guard 
> item).

Isn't that exactly what I said???

My point was, that if you add a simple boolean flag to the FIFO structure,
you CAN completely fill it, and this actually has negligible impact on the
overall code size, and does NOT affect its thread-safe properties.

> NOTE: the way a temporary index is used to protect the location of fifo 
> item while it is being manipulated and consequently these routines are
> *** INTERRUPT SAFE ***

Actually, your temporary variable has nothing to do with making the
routines thread/interrupt-safe. I don't know the complete specs of the
language you're using, but using the variable appears to be just a
performance optimization (to avoid having to recalculate the updated
value at the end of fifo_write()) and/or a workaround for the limited
expression-handling capabilities of the compiler.

You could just as well have called fifo_full() from fifo_write(), and
fifo_empty() from fifo_read().

-- Dave Tweed
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist