Re: [RFC, 2.6] a simple FIFO implementation

[Buddy Lucas]

On Thu, 16 Sep 2004 17:52:47 +0200, Stelian Pop <stelian@xxxxxxxxxx> wrote:
> On Thu, Sep 16, 2004 at 05:51:04PM +0200, Buddy Lucas wrote:
> 
> > > No, because the type is *unsigned* int.
> >
> > Indeed, that would exactly be the reason *why* this would fail. ;-)
> >
> > The expression fifo->size - fifo->tail + fifo->head might be negative
> > at some point, right? (fifo->head has wrapped to some small value and
> > fifo->tail > fifo->size)
> 
> And what is the value of an unsigned int holding that 'negative' value ? :)
> 

Which unsigned int?! ;-) The expression a - b is negative for unsigned
ints a and b where a < b. So, your unsigned ints "total" and
"remaining" won't be negative of
course, but they won't reflect what is actually left in the buffer;
they will equal the
value of len (in some cases) after fifo->head has wrapped (because of the 
unsignedness) and fifo->tail has not. Which would not be correct.


Cheers,
Buddy

> 
> 
> Stelian.
> --
> Stelian Pop <stelian@xxxxxxxxxx>
>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/