Problem with reordering of memory accesses

[Franz Sirl]

At 19:12 18.10.00, Michael Schwingen wrote:
>On Wed, Oct 18, 2000 at 06:52:43PM +0200, Franz Sirl wrote:
> > I don't see a reason why gcc should order these accesses. Unless you meant
> > to write:
> >
> > typedef struct {
> >          int x1;
> >          short x2;
> >          void * volatile x3;
> >          void *x4;
> >          void * volatile x5;
> >          void *x6;
> >          void *x7;
> > } foo;
>
>I guess that is indeed what I really wanted (and it produces the right
>results).
>
>Although I must confess that I don´t quite understand the difference in the
>declaration.

volatile void * x5; // the stuff x5 points to is volatile
void * volatile x5; // x5 itself is volatile

Franz.