metag: fix memory barriers
authorMikulas Patocka <mpatocka@redhat.com>
Thu, 8 May 2014 19:51:37 +0000 (15:51 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 7 Jun 2014 20:25:37 +0000 (13:25 -0700)
commit44563045712ee4f385b5f3814d69fc73b5f22288
tree2e6bcadf007ed599527b1c45a62ddececd7c5fd6
parentaece7dc95409f8934281954a7e82ddf55b765913
metag: fix memory barriers

commit 2425ce84026c385b73ae72039f90d042d49e0394 upstream.

Volatile access doesn't really imply the compiler barrier. Volatile access
is only ordered with respect to other volatile accesses, it isn't ordered
with respect to general memory accesses. Gcc may reorder memory accesses
around volatile access, as we can see in this simple example (if we
compile it with optimization, both increments of *b will be collapsed to
just one):

void fn(volatile int *a, long *b)
{
(*b)++;
*a = 10;
(*b)++;
}

Consequently, we need the compiler barrier after a write to the volatile
variable, to make sure that the compiler doesn't reorder the volatile
write with something else.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/metag/include/asm/barrier.h