On Tue, Nov 29, 2011 at 04:15:36PM +1030, Alan Modra wrote:
> On Mon, Nov 28, 2011 at 05:42:15PM -0800, Richard Henderson wrote:
> > On 11/28/2011 06:02 AM, Alan Modra wrote:
> > > - unsigned int ret = bar->generation & ~3;
> > > - /* Do we need any barrier here or is __sync_add_and_fetch acting
On Mon, Nov 28, 2011 at 05:42:15PM -0800, Richard Henderson wrote:
> On 11/28/2011 06:02 AM, Alan Modra wrote:
> > - unsigned int ret = bar->generation & ~3;
> > - /* Do we need any barrier here or is __sync_add_and_fetch acting
> > - as the needed LoadLoad barrier already? */
> > - ret +=
On 11/28/2011 06:02 AM, Alan Modra wrote:
> - unsigned int ret = bar->generation & ~3;
> - /* Do we need any barrier here or is __sync_add_and_fetch acting
> - as the needed LoadLoad barrier already? */
> - ret += __sync_add_and_fetch (&bar->awaited, -1) == 0;
> + unsigned int ret = __atom
On Mon, Nov 28, 2011 at 08:09:01PM +1030, Alan Modra wrote:
> On Mon, Nov 28, 2011 at 09:13:27AM +0100, Jakub Jelinek wrote:
> > On Mon, Nov 28, 2011 at 10:12:09AM +1030, Alan Modra wrote:
> > > --- libgomp/config/linux/bar.c(revision 181718)
> > > +++ libgomp/config/linux/bar.c(working cop
On Mon, Nov 28, 2011 at 09:13:27AM +0100, Jakub Jelinek wrote:
> On Mon, Nov 28, 2011 at 10:12:09AM +1030, Alan Modra wrote:
> > --- libgomp/config/linux/bar.c (revision 181718)
> > +++ libgomp/config/linux/bar.c (working copy)
> > @@ -36,18 +36,15 @@ gomp_barrier_wait_end (gomp_barrier_t
On Mon, Nov 28, 2011 at 10:12:09AM +1030, Alan Modra wrote:
> --- libgomp/config/linux/bar.c(revision 181718)
> +++ libgomp/config/linux/bar.c(working copy)
> @@ -36,18 +36,15 @@ gomp_barrier_wait_end (gomp_barrier_t *b
>if (__builtin_expect ((state & 1) != 0, 0))
> {
>
This patch cures the remaining libgomp failures I see on power7. I'll
admit to approaching this fix with a big hammer at first, liberally
using MEMMODEL_SEQ_CST barriers all over bar.h and bar.c, then
gradually reducing the number of places and strictness of the
barriers. This may still have a fe
