Re: freezer: should barriers be smp ?

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Thursday, April 14, 2011, Mike Frysinger wrote:
> On Wed, Apr 13, 2011 at 18:49, Rafael J. Wysocki wrote:
> > On Thursday, April 14, 2011, Mike Frysinger wrote:
> >> i guess the trouble for us is that you have one CPU posting writes to
> >> task->flags (and doing so by grabbing the task's spinlock), but the
> >> other CPU is simply reading those flags.  there are no SMP barriers in
> >> between the read and write steps, nor is the reading CPU grabbing any
> >> locks which would be an implicit SMP barrier.  since the Blackfin SMP
> >> port lacks hardware cache coherency, there is no way for us to know
> >> "we've got to sync the caches before we can do this read".  by using
> >> the patch i posted above, we have that signal and so things work
> >> correctly.,
> >
> > In theory I wouldn't expect the patch to work correctly, because it replaces
> > _stronger_ memory barriers with _weaker_ SMP barriers.  However, looking at
> > the blackfin's definitions of SMP barriers I see that it uses extra stuff that
> > should _also_ be used in the definitions of the mandatory barriers.
> >
> > In my opinion is an architecture problem, not the freezer code problem.
> 
> OK, we have a patch pending locally which populates all barriers with
> this logic, but based on my understanding of things, that didnt seem
> correct.  i guess i'm reading too much into the names ... i'd expect
> the opposite behavior where "rmb" is only for UP needs while "smp_rmb"
> is a rmb which additionally covers SMP.

Well, I guess the naming is for historical reasons, ie. mb(), rmb() and wmb()
were there first and it probably was regarded cleaner to use new names for the
optimized smp_ variants than to rename all instances already in the code and
then repurpose the old names.

Thanks,
Rafael