Re: [PATCH RFC tip/lkmm 08/10] EXP Remove understore from smp_mb__before_atomic() workings

From: Paul E. McKenney
Date: Thu Feb 15 2018 - 18:32:53 EST


On Fri, Feb 16, 2018 at 12:19:58AM +0100, Andrea Parri wrote:
> On Thu, Feb 15, 2018 at 02:49:37PM -0800, Paul E. McKenney wrote:
> > On Thu, Feb 15, 2018 at 11:30:09PM +0100, Andrea Parri wrote:
> > > On Fri, Feb 09, 2018 at 06:20:29AM -0800, Paul E. McKenney wrote:
> > > > Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
> > >
> > > nit: s/understore/underscore (a very understandable lapsus, BTW ;-)
> >
> > ;-)
> >
> > Will fix when squashing.
> >
> > > According to the cover letter, patches 8-10 of this series are intended
> > > to be squashed together; it could be useful to add a description then:
> > > the motivation for these changes might appear "trivial" today; months/
> > > years from now? mmh, I can't put my finger on it ;-)
> >
> > Assuming we keep these patches, agreed. Something like this for the
> > aggregated single commit, perhaps:
> >
> > tools/memory-model: Convert underscores to hyphens
> >
> > Typical cat-language code uses hyphens for word separators in
> > identifiers, but several LKMM identifiers use underscores instead.
> > This commit therefore converts the underscores in the .bell and
> > .cat files for smp_mb__before_atomic(), smp_mb__after_atomic(),
> > and smp_mb__after_spinlock() to hyphens.
> >
> > > IMO, this descrition should also mention why this solution "is the way
> > > to go" (e.g., compare with the solution recently implemented by Akira).
> >
> > You lost me here. Or are you asking for a commit log similar to that
> > provided by Akira for his forward-port of Alan's unrelated patch? If
> > so, please see above.
>
> I was mainly thinking "but rb_dep..."; or you're also planning to
> rename this annotation and, maybe, the "rb-dep" relation (in order
> to avoid the conflict)?

The rb_dep relation goes away completely once we catch LKMM up with
the recent changes that remove smp_read_barrier_depends() from the
core kernel. So my current thought is hold back the "_"->"-" commit
until that catch-up happens.

Thanx, Paul

> Andrea
>
>
> >
> > Thanx, Paul
> >
> > > Andrea
> > >
> > >
> > > > ---
> > > > tools/memory-model/linux-kernel.bell | 2 +-
> > > > tools/memory-model/linux-kernel.cat | 2 +-
> > > > tools/memory-model/linux-kernel.def | 2 +-
> > > > 3 files changed, 3 insertions(+), 3 deletions(-)
> > > >
> > > > diff --git a/tools/memory-model/linux-kernel.bell b/tools/memory-model/linux-kernel.bell
> > > > index b984bbda01a5..558bbef9e08a 100644
> > > > --- a/tools/memory-model/linux-kernel.bell
> > > > +++ b/tools/memory-model/linux-kernel.bell
> > > > @@ -28,7 +28,7 @@ enum Barriers = 'wmb (*smp_wmb*) ||
> > > > 'rcu-lock (*rcu_read_lock*) ||
> > > > 'rcu-unlock (*rcu_read_unlock*) ||
> > > > 'sync-rcu (*synchronize_rcu*) ||
> > > > - 'before_atomic (*smp_mb__before_atomic*) ||
> > > > + 'before-atomic (*smp_mb__before_atomic*) ||
> > > > 'after_atomic (*smp_mb__after_atomic*) ||
> > > > 'after_spinlock (*smp_mb__after_spinlock*)
> > > > instructions F[Barriers]
> > > > diff --git a/tools/memory-model/linux-kernel.cat b/tools/memory-model/linux-kernel.cat
> > > > index babe2b3b0bb3..31c08a44c088 100644
> > > > --- a/tools/memory-model/linux-kernel.cat
> > > > +++ b/tools/memory-model/linux-kernel.cat
> > > > @@ -29,7 +29,7 @@ let rb-dep = [R] ; fencerel(Rb_dep) ; [R]
> > > > let rmb = [R \ Noreturn] ; fencerel(Rmb) ; [R \ Noreturn]
> > > > let wmb = [W] ; fencerel(Wmb) ; [W]
> > > > let mb = ([M] ; fencerel(Mb) ; [M]) |
> > > > - ([M] ; fencerel(Before_atomic) ; [RMW] ; po? ; [M]) |
> > > > + ([M] ; fencerel(Before-atomic) ; [RMW] ; po? ; [M]) |
> > > > ([M] ; po? ; [RMW] ; fencerel(After_atomic) ; [M]) |
> > > > ([M] ; po? ; [LKW] ; fencerel(After_spinlock) ; [M])
> > > > let gp = po ; [Sync-rcu] ; po?
> > > > diff --git a/tools/memory-model/linux-kernel.def b/tools/memory-model/linux-kernel.def
> > > > index a397387f77cc..ff4b756f4a44 100644
> > > > --- a/tools/memory-model/linux-kernel.def
> > > > +++ b/tools/memory-model/linux-kernel.def
> > > > @@ -21,7 +21,7 @@ smp_mb() { __fence{mb} ; }
> > > > smp_rmb() { __fence{rmb} ; }
> > > > smp_wmb() { __fence{wmb} ; }
> > > > smp_read_barrier_depends() { __fence{rb_dep}; }
> > > > -smp_mb__before_atomic() { __fence{before_atomic} ; }
> > > > +smp_mb__before_atomic() { __fence{before-atomic} ; }
> > > > smp_mb__after_atomic() { __fence{after_atomic} ; }
> > > > smp_mb__after_spinlock() { __fence{after_spinlock} ; }
> > > >
> > > > --
> > > > 2.5.2
> > > >
> > >
> >
>