Re: Question on smp_mb__before_spinlock

From	Will Deacon <will.deacon@arm.com>
Newsgroups	linux.kernel
Subject	Re: Question on smp_mb__before_spinlock
Date	2016-09-06 19:50 +0200
Message-ID	<sesFs-52m-35@gated-at.bofh.it> (permalink)
References	<sdYxH-1S9-3@gated-at.bofh.it> <sdZaq-2o0-19@gated-at.bofh.it> <semA2-1hJ-7@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

On Tue, Sep 06, 2016 at 01:17:53PM +0200, Peter Zijlstra wrote:
> On Mon, Sep 05, 2016 at 11:10:22AM +0100, Will Deacon wrote:
> 
> > > The second issue I wondered about is spinlock transitivity. All except
> > > powerpc have RCsc locks, and since Power already does a full mb, would
> > > it not make sense to put it _after_ the spin_lock(), which would provide
> > > the same guarantee, but also upgrades the section to RCsc.
> > > 
> > > That would make all schedule() calls fully transitive against one
> > > another.
> > 
> > It would also match the way in which the arm64 atomic_*_return ops
> > are implemented, since full barrier semantics are required there.
> 
> Hmm, are you sure; the way I read arch/arm64/include/asm/atomic_ll_sc.h
> is that you do ll/sc-rel + mb.

Yes, all I meant was that we put the full barrier at the end, but the
two things are indeed different sequences.

Will

Thread

Question on smp_mb__before_spinlock Peter Zijlstra <peterz@infradead.org> - 2016-09-05 11:40 +0200
  Re: Question on smp_mb__before_spinlock kbuild test robot <lkp@intel.com> - 2016-09-05 12:00 +0200
    Re: Question on smp_mb__before_spinlock Peter Zijlstra <peterz@infradead.org> - 2016-09-05 12:20 +0200
      Re: Question on smp_mb__before_spinlock Fengguang Wu <lkp@intel.com> - 2016-09-05 13:30 +0200
  Re: Question on smp_mb__before_spinlock Will Deacon <will.deacon@arm.com> - 2016-09-05 12:20 +0200
    Re: Question on smp_mb__before_spinlock Peter Zijlstra <peterz@infradead.org> - 2016-09-06 13:20 +0200
      Re: Question on smp_mb__before_spinlock Will Deacon <will.deacon@arm.com> - 2016-09-06 19:50 +0200
  Re: Question on smp_mb__before_spinlock "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2016-09-05 12:40 +0200
    Re: Question on smp_mb__before_spinlock Peter Zijlstra <peterz@infradead.org> - 2016-09-05 13:40 +0200
      Re: Question on smp_mb__before_spinlock "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2016-09-05 16:00 +0200
  Re: Question on smp_mb__before_spinlock Nicholas Piggin <npiggin@gmail.com> - 2016-09-07 14:20 +0200
    Re: Question on smp_mb__before_spinlock Peter Zijlstra <peterz@infradead.org> - 2016-09-07 15:30 +0200
      Re: Question on smp_mb__before_spinlock Will Deacon <will.deacon@arm.com> - 2016-09-07 16:00 +0200

csiph-web