[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160527090221.GY3193@twins.programming.kicks-ass.net>
Date: Fri, 27 May 2016 11:02:21 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Martin Schwidefsky <schwidefsky@...ibm.com>
Cc: linux-kernel@...r.kernel.org, torvalds@...ux-foundation.org,
manfred@...orfullife.com, dave@...olabs.net,
paulmck@...ux.vnet.ibm.com, will.deacon@....com,
boqun.feng@...il.com, Waiman.Long@....com, tj@...nel.org,
pablo@...filter.org, kaber@...sh.net, davem@...emloft.net,
oleg@...hat.com, netfilter-devel@...r.kernel.org,
sasha.levin@...cle.com, hofrat@...dl.org, rth@...ddle.net,
vgupta@...opsys.com, linux@...linux.org.uk, realmz6@...il.com,
rkuo@...eaurora.org, tony.luck@...el.com, james.hogan@...tec.com,
ralf@...ux-mips.org, dhowells@...hat.com, jejb@...isc-linux.org,
mpe@...erman.id.au, ysato@...rs.sourceforge.jp,
cmetcalf@...lanox.com, chris@...kel.net,
Heiko Carstens <heiko.carstens@...ibm.com>
Subject: Re: [PATCH -v2 4/6] locking, arch: Update spin_unlock_wait()
On Fri, May 27, 2016 at 08:46:49AM +0200, Martin Schwidefsky wrote:
> > This fixes a number of spin_unlock_wait() users that (not
> > unreasonably) rely on this.
>
> All that is missing is an smp_rmb(), no?
Indeed.
> > --- a/arch/s390/include/asm/spinlock.h
> > +++ b/arch/s390/include/asm/spinlock.h
> > @@ -95,8 +95,7 @@ static inline void arch_spin_unlock(arch
> >
> > static inline void arch_spin_unlock_wait(arch_spinlock_t *lock)
> > {
> > - while (arch_spin_is_locked(lock))
> > - arch_spin_relax(lock);
> > + smp_cond_load_acquire(&lock->lock, !VAL);
> > }
> >
> > /*
>
> This change adds the smp_rmb() at the end of the waiting loop, but
> it also replaces arch_spin_relax() alias arch_lock_relax() with a
> cpu_relax(). This is not good, these two functions do *very* different
> things. cpu_relax() does an undirected yield with diagnose 0x44 but
> only if the system is non-SMT. arch_lock_relax() does an additional
> cpu_is_preempted() to test if the target cpu is running and does a
> directed yield with diagnose 0x9c.
>
> Why can't we just add the smp_rmb() to the arch_spin_unlock_wait()?
We can; I forgot about the special cpu_relax on s390, will fix.
Powered by blists - more mailing lists