From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCHv7 2/2] memory barrier: adding smp_mb__after_lock Date: Thu, 09 Jul 2009 06:25:55 +0200 Message-ID: <4A557153.8000108@gmail.com> References: <20090708220717.GB2945@jolsa.redhat.com> <20090708221031.GD2945@jolsa.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: netdev@vger.kernel.org, mathieu.desnoyers@polymtl.ca, mingo@elte.hu, a.p.zijlstra@chello.nl, linux-kernel@vger.kernel.org, fbl@redhat.com, nhorman@redhat.com, htejun@gmail.com, jarkao2@gmail.com, oleg@redhat.com, davidel@xmailserver.org, davem@davemloft.net To: Jiri Olsa Return-path: Received: from gw1.cosmosbay.com ([212.99.114.194]:55142 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751739AbZGIE02 (ORCPT ); Thu, 9 Jul 2009 00:26:28 -0400 In-Reply-To: <20090708221031.GD2945@jolsa.redhat.com> Sender: netdev-owner@vger.kernel.org List-ID: Jiri Olsa a =E9crit : > Adding smp_mb__after_lock define to be used as a smp_mb call after > a lock. >=20 > Making it nop for x86, since {read|write|spin}_lock() on x86 are > full memory barriers. >=20 > wbr, > jirka >=20 >=20 > Signed-off-by: Jiri Olsa Signed-off-by: Eric Dumazet > --- > arch/x86/include/asm/spinlock.h | 4 ++++ > include/linux/spinlock.h | 5 +++++ > include/net/sock.h | 5 ++++- > 3 files changed, 13 insertions(+), 1 deletions(-) >=20 > diff --git a/arch/x86/include/asm/spinlock.h b/arch/x86/include/asm/s= pinlock.h > index b7e5db8..4e77853 100644 > --- a/arch/x86/include/asm/spinlock.h > +++ b/arch/x86/include/asm/spinlock.h > @@ -302,4 +302,8 @@ static inline void __raw_write_unlock(raw_rwlock_= t *rw) > #define _raw_read_relax(lock) cpu_relax() > #define _raw_write_relax(lock) cpu_relax() > =20 > +/* The {read|write|spin}_lock() on x86 are full memory barriers. */ > +static inline void smp_mb__after_lock(void) { } > +#define ARCH_HAS_SMP_MB_AFTER_LOCK > + > #endif /* _ASM_X86_SPINLOCK_H */ > diff --git a/include/linux/spinlock.h b/include/linux/spinlock.h > index 252b245..4be57ab 100644 > --- a/include/linux/spinlock.h > +++ b/include/linux/spinlock.h > @@ -132,6 +132,11 @@ do { \ > #endif /*__raw_spin_is_contended*/ > #endif > =20 > +/* The lock does not imply full memory barrier. */ > +#ifndef ARCH_HAS_SMP_MB_AFTER_LOCK > +static inline void smp_mb__after_lock(void) { smp_mb(); } > +#endif > + > /** > * spin_unlock_wait - wait until the spinlock gets unlocked > * @lock: the spinlock in question. > diff --git a/include/net/sock.h b/include/net/sock.h > index 4eb8409..2c0da92 100644 > --- a/include/net/sock.h > +++ b/include/net/sock.h > @@ -1271,6 +1271,9 @@ static inline int sk_has_allocations(const stru= ct sock *sk) > * in its cache, and so does the tp->rcv_nxt update on CPU2 side. T= he CPU1 > * could then endup calling schedule and sleep forever if there are = no more > * data on the socket. > + * > + * The sk_has_sleeper is always called right after a call to read_lo= ck, so we > + * can use smp_mb__after_lock barrier. > */ > static inline int sk_has_sleeper(struct sock *sk) > { > @@ -1280,7 +1283,7 @@ static inline int sk_has_sleeper(struct sock *s= k) > * > * This memory barrier is paired in the sock_poll_wait. > */ > - smp_mb(); > + smp_mb__after_lock(); > return sk->sk_sleep && waitqueue_active(sk->sk_sleep); > } > =20