From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Zijlstra Subject: [RFC][PATCH 1/5] ia64: Fix up smp_mb__{before,after}_clear_bit Date: Thu, 06 Feb 2014 14:48:26 +0100 Message-ID: <20140206135001.059111455@infradead.org> References: <20140206134825.305510953@infradead.org> Return-path: Content-Disposition: inline; filename=peterz-ia64-atomics.patch Sender: linux-kernel-owner@vger.kernel.org To: linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org Cc: torvalds@linux-foundation.org, akpm@linux-foundation.org, mingo@kernel.org, will.deacon@arm.com, paulmck@linux.vnet.ibm.com, Peter Zijlstra List-Id: linux-arch.vger.kernel.org IA64 doesn't actually have acquire/release barriers, its a lie! Add a comment explaining this and fix up the bitop barriers. Signed-off-by: Peter Zijlstra --- arch/ia64/include/asm/bitops.h | 7 ++----- arch/ia64/include/uapi/asm/cmpxchg.h | 9 +++++++++ 2 files changed, 11 insertions(+), 5 deletions(-) --- a/arch/ia64/include/asm/bitops.h +++ b/arch/ia64/include/asm/bitops.h @@ -65,11 +65,8 @@ __set_bit (int nr, volatile void *addr) *((__u32 *) addr + (nr >> 5)) |= (1 << (nr & 31)); } -/* - * clear_bit() has "acquire" semantics. - */ -#define smp_mb__before_clear_bit() smp_mb() -#define smp_mb__after_clear_bit() do { /* skip */; } while (0) +#define smp_mb__before_clear_bit() barrier(); +#define smp_mb__after_clear_bit() barrier(); /** * clear_bit - Clears a bit in memory --- a/arch/ia64/include/uapi/asm/cmpxchg.h +++ b/arch/ia64/include/uapi/asm/cmpxchg.h @@ -118,6 +118,15 @@ extern long ia64_cmpxchg_called_with_bad #define cmpxchg_rel(ptr, o, n) \ ia64_cmpxchg(rel, (ptr), (o), (n), sizeof(*(ptr))) +/* + * Worse still - early processor implementations actually just ignored + * the acquire/release and did a full fence all the time. Unfortunately + * this meant a lot of badly written code that used .acq when they really + * wanted .rel became legacy out in the wild - so when we made a cpu + * that strictly did the .acq or .rel ... all that code started breaking - so + * we had to back-pedal and keep the "legacy" behavior of a full fence :-( + */ + /* for compatibility with other platforms: */ #define cmpxchg(ptr, o, n) cmpxchg_acq((ptr), (o), (n)) #define cmpxchg64(ptr, o, n) cmpxchg_acq((ptr), (o), (n)) From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from merlin.infradead.org ([205.233.59.134]:49159 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932695AbaBFNwT (ORCPT ); Thu, 6 Feb 2014 08:52:19 -0500 Message-ID: <20140206135001.059111455@infradead.org> Date: Thu, 06 Feb 2014 14:48:26 +0100 From: Peter Zijlstra Subject: [RFC][PATCH 1/5] ia64: Fix up smp_mb__{before,after}_clear_bit References: <20140206134825.305510953@infradead.org> Content-Disposition: inline; filename=peterz-ia64-atomics.patch Sender: linux-arch-owner@vger.kernel.org List-ID: To: linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org Cc: torvalds@linux-foundation.org, akpm@linux-foundation.org, mingo@kernel.org, will.deacon@arm.com, paulmck@linux.vnet.ibm.com, Peter Zijlstra Message-ID: <20140206134826.IkNRZnOu6qlmvJ9gA-uCVzEa16e0wCCA9pP0fCOqqes@z> IA64 doesn't actually have acquire/release barriers, its a lie! Add a comment explaining this and fix up the bitop barriers. Signed-off-by: Peter Zijlstra --- arch/ia64/include/asm/bitops.h | 7 ++----- arch/ia64/include/uapi/asm/cmpxchg.h | 9 +++++++++ 2 files changed, 11 insertions(+), 5 deletions(-) --- a/arch/ia64/include/asm/bitops.h +++ b/arch/ia64/include/asm/bitops.h @@ -65,11 +65,8 @@ __set_bit (int nr, volatile void *addr) *((__u32 *) addr + (nr >> 5)) |= (1 << (nr & 31)); } -/* - * clear_bit() has "acquire" semantics. - */ -#define smp_mb__before_clear_bit() smp_mb() -#define smp_mb__after_clear_bit() do { /* skip */; } while (0) +#define smp_mb__before_clear_bit() barrier(); +#define smp_mb__after_clear_bit() barrier(); /** * clear_bit - Clears a bit in memory --- a/arch/ia64/include/uapi/asm/cmpxchg.h +++ b/arch/ia64/include/uapi/asm/cmpxchg.h @@ -118,6 +118,15 @@ extern long ia64_cmpxchg_called_with_bad #define cmpxchg_rel(ptr, o, n) \ ia64_cmpxchg(rel, (ptr), (o), (n), sizeof(*(ptr))) +/* + * Worse still - early processor implementations actually just ignored + * the acquire/release and did a full fence all the time. Unfortunately + * this meant a lot of badly written code that used .acq when they really + * wanted .rel became legacy out in the wild - so when we made a cpu + * that strictly did the .acq or .rel ... all that code started breaking - so + * we had to back-pedal and keep the "legacy" behavior of a full fence :-( + */ + /* for compatibility with other platforms: */ #define cmpxchg(ptr, o, n) cmpxchg_acq((ptr), (o), (n)) #define cmpxchg64(ptr, o, n) cmpxchg_acq((ptr), (o), (n))