Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Received-SPF: Pass (protection.outlook.com: domain of nokia.com designates
 131.228.2.8 as permitted sender) receiver=protection.outlook.com;
 client-ip=131.228.2.8; helo=fihe3nok0734.emea.nsn-net.net;
From:   Alexander A Sverdlin <alexander.sverdlin@nokia.com>
To:     Paul Burton <paul.burton@imgtec.com>, linux-mips@vger.kernel.org
Cc:     Alexander Sverdlin <alexander.sverdlin@nokia.com>,
        Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
        Will Deacon <will@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Boqun Feng <boqun.feng@gmail.com>,
        Ingo Molnar <mingo@redhat.com>, linux-kernel@vger.kernel.org
Subject: [PATCH 1/6] MIPS: Octeon: Implement __smp_store_release()
Date:   Wed, 27 Jan 2021 21:36:22 +0100
Message-Id: <20210127203627.47510-2-alexander.sverdlin@nokia.com>
In-Reply-To: <20210127203627.47510-1-alexander.sverdlin@nokia.com>
References: <20210127203627.47510-1-alexander.sverdlin@nokia.com>
MIME-Version: 1.0
Content-Type: text/plain
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 27 Jan 2021 20:36:34.6757
 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 869a0cb6-4be5-42cd-3a58-08d8c3033cf3
X-MS-Exchange-CrossTenant-Id: 5d471751-9675-428d-917b-70f44f9630b0
X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=5d471751-9675-428d-917b-70f44f9630b0;Ip=[131.228.2.8];Helo=[fihe3nok0734.emea.nsn-net.net]
X-MS-Exchange-CrossTenant-AuthSource: DB5EUR03FT043.eop-EUR03.prod.protection.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem
X-MS-Exchange-Transport-CrossTenantHeadersStamped: DB7PR07MB3947
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

From: Alexander Sverdlin <alexander.sverdlin@nokia.com>

On Octeon smp_mb() translates to SYNC while wmb+rmb translates to SYNCW
only. This brings around 10% performance on tight uncontended spinlock
loops.

Refer to commit 500c2e1fdbcc ("MIPS: Optimize spinlocks.") and the link
below.

On 6-core Octeon machine:
sysbench --test=mutex --num-threads=64 --memory-scope=local run

w/o patch:	1.60s
with patch:	1.51s

Link: https://lore.kernel.org/lkml/5644D08D.4080206@caviumnetworks.com/
Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
---
 arch/mips/include/asm/barrier.h | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/arch/mips/include/asm/barrier.h b/arch/mips/include/asm/barrier.h
index 49ff172..24c3f2c 100644
--- a/arch/mips/include/asm/barrier.h
+++ b/arch/mips/include/asm/barrier.h
@@ -113,6 +113,15 @@ static inline void wmb(void)
 					    ".set arch=octeon\n\t"	\
 					    "syncw\n\t"			\
 					    ".set pop" : : : "memory")
+
+#define __smp_store_release(p, v)					\
+do {									\
+	compiletime_assert_atomic_type(*p);				\
+	__smp_wmb();							\
+	__smp_rmb();							\
+	WRITE_ONCE(*p, v);						\
+} while (0)
+
 #else
 #define smp_mb__before_llsc() smp_llsc_mb()
 #define __smp_mb__before_llsc() smp_llsc_mb()
-- 
2.10.2