Received-SPF: pass (google.com: domain of linux-kernel+bounces-138566-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
From: Huacai Chen <chenhuacai@loongson.cn>
To: Huacai Chen <chenhuacai@kernel.org>
Cc: loongarch@lists.linux.dev,
	Xuefeng Li <lixuefeng@loongson.cn>,
	Guo Ren <guoren@kernel.org>,
	Xuerui Wang <kernel@xen0n.name>,
	Jiaxun Yang <jiaxun.yang@flygoat.com>,
	linux-kernel@vger.kernel.org,
	loongson-kernel@lists.loongnix.cn,
	Huacai Chen <chenhuacai@loongson.cn>
Subject: [PATCH] LoongArch: Select THP_SWAP if HAVE_ARCH_TRANSPARENT_HUGEPAGE
Date: Wed, 10 Apr 2024 21:12:56 +0800
Message-ID: <20240410131256.2165746-1-chenhuacai@loongson.cn>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

THP_SWAP has been proven to improve the swap throughput significantly on
x86_64 system according to commit bd4c82c22c367e0 ("mm, THP, swap: delay
splitting THP after swapped out"), on ARM64 system according to commit
d0637c505f8a1d ("arm64: enable THP_SWAP for arm64") and on RISC-V system
according to commit 87f81e66e2e84c7 ("riscv: enable THP_SWAP for RV64").

Enable THP_SWAP for LoongArch, testing the micro-benchmark which is
introduced by commit d0637c505f8a1d ("arm64: enable THP_SWAP for arm64")
shows below numbers on the Loongson-3A5000 board:

swp out bandwidth w/o patch: 1815716 bytes/ms (mean of 10 tests)
swp out bandwidth w/  patch: 3410003 bytes/ms (mean of 10 tests)

Improved by 46.75%!

Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
---
 arch/loongarch/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig
index d52a95195e7f..41ddc39e1fc8 100644
--- a/arch/loongarch/Kconfig
+++ b/arch/loongarch/Kconfig
@@ -69,6 +69,7 @@ config LOONGARCH
 	select ARCH_WANT_LD_ORPHAN_WARN
 	select ARCH_WANT_OPTIMIZE_HUGETLB_VMEMMAP
 	select ARCH_WANTS_NO_INSTR
+	select ARCH_WANTS_THP_SWAP if HAVE_ARCH_TRANSPARENT_HUGEPAGE
 	select BUILDTIME_TABLE_SORT
 	select COMMON_CLK
 	select CPU_PM
-- 
2.43.0