Received-SPF: pass (google.com: domain of linux-kernel+bounces-142408-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Subject: Re: [PATCH v3 1/4] LoongArch: Move CONFIG_HAVE_SETUP_PER_CPU_AREA
 related code to smp.c
To: Huacai Chen <chenhuacai@kernel.org>
References: <20240411010510.22135-1-yangtiezhu@loongson.cn>
 <20240411010510.22135-2-yangtiezhu@loongson.cn>
 <CAAhV-H6TVEEB=xdr_ymuUnDqPDXwtM7R7yJBj03YhWpCwqjAiA@mail.gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>, Arnd Bergmann <arnd@arndb.de>,
 Marc Zyngier <maz@kernel.org>, loongarch@lists.linux.dev,
 linux-kernel@vger.kernel.org, loongson-kernel@lists.loongnix.cn
From: Tiezhu Yang <yangtiezhu@loongson.cn>
Message-ID: <dcf6e475-8219-6b1e-a99b-2cdf1e86289c@loongson.cn>
Date: Fri, 12 Apr 2024 17:27:31 +0800
User-Agent: Mozilla/5.0 (X11; Linux mips64; rv:45.0) Gecko/20100101
 Thunderbird/45.4.0
Precedence: bulk
MIME-Version: 1.0
In-Reply-To: <CAAhV-H6TVEEB=xdr_ymuUnDqPDXwtM7R7yJBj03YhWpCwqjAiA@mail.gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit


On 04/12/2024 12:12 PM, Huacai Chen wrote:
> On Thu, Apr 11, 2024 at 9:05 AM Tiezhu Yang <yangtiezhu@loongson.cn> wrote:
>>
>> Currently, if CONFIG_NUMA is not set but CONFIG_SMP is set, the arch
>> specified setup_per_cpu_areas() in arch/loongarch/kernel/numa.c will
>> not be built and the generic setup_per_cpu_areas() in mm/percpu.c is
>> actually used, this is not reasonable and does not work as intended.
> Why is the generic version not reasonable? We need a custom version
> just because it can put the percpu variable in the best node. If NUMA
> disabled, software can only see one node, how to optimize?

The initial aim is to use the arch specific setup_per_cpu_areas()
in any case under CONFIG_SMP, this patch can be dropped if it is
meaningless for the case of !CONFIG_NUMA and CONFIG_SMP.

Thanks,
Tiezhu