Received-SPF: pass (google.com: domain of linux-kernel+bounces-185204-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Tue, 21 May 2024 19:10:07 +0300
From: Mike Rapoport <rppt@kernel.org>
To: Eric Chanudet <echanude@redhat.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>, Baoquan He <bhe@redhat.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Zhen Lei <thunder.leizhen@huawei.com>,
	Yajun Deng <yajun.deng@linux.dev>,
	Zhang Jianhua <chris.zjh@huawei.com>,
	linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
	Nick Piggin <npiggin@gmail.com>,
	Michael Ellerman <mpe@ellerman.id.au>
Subject: Re: [PATCH] arm64: init: override deferred_page_init_max_threads
Message-ID: <ZkzHX-4yYybEJdEQ@kernel.org>
References: <20240520231555.395979-5-echanude@redhat.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240520231555.395979-5-echanude@redhat.com>

(added powerpc folks)

On Mon, May 20, 2024 at 07:15:59PM -0400, Eric Chanudet wrote:
> This was the behavior prior to making the function arch-specific with
> commit ecd096506922 ("mm: make deferred init's max threads
> arch-specific")
> 
> Architectures can override the generic implementation that uses only one
> CPU. Setting DEFERRED_STRUCT_PAGE_INIT and testing on a few arm64
> platforms shows faster deferred_init_memmap completions:
> 
> |         | x13s        | SA8775p-ride | Ampere R137-P31 | Ampere HR330 |
> |         | Metal, 32GB | VM, 36GB     | VM, 58GB        | Metal, 128GB |
> |         | 8cpus       | 8cpus        | 8cpus           | 32cpus       |
> |---------|-------------|--------------|-----------------|--------------|
> | threads |  ms     (%) | ms       (%) |  ms         (%) |  ms      (%) |
> |---------|-------------|--------------|-----------------|--------------|
> | 1       | 108    (0%) | 72      (0%) | 224        (0%) | 324     (0%) |
> | cpus    |  24  (-77%) | 36    (-50%) |  40      (-82%) |  56   (-82%) |
> 
> Signed-off-by: Eric Chanudet <echanude@redhat.com>
> ---
>  arch/arm64/mm/init.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c
> index 9b5ab6818f7f..71f5188fe63d 100644
> --- a/arch/arm64/mm/init.c
> +++ b/arch/arm64/mm/init.c
> @@ -158,6 +158,13 @@ static void __init zone_sizes_init(void)
>  	free_area_init(max_zone_pfns);
>  }
>  
> +#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
> +int __init deferred_page_init_max_threads(const struct cpumask *node_cpumask)
> +{
> +	return max_t(int, cpumask_weight(node_cpumask), 1);
> +}
> +#endif
> +

Maybe we should make this default and let architectures that want a single
thread override deferred_page_init_max_threads() to return 1?

>  int pfn_is_map_memory(unsigned long pfn)
>  {
>  	phys_addr_t addr = PFN_PHYS(pfn);
> -- 
> 2.44.0
> 

-- 
Sincerely yours,
Mike.