Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <a0182cf5-5d1d-8a3d-2cf7-8862a70edd11@suse.cz>
Date:   Tue, 2 May 2023 12:00:48 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.10.1
Subject: Re: [PATCH 1/1] mm: change per-VMA lock statistics to be disabled by
 default
Content-Language: en-US
To:     Suren Baghdasaryan <surenb@google.com>, akpm@linux-foundation.org
Cc:     torvalds@linux-foundation.org, lstoakes@gmail.com,
        linux-mm@kvack.org, linux-kernel@vger.kernel.org,
        kernel-team@android.com
References: <20230428173533.18158-1-surenb@google.com>
From:   Vlastimil Babka <vbabka@suse.cz>
In-Reply-To: <20230428173533.18158-1-surenb@google.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 4/28/23 19:35, Suren Baghdasaryan wrote:
> Change CONFIG_PER_VMA_LOCK_STATS to be disabled by default, as most
> users don't need it. Add configuration help to clarify its usage.
> 
> Fixes: 52f238653e45 ("mm: introduce per-VMA lock statistics")
> Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
> Signed-off-by: Suren Baghdasaryan <surenb@google.com>

Acked-by: Vlastimil Babka <vbabka@suse.cz>

> ---
>  mm/Kconfig.debug | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/Kconfig.debug b/mm/Kconfig.debug
> index 6dae63b46368..a925415b4d10 100644
> --- a/mm/Kconfig.debug
> +++ b/mm/Kconfig.debug
> @@ -274,6 +274,12 @@ config DEBUG_KMEMLEAK_AUTO_SCAN
>  config PER_VMA_LOCK_STATS
>  	bool "Statistics for per-vma locks"
>  	depends on PER_VMA_LOCK
> -	default y
>  	help
> -	  Statistics for per-vma locks.
> +	  Say Y here to enable success, retry and failure counters of page
> +	  faults handled under protection of per-vma locks. When enabled, the
> +	  counters are exposed in /proc/vmstat. This information is useful for
> +	  kernel developers to evaluate effectiveness of per-vma locks and to
> +	  identify pathological cases. Counting these events introduces a small
> +	  overhead in the page fault path.
> +
> +	  If in doubt, say N.