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;
Sender: Tejun Heo <htejun@gmail.com>
Date:   Mon, 17 Oct 2022 09:15:39 -1000
From:   Tejun Heo <tj@kernel.org>
To:     Kemeng Shi <shikemeng@huawei.com>
Cc:     axboe@kernel.dk, linux-block@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 6/8] blk-iocost: Avoid to call current_hweight_max if
 iocg->inuse == iocg->active
Message-ID: <Y02p23KdUZfkYyzI@slm.duckdns.org>
References: <20221017020011.25016-1-shikemeng@huawei.com>
 <20221017020011.25016-7-shikemeng@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20221017020011.25016-7-shikemeng@huawei.com>
Precedence: bulk

On Mon, Oct 17, 2022 at 10:00:09AM +0800, Kemeng Shi wrote:
> The old_hwi is already max hweight_inuse if iocg->inuse == iocg->active.
> Remove unnecessary calculation.
> 
> Signed-off-by: Kemeng Shi <shikemeng@huawei.com>
> ---
>  block/blk-iocost.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/block/blk-iocost.c b/block/blk-iocost.c
> index 96c1571a8a1d..fa90f471dfdc 100644
> --- a/block/blk-iocost.c
> +++ b/block/blk-iocost.c
> @@ -2299,7 +2299,10 @@ static void ioc_timer_fn(struct timer_list *timer)
>  			 * Determine the donation amount.
>  			 */
>  			current_hweight(iocg, &hwa, &old_hwi);
> -			hwm = current_hweight_max(iocg);
> +			if (iocg->inuse == iocg->active)
> +				hwm = old_hwi;
> +			else
> +				hwm = current_hweight_max(iocg);

I don't think this is correct. The intermediate nodes might be donating.
This also isn't a meaningful optimization given that it's in the cold
periodic timer path. I'd much rather keep the code simpler unless the
performance benfeit can be clearly demonstrated.

Thanks.

-- 
tejun