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: <f0dd9208-05ec-b2fa-4b29-5fa140486fa3@deltatee.com>
Date:   Wed, 14 Sep 2022 10:16:31 -0600
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.13.0
Content-Language: en-US
To:     Yu Kuai <yukuai1@huaweicloud.com>, song@kernel.org,
        guoqing.jiang@linux.dev, pmenzel@molgen.mpg.de
Cc:     linux-raid@vger.kernel.org, linux-kernel@vger.kernel.org,
        yukuai3@huawei.com, yi.zhang@huawei.com
References: <20220914014914.398712-1-yukuai1@huaweicloud.com>
 <20220914014914.398712-2-yukuai1@huaweicloud.com>
From:   Logan Gunthorpe <logang@deltatee.com>
In-Reply-To: <20220914014914.398712-2-yukuai1@huaweicloud.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Subject: Re: [PATCH v2 1/4] md/raid10: cleanup wait_barrier()
Precedence: bulk


On 2022-09-13 19:49, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Currently the nasty condition is wait_barrier() is hard to read. This
> patch factor out the condition into a function.
> 
> There are no functional changes.
> 
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> ---
>  drivers/md/raid10.c | 56 ++++++++++++++++++++++++++-------------------
>  1 file changed, 32 insertions(+), 24 deletions(-)
> 
> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index 64d6e4cd8a3a..56458a53043d 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -957,44 +957,52 @@ static void lower_barrier(struct r10conf *conf)
>  	wake_up(&conf->wait_barrier);
>  }
>  
> +static bool stop_waiting_barrier(struct r10conf *conf)
> +{
> +	/* barrier is dropped */
> +	if (!conf->barrier)
> +		return true;
> +
> +	/*
> +	 * If there are already pending requests (preventing the barrier from
> +	 * rising completely), and the pre-process bio queue isn't empty, then
> +	 * don't wait, as we need to empty that queue to get the nr_pending
> +	 * count down.
> +	 */
> +	if (atomic_read(&conf->nr_pending)) {
> +		struct bio_list *bio_list = current->bio_list;

I'd probably just put the bio_list declaration at the top of this
function, then the nested if statements are not necessary. The compiler
should be able to optimize the access just fine.

>  	if (conf->barrier) {
> -		struct bio_list *bio_list = current->bio_list;
> -		conf->nr_waiting++;
> -		/* Wait for the barrier to drop.
> -		 * However if there are already pending
> -		 * requests (preventing the barrier from
> -		 * rising completely), and the
> -		 * pre-process bio queue isn't empty,
> -		 * then don't wait, as we need to empty
> -		 * that queue to get the nr_pending
> -		 * count down.
> -		 */
>  		/* Return false when nowait flag is set */
>  		if (nowait) {
>  			ret = false;
>  		} else {
> +			conf->nr_waiting++;

Technically speaking, I think moving nr_waiting counts as a functional
change. As best as I can see it is correct, but it should probably be at
least mentioned in the commit message, or maybe done as a separate
commit with it's own justification. That way if it causes problems down
the road, a bisect will make the issue clearer.

Thanks,

Logan