Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 27 Jan 2022 09:17:01 -0500
From:   "Alex Xu (Hello71)" <alex_y_xu@yahoo.ca>
Subject: Re: [PATCH -next] blk-mq: Fix wrong wakeup batch configuration which
 will cause hang
To:     andriy.shevchenko@linux.intel.com, axboe@kernel.dk,
        Laibin Qiu <qiulaibin@huawei.com>
Cc:     linux-block@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20220127100047.1763746-1-qiulaibin@huawei.com>
In-Reply-To: <20220127100047.1763746-1-qiulaibin@huawei.com>
MIME-Version: 1.0
Message-Id: <1643292980.cunltl8w6q.none@localhost>
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Content-Length: 1854
Precedence: bulk

Excerpts from Laibin Qiu's message of January 27, 2022 5:00 am:
> Commit 180dccb0dba4f ("blk-mq: fix tag_get wait task can't be
> awakened") will recalculating wake_batch when inc or dec active_queues
> to avoid wake_batch is > hctx_max_depth. At the same time, in order to
> not affect performance as much as possible, the minimum wakeup batch is
> set to 4. But when the QD is small (such as QD=3D1), if inc or dec
> active_queues will increase wakeup batch, which will lead to hang.
>=20
> Fix this problem with the following strategies:
> QD          :  >=3D 32 | < 32
> ---------------------------------
> wakeup batch:  8~4   | 3~1
>=20
> Fixes: 180dccb0dba4f ("blk-mq: fix tag_get wait task can't be awakened")
> Link: https://lore.kernel.org/linux-block/78cafe94-a787-e006-8851-69906f0=
c2128@huawei.com/T/#t
> Reported-by: Alex Xu (Hello71) <alex_y_xu@yahoo.ca>
> Signed-off-by: Laibin Qiu <qiulaibin@huawei.com>
> ---
>  lib/sbitmap.c | 8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
>=20
> diff --git a/lib/sbitmap.c b/lib/sbitmap.c
> index 6220fa67fb7e..09d293c30fd2 100644
> --- a/lib/sbitmap.c
> +++ b/lib/sbitmap.c
> @@ -488,9 +488,13 @@ void sbitmap_queue_recalculate_wake_batch(struct sbi=
tmap_queue *sbq,
>  					    unsigned int users)
>  {
>  	unsigned int wake_batch;
> +	unsigned int min_batch;
> +	unsigned int depth =3D (sbq->sb.depth + users - 1) / users;
> =20
> -	wake_batch =3D clamp_val((sbq->sb.depth + users - 1) /
> -			users, 4, SBQ_WAKE_BATCH);
> +	min_batch =3D sbq->sb.depth >=3D (4 * SBQ_WAIT_QUEUES) ? 4 : 1;
> +
> +	wake_batch =3D clamp_val(depth / SBQ_WAIT_QUEUES,
> +			min_batch, SBQ_WAKE_BATCH);
>  	__sbitmap_queue_update_wake_batch(sbq, wake_batch);
>  }
>  EXPORT_SYMBOL_GPL(sbitmap_queue_recalculate_wake_batch);
> --=20
> 2.22.0
>=20
>=20

Tested-by: Alex Xu (Hello71) <alex_y_xu@yahoo.ca>