Re: [PATCH V2 5/5] blk-mq: Replace tags->lock with SRCU for tag iterators - Rockchip UFS regression

From: Jens Axboe <axboe@kernel.dk>
To: Bart Van Assche <bvanassche@acm.org>,
	Ming Lei <ming.lei@redhat.com>,
	Sebastian Reichel <sebastian.reichel@collabora.com>
Cc: linux-block@vger.kernel.org, Hannes Reinecke <hare@suse.de>,
	Yu Kuai <yukuai3@huawei.com>,
	kernel@collabora.com, linux-rockchip@lists.infradead.org
Subject: Re: [PATCH V2 5/5] blk-mq: Replace tags->lock with SRCU for tag iterators - Rockchip UFS regression
Date: Mon, 6 Oct 2025 17:03:19 -0600	[thread overview]
Message-ID: <07dbe66c-86d4-4a2d-ab7d-e0a351b3ae9f@kernel.dk> (raw)
In-Reply-To: <4f16e17d-796b-4896-9d4c-6d227514f1ca@acm.org>

On 10/6/25 12:02 PM, Bart Van Assche wrote:
> On 10/4/25 7:42 AM, Ming Lei wrote:
>>    One possible fix is to check hba->scsi_host_added before calling
>> scsi_host_busy():
>>
>>    dev_err(hba->dev, "%d outstanding reqs, tasks=0x%lx\n",
>>        hba->scsi_host_added ? scsi_host_busy(hba->host) : 0,
>>        hba->outstanding_tasks);
> 
> My preference is to add a check inside scsi_host_busy(), e.g. as follows
> (entirely untested):
> 
> diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
> index cc5d05dc395c..17173239301e 100644
> --- a/drivers/scsi/hosts.c
> +++ b/drivers/scsi/hosts.c
> @@ -611,8 +611,9 @@ int scsi_host_busy(struct Scsi_Host *shost)
>  {
>      int cnt = 0;
> 
> -    blk_mq_tagset_busy_iter(&shost->tag_set,
> -                scsi_host_check_in_flight, &cnt);
> +    if (shost->tag_set.ops)
> +        blk_mq_tagset_busy_iter(&shost->tag_set,
> +                    scsi_host_check_in_flight, &cnt);
>      return cnt;
>  }
>  EXPORT_SYMBOL(scsi_host_busy);

Agree, that's a nicer way to solve it, putting that stuff outside
of actual drivers.

-- 
Jens Axboe