public inbox for linux-scsi@vger.kernel.org
 help / color / mirror / Atom feed
From: "Asutosh Das (asd)" <asutoshd@codeaurora.org>
To: Bart Van Assche <bvanassche@acm.org>,
	"Martin K . Petersen" <martin.petersen@oracle.com>
Cc: Jaegeuk Kim <jaegeuk@kernel.org>,
	Adrian Hunter <adrian.hunter@intel.com>,
	linux-scsi@vger.kernel.org,
	"James E.J. Bottomley" <jejb@linux.ibm.com>,
	Bean Huo <beanhuo@micron.com>, Avri Altman <avri.altman@wdc.com>,
	Can Guo <cang@codeaurora.org>,
	Stanley Chu <stanley.chu@mediatek.com>,
	Keoseong Park <keosung.park@samsung.com>
Subject: Re: [PATCH v4 16/17] scsi: ufs: Optimize the command queueing code
Date: Mon, 6 Dec 2021 14:41:50 -0800	[thread overview]
Message-ID: <0ba5c50f-3e79-2cae-c502-59f70812cca3@codeaurora.org> (raw)
In-Reply-To: <20211203231950.193369-17-bvanassche@acm.org>

On 12/3/2021 3:19 PM, Bart Van Assche wrote:
> Remove the clock scaling lock from ufshcd_queuecommand() since it is a
> performance bottleneck. Instead check the SCSI device budget bitmaps in
> the code that waits for ongoing ufshcd_queuecommand() calls. A bit is
> set in sdev->budget_map just before scsi_queue_rq() is called and a bit
> is cleared from that bitmap if scsi_queue_rq() does not submit the
> request or after the request has finished. See also the
> blk_mq_{get,put}_dispatch_budget() calls in the block layer.
> 
> There is no risk for a livelock since the block layer delays queue
> reruns if queueing a request fails because the SCSI host has been
> blocked.
> 
> Cc: Asutosh Das (asd) <asutoshd@codeaurora.org>
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
> ---

Reviewed-by: Asutosh Das <asutoshd@codeaurora.org>

>   drivers/scsi/ufs/ufshcd.c | 33 +++++++++++++++++++++++----------
>   drivers/scsi/ufs/ufshcd.h |  1 +
>   2 files changed, 24 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 9f0a1f637030..650dddf960c2 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -1070,13 +1070,31 @@ static bool ufshcd_is_devfreq_scaling_required(struct ufs_hba *hba,
>   	return false;
>   }
>   
> +/*
> + * Determine the number of pending commands by counting the bits in the SCSI
> + * device budget maps. This approach has been selected because a bit is set in
> + * the budget map before scsi_host_queue_ready() checks the host_self_blocked
> + * flag. The host_self_blocked flag can be modified by calling
> + * scsi_block_requests() or scsi_unblock_requests().
> + */
> +static u32 ufshcd_pending_cmds(struct ufs_hba *hba)
> +{
> +	struct scsi_device *sdev;
> +	u32 pending = 0;
> +
> +	shost_for_each_device(sdev, hba->host)
> +		pending += sbitmap_weight(&sdev->budget_map);
> +
> +	return pending;
> +}
> +
>   static int ufshcd_wait_for_doorbell_clr(struct ufs_hba *hba,
>   					u64 wait_timeout_us)
>   {
>   	unsigned long flags;
>   	int ret = 0;
>   	u32 tm_doorbell;
> -	u32 tr_doorbell;
> +	u32 tr_pending;
>   	bool timeout = false, do_last_check = false;
>   	ktime_t start;
>   
> @@ -1094,8 +1112,8 @@ static int ufshcd_wait_for_doorbell_clr(struct ufs_hba *hba,
>   		}
>   
>   		tm_doorbell = ufshcd_readl(hba, REG_UTP_TASK_REQ_DOOR_BELL);
> -		tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
> -		if (!tm_doorbell && !tr_doorbell) {
> +		tr_pending = ufshcd_pending_cmds(hba);
> +		if (!tm_doorbell && !tr_pending) {
>   			timeout = false;
>   			break;
>   		} else if (do_last_check) {
> @@ -1115,12 +1133,12 @@ static int ufshcd_wait_for_doorbell_clr(struct ufs_hba *hba,
>   			do_last_check = true;
>   		}
>   		spin_lock_irqsave(hba->host->host_lock, flags);
> -	} while (tm_doorbell || tr_doorbell);
> +	} while (tm_doorbell || tr_pending);
>   
>   	if (timeout) {
>   		dev_err(hba->dev,
>   			"%s: timedout waiting for doorbell to clear (tm=0x%x, tr=0x%x)\n",
> -			__func__, tm_doorbell, tr_doorbell);
> +			__func__, tm_doorbell, tr_pending);
>   		ret = -EBUSY;
>   	}
>   out:
> @@ -2681,9 +2699,6 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
>   
>   	WARN_ONCE(tag < 0, "Invalid tag %d\n", tag);
>   
> -	if (!down_read_trylock(&hba->clk_scaling_lock))
> -		return SCSI_MLQUEUE_HOST_BUSY;
> -
>   	/*
>   	 * Allows the UFS error handler to wait for prior ufshcd_queuecommand()
>   	 * calls.
> @@ -2772,8 +2787,6 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
>   out:
>   	rcu_read_unlock();
>   
> -	up_read(&hba->clk_scaling_lock);
> -
>   	if (ufs_trigger_eh()) {
>   		unsigned long flags;
>   
> diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
> index 8e942762e668..88c20f3608c2 100644
> --- a/drivers/scsi/ufs/ufshcd.h
> +++ b/drivers/scsi/ufs/ufshcd.h
> @@ -778,6 +778,7 @@ struct ufs_hba_monitor {
>    * @clk_list_head: UFS host controller clocks list node head
>    * @pwr_info: holds current power mode
>    * @max_pwr_info: keeps the device max valid pwm
> + * @clk_scaling_lock: used to serialize device commands and clock scaling
>    * @desc_size: descriptor sizes reported by device
>    * @urgent_bkops_lvl: keeps track of urgent bkops level for device
>    * @is_urgent_bkops_lvl_checked: keeps track if the urgent bkops level for
> 


-- 
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
Linux Foundation Collaborative Project

  reply	other threads:[~2021-12-06 22:42 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-03 23:19 [PATCH v4 00/17] UFS patches for kernel v5.17 Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 01/17] scsi: core: Fix scsi_device_max_queue_depth() Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 02/17] scsi: ufs: Rename a function argument Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 03/17] scsi: ufs: Remove is_rpmb_wlun() Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 04/17] scsi: ufs: Remove the sdev_rpmb member Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 05/17] scsi: ufs: Remove dead code Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 06/17] scsi: ufs: Fix race conditions related to driver data Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 07/17] scsi: ufs: Remove ufshcd_any_tag_in_use() Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 08/17] scsi: ufs: Rework ufshcd_change_queue_depth() Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 09/17] scsi: ufs: Fix a deadlock in the error handler Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 10/17] scsi: ufs: Remove hba->cmd_queue Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 11/17] scsi: ufs: Remove the 'update_scaling' local variable Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 12/17] scsi: ufs: Introduce ufshcd_release_scsi_cmd() Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 13/17] scsi: ufs: Improve SCSI abort handling further Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 14/17] scsi: ufs: Fix a kernel crash during shutdown Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 15/17] scsi: ufs: Stop using the clock scaling lock in the error handler Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 16/17] scsi: ufs: Optimize the command queueing code Bart Van Assche
2021-12-06 22:41   ` Asutosh Das (asd) [this message]
2021-12-08 17:28     ` Asutosh Das (asd)
2021-12-08 17:53       ` Bart Van Assche
2021-12-14  4:04   ` Bjorn Andersson
2021-12-14  4:57     ` Bart Van Assche
2021-12-15  3:52       ` Bjorn Andersson
2021-12-15 18:44         ` Bart Van Assche
2021-12-03 23:19 ` [PATCH v4 17/17] scsi: ufs: Implement polling support Bart Van Assche
2021-12-07  3:31 ` [PATCH v4 00/17] UFS patches for kernel v5.17 Martin K. Petersen
2021-12-14  4:40 ` Martin K. Petersen
2021-12-14  7:14   ` Avri Altman
2021-12-14  7:18     ` Martin K. Petersen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=0ba5c50f-3e79-2cae-c502-59f70812cca3@codeaurora.org \
    --to=asutoshd@codeaurora.org \
    --cc=adrian.hunter@intel.com \
    --cc=avri.altman@wdc.com \
    --cc=beanhuo@micron.com \
    --cc=bvanassche@acm.org \
    --cc=cang@codeaurora.org \
    --cc=jaegeuk@kernel.org \
    --cc=jejb@linux.ibm.com \
    --cc=keosung.park@samsung.com \
    --cc=linux-scsi@vger.kernel.org \
    --cc=martin.petersen@oracle.com \
    --cc=stanley.chu@mediatek.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox