linux-block.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bart Van Assche <bvanassche@acm.org>
To: Jens Axboe <axboe@kernel.dk>
Cc: linux-block@vger.kernel.org, linux-scsi@vger.kernel.org,
	Jaegeuk Kim <jaegeuk@kernel.org>, Christoph Hellwig <hch@lst.de>,
	Bart Van Assche <bvanassche@acm.org>,
	"Martin K . Petersen" <martin.petersen@oracle.com>,
	Can Guo <quic_cang@quicinc.com>,
	Avri Altman <avri.altman@wdc.com>,
	Damien Le Moal <dlemoal@kernel.org>,
	Ming Lei <ming.lei@redhat.com>,
	"James E.J. Bottomley" <jejb@linux.ibm.com>,
	Stanley Chu <stanley.chu@mediatek.com>,
	Bean Huo <beanhuo@micron.com>,
	Asutosh Das <quic_asutoshd@quicinc.com>,
	"Bao D. Nguyen" <quic_nguyenb@quicinc.com>,
	Arthur Simchaev <Arthur.Simchaev@wdc.com>
Subject: [PATCH v7 7/7] scsi: ufs: Disable zone write locking
Date: Wed,  9 Aug 2023 13:23:48 -0700	[thread overview]
Message-ID: <20230809202355.1171455-8-bvanassche@acm.org> (raw)
In-Reply-To: <20230809202355.1171455-1-bvanassche@acm.org>

From the UFSHCI 4.0 specification, about the legacy (single queue) mode:
"The host controller always process transfer requests in-order according
to the order submitted to the list. In case of multiple commands with
single doorbell register ringing (batch mode), The dispatch order for
these transfer requests by host controller will base on their index in
the List. A transfer request with lower index value will be executed
before a transfer request with higher index value."

From the UFSHCI 4.0 specification, about the MCQ mode:
"Command Submission
1. Host SW writes an Entry to SQ
2. Host SW updates SQ doorbell tail pointer

Command Processing
3. After fetching the Entry, Host Controller updates SQ doorbell head
   pointer
4. Host controller sends COMMAND UPIU to UFS device"

In other words, for both legacy and MCQ mode, UFS controllers are
required to forward commands to the UFS device in the order these
commands have been received from the host.

Notes:
- For legacy mode this is only correct if the host submits one
  command at a time. The UFS driver does this.
- Also in legacy mode, the command order is not preserved if
  auto-hibernation is enabled in the UFS controller. Hence, enable
  zone write locking if auto-hibernation is enabled.

This patch improves performance as follows on my test setup:
- With the mq-deadline scheduler: 2.5x more IOPS for small writes.
- When not using an I/O scheduler compared to using mq-deadline with
  zone locking: 4x more IOPS for small writes.

Cc: Martin K. Petersen <martin.petersen@oracle.com>
Cc: Can Guo <quic_cang@quicinc.com>
Cc: Avri Altman <avri.altman@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Damien Le Moal <dlemoal@kernel.org>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 37 ++++++++++++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index ae7b868f9c26..ae6b63b02930 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -4337,23 +4337,50 @@ int ufshcd_uic_hibern8_exit(struct ufs_hba *hba)
 }
 EXPORT_SYMBOL_GPL(ufshcd_uic_hibern8_exit);
 
+static void ufshcd_update_zone_write_lock(struct ufs_hba *hba,
+					  bool use_zone_write_lock)
+{
+	struct scsi_device *sdev;
+
+	shost_for_each_device(sdev, hba->host)
+		blk_freeze_queue_start(sdev->request_queue);
+	shost_for_each_device(sdev, hba->host) {
+		struct request_queue *q = sdev->request_queue;
+
+		blk_mq_freeze_queue_wait(q);
+		q->limits.use_zone_write_lock = use_zone_write_lock;
+		blk_mq_unfreeze_queue(q);
+	}
+}
+
 void ufshcd_auto_hibern8_update(struct ufs_hba *hba, u32 ahit)
 {
 	unsigned long flags;
-	bool update = false;
+	bool prev_state, new_state, update = false;
 
 	if (!ufshcd_is_auto_hibern8_supported(hba))
 		return;
 
 	spin_lock_irqsave(hba->host->host_lock, flags);
+	prev_state = ufshcd_is_auto_hibern8_enabled(hba);
 	if (hba->ahit != ahit) {
 		hba->ahit = ahit;
 		update = true;
 	}
+	new_state = ufshcd_is_auto_hibern8_enabled(hba);
 	spin_unlock_irqrestore(hba->host->host_lock, flags);
 
 	if (!update)
 		return;
+	if (!is_mcq_enabled(hba) && !prev_state && new_state) {
+		/*
+		 * Auto-hibernation will be enabled. Enable write locking for
+		 * zoned writes since auto-hibernation may cause reordering of
+		 * zoned writes when using the legacy mode of the UFS host
+		 * controller.
+		 */
+		ufshcd_update_zone_write_lock(hba, true);
+	}
 	if (!pm_runtime_suspended(&hba->ufs_device_wlun->sdev_gendev)) {
 		ufshcd_rpm_get_sync(hba);
 		ufshcd_hold(hba);
@@ -4361,6 +4388,13 @@ void ufshcd_auto_hibern8_update(struct ufs_hba *hba, u32 ahit)
 		ufshcd_release(hba);
 		ufshcd_rpm_put_sync(hba);
 	}
+	if (!is_mcq_enabled(hba) && prev_state && !new_state) {
+		/*
+		 * Auto-hibernation has been disabled. Disable write locking
+		 * for zoned writes.
+		 */
+		ufshcd_update_zone_write_lock(hba, false);
+	}
 }
 EXPORT_SYMBOL_GPL(ufshcd_auto_hibern8_update);
 
@@ -5140,6 +5174,7 @@ static int ufshcd_slave_configure(struct scsi_device *sdev)
 
 	ufshcd_hpb_configure(hba, sdev);
 
+	q->limits.use_zone_write_lock = false;
 	blk_queue_update_dma_pad(q, PRDT_DATA_BYTE_COUNT_PAD - 1);
 	if (hba->quirks & UFSHCD_QUIRK_4KB_DMA_ALIGNMENT)
 		blk_queue_update_dma_alignment(q, SZ_4K - 1);

      parent reply	other threads:[~2023-08-09 20:24 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-09 20:23 [PATCH v7 0/7] Improve performance for zoned UFS devices Bart Van Assche
2023-08-09 20:23 ` [PATCH v7 1/7] block: Introduce the use_zone_write_lock member variable Bart Van Assche
2023-08-10  1:33   ` Damien Le Moal
2023-08-10 14:02     ` Bart Van Assche
2023-08-11  0:39       ` Damien Le Moal
2023-08-11 15:41         ` Bart Van Assche
2023-08-12  2:44           ` Damien Le Moal
2023-08-09 20:23 ` [PATCH v7 2/7] block/mq-deadline: Only use zone locking if necessary Bart Van Assche
2023-08-10  1:36   ` Damien Le Moal
2023-08-10 14:00     ` Bart Van Assche
2023-08-11  0:45       ` Damien Le Moal
2023-08-11 15:49         ` Bart Van Assche
2023-08-12  2:49           ` Damien Le Moal
2023-08-09 20:23 ` [PATCH v7 3/7] scsi: core: Retry unaligned zoned writes Bart Van Assche
2023-08-11 13:29   ` Christoph Hellwig
2023-08-09 20:23 ` [PATCH v7 4/7] scsi: scsi_debug: Support disabling zone write locking Bart Van Assche
2023-08-09 20:23 ` [PATCH v7 5/7] scsi: scsi_debug: Support injecting unaligned write errors Bart Van Assche
2023-08-09 20:23 ` [PATCH v7 6/7] scsi: ufs: Split an if-condition Bart Van Assche
2023-08-09 20:23 ` Bart Van Assche [this message]

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=20230809202355.1171455-8-bvanassche@acm.org \
    --to=bvanassche@acm.org \
    --cc=Arthur.Simchaev@wdc.com \
    --cc=avri.altman@wdc.com \
    --cc=axboe@kernel.dk \
    --cc=beanhuo@micron.com \
    --cc=dlemoal@kernel.org \
    --cc=hch@lst.de \
    --cc=jaegeuk@kernel.org \
    --cc=jejb@linux.ibm.com \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=martin.petersen@oracle.com \
    --cc=ming.lei@redhat.com \
    --cc=quic_asutoshd@quicinc.com \
    --cc=quic_cang@quicinc.com \
    --cc=quic_nguyenb@quicinc.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;
as well as URLs for NNTP newsgroup(s).