The Linux Kernel Mailing List
 help / color / mirror / Atom feed
From: John Garry <john.garry@huawei.com>
To: <jejb@linux.ibm.com>, <martin.petersen@oracle.com>
Cc: <linux-scsi@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<linuxarm@huawei.com>, Qi Liu <liuqi115@huawei.com>,
	John Garry <john.garry@huawei.com>
Subject: [PATCH 7/8] scsi: hisi_sas: Fix phyup timeout on FPGA
Date: Wed, 15 Dec 2021 22:37:40 +0800	[thread overview]
Message-ID: <1639579061-179473-8-git-send-email-john.garry@huawei.com> (raw)
In-Reply-To: <1639579061-179473-1-git-send-email-john.garry@huawei.com>

From: Qi Liu <liuqi115@huawei.com>

The OOB interrupt and phyup interrupt handlers may run out-of-order in
high CPU usage scenarios. Since the hisi_sas_phy.timer is added in
hisi_sas_phy_oob_ready() and disarmed in phy_up_v3_hw(), this
out-of-order execution will cause hisi_sas_phy.timer timeout to trigger.

To solve, protect hisi_sas_phy.timer and .attached with a lock, and ensure
that the timer won't be added after phyup handler completes.

Signed-off-by: Qi Liu <liuqi115@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_main.c  | 18 +++++++++++++-----
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 10 ++++++++--
 2 files changed, 21 insertions(+), 7 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index 0e14f90dbb1e..66e63a336770 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -909,10 +909,14 @@ void hisi_sas_phy_oob_ready(struct hisi_hba *hisi_hba, int phy_no)
 {
 	struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
 	struct device *dev = hisi_hba->dev;
+	unsigned long flags;
 
 	dev_dbg(dev, "phy%d OOB ready\n", phy_no);
-	if (phy->phy_attached)
+	spin_lock_irqsave(&phy->lock, flags);
+	if (phy->phy_attached) {
+		spin_unlock_irqrestore(&phy->lock, flags);
 		return;
+	}
 
 	if (!timer_pending(&phy->timer)) {
 		if (phy->wait_phyup_cnt < HISI_SAS_WAIT_PHYUP_RETRIES) {
@@ -920,13 +924,17 @@ void hisi_sas_phy_oob_ready(struct hisi_hba *hisi_hba, int phy_no)
 			phy->timer.expires = jiffies +
 					     HISI_SAS_WAIT_PHYUP_TIMEOUT;
 			add_timer(&phy->timer);
-		} else {
-			dev_warn(dev, "phy%d failed to come up %d times, giving up\n",
-				 phy_no, phy->wait_phyup_cnt);
-			phy->wait_phyup_cnt = 0;
+			spin_unlock_irqrestore(&phy->lock, flags);
+			return;
 		}
+
+		dev_warn(dev, "phy%d failed to come up %d times, giving up\n",
+			 phy_no, phy->wait_phyup_cnt);
+		phy->wait_phyup_cnt = 0;
 	}
+	spin_unlock_irqrestore(&phy->lock, flags);
 }
+
 EXPORT_SYMBOL_GPL(hisi_sas_phy_oob_ready);
 
 static void hisi_sas_phy_init(struct hisi_hba *hisi_hba, int phy_no)
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 11a44d9dd9b2..0239e2b4b84f 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -1484,7 +1484,6 @@ static irqreturn_t phy_up_v3_hw(int phy_no, struct hisi_hba *hisi_hba)
 	struct asd_sas_phy *sas_phy = &phy->sas_phy;
 	struct device *dev = hisi_hba->dev;
 
-	del_timer(&phy->timer);
 	hisi_sas_phy_write32(hisi_hba, phy_no, PHYCTRL_PHY_ENA_MSK, 1);
 
 	port_id = hisi_sas_read32(hisi_hba, PHY_PORT_NUM_MA);
@@ -1561,9 +1560,16 @@ static irqreturn_t phy_up_v3_hw(int phy_no, struct hisi_hba *hisi_hba)
 	}
 
 	phy->port_id = port_id;
-	phy->phy_attached = 1;
+
 	hisi_sas_notify_phy_event(phy, HISI_PHYE_PHY_UP);
+
 	res = IRQ_HANDLED;
+
+	spin_lock(&phy->lock);
+	/* Delete timer and set phy_attached atomically */
+	del_timer(&phy->timer);
+	phy->phy_attached = 1;
+	spin_unlock(&phy->lock);
 end:
 	if (phy->reset_completion)
 		complete(phy->reset_completion);
-- 
2.26.2


  parent reply	other threads:[~2021-12-15 14:43 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-15 14:37 [PATCH 0/8] hisi_sas: Some misc patches John Garry
2021-12-15 14:37 ` [PATCH 1/8] scsi: hisi_sas: Start delivery hisi_sas_task_exec() directly John Garry
2021-12-15 14:37 ` [PATCH 2/8] scsi: hisi_sas: Make internal abort have no task proto John Garry
2021-12-15 14:37 ` [PATCH 3/8] scsi: hisi_sas: Pass abort structure for internal abort John Garry
2021-12-15 14:37 ` [PATCH 4/8] scsi: hisi_sas: Factor out task prep and delivery code John Garry
2021-12-15 14:37 ` [PATCH 5/8] scsi: hisi_sas: Prevent parallel controller reset and control phy command John Garry
2021-12-15 14:37 ` [PATCH 6/8] scsi: hisi_sas: Prevent parallel FLR and controller reset John Garry
2021-12-15 14:37 ` John Garry [this message]
2021-12-15 14:37 ` [PATCH 8/8] scsi: libsas: Decode SAM status and host byte codes John Garry
2021-12-17  4:00 ` [PATCH 0/8] hisi_sas: Some misc patches Martin K. Petersen
2021-12-23  5:09 ` 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=1639579061-179473-8-git-send-email-john.garry@huawei.com \
    --to=john.garry@huawei.com \
    --cc=jejb@linux.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=linuxarm@huawei.com \
    --cc=liuqi115@huawei.com \
    --cc=martin.petersen@oracle.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