scsi: ufs: Allowing power mode change

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 8 +++++++-
 drivers/scsi/ufs/ufshcd.h | 7 +++++++
 2 files changed, 14 insertions(+), 1 deletion(-)

drivers/scsi/ufs/ufshcd.c +7 -1

index 5bc9dc1..f3083fe 100644
@@ -4162,9 +4162,15 @@ static int ufshcd_link_startup(struct ufs_hba *hba)
 			goto out;
 	} while (ret && retries--);
 
-	if (ret)
+	if (ret) {
 		/* failed to get the link up... retire */
 		goto out;
+	}
+
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE) {
+		ufshcd_dme_set(hba, UIC_ARG_MIB(TX_LCC_ENABLE), 0);
+		ufshcd_dme_set(hba, UIC_ARG_MIB(TX_LCC_ENABLE), 1);
+	}
 
 	if (link_startup_again) {
 		link_startup_again = false;

drivers/scsi/ufs/ufshcd.h +7 -0

index cbe46f6..bb4ecfb 100644
@@ -591,6 +591,13 @@ struct ufs_hba {
 	 */
 	#define UFSHCD_QUIRK_PRDT_BYTE_GRAN			UFS_BIT(7)
 
+	/*
+	 * Needs to be enabled if moving from HS to any other gear/mode or even
+	 * hibern8 causes unpredicted behavior of device.
+	 * If this quirk is enabled, standard UFS driver will disable/enable
+	 * TX_LCC.
+	 */
+	#define UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE		UFS_BIT(8)
 	unsigned int quirks;	/* Deviations from standard UFSHCI spec. */
 
 	/* Device deviations from standard UFS device spec. */

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

2018-05-16 21:09 UTC #2 lore raw

Quoted 49 lines · Asutosh Das

From: Yaniv Gardi <[email protected]>
Due to M-PHY issues, moving from HS to any other mode or gear or even Hibern8 may cause some un-predicted behavior
of the device.
This patch adds provides a quirk to address that.
Signed-off-by Yaniv Gardi <[email protected]>
Signed-off-by Subhash Jadavani <[email protected]>
Signed-off-by Can Guo <[email protected]>
Signed-off-by Asutosh Das <[email protected]>
---
 drivers/scsi/ufs/ufshcd.c | 8 +++++++-
 drivers/scsi/ufs/ufshcd.h | 7 +++++++
 2 files changed, 14 insertions(+), 1 deletion(-)
drivers/scsi/ufs/ufshcd.c +7 -1
index 5bc9dc1..f3083fe 100644
@@ -4162,9 +4162,15 @@ static int ufshcd_link_startup(struct ufs_hba 
 *hba)
 			goto out;
 	} while (ret && retries--);
 
-	if (ret)
+	if (ret) {
 		/* failed to get the link up... retire */
 		goto out;
+	}
+
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE) {
+		ufshcd_dme_set(hba, UIC_ARG_MIB(TX_LCC_ENABLE), 0);
+		ufshcd_dme_set(hba, UIC_ARG_MIB(TX_LCC_ENABLE), 1);
+	}
 
 	if (link_startup_again) {
drivers/scsi/ufs/ufshcd.c +0 -0
 		link_startup_again = false;
drivers/scsi/ufs/ufshcd.h +2 -0
index cbe46f6..bb4ecfb 100644
@@ -591,6 +591,13 @@ struct ufs_hba {
 	 */
 	#define UFSHCD_QUIRK_PRDT_BYTE_GRAN			UFS_BIT(7)
 
+	/*
+	 * Needs to be enabled if moving from HS to any other gear/mode or 
even

drivers/scsi/ufs/ufshcd.h struct ufs_hba { · line 603

Quoted 8 lines

drivers/scsi/ufs/ufshcd.h +5 -0

+	 * hibern8 causes unpredicted behavior of device.
+	 * If this quirk is enabled, standard UFS driver will disable/enable
+	 * TX_LCC.
+	 */
+	#define UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE		UFS_BIT(8)
 	unsigned int quirks;	/* Deviations from standard UFSHCI spec. */
 
 	/* Device deviations from standard UFS device spec. */

This was probably needed in early generation of UFS devices (non commercial) and may not be really needed now. I believe we can skip this patch.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #3 lore raw

PATCH v2 2/10 scsi: ufs-qcom: Enable UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE quirk

Enable UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE quirk to avoid failures in seen on some UFS devices.

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufs-qcom.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/scsi/ufs/ufs-qcom.c +1 -1

index 4563d2e..d9edef8 100644
@@ -1105,7 +1105,7 @@ static void ufs_qcom_advertise_quirks(struct ufs_hba *hba)
 
 	if (host->hw_ver.major >= 0x2) {
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_UFS_HCI_VERSION;
-
+		hba->quirks |= UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE;
 		if (!ufs_qcom_cap_qunipro(host))
 			/* Legacy UniPro mode still need following quirks */
 			hba->quirks |= (UFSHCD_QUIRK_DELAY_BEFORE_DME_CMDS

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

2018-05-16 21:10 UTC #4 lore raw

Re: [PATCH v2 02/10] scsi: ufs-qcom: Enable UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE quirk

Quoted 21 lines · Asutosh Das

Enable UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE quirk to avoid failures in seen on some UFS devices.

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufs-qcom.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/scsi/ufs/ufs-qcom.c +1 -1

index 4563d2e..d9edef8 100644
@@ -1105,7 +1105,7 @@ static void ufs_qcom_advertise_quirks(struct 
 ufs_hba *hba)
 
 	if (host->hw_ver.major >= 0x2) {
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_UFS_HCI_VERSION;
-
+		hba->quirks |= UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE;
 		if (!ufs_qcom_cap_qunipro(host))
 			/* Legacy UniPro mode still need following quirks */

Quoted 1 line

hba->quirks |= (UFSHCD_QUIRK_DELAY_BEFORE_DME_CMDS

We may not need this.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #5 lore raw

PATCH v2 3/10 scsi: ufs: Add LCC quirk for host and device

LCC (Line Control Command) is being used for communication between UFS host and UFS device. But some hosts might have the issue with issuing the LCC commands to UFS device and in this case LCC could be explicitly disabled.

But there could be a need where we don't want to disable the LCC on both host & device; hence this change splits the quirk in 2 parts one for host and one for device.

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 16 ++++++++++++++++
 drivers/scsi/ufs/ufshcd.h | 12 ++++++++++++
 2 files changed, 28 insertions(+)

drivers/scsi/ufs/ufshcd.c +16 -0

index f3083fe..6dabce8 100644
@@ -4116,6 +4116,11 @@ static int ufshcd_disable_tx_lcc(struct ufs_hba *hba, bool peer)
 	return err;
 }
 
+static inline int ufshcd_disable_host_tx_lcc(struct ufs_hba *hba)
+{
+	return ufshcd_disable_tx_lcc(hba, false);
+}
+
 static inline int ufshcd_disable_device_tx_lcc(struct ufs_hba *hba)
 {
 	return ufshcd_disable_tx_lcc(hba, true);

@@ -4172,6 +4177,17 @@ static int ufshcd_link_startup(struct ufs_hba *hba)
 		ufshcd_dme_set(hba, UIC_ARG_MIB(TX_LCC_ENABLE), 1);
 	}
 
+	if (hba->quirks & UFSHCD_BROKEN_LCC_PROCESSING_ON_HOST) {
+		ret = ufshcd_disable_device_tx_lcc(hba);
+		if (ret)
+			goto out;
+	}
+
+	if (hba->quirks & UFSHCD_BROKEN_LCC_PROCESSING_ON_DEVICE) {
+		ret = ufshcd_disable_host_tx_lcc(hba);
+		if (ret)
+			goto out;
+	}
 	if (link_startup_again) {
 		link_startup_again = false;
 		retries = DME_LINKSTARTUP_RETRIES;

drivers/scsi/ufs/ufshcd.h +12 -0

index bb4ecfb..0417c42 100644
@@ -598,6 +598,18 @@ struct ufs_hba {
 	 * TX_LCC.
 	 */
 	#define UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE		UFS_BIT(8)
+
+	/*
+	 * If UFS device is having issue in processing LCC (Line Control
+	 * Command) coming from UFS host controller then enable this quirk.
+	 * When this quirk is enabled, host controller driver should disable
+	 * the LCC transmission on UFS host controller (by clearing
+	 * TX_LCC_ENABLE attribute of host to 0).
+	 */
+	#define UFSHCD_BROKEN_LCC_PROCESSING_ON_DEVICE		0x100
+
+	#define UFSHCD_BROKEN_LCC_PROCESSING_ON_HOST		0x200
+
 	unsigned int quirks;	/* Deviations from standard UFSHCI spec. */
 
 	/* Device deviations from standard UFS device spec. */

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

2018-05-16 21:11 UTC #6 lore raw

Re: [PATCH v2 03/10] scsi: ufs: Add LCC quirk for host and device

Quoted 67 lines · Asutosh Das

But there could be a need where we don't want to disable the LCC on both host & device; hence this change splits the quirk in 2 parts one for host and one for device.

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 16 ++++++++++++++++
 drivers/scsi/ufs/ufshcd.h | 12 ++++++++++++
 2 files changed, 28 insertions(+)

drivers/scsi/ufs/ufshcd.c +5 -0

index f3083fe..6dabce8 100644
@@ -4116,6 +4116,11 @@ static int ufshcd_disable_tx_lcc(struct ufs_hba
 *hba, bool peer)
 	return err;
 }
 
+static inline int ufshcd_disable_host_tx_lcc(struct ufs_hba *hba)
+{
+	return ufshcd_disable_tx_lcc(hba, false);
+}
+
 static inline int ufshcd_disable_device_tx_lcc(struct ufs_hba *hba)
 {

drivers/scsi/ufs/ufshcd.c +11 -0

 	return ufshcd_disable_tx_lcc(hba, true);

@@ -4172,6 +4177,17 @@ static int ufshcd_link_startup(struct ufs_hba 
 *hba)
 		ufshcd_dme_set(hba, UIC_ARG_MIB(TX_LCC_ENABLE), 1);
 	}
 
+	if (hba->quirks & UFSHCD_BROKEN_LCC_PROCESSING_ON_HOST) {
+		ret = ufshcd_disable_device_tx_lcc(hba);
+		if (ret)
+			goto out;
+	}
+
+	if (hba->quirks & UFSHCD_BROKEN_LCC_PROCESSING_ON_DEVICE) {
+		ret = ufshcd_disable_host_tx_lcc(hba);
+		if (ret)
+			goto out;
+	}
 	if (link_startup_again) {
 		link_startup_again = false;

drivers/scsi/ufs/ufshcd.c +0 -0

 		retries = DME_LINKSTARTUP_RETRIES;

drivers/scsi/ufs/ufshcd.h +5 -0

index bb4ecfb..0417c42 100644
@@ -598,6 +598,18 @@ struct ufs_hba {
 	 * TX_LCC.
 	 */
 	#define UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE		UFS_BIT(8)
+
+	/*
+	 * If UFS device is having issue in processing LCC (Line Control
+	 * Command) coming from UFS host controller then enable this quirk.
+	 * When this quirk is enabled, host controller driver should disable

drivers/scsi/ufs/ufshcd.h struct ufs_hba { · line 615

Quoted 10 lines

drivers/scsi/ufs/ufshcd.h +7 -0

+	 * the LCC transmission on UFS host controller (by clearing
+	 * TX_LCC_ENABLE attribute of host to 0).
+	 */
+	#define UFSHCD_BROKEN_LCC_PROCESSING_ON_DEVICE		0x100
+
+	#define UFSHCD_BROKEN_LCC_PROCESSING_ON_HOST		0x200
+
 	unsigned int quirks;	/* Deviations from standard UFSHCI spec. */
 
 	/* Device deviations from standard UFS device spec. */

Please check if these quirks are really needed for commercial version of UFS host controllers.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #7 lore raw

PATCH v2 4/10 scsi: ufs: fix exception event handling

From: Maya Erez <[email protected]>

The device can set the exception event bit in one of the response UPIU, for example to notify the need for urgent BKOPs operation.
In such a case the host driver calls ufshcd_exception_event_handler to handle this notification.
When trying to check the exception event status (for finding the cause for the exception event), the device may be busy with additional SCSI commands handling and may not respond within the 100ms timeout.

To prevent that, we need to block SCSI commands during handling of exception events and allow retransmissions of the query requests, in case of timeout.

Signed-off-by Maya Erez <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 2 ++
 1 file changed, 2 insertions(+)

drivers/scsi/ufs/ufshcd.c +2 -0

index 6dabce8..838ba8f0 100644
@@ -4967,6 +4967,7 @@ static void ufshcd_exception_event_handler(struct work_struct *work)
 	hba = container_of(work, struct ufs_hba, eeh_work);
 
 	pm_runtime_get_sync(hba->dev);
+	scsi_block_requests(hba->host);
 	err = ufshcd_get_ee_status(hba, &status);
 	if (err) {
 		dev_err(hba->dev, "%s: failed to get exception status %d\n",

@@ -4980,6 +4981,7 @@ static void ufshcd_exception_event_handler(struct work_struct *work)
 		ufshcd_bkops_exception_event_handler(hba);
 
 out:
+	scsi_unblock_requests(hba->host);
 	pm_runtime_put_sync(hba->dev);
 	return;
 }

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

Reviewed-by on v2 4/10

2018-05-16 21:12 UTC #8 lore raw

Re: [PATCH v2 04/10] scsi: ufs: fix exception event handling

Quoted 43 lines · Asutosh Das

From: Maya Erez <[email protected]>
The device can set the exception event bit in one of the response UPIU, for example to notify the need for urgent BKOPs operation.
In such a case the host driver calls ufshcd_exception_event_handler to handle this notification.
When trying to check the exception event status (for finding the cause for the exception event), the device may be busy with additional SCSI commands handling and may not respond within the 100ms timeout.
To prevent that, we need to block SCSI commands during handling of exception events and allow retransmissions of the query requests, in case of timeout.
Signed-off-by Subhash Jadavani <[email protected]>
Signed-off-by Maya Erez <[email protected]>
Signed-off-by Can Guo <[email protected]>
Signed-off-by Asutosh Das <[email protected]>
---
 drivers/scsi/ufs/ufshcd.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/ufs/ufshcd.c +1 -0
index 6dabce8..838ba8f0 100644
@@ -4967,6 +4967,7 @@ static void
 ufshcd_exception_event_handler(struct work_struct *work)
 	hba = container_of(work, struct ufs_hba, eeh_work);
 
 	pm_runtime_get_sync(hba->dev);
+	scsi_block_requests(hba->host);
 	err = ufshcd_get_ee_status(hba, &status);
 	if (err) {
drivers/scsi/ufs/ufshcd.c +1 -0
 		dev_err(hba->dev, "%s: failed to get exception status %d\n",
@@ -4980,6 +4981,7 @@ static void
 ufshcd_exception_event_handler(struct work_struct *work)
 		ufshcd_bkops_exception_event_handler(hba);
 
 out:
+	scsi_unblock_requests(hba->host);
 	pm_runtime_put_sync(hba->dev);
 	return;

Quoted 1 line

}

Looks good to me.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #9 lore raw

PATCH v2 5/10 scsi: ufshcd: fix possible unclocked register access

From: Subhash Jadavani <[email protected]>

vendor specific setup_clocks ops may depend on clocks managed by ufshcd driver so if the vendor specific setup_clocks callback is called when the required clocks are turned off, it results into unclocked register access.

This change make sure that required clocks are enabled before vendor specific setup_clocks callback is called.

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 26 ++++++++++++++++++++------
 1 file changed, 20 insertions(+), 6 deletions(-)

drivers/scsi/ufs/ufshcd.c +20 -6

index 838ba8f0..dfeb194 100644
@@ -6780,9 +6780,16 @@ static int __ufshcd_setup_clocks(struct ufs_hba *hba, bool on,
 	if (list_empty(head))
 		goto out;
 
-	ret = ufshcd_vops_setup_clocks(hba, on, PRE_CHANGE);
-	if (ret)
-		return ret;
+	/*
+	 * vendor specific setup_clocks ops may depend on clocks managed by
+	 * this standard driver hence call the vendor specific setup_clocks
+	 * before disabling the clocks managed here.
+	 */
+	if (!on) {
+		ret = ufshcd_vops_setup_clocks(hba, on, PRE_CHANGE);
+		if (ret)
+			return ret;
+	}
 
 	list_for_each_entry(clki, head, list) {
 		if (!IS_ERR_OR_NULL(clki->clk)) {

@@ -6806,9 +6813,16 @@ static int __ufshcd_setup_clocks(struct ufs_hba *hba, bool on,
 		}
 	}
 
-	ret = ufshcd_vops_setup_clocks(hba, on, POST_CHANGE);
-	if (ret)
-		return ret;
+	/*
+	 * vendor specific setup_clocks ops may depend on clocks managed by
+	 * this standard driver hence call the vendor specific setup_clocks
+	 * after enabling the clocks managed here.
+	 */
+	if (on) {
+		ret = ufshcd_vops_setup_clocks(hba, on, POST_CHANGE);
+		if (ret)
+			return ret;
+	}
 
 out:
 	if (ret) {

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

2018-05-16 21:12 UTC #10 lore raw

Re: [PATCH v2 05/10] scsi: ufshcd: fix possible unclocked register access

Quoted 63 lines · Asutosh Das

From: Subhash Jadavani <[email protected]>

This change make sure that required clocks are enabled before vendor specific setup_clocks callback is called.

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 26 ++++++++++++++++++++------
 1 file changed, 20 insertions(+), 6 deletions(-)

drivers/scsi/ufs/ufshcd.c +10 -3

index 838ba8f0..dfeb194 100644
@@ -6780,9 +6780,16 @@ static int __ufshcd_setup_clocks(struct ufs_hba
 *hba, bool on,
 	if (list_empty(head))
 		goto out;
 
-	ret = ufshcd_vops_setup_clocks(hba, on, PRE_CHANGE);
-	if (ret)
-		return ret;
+	/*
+	 * vendor specific setup_clocks ops may depend on clocks managed by
+	 * this standard driver hence call the vendor specific setup_clocks
+	 * before disabling the clocks managed here.
+	 */
+	if (!on) {
+		ret = ufshcd_vops_setup_clocks(hba, on, PRE_CHANGE);
+		if (ret)
+			return ret;
+	}
 
 	list_for_each_entry(clki, head, list) {

drivers/scsi/ufs/ufshcd.c +10 -3

 		if (!IS_ERR_OR_NULL(clki->clk)) {

@@ -6806,9 +6813,16 @@ static int __ufshcd_setup_clocks(struct ufs_hba
 *hba, bool on,
 		}
 	}
 
-	ret = ufshcd_vops_setup_clocks(hba, on, POST_CHANGE);
-	if (ret)
-		return ret;
+	/*
+	 * vendor specific setup_clocks ops may depend on clocks managed by
+	 * this standard driver hence call the vendor specific setup_clocks
+	 * after enabling the clocks managed here.
+	 */
+	if (on) {
+		ret = ufshcd_vops_setup_clocks(hba, on, POST_CHANGE);
+		if (ret)
+			return ret;
+	}
 
 out:

Quoted 1 line

if (ret) {

Looks good to me.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #11 lore raw

PATCH v2 6/10 scsi: ufs: add reference counting for scsi block requests

From: Subhash Jadavani <[email protected]>

Currently we call the scsi_block_requests()/scsi_unblock_requests() whenever we want to block/unblock scsi requests but as there is no reference counting, nesting of these calls could leave us in undesired state sometime. Consider following call flow sequence:
1. func1() calls scsi_block_requests() but calls func2() before calling scsi_unblock_requests()
2. func2() calls scsi_block_requests()
3. func2() calls scsi_unblock_requests()
4. func1() calls scsi_unblock_requests()

As there is no reference counting, we will have scsi requests unblocked after #3 instead of it to be unblocked only after #4. Though we may not have failures seen with this, we might run into some failures in future. Better solution would be to fix this by adding reference counting.

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 28 ++++++++++++++++++++--------
 drivers/scsi/ufs/ufshcd.h |  2 ++
 2 files changed, 22 insertions(+), 8 deletions(-)

drivers/scsi/ufs/ufshcd.c +20 -8

index dfeb194..c35a076 100644
@@ -264,6 +264,18 @@ static inline void ufshcd_disable_irq(struct ufs_hba *hba)
 	}
 }
 
+static void ufshcd_scsi_unblock_requests(struct ufs_hba *hba)
+{
+	if (atomic_dec_and_test(&hba->scsi_block_reqs_cnt))
+		scsi_unblock_requests(hba->host);
+}
+
+static void ufshcd_scsi_block_requests(struct ufs_hba *hba)
+{
+	if (atomic_inc_return(&hba->scsi_block_reqs_cnt) == 1)
+		scsi_block_requests(hba->host);
+}
+
 /* replace non-printable or non-ASCII characters with spaces */
 static inline void ufshcd_remove_non_printable(char *val)
 {

@@ -1077,12 +1089,12 @@ static int ufshcd_clock_scaling_prepare(struct ufs_hba *hba)
 	 * make sure that there are no outstanding requests when
 	 * clock scaling is in progress
 	 */
-	scsi_block_requests(hba->host);
+	ufshcd_scsi_block_requests(hba);
 	down_write(&hba->clk_scaling_lock);
 	if (ufshcd_wait_for_doorbell_clr(hba, DOORBELL_CLR_TOUT_US)) {
 		ret = -EBUSY;
 		up_write(&hba->clk_scaling_lock);
-		scsi_unblock_requests(hba->host);
+		ufshcd_scsi_unblock_requests(hba);
 	}
 
 	return ret;

@@ -1091,7 +1103,7 @@ static int ufshcd_clock_scaling_prepare(struct ufs_hba *hba)
 static void ufshcd_clock_scaling_unprepare(struct ufs_hba *hba)
 {
 	up_write(&hba->clk_scaling_lock);
-	scsi_unblock_requests(hba->host);
+	ufshcd_scsi_unblock_requests(hba);
 }
 
 /**

@@ -1411,7 +1423,7 @@ static void ufshcd_ungate_work(struct work_struct *work)
 		hba->clk_gating.is_suspended = false;
 	}
 unblock_reqs:
-	scsi_unblock_requests(hba->host);
+	ufshcd_scsi_unblock_requests(hba);
 }
 
 /**

@@ -1467,7 +1479,7 @@ int ufshcd_hold(struct ufs_hba *hba, bool async)
 		 * work and to enable clocks.
 		 */
 	case CLKS_OFF:
-		scsi_block_requests(hba->host);
+		ufshcd_scsi_block_requests(hba);
 		hba->clk_gating.state = REQ_CLKS_ON;
 		trace_ufshcd_clk_gating(dev_name(hba->dev),
 					hba->clk_gating.state);

@@ -5192,7 +5204,7 @@ static void ufshcd_err_handler(struct work_struct *work)
 
 out:
 	spin_unlock_irqrestore(hba->host->host_lock, flags);
-	scsi_unblock_requests(hba->host);
+	ufshcd_scsi_unblock_requests(hba);
 	ufshcd_release(hba);
 	pm_runtime_put_sync(hba->dev);
 }

@@ -5294,7 +5306,7 @@ static void ufshcd_check_errors(struct ufs_hba *hba)
 		/* handle fatal errors only when link is functional */
 		if (hba->ufshcd_state == UFSHCD_STATE_OPERATIONAL) {
 			/* block commands from scsi mid-layer */
-			scsi_block_requests(hba->host);
+			ufshcd_scsi_block_requests(hba);
 
 			hba->ufshcd_state = UFSHCD_STATE_EH_SCHEDULED;

@@ -8017,7 +8029,7 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 
 	/* Hold auto suspend until async scan completes */
 	pm_runtime_get_sync(dev);
-
+	atomic_set(&hba->scsi_block_reqs_cnt, 0);
 	/*
 	 * We are assuming that device wasn't put in sleep/power-down
 	 * state exclusively during the boot stage before kernel.

drivers/scsi/ufs/ufshcd.h +2 -0

index 0417c42..76c31d5 100644
@@ -498,6 +498,7 @@ struct ufs_stats {
  * @urgent_bkops_lvl: keeps track of urgent bkops level for device
  * @is_urgent_bkops_lvl_checked: keeps track if the urgent bkops level for
  *  device is known or not.
+ * @scsi_block_reqs_cnt: reference counting for scsi block requests
  */
 struct ufs_hba {
 	void __iomem *mmio_base;

@@ -698,6 +699,7 @@ struct ufs_hba {
 
 	struct rw_semaphore clk_scaling_lock;
 	struct ufs_desc_size desc_size;
+	atomic_t scsi_block_reqs_cnt;
 };
 
 /* Returns true if clocks can be gated. Otherwise false */

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

2018-05-16 21:13 UTC #12 lore raw

Re: [PATCH v2 06/10] scsi: ufs: add reference counting for scsi block requests

Quoted 138 lines · Asutosh Das

From: Subhash Jadavani <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 28 ++++++++++++++++++++--------
 drivers/scsi/ufs/ufshcd.h |  2 ++
 2 files changed, 22 insertions(+), 8 deletions(-)

drivers/scsi/ufs/ufshcd.c +12 -0

index dfeb194..c35a076 100644
@@ -264,6 +264,18 @@ static inline void ufshcd_disable_irq(struct 
 ufs_hba *hba)
 	}
 }
 
+static void ufshcd_scsi_unblock_requests(struct ufs_hba *hba)
+{
+	if (atomic_dec_and_test(&hba->scsi_block_reqs_cnt))
+		scsi_unblock_requests(hba->host);
+}
+
+static void ufshcd_scsi_block_requests(struct ufs_hba *hba)
+{
+	if (atomic_inc_return(&hba->scsi_block_reqs_cnt) == 1)
+		scsi_block_requests(hba->host);
+}
+
 /* replace non-printable or non-ASCII characters with spaces */
 static inline void ufshcd_remove_non_printable(char *val)

{

drivers/scsi/ufs/ufshcd.c +2 -2

@@ -1077,12 +1089,12 @@ static int ufshcd_clock_scaling_prepare(struct
 ufs_hba *hba)
 	 * make sure that there are no outstanding requests when
 	 * clock scaling is in progress
 	 */
-	scsi_block_requests(hba->host);
+	ufshcd_scsi_block_requests(hba);
 	down_write(&hba->clk_scaling_lock);
 	if (ufshcd_wait_for_doorbell_clr(hba, DOORBELL_CLR_TOUT_US)) {
 		ret = -EBUSY;
 		up_write(&hba->clk_scaling_lock);
-		scsi_unblock_requests(hba->host);
+		ufshcd_scsi_unblock_requests(hba);
 	}

drivers/scsi/ufs/ufshcd.c +1 -1

 	return ret;

@@ -1091,7 +1103,7 @@ static int ufshcd_clock_scaling_prepare(struct
 ufs_hba *hba)
 static void ufshcd_clock_scaling_unprepare(struct ufs_hba *hba)
 {
 	up_write(&hba->clk_scaling_lock);
-	scsi_unblock_requests(hba->host);
+	ufshcd_scsi_unblock_requests(hba);
 }

/**

drivers/scsi/ufs/ufshcd.c +1 -1

@@ -1411,7 +1423,7 @@ static void ufshcd_ungate_work(struct work_struct 
 *work)
 		hba->clk_gating.is_suspended = false;
 	}
 unblock_reqs:
-	scsi_unblock_requests(hba->host);
+	ufshcd_scsi_unblock_requests(hba);
 }

/**

drivers/scsi/ufs/ufshcd.c +2 -2

@@ -1467,7 +1479,7 @@ int ufshcd_hold(struct ufs_hba *hba, bool async)
 		 * work and to enable clocks.
 		 */
 	case CLKS_OFF:
-		scsi_block_requests(hba->host);
+		ufshcd_scsi_block_requests(hba);
 		hba->clk_gating.state = REQ_CLKS_ON;
 		trace_ufshcd_clk_gating(dev_name(hba->dev),
 					hba->clk_gating.state);

@@ -5192,7 +5204,7 @@ static void ufshcd_err_handler(struct work_struct 
 *work)
 
 out:
 	spin_unlock_irqrestore(hba->host->host_lock, flags);
-	scsi_unblock_requests(hba->host);
+	ufshcd_scsi_unblock_requests(hba);
 	ufshcd_release(hba);
 	pm_runtime_put_sync(hba->dev);

}

drivers/scsi/ufs/ufshcd.c +2 -2

@@ -5294,7 +5306,7 @@ static void ufshcd_check_errors(struct ufs_hba 
 *hba)
 		/* handle fatal errors only when link is functional */
 		if (hba->ufshcd_state == UFSHCD_STATE_OPERATIONAL) {
 			/* block commands from scsi mid-layer */
-			scsi_block_requests(hba->host);
+			ufshcd_scsi_block_requests(hba);
 
 			hba->ufshcd_state = UFSHCD_STATE_EH_SCHEDULED;

@@ -8017,7 +8029,7 @@ int ufshcd_init(struct ufs_hba *hba, void
 __iomem *mmio_base, unsigned int irq)
 
 	/* Hold auto suspend until async scan completes */
 	pm_runtime_get_sync(dev);
-
+	atomic_set(&hba->scsi_block_reqs_cnt, 0);
 	/*
 	 * We are assuming that device wasn't put in sleep/power-down

drivers/scsi/ufs/ufshcd.c +0 -0

 	 * state exclusively during the boot stage before kernel.

drivers/scsi/ufs/ufshcd.h +0 -0

index 0417c42..76c31d5 100644
@@ -498,6 +498,7 @@ struct ufs_stats {
  * @urgent_bkops_lvl: keeps track of urgent bkops level for device
  * @is_urgent_bkops_lvl_checked: keeps track if the urgent bkops level

for
* device is known or not.
+ * @scsi_block_reqs_cnt: reference counting for scsi block requests */
struct ufs_hba {

drivers/scsi/ufs/ufshcd.h +0 -0

 	void __iomem *mmio_base;

drivers/scsi/ufs/ufshcd.h struct ufs_hba { · line 705

Quoted 7 lines

drivers/scsi/ufs/ufshcd.h +1 -0

@@ -698,6 +699,7 @@ struct ufs_hba {
 
 	struct rw_semaphore clk_scaling_lock;
 	struct ufs_desc_size desc_size;
+	atomic_t scsi_block_reqs_cnt;
 };
 
 /* Returns true if clocks can be gated. Otherwise false */

Looks good to me.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #13 lore raw

PATCH v2 7/10 scsi: ufs-qcom: remove broken hci version quirk

From: Subhash Jadavani <[email protected]>

UFSHCD_QUIRK_BROKEN_UFS_HCI_VERSION is only applicable for QCOM UFS host controller version 2.x.y and this has been fixed from version 3.x.y onwards, hence this change removes this quirk for version 3.x.y onwards.

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufs-qcom.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/scsi/ufs/ufs-qcom.c +1 -1

index d9edef8..27be327 100644
@@ -1103,7 +1103,7 @@ static void ufs_qcom_advertise_quirks(struct ufs_hba *hba)
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_LCC;
 	}
 
-	if (host->hw_ver.major >= 0x2) {
+	if (host->hw_ver.major == 0x2) {
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_UFS_HCI_VERSION;
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE;
 		if (!ufs_qcom_cap_qunipro(host))

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

2018-05-16 21:13 UTC #14 lore raw

Re: [PATCH v2 07/10] scsi: ufs-qcom: remove broken hci version quirk

Quoted 25 lines · Asutosh Das

From: Subhash Jadavani <[email protected]>
UFSHCD_QUIRK_BROKEN_UFS_HCI_VERSION is only applicable for QCOM UFS host controller version 2.x.y and this has been fixed from version 3.x.y onwards, hence this change removes this quirk for version 3.x.y onwards.
Signed-off-by Subhash Jadavani <[email protected]>
Signed-off-by Asutosh Das <[email protected]>
---
 drivers/scsi/ufs/ufs-qcom.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/ufs/ufs-qcom.c +1 -1
index d9edef8..27be327 100644
@@ -1103,7 +1103,7 @@ static void ufs_qcom_advertise_quirks(struct 
 ufs_hba *hba)
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_LCC;
 	}
 
-	if (host->hw_ver.major >= 0x2) {
+	if (host->hw_ver.major == 0x2) {
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_UFS_HCI_VERSION;
 		hba->quirks |= UFSHCD_QUIRK_BROKEN_PWR_MODE_CHANGE;

Quoted 1 line

if (!ufs_qcom_cap_qunipro(host))

Looks good to me.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #15 lore raw

PATCH v2 8/10 scsi: ufs: make sure all interrupts are processed

From: Venkat Gopalakrishnan <[email protected]>

As multiple requests are submitted to the ufs host controller in parallel there could be instances where the command completion interrupt arrives later for a request that is already processed earlier as the corresponding doorbell was cleared when handling the previous interrupt. Read the interrupt status in a loop after processing the received interrupt to catch such interrupts and handle it.

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 27 +++++++++++++++++++--------
 1 file changed, 19 insertions(+), 8 deletions(-)

drivers/scsi/ufs/ufshcd.c +19 -8

index c35a076..09b7a3f 100644
@@ -5383,19 +5383,30 @@ static irqreturn_t ufshcd_intr(int irq, void *__hba)
 	u32 intr_status, enabled_intr_status;
 	irqreturn_t retval = IRQ_NONE;
 	struct ufs_hba *hba = __hba;
+	int retries = hba->nutrs;
 
 	spin_lock(hba->host->host_lock);
 	intr_status = ufshcd_readl(hba, REG_INTERRUPT_STATUS);
-	enabled_intr_status =
-		intr_status & ufshcd_readl(hba, REG_INTERRUPT_ENABLE);
 
-	if (intr_status)
-		ufshcd_writel(hba, intr_status, REG_INTERRUPT_STATUS);
+	/*
+	 * There could be max of hba->nutrs reqs in flight and in worst case
+	 * if the reqs get finished 1 by 1 after the interrupt status is
+	 * read, make sure we handle them by checking the interrupt status
+	 * again in a loop until we process all of the reqs before returning.
+	 */
+	do {
+		enabled_intr_status =
+			intr_status & ufshcd_readl(hba, REG_INTERRUPT_ENABLE);
+		if (intr_status)
+			ufshcd_writel(hba, intr_status, REG_INTERRUPT_STATUS);
+		if (enabled_intr_status) {
+			ufshcd_sl_intr(hba, enabled_intr_status);
+			retval = IRQ_HANDLED;
+		}
+
+		intr_status = ufshcd_readl(hba, REG_INTERRUPT_STATUS);
+	} while (intr_status && --retries);
 
-	if (enabled_intr_status) {
-		ufshcd_sl_intr(hba, enabled_intr_status);
-		retval = IRQ_HANDLED;
-	}
 	spin_unlock(hba->host->host_lock);
 	return retval;
 }

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

Reviewed-by on v2 8/10

2018-05-16 21:14 UTC #16 lore raw

Re: [PATCH v2 08/10] scsi: ufs: make sure all interrupts are processed

Quoted 59 lines · Asutosh Das

From: Venkat Gopalakrishnan <[email protected]>

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 27 +++++++++++++++++++--------
 1 file changed, 19 insertions(+), 8 deletions(-)

drivers/scsi/ufs/ufshcd.c +19 -8

index c35a076..09b7a3f 100644
@@ -5383,19 +5383,30 @@ static irqreturn_t ufshcd_intr(int irq, void 
 *__hba)
 	u32 intr_status, enabled_intr_status;
 	irqreturn_t retval = IRQ_NONE;
 	struct ufs_hba *hba = __hba;
+	int retries = hba->nutrs;
 
 	spin_lock(hba->host->host_lock);
 	intr_status = ufshcd_readl(hba, REG_INTERRUPT_STATUS);
-	enabled_intr_status =
-		intr_status & ufshcd_readl(hba, REG_INTERRUPT_ENABLE);
 
-	if (intr_status)
-		ufshcd_writel(hba, intr_status, REG_INTERRUPT_STATUS);
+	/*
+	 * There could be max of hba->nutrs reqs in flight and in worst case
+	 * if the reqs get finished 1 by 1 after the interrupt status is
+	 * read, make sure we handle them by checking the interrupt status
+	 * again in a loop until we process all of the reqs before returning.
+	 */
+	do {
+		enabled_intr_status =
+			intr_status & ufshcd_readl(hba, REG_INTERRUPT_ENABLE);
+		if (intr_status)
+			ufshcd_writel(hba, intr_status, REG_INTERRUPT_STATUS);
+		if (enabled_intr_status) {
+			ufshcd_sl_intr(hba, enabled_intr_status);
+			retval = IRQ_HANDLED;
+		}
+
+		intr_status = ufshcd_readl(hba, REG_INTERRUPT_STATUS);
+	} while (intr_status && --retries);
 
-	if (enabled_intr_status) {
-		ufshcd_sl_intr(hba, enabled_intr_status);
-		retval = IRQ_HANDLED;
-	}
 	spin_unlock(hba->host->host_lock);
 	return retval;

Quoted 1 line

}

Looks good to me.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #17 lore raw

PATCH v2 9/10 scsi: ufs: fix irq return code

From: Venkat Gopalakrishnan <[email protected]>

Return IRQ_HANDLED only if the irq is really handled, this will help in catching spurious interrupts that go unhandled.

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 137 ++++++++++++++++++++++++++++++++++------------
 1 file changed, 101 insertions(+), 36 deletions(-)

drivers/scsi/ufs/ufshcd.c +101 ~~-36~~

index 09b7a3f..557d538 100644
@@ -211,7 +211,7 @@ enum {
 	END_FIX
 };
 
-static void ufshcd_tmc_handler(struct ufs_hba *hba);
+static irqreturn_t ufshcd_tmc_handler(struct ufs_hba *hba);
 static void ufshcd_async_scan(void *data, async_cookie_t cookie);
 static int ufshcd_reset_and_restore(struct ufs_hba *hba);
 static int ufshcd_eh_host_reset_handler(struct scsi_cmnd *cmd);

@@ -4609,19 +4609,29 @@ static int ufshcd_task_req_compl(struct ufs_hba *hba, u32 index, u8 *resp)
  * ufshcd_uic_cmd_compl - handle completion of uic command
  * @hba: per adapter instance
  * @intr_status: interrupt status generated by the controller
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 intr_status)
+static irqreturn_t ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 intr_status)
 {
+	irqreturn_t retval = IRQ_NONE;
+
 	if ((intr_status & UIC_COMMAND_COMPL) && hba->active_uic_cmd) {
 		hba->active_uic_cmd->argument2 |=
 			ufshcd_get_uic_cmd_result(hba);
 		hba->active_uic_cmd->argument3 =
 			ufshcd_get_dme_attr_val(hba);
 		complete(&hba->active_uic_cmd->done);
+		retval = IRQ_HANDLED;
 	}
 
-	if ((intr_status & UFSHCD_UIC_PWR_MASK) && hba->uic_async_done)
+	if ((intr_status & UFSHCD_UIC_PWR_MASK) && hba->uic_async_done) {
 		complete(hba->uic_async_done);
+		retval = IRQ_HANDLED;
+	}
+	return retval;
 }
 
 /**

@@ -4675,8 +4685,12 @@ static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
 /**
  * ufshcd_transfer_req_compl - handle SCSI and query command completion
  * @hba: per adapter instance
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
+static irqreturn_t ufshcd_transfer_req_compl(struct ufs_hba *hba)
 {
 	unsigned long completed_reqs;
 	u32 tr_doorbell;

@@ -4694,7 +4708,12 @@ static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
 	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
 	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
 
-	__ufshcd_transfer_req_compl(hba, completed_reqs);
+	if (completed_reqs) {
+		__ufshcd_transfer_req_compl(hba, completed_reqs);
+		return IRQ_HANDLED;
+	} else {
+		return IRQ_NONE;
+	}
 }
 
 /**

@@ -5220,16 +5239,21 @@ static void ufshcd_update_uic_reg_hist(struct ufs_uic_err_reg_hist *reg_hist,
 /**
  * ufshcd_update_uic_error - check and set fatal UIC error flags.
  * @hba: per-adapter instance
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_update_uic_error(struct ufs_hba *hba)
+static irqreturn_t ufshcd_update_uic_error(struct ufs_hba *hba)
 {
 	u32 reg;
+	irqreturn_t retval = IRQ_NONE;
 
 	/* PHY layer lane error */
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_PHY_ADAPTER_LAYER);
 	/* Ignore LINERESET indication, as this is not an error */
 	if ((reg & UIC_PHY_ADAPTER_LAYER_ERROR) &&
-			(reg & UIC_PHY_ADAPTER_LAYER_LANE_ERR_MASK)) {
+			(reg & UIC_PHY_ADAPTER_LAYER_ERROR_CODE_MASK)) {
 		/*
 		 * To know whether this error is fatal or not, DB timeout
 		 * must be checked but this error is handled separately.

@@ -5240,57 +5264,73 @@ static void ufshcd_update_uic_error(struct ufs_hba *hba)
 
 	/* PA_INIT_ERROR is fatal and needs UIC reset */
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_DATA_LINK_LAYER);
-	if (reg)
+	if ((reg & UIC_DATA_LINK_LAYER_ERROR) &&
+	    (reg & UIC_DATA_LINK_LAYER_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.dl_err, reg);
-
-	if (reg & UIC_DATA_LINK_LAYER_ERROR_PA_INIT)
-		hba->uic_error |= UFSHCD_UIC_DL_PA_INIT_ERROR;
-	else if (hba->dev_quirks &
-		   UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS) {
-		if (reg & UIC_DATA_LINK_LAYER_ERROR_NAC_RECEIVED)
-			hba->uic_error |=
-				UFSHCD_UIC_DL_NAC_RECEIVED_ERROR;
-		else if (reg & UIC_DATA_LINK_LAYER_ERROR_TCx_REPLAY_TIMEOUT)
-			hba->uic_error |= UFSHCD_UIC_DL_TCx_REPLAY_ERROR;
+		if (reg & UIC_DATA_LINK_LAYER_ERROR_PA_INIT) {
+			hba->uic_error |= UFSHCD_UIC_DL_PA_INIT_ERROR;
+		} else if (hba->dev_quirks &
+			   UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS) {
+			if (reg & UIC_DATA_LINK_LAYER_ERROR_NAC_RECEIVED)
+				hba->uic_error |=
+					UFSHCD_UIC_DL_NAC_RECEIVED_ERROR;
+			else if (reg &
+				UIC_DATA_LINK_LAYER_ERROR_TCx_REPLAY_TIMEOUT)
+				hba->uic_error |=
+					UFSHCD_UIC_DL_TCx_REPLAY_ERROR;
+		}
+		retval |= IRQ_HANDLED;
 	}
 
 	/* UIC NL/TL/DME errors needs software retry */
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_NETWORK_LAYER);
-	if (reg) {
+	if ((reg & UIC_NETWORK_LAYER_ERROR) &&
+	    (reg & UIC_NETWORK_LAYER_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.nl_err, reg);
 		hba->uic_error |= UFSHCD_UIC_NL_ERROR;
+		retval |= IRQ_HANDLED;
 	}
 
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_TRANSPORT_LAYER);
-	if (reg) {
+	if ((reg & UIC_TRANSPORT_LAYER_ERROR) &&
+	    (reg & UIC_TRANSPORT_LAYER_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.tl_err, reg);
 		hba->uic_error |= UFSHCD_UIC_TL_ERROR;
+		retval |= IRQ_HANDLED;
 	}
 
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_DME);
-	if (reg) {
+	if ((reg & UIC_DME_ERROR) &&
+	    (reg & UIC_DME_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.dme_err, reg);
 		hba->uic_error |= UFSHCD_UIC_DME_ERROR;
+		retval |= IRQ_HANDLED;
 	}
 
 	dev_dbg(hba->dev, "%s: UIC error flags = 0x%08x\n",
 			__func__, hba->uic_error);
+	return retval;
 }
 
 /**
  * ufshcd_check_errors - Check for errors that need s/w attention
  * @hba: per-adapter instance
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_check_errors(struct ufs_hba *hba)
+static irqreturn_t ufshcd_check_errors(struct ufs_hba *hba)
 {
 	bool queue_eh_work = false;
+	irqreturn_t retval = IRQ_NONE;
 
 	if (hba->errors & INT_FATAL_ERRORS)
 		queue_eh_work = true;
 
 	if (hba->errors & UIC_ERROR) {
 		hba->uic_error = 0;
-		ufshcd_update_uic_error(hba);
+		retval = ufshcd_update_uic_error(hba);
 		if (hba->uic_error)
 			queue_eh_work = true;
 	}

@@ -5327,6 +5367,7 @@ static void ufshcd_check_errors(struct ufs_hba *hba)
 			}
 			schedule_work(&hba->eh_work);
 		}
+		retval |= IRQ_HANDLED;
 	}
 	/*
 	 * if (!queue_eh_work) -

@@ -5334,40 +5375,58 @@ static void ufshcd_check_errors(struct ufs_hba *hba)
 	 * itself without s/w intervention or errors that will be
 	 * handled by the SCSI core layer.
 	 */
+	return retval;
 }
 
 /**
  * ufshcd_tmc_handler - handle task management function completion
  * @hba: per adapter instance
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_tmc_handler(struct ufs_hba *hba)
+static irqreturn_t ufshcd_tmc_handler(struct ufs_hba *hba)
 {
 	u32 tm_doorbell;
 
 	tm_doorbell = ufshcd_readl(hba, REG_UTP_TASK_REQ_DOOR_BELL);
 	hba->tm_condition = tm_doorbell ^ hba->outstanding_tasks;
-	wake_up(&hba->tm_wq);
+	if (hba->tm_condition) {
+		wake_up(&hba->tm_wq);
+		return IRQ_HANDLED;
+	} else {
+		return IRQ_NONE;
+	}
 }
 
 /**
  * ufshcd_sl_intr - Interrupt service routine
  * @hba: per adapter instance
  * @intr_status: contains interrupts generated by the controller
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_sl_intr(struct ufs_hba *hba, u32 intr_status)
+static irqreturn_t ufshcd_sl_intr(struct ufs_hba *hba, u32 intr_status)
 {
+	irqreturn_t retval = IRQ_NONE;
+
 	hba->errors = UFSHCD_ERROR_MASK & intr_status;
 	if (hba->errors)
-		ufshcd_check_errors(hba);
+		retval |= ufshcd_check_errors(hba);
 
 	if (intr_status & UFSHCD_UIC_MASK)
-		ufshcd_uic_cmd_compl(hba, intr_status);
+		retval |= ufshcd_uic_cmd_compl(hba, intr_status);
 
 	if (intr_status & UTP_TASK_REQ_COMPL)
-		ufshcd_tmc_handler(hba);
+		retval |= ufshcd_tmc_handler(hba);
 
 	if (intr_status & UTP_TRANSFER_REQ_COMPL)
-		ufshcd_transfer_req_compl(hba);
+		retval |= ufshcd_transfer_req_compl(hba);
+
+	return retval;
 }
 
 /**

@@ -5375,8 +5434,9 @@ static void ufshcd_sl_intr(struct ufs_hba *hba, u32 intr_status)
  * @irq: irq number
  * @__hba: pointer to adapter instance
  *
- * Returns IRQ_HANDLED - If interrupt is valid
- *		IRQ_NONE - If invalid interrupt
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
 static irqreturn_t ufshcd_intr(int irq, void *__hba)
 {

@@ -5399,14 +5459,19 @@ static irqreturn_t ufshcd_intr(int irq, void *__hba)
 			intr_status & ufshcd_readl(hba, REG_INTERRUPT_ENABLE);
 		if (intr_status)
 			ufshcd_writel(hba, intr_status, REG_INTERRUPT_STATUS);
-		if (enabled_intr_status) {
-			ufshcd_sl_intr(hba, enabled_intr_status);
-			retval = IRQ_HANDLED;
-		}
+		if (enabled_intr_status)
+			retval |= ufshcd_sl_intr(hba, enabled_intr_status);
 
 		intr_status = ufshcd_readl(hba, REG_INTERRUPT_STATUS);
 	} while (intr_status && --retries);
 
+	if (retval == IRQ_NONE) {
+		dev_err(hba->dev, "%s: Unhandled interrupt 0x%08x\n",
+			__func__, intr_status);
+		ufshcd_hex_dump("host regs: ", hba->mmio_base,
+				UFSHCI_REG_SPACE_SIZE);
+	}
+
 	spin_unlock(hba->host->host_lock);
 	return retval;
 }

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

Reviewed-by on v2 9/10

2018-06-28 22:26 UTC #18 lore raw

Re: [PATCH v2 09/10] scsi: ufs: fix irq return code

Quoted 321 lines · Asutosh Das

From: Venkat Gopalakrishnan <[email protected]>

Return IRQ_HANDLED only if the irq is really handled, this will help in catching spurious interrupts that go unhandled.

Signed-off-by Venkat Gopalakrishnan <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 137 
++++++++++++++++++++++++++++++++++------------
 1 file changed, 101 insertions(+), 36 deletions(-)

drivers/scsi/ufs/ufshcd.c +13 -3

index 09b7a3f..557d538 100644
@@ -211,7 +211,7 @@ enum {
 	END_FIX
 };
 
-static void ufshcd_tmc_handler(struct ufs_hba *hba);
+static irqreturn_t ufshcd_tmc_handler(struct ufs_hba *hba);
 static void ufshcd_async_scan(void *data, async_cookie_t cookie);
 static int ufshcd_reset_and_restore(struct ufs_hba *hba);
 static int ufshcd_eh_host_reset_handler(struct scsi_cmnd *cmd);

@@ -4609,19 +4609,29 @@ static int ufshcd_task_req_compl(struct
 ufs_hba *hba, u32 index, u8 *resp)
  * ufshcd_uic_cmd_compl - handle completion of uic command
  * @hba: per adapter instance
  * @intr_status: interrupt status generated by the controller
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 intr_status)
+static irqreturn_t ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 
 intr_status)
 {
+	irqreturn_t retval = IRQ_NONE;
+
 	if ((intr_status & UIC_COMMAND_COMPL) && hba->active_uic_cmd) {
 		hba->active_uic_cmd->argument2 |=
 			ufshcd_get_uic_cmd_result(hba);
 		hba->active_uic_cmd->argument3 =
 			ufshcd_get_dme_attr_val(hba);
 		complete(&hba->active_uic_cmd->done);
+		retval = IRQ_HANDLED;
 	}
 
-	if ((intr_status & UFSHCD_UIC_PWR_MASK) && hba->uic_async_done)
+	if ((intr_status & UFSHCD_UIC_PWR_MASK) && hba->uic_async_done) {
 		complete(hba->uic_async_done);
+		retval = IRQ_HANDLED;
+	}
+	return retval;
 }

/**

drivers/scsi/ufs/ufshcd.c +0 -0

@@ -4675,8 +4685,12 @@ static void __ufshcd_transfer_req_compl(struct
 ufs_hba *hba,
 /**
  * ufshcd_transfer_req_compl - handle SCSI and query command

completion
* @hba: per adapter instance
+ *
+ * Returns
+ * IRQ_HANDLED - If interrupt is valid
+ * IRQ_NONE - If invalid interrupt
*/

drivers/scsi/ufs/ufshcd.c +7 -2

-static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
+static irqreturn_t ufshcd_transfer_req_compl(struct ufs_hba *hba)
 {
 	unsigned long completed_reqs;
 	u32 tr_doorbell;

@@ -4694,7 +4708,12 @@ static void ufshcd_transfer_req_compl(struct
 ufs_hba *hba)
 	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
 	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
 
-	__ufshcd_transfer_req_compl(hba, completed_reqs);
+	if (completed_reqs) {
+		__ufshcd_transfer_req_compl(hba, completed_reqs);
+		return IRQ_HANDLED;
+	} else {
+		return IRQ_NONE;
+	}
 }

/**

drivers/scsi/ufs/ufshcd.c +7 -2

@@ -5220,16 +5239,21 @@ static void ufshcd_update_uic_reg_hist(struct
 ufs_uic_err_reg_hist *reg_hist,
 /**
  * ufshcd_update_uic_error - check and set fatal UIC error flags.
  * @hba: per-adapter instance
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_update_uic_error(struct ufs_hba *hba)
+static irqreturn_t ufshcd_update_uic_error(struct ufs_hba *hba)
 {
 	u32 reg;
+	irqreturn_t retval = IRQ_NONE;
 
 	/* PHY layer lane error */
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_PHY_ADAPTER_LAYER);
 	/* Ignore LINERESET indication, as this is not an error */
 	if ((reg & UIC_PHY_ADAPTER_LAYER_ERROR) &&
-			(reg & UIC_PHY_ADAPTER_LAYER_LANE_ERR_MASK)) {
+			(reg & UIC_PHY_ADAPTER_LAYER_ERROR_CODE_MASK)) {
 		/*
 		 * To know whether this error is fatal or not, DB timeout

drivers/scsi/ufs/ufshcd.c +32 ~~-16~~

 		 * must be checked but this error is handled separately.

@@ -5240,57 +5264,73 @@ static void ufshcd_update_uic_error(struct 
 ufs_hba *hba)
 
 	/* PA_INIT_ERROR is fatal and needs UIC reset */
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_DATA_LINK_LAYER);
-	if (reg)
+	if ((reg & UIC_DATA_LINK_LAYER_ERROR) &&
+	    (reg & UIC_DATA_LINK_LAYER_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.dl_err, reg);
-
-	if (reg & UIC_DATA_LINK_LAYER_ERROR_PA_INIT)
-		hba->uic_error |= UFSHCD_UIC_DL_PA_INIT_ERROR;
-	else if (hba->dev_quirks &
-		   UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS) {
-		if (reg & UIC_DATA_LINK_LAYER_ERROR_NAC_RECEIVED)
-			hba->uic_error |=
-				UFSHCD_UIC_DL_NAC_RECEIVED_ERROR;
-		else if (reg & UIC_DATA_LINK_LAYER_ERROR_TCx_REPLAY_TIMEOUT)
-			hba->uic_error |= UFSHCD_UIC_DL_TCx_REPLAY_ERROR;
+		if (reg & UIC_DATA_LINK_LAYER_ERROR_PA_INIT) {
+			hba->uic_error |= UFSHCD_UIC_DL_PA_INIT_ERROR;
+		} else if (hba->dev_quirks &
+			   UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS) {
+			if (reg & UIC_DATA_LINK_LAYER_ERROR_NAC_RECEIVED)
+				hba->uic_error |=
+					UFSHCD_UIC_DL_NAC_RECEIVED_ERROR;
+			else if (reg &
+				UIC_DATA_LINK_LAYER_ERROR_TCx_REPLAY_TIMEOUT)
+				hba->uic_error |=
+					UFSHCD_UIC_DL_TCx_REPLAY_ERROR;
+		}
+		retval |= IRQ_HANDLED;
 	}
 
 	/* UIC NL/TL/DME errors needs software retry */
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_NETWORK_LAYER);
-	if (reg) {
+	if ((reg & UIC_NETWORK_LAYER_ERROR) &&
+	    (reg & UIC_NETWORK_LAYER_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.nl_err, reg);
 		hba->uic_error |= UFSHCD_UIC_NL_ERROR;
+		retval |= IRQ_HANDLED;
 	}
 
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_TRANSPORT_LAYER);
-	if (reg) {
+	if ((reg & UIC_TRANSPORT_LAYER_ERROR) &&
+	    (reg & UIC_TRANSPORT_LAYER_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.tl_err, reg);
 		hba->uic_error |= UFSHCD_UIC_TL_ERROR;
+		retval |= IRQ_HANDLED;
 	}
 
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_DME);
-	if (reg) {
+	if ((reg & UIC_DME_ERROR) &&
+	    (reg & UIC_DME_ERROR_CODE_MASK)) {
 		ufshcd_update_uic_reg_hist(&hba->ufs_stats.dme_err, reg);
 		hba->uic_error |= UFSHCD_UIC_DME_ERROR;
+		retval |= IRQ_HANDLED;
 	}
 
 	dev_dbg(hba->dev, "%s: UIC error flags = 0x%08x\n",
 			__func__, hba->uic_error);
+	return retval;
 }
 
 /**
  * ufshcd_check_errors - Check for errors that need s/w attention
  * @hba: per-adapter instance
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_check_errors(struct ufs_hba *hba)
+static irqreturn_t ufshcd_check_errors(struct ufs_hba *hba)
 {
 	bool queue_eh_work = false;
+	irqreturn_t retval = IRQ_NONE;
 
 	if (hba->errors & INT_FATAL_ERRORS)
 		queue_eh_work = true;
 
 	if (hba->errors & UIC_ERROR) {
 		hba->uic_error = 0;
-		ufshcd_update_uic_error(hba);
+		retval = ufshcd_update_uic_error(hba);
 		if (hba->uic_error)
 			queue_eh_work = true;

drivers/scsi/ufs/ufshcd.c +1 -0

@@ -5327,6 +5367,7 @@ static void ufshcd_check_errors(struct ufs_hba 
 *hba)
 			}
 			schedule_work(&hba->eh_work);
 		}
+		retval |= IRQ_HANDLED;
 	}
 	/*

drivers/scsi/ufs/ufshcd.c +25 -7

 	 * if (!queue_eh_work) -

@@ -5334,40 +5375,58 @@ static void ufshcd_check_errors(struct ufs_hba 
 *hba)
 	 * itself without s/w intervention or errors that will be
 	 * handled by the SCSI core layer.
 	 */
+	return retval;
 }
 
 /**
  * ufshcd_tmc_handler - handle task management function completion
  * @hba: per adapter instance
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_tmc_handler(struct ufs_hba *hba)
+static irqreturn_t ufshcd_tmc_handler(struct ufs_hba *hba)
 {
 	u32 tm_doorbell;
 
 	tm_doorbell = ufshcd_readl(hba, REG_UTP_TASK_REQ_DOOR_BELL);
 	hba->tm_condition = tm_doorbell ^ hba->outstanding_tasks;
-	wake_up(&hba->tm_wq);
+	if (hba->tm_condition) {
+		wake_up(&hba->tm_wq);
+		return IRQ_HANDLED;
+	} else {
+		return IRQ_NONE;
+	}
 }
 
 /**
  * ufshcd_sl_intr - Interrupt service routine
  * @hba: per adapter instance
  * @intr_status: contains interrupts generated by the controller
+ *
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
-static void ufshcd_sl_intr(struct ufs_hba *hba, u32 intr_status)
+static irqreturn_t ufshcd_sl_intr(struct ufs_hba *hba, u32 
 intr_status)
 {
+	irqreturn_t retval = IRQ_NONE;
+
 	hba->errors = UFSHCD_ERROR_MASK & intr_status;
 	if (hba->errors)
-		ufshcd_check_errors(hba);
+		retval |= ufshcd_check_errors(hba);
 
 	if (intr_status & UFSHCD_UIC_MASK)
-		ufshcd_uic_cmd_compl(hba, intr_status);
+		retval |= ufshcd_uic_cmd_compl(hba, intr_status);
 
 	if (intr_status & UTP_TASK_REQ_COMPL)
-		ufshcd_tmc_handler(hba);
+		retval |= ufshcd_tmc_handler(hba);
 
 	if (intr_status & UTP_TRANSFER_REQ_COMPL)
-		ufshcd_transfer_req_compl(hba);
+		retval |= ufshcd_transfer_req_compl(hba);
+
+	return retval;
 }

/**

drivers/scsi/ufs/ufshcd.c +3 -2

@@ -5375,8 +5434,9 @@ static void ufshcd_sl_intr(struct ufs_hba *hba,
 u32 intr_status)
  * @irq: irq number
  * @__hba: pointer to adapter instance
  *
- * Returns IRQ_HANDLED - If interrupt is valid
- *		IRQ_NONE - If invalid interrupt
+ * Returns
+ *  IRQ_HANDLED - If interrupt is valid
+ *  IRQ_NONE    - If invalid interrupt
  */
 static irqreturn_t ufshcd_intr(int irq, void *__hba)

{

drivers/scsi/ufs/ufshcd.c +9 -4

@@ -5399,14 +5459,19 @@ static irqreturn_t ufshcd_intr(int irq, void 
 *__hba)
 			intr_status & ufshcd_readl(hba, REG_INTERRUPT_ENABLE);
 		if (intr_status)
 			ufshcd_writel(hba, intr_status, REG_INTERRUPT_STATUS);
-		if (enabled_intr_status) {
-			ufshcd_sl_intr(hba, enabled_intr_status);
-			retval = IRQ_HANDLED;
-		}
+		if (enabled_intr_status)
+			retval |= ufshcd_sl_intr(hba, enabled_intr_status);
 
 		intr_status = ufshcd_readl(hba, REG_INTERRUPT_STATUS);
 	} while (intr_status && --retries);
 
+	if (retval == IRQ_NONE) {
+		dev_err(hba->dev, "%s: Unhandled interrupt 0x%08x\n",
+			__func__, intr_status);
+		ufshcd_hex_dump("host regs: ", hba->mmio_base,
+				UFSHCI_REG_SPACE_SIZE);
+	}
+
 	spin_unlock(hba->host->host_lock);
 	return retval;

Quoted 1 line

}

Looks good to me.

Signature · 2 lines

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

Asutosh Das

author

2018-05-03 11:07 UTC #19 lore raw

PATCH v2 10/10 scsi: ufs: Add clock ungating to a separate workqueue

From: Vijay Viswanath <[email protected]>

UFS driver can receive a request during memory reclaim by kswapd. So when ufs driver puts the ungate work in queue, and if there are no idle workers, kthreadd is invoked to create a new kworker. Since kswapd task holds a mutex which kthreadd also needs, this can cause a deadlock situation. So ungate work must be done in a separate work queue with WQ_MEM_RECLAIM flag enabled.
Such a workqueue will have a rescue thread which will be called when the above deadlock condition is possible.

Signed-off-by Vijay Viswanath <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 11 ++++++++++-
 drivers/scsi/ufs/ufshcd.h |  1 +
 2 files changed, 11 insertions(+), 1 deletion(-)

drivers/scsi/ufs/ufshcd.c +10 -1

index 557d538..3be61b7 100644
@@ -1483,7 +1483,8 @@ int ufshcd_hold(struct ufs_hba *hba, bool async)
 		hba->clk_gating.state = REQ_CLKS_ON;
 		trace_ufshcd_clk_gating(dev_name(hba->dev),
 					hba->clk_gating.state);
-		schedule_work(&hba->clk_gating.ungate_work);
+		queue_work(hba->clk_gating.clk_gating_workq,
+			   &hba->clk_gating.ungate_work);
 		/*
 		 * fall through to check if we should wait for this
 		 * work to be done or not.

@@ -1669,6 +1670,8 @@ static ssize_t ufshcd_clkgate_enable_store(struct device *dev,
 
 static void ufshcd_init_clk_gating(struct ufs_hba *hba)
 {
+	char wq_name[sizeof("ufs_clk_gating_00")];
+
 	if (!ufshcd_is_clkgating_allowed(hba))
 		return;

@@ -1676,6 +1679,11 @@ static void ufshcd_init_clk_gating(struct ufs_hba *hba)
 	INIT_DELAYED_WORK(&hba->clk_gating.gate_work, ufshcd_gate_work);
 	INIT_WORK(&hba->clk_gating.ungate_work, ufshcd_ungate_work);
 
+	snprintf(wq_name, ARRAY_SIZE(wq_name), "ufs_clk_gating_%d",
+		 hba->host->host_no);
+	hba->clk_gating.clk_gating_workq = alloc_ordered_workqueue(wq_name,
+							   WQ_MEM_RECLAIM);
+
 	hba->clk_gating.is_enabled = true;
 
 	hba->clk_gating.delay_attr.show = ufshcd_clkgate_delay_show;

@@ -1703,6 +1711,7 @@ static void ufshcd_exit_clk_gating(struct ufs_hba *hba)
 	device_remove_file(hba->dev, &hba->clk_gating.enable_attr);
 	cancel_work_sync(&hba->clk_gating.ungate_work);
 	cancel_delayed_work_sync(&hba->clk_gating.gate_work);
+	destroy_workqueue(hba->clk_gating.clk_gating_workq);
 }
 
 /* Must be called with host lock acquired */

drivers/scsi/ufs/ufshcd.h +1 -0

index 76c31d5..2e6bdc0 100644
@@ -361,6 +361,7 @@ struct ufs_clk_gating {
 	struct device_attribute enable_attr;
 	bool is_enabled;
 	int active_reqs;
+	struct workqueue_struct *clk_gating_workq;
 };
 
 struct ufs_saved_pwr_info {

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc. 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Subhash Jadavani

Reviewed-by on v2 10/10

2018-05-16 21:14 UTC #20 lore raw

Re: [PATCH v2 10/10] scsi: ufs: Add clock ungating to a separate workqueue

Quoted 70 lines · Asutosh Das

From: Vijay Viswanath <[email protected]>

Signed-off-by Vijay Viswanath <[email protected]>

Signed-off-by Can Guo <[email protected]>

Signed-off-by Asutosh Das <[email protected]>

---
 drivers/scsi/ufs/ufshcd.c | 11 ++++++++++-
 drivers/scsi/ufs/ufshcd.h |  1 +
 2 files changed, 11 insertions(+), 1 deletion(-)

drivers/scsi/ufs/ufshcd.c +9 -1

index 557d538..3be61b7 100644
@@ -1483,7 +1483,8 @@ int ufshcd_hold(struct ufs_hba *hba, bool async)
 		hba->clk_gating.state = REQ_CLKS_ON;
 		trace_ufshcd_clk_gating(dev_name(hba->dev),
 					hba->clk_gating.state);
-		schedule_work(&hba->clk_gating.ungate_work);
+		queue_work(hba->clk_gating.clk_gating_workq,
+			   &hba->clk_gating.ungate_work);
 		/*
 		 * fall through to check if we should wait for this
 		 * work to be done or not.

@@ -1669,6 +1670,8 @@ static ssize_t
 ufshcd_clkgate_enable_store(struct device *dev,
 
 static void ufshcd_init_clk_gating(struct ufs_hba *hba)
 {
+	char wq_name[sizeof("ufs_clk_gating_00")];
+
 	if (!ufshcd_is_clkgating_allowed(hba))
 		return;

@@ -1676,6 +1679,11 @@ static void ufshcd_init_clk_gating(struct 
 ufs_hba *hba)
 	INIT_DELAYED_WORK(&hba->clk_gating.gate_work, ufshcd_gate_work);
 	INIT_WORK(&hba->clk_gating.ungate_work, ufshcd_ungate_work);
 
+	snprintf(wq_name, ARRAY_SIZE(wq_name), "ufs_clk_gating_%d",
+		 hba->host->host_no);
+	hba->clk_gating.clk_gating_workq = alloc_ordered_workqueue(wq_name,
+							   WQ_MEM_RECLAIM);
+
 	hba->clk_gating.is_enabled = true;

drivers/scsi/ufs/ufshcd.c +1 -0

 	hba->clk_gating.delay_attr.show = ufshcd_clkgate_delay_show;

@@ -1703,6 +1711,7 @@ static void ufshcd_exit_clk_gating(struct ufs_hba 
 *hba)
 	device_remove_file(hba->dev, &hba->clk_gating.enable_attr);
 	cancel_work_sync(&hba->clk_gating.ungate_work);
 	cancel_delayed_work_sync(&hba->clk_gating.gate_work);
+	destroy_workqueue(hba->clk_gating.clk_gating_workq);
 }

/* Must be called with host lock acquired */

drivers/scsi/ufs/ufshcd.h +0 -0

index 76c31d5..2e6bdc0 100644

drivers/scsi/ufs/ufshcd.h struct ufs_clk_gating { · line 367

Quoted 7 lines

drivers/scsi/ufs/ufshcd.h +1 -0

@@ -361,6 +361,7 @@ struct ufs_clk_gating {
 	struct device_attribute enable_attr;
 	bool is_enabled;
 	int active_reqs;
+	struct workqueue_struct *clk_gating_workq;
 };
 
 struct ufs_saved_pwr_info {

Looks good to me.