[PATCH v2] ufs: core: wlun suspend SSU/enter hibern8 fail recovery

[PATCH v2] ufs: core: wlun suspend SSU/enter hibern8 fail recovery

[peter.wang]

From: Peter Wang <peter.wang@mediatek.com>

When SSU/enter hibern8 fail in wlun suspend flow, trigger error
handler and return busy to break the suspend.
If not, wlun runtime pm status become error and the consumer will
stuck in runtime suspend status.

Signed-off-by: Peter Wang <peter.wang@mediatek.com>
---
 drivers/ufs/core/ufshcd.c | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index b1f59a5fe632..8e5f7d9fe7c2 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -8970,6 +8970,7 @@ static int __ufshcd_wl_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
 	enum ufs_pm_level pm_lvl;
 	enum ufs_dev_pwr_mode req_dev_pwr_mode;
 	enum uic_link_state req_link_state;
+	unsigned long flags;
 
 	hba->pm_op_in_progress = true;
 	if (pm_op != UFS_SHUTDOWN_PM) {
@@ -9049,6 +9050,20 @@ static int __ufshcd_wl_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
 
 		if (!hba->dev_info.b_rpm_dev_flush_capable) {
 			ret = ufshcd_set_dev_pwr_mode(hba, req_dev_pwr_mode);
+			if ((ret) && (pm_op != UFS_SHUTDOWN_PM)) {
+				/*
+				 * If return err in suspend flow, IO will hang.
+				 * Trigger error handler and break suspend for
+				 * error recovery.
+				 */
+				spin_lock_irqsave(hba->host->host_lock, flags);
+				hba->force_reset = true;
+				ufshcd_schedule_eh_work(hba);
+				spin_unlock_irqrestore(hba->host->host_lock,
+					flags);
+
+				ret = -EBUSY;
+			}
 			if (ret)
 				goto enable_scaling;
 		}
@@ -9060,6 +9075,20 @@ static int __ufshcd_wl_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
 	 */
 	check_for_bkops = !ufshcd_is_ufs_dev_deepsleep(hba);
 	ret = ufshcd_link_state_transition(hba, req_link_state, check_for_bkops);
+	if ((ret) && (pm_op != UFS_SHUTDOWN_PM)) {
+		/*
+		 * If return err in suspend flow, IO will hang.
+		 * Trigger error handler and break suspend for
+		 * error recovery.
+		 */
+		spin_lock_irqsave(hba->host->host_lock, flags);
+		hba->force_reset = true;
+		ufshcd_schedule_eh_work(hba);
+		spin_unlock_irqrestore(hba->host->host_lock,
+			flags);
+
+		ret = -EBUSY;
+	}
 	if (ret)
 		goto set_dev_active;
 
-- 
2.18.0

Re: [PATCH v2] ufs: core: wlun suspend SSU/enter hibern8 fail recovery

[Bart Van Assche]

On 12/1/22 01:43, peter.wang@mediatek.com wrote:
> @@ -9049,6 +9050,20 @@ static int __ufshcd_wl_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
>   
>   		if (!hba->dev_info.b_rpm_dev_flush_capable) {
>   			ret = ufshcd_set_dev_pwr_mode(hba, req_dev_pwr_mode);
> +			if ((ret) && (pm_op != UFS_SHUTDOWN_PM)) {

Please remove the superfluous parentheses from the above statement.

> +				/*
> +				 * If return err in suspend flow, IO will hang.
> +				 * Trigger error handler and break suspend for
> +				 * error recovery.
> +				 */
> +				spin_lock_irqsave(hba->host->host_lock, flags);

__ufshcd_wl_suspend() is allowed to sleep. Please change 
spin_lock_irqsave() into spin_lock_irq().

> +				hba->force_reset = true;
> +				ufshcd_schedule_eh_work(hba);
> +				spin_unlock_irqrestore(hba->host->host_lock,
> +					flags);
> +
> +				ret = -EBUSY;

Why is the value of 'ret' changed into -EBUSY? Can the above code be 
left out?

> @@ -9060,6 +9075,20 @@ static int __ufshcd_wl_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
>   	 */
>   	check_for_bkops = !ufshcd_is_ufs_dev_deepsleep(hba);
>   	ret = ufshcd_link_state_transition(hba, req_link_state, check_for_bkops);
> +	if ((ret) && (pm_op != UFS_SHUTDOWN_PM)) {
> +		/*
> +		 * If return err in suspend flow, IO will hang.
> +		 * Trigger error handler and break suspend for
> +		 * error recovery.
> +		 */
> +		spin_lock_irqsave(hba->host->host_lock, flags);
> +		hba->force_reset = true;
> +		ufshcd_schedule_eh_work(hba);
> +		spin_unlock_irqrestore(hba->host->host_lock,
> +			flags);
> +
> +		ret = -EBUSY;
> +	}

Same comments as above for this code block.

Thanks,

Bart.

Re: [PATCH v2] ufs: core: wlun suspend SSU/enter hibern8 fail recovery

[Peter Wang (王信友)]

On Thu, 2022-12-01 at 10:44 -0800, Bart Van Assche wrote:
> On 12/1/22 01:43, peter.wang@mediatek.com wrote:
> > @@ -9049,6 +9050,20 @@ static int __ufshcd_wl_suspend(struct
> > ufs_hba *hba, enum ufs_pm_op pm_op)
> >   
> >   		if (!hba->dev_info.b_rpm_dev_flush_capable) {
> >   			ret = ufshcd_set_dev_pwr_mode(hba,
> > req_dev_pwr_mode);
> > +			if ((ret) && (pm_op != UFS_SHUTDOWN_PM)) {
> 
> Please remove the superfluous parentheses from the above statement.
> 
Hi Bart,

Okay, will change next version.

> > +				/*
> > +				 * If return err in suspend flow, IO
> > will hang.
> > +				 * Trigger error handler and break
> > suspend for
> > +				 * error recovery.
> > +				 */
> > +				spin_lock_irqsave(hba->host->host_lock, 
> > flags);
> 
> __ufshcd_wl_suspend() is allowed to sleep. Please change 
> spin_lock_irqsave() into spin_lock_irq().
> 
> > +				hba->force_reset = true;
> > +				ufshcd_schedule_eh_work(hba);
> > +				spin_unlock_irqrestore(hba->host-
> > >host_lock,
> > +					flags);
> > +
> > +				ret = -EBUSY;
> 
> Why is the value of 'ret' changed into -EBUSY? Can the above code be 
> left out?
> 

Because in runtime pm framework, return EBUSY is okay for next time
runtime suspend retry.
Return ETIMEOUT/EIO or other then EBUSY will keep error code in
veriable "runtime_error". (supplier first then consumer resume block by
supplier also get runtime_error too). Finally, lead to IO hang.


> > @@ -9060,6 +9075,20 @@ static int __ufshcd_wl_suspend(struct
> > ufs_hba *hba, enum ufs_pm_op pm_op)
> >   	 */
> >   	check_for_bkops = !ufshcd_is_ufs_dev_deepsleep(hba);
> >   	ret = ufshcd_link_state_transition(hba, req_link_state,
> > check_for_bkops);
> > +	if ((ret) && (pm_op != UFS_SHUTDOWN_PM)) {
> > +		/*
> > +		 * If return err in suspend flow, IO will hang.
> > +		 * Trigger error handler and break suspend for
> > +		 * error recovery.
> > +		 */
> > +		spin_lock_irqsave(hba->host->host_lock, flags);
> > +		hba->force_reset = true;
> > +		ufshcd_schedule_eh_work(hba);
> > +		spin_unlock_irqrestore(hba->host->host_lock,
> > +			flags);
> > +
> > +		ret = -EBUSY;
> > +	}
> 
> Same comments as above for this code block.
> 
> Thanks,
> 
> Bart.

Okay, will change next version.

Thanks
BR
Peter


>