Skip to contentSkip to conversation

scsi: sd: fix sd shutdown to issue START STOP UNIT command appropriately

PATCH Superseded Salomon Dushimirimana 4 participants 9 messages Jul 24, 2025 – Jul 28, 2025 SCSI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 scsi: sd: fix sd shutdown to issue START STOP UNIT command appropriately

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

---
 drivers/scsi/sd.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +3 -1
index eeaa6af294b81..282000c761f8e 100644
@@ -4173,7 +4173,9 @@ static void sd_shutdown(struct device *dev)
 	if ((system_state != SYSTEM_RESTART &&
 	     sdkp->device->manage_system_start_stop) ||
 	    (system_state == SYSTEM_POWER_OFF &&
-	     sdkp->device->manage_shutdown)) {
+	     sdkp->device->manage_shutdown) ||
+	    (system_state == SYSTEM_RUNNING &&
+	     sdkp->device->manage_runtime_start_stop)) {
 		sd_printk(KERN_NOTICE, sdkp, "Stopping disk\n");
 		sd_start_stop_device(sdkp, 0);
 	}
Signature · 1 line
2.50.1.470.g6ba607880d-goog
#1 lore raw
Quoted 25 lines · Salomon Dushimirimana

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

---
  drivers/scsi/sd.c | 4 +++-
  1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +0 -0
index eeaa6af294b81..282000c761f8e 100644

drivers/scsi/sd.c static void sd_shutdown(struct device *dev) · line 4181

Quoted 10 lines
drivers/scsi/sd.c +3 -1
@@ -4173,7 +4173,9 @@ static void sd_shutdown(struct device *dev)
  	if ((system_state != SYSTEM_RESTART &&
  	     sdkp->device->manage_system_start_stop) ||
  	    (system_state == SYSTEM_POWER_OFF &&
-	     sdkp->device->manage_shutdown)) {
+	     sdkp->device->manage_shutdown) ||
+	    (system_state == SYSTEM_RUNNING &&
+	     sdkp->device->manage_runtime_start_stop)) {
  		sd_printk(KERN_NOTICE, sdkp, "Stopping disk\n");
  		sd_start_stop_device(sdkp, 0);
  	}

A Fixes: tag is missing.

manage_runtime_start_stop is for runtime power management. /sys/block/*/device/delete is not related to runtime power management. Isn't manage_system_start_stop more appropriate here than manage_runtime_start_stop? Shouldn't sd_shutdown() calls triggered by writing into /sys/block/*/device/delete already be covered by this
test: system_state != SYSTEM_RESTART &&
sdkp->device->manage_system_start_stop?

Thanks,

Bart.

#2 lore raw

Hi Bart,

Thanks for the timely feedback! I also thought that the manage_system_start_stop flag is more appropriate.

But I wasn't sure why the commit mentioned in the description removed the manage_system_start_stop from ata_scsi_dev_config() function in the first place. So other candidates for a fix were:
1. Adding the setting of the manage_system_start_stop flag back to ata_scsi_dev_config()
2. Configuring our various drivers to individually enable the manage_system_start_stop flag (less preferable)

If there was no good reason to remove setting the flag, we can go with option 1.

Thanks,

Quoted 51 lines · Bart Van Assche
Quoted 25 lines · Salomon Dushimirimana

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

---
  drivers/scsi/sd.c | 4 +++-
  1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +0 -0
index eeaa6af294b81..282000c761f8e 100644

drivers/scsi/sd.c static void sd_shutdown(struct device *dev) · line 4181

Quoted 10 lines
drivers/scsi/sd.c +3 -1
@@ -4173,7 +4173,9 @@ static void sd_shutdown(struct device *dev)
      if ((system_state != SYSTEM_RESTART &&
           sdkp->device->manage_system_start_stop) ||
          (system_state == SYSTEM_POWER_OFF &&
-          sdkp->device->manage_shutdown)) {
+          sdkp->device->manage_shutdown) ||
+         (system_state == SYSTEM_RUNNING &&
+          sdkp->device->manage_runtime_start_stop)) {
              sd_printk(KERN_NOTICE, sdkp, "Stopping disk\n");
              sd_start_stop_device(sdkp, 0);
      }

A Fixes: tag is missing.

manage_runtime_start_stop is for runtime power management. /sys/block/*/device/delete is not related to runtime power management. Isn't manage_system_start_stop more appropriate here than manage_runtime_start_stop? Shouldn't sd_shutdown() calls triggered by writing into /sys/block/*/device/delete already be covered by this
test: system_state != SYSTEM_RESTART &&
sdkp->device->manage_system_start_stop?

Thanks,

Bart.

#3 lore raw

PATCH v2 1/1 scsi: sd: fix sd shutdown to issue START STOP UNIT command appropriately

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

Fixes aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop")
Change-Id I820269189d1a2ee03795b8c0db41aa50c0cb484d
---
 drivers/scsi/sd.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +3 -1
index eeaa6af294b81..282000c761f8e 100644
@@ -4173,7 +4173,9 @@ static void sd_shutdown(struct device *dev)
 	if ((system_state != SYSTEM_RESTART &&
 	     sdkp->device->manage_system_start_stop) ||
 	    (system_state == SYSTEM_POWER_OFF &&
-	     sdkp->device->manage_shutdown)) {
+	     sdkp->device->manage_shutdown) ||
+	    (system_state == SYSTEM_RUNNING &&
+	     sdkp->device->manage_runtime_start_stop)) {
 		sd_printk(KERN_NOTICE, sdkp, "Stopping disk\n");
 		sd_start_stop_device(sdkp, 0);
 	}
Signature · 1 line
2.50.1.470.g6ba607880d-goog
#4 lore raw

PATCH v3 1/1 scsi: sd: fix sd shutdown to issue START STOP UNIT command appropriately

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

Fixes aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop")

---
Changes in v3:
- Removed unnecessary tag

 drivers/scsi/sd.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +3 -1
index eeaa6af294b81..282000c761f8e 100644
@@ -4173,7 +4173,9 @@ static void sd_shutdown(struct device *dev)
 	if ((system_state != SYSTEM_RESTART &&
 	     sdkp->device->manage_system_start_stop) ||
 	    (system_state == SYSTEM_POWER_OFF &&
-	     sdkp->device->manage_shutdown)) {
+	     sdkp->device->manage_shutdown) ||
+	    (system_state == SYSTEM_RUNNING &&
+	     sdkp->device->manage_runtime_start_stop)) {
 		sd_printk(KERN_NOTICE, sdkp, "Stopping disk\n");
 		sd_start_stop_device(sdkp, 0);
 	}
Signature · 1 line
2.50.1.470.g6ba607880d-goog
Tested-by Reviewed-by on v3 1/1
#5 lore raw
Quoted 15 lines · Salomon Dushimirimana

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

Quoted 2 lines
Fixes aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop")
Signature · 2 lines
Damien Le Moal
Western Digital Research
#6 lore raw
In reply to Salomon Dushimirimana

Salomon,

Quoted 7 lines

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

Applied to 6.17/scsi-staging, thanks!

Signature · 1 line
Martin K. Petersen
#7 lore raw
In reply to Salomon Dushimirimana
Quoted 29 lines · Salomon Dushimirimana

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

Fixes aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop")

---
Changes in v3:
- Removed unnecessary tag

  drivers/scsi/sd.c | 4 +++-
  1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +0 -0
index eeaa6af294b81..282000c761f8e 100644

drivers/scsi/sd.c static void sd_shutdown(struct device *dev) · line 4181

Quoted 10 lines
drivers/scsi/sd.c +3 -1
@@ -4173,7 +4173,9 @@ static void sd_shutdown(struct device *dev)
  	if ((system_state != SYSTEM_RESTART &&
  	     sdkp->device->manage_system_start_stop) ||
  	    (system_state == SYSTEM_POWER_OFF &&
-	     sdkp->device->manage_shutdown)) {
+	     sdkp->device->manage_shutdown) ||
+	    (system_state == SYSTEM_RUNNING &&
+	     sdkp->device->manage_runtime_start_stop)) {
  		sd_printk(KERN_NOTICE, sdkp, "Stopping disk\n");
  		sd_start_stop_device(sdkp, 0);
  	}

Runtime power management is not related at all to deleting a LUN through sysfs. This patch makes it impossible to understand the sd_shutdown() implementation without studying the ATA subsystem and its history. Additionally, this patch makes the documentation of .manage_runtime_start_stop incorrect.

There are only two drivers that set .manage_runtime_start_stop: libata and the unmaintained sbp2 driver. Has it been considered to test sdkp->device->is_ata instead of sdkp->device->manage_runtime_start_stop?

Thanks,

Bart.

#8 lore raw
Quoted 47 lines · Bart Van Assche
Quoted 29 lines · Salomon Dushimirimana

Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") enabled libata EH to manage device power mode trasitions for system suspend/resume and removed the flag from ata_scsi_dev_config. However, since the sd_shutdown() function still relies on the manage_system_start_stop flag, a spin-down command is not issued to the disk with command "echo 1 > /sys/block/sdb/device/delete"

sd_shutdown() can be called for both system/runtime start stop operations, so utilize the manage_run_time_start_stop flag set in the ata_scsi_dev_config and issue a spin-down command during disk removal when the system is running. This is in addition to when the system is powering off and manage_shutdown flag is set. The manage_system_start_stop flag will still be used for drivers that still set the flag.

Fixes aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop")

---
Changes in v3:
- Removed unnecessary tag

  drivers/scsi/sd.c | 4 +++-
  1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +0 -0
index eeaa6af294b81..282000c761f8e 100644

drivers/scsi/sd.c static void sd_shutdown(struct device *dev) · line 4181

Quoted 10 lines
drivers/scsi/sd.c +3 -1
@@ -4173,7 +4173,9 @@ static void sd_shutdown(struct device *dev)
  	if ((system_state != SYSTEM_RESTART &&
  	     sdkp->device->manage_system_start_stop) ||
  	    (system_state == SYSTEM_POWER_OFF &&
-	     sdkp->device->manage_shutdown)) {
+	     sdkp->device->manage_shutdown) ||
+	    (system_state == SYSTEM_RUNNING &&
+	     sdkp->device->manage_runtime_start_stop)) {
  		sd_printk(KERN_NOTICE, sdkp, "Stopping disk\n");
  		sd_start_stop_device(sdkp, 0);
  	}

Runtime power management is not related at all to deleting a LUN through sysfs. This patch makes it impossible to understand the sd_shutdown() implementation without studying the ATA subsystem and its history. Additionally, this patch makes the documentation of .manage_runtime_start_stop incorrect.

Quoted 3 lines

There are only two drivers that set .manage_runtime_start_stop: libata and the unmaintained sbp2 driver. Has it been considered to test sdkp->device->is_ata instead of sdkp->device->manage_runtime_start_stop?

I would rather prefer to not spread the use of device->is_ata further than it already is. We are not even supposed to be needing that one in the first place, but SAT is not a perfect specification.

But nevertheless, feel free to send cleanup patches to make this code prettier :)

Signature · 2 lines
Damien Le Moal
Western Digital Research