Skip to contentSkip to conversation

Make libata not powerdown drivers on PM_EVENT_FREEZE.

PATCH In review Nigel Cunningham 6 participants 10 messages Mar 23, 2006 – Mar 26, 2006

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Make libata not powerdown drivers on PM_EVENT_FREEZE.

Hi.

At the moment libata doesn't pass pm_message_t down ata_device_suspend. This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Prepared against git at the time of writing. Please apply.

 drivers/scsi/libata-core.c |    5 +++--
 drivers/scsi/libata-scsi.c |    2 +-
 drivers/scsi/scsi_sysfs.c  |    2 +-
 include/linux/libata.h     |    4 ++--
 include/scsi/scsi_host.h   |    2 +-
 5 files changed, 8 insertions(+), 7 deletions(-)
drivers/scsi/libata-core.c +3 -2
diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/libata-core.c 9904-libata-freeze.patch-new/drivers/scsi/libata-core.c
@@ -4506,14 +4506,15 @@ int ata_device_resume(struct ata_port *a
  *	Flush the cache on the drive, if appropriate, then issue a
  *	standbynow command.
  */
-int ata_device_suspend(struct ata_port *ap, struct ata_device *dev)
+int ata_device_suspend(struct ata_port *ap, struct ata_device *dev, pm_message_t state)
 {
 	if (!ata_dev_present(dev))
 		return 0;
 	if (dev->class == ATA_DEV_ATA)
 		ata_flush_cache(ap, dev);
 
-	ata_standby_drive(ap, dev);
+	if (state.event != PM_EVENT_FREEZE)
+		ata_standby_drive(ap, dev);
 	ap->flags |= ATA_FLAG_SUSPENDED;
 	return 0;
 }
drivers/scsi/libata-scsi.c +1 -1
diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/libata-scsi.c 9904-libata-freeze.patch-new/drivers/scsi/libata-scsi.c
@@ -419,7 +419,7 @@ int ata_scsi_device_suspend(struct scsi_
 	struct ata_port *ap = (struct ata_port *) &sdev->host->hostdata[0];
 	struct ata_device *dev = &ap->device[sdev->id];
 
-	return ata_device_suspend(ap, dev);
+	return ata_device_suspend(ap, dev, state);
 }
 
 /**
drivers/scsi/scsi_sysfs.c +1 -1
diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/scsi_sysfs.c 9904-libata-freeze.patch-new/drivers/scsi/scsi_sysfs.c
@@ -286,7 +286,7 @@ static int scsi_bus_suspend(struct devic
 		return err;
 
 	if (sht->suspend)
-		err = sht->suspend(sdev);
+		err = sht->suspend(sdev, state);
 
 	return err;
 }
include/linux/libata.h +2 -2
diff -ruNp 9904-libata-freeze.patch-old/include/linux/libata.h 9904-libata-freeze.patch-new/include/linux/libata.h
@@ -515,9 +515,9 @@ extern void ata_eh_qc_retry(struct ata_q
 extern int ata_scsi_release(struct Scsi_Host *host);
 extern unsigned int ata_host_intr(struct ata_port *ap, struct ata_queued_cmd *qc);
 extern int ata_scsi_device_resume(struct scsi_device *);
-extern int ata_scsi_device_suspend(struct scsi_device *);
+extern int ata_scsi_device_suspend(struct scsi_device *, pm_message_t state);
 extern int ata_device_resume(struct ata_port *, struct ata_device *);
-extern int ata_device_suspend(struct ata_port *, struct ata_device *);
+extern int ata_device_suspend(struct ata_port *, struct ata_device *, pm_message_t state);
 extern int ata_ratelimit(void);
 extern unsigned int ata_busy_sleep(struct ata_port *ap,
 				   unsigned long timeout_pat,
include/scsi/scsi_host.h +1 -1
diff -ruNp 9904-libata-freeze.patch-old/include/scsi/scsi_host.h 9904-libata-freeze.patch-new/include/scsi/scsi_host.h
@@ -286,7 +286,7 @@ struct scsi_host_template {
 	 * suspend support
 	 */
 	int (*resume)(struct scsi_device *);
-	int (*suspend)(struct scsi_device *);
+	int (*suspend)(struct scsi_device *, pm_message_t state);
 
 	/*
 	 * Name of proc directory
#1 lore raw
Quoted 6 lines · Nigel Cunningham

At the moment libata doesn't pass pm_message_t down ata_device_suspend.
This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Does this explain http://bugzilla.kernel.org/show_bug.cgi?id=6264 ?

This might be 2.6.16.1 material - how irritating is it?

#2 lore raw

Hi.

Quoted 8 lines · Andrew Morton
Quoted 5 lines · Nigel Cunningham

At the moment libata doesn't pass pm_message_t down ata_device_suspend.
This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Does this explain http://bugzilla.kernel.org/show_bug.cgi?id=6264 ?

Yes, it does.

Quoted 1 line

This might be 2.6.16.1 material - how irritating is it?

Very. It extends the time to write the image, but as mentioned above, I'm more concerned by the fact that (assuming I understand correctly), it's using up the limited number of power cycles a drive can handle.

Regards,

Nigel

#4 lore raw
In reply to Nigel Cunningham

drivers/scsi/libata-scsi.c int ata_scsi_device_suspend(struct scsi_ · line 423

Quoted 10 lines · Nigel Cunningham

diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/libata-scsi.c 9904-libata-freeze.patch-new/drivers/scsi/libata-scsi.c

drivers/scsi/libata-scsi.c +1 -1
@@ -419,7 +419,7 @@ int ata_scsi_device_suspend(struct scsi_
 	struct ata_port *ap = (struct ata_port *) &sdev->host->hostdata[0];
 	struct ata_device *dev = &ap->device[sdev->id];
 
-	return ata_device_suspend(ap, dev);
+	return ata_device_suspend(ap, dev, state);
 }

Eh this looks odd, you are forgetting to add pm_message_t state as an extra argument to that function.

Apart from that, patch looks good. It should definitely be included asap.

Signature · 1 line
Jens Axboe
#5 lore raw

Hi.

Quoted 18 lines · Jens Axboe

drivers/scsi/libata-scsi.c int ata_scsi_device_suspend(struct scsi_ · line 423

Quoted 13 lines · Nigel Cunningham

diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/libata-scsi.c 9904-libata-freeze.patch-new/drivers/scsi/libata-scsi.c ---
9904-libata-freeze.patch-old/drivers/scsi/libata-scsi.c 2006-03-23
21:16:22.000000000 +1000 +++
9904-libata-freeze.patch-new/drivers/scsi/libata-scsi.c 2006-03-23
21:24:06.000000000 +1000 @@ -419,7 +419,7 @@ int ata_scsi_device_suspend(struct scsi_

drivers/scsi/libata-scsi.c +1 -1
 	struct ata_port *ap = (struct ata_port *) &sdev->host->hostdata[0];
 	struct ata_device *dev = &ap->device[sdev->id];
 
-	return ata_device_suspend(ap, dev);
+	return ata_device_suspend(ap, dev, state);
 }

Eh this looks odd, you are forgetting to add pm_message_t state as an extra argument to that function.

Quoted 2 lines

Apart from that, patch looks good. It should definitely be included asap.

Good catch! Fixed, will resend.

Thanks!

Nigel

#6 lore raw

PATCH 1/1 Make libata not powerdown drivers on PM_EVENT_FREEZE.

In reply to Jens Axboe

Hi.

At the moment libata doesn't pass pm_message_t down ata_device_suspend. This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Prepared against git at the time of writing. Please apply.

 drivers/scsi/libata-core.c |    5 +++--
 drivers/scsi/libata-scsi.c |    4 ++--
 drivers/scsi/scsi_sysfs.c  |    2 +-
 include/linux/libata.h     |    4 ++--
 include/scsi/scsi_host.h   |    2 +-
 5 files changed, 9 insertions(+), 8 deletions(-)
drivers/scsi/libata-core.c +3 -2
diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/libata-core.c 9904-libata-freeze.patch-new/drivers/scsi/libata-core.c
@@ -4506,14 +4506,15 @@ int ata_device_resume(struct ata_port *a
  *	Flush the cache on the drive, if appropriate, then issue a
  *	standbynow command.
  */
-int ata_device_suspend(struct ata_port *ap, struct ata_device *dev)
+int ata_device_suspend(struct ata_port *ap, struct ata_device *dev, pm_message_t state)
 {
 	if (!ata_dev_present(dev))
 		return 0;
 	if (dev->class == ATA_DEV_ATA)
 		ata_flush_cache(ap, dev);
 
-	ata_standby_drive(ap, dev);
+	if (state.event != PM_EVENT_FREEZE)
+		ata_standby_drive(ap, dev);
 	ap->flags |= ATA_FLAG_SUSPENDED;
 	return 0;
 }
drivers/scsi/libata-scsi.c +2 -2
diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/libata-scsi.c 9904-libata-freeze.patch-new/drivers/scsi/libata-scsi.c
@@ -414,12 +414,12 @@ int ata_scsi_device_resume(struct scsi_d
 	return ata_device_resume(ap, dev);
 }
 
-int ata_scsi_device_suspend(struct scsi_device *sdev)
+int ata_scsi_device_suspend(struct scsi_device *sdev, pm_message_t state)
 {
 	struct ata_port *ap = (struct ata_port *) &sdev->host->hostdata[0];
 	struct ata_device *dev = &ap->device[sdev->id];
 
-	return ata_device_suspend(ap, dev);
+	return ata_device_suspend(ap, dev, state);
 }
 
 /**
drivers/scsi/scsi_sysfs.c +1 -1
diff -ruNp 9904-libata-freeze.patch-old/drivers/scsi/scsi_sysfs.c 9904-libata-freeze.patch-new/drivers/scsi/scsi_sysfs.c
@@ -286,7 +286,7 @@ static int scsi_bus_suspend(struct devic
 		return err;
 
 	if (sht->suspend)
-		err = sht->suspend(sdev);
+		err = sht->suspend(sdev, state);
 
 	return err;
 }
include/linux/libata.h +2 -2
diff -ruNp 9904-libata-freeze.patch-old/include/linux/libata.h 9904-libata-freeze.patch-new/include/linux/libata.h
@@ -515,9 +515,9 @@ extern void ata_eh_qc_retry(struct ata_q
 extern int ata_scsi_release(struct Scsi_Host *host);
 extern unsigned int ata_host_intr(struct ata_port *ap, struct ata_queued_cmd *qc);
 extern int ata_scsi_device_resume(struct scsi_device *);
-extern int ata_scsi_device_suspend(struct scsi_device *);
+extern int ata_scsi_device_suspend(struct scsi_device *, pm_message_t state);
 extern int ata_device_resume(struct ata_port *, struct ata_device *);
-extern int ata_device_suspend(struct ata_port *, struct ata_device *);
+extern int ata_device_suspend(struct ata_port *, struct ata_device *, pm_message_t state);
 extern int ata_ratelimit(void);
 extern unsigned int ata_busy_sleep(struct ata_port *ap,
 				   unsigned long timeout_pat,
include/scsi/scsi_host.h +1 -1
diff -ruNp 9904-libata-freeze.patch-old/include/scsi/scsi_host.h 9904-libata-freeze.patch-new/include/scsi/scsi_host.h
@@ -286,7 +286,7 @@ struct scsi_host_template {
 	 * suspend support
 	 */
 	int (*resume)(struct scsi_device *);
-	int (*suspend)(struct scsi_device *);
+	int (*suspend)(struct scsi_device *, pm_message_t state);
 
 	/*
 	 * Name of proc directory
#7 lore raw
Quoted 11 lines · Nigel Cunningham

Hi.

At the moment libata doesn't pass pm_message_t down ata_device_suspend. This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Prepared against git at the time of writing. Please apply.

I'll put this into the queue for review.

As the top of each source file requests, please CC [email protected] and myself on libata changes.

Jeff

#8 lore raw

Hi.

Quoted 9 lines · Jeff Garzik
Quoted 8 lines · Nigel Cunningham

Hi.

At the moment libata doesn't pass pm_message_t down ata_device_suspend. This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Quoted 8 lines
Quoted 3 lines

Prepared against git at the time of writing. Please apply.

I'll put this into the queue for review.

As the top of each source file requests, please CC [email protected] and myself on libata changes.

Ok. Thanks.

Nigel

#9 lore raw
Quoted 15 lines · Nigel Cunningham

Hi.

Quoted 12 lines · Jeff Garzik
Quoted 11 lines · Nigel Cunningham

Hi.

At the moment libata doesn't pass pm_message_t down ata_device_suspend. This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Prepared against git at the time of writing. Please apply.

Works very well here for me, and long overdue it was!

Thanks, Nigel!

#3 lore raw
In reply to Andrew Morton

Gi!

Quoted 9 lines
Quoted 5 lines

At the moment libata doesn't pass pm_message_t down ata_device_suspend.
This causes drives to be powered down when we just want a freeze, causing unnecessary wear and tear. This patch gets pm_message_t passed down so that it can be used to determine whether to power down the drive.

Does this explain http://bugzilla.kernel.org/show_bug.cgi?id=6264 ?

This might be 2.6.16.1 material - how irritating is it?

I'd say patch is okay for mainline, but not worth it for 2.6.16.1 --- we have lived with similar problem for quite a long time for PATA.

Signature · 1 line
Thanks, Sharp!