Skip to contentSkip to conversation

Spinup command fails on WD3200JS

PATCH In review Ryan Power 3 participants 9 messages Jul 13, 2007 – Jul 16, 2007 libata

lore
Share & tools
lore

Ryan Power

author
#0 lore raw

Hi.
I've run into a problem with the lastest kernel (2.6.22.1) and a drive set to spinup on set_features.

hdparm reports:
/dev/sdf:

ATA device, with non-removable media powers-up in standby; SET FEATURES subcmd spins-up.

        Model Number:       WDC WD3200JS-57PDB0
        Serial Number:      X
        Firmware Revision:  21.00M21

...
Commands/features:
Enabled Supported:
* Power-Up In Standby feature set
* SET_FEATURES required to spinup after power up
...

I get the following messages when booting:

Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: ATA-7: WDC WD3200JS-57PDB0, 21.00M21, max UDMA/133
Jul 12 23:47:22 sysreset kernel: ata6.00: 625142448 sectors, multi 0: LBA48 NCQ (depth 1)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6: failed to recover some devices, retrying in 5 secs
Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6.00: limiting speed to UDMA/133:PIO3
Jul 12 23:47:22 sysreset kernel: ata6: failed to recover some devices, retrying in 5 secs
Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6.00: disabled

The interesting thing is that after displaying that, the drive has in fact spun up. I'm not sure why it's getting an error returned, but chaning the err_mask line to read "if (err_mask && err_mask != 0x01)" corrects the problem and the drive spins up and ID's normally:
ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata6.00: ATA-7: WDC WD3200JS-57PDB0, 21.00M21, max UDMA/133
ata6.00: 625142448 sectors, multi 0: LBA48 NCQ (depth 1)
ata6.00: configured for UDMA/133

If you'd like me to do any more testing or need more information from me please let me know. Thank you.

-Ryan Power

#1 lore raw

Mark?

Quoted 58 lines · Ryan Power

Hi.
I've run into a problem with the lastest kernel (2.6.22.1) and a drive set to spinup on set_features.

hdparm reports:
/dev/sdf:

ATA device, with non-removable media powers-up in standby; SET FEATURES subcmd spins-up.

        Model Number:       WDC WD3200JS-57PDB0
        Serial Number:      X
        Firmware Revision:  21.00M21

...
Commands/features:
Enabled Supported:
* Power-Up In Standby feature set
* SET_FEATURES required to spinup after power up
...

I get the following messages when booting:

Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus
113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: ATA-7: WDC WD3200JS-57PDB0,
21.00M21, max UDMA/133
Jul 12 23:47:22 sysreset kernel: ata6.00: 625142448 sectors, multi 0:
LBA48 NCQ (depth 1)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP
failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6: failed to recover some devices,
retrying in 5 secs
Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus
113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP
failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6.00: limiting speed to UDMA/133:PIO3
Jul 12 23:47:22 sysreset kernel: ata6: failed to recover some devices,
retrying in 5 secs
Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus
113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP
failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6.00: disabled

The interesting thing is that after displaying that, the drive has in fact spun up. I'm not sure why it's getting an error returned, but chaning the err_mask line to read "if (err_mask && err_mask != 0x01)" corrects the problem and the drive spins up and ID's normally:
ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300) ata6.00: ATA-7: WDC WD3200JS-57PDB0, 21.00M21, max UDMA/133 ata6.00: 625142448 sectors, multi 0: LBA48 NCQ (depth 1) ata6.00: configured for UDMA/133

If you'd like me to do any more testing or need more information from me please let me know.
Thank you.

Signature · 1 line
tejun
#2 lore raw
In reply to Ryan Power
Quoted 2 lines

Hi.
I've run into a problem with the lastest kernel (2.6.22.1) and a drive set to spinup on set_features.

..

Quoted 9 lines

ATA device, with non-removable media powers-up in standby; SET FEATURES subcmd spins-up.

        Model Number:       WDC WD3200JS-57PDB0
        Serial Number:      X
        Firmware Revision:  21.00M21

...
* Power-Up In Standby feature set
* SET_FEATURES required to spinup after power up
...

Quoted 15 lines
Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: ATA-7: WDC WD3200JS-57PDB0, 21.00M21, max UDMA/133
Jul 12 23:47:22 sysreset kernel: ata6.00: 625142448 sectors, multi 0: LBA48 NCQ (depth 1)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6: failed to recover some devices, retrying in 5 secs
Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6.00: limiting speed to UDMA/133:PIO3
Jul 12 23:47:22 sysreset kernel: ata6: failed to recover some devices, retrying in 5 secs
Jul 12 23:47:22 sysreset kernel: ata6: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
Jul 12 23:47:22 sysreset kernel: ata6.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
Jul 12 23:47:22 sysreset kernel: ata6.00: revalidation failed (errno=-5)
Jul 12 23:47:22 sysreset kernel: ata6.00: disabled

It would be handy to see those messages again with real timestamps on them. Ryan, could you add CONFIG_PRINTK_TIME=y into your .config and rebuild/retry ?

Quoted 4 lines

The interesting thing is that after displaying that, the drive has in fact spun up. I'm not sure why it's getting an error returned, but chaning the err_mask line to read "if (err_mask && err_mask != 0x01)" corrects the problem and the drive spins up and ID's normally:

Yeah, I believe that for WD drives. Their firmware has *always* been peculiar. In fact, the WD Raptor drives I used to have here had similar, though not identical, problems with the "jumper-enabled" power-on-in-standby version. I never did figure those ones out.

But yes, perhaps your suggestion of just ignoring the device's response may be best. Are you sure it spins up on the very first attempt? Also, could you add this printk() line to just before your err_mask hack:

  printk("id[2]=0x%04x\n", id[2]);

If we see 0x738c reported from that, then we *know* we should just continue rather than err'ing out as is presently done.

Thanks

author
#3 lore raw

Here's the section with the modifications you requested:

[    2.212000] sata_sil 0000:01:08.0: version 2.2
[    2.216000] ACPI: PCI Interrupt Link [APC3] enabled at IRQ 18
[    2.216000] ACPI: PCI Interrupt 0000:01:08.0[A] -> Link [APC3] -> GSI 18 (level, low) -> IRQ 16
[    2.216000] PCI: Setting latency timer of device 0000:01:08.0 to 64
[    2.216000] scsi0 : sata_sil
[    2.216000] scsi1 : sata_sil
[    2.216000] ata1: SATA max UDMA/100 cmd 0xf880c080 ctl 0xf880c08a bmdma 0xf880c000 irq 16
[    2.216000] ata2: SATA max UDMA/100 cmd 0xf880c0c0 ctl 0xf880c0ca bmdma 0xf880c008 irq 16
[    2.684000] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[    2.692000] ata1.00: ATA-7: Maxtor 6L320S0, BACE1G20, max UDMA/133
[    2.692000] ata1.00: 625142448 sectors, multi 16: LBA48 NCQ (depth 0/32)
[    2.708000] ata1.00: configured for UDMA/100
[    3.176000] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[   11.020000] id[2]=0x37c8
[   11.036000] ata2.00: ATA-7: WDC WD3200JS-57PDB0, 21.00M21, max UDMA/133
[   11.036000] ata2.00: 625142448 sectors, multi 0: LBA48 NCQ (depth 0/1)
[   11.052000] id[2]=0x738c
[   11.052000] ata2.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
[   11.052000] ata2.00: revalidation failed (errno=-5)
[   11.052000] ata2: failed to recover some devices, retrying in 5 secs
[   16.532000] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[   16.540000] id[2]=0x738c
[   16.540000] ata2.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
[   16.540000] ata2.00: revalidation failed (errno=-5)
[   16.540000] ata2.00: limiting speed to UDMA/100:PIO3
[   16.540000] ata2: failed to recover some devices, retrying in 5 secs
[   22.020000] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[   22.028000] id[2]=0x738c
[   22.028000] ata2.00: failed to IDENTIFY (SPINUP failed, err_mask=0x1)
[   22.028000] ata2.00: revalidation failed (errno=-5)
[   22.028000] ata2.00: disabled
[   22.532000] scsi 0:0:0:0: Direct-Access     ATA      Maxtor 6L320S0   BACE PQ: 0 ANSI: 5
[   22.532000] sd 0:0:0:0: [sda] 625142448 512-byte hardware sectors (320073 MB)
[   22.532000] sd 0:0:0:0: [sda] Write Protect is off
[   22.532000] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00

I'm positive that it's spun up after the first command is execicuted. I can hear the drive spinup, and doing a soft reboot allows it to be detected normally. The same section with "if (err_mask && id[2] != 0x738c)" in place of the previous modification:

[    2.216000] scsi0 : sata_sil
[    2.216000] scsi1 : sata_sil
[    2.216000] ata1: SATA max UDMA/100 cmd 0xf880c080 ctl 0xf880c08a bmdma 0xf880c000 irq 16
[    2.216000] ata2: SATA max UDMA/100 cmd 0xf880c0c0 ctl 0xf880c0ca bmdma 0xf880c008 irq 16
[    2.684000] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[    2.692000] ata1.00: ATA-7: Maxtor 6L320S0, BACE1G20, max UDMA/133
[    2.692000] ata1.00: 625142448 sectors, multi 16: LBA48 NCQ (depth 0/32)
[    2.708000] ata1.00: configured for UDMA/100
[    3.176000] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[   10.944000] id[2]=0x37c8
[   10.964000] ata2.00: ATA-7: WDC WD3200JS-57PDB0, 21.00M21, max UDMA/133
[   10.964000] ata2.00: 625142448 sectors, multi 0: LBA48 NCQ (depth 0/1)
[   10.972000] id[2]=0x738c
[   10.972000] ata2.00: configured for UDMA/100

I'm uncertain as to which would be better to do; ignoring the error or looking for the STBY_ID value.

Thanks.
-Ryan Power

Quoted 27 lines · Mark Lord
Quoted 2 lines

Hi.
I've run into a problem with the lastest kernel (2.6.22.1) and a drive set to spinup on set_features.

..

It would be handy to see those messages again with real timestamps on them. Ryan, could you add CONFIG_PRINTK_TIME=y into your .config and rebuild/retry ?

Quoted 4 lines

The interesting thing is that after displaying that, the drive has in fact spun up. I'm not sure why it's getting an error returned, but chaning the err_mask line to read "if (err_mask && err_mask != 0x01)" corrects the problem and the drive spins up and ID's normally:

Yeah, I believe that for WD drives. Their firmware has *always* been peculiar. In fact, the WD Raptor drives I used to have here had similar, though not identical, problems with the "jumper-enabled" power-on-in-standby version. I never did figure those ones out.

But yes, perhaps your suggestion of just ignoring the device's response may be best. Are you sure it spins up on the very first attempt? Also, could you add this printk() line to just before your err_mask hack:

  printk("id[2]=0x%04x\n", id[2]);

If we see 0x738c reported from that, then we *know* we should just continue rather than err'ing out as is presently done.

Thanks

#4 lore raw
Quoted 1 line · Ryan Power

Here's the section with the modifications you requested:

..

Quoted 11 lines

The same section with "if (err_mask && id[2] != 0x738c)" in place of the previous modification:

[    2.216000] scsi0 : sata_sil
[    2.216000] scsi1 : sata_sil
[    2.216000] ata1: SATA max UDMA/100 cmd 0xf880c080 ctl 0xf880c08a bmdma 0xf880c000 irq 16
[    2.216000] ata2: SATA max UDMA/100 cmd 0xf880c0c0 ctl 0xf880c0ca bmdma 0xf880c008 irq 16
[    2.684000] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[    2.692000] ata1.00: ATA-7: Maxtor 6L320S0, BACE1G20, max UDMA/133
[    2.692000] ata1.00: 625142448 sectors, multi 16: LBA48 NCQ (depth 0/32)
[    2.708000] ata1.00: configured for UDMA/100
[    3.176000] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[   10.944000] id[2]=0x37c8

Okay, above we can see the drive requires a spin-up command, which now gets issued, and then we loop to retry the IDENTIFY.

Quoted 4 lines
[   10.964000] ata2.00: ATA-7: WDC WD3200JS-57PDB0, 21.00M21, max UDMA/133
[   10.964000] ata2.00: 625142448 sectors, multi 0: LBA48 NCQ (depth 0/1)
[   10.972000] id[2]=0x738c
[   10.972000] ata2.00: configured for UDMA/100

There we can see that, after the spin-up command, the IDENTIFY succeeds, and the new signature value of 0x728c indicates that full IDENTIFY data is now present. So we can just continue at that point.

Good!

Now.. time to get your name in lights and all that! Would you like to package up that new if-stmt (below) as a nice and proper kernel patch? Just follow the advice in the SubmittingPatches file, or model things after other patches you've seen here, and post the result with a proper Signed-Off-By line, and I'll ack it.

		if (err_mask && id[2] != 0x738c)

Or leave it to me, and I'll steal your fame and fortune! :)

Cheers
Thanks Ryan!

author
#5 lore raw

PATCH 1/1 libata: Adjust libata to ignore errors after spinup

From: Ryan Power <[email protected]>

Adjust libata to ignore errors after spinup

This patch is to ignore errors from the spinup attempt if the drive is in the "standby id" state.

---

Index: drivers.new/ata/libata-core.c

ata/libata-core.c 2007-07-15 01:58:49.000000000 -0600 renamed from ata/libata-core.c 2007-07-10 12:56:30.000000000 -0600 +0 -0
@@ -1750,7 +1750,7 @@ int ata_dev_read_id(struct ata_device *d
                tf.protocol = ATA_PROT_NODATA;
                tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_DEVICE;
                err_mask = ata_exec_internal(dev, &tf, NULL, DMA_NONE, 

NULL, 0);
- if (err_mask) {

+               if (err_mask && id[2] != 0x738c) {
                        rc = -EIO;
                        reason = "SPINUP failed";
                        goto err_out;
#6 lore raw

Re: [PATCH 2.6.22.1] libata: Adjust libata to ignore errors after spinup

Quoted 6 lines · Ryan Power

From: Ryan Power <[email protected]>

Adjust libata to ignore errors after spinup

This patch is to ignore errors from the spinup attempt if the drive is in the "standby id" state.

Jeff / Tejun: here is Ryan's patch for his WD drive spinup problems. This copy has the pathnames and whitespace repaired.

---
Index: linux/drivers/ata/libata-core.c

drivers/ata/libata-core.c 2007-07-15 01:58:49.000000000 -0600 renamed from drivers/ata/libata-core.c 2007-07-10 12:56:30.000000000 -0600 +0 -0
@@ -1750,7 +1750,7 @@ int ata_dev_read_id(struct ata_device *d
		tf.protocol = ATA_PROT_NODATA;
		tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_DEVICE;
		err_mask = ata_exec_internal(dev, &tf, NULL, DMA_NONE, NULL, 0);
-		if (err_mask) {
+		if (err_mask && id[2] != 0x738c) {
			rc = -EIO;
			reason = "SPINUP failed";
			goto err_out;
#7 lore raw

Re: [PATCH 2.6.22.1] libata: Adjust libata to ignore errors after spinup

Quoted 10 lines · Mark Lord
Quoted 6 lines · Ryan Power

From: Ryan Power <[email protected]>

Adjust libata to ignore errors after spinup

This patch is to ignore errors from the spinup attempt if the drive is in the "standby id" state.

Jeff / Tejun: here is Ryan's patch for his WD drive spinup problems. This copy has the pathnames and whitespace repaired.

Mmm.. something weird happened there.
Here it is again, with corrected whitespace.

---
Index: linux/drivers/ata/libata-core.c

drivers/ata/libata-core.c 2007-07-15 01:58:49.000000000 -0600 renamed from drivers/ata/libata-core.c 2007-07-10 12:56:30.000000000 -0600 +0 -0
@@ -1750,7 +1750,7 @@ int ata_dev_read_id(struct ata_device *d
		tf.protocol = ATA_PROT_NODATA;
		tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_DEVICE;
		err_mask = ata_exec_internal(dev, &tf, NULL, DMA_NONE, NULL, 0);
-		if (err_mask) {
+		if (err_mask && id[2] != 0x738c) {
			rc = -EIO;
			reason = "SPINUP failed";
			goto err_out;
#8 lore raw

Re: [PATCH 2.6.22.1] libata: Adjust libata to ignore errors after spinup

Quoted 10 lines · Mark Lord
Quoted 6 lines · Ryan Power

From: Ryan Power <[email protected]>

Adjust libata to ignore errors after spinup

This patch is to ignore errors from the spinup attempt if the drive is in the "standby id" state.

Jeff / Tejun: here is Ryan's patch for his WD drive spinup problems. This copy has the pathnames and whitespace repaired.

Mmm.. Thunderbird-2.x seems to no longer keep my whitespace intact, even when using the external editor plugin..

Hopefully this one (posted with thunderbird-1.5 + exteditor.xpi) works:

---
Index: linux/drivers/ata/libata-core.c

drivers/ata/libata-core.c 2007-07-15 01:58:49.000000000 -0600 renamed from drivers/ata/libata-core.c 2007-07-10 12:56:30.000000000 -0600 +0 -0
@@ -1750,7 +1750,7 @@ int ata_dev_read_id(struct ata_device *d
		tf.protocol = ATA_PROT_NODATA;
		tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_DEVICE;
		err_mask = ata_exec_internal(dev, &tf, NULL, DMA_NONE, NULL, 0);
-		if (err_mask) {
+		if (err_mask && id[2] != 0x738c) {
			rc = -EIO;
			reason = "SPINUP failed";
			goto err_out;