[PATCH 1/1] libata: blacklist NCQ on WDC WD740ADFD-00NLR4

[PATCH 1/1] libata: blacklist NCQ on WDC WD740ADFD-00NLR4

[Simon J. Rowe]

I've been seeing intermittent issues on my desktop (running 2.6.38.7) recently

Jan 16 20:54:08 gilraen kernel: ata5.00: exception Emask 0x10 SAct 0x807f SErr 
0x400100 action 0x6 frozen
Jan 16 20:54:08 gilraen kernel: ata5.00: irq_stat 0x08000000, interface fatal 
error
Jan 16 20:54:08 gilraen kernel: ata5: SError: { UnrecovData Handshk }
Jan 16 20:54:08 gilraen kernel: ata5.00: failed command: WRITE FPDMA QUEUED
Jan 16 20:54:08 gilraen kernel: ata5.00: cmd 
61/d8:00:ef:13:d8/00:00:00:00:00/40 tag 0 ncq 110592 out
Jan 16 20:54:08 gilraen kernel:         res 
40/00:20:97:13:d8/00:00:00:00:00/40 Emask 0x10 (ATA bus error)
Jan 16 20:54:08 gilraen kernel: ata5.00: status: { DRDY }

The disk is a WD Raptor WD740 which I noticed has NCQ blacklisted. The exact 
model isn't present in the blacklist so I added it (patch below). The errors 
then came out slightly differently

Jan 16 21:17:08 gilraen kernel: ata5.00: exception Emask 0x10 SAct 0x0 SErr 
0x400100 action 0x6 frozen
Jan 16 21:17:08 gilraen kernel: ata5.00: irq_stat 0x08000000, interface fatal 
error
Jan 16 21:17:08 gilraen kernel: ata5: SError: { UnrecovData Handshk }
Jan 16 21:17:08 gilraen kernel: ata5.00: failed command: WRITE DMA EXT
Jan 16 21:17:08 gilraen kernel: ata5.00: cmd 
35/00:68:6f:5d:9a/00:01:00:00:00/e0 tag 0 dma 184320 out
Jan 16 21:17:08 gilraen kernel:         res 
50/00:00:66:5d:9a/00:00:00:00:00/e0 Emask 0x10 (ATA bus error)
Jan 16 21:17:08 gilraen kernel: ata5.00: status: { DRDY }
Jan 16 21:17:08 gilraen kernel: ata5: hard resetting link
Jan 16 21:17:08 gilraen kernel: ata5: SATA link up 1.5 Gbps (SStatus 113 
SControl 300)
Jan 16 21:17:08 gilraen kernel: ata5.00: configured for UDMA/133
Jan 16 21:17:08 gilraen kernel: ata5: EH complete

Googling revealed that this is often caused by faulty cabling. I moved the 
power and SATA cables and the problem has now gone but I think the disk should 
still be added to the blacklist,

Thanks
		Simon


>From 5df7fe5040b02281617965f99931a9a8ef30b9cb Mon Sep 17 00:00:00 2001
From: Simon J. Rowe <srowe@mose.org.uk>
Date: Wed, 18 Jan 2012 10:14:02 +0000
Subject: [PATCH 1/1] libata: blacklist NCQ on WDC WD740ADFD-00NLR4

Add another Western Digital WD740 model to the blacklist.

Signed-off-by: Simon J. Rowe <srowe@mose.org.uk>
---
 drivers/ata/libata-core.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c
index 76c3c15..1d06f37 100644
--- a/drivers/ata/libata-core.c
+++ b/drivers/ata/libata-core.c
@@ -4077,6 +4077,7 @@ static const struct ata_blacklist_entry 
ata_device_blacklist [] = {
 	/* NCQ is slow */
 	{ "WDC WD740ADFD-00",	NULL,		ATA_HORKAGE_NONCQ },
 	{ "WDC WD740ADFD-00NLR1", NULL,		ATA_HORKAGE_NONCQ, },
+	{ "WDC WD740ADFD-00NLR4", NULL,		ATA_HORKAGE_NONCQ, },
 	/* http://thread.gmane.org/gmane.linux.ide/14907 */
 	{ "FUJITSU MHT2060BH",	NULL,		ATA_HORKAGE_NONCQ },
 	/* NCQ is broken */
-- 
1.7.2.5

Re: [PATCH 1/1] libata: blacklist NCQ on WDC WD740ADFD-00NLR4

[Alan Cox]

> Jan 16 21:17:08 gilraen kernel: ata5: SError: { UnrecovData Handshk }
> Jan 16 21:17:08 gilraen kernel: ata5.00: failed command: WRITE DMA EXT
> Jan 16 21:17:08 gilraen kernel: ata5.00: cmd 
> 35/00:68:6f:5d:9a/00:01:00:00:00/e0 tag 0 dma 184320 out
> Jan 16 21:17:08 gilraen kernel:         res 
> 50/00:00:66:5d:9a/00:00:00:00:00/e0 Emask 0x10 (ATA bus error)
> Jan 16 21:17:08 gilraen kernel: ata5.00: status: { DRDY }
> Jan 16 21:17:08 gilraen kernel: ata5: hard resetting link
> Jan 16 21:17:08 gilraen kernel: ata5: SATA link up 1.5 Gbps (SStatus 113 
> SControl 300)
> Jan 16 21:17:08 gilraen kernel: ata5.00: configured for UDMA/133
> Jan 16 21:17:08 gilraen kernel: ata5: EH complete
> 
> Googling revealed that this is often caused by faulty cabling. I moved the 
> power and SATA cables and the problem has now gone but I think the disk should 
> still be added to the blacklist,

We add things to the NCQ blacklist because we have multiple clear cases
showing NCQ problems and preferably vendor confirmation. You have neither
and as you said yourself when you fixed the cables it went away.

Your drive is report UnrecovData Handshk, and its' reporting it on a non
NCQ command in your second example. You had a cable problem.

So NAK this patch

Alan