SATA DOM is not identified by ata_piix module

SATA DOM is not identified by ata_piix module

[Mao Rui]

Hi,

I have a PQI Turbo SATA DOM. It works well under Windows. I installed it in
a SuperMicro motherboard, Intel 5000P chipset. The OS is Ubuntu 7.04, kernel
2.6.20-15. But the DOM is not appeared as a device node, and I found several
error messages in kernel log.

[   67.124299] ata2.00: qc timeout (cmd 0x91)
[   67.124306] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
err_mask=0x4)
[   67.124309] ata2: failed to recover some devices, retrying in 5 secs
[  102.258334] ata2.00: qc timeout (cmd 0x91)
[  102.258341] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
err_mask=0x4)
[  102.258344] ata2: failed to recover some devices, retrying in 5 secs
[  137.392654] ata2.00: qc timeout (cmd 0x91)
[  137.392661] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
err_mask=0x4)

Actually I also have a PQI IDE DOM, it have same error with Ubuntu 7.04 /
kernel 2.6.20-15. After I upgrade kernel to 2.6.22, the IDE DOM can work.
But as for SATA DOM, I have no luck this time. Even with the newest kernel,
2.6.23.14 and 2.6.24-rc7-git6, the problem still exists.
I also found an interesting matter. The Ubuntu 6.10, say kernel 2.6.15-28,
can identify both IDE DOM and SATA DOM. But due to other critical problem, I
can not go back to this old kernel.

What’s going on? The new kernel is even worse than old one. It seems the
problem is existing in the ata_piix driver.

Best regards,
Mao Rui

Re: SATA DOM is not identified by ata_piix module

[Alan Cox]

On Tue, 15 Jan 2008 18:11:25 +0800
"Mao Rui" <maorui@yuvad.com> wrote:

> Hi,
> 
> I have a PQI Turbo SATA DOM. It works well under Windows. I installed it in
> a SuperMicro motherboard, Intel 5000P chipset. The OS is Ubuntu 7.04, kernel
> 2.6.20-15. But the DOM is not appeared as a device node, and I found several
> error messages in kernel log.

Generally it is a good idea to report problems with vendor built kernels
to the vendor and their support, especially one that is 3 releases behind.
They have a much better idea what is in that kernel and who else has seen
problems.

> [   67.124299] ata2.00: qc timeout (cmd 0x91)
> [   67.124306] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
> err_mask=0x4)

We issued commands, it didn't respond. That could be a libata problem but
actually looks more like an IRQ routing problem.

> Actually I also have a PQI IDE DOM, it have same error with Ubuntu 7.04 /

The PQI DOM is a bit odd, it is however known to work with libata at
least for the PATA one, and the versions which don't understand
SET_XFER_MODE to work with current kernels. (Your failure mode isn't the
SET_XFER_MODE one - it hasn't got that far).

Alan

RE: SATA DOM is not identified by ata_piix module

[Mao Rui]

I tried to nail down when the problem was introduced. I compiled some
official kernel release. Here is the result.
2.6.17.14   IDE -- Failed   SATA -- passed
2.6.18   IDE -- Failed   SATA -- failed
2.6.18.8   IDE -- Failed   SATA -- failed
2.6.24-rc7-git6  IDE -- passed   SATA DOM -- failed
linux-2.6.24-rc8-git1   IDE -- passed    SATA -- failed
All IDE failed reason is xfermode error, and all SATA failure is IDENTIFY
error.

As you can find out, the failure of SATA DOM was introduced from kernel
2.6.18.

I'm not good at low level driver programming, so I cannot find out the root
cause by myself. But if Alan or someone else needs more info or wants to
test the patch, I'm glad to do it in my platform.

Best regards,
Mao Rui

> -----Original Message-----
> From: Alan Cox [mailto:alan@lxorguk.ukuu.org.uk]
> Sent: Tuesday, January 15, 2008 7:02 PM
> To: Mao Rui
> Cc: linux-kernel@vger.kernel.org
> Subject: Re: SATA DOM is not identified by ata_piix module
> 
> On Tue, 15 Jan 2008 18:11:25 +0800
> "Mao Rui" <maorui@yuvad.com> wrote:
> 
> > Hi,
> >
> > I have a PQI Turbo SATA DOM. It works well under Windows. I installed it
in
> > a SuperMicro motherboard, Intel 5000P chipset. The OS is Ubuntu 7.04,
kernel
> > 2.6.20-15. But the DOM is not appeared as a device node, and I found
several
> > error messages in kernel log.
> 
> Generally it is a good idea to report problems with vendor built kernels
> to the vendor and their support, especially one that is 3 releases behind.
> They have a much better idea what is in that kernel and who else has seen
> problems.
> 
> > [   67.124299] ata2.00: qc timeout (cmd 0x91)
> > [   67.124306] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
> > err_mask=0x4)
> 
> We issued commands, it didn't respond. That could be a libata problem but
> actually looks more like an IRQ routing problem.
> 
> > Actually I also have a PQI IDE DOM, it have same error with Ubuntu 7.04
/
> 
> The PQI DOM is a bit odd, it is however known to work with libata at
> least for the PATA one, and the versions which don't understand
> SET_XFER_MODE to work with current kernels. (Your failure mode isn't the
> SET_XFER_MODE one - it hasn't got that far).
> 
> Alan

RE: SATA DOM is not identified by ata_piix module

[Mao Rui]

Hmm... Does anybody own this bug?

Best regards,

Mao Rui


> -----Original Message-----
> From: linux-kernel-owner@vger.kernel.org
> [mailto:linux-kernel-owner@vger.kernel.org] On Behalf Of Mao Rui
> Sent: Friday, January 18, 2008 2:12 PM
> To: 'Alan Cox'
> Cc: linux-kernel@vger.kernel.org
> Subject: RE: SATA DOM is not identified by ata_piix module
> 
> I tried to nail down when the problem was introduced. I compiled some
> official kernel release. Here is the result.
> 2.6.17.14   IDE -- Failed   SATA -- passed
> 2.6.18   IDE -- Failed   SATA -- failed
> 2.6.18.8   IDE -- Failed   SATA -- failed
> 2.6.24-rc7-git6  IDE -- passed   SATA DOM -- failed
> linux-2.6.24-rc8-git1   IDE -- passed    SATA -- failed
> All IDE failed reason is xfermode error, and all SATA failure is IDENTIFY
> error.
> 
> As you can find out, the failure of SATA DOM was introduced from kernel
> 2.6.18.
> 
> I'm not good at low level driver programming, so I cannot find out the
root
> cause by myself. But if Alan or someone else needs more info or wants to
> test the patch, I'm glad to do it in my platform.
> 
> Best regards,
> Mao Rui
> 
> > -----Original Message-----
> > From: Alan Cox [mailto:alan@lxorguk.ukuu.org.uk]
> > Sent: Tuesday, January 15, 2008 7:02 PM
> > To: Mao Rui
> > Cc: linux-kernel@vger.kernel.org
> > Subject: Re: SATA DOM is not identified by ata_piix module
> >
> > On Tue, 15 Jan 2008 18:11:25 +0800
> > "Mao Rui" <maorui@yuvad.com> wrote:
> >
> > > Hi,
> > >
> > > I have a PQI Turbo SATA DOM. It works well under Windows. I installed
it
> in
> > > a SuperMicro motherboard, Intel 5000P chipset. The OS is Ubuntu 7.04,
> kernel
> > > 2.6.20-15. But the DOM is not appeared as a device node, and I found
> several
> > > error messages in kernel log.
> >
> > Generally it is a good idea to report problems with vendor built kernels
> > to the vendor and their support, especially one that is 3 releases
behind.
> > They have a much better idea what is in that kernel and who else has
seen
> > problems.
> >
> > > [   67.124299] ata2.00: qc timeout (cmd 0x91)
> > > [   67.124306] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
> > > err_mask=0x4)
> >
> > We issued commands, it didn't respond. That could be a libata problem
but
> > actually looks more like an IRQ routing problem.
> >
> > > Actually I also have a PQI IDE DOM, it have same error with Ubuntu
7.04
> /
> >
> > The PQI DOM is a bit odd, it is however known to work with libata at
> > least for the PATA one, and the versions which don't understand
> > SET_XFER_MODE to work with current kernels. (Your failure mode isn't the
> > SET_XFER_MODE one - it hasn't got that far).
> >
> > Alan
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: SATA DOM is not identified by ata_piix module

[Alan Cox]

On Wed, 30 Jan 2008 10:42:12 +0800
"Mao Rui" <maorui@yuvad.com> wrote:

> Hmm... Does anybody own this bug?

No hardware so its really up to someone with hardware to fix. I have a
long list of things to sort out which affect more than one person and for
which I do posess hardware so its priority is pretty low right now.

Re: SATA DOM is not identified by ata_piix module

[Tejun Heo]

[-- Attachment #1: Type: text/plain, Size: 814 bytes --]

Mao Rui wrote:
> I tried to nail down when the problem was introduced. I compiled some
> official kernel release. Here is the result.
> 2.6.17.14   IDE -- Failed   SATA -- passed
> 2.6.18   IDE -- Failed   SATA -- failed
> 2.6.18.8   IDE -- Failed   SATA -- failed
> 2.6.24-rc7-git6  IDE -- passed   SATA DOM -- failed
> linux-2.6.24-rc8-git1   IDE -- passed    SATA -- failed
> All IDE failed reason is xfermode error, and all SATA failure is IDENTIFY
> error.
> 
> As you can find out, the failure of SATA DOM was introduced from kernel
> 2.6.18.
> 
> I'm not good at low level driver programming, so I cannot find out the root
> cause by myself. But if Alan or someone else needs more info or wants to
> test the patch, I'm glad to do it in my platform.

Does the attached patch make any difference?

-- 
tejun

[-- Attachment #2: ata_dev_init_params-debug.patch --]
[-- Type: text/x-patch, Size: 997 bytes --]

diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c
index ce803d1..76d779a 100644
--- a/drivers/ata/libata-core.c
+++ b/drivers/ata/libata-core.c
@@ -1975,17 +1975,21 @@ int ata_dev_read_id(struct ata_device *dev, unsigned int *p_class,
 		 */
 		if (ata_id_major_version(id) < 4 || !ata_id_has_lba(id)) {
 			err_mask = ata_dev_init_params(dev, id[3], id[6]);
-			if (err_mask) {
+			if (err_mask == 0) {
+				/* current CHS translation info (id[53-58])
+				 * might be changed. reread the identify
+				 * device info.
+				 */
+				flags &= ~ATA_READID_POSTRESET;
+				goto retry;
+			} else if (err_mask == AC_ERR_DEV) {
+				ata_dev_printk(dev, KERN_INFO,
+					"INIT_DEV_PARAMS rejected by device\n");
+			} else {
 				rc = -EIO;
 				reason = "INIT_DEV_PARAMS failed";
 				goto err_out;
 			}
-
-			/* current CHS translation info (id[53-58]) might be
-			 * changed. reread the identify device info.
-			 */
-			flags &= ~ATA_READID_POSTRESET;
-			goto retry;
 		}
 	}
 

RE: SATA DOM is not identified by ata_piix module

[Mao Rui]

Hi,

No difference... :-(

Here is the log.

[   26.729011] scsi0 : ata_piix
[   26.729067] scsi1 : ata_piix
[   26.729095] ata1: SATA max UDMA/133 cmd 0x1880 ctl 0x1874 bmdma 0x1860
irq 18
[   26.729097] ata2: SATA max UDMA/133 cmd 0x1878 ctl 0x1870 bmdma 0x1868
irq 18
[   26.894043] ata1.00: ATA-6: WDC WD2500SD-01KCB0, 08.02D08, max UDMA/133
[   26.894046] ata1.00: 488397168 sectors, multi 16: LBA48
[   26.910218] ata1.00: configured for UDMA/133
[   57.043651] ata2.00: qc timeout (cmd 0x91)
[   57.043656] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
err_mask=0x4)
[   57.043659] ata2: failed to recover some devices, retrying in 5 secs
[   92.177684] ata2.00: qc timeout (cmd 0x91)
[   92.177688] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
err_mask=0x4)
[   92.177690] ata2: failed to recover some devices, retrying in 5 secs
[  127.312003] ata2.00: qc timeout (cmd 0x91)
[  127.312007] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
err_mask=0x4)
[  127.312010] ata2: failed to recover some devices, retrying in 5 secs
[  132.467196] scsi 0:0:0:0: Direct-Access     ATA      WDC WD2500SD-01K
08.0 PQ: 0 ANSI: 5
[  132.472308] Driver 'sd' needs updating - please use bus_type methods
[  132.472389] sd 0:0:0:0: [sda] 488397168 512-byte hardware sectors (250059
MB)
[  132.472400] sd 0:0:0:0: [sda] Write Protect is off
[  132.472403] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[  132.472416] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled,
doesn't support DPO or FUA
[  132.472459] sd 0:0:0:0: [sda] 488397168 512-byte hardware sectors (250059
MB)
[  132.472466] sd 0:0:0:0: [sda] Write Protect is off
[  132.472468] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[  132.472481] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled,
doesn't support DPO or FUA
[  132.472484]  sda: sda1 sda2 < sda5 >
[  132.512151] sd 0:0:0:0: [sda] Attached SCSI disk

Best regards,

Mao Rui


> -----Original Message-----
> From: Tejun Heo [mailto:htejun@gmail.com]
> Sent: Saturday, February 02, 2008 4:42 PM
> To: Mao Rui
> Cc: 'Alan Cox'; linux-kernel@vger.kernel.org
> Subject: Re: SATA DOM is not identified by ata_piix module
> 
> Mao Rui wrote:
> > I tried to nail down when the problem was introduced. I compiled some
> > official kernel release. Here is the result.
> > 2.6.17.14   IDE -- Failed   SATA -- passed
> > 2.6.18   IDE -- Failed   SATA -- failed
> > 2.6.18.8   IDE -- Failed   SATA -- failed
> > 2.6.24-rc7-git6  IDE -- passed   SATA DOM -- failed
> > linux-2.6.24-rc8-git1   IDE -- passed    SATA -- failed
> > All IDE failed reason is xfermode error, and all SATA failure is
IDENTIFY
> > error.
> >
> > As you can find out, the failure of SATA DOM was introduced from kernel
> > 2.6.18.
> >
> > I'm not good at low level driver programming, so I cannot find out the
root
> > cause by myself. But if Alan or someone else needs more info or wants to
> > test the patch, I'm glad to do it in my platform.
> 
> Does the attached patch make any difference?
> 
> --
> tejun

Re: SATA DOM is not identified by ata_piix module

[Tejun Heo]

Mao Rui wrote:
> Hi,
> 
> No difference... :-(
> 
> Here is the log.
> 
> [   26.729011] scsi0 : ata_piix
> [   26.729067] scsi1 : ata_piix
> [   26.729095] ata1: SATA max UDMA/133 cmd 0x1880 ctl 0x1874 bmdma 0x1860
> irq 18
> [   26.729097] ata2: SATA max UDMA/133 cmd 0x1878 ctl 0x1870 bmdma 0x1868
> irq 18
> [   26.894043] ata1.00: ATA-6: WDC WD2500SD-01KCB0, 08.02D08, max UDMA/133
> [   26.894046] ata1.00: 488397168 sectors, multi 16: LBA48
> [   26.910218] ata1.00: configured for UDMA/133
> [   57.043651] ata2.00: qc timeout (cmd 0x91)
> [   57.043656] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
> err_mask=0x4)
> [   57.043659] ata2: failed to recover some devices, retrying in 5 secs
> [   92.177684] ata2.00: qc timeout (cmd 0x91)
> [   92.177688] ata2.00: failed to IDENTIFY (INIT_DEV_PARAMS failed,
> err_mask=0x4)

Ah.. Timeout.  Hmmm... Interesting.  Maybe the device just should be
blacklisted such that INIT_DEV_PARAMS is not issued.  Can you please
just comment out the INIT_DEV_PARAMS block and see how it works and post
full boot log?

-- 
tejun