Re: External USB3 disk fails with "Invalid field in cdb"

[Alan Stern]

On Fri, 27 Jun 2014, Michael Büsch wrote:

> On Fri, 27 Jun 2014 08:48:01 -0700
> James Bottomley <James.B...@HansenPartnership.com> wrote:
>
> > On Fri, 2014-06-27 at 17:34 +0200, Michael Büsch wrote:
> > > I tried the following patch:
> > >
> > > Index: linux/drivers/scsi/sd.c
> > > ===================================================================
> > > --- linux.orig/drivers/scsi/sd.c 2014-06-26 18:40:39.214696552 +0200
> > > +++ linux/drivers/scsi/sd.c 2014-06-27 15:52:30.776159456 +0200
> > > @@ -2440,7 +2440,7 @@ sd_read_cache_type(struct scsi_disk *sdk
> > > sdkp->RCD = 0;
> > > }
> > >
> > > - sdkp->DPOFUA = (data.device_specific & 0x10) != 0;
> > > + sdkp->DPOFUA = 0;
> > > if (sdkp->DPOFUA && !sdkp->device->use_10_for_rw) {
> > > sd_first_printk(KERN_NOTICE, sdkp,
> > > "Uses READ/WRITE(6), disabling FUA\n");
> > >
> > >
> > > This obviously is not the correct thing to do, but it makes the disk usable:
> > >
> > > Jun 27 17:26:50 marge kernel: [ 523.909815] usb 2-1: new SuperSpeed USB device number 2 using xhci_hcd
> > > Jun 27 17:26:50 marge kernel: [ 523.929246] usb 2-1: New USB device found, idVendor=152d, idProduct=0567
> > > Jun 27 17:26:50 marge kernel: [ 523.929258] usb 2-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
> > > Jun 27 17:26:50 marge kernel: [ 523.929265] usb 2-1: Product: USB to ATA/ATAPI Bridge
> > > Jun 27 17:26:50 marge kernel: [ 523.929271] usb 2-1: Manufacturer: JMicron
> > > Jun 27 17:26:50 marge kernel: [ 523.929276] usb 2-1: SerialNumber: xxx
> > > Jun 27 17:26:50 marge kernel: [ 523.930999] usb-storage 2-1:1.0: USB Mass Storage device detected
> > > Jun 27 17:26:50 marge kernel: [ 523.931237] scsi12 : usb-storage 2-1:1.0
> > > Jun 27 17:26:51 marge kernel: [ 524.930451] scsi 12:0:0:0: Direct-Access JMicron Generic 0114 PQ: 0 ANSI: 6
> > > Jun 27 17:26:51 marge kernel: [ 524.931228] sd 12:0:0:0: Attached scsi generic sg3 type 0
> > > Jun 27 17:26:51 marge kernel: [ 524.932964] sd 12:0:0:0: [sdd] Spinning up disk...
> > > Jun 27 17:26:53 marge kernel: [ 525.937848] ..ready
> > > Jun 27 17:26:53 marge kernel: [ 526.942395] sd 12:0:0:0: [sdd] 976773168 512-byte logical blocks: (500 GB/465 GiB)
> > > Jun 27 17:26:53 marge kernel: [ 526.943037] sd 12:0:0:0: [sdd] Write Protect is off
> > > Jun 27 17:26:53 marge kernel: [ 526.943048] sd 12:0:0:0: [sdd] Mode Sense: 47 00 10 08
> > > Jun 27 17:26:53 marge kernel: [ 526.943662] sd 12:0:0:0: [sdd] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
> > > Jun 27 17:26:53 marge kernel: [ 526.987919] sdd: sdd1
> > > Jun 27 17:26:53 marge kernel: [ 526.990055] sd 12:0:0:0: [sdd] Attached SCSI disk
> > > Jun 27 17:27:29 marge kernel: [ 563.227849] EXT4-fs (sdd1): warning: mounting fs with errors, running e2fsck is recommended
> > > Jun 27 17:27:29 marge kernel: [ 563.228043] EXT4-fs (sdd1): mounted filesystem with ordered data mode. Opts: (null)
> > >
> > > (This is on another machine, but it shows the same behavior without the patch.)
> > >
> > > Does somebody have a hint for a real fix?
> >
> > I suspect the problem is in the USB bridge: The device reports FUA
> > support in its initial IDENTIFY, which gets translated to the correct
> > mode page bits, but when we send a READ/WRITE with the FUA bit set, the
> > bridge doesn't know how to translate it and gives us the error. There's
> > probably some way in USB to blacklist the bridge, but we need to
> > understand the consequences of the blacklist: The device is claiming a
> > writeback cache which means if it won't do either FUA or SYNC CACHE,
> > there's going to be a data integrity failure.
> >
> > The fact that your initial hack works suggests that sync cache is
> > accepted by it, so we probably just need a USB blacklist fixing the FUA
> > problem.

Michael, can you post the "lsusb -v" output for this device? I see it
is made by JMicron; they are notorious for buggy USB-ATA bridges.

Alan Stern

[Michael Büsch]

On Fri, 27 Jun 2014 14:42:01 -0400 (EDT)
Alan Stern <st...@rowland.harvard.edu> wrote:

> Michael, can you post the "lsusb -v" output for this device? I see it
> is made by JMicron; they are notorious for buggy USB-ATA bridges.

Of course. Here you go:

Bus 004 Device 009: ID 152d:0567 JMicron Technology Corp. / JMicron USA Technology Corp.
Device Descriptor:
bLength 18
bDescriptorType 1
bcdUSB 3.00
bDeviceClass 0 (Defined at Interface level)
bDeviceSubClass 0
bDeviceProtocol 0
bMaxPacketSize0 9
idVendor 0x152d JMicron Technology Corp. / JMicron USA Technology Corp.
idProduct 0x0567
bcdDevice 1.14
iManufacturer 1 JMicron
iProduct 2 USB to ATA/ATAPI Bridge
iSerial 3 xxx
bNumConfigurations 1
Configuration Descriptor:
bLength 9
bDescriptorType 2
wTotalLength 121
bNumInterfaces 1
bConfigurationValue 1
iConfiguration 4 USB Mass Storage
bmAttributes 0xc0
Self Powered
MaxPower 2mA
Interface Descriptor:
bLength 9
bDescriptorType 4
bInterfaceNumber 0
bAlternateSetting 0
bNumEndpoints 2
bInterfaceClass 8 Mass Storage
bInterfaceSubClass 6 SCSI
bInterfaceProtocol 80 Bulk-Only
iInterface 6 MSC Bulk-Only Transport
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x81 EP 1 IN
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0400 1x 1024 bytes
bInterval 0
bMaxBurst 15
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x02 EP 2 OUT
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0400 1x 1024 bytes
bInterval 0
bMaxBurst 15
Interface Descriptor:
bLength 9
bDescriptorType 4
bInterfaceNumber 0
bAlternateSetting 1
bNumEndpoints 4
bInterfaceClass 8 Mass Storage
bInterfaceSubClass 6 SCSI
bInterfaceProtocol 98
iInterface 10 MSC USB Attached SCSI
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x01 EP 1 OUT
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0400 1x 1024 bytes
bInterval 0
bMaxBurst 0
Command pipe (0x01)
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x82 EP 2 IN
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0400 1x 1024 bytes
bInterval 0
bMaxBurst 0
MaxStreams 16
Status pipe (0x02)
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x83 EP 3 IN
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0400 1x 1024 bytes
bInterval 0
bMaxBurst 15
MaxStreams 16
Data-in pipe (0x03)
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x04 EP 4 OUT
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0400 1x 1024 bytes
bInterval 0
bMaxBurst 15
MaxStreams 16
Data-out pipe (0x04)
Binary Object Store Descriptor:
bLength 5
bDescriptorType 15
wTotalLength 22
bNumDeviceCaps 2
USB 2.0 Extension Device Capability:
bLength 7
bDescriptorType 16
bDevCapabilityType 2
bmAttributes 0x00000002
Link Power Management (LPM) Supported
SuperSpeed USB Device Capability:
bLength 10
bDescriptorType 16
bDevCapabilityType 3
bmAttributes 0x00
wSpeedsSupported 0x000e
Device can operate at Full Speed (12Mbps)
Device can operate at High Speed (480Mbps)
Device can operate at SuperSpeed (5Gbps)
bFunctionalitySupport 1
Lowest fully-functional device speed is Full Speed (12Mbps)
bU1DevExitLat 10 micro seconds
bU2DevExitLat 32 micro seconds
Device Status: 0x000d
Self Powered
U1 Enabled
U2 Enabled



MaxPower=2mA is a nice guess for a hard disk. ;)

--
Michael

[Michael Büsch]

[Added USB lists to Cc]


--
Michael

[Alan Stern]

On Fri, 27 Jun 2014, Michael Büsch wrote:

...

> MaxPower=2mA is a nice guess for a hard disk. ;)

That refers to the amount of power the device draws from the USB bus.
Since the disk drive is self-powered, it doesn't use much bus power.

Does the patch below do what you and James want?

Alan Stern



Index: usb-3.16/include/linux/usb_usual.h
===================================================================
--- usb-3.16.orig/include/linux/usb_usual.h
+++ usb-3.16/include/linux/usb_usual.h
@@ -70,7 +70,9 @@
US_FLAG(NEEDS_CAP16, 0x00400000) \
/* cannot handle READ_CAPACITY_10 */ \
US_FLAG(IGNORE_UAS, 0x00800000) \
- /* Device advertises UAS but it is broken */
+ /* Device advertises UAS but it is broken */ \
+ US_FLAG(BROKEN_FUA, 0x01000000) \
+ /* Cannot handle FUA in WRITE CDBs */ \

#define US_FLAG(name, value) US_FL_##name = value ,
enum { US_DO_ALL_FLAGS };
Index: usb-3.16/include/scsi/scsi_device.h
===================================================================
--- usb-3.16.orig/include/scsi/scsi_device.h
+++ usb-3.16/include/scsi/scsi_device.h
@@ -173,6 +173,7 @@ struct scsi_device {
unsigned is_visible:1; /* is the device visible in sysfs */
unsigned wce_default_on:1; /* Cache is ON by default */
unsigned no_dif:1; /* T10 PI (DIF) should be disabled */
+ unsigned broken_fua:1; /* Don't set FUA bit */

atomic_t disk_events_disable_depth; /* disable depth for disk events */

Index: usb-3.16/drivers/usb/storage/unusual_devs.h
===================================================================
--- usb-3.16.orig/drivers/usb/storage/unusual_devs.h
+++ usb-3.16/drivers/usb/storage/unusual_devs.h
@@ -1936,6 +1936,13 @@ UNUSUAL_DEV( 0x14cd, 0x6600, 0x0201, 0x
USB_SC_DEVICE, USB_PR_DEVICE, NULL,
US_FL_IGNORE_RESIDUE ),

+/* Reported by Michael Büsch <m...@bues.ch> */
+UNUSUAL_DEV( 0x152d, 0x0567, 0x0114, 0x0114,
+ "JMicron",
+ "USB to ATA/ATAPI Bridge",
+ USB_SC_DEVICE, USB_PR_DEVICE, NULL,
+ US_FL_BROKEN_FUA ),
+
/* Reported by Alexandre Oliva <ol...@lsd.ic.unicamp.br>
* JMicron responds to USN and several other SCSI ioctls with a
* residue that causes subsequent I/O requests to fail. */
Index: usb-3.16/drivers/usb/storage/scsiglue.c
===================================================================
--- usb-3.16.orig/drivers/usb/storage/scsiglue.c
+++ usb-3.16/drivers/usb/storage/scsiglue.c
@@ -256,6 +256,10 @@ static int slave_configure(struct scsi_d
if (us->fflags & US_FL_WRITE_CACHE)
sdev->wce_default_on = 1;

+ /* A few buggy USB-ATA bridges don't understand FUA */
+ if (us->fflags & US_FL_BROKEN_FUA)
+ sdev->broken_fua = 1;
+
} else {

/* Non-disk-type devices don't need to blacklist any pages
Index: usb-3.16/drivers/scsi/sd.c
===================================================================
--- usb-3.16.orig/drivers/scsi/sd.c
+++ usb-3.16/drivers/scsi/sd.c
@@ -2441,7 +2441,10 @@ sd_read_cache_type(struct scsi_disk *sdk

}

sdkp->DPOFUA = (data.device_specific & 0x10) != 0;

- if (sdkp->DPOFUA && !sdkp->device->use_10_for_rw) {
+ if (sdp->broken_fua) {
+ sd_first_printk(KERN_NOTICE, sdkp, "Disabling FUA\n");

+ sdkp->DPOFUA = 0;

+ } else if (sdkp->DPOFUA && !sdkp->device->use_10_for_rw) {

sd_first_printk(KERN_NOTICE, sdkp,
"Uses READ/WRITE(6), disabling FUA\n");

sdkp->DPOFUA = 0;

[Michael Büsch]

On Fri, 27 Jun 2014 15:23:42 -0400 (EDT)
Alan Stern <st...@rowland.harvard.edu> wrote:

> > MaxPower=2mA is a nice guess for a hard disk. ;)
>
> That refers to the amount of power the device draws from the USB bus.
> Since the disk drive is self-powered, it doesn't use much bus power.

Unless the device uses free energy or some other perpetual motion power source,
it seems that it actually is bus-powered.
So the self-powered from the descriptor is a lie, too.

> Does the patch below do what you and James want?

Yes, this does work. Thanks!

You can add my
Tested-by: Michael Büsch <m...@bues.ch>

--
Michael

[Michael Büsch]

Does somebody pick this up for inclusion?
It'd also be good, if this hit stable.


--
Michael

[James Bottomley]

Yes, that's the usual annoying additions to our blacklist. You can add
my acked-by and could you cc stable?

Thanks,

James

[Alan Stern]

On Mon, 30 Jun 2014, James Bottomley wrote:

> > Does the patch below do what you and James want?
>
> Yes, that's the usual annoying additions to our blacklist. You can add
> my acked-by and could you cc stable?

Will do.

Alan Stern

[Alan Stern]

Some buggy JMicron USB-ATA bridges don't know how to translate the FUA
bit in READs or WRITEs. This patch adds an entry in unusual_devs.h
and a blacklist flag to tell the sd driver not to use FUA.

Signed-off-by: Alan Stern <st...@rowland.harvard.edu>
Reported-by: Michael Büsch <m...@bues.ch>
Tested-by: Michael Büsch <m...@bues.ch>
Acked-by: James Bottomley <James.B...@HansenPartnership.com>
CC: Matthew Dharm <mdhar...@one-eyed-alien.net>
CC: <sta...@vger.kernel.org>

---


[as1751]


drivers/scsi/sd.c | 5 ++++-
drivers/usb/storage/scsiglue.c | 4 ++++
drivers/usb/storage/unusual_devs.h | 7 +++++++
include/linux/usb_usual.h | 4 +++-
include/scsi/scsi_device.h | 1 +
5 files changed, 19 insertions(+), 2 deletions(-)

Index: usb-3.16/include/linux/usb_usual.h
===================================================================
--- usb-3.16.orig/include/linux/usb_usual.h
+++ usb-3.16/include/linux/usb_usual.h
@@ -70,7 +70,9 @@
US_FLAG(NEEDS_CAP16, 0x00400000) \
/* cannot handle READ_CAPACITY_10 */ \
US_FLAG(IGNORE_UAS, 0x00800000) \
- /* Device advertises UAS but it is broken */
+ /* Device advertises UAS but it is broken */ \
+ US_FLAG(BROKEN_FUA, 0x01000000) \

+ /* Cannot handle FUA in WRITE or READ CDBs */ \

[Alan Stern]

On Mon, 30 Jun 2014, Alan Stern wrote:

> Some buggy JMicron USB-ATA bridges don't know how to translate the FUA
> bit in READs or WRITEs. This patch adds an entry in unusual_devs.h
> and a blacklist flag to tell the sd driver not to use FUA.
>
> Signed-off-by: Alan Stern <st...@rowland.harvard.edu>
> Reported-by: Michael Büsch <m...@bues.ch>
> Tested-by: Michael Büsch <m...@bues.ch>
> Acked-by: James Bottomley <James.B...@HansenPartnership.com>
> CC: Matthew Dharm <mdhar...@one-eyed-alien.net>
> CC: <sta...@vger.kernel.org>

Greg:

I forgot to add your name to the recipient list on this patch. Since
James has already Ack'ed it, would you like to take it for your USB
tree?

Alan Stern

[Greg KH]

Sure, I can take it, thanks.

greg k-h