Re: usb device problem

[Mark Lord]

Jiri Slaby wrote:
> Mark Lord napsal(a):
> > Jiri Slaby wrote:
> > > Hello,
> > >
> > > I get this with 2.6.17-rc5-mm3 kernel:
> > ..
> > > usb-storage: device found at 10
> > > usb-storage: waiting for device to settle before scanning
> > >   Vendor:           Model:                   Rev:
> > >   Type:   Direct-Access                      ANSI SCSI revision: 00
> > > SCSI device sdb: 245920 512-byte hdwr sectors (126 MB)
> > ..
> > > now read and write and sync or umount, then:
> > > ---
> > > sd 10:0:0:0: SCSI error: return code = 0x10070000
> > > end_request: I/O error, dev sdb, sector 1575
> > > sd 10:0:0:0: SCSI error: return code = 0x10070000
> > > end_request: I/O error, dev sdb, sector 1583
> > > sd 10:0:0:0: SCSI error: return code = 0x10070000
> > > end_request: I/O error, dev sdb, sector 1591
> > > sd 10:0:0:0: SCSI error: return code = 0x10070000
> > > end_request: I/O error, dev sdb, sector 1599
> > > sd 10:0:0:0: SCSI error: return code = 0x10070000
> > > end_request: I/O error, dev sdb, sector 1607
> > > sd 10:0:0:0: SCSI error: return code = 0x10070000
> > > end_request: I/O error, dev sdb, sector 1615
> > > ... and so on. data are maybe there, but it takes so long to write a
> > > meg file.
> > > sometimes
> > ..
> >
> > This *looks* like maybe the drive reported a sector read error,
> > and the standard "fail the whole request one block at a time"
> > error mechanism from sd.c has kicked in.
>
> Do you mean something like seek error, i.e. error in hardware, or how to call
> this? This is brand new minisd card, it is possible to be waster, but it's
> rather something bad in the software (writing by the device itself is perfomed
> and data are ok). The error occurs accurately every 8 sectors...

The "every 8 sectors" corresponds to the Linux page size (32-bit) of 4KBytes,
which is the basic block I/O unit in real life.

The 0x1007000 is broken down as:

07 == "host byte"   = DID_ERROR = "internal error"
10 == "driver byte" = SUGGEST_RETRY

So it could just be some kind of internal soft error within the device driver.
The messages certainly lack end-user clarity, though.

> > I have a patch to fix this behaviour (in sd.c), but it has not yet
> > been decided whether to go upstream with it or not.
>
> Could you post me a copy, please?

Probably tomorrow.  I haven't ported it forward yet (from a much older kernel).
But I don't think it will help here now, as these errors
don't really look like bad media -- gotta look inside the usb-storage code to find out.

Cheers

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/