Re: [2.6.30-rc2] usb reset during big file transfer and ext3 error

[Robert Hancock <hancockrwd@gmail.com>]

(ccing linux-usb)

Rogério Brito wrote:
> Hi there.
> 
> I have an amd64 system running Debian's sid distribution and I installed
> Linux 2.6.30-rc2 on it, as a way to get improvements for the i915
> subsystem.
> 
> Unfortunately, when I was transferring the contents of 2 DVDs from the
> main IDE HD to a USB external HD, I got errors from the USB host, the
> writes on the external HD become failures and the ext3 filesystem there
> enters into error mode, going read-only.
> 
> I eventually lose the access to the device (i.e., the /dev/sd??? device
> isn't there anymore) and I then have to re-run fsck on the given
> filesystem.
> 
> This has already happened 2 or 3 times already and I observed that it
> only occurs when there is high traffic---if I am, say, compiling the
> kernel on that external HD, I don't see any problems.
> 
> Attached is part of the dmesg log that shows the problem. I put the
> whole dmesg at <http://rb.doesntexist.org/linux/>.
> 
> As always, if any further information is needed, please let me know.

You're seeing these:

[103051.265045] ehci_hcd 0000:00:1d.7: detected XactErr len 1536/4096 
retry 1
[103051.265156] ehci_hcd 0000:00:1d.7: detected XactErr len 1536/4096 
retry 2
[103051.265281] ehci_hcd 0000:00:1d.7: detected XactErr len 1536/4096 
retry 3
[103051.265406] ehci_hcd 0000:00:1d.7: detected XactErr len 1536/4096 
retry 4

According to the EHCI spec, XactErr is "Set to a one by the Host 
Controller during status update in the case where the host did not 
receive a valid response from the device (Timeout, CRC, Bad PID, etc.)"

Quite likely this is some kind of hardware problem - maybe the USB port 
doesn't quite provide enough power for the drive, etc. A lot of these 
USB enclosure devices are also rather poor quality in general..