Re: USB Storage Device Not Working

On Wed, Nov 07, 2007 at 02:57:20PM -0500, Alan Stern wrote:
> Please reply to both the mailing and to me.
> 
> On Thu, 8 Nov 2007, balajirrao wrote:
> 
> > On Wed, Nov 07, 2007 at 02:10:17PM -0500, Alan Stern wrote:
> > > On Wed, 7 Nov 2007, balajirrao wrote:
> > > 
> > > > Hello people,
> > > > 
> > > > My usb storage device suddenly stopped being detected in Linux. It
> > > > continued to work properly in Windows and Solaris. While trying to find
> > > > out more about the problem, i tried rmmod ehci_hcd, and then the device
> > > > started working properly in Full Speed (it is actually a high speed
> > > > device). And, in windows, i get a warning saying that the device is
> > > > plugged into a non high speed port.
> > > 
> > > I don't get it.  You say the device works properly under Windows, but
> > > you also say that Windows displays a warning message about a problem
> > > with the device.  These statements can't both be true.
> > > 
> > It works properly in the sense, it works under OHCI (Full Speed) but.
> > Windows only complains that a high speed usb device is plugged into a
> > non high speed usb port, and that it can function faster if plugged into
> > a high speed port. This status message is not true, as i have been using
> > the same port and the same device for a long time.
> 
> You mean the port is connected to both an EHCI and an OHCI controller, 
> but Windows assigns the device to the OHCI controller and then 
> complains that it isn't running at high speed, right?
> 
> > > Have you tried using the device on a different computer?  Does it run 
> > > okay at high speed on the other machine?
> > > 
> > > Do other high speed devices work okay on your computer?
> > Yes. Absolutely.
> 
> There are three questions above.  Which question does your "Yes. 
> Absolutely." answer, and what are the answers to the other two 
> questions?
> 
> > > > I run kernel version 2.6.24-rc2. I get the 3strikes error quite number
> > > > of times. I tried a lot to debug this by going through the sources, but
> > > > i was unsuccessful.
> > > > 
> > > > I think this is case of hardware malfunction which the kernel does not
> > > > handle. I guess the device needed to be handed over to the companion
> > > > controllers on detection of this condition.
> > > 
> > > It does sound like a hardware malfunction, but from your description 
> > > there's no way to tell whether the malfunction is in the device or in 
> > > your computer's EHCI controller.
> > Right. But i have tried other high speed devices on my machine and it
> > works fine.
> 
> Which indicates that your EHCI controller is okay and hence the device 
> must be broken.
> 
> > > Why do you think the kernel detected a hardware malfunction?  The log 
> > > messages you included don't indicate any such detection.
> > > 
> > Oh! is it ? Does'nt the 3strikes message indicate that ? I saw in the
> > kernel sources that it indicates a status of -EPROTO. Does it not
> > indicate a hardware malfunction ? Also, in ehci-q.c, a comment above the
> > printing code talks about 'serious "cant proceed faults" reported by
> > hardware"'. Hence i guessed that it was a hardware malfunction. CMIIW.
> 
> It's a hardware fault (i.e., unexpected result), but not necessarily a
> malfunction.  For example, it could simply indicate that there's some
> temporary noise in the USB data signals.  It doesn't always mean that
> the device is broken.  Retrying the request will sometimes work.
> 
> > > > Please help me fix this bug. Thank you in advance for any help. Please
> > > > CC me, as i am not subscribed to the list.
> > > 
> > > Are you sure this is a fixable bug?  We can't do anything to repair
> > > hardware failures.
> > > 
> > No. I am sure its a fixable bug.
> > * Other high speed devices work in my computer.
> > * The device does not get detected in Linux in all the computers i have
> >  tried.
> > * On all the machines, the hardware was usable in windows but only at
> >  full speed.
> 
> Ah, so these are the answers to my questions above.  They indicate that 
> the device really is broken, since it no longer runs at high speed on 
> any machine.  I don't see how we can fix your device.
> 
No! It's not what i wanted from linux-usb-devel either :)
> Right now the kernel doesn't have any means for automatically forcing
> devices to run at full speed when they fail to enumerate at high speed,
> although adding such a feature might not be a bad idea.  However you
> can force a particular USB port to run at full speed by hand.  The
> command is:
> 
> 	echo P >/sys/class/usb_host/usb_hostN/companion
> 
> where P is the port number and N is the bus number of the EHCI 
> controller.
> 
Thank you very much for the command. It worked fine.
And, about that feature - I am interested in developing it. Can you
please help me as to how to go about it ? 

Thank you.
> Alan Stern
> 
regards,

Balaji Rao

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel