option to ignore flash memory device at USB1.1 "full" speed

Hans de Goede hdegoede at redhat.com
Tue Jun 18 06:26:54 UTC 2013


Hi,

On 06/17/2013 10:37 PM, Adam Williamson wrote:
> On Sun, 2013-06-16 at 22:33 +0100, Matthew Garrett wrote:
>> On Sun, Jun 16, 2013 at 10:11:42PM +0100, David Woodhouse wrote:
>>> On Sun, 2013-06-16 at 05:38 +0100, Matthew Garrett wrote:
>>>> On Sat, Jun 15, 2013 at 08:24:33AM -0700, John Reiser wrote:
>>>>> How can I force the system not to recognize a USB2.0 flash memory
>>>> device at USB1.1 speed?
>>>>
>>>> You can't - it's negotiated at the host controller level, the OS isn't
>>>> involved.
>>>
>>> You can't force it to use USB2 mode when for some reason it's negotiated
>>> something slower. But you can *detect* that it's connected as a USB1
>>> device and refuse to mount it, surely? And then the user will unplug it
>>> and plug it in again, until it works correctly.
>>
>> Yeah, I guess you could write a udev rule that detected that case and
>> flagged it such that it didn't get automounted.
>
> IIRC, Windows pops up one of its little yellow warnings associated with
> a notification tray icon when this happens - the medium is mounted but
> you get a warning that it's running at a slow speed. That seems
> reasonable.

And IIRC the kernel will log a message when plugging a usb-2 device into
a port which is not usb-2 capable. But if I understand correctly, that is
not the issue here?

The issue seems to be that sometimes a usb-2 device connects at usb-1 speed
even though plugged into a usb-2 port, right ?

That is just buggy hardware, and I don't think that warrants any special
handling. I would try cleaning the contacts of both the usb-port and
the usb-stick. Also if a usb-extension cable is involved, try replacing it,
or taking it out of the loop all-together.

I've seen similar issue in the past and sofar it has always been bad hardware.

Sometimes things like cleaning contacts help, sometimes the contacts on the
usb-port side are simply worn out / too loose.

Regards,

Hans


More information about the devel mailing list