Bug 8977

Created attachment 12687 [details]
kern.log

Created attachment 12688 [details]
lspci

Reply-To: akpm@linux-foundation.org

> On Mon,  3 Sep 2007 19:12:56 -0700 (PDT) bugme-daemon@bugzilla.kernel.org
> wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=8977
> 
>            Summary: after update from 2.6.20 to 2.6.21 or higher usb doesn't
>                     work
>            Product: Other
>            Version: 2.5
>      KernelVersion: 2.6.21 or higher
>           Platform: All
>         OS/Version: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: normal
>           Priority: P1
>          Component: Other
>         AssignedTo: other_other@kernel-bugs.osdl.org
>         ReportedBy: hansen12345@gmx.de
> 
> 
> hello,
> after update the kernel version from 2.6.20 to 2.6.21 or higher the usb
> connector doesn't work.
> when i start kernel 2.6.21 or 2.6.23-rc4, the loading stop for four or five
> seconds with a message: "usb can`t set config #0, error -110".
> the system starts nevertheless but not usb device is recognised.
> 
> i have build the kernel 2.6.23-rc4 with CONFIG_USB_DEBUG on for a better
> logfile.
> 
> 
> -- 
> Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug, or are watching someone who is.

It looks like a problem in your USB device, not a problem in the kernel.  Does the device still work if you boot 2.6.20?

If you want to get more information, use the usbmon facility.  Instructions are in the kernel source file Documentation/usb/usbmon.txt.  Start the usbmon monitoring before you plug in the device, and attach the output to this bug report.

(In reply to comment #4)
> It looks like a problem in your USB device, not a problem in the kernel. 
> Does
> the device still work if you boot 2.6.20?
> 

thanks for the answer,
yes, the usb works with kernel 2.6.20 and lower.


a new discovery:

when the system boot and my two usb devices are connected, not any usb device works.
when I boot without usb devices and then plug in at first the usb "dsl modem" and then the usb mouse, all devices work perfectly.
but, when I at first connect or alone the usb mouse, then no usb device works more.

Created attachment 12704 [details]
usbmo modem and mouse

Created attachment 12705 [details]
usbmon mouse alone

Created attachment 12706 [details]
usbmon dsl modem alone

The usbmon logs show clearly that the mouse doesn't work right.  The two logs in comments #6 and #7 show the mouse generating a babble response (error -75).  But then they differ: The first log shows the mouse generating another error and then it starts working, whereas the second log shows no more data from the mouse at all.

Can you attach a similar usbmon log, for just the mouse alone, under 2.6.20?

Have you tried plugging the mouse into different ports?

(In reply to comment #9)
> Have you tried plugging the mouse into different ports?
>

Yes, I have tried another usb port but without success.

Created attachment 12746 [details]
usbmon mouse 2.6.20

That is strange.  Under 2.6.20, the mouse alone behaved exactly the same way as it did when the modem was also attached under 2.6.23.  There's no obvious reason why it didn't behave the same when it was the only device attached under 2.6.23; the computer sent exactly the same data to the mouse in each case.

If you're really curious to find out the reason, you'll have to install Git and do a bisection between kernel versions 2.6.20 and 2.6.21 to find which patch caused the change in the mouse's behavior.

If you don't want to go to all that trouble then you should just get another mouse.  It's quite clear that this one doesn't work right.  For example, even under 2.6.20 the log shows that it sends an invalid response to a Get-String-Descriptor request, and its bConfigurationValue is set to 0 (which is illegal).

On the bright side, at least you know that the computer's USB connector _does_ work under 2.6.23, because the modem is okay.  Only the mouse is broken.

I installed the programm git and did the bisection.
the result of the search is shown in the attachment.

I was surprised that you were the hero who changed the relevant patches:

17230acdc71137622ca7dfd789b3944c75d39404 is first bad commit
commit 17230acdc71137622ca7dfd789b3944c75d39404
Author: Alan Stern <stern@rowland.harvard.edu>
Date:   Mon Feb 19 15:52:45 2007 -0500

My question is, how to use my "old" mouse for version 2.6.21 and higher?

Created attachment 12806 [details]
git bisection

I certainly admire the persistence and energy you have put into tracking this down!

Yes, I wrote that patch.  It did turn out to have an error, and a later patch was added to fix it:

  commit e009f1b202219c62ea7e277adbb953d703dac983
  UHCI: Fix problem caused by lack of terminating QH

Do you still see your mouse problem if you apply this commit on top of the other one?  I still can't tell exactly what the problem is, but at least now we know where to look.

By the way, have you tried borrowing another mouse to see if it works any better?  Even without the problems caused by these patches, the mouse itself doesn't behaving correctly.

Hansen, do you have any update on this, have you had chance for testing as suggested in #15?
Thanks.

Pinging the reporter again... any news from your side, have you tried Alan's recommendations in #15?

I'm afraid I have to close this bug due to inactivity of the reporter. Besides Alan has fixed the bug anyway.
Hansen, you're welcome to reopen the bug if this is still a problem for you.
thanks.