Mailing List Archive: Kernel-3.8 USB Disconnect

Kernel-3.8 USB Disconnect

Mar 8, 2013, 12:27 PM

Post #1 of 6 (4165 views)

Hello,

After compiling kernel-3.8 (plain vanilla) I begin to see lots of these
entries in the kernel log:

usb 1-1.3: USB disconnect, device number 4
usb 1-1.3: new low-speed USB device number 5 using ehci-pci
input: Microsoft Microsoft Basic Optical Mouse v2.0 as /devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.3/1-1.3:1.0/input/input2

These entries are repeated, with the device number increasing (+1)
each time.

The odd thing is that these USB disconnects occur only in console mode.
My system boots up to the console and the messages appear roughly every
60 seconds. But when I start X, the messages suddenly stop. There may
be only one or two USB disconnects over the next 12-16 hours. The only
difference between console mode and X is that the mouse is only active
under X.

The system remains stable, but for some reason, the USB mouse will repeatedly
disconnect in console mode, but also rarely in X.

I have searched for references to the problem but could only find a
possible similar issue with 3.8-rc1

https://lkml.org/lkml/2012/12/23/100

Has anyone else seen this "USB disconnect" problem with kernel-3.8.x?

It could be that my motherboard is failing (it's approx. 2 years old) but
then why would the USB disconnects suddenly halt under X.

USB is built in to the kernel; there are no USB modules loaded.

Any comments on how I could investigate this further would be appreciated.

Frank Peters

Re: Kernel-3.8 USB Disconnect [ In reply to ]

paul.hartman+gentoo at gmail

Mar 8, 2013, 12:34 PM

Post #2 of 6 (4061 views)

Permalink

On Fri, Mar 8, 2013 at 2:27 PM, Frank Peters <frank.peters@comcast.net> wrote:
> Hello,
>
> After compiling kernel-3.8 (plain vanilla) I begin to see lots of these
> entries in the kernel log:
>
> usb 1-1.3: USB disconnect, device number 4
> usb 1-1.3: new low-speed USB device number 5 using ehci-pci
> input: Microsoft Microsoft Basic Optical Mouse v2.0 as /devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.3/1-1.3:1.0/input/input2
>
> These entries are repeated, with the device number increasing (+1)
> each time.
>
> The odd thing is that these USB disconnects occur only in console mode.
> My system boots up to the console and the messages appear roughly every
> 60 seconds. But when I start X, the messages suddenly stop. There may
> be only one or two USB disconnects over the next 12-16 hours. The only
> difference between console mode and X is that the mouse is only active
> under X.
>
> The system remains stable, but for some reason, the USB mouse will repeatedly
> disconnect in console mode, but also rarely in X.
>
> I have searched for references to the problem but could only find a
> possible similar issue with 3.8-rc1
>
> https://lkml.org/lkml/2012/12/23/100
>
> Has anyone else seen this "USB disconnect" problem with kernel-3.8.x?
>
> It could be that my motherboard is failing (it's approx. 2 years old) but
> then why would the USB disconnects suddenly halt under X.
>
> USB is built in to the kernel; there are no USB modules loaded.
>
> Any comments on how I could investigate this further would be appreciated.

I do not know the cause of your problem, but if it's a machine you can
afford to reboot a few times, I would try doing a git bisect between
that kernel and the previous working one, and e-mailing your results
to the person whose name is on the offending commit.

Re: Kernel-3.8 USB Disconnect [ In reply to ]

rich0 at gentoo

Mar 8, 2013, 12:42 PM

Post #3 of 6 (4060 views)

Permalink

On Fri, Mar 8, 2013 at 3:34 PM, Paul Hartman
<paul.hartman+gentoo@gmail.com> wrote:
> I do not know the cause of your problem, but if it's a machine you can
> afford to reboot a few times, I would try doing a git bisect between
> that kernel and the previous working one, and e-mailing your results
> to the person whose name is on the offending commit.

Always a good idea. It doesn't always get you to the root cause
though - I had a QT/xcb/xserver/mesa issue that I bisected but it
seems like the change I found only fixed a bug that apparently was
forcing xserver into software rendering mode which masked the error.

As far as the bug goes - could it be a result of the hardware being
told to go to sleep due to non-use?

Rich

Re: Kernel-3.8 USB Disconnect [ In reply to ]

frank.peters at comcast

Mar 8, 2013, 3:22 PM

Post #4 of 6 (4042 views)

Permalink

On Fri, 8 Mar 2013 15:42:18 -0500
Rich Freeman <rich0@gentoo.org> wrote:

>
> As far as the bug goes - could it be a result of the hardware being
> told to go to sleep due to non-use?
>

Thanks for the replies.

I booted with Gentoo LiveDVD and did not see the "USB disconnect" messages,
and I also noticed that Gentoo starts the gpm mouse at boot up.

I then added the gpm start to my init scripts and when I boot my system
the problem disappears. I now do not see the "USB disconnect" messages
in console mode.

Here is the start command I added:

/usr/sbin/gpm -m /dev/input/mice -t imps2

Also, the problem seems not limited to kernel-3.8.x. I always keep the
previous kernel available and the same problem happens when booting with
kernel-3.7. If I can find the time I may try to compile some earlier kernels
and test those.

But running gpm from the boot script seems to prevent whatever event was
producing those "USB disconnect" messages. If the mouse is not "grabbed"
by something it somehow disconnects.

Frank Peters

Re: Kernel-3.8 USB Disconnect [ In reply to ]

philip at turmel

Mar 9, 2013, 8:24 AM

Post #5 of 6 (4058 views)

Permalink

On 03/08/2013 06:22 PM, Frank Peters wrote:

> But running gpm from the boot script seems to prevent whatever event was
> producing those "USB disconnect" messages. If the mouse is not "grabbed"
> by something it somehow disconnects.

Hi Frank,

There's been on-going work on usb auto powersave for devices that
support it, and some reports like yours have popped up (I lurk on lkml).
I suggest you search the lkml archives with those keywords, with
particular attention to traffic from Alan Stern and Rafael Wysocki.

HTH,

Phil

Re: Kernel-3.8 USB Disconnect [ In reply to ]

frank.peters at comcast

Mar 11, 2013, 9:27 AM

Post #6 of 6 (4046 views)

Permalink

On Sat, 09 Mar 2013 11:24:49 -0500
Phil Turmel <philip@turmel.org> wrote:

>
> There's been on-going work on usb auto powersave for devices that
> support it, and some reports like yours have popped up
>

Thanks a lot for this info. It does make sense.

I can reproduce the problem at will. In console mode if I start
gpm the disconnects will stop. If I then stop gpm, the disconnects
start up again.

I'm reporting this behavior to the LKML. It should not be happening.

Frank Peters