Bug 4724 - USB error causes system to become unresponsive to input
USB error causes system to become unresponsive to input
Status: REJECTED INSUFFICIENT_DATA
Product: Drivers
Classification: Unclassified
Component: USB
i386 Linux
: P2 normal
Assigned To: Vojtech Pavlik
:
Depends on:
Blocks: USB
  Show dependency treegraph
 
Reported: 2005-06-08 12:47 UTC by Bharath Ramesh
Modified: 2007-07-28 12:35 UTC (History)
5 users (show)

See Also:
Kernel Version: 2.6.x
Tree: Mainline
Regression: ---


Attachments
lsusb output (2.96 KB, text/plain)
2005-08-14 16:44 UTC, Michiel ten Hagen
Details

Description Bharath Ramesh 2005-06-08 12:47:29 UTC
I am using Debian's unofficial amd64 distribution. I am running the kernel.org
2.6.x kernels. Every once in a while my system becomes unresponsive to USB
keyboard/mice. I am using the Microsoft Wireless Desktop Keyboard and Mice. I
can still access the machine remotely. I have to reboot it either remotely or by
hitting the switch. dmesg is full of the following message.

drivers/usb/input/hid-core.c: input irq status -75 received
Comment 1 Greg Kroah-Hartman 2005-06-08 13:12:00 UTC
What specific kernel version does this happen on?

Can you try 2.6.12-rc6?  
Comment 2 Bharath Ramesh 2005-06-08 14:13:00 UTC
I have faced this problem with all the kernels that I have used till now in the
2.6 series. I am currently running the 2.6.9 kernel. I will test this out with
the 2.6.12-rc6 and post what happens.
Comment 3 Bharath Ramesh 2005-06-09 10:44:55 UTC
Tested 2.6.12-rc6 bug still exists. dmesg is still full of the message

drivers/usb/input/hid-core.c: input irq status -75 received
Comment 4 Bharath Ramesh 2005-06-22 17:13:39 UTC
I just installed the 2.6.12 kernel. I am still facing the same problems with
system becoming unresponsive to input from keyboard or mouse.

dmesg is full of the following messages.

drivers/usb/input/hid-core.c: input irq status -75 received

When I reboot I see the same messages once I load usbhid module. I need to
reboot again and then things settle down only till the next freeze.

I am using the Microsoft Desktop Pro keyboard.
Comment 5 Benjamin Grossmann Geese 2005-07-21 13:03:14 UTC
Im using Fedora Core 4 ( Kernel 2.6.12-1.1398_FC4 ) and i experience the same
problem.
Im using Microsoft Wireless Keyboad and Mice.
The Keyboard still works fine.
Sometimes unplug/ replug the connector another usb port helps,
most times i have to reboot the machine.
Error appears with almost every kernel version i tried since 2.6.8

Comment 6 Greg Kroah-Hartman 2005-08-04 13:17:36 UTC
Looks like an input issue...
Comment 7 D. Hugh Redelmeier 2005-08-08 15:05:31 UTC
I am experiencing the same problem on FC4 with their kernel 2.6.12-1.1398_FC4
(32-bit athlon).  Summary:
- constant stream of: drivers/usb/input/hid-core.c: input irq status -75 received
- keyboard / mouse is Microsoft Wireless Desktop
- only on some boots, not all
- keyboard works in this state, but mouse does not.

Unplugging, waiting 10 seconds, and replugging in the same USB socket does not
fix things.

I am not a kernel hacker, but here are some observations from the souce code.

- -75 is EOVERFLOW

- host/uhci-q.c: uhci_map_status maps TD_CTRL_BABBLE to EOVERFLOW

- Documentation/usb/error-codes.txt says:
  EOVERFLOW (*)          The amount of data returned by the endpoint was
                        greater than either the max packet size of the
                        endpoint or the remaining buffer size.  "Babble".

  (*) Error codes like -EPROTO, -EILSEQ and -EOVERFLOW normally indicate
  hardware problems such as bad devices (including firmware) or cables.

I am not convinced this is a hardware error, but it might be the case.
Comment 8 D. Hugh Redelmeier 2005-08-08 19:55:31 UTC
My system (see above) got into this state again and I did a little bit of testing.

- the messages stop when the keyboard/mouse is unplugged.

- they start up again when it is plugged in again.  Even if it is to a different
USB port.

- after a reboot (actually, power cycle) of the computer, all is well

To me, this sounds as if the kernel gets into a bad state that won't clear.  The
original error might be in the device, but its persistance seems to be in the
kernel.  Just a guess, of course.
Comment 9 Michiel ten Hagen 2005-08-14 16:44:11 UTC
Created attachment 5638 [details]
lsusb output

Ubuntu Linux 2.6.10-5-386 

This happens to me aswell, it looks like a key is stuck, unplugging the usb
devices solves it. It appears to happen mostly when pressing some of the
"multimedia buttons". This device only works if it is connected thru my usb
hub, directly on to any of the ports on the pc the keyboard doesnt function at
all.

Attached is the lsusb output.
Comment 10 Greg Kroah-Hartman 2005-08-18 16:29:37 UTC
Still present on 2.6.13-rc6?  If so, please reopen the bug with the new kernel
log information.
Comment 11 Bharath Ramesh 2006-01-25 18:26:37 UTC
It is still present in the latest 2.6.15.1 kernel. dmesg is just full of

drivers/usb/input/hid-core.c: input irq status -75 received

The mouse becomes unresponsive then. Only a power cycle fixes it.
Comment 12 Greg Kroah-Hartman 2006-03-06 10:39:47 UTC
Still an issue on 2.6.15 or 2.6.16-rc5?
Comment 13 Bharath Ramesh 2006-03-06 11:02:52 UTC
Its still a issue with the stable 2.6.15.1 kernel. I havent upgraded after that.
I  will try the latest rc5 kernel and post if its an issue.
Comment 14 Bharath Ramesh 2006-03-07 22:24:56 UTC
The latest 2.6.16-rc5 has the same problem. Rebooting from Windows to Linux
causes the problem on every switch. Power cycle fixes the issue, sometimes more
than one power cycle is needed. I havent used the rc5 kernel long enough to see
how long it takes for it break just by using Linux.
Comment 15 Justin M. Forbes 2006-03-08 15:00:25 UTC
I have seen a similar problem in 2.6.14 through 2.6.16.5.  I do not dual boot,
and use a Saitek Eclipse keyboard and Logitech Trackman Wheel... The amount of
time required to trigger this seems to vary, from less than a day to several
days. It has happened while the box is currently in use, and while the box is
idle overnight.  I have tried .configs with USB suspend enabled and disabled. 
This problem only seems present on my VIA MASTER2-FAR dual Opteron board, and I
have not seen it on any of my other systems using AMD ATI or Nvidia chipsets.
Comment 16 Bharath Ramesh 2006-03-08 19:59:05 UTC
I noticed another thing that switching from different versions of kernel also
seeming to cause the problem. I also do have the MSI K8T Neo-FIS2R motherboard
based on VIA chipset.
Comment 17 Bharath Ramesh 2006-09-21 19:03:49 UTC
Just compiled the latest 2.6.18 kernel. The problem is still present. Power
cycle helps to stop this problem. Mouse doesnt respond at all while the keyboard
works fine.
Comment 18 Adam Sulmicki 2006-10-05 07:45:17 UTC
Oct  4 13:41:02 ophir kernel: drivers/usb/input/hid-core.c: input irq status -75
received

On boot I get errors like the one above until I unplug the USB cable.
Re-plugging the USB cable helps, in that the errors usually go away.

# uname -sr
Linux 2.6.17-1.2187_FC5

# dmesg | grep input
input: Microsoft Microsoft Wireless Optical Desktop� 1.00 as /class/input/input0
input: USB HID v1.11 Keyboard [Microsoft Microsoft Wireless Optical Desktop�
1.00] on usb-0000:00:10.0-2
input: Microsoft Microsoft Wireless Optical Desktop� 1.00 as /class/input/input1
input: USB HID v1.11 Mouse [Microsoft Microsoft Wireless Optical Desktop� 1.00]
on usb-0000:00:10.0-2

# lspci | grep VIA | egrep -v "Multimedia|IDE|ISA"
00:00.0 Host bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo KT266/A/333]
00:01.0 PCI bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo KT266/A/333 AGP]
00:10.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller
(rev 50)
00:10.1 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller
(rev 50)
00:10.2 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 51)
00:11.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller
(rev 23)
00:11.3 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller
(rev 23)
Comment 19 Adam Sulmicki 2006-10-15 11:54:23 UTC
This is weird. I just installed vanilla kernel 2.6.18 and the problems seems to
have gone away. Maybe it is related to the fact that I have tried to build
minimal kernel with as little stuff enabled as possible.

Either way so far I was unable to observe this problem in my 2.6.18.

Rebooting back to Linux 2.6.17-1.2187_FC5 makes the bug come back with vengeance.
Comment 20 Bharath Ramesh 2006-10-15 16:17:09 UTC
I thought the same when I compiled the kernel. I have a dual boot, booting back
from windows made it come back immediately. For now I am running the 2.6.18
kernel After the initial breakage I havent seen it till now. The other thing
being I havent rebooted into windows yet since then.
Comment 21 Alan Stern 2007-04-29 07:56:50 UTC
If anyone is still having problems related to this bug, please try using either
2.6.20 or 2.6.21 and follow the instructions in the kernel source file
Documentation/usb/usbmon.txt.  Attach the usbmon trace you get showing what
happens when the -75 errors start to occur.
Comment 22 Adrian Bunk 2007-07-28 12:35:52 UTC
Please reopen this bug if:
- it is still present with kernel 2.6.22 and
- you are able to provide the requested information.

Note You need to log in before you can comment on or make changes to this bug.