Over-Current on Port..... USB Error

kiyiya · 02-10-2007, 07:43 AM

I have a Dell Inspiron 8500 on which I installed Fedora Core 6 (2.6.18-1.2798.fc6). The installation went without a hitch and everything appears to be working except USB. I get the following error:

hub 2-0:1.0: over-current change on port 1
hub 4-0:1.0: over-current change on port 3
This just goes and on in the logs and is the output of dmesg.

lspci output for USB is:

00:1d.0 USB Controller: Intel Corporation 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #1 (rev 03)
00:1d.1 USB Controller: Intel Corporation 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #2 (rev 03)
00:1d.2 USB Controller: Intel Corporation 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #3 (rev 03)

My hdd is constantly writing to the log file with this error and the system seems to run slow. I must point out that this error (or a similar one) was present in Windows XP as well but I was able to disable the offending ports and go on without issue.

I know this type of error usually indicates a short or something else wrong with the hardware but I have completely taken this laptop apart and see nothing "obviously" wrong like a piece of metal touching the motherboard or anything.

I guess my question is this: how can I disable the offending ports so Linux does not want to use them?

I appreciate your help.

Kiyiya
-

macemoneta · 02-11-2007, 04:18 PM

Apply maintenance:

yum -y update

Reboot to use the newly installed kernel. Do you still get the messages?

macemoneta · 02-11-2007, 04:34 PM

If the update doesn't resolve the problem, you are likely seeing a hardware problem. You can look at this bug report for details. You can suppress the messages with the workarounds in this bug report, but the situation is not ideal.

kiyiya · 02-13-2007, 12:37 AM

Thanks for your reply. The updates (all 268 of them) did not work. The workaround I am still working with but I was hoping there was some way to "turn off" the offending hub like I have been able to do in Windows. I figured I had a hardware problem but I am not just quite sure how it went bad. Any other suggestions?

Kiyiya

Electro · 02-13-2007, 01:01 AM

Probably a paper clip, stapler, graphite fillings, or anything that is metal is shorting out USB. A good cleaning with compress air should fix the problem unless your notebook contains a defective component.

kiyiya · 02-13-2007, 06:24 AM

I did a good cleaning on it with compressed air right after I got these messages. I took the entire laptop apart and cleaned the motherboard, inspected for any sign of metal contact and/or damage but could not find anything. This port used to work fine and all of a sudden it started to give this same type of message in Windows after an update I performed. I think I may have a bad controller,however, the other port seems to work fine. At least in Windows I could stop the annoying messages.

Kiyiya

macemoneta · 02-13-2007, 09:35 AM

The workarounds in the bug reports didn't work? You tried blacklisting uhci-hcd and ohci-hcd?

kiyiya · 02-13-2007, 05:19 PM

The workarounds listed below are not available on my system. I do not have a hotplug directory (do I need to create it?) and the KLOGD daemon is not where it is supposed to be. I would rather not turn off logging of the kernel and hope the first solution is achieveable.

How do I know if Hotplug is being used?

From the bug report:

This can be circumvented (if using hotplug) by adding uhci_hcd to
/etc/hotplug/blacklist. Of course, if the module is loaded, problem will appears
again.

It is possible to filter the messages by adding switch "-c 2" to kernel logging
daemon. Just edit /etc/init.d/klogd:

Change

KLOGD=""

to

KLOGD="-c 2"

Unfortunately this filters also many other useful kernel messages.

macemoneta · 02-13-2007, 05:32 PM

If you are not using hotplug, you can just:

/sbin/modprobe -r uhci_hcd
/sbin/modprobe -r ohci_hcd

If this works, add the commands to your /etc/rc.local (or the equivalent on your distribution).

kiyiya · 02-13-2007, 11:09 PM

I ran modprobe and was able to stop the constant thrashing of my HDD and I modified the rc.local file so hopefully it will survive a reboot. However, when I checked the log it was only able to stop the messages for the port 1 error not the port 3.

Here is a sample of the log file:

Feb 13 22:03:22 kiyiya kernel: hub 4-0:1.0: over-current change on port 3
Feb 13 22:03:53 kiyiya last message repeated 121 times
Feb 13 22:04:54 kiyiya last message repeated 244 times
Feb 13 22:05:55 kiyiya last message repeated 244 times
Feb 13 22:06:56 kiyiya last message repeated 244 times
Feb 13 22:07:57 kiyiya last message repeated 244 times
Feb 13 22:08:58 kiyiya last message repeated 244 times
Feb 13 22:09:59 kiyiya last message repeated 245 times
(sample)

It now appears to write the message every minute instead of multiple times a second. This is definitely a step in the right direction.

kiyiya · 02-16-2007, 11:37 PM

Running the commands also took out my CD Drive.

Electro · 02-17-2007, 03:51 PM

Try specifying nousb at the bootloader.

Disabling USB is only a software workaround. If your notebook computer is still under warranty, it is best to see if Dell can replace it. If it is not, get a new notebook.

UhhMaybe · 02-17-2007, 07:57 PM

IF it is possible, go to BIOS and disable USB. Do you need both USB and USB legacy to run your laptop? This at least offers you two more options.

kiyiya · 02-19-2007, 10:08 PM

This notebook is a little over 3 years old so the warranty is out. I can not find out what the problems are with the ports but Windows at least has functionality for one of them. I know that this would be a "workaround" but it beats giving up on the machine entirely. This thing has been a real work horse! I guess I could go back to Windows where I can manage the bad port and still use the other one and my CD/DVD combo drive. I just thought there was a way in Linux to have more manageability of the hardware.

Thanks for your responses.