Unlink after no-IRQ

[Daniel Ashbrook]

I just installed Breezy on a new machine, and I'm getting these messages when I
insert a USB2 memory stick:

usb 3-1: new high speed USB device using ehci_hcd and address 4
ehci_hcd 0000:00:0c.2: Unlink after no-IRQ? Controller is probably using the
wrong IRQ.
usb 3-1: device not accepting address 4, error -110
usb 3-1: new high speed USB device using ehci_hcd and address 5
usb 3-1: device not accepting address 5, error -110
usb 3-1: new high speed USB device using ehci_hcd and address 6
usb 3-1: device not accepting address 6, error -110

...and so on. I saw a suggestion that passing "noapic" to the kernel would fix
this, but no luck. Passing "acpi=off" fixes this, but it's not really a good
solution.

[Ben Collins (ben-collins)]

Doubtful that this will get fixed in breezy. Could you try dapper's 2.6.15 kernel?

[Dennis Kaarsemaker (dennis)]

acpi=routeirq may help too, I've seen a similar problem on the dutch forums
where acpi=routeirq was the solution.

[Carthik Sharma (carthik)]

Daniel,
Thanks for reporting this bug. Can you please confirm whether this bug persists with the latest Dapper packages. You could try reinstalling using the Dapper Flight 6 CD, or do a dist-upgrade to dapper.

Thanks again.

[Gilles Gravier (gillesgravier)]

Same problem with a USB attached IDE hard disk. And acpi=off didn't fix it.

Machine is SHUTTLE XPC ST20G5 with nVidia nForce chipset and AMD CPU. Running in 32 bit mode.

[rak (rak-rkerbitz)]

I have the same Problem with a USB attached IDE harddisk and I am using a fresh installed and up to date Dapper 6.06 LTS with k7-kernel. It seems to me that the problem came with one of the newer kernel updates, because before I reinstalled my computer I could switch back to an older kernel and the drive worked fine. But I don't know which the newest kernel image it was, that worked with the drive.

[TRiSS (triss)]

I have the same problem, the ehci_hcd still worked in kernel 2.6.15-25-386, but since 2.6.15-26-386 neither the -386 or -k7 kernel works:

[17179892.948000] usb 4-2: new high speed USB device using ehci_hcd and address 5
[17179893.948000] ehci_hcd 0000:00:10.3: Unlink after no-IRQ? Controller is probably using the wrong IRQ.
[17179904.492000] usb 4-2: device not accepting address 5, error -110
[17179904.792000] usb 4-2: new high speed USB device using ehci_hcd and address 7
[17179916.336000] usb 4-2: device not accepting address 7, error -110
[17179916.636000] usb 4-2: new high speed USB device using ehci_hcd and address 9

This is a dapper 6.06.... Hope this gets patched soon, since it's quite annoying!

[QuinnHarris (ubuntu-quinnh)]

I can verify this problem isn't in 2.6.15-25-386 but is in 2.6.15-26-386 and edgy 2.6.17-7-386.

"sudo rmmod ehci_hcd" allows the device to work but at USB 1.1 speeds.

This is on an Averatec 3250 laptop.
Via south bridge, USB 2.0 PCI ID 1106:3104

[Tero Karvinen (karvinen+launchpad)]

I have the same problem. With default settings, USB key lights are not lit when connected, and nothing is printed on syslog. With some changes, I have succeeded to get lights lit and lines on syslog, but mounting is not possible. "sudo rmmod ehci_hcd" does not help on this system. USB key works on another Dapper installation with older updates.

I see no unanswered questions in this bug report, so I change "Needs Info" to "Unconfirmed".

It is very uncomfortable to see updates breaking basic things in production version. I feel that importance of this bug should be raised.

Tested on
$ grep DESC /etc/lsb-release
DISTRIB_DESCRIPTION="Ubuntu 6.06.1 LTS"
$ uname -srvm
Linux 2.6.15-26-k7 #1 SMP PREEMPT Fri Sep 8 20:38:34 UTC 2006 i686
$ sudo apt-get update && sudo apt-get upgrade # 2006-09-17
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

[QuinnHarris (ubuntu-quinnh)]

I agree with Tero. This really should be a higher priority. This is a major regression that would effectively completely break USB for a typical user with this hardware. I think this problem suggests it could be prudent to either be more conservative with kernel updates or provide a mechanism to get these changes tested by a wider audience before they make it to an "average user". If a typical Windows user installed Ubuntu then USB just broke, they would consider Ubuntu a flaky piece of shit.

Every time I google this I find more people and more Ubuntu bugs that all look like the same problem.
Check
Bug 53972
Bug 54273

Is this a fix?
http://archives.free.net.ph/message/20060829.011900.bece9a96.en.html

[QuinnHarris (ubuntu-quinnh)]

This has been fixed in Edgy 2.6.17-8-generic.
Not sure about Dapper.

Thanks.

[TRiSS (triss)]

It seems fixed with the latest kernel update for dapper, i believe it was 2.6.17-27-k7 (can't check version right now)

[Tero Karvinen (karvinen+launchpad)]

> It seems fixed with the latest kernel update
> for dapper, i believe it was 2.6.17-27-k7

There is no 2.6.17 in Dapper. Do you mean 2.6.15-27?

I am running Dapper and still have no usb working - my problem might be different than this bug (Bug #36619).

[TRiSS (triss)]

>> It seems fixed with the latest kernel update
>> for dapper, i believe it was 2.6.17-27-k7
>
>There is no 2.6.17 in Dapper. Do you mean 2.6.15-27?
>
>I am running Dapper and still have no usb working - my problem might be >different than this bug (Bug #36619).

sorry, my bad, my current kernel, with working usb2.0 is indeed 2.6.15-27-k7!

[Ben Collins (ben-collins)]

From reading the comments, this was reported fixed in Edgy.

[Ben Collins (ben-collins)]

From reading the comments this was reported fixed in latest dapper kernel.