Re: USB timeout / "control timeout on ep0out"

[David Wood <obsidian@panix.com>]

I have this exact same problem. I've had it ever since I went to amd64, 
for a long long time now. Don't know if it's an amd64 thing, a 
particular-USB-chip thing, a debian/discover/hotplug thing, or a recent 
Linux kernel thing altogether. Interestingly, though, Graham's motherboard 
is completely different from mine, an Asus A8V deluxe (i.e. Via versus 
Nvidia chipsets)...

It used to freeze every 1 boot in 2, rather than 1 in 15! This was ugly, 
and I was morosely starting to learn the USB code in the kernel and 
watching kernel traffic/patches constantly for USB-related changes. Recent 
kernel upgrades have drastically reduced the frequency, but it still does 
happen once in a while. These intermittent problems are absolutely awful. 
I actually considered it fixed by now, but last night I saw it again for 
the first time in about three weeks.

Another thing I established about this problem: failure frequency seems to 
react positively or negatively (but in no way I could ever understand) to 
moving devices between ports, removing, or adding devices, hubs, etc. Of 
course, even if it's some device that's misbehaving, that's no reason for 
this kind of failure...

On Thu, 26 May 2005, Javier Kohen wrote:

> El jue, 26-05-2005 a las 18:09 +0100, Graham Smith escribió:
> > Hi,
> > About 1 time in 15 the boot process fails and locks solid due to, I
> > think, a time-out problem while initializing the USB system. The
> > following are the last (useful looking) messages
> >
> > ehci_hcd 0000:00:02.1: USB 2.0 enabled EHCI 1.00, driver 2004-may-10
> > usb 1-8: device not accepting address 2, error -110
> > usb 1-8: new full speed USB device using address 4
> > usb 1-8: control timeout on ep0out
>
> > Linux doozer 2.6.8-11-amd64-k8 #1 Wed May 4 20:27:04 UTC 2005 x86_64
> > GNU/Linux
>
> Try to reproduce the problem with a more recent kernel. I guess that you
> can force the problematic step by repeatedly removing and re-installing
> the ehci module. To be on the safe side, try that first on your current
> kernel and see if you can hang it. Or just reboot a few dozen times with
> kernel 2.6.12-rc4 if you have the time.
>
> Greetings,
> --
> Javier Kohen <jkohen@users.sourceforge.net>
> ICQ: blashyrkh #2361802
> Jabber: jkohen@jabber.org