[Nut-upsuser] Tripplite OmniPlus1000LCD Problem on VIA system

[Craig Van Degrift]

Charles,

Thanks for confirming that both OHCI and UHCI support 1.0 and 1.1.  Also, at 
your suggestion I checked my Sony Mavica camera.  It is USB 1.1 and works 
fine with the UHCI.  I have joined the linux-usb mail list and submitted my 
problem including syslog output made with USB debug compiled in the kernel.  
I had incorrectly assumed that the low-level code was rock solid, but from 
the activity on linux-usb, I see that there are lots of fixes underway.

I am going to assume that it is a problem that will be resolved by future 
updates to the linux kernel.  The UPS was $100 at CostCo and I will refrain 
from taking it back.  In the worst case, I can help the usb developers 
isolate the problem. Other Linux users will no doubt be buying these units.

/proc/interrupts shows ehci_hcd, and the four uhci_hcd devices, but they share 
interrupts with other things so I don't see how to separate out their 
activity.  The counts rise rapidly even when I do nothing with USB gadgets.

On Tuesday 12 August 2008 19:07:25 Charles Lepple wrote:
> On Sun, Aug 10, 2008 at 4:51 AM, Craig Van Degrift
>
> <craig at yosemitefoothills.com> wrote:
> > I have two computers running Ubuntu 8.04 with all updates.  One with EHCI
> > and OHCI host controllers recognizes my Tripplite OmniPlus1000LCD UPS,
> > but the other (VIA chipset) with EHCI and UHCI does not.  When I compiled
> > the kernel with USB debugging enabled, I could see that the handoff from
> > EHCI to the low speed UHCI controller was failing leading to endless
> > looping as the EHCI was repeated retried.
> >
> > Should the difference between OCHI and UHCI really matter in this case? 
> > I thought both supported USB 1.0 and USB 1.1 protocols, but I am quite
> > fuzzy on such matters.
> >
> > I have compiled and installed the newer 2.2.1-2 nut usbhid-ups driver,
> > but this problem prevents me from even getting the OmniPlus1000LCD
> > registered by the USB low level drivers.  lsusb does not show it.
>
> If lsusb doesn't see the device (possibly run as root to avoid
> permission problems), then it does sound like a lower-level issue.
>
> OHCI and UHCI both support USB 1.0 and 1.1.
>
> It could be that the VIA board is not routing interrupts correctly. Do
> you see any UHCI activity in /proc/interrupts? Do other USB 1.1
> devices work on that particular port? There are a few debug options
> that can be passed to the kernel that will use other means to
> configure the interrupt routing.
>
> It sounds like a question that might be better answered on a generic
> Linux USB mailing list or an Ubuntu forum, though.