-CURRENT, EHCI, and the iPod

[Julian Elischer]

On Mon, 8 Mar 2004, Eric Anderson wrote:

> Julian Elischer wrote:
> [..snip..]
> >>Have you been able to successfully connect a USB 2 device to this controller?  
> >>The behavior I've seen on a couple ICH4 boards is that the controller gets 
> >>recognized fine and USB 1 devices work (through UHCI).   However, as soon as 
> >>I connect a USB 2 device, I get the "controller halted" and "disabling port" 
> >>messages.
> > 
> > 
> > yes I have read a USB-2 flash 'keyring' drive at 54Mb/sec. well above
> > the 12Mb/sec that USB1 is capable of..

Correction.. I have a test machine that exhibits the problem you
describe. it is also an intel EHCI controller.



> > 
> > 
> > 
> > 
> > 
> >>Tried to debug it but didn't have much luck.  I suspect it's some subtle 
> >>timing bug or other weirdness with Intel's implementation. . .
> >>
> >>My two NEC-based ones work ok -- one PCI and one CardBus.  I do
> >>sometimes have panics with umass, and occassionally the laptop
> >>somehow loses all ability to deliver interrupts, to anything (almost
> >>always when accessing the USB hard drive for the first time after a
> >>reboot).
> > 
> > 
> > Note this is with the MFC'd -stable code.
> 
> So it should work in -current then, right?
> 
> I'll build a new kernel with EHCI support and give it a try..

Try with acpi enabled and disabled..

> 
> Eric
> 
> 
> -- 
> ------------------------------------------------------------------
> Eric Anderson     Sr. Systems Administrator    Centaur Technology
> Today is the tomorrow you worried about yesterday.
> ------------------------------------------------------------------
>