Re: Kernel panic with PCI EHCI on AMCC Sequoia

On Thursday 23 August 2007, Mike Nuss wrote:
> Stefan Roese wrote:
> > 
> > Might I ask you to do one last test, to verify, that this is 
> > a PPC specific 
> > problem? Please test this setup with the same PCI card and 
> > the same USB hub 
> > in a standard x86 PC (using the same kernel version).
> 
> The same kernel built for x86 and running on a PIII works fine with this
> controller.
> 
> Maybe the EHCI driver is still making some assumptions about the byte
> order of the host? After all, the controller and the x86 are both
> little-endian. I haven't tracked it down much further yet, but I've
> determined the crash is happening the first time scan_periodic is
> called. I'd guess it's reading in a value from the controller and trying
> to use it as an address without properly converting it.

It's been years since I've heard of bugs in this code on pure
big-endian systems though ... makes me think the issue relates
to the mixed byteorder.  (Although I couldn't rule out a bug
that just hasn't been reported.)

The driver should never be mapping hardware addresses into
CPU pointers, so that hypothesis doesn't ring true.

- Dave

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel