scotte@warped.com wrote:

> The behavior seems very similar to what is described by your final post 
> on that PR.  I don't think it's a gcc4 problem (although I am using 
> gcc4) since I believe I saw the issue before amd64 switched over, 
> although I could be mistaken.
> 
> I'm honestly clueless as to how to debug this further, and have just 
> been living with it, since I can usually re-plug the devices to make 
> them show up. :-)

On macppc VIA UHCI was mostly useless but *sometimes* worked.
I've put some (a bit old) debug logs:

http://www.ceres.dti.ne.jp/~tsutsui/netbsd/via-uhci-fail1.log
http://www.ceres.dti.ne.jp/~tsutsui/netbsd/via-uhci-fail2.log
http://www.ceres.dti.ne.jp/~tsutsui/netbsd/via-uhci-fail3.log
http://www.ceres.dti.ne.jp/~tsutsui/netbsd/via-uhci-fail4.log
http://www.ceres.dti.ne.jp/~tsutsui/netbsd/via-uhci-fail5.log
http://www.ceres.dti.ne.jp/~tsutsui/netbsd/via-uhci-success.log

Looks timing issue (since each failure log is different from others).


Just FYI, Microsoft also has some fixes for VIA UHCI:
http://support.microsoft.com/kb/240075/en-us

 "This problem occurs because there is a timing-specific problem
  in the Universal Host Controller driver (Uhcd.sys) that may
  prevent USB devices from being properly enumerated under
  specific timing conditions."

http://support.microsoft.com/kb/232889/en-us

 "This issue can occur if the VIA hardware chipset for the device
  does not initialize properly."
---
Izumi Tsutsui