On Fri, Mar 24, 2000, Benjamin Herrenschmidt <[EMAIL PROTECTED]> wrote:
>>Aki's fix is to switch steps 3 and 4 above so that the
>>HcControlHeadED or HcBulkHeadED registers get initialized with
>>the address of a valid endpoint descriptor before the
>>corresponding list is started. This is the right thing to
>
>Yes, indeed. That would be a good explanation. Address 0 is perfectly
>valid for a PCI device, the controller has no good reason not to consider
>it as a valid ED.
A bit late (I didn't have mail this we), I confirm that the crashing OHCI
bug on the G4 is gone with the recent OHCI patches applied.
There's still a minor issue: occasionally, devices plugged at startup
won't be detected properly (timeout error). I'll investigate more later
this week. The problem is intermittent and seems to happen more often
since I removed part of the debug output of the various USB drivers (I
needed some room in my kernel message buffer to track something else ;).
So I beleive it's somewhat timing-related. Unplugging and plugging back
the device works.
Ben.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
