Georg Acher wrote:
...
>It looks like your case (stall in data phase) is treated as a short packet
>with a manually executed status phase, which is wrong. The "stalled" state
>is detected (and IMHO the endpoint state of the device-structure should
>reflect this), but it doesn't survive in the end...

Interesting;  I didn't look at the code (sorry, I was in a hurry).

>I will fix it..

excellent!

any idea what uhci.c does in this case?  (it looks like it will return
an error and not do the status phase, and usb-ohci.c looks the same)

Interesting, if you fix this my audio box will stop working.  It seems
to respond to one of the control messages with a "protocol stall" (as
opposed to a function stall).  note that protocol stall = stall on
control pipe.  This seems to indication a function which is not
implemented or does not work...

I'm wondering if, in fact, it's ok/better/recommended to include the
status phase for an 'protocol stall', but not a function stall...  I'd
be curious to hear what others think.

I've had problems getting audio.c to work with different audio
devices.  Some of the behavior has been hci driver dependent
(i.e. ohci != uhci != usb-uhci) on some of these finer points.
Humm... maybe this is why usb-ohci worked better :-)

-brad




_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to