USB host on Ingenic CPUs

Daniel Glöckner daniel-gl at
Sat Sep 8 07:45:13 EDT 2012


for analyzing the problems observed with a cheap no name USB serial
adapter in combination with my JZ4730 based Letux 400, I triggered
an oscilloscope once the OHCI controller aborts an URB with an error.
The capture clearly shows several NACK'ed IN transactions followed by
a single IN transaction without response from the USB device.
The error counter in the transfer descriptor had the value 3.

Here the missing response is not the problem. The problem is that
the USB spec states that transactions must be aborted after three
errors _in_a_row_. I suspect that the Ingenic OHCI controller fails
to reset the error counter on a NACK'ed transaction. 

Can anyone confirm my observations?
What about other Ingenic chips?


More information about the jz47xx-kernel mailing list