Hi,

we are writing a linux usb driver for a ISDN device that has 3 ISO IN and 3 ISO out endpoints. Therefore we have 6 iso chains running during transfers.
We programmed the iso chains that we send 8 iso requests in a single URB and everytime we have one URB request still pending.
In the CATC USB analyzer we see that we have periods of time where in up to 20 SOF's (or equally 20 msec) there is no ISO traffic. This "gaps" in the USB traffic produces no submit or ISO packet status errors. It looks like as if the kernel found no time to call the completion handler that resubmits the ISO URB requests. Since the USB ISDN chip has only 128 bytes deep FIFO we lose data when traffic stops more than 16 msec.
It seems that behaviour is slightly different from machine to machine and depends on the type of usb host controller we use. We see these large gaps in the traffic only with OHCI host controller.


Has anybody an explanation for this? Is there a way to avoid this proplem?

Regards

Peter Sprenger



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to