(again - deleted html mime contents)
(resent again, subject line was not accepted by "moderator" when sent from 
Outlook)
(resent another , case)

I have to little knowledge of USB and OHCI to judge this errata with
it suggested workaround myself. Will this work or should we design in
an external USB host?

"14 Bit stuffing right before the EOP (USB)

Problem;

When there is a bit stuffing at the end of the CRC, just before the
EOP, (ie CRC = XXXX XXXX 1111 11XX), one byte is added to
those already received, giving the first CRC byte received as the
last data byte of the received data packet. This delays the EOP recognition
by one cycle and generates an extra byte write pulse. As
a consequence, the first CRC byte is perceived as a data byte.

Implication; Data coming from the USB device may be corrupted

Workaround;
If the size of the data transfers are known, set the buffer length to
the exact number of bytes to be received."

Can this workaround be functional? It requires that
* Message length is known in advance - are they? When doesn't this hold?
  Can the OHCI driver be modified or is it written in this way already?
* Messages with unknown length does not contain several ones at the end
 - can this be trusted on?
* Retransmit of those messages does not contain the same data at the end
 - I would be supriced if this is the case.

Is another workarounds possible?
* Ignore the problem
  Will this extra data byte cause any problem?
  Checksum error? Data size errors?
  Probably not possible - then there would not be this errata in the first
place.

* If recived message is one to big, check if the last expected byte
  contains low ones. And if this is the case shorten the message by one.
  Will this approach result in a check sum errors? Then what about the next
  approach?

* If received messages results in checksum errors, try to receive a one byte
  shorter message.
  How is retransmit after checksum error handled?

/RogerL







-------------------------------------------------------
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