Hello, I've got a further question concerning the occurrence of an excessive deferral. When generating such a condition (I just keep CSR='1' long enough for the nibble counter to pass h17b7 while TX wants to send a frame), I noticed the TX is not able to come back from the "StateDefer". I See that TxEthMac drives TxAbort='1', however, this is filtered out within MacControl because TxStartFrmIn='1' and with that, TxAbortOut is kept low. This causes the host interface to wait forever to be able to send the frame and the TX to stay in "StateDefer". Has anyone experienced such behaviour, too? Regards, Peter By the way: Has anybody ever answered thread [ethmac] DeferLatched signal of the newsgroup (Sandhya Sastry, 14-Jun-2004), as I experience the same thing and did not find additional comments to this threat? ======================================================================== Kaeser Peter Senior Engineer BridgeCo AG Direct +41 1 802 33 47 Ringstrasse 14 Phone +41 1 802 33 33 CH-8600 Duebendorf Fax +41 1 802 33 39 Switzerland mailto:peter.kaeser@bridgeco.net http://www.bridgeco.net