<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2654.45">
<TITLE>RE: [openib-general] ib_mad shutdown WC status code ?</TITLE>
</HEAD>
<BODY>
<BR>
<BR>
<P><FONT SIZE=2>> -----Original Message-----</FONT>
<BR><FONT SIZE=2>> From: Sean Hefty [<A HREF="mailto:mshefty@ichips.intel.com">mailto:mshefty@ichips.intel.com</A>] </FONT>
<BR><FONT SIZE=2>> Sent: Friday, September 24, 2004 6:41 PM</FONT>
</P>
<P><FONT SIZE=2>> I think that flushed makes sense. With queuing in the access </FONT>
<BR><FONT SIZE=2>> layer, we should be able to stop and restart the QP in most </FONT>
<BR><FONT SIZE=2>> error cases without affecting the user. If the link goes </FONT>
<BR><FONT SIZE=2>> down, flushing the requests seems reasonable. What other </FONT>
<BR><FONT SIZE=2>> cases would cause the MAD layer to shut down?</FONT>
<BR><FONT SIZE=2>> </FONT>
</P>
<P><FONT SIZE=2>If link goes down, UD QP will go on "sending" the packets as</FONT>
<BR><FONT SIZE=2>if link was up all the time. You wouldn't get flushed WQEs in this</FONT>
<BR><FONT SIZE=2>case for a UD QP.</FONT>
<BR><FONT SIZE=2>In general, for UD QPs, you'll get flushed WQEs when the QP</FONT>
<BR><FONT SIZE=2>moves to the error state and that seldom happens. The exact</FONT>
<BR><FONT SIZE=2>scenarios for this to happen (hopefully I don't miss anything):</FONT>
<BR><FONT SIZE=2>- s/g element points at a wrong address/protection error</FONT>
<BR><FONT SIZE=2>- HCA can't fetch a WQE (e.g. bus error)</FONT>
</P>
<P><FONT SIZE=2>-Dror</FONT>
</P>
<P><FONT SIZE=2>> - Sean</FONT>
</P>
</BODY>
</HTML>