<!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] Re: [PATCH] [2/2] change QP state to SQE</TITLE>
</HEAD>
<BODY>
<BR>
<BR>
<P><FONT SIZE=2>> -----Original Message-----</FONT>
<BR><FONT SIZE=2>> From: Roland Dreier [<A HREF="mailto:roland@topspin.com">mailto:roland@topspin.com</A>] </FONT>
<BR><FONT SIZE=2>> Sent: Friday, November 12, 2004 7:42 PM</FONT>
<BR><FONT SIZE=2>> </FONT>
<BR><FONT SIZE=2>> </FONT>
<BR><FONT SIZE=2>> I thought about this a little, and it seems that having the </FONT>
<BR><FONT SIZE=2>> CQ poll operation update the QP state is not the right </FONT>
<BR><FONT SIZE=2>> solution. It seems it would be better to add support for the </FONT>
<BR><FONT SIZE=2>> "Current QP state" modifier for the modify QP operation and </FONT>
<BR><FONT SIZE=2>> expect the consumer to use that to indicate that the QP is in </FONT>
<BR><FONT SIZE=2>> SQE state.</FONT>
<BR><FONT SIZE=2>> </FONT>
</P>
<P><FONT SIZE=2>Actually I recall adding "current QP state" as an input modifier</FONT>
<BR><FONT SIZE=2>to the modify QP verb as part of the IB 1.1 errata (if I remember</FONT>
<BR><FONT SIZE=2>correctly). The main intention is to avoid the ambiguity when</FONT>
<BR><FONT SIZE=2>a consumer moves a QP into RTS state but can't tell if the QP</FONT>
<BR><FONT SIZE=2>was in SQError/Error or SQDrain. According to the spec, current</FONT>
<BR><FONT SIZE=2>QP state should only be valid when moving QP into RTS state.</FONT>
<BR><FONT SIZE=2>Hope that helps.</FONT>
</P>
<P><FONT SIZE=2>-Dror</FONT>
</P>
</BODY>
</HTML>