<html><head><meta http-equiv="content-type" content="text/html; charset=utf-8"></head><body dir="auto"><div><div style="direction: rtl;"><br></div></div><blockquote type="cite"><div><span style="-webkit-text-size-adjust: auto;">Thanks Roland to clarify our confusion.</span><br><span style="-webkit-text-size-adjust: auto;"></span><br><span style="-webkit-text-size-adjust: auto;">So looks ping-pong mechanism is the way to go.</span><br><div style="direction: ltr;"><span style="-webkit-text-size-adjust: auto;"><br></span></div></div></blockquote><div><br></div><div>Not sure if it will work for your solution, but you can also register to SM traps.</div><br><blockquote type="cite"><div><span style="-webkit-text-size-adjust: auto;"></span><span style="-webkit-text-size-adjust: auto;">Regards,</span><br><span style="-webkit-text-size-adjust: auto;">Jack</span><br><span style="-webkit-text-size-adjust: auto;"></span><br><span style="-webkit-text-size-adjust: auto;">2014-10-23 20:43 GMT+02:00 Roland Dreier <<a href="mailto:roland@purestorage.com">roland@purestorage.com</a>>:</span><br><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>On Thu, Oct 23, 2014 at 6:50 AM, Jack Wang <<a href="mailto:xjtuwjp@gmail.com">xjtuwjp@gmail.com</a>> wrote:</span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><blockquote type="cite"><blockquote type="cite"><span>I expected that RDMA-Write operations will fail if the other crashes.</span><br></blockquote></blockquote></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><blockquote type="cite"><blockquote type="cite"><span>Also I hoped that an event is generated when a host is crashed. The subnet</span><br></blockquote></blockquote></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><blockquote type="cite"><blockquote type="cite"><span>manager should notice it and notify every other device in the network.</span><br></blockquote></blockquote></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><blockquote type="cite"><blockquote type="cite"><span></span><br></blockquote></blockquote></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><blockquote type="cite"><blockquote type="cite"><span>Are we missing something in our modules?</span><br></blockquote></blockquote></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><blockquote type="cite"><blockquote type="cite"><span>Is there a way to determine that a RC peer crashed without implementing a</span><br></blockquote></blockquote></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><blockquote type="cite"><blockquote type="cite"><span>ping-pong mechanism?</span><br></blockquote></blockquote></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span></span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>If the remote system crashes then any memory regions, QPs, etc. are</span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>still valid with the remote HCA, and RDMA read/write operations will</span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>continue to succeed. (Unless the system reboots and reinitializes the</span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>adapter or something like that).</span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span></span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>There isn't a way to detect a remote crash unless that remote crash</span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>disconnects your QP or otherwise affects the HCA on the crashed</span><br></blockquote><blockquote type="cite" style="-webkit-text-size-adjust: auto;"><span>system.</span><br></blockquote><span style="-webkit-text-size-adjust: auto;">--</span><br><span style="-webkit-text-size-adjust: auto;">To unsubscribe from this list: send the line "unsubscribe linux-rdma" in</span><br><span style="-webkit-text-size-adjust: auto;">the body of a message to <a href="mailto:majordomo@vger.kernel.org">majordomo@vger.kernel.org</a></span><br><span style="-webkit-text-size-adjust: auto;">More majordomo info at <a href="http://vger.kernel.org/majordomo-info.html">http://vger.kernel.org/majordomo-info.html</a></span><br></div></blockquote></body></html>