<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">I’ve made a number of bug fixes to the Linux DECnet dnlogin program. These were found trying to login to TOPS-10 but it’s quite possible that they could be generated by logging in to other operating systems depending on the usage (e.g. timing out read requests was totally broken and generated a message that TOPS-10 was not expecting). I have no idea if this is related to the problem you are seeing but it may be worth a try. My version of the code is available on Erik Olofsen’s site (<a href="http://rullf2.xs4all.nl/decnet/" class="">http://rullf2.xs4all.nl/decnet/</a>) as version 2.68 (this version also includes a number of fixes to the DAP utilities).<div class=""><br class=""></div><div class=""> John.<br class=""><div class=""><br class=""></div><div class="">On Oct 30, 2018, at 9:18 PM, Jeffrey H. Johnson <<a href="mailto:jhj@trnsz.com" class="">jhj@trnsz.com</a>> wrote:<div><blockquote type="cite" class=""><br class="Apple-interchange-newline"><div class=""><div class="">It's actually the Linux DECnet suite, all running as part of the user process on the (OpenBSD) host, as part of the Multics FNP simulation - new incoming CTERM connections (yes, using the essentially unmodified Linux DECnet implementation) are treated as dial-in connections over simulated serial lines ... the whole thing is rather odd, but essentially the setup is the same way a foreign network of this type would would have been accessed via protocol converters on real iron.<br class=""><br class="">There surely are still bugs of my own, especially in the part creates Ethernet frames and provides a TAP interface. I know there are some ways to make the whole thing crash that don't happen in the Linux code, but those situations are impossible to trigger by end users. <br class=""><br class="">Unfortunately, it seems the particular bug I'm looking for exists in the original Linux DECnet as well, so it's in code I'm not overly familiar with and probably haven't even looked at. <br class=""><br class="">None of those issues are related to performance, however. I'm going to actively investigate moving to that Python-based DECnet implementation at some point in the future. There is also the possibility of a hardcore/ring-0 based implementation in the further future as well, provided I continue to get more comfortable with PL/I - such an implementation would be able to be more efficient.<br class=""><br class="">Right now, however, my main concern is it being a good citizen on the network.<br class=""><br class="">--<br class="">Jeffrey H. Johnson<br class=""><a href="mailto:jhj@trnsz.com" class="">jhj@trnsz.com</a><br class="">https://ban.ai/multics<br class=""></div></div></blockquote></div><br class=""></div></div></body></html>