From: MERC::"uunet!syr.ge.com!PAVLIN" 19-MAY-1992 21:37:39.06 To: galaxy::GleEve, GELD@mail.physics.lsa.umich.edu, SEB@lns61.tn.cornell.edu, OSUDAR@noah.cmt.anl.gov, VERSTEEGEN@rulfsw.LeidenUniv.nl, KROON@KVI.nl, e05004@nl.usna.navy.mil, Robert.Palamara%minenco.otc.au@clix.aarnet.edu.au, 71071.1411@compuserve.com, KXGLW@ACVAX.INRE.ASU.EDU, pwa-b!saldi, balr!vpnet!cmf, greg_a@npri6.npri.com, moeller%gwdgv1.dnet.gwdg.de@cunyvm.cuny.edu, MACALLSTR@physics, oxford.ac.uk@syr.ge.com, rbn@ralph.rtpnc.epa.gov, vcs!gary_p@ucdavis.edu, YOUNG@eisner.decus.org, CC: Subj: new release of ETHERMON (beta-test) Greetings. Thank you for your problem reports on ETHERMON with regards to VMS V5.4-3 and up. A new version of ETHERMON which does work under VMS V5.4-3 successfully is now available for beta-testing, if you are interested. If so, please send an e-mail message to pavlin@syr.ge.com, and I will tell you how to get the beta-test version. I would appreciate any comments regarding its performance, features, etc., before I close ETHERMON V2.4 to a production release. A warning: ETHERMON V2.4 under VMS V5.4-3 and up requires CMKRNL privilege (the privilege is preferred for earlier versions). An explanation for why is in the revised user's manual, but, briefly: DEC changed the way the P5 buffer (containing the Ethernet packet header) was returned to the user application. The way this change impacted on promiscuous mode operation prevented Ethernet packets from being deciphered correctly in IEEE 802.3 mode, and vice versa. To correct the problem, ETHERMON now intercepts the $QIO service call's completion to properly return the raw packet header (unmangled by the updated device drivers), so as to let the user-mode code in ETHERMON decide how to interpret the packet. This requires kernel-mode code to intercept I/O postprocessing of the read requests, hence the need for CMKRNL privilege. The source of the code that does the intercepting is included in the ETHERMON distribution kit (it HAS to be included because that part has to be recompiled to work under different versions of VMS), so you can examine how the intercept works and decide if you want this code on your system. The user's manual explains what precautions were taken to prevent malicious users and programming errors from taking advantage of this intercept code. Note that this intercept is only invoked for Ethernet read requests posted by ETHERMON; it does NOT affect general I/O postprocessing on the system (including other I/O calls by ETHERMON). I hope that this new release will give you what new features you are looking for in ETHERMON. Thanks for using my product! Andrew Pavlin, TCP/IP network administrator General Electric Co., Aerospace Information Technology Syracuse, NY 13221 U.S.A. phone: 315-456-6212 (7:30AM to 3:30PM, U.S.A. Eastern Daylight Savings Time, or leave voice-mail with time to return call) email: pavlin@syr.ge.com