From: CSBVAX::MRGATE!RELAY-INFO-VAX@CRVAX.SRI.COM@SMTP 21-SEP-1988 04:50 To: ARISIA::EVERHART Subj: MON ETHERNET update Received: From KL.SRI.COM by CRVAX.SRI.COM with TCP; Tue, 20 SEP 88 16:04:59 PDT Received: from uhnix1.uh.edu ([129.7.1.14]) by KL.SRI.COM with TCP; Tue, 20 Sep 88 15:46:42 PDT Received: from uhvax1.uh.edu by uhnix1.uh.edu (5.59/1.0) id AA20811; Tue, 20 Sep 88 11:54:52 CDT Message-Id: <8809201654.AA20811@uhnix1.uh.edu> Date: Tue, 20 Sep 88 11:54 CDT From: COSC35N@uhvax1.uh.edu Subject: MON ETHERNET update To: INFO-VAX@KL.SRI.COM X-Vms-To: IN%"INFO-VAX@KL.SRI.COM" Here are the VMS MONITOR patches,updated for 4.6 and 4.7. VMS 5.0 patches are not yet available. These add the following display classes to the MONITOR utility. JOURNAL RU JDEVICE VMS1 ETHERNET $ set def sys$common:[sysexe] $ patch monitor.exe replace 4158+33e 211 EXIT 11 EXIT replace 4158+391 211 EXIT 11 EXIT replace 4158+48A 235 EXIT 35 EXIT replace 4158+583 211 EXIT 11 EXIT replace 4158+629 211 EXIT 11 EXIT replace 170f6 0 exit 6006 exit UPDATE EXIT $ install:==$install/comm $ install replace sys$system:monitor.exe It was necessary to patch one other longword to make MONITOR ETHERNET work on VMS 4.6 and 4.7. In VMS 4.6, the Ethernet device drivers started enforcing checking of the protocol type field in QIOs directed to them. MONITOR ETHERNET QIOs use promiscuous mode to collect their data. Although the protocol type field is a required parameter, it is supposed to be ignored when promiscuous mode is in effect. This was the pre VMS 4.6 behaviour. Starting in VMS 4.6, the protocol type field must be a valid type, even when you are using promiscuous mode. The protocol type field for the QIOs is at address 170F6. It was previously 0, which is not a valid protocol type. To use MONITOR ETHERNET on later versions, I deposit a longword 6006 (hex) at this location. This protocol type is documented as reserved for the customer - and that's us! The ETHERNET class gives a monitor page that displays the statistics from the ETHERNET adapter. It does not, however, work on the Ethernet devices of newer BI class machines. If you have a BI VAX, and your Ethernet device calls itself ETA0, this display will always have zeroes in all fields. If you are trying to MON ETHERNET a device other than XEA0, or XQA0, then you will need to define a logical name, NI$DEVICE, to specify what device to monitor. For example, $ def/system ni$device XEB0 to monitor a second DEUNA, rather than the 1st one. Also, PHY_IO privilege is required to use MONITOR ETHERNET. USE THESE PATCHES AT YOUR OWN RISK!!!! KEEP A COPY OF THE UNALTERED IMAGE!!! They haven't caused any problem on all of my VAXen (of all sizes), but, you're crazy if you think I will be responsible for anything they do on your systems. I have used these displays on VMS 4.4, 4.5, 4.6, and 4.7. VMS 5.0? Who can say? I'll be looking for them there, as soon as my copy arrives. Since VMS 4.4 supplied a new MONITOR image, odds are these patches wouldn't be correct on version 4.3 and older. Lee K. Gleason Control-G Consultants 2416 Branard #D Houston TX 77098 713/528-1859 COSC35N@UHVAX1.UH.EDU