RTE 6.2 Patch Information **************************************************************************** 12/01/95 COPYL can now handle tracks over 32767 PROBLEM DESCRIPTION: SR#1653146159 COPYL cannot handle tracks over 32767 FIX DESCRIPTION: ©L has been changed to use double integer values for tracks and sectors where appropriate in its calculations. COPYL is now able to make its disk-to-disk copy using disk lus with up to 65534 tracks. Fixed in the 6.21 patch release. FILES: COPYL.READ (This File) %COPYL REV.6201 951201 INSTALLATION: Relink COPYL using #COPYL from the RTE-A product directory. OS Revs Supported: 6.1 and 6.2 USING THE NEW FEATURE: Use COPYL as always. INSTALLATION RECOMMENDATION: Install if you need to run COPYL on disk lus with more than 32767 tracks. **************************************************************************** 11/10/95 New relocatable for %dvs23 that corrects IO14 in FST backups using DAT tapes. PROBLEM DESCRIPTION: SR#4701308213 RTE-6 FST backup with verify gets IO14 on the terminal and IOTO on the system terminal. INSTALLATION: Replace the old file %dvs23 with the new rev. 6201 %dvs23 and re-generate the system. If you are an expert user of RTE-6, you may patch the driver in physical memory by using CMM6 in the following manner. 1. Locate the physical page of dvs23 in the generation map. 2. Use CMM6 to display the driver entry point IS23. example: le,IS23 3. Use CMM6 to display location 121B from the entry point IS23. Location 121B should contain 176030B which is DEC -1000. 4. To confirm the correct location use CMM6 to display two locations above and four locations below 121B. They should be the following values or instructions that have been relocated. 117 000010 OCT 10 120 000037 OCT 37 121 176030 DEC -1000 122 161667 LDA EQT8,I 123 040137 ADA M3 124 002020 SSA 125 024063 JMP REJ01 5. Use CMM6 to change Location 121 from -1000 to -2000. 6. Use CMM6 with caution and be sure to re-generate your system with the new %dvs23 for a permanent fix. FILES: dvs23.read (This file) %dvs23 rev.6201 OS Revs Supported: 6.1 6.2 Others: Keep factory informed if this corrected the problem. **************************************************************************** ************************************************************************ FOWN.REL 5 May 1996 FOWN now correctly displays > 32767 files 5 May 1996 Patch to allow FOWN to properly Fixed SRs: KPR:5003-306688 report the number of files if the number of files > 32767 PROBLEM DESCRIPTION> FOWN reports negative number of files if number of files is greater than 32767. --------------------------------------------------------------------------- INSTALLATION: Reload FOWN.RUN using FOWN.LOD from the RTE-A product directory and the 6201 patched version of FOWN.REL. FILES: FOWN.READ (This file) FOWN.REL FOWN relocatable file REV. 6201 . . . OS Revs Supported: 6.1 and 6.2 INSTALLATION RECOMMENDATION: Install if desired to fix the described problem. ************************************************************************ **************************************************************************** 2/02/96 New relocatable file to correct problem in which GRUMP aborts when a response to one of its questions begins with "/A", even when providing name of logon or logof command file. PROBLEM DESCRIPTION: SR#4701312868 GRUMP aborts when response begins with "/A" even for logon cmd file FIX DESCRIPTION: There is a new GRUMP.REL. GRUMP's routine to check for the character string "/A" (which indicates the user wants to abort GRUMP) now checks the response string to see if there are charac- ters after the "/A", which would indicate that the response is a file descriptor. INSTALLATION: Copy the new relocatable to the VCPLUS product directory. Run link on load file GRUMP.LOD to create the new version of GRUMP. FILES: GRUMP.REL REV.6201 960202 OS Revs Supported: 6.1 and 6.2 ************************************************************************ Date: 22 June 1995 SR# 5003270108 ID100 does multiple program schedules for one unsolicited character PROBLEM TEXT: ID100 was setting the schedule program bit in DV20 on completion of each write following receipt of an unsolicited character until a read request was executed. That would cause DDC00 to schedule the primary or secondary program. This problem was introduced in 6.1. CAUSE TEXT: The Data Received (DR) bit in the card status word was not being cleared after an unsolicited interrupt. On completion of a write request this bit is checked to determine if a key was struck during the write. SR# 4701295196 With ID100, unsolicited characters are echoed in some cases PROBLEM TEXT: Unsolicited characters are echoed if the last request that ID100 processed was not a write request. These characters should cause a program schedule, but should not be echoed. This problem was introduced in the 6.1 release. CAUSE TEXT: The ECHO bit in the card control word was not being cleared after completion of a read request. FILES: ID100.REL REV.6202 950622 ID101.REL REV.6202 950622 INSTALLATION: The system must be regenerated using RTAGN to install this patch. OS Revs Supported: 6.1 and 6.2 INSTALLATION RECOMMENDATION: Install if desired to fix the above problems. ************************************************************************** **************************************************************************** 1/18/96 New relocatables and load files to allow linking non-EMA/VMA LI. These files also correct problem of non-EMA/VMA LI only running once on a memory-based system. PROBLEM DESCRIPTION: SR#4701302190 LI fails to link as non-EMA/VMA without LI_VMA.REL PROBLEM DESCRIPTION SR#5003282590 LI only runs once in a memory-based system FIX DESCRIPTION: There are two new relocatable files for non-EMA/VMA LI, LI_NONEMA.REL and LI_CDS_NONEMA.REL. These files use the non-EMA/VMA versions of the routines contained in LI_VMA.REL and LI_VMA.RELC. Additionally, a fix has been made to LI.REL and LI_CDS.REL to allow multiple runs of LI on a memory-based system. LI.LOD (for RTE-A) and LI_CDS.LOD (for VC+) have been updated to account for these changes, including changes to the documentation describing how to implement the non-EMA/VMA load. INSTALLATION: Copy the new relocatables and altered load files to the appropriate product directories. For a non-CDS system, the new LI.LOD and LI_NONEMA.REL should be placed in your RTE-A product directory; for a CDS system, LI_CDS.LOD and LI_CDS_NONEMA.REL should be place in your VC+ product directory. Follow the edit instructions in the appropriate load file to link a non-EMA/VMA version of LI, commenting out the appropriate command lines. Now it is just a matter of following the standard steps for loading system files (such as using RTE_INSTALL.CMD). FILES: LI_NONEMA.REL REV.6201 960118 LI.LOD REV.6201 960118 OS Revs Supported: 6.1 and 6.2 NOTE: The CDS versions of these files can be found in the VCPLUS patch directory. **************************************************************************** 10/19/95 New relocatables for the LP spooler correcting the hostname greater than 8 character and/or containing a "." in the hostname. PROBLEM DESCRIPTION: SR#5003262790 LP fails to print local files if remote hostname is >8 characters FIX DESCRIPTION: The new LP code will handle remote names up to the first 10 characters. If the first ten characters of the remote hostname contains a ".", then the spooler will replace the "." with a ")". For example, the remote hostname of bones2.mayfield.hp.com will be identified as bones2)may in the ID field of the spool job. INSTALLATION: Replaced all the old relocatables in the LP directory with the new relocatables and excute the MKLP command to reload all the programs. Be sure to set the CI variables needed for the MKLP command. FILES: lp.read (This file) lp.rel rev.6201 rlpdaemon.rel rev.6201 lpstat.rel rev.6201 lpsched.rel rev.6201 lpout.rel rev.6201 lpmove.rel rev.6201 cancel.rel rev.6201 OS Revs Supported: 6.1 and 6.2 Others: Keep factory inform if this corrected the problem. **************************************************************************** This file describes all of the patches in the directories /patches/r6200/ns1000/@ Group Directory: /PATCHES/R6100/NS1000/LIB/ NSSYS_CDS.LIB 91790-12013 6201 950524 A Directory: /PATCHES/R6100/NS1000/REL/ FSRV.REL 91790-16339 6202 950725 C FTP.REL 98170-16046 6204 951219 B INPRO.REL 91790-16087 6204 950524 A NOTE: The files followed by a group designation letter contain interdependent changes. To install a patch which uses any one of the files in a group, you must retrieve all the files in the group. Then all programs associated with the group (listed below) should be linked. Group A: INPRO and OUTPRO Group B: FTP Group C: FSRV ************************************************************************** 23 May 1995 SR# 4701293084 NFS file access fails for some files in a large directory PROBLEM TEXT: NFS file access fails when the file's directory entry is in an extent with a disk block address that is less than the address of the directory's main. A stale file handle error is returned to the client. FILES: REL/FSRV.REL REV.6201 950523 INSTALLATION: Place new FSRV.REL in /NS1000/REL. Dismount the RTE-A file systems from NFS client systems. Relink FSRV. OF,FSRV WD /NS1000 LINK LOD/FSRV.LOD /PROGRAMS/FSRV.RUN Restart FSRV. FSRV [options] The RTE-A file systems can again be mounted by NFS clients. **************************************************************************** 24 May 1995 FTP / TCP problems SR# 5003263681 FTP can hang occasionally during put/send command PROBLEM TEXT: The FTP user program hangs occasionally when executing a put or send command. When this occurs, the FTP program is RN suspended. NSINF will show that the FTP data connection is in the fin-wait-1 state, but the remote system indicates that the connection is in the established state. This problem only occurs for files that are less than 16 Kbytes. SR# 4701292623 TCP can indicate that remote side closed connection too early PROBLEM TEXT: The TCP code within inpro can occasionally signal that the remote system has closed a connection before all the data has arrived. If this happens during an FTP file transfer, the destination file will be missing data from the end of the file. This problem exists in the Rev.6106 patch version of INPRO.REL and the 6.2 version. FILES: LIB/NSSYS_CDS.LIB REV.6201 950524 REL/INPRO.REL REV.6201 950524 INSTALLATION: Both INPRO and OUTPRO must be relinked for this patch. Place new NSSYS_CDS.LIB in /NS1000/LIB and INPRO.REL in /NS1000/REL. Shutdown NS-ARPA with NSINIT. OF INPRO ID OF OUTPRO ID WD /NS1000 LINK LOD/INPRO.LOD /PROGRAMS/INPRO.RUN LINK LOD/OUTPRO.LOD /PROGRAMS/OUTPRO.RUN RP /PROGRAMS/INPRO.RUN RP /PROGRAMS/OUTPRO.RUN Restart NS-ARPA with NSINIT. *************************************************************************** 24 May 1995 SR# 4701292631 FTP can hang during get/recv command PROBLEM TEXT: FTP can occasionally hang while executing a get or recv command. It stops after reporting the following messages. ftp> get test3.txt 200 PORT command successful. 150-Opening BINARY mode data connection for 150 TEST3.TXT:::4:2:128 226 Closing data connection. At this point, FTP is RN suspended. NSINF shows the FTP data connection in the close-wait state. FILES: REL/FTP.REL REV.6201 950524 INSTALLATION: Link FTP as follows. Place new FTP.REL (and FTPULIB.REL) in /NS1000/REL. Be sure that FTP is not being used. WD /NS1000 LINK LOD/FTP.LOD /PROGRAMS/FTP.RUN ****************************************************************************** 11 June 1995 SR# 1653132522 ICMP rerouting fails on HP1000 PROBLEM TEXT: ICMP redirect messages are not processed correctly and can cause path records to be corrupted. ICMP redirect messages are sent by gateways to advise a source node that a different gateway should be used to send IP datagrams to a remote host or network. This problem can cause any of the following event log error messages. Error IEEE-802 14008 156 257 171b OUTPR/0 4004 Error IEEE-802 7003 258 0 171b INPRO/0 4004 Error IEEE-802 14006 4004 0 171b OUTPR/0 16 19 449 7 4004 FILES: REL/INPRO.REL REV.6201 950524 INSTALLATION: See note at the beginning of this file regarding the requirements for installing previous patches. INPRO must be relinked for this patch. Place new INPRO.REL in /NS1000/REL. Shutdown NS-ARPA with NSINIT. OF INPRO ID WD /NS1000 LINK LOD/INPRO.LOD /PROGRAMS/INPRO.RUN RP /PROGRAMS/INPRO.RUN Restart NS-ARPA with NSINIT. ****************************************************************************** 25 July 1995 SR# 4701297473 FSRV can hang on instruction with infinite indirect addressing PROBLEM TEXT: FSRV can hang executing an instruction with an infinite loop of indirect addresses. This problem occurs intermittently when FSRV reads or searches directories. FILES: REL/FSRV.REL REV.6202 950725 INSTALLATION: Place new FSRV.REL in /NS1000/REL. Dismount the RTE-A file systems from NFS client systems. Relink FSRV. OF,FSRV WD /NS1000 LINK LOD/FSRV.LOD /PROGRAMS/FSRV.RUN Restart FSRV. FSRV [options] The RTE-A file systems can again be mounted by NFS clients. ****************************************************************************** 11 October 1995 SR# 4701304873 Gateway does not forward IP datagrams for some destination addresses PROBLEM TEXT: A system connected to more than one network, a gateway, will not forward IP datagrams if the third byte of the destination IP address is zero. This problem occurs in 6.1 and 6.2 systems. FILES: REL/INPRO.REL REV.6202 951011 INSTALLATION: See note at the beginning of this file regarding the requirements for installing previous patches. INPRO must be relinked for this patch. Place new INPRO.REL in /NS1000/REL. Shutdown NS-ARPA with NSINIT. OF INPRO ID WD /NS1000 LINK LOD/INPRO.LOD /PROGRAMS/INPRO.RUN RP /PROGRAMS/INPRO.RUN Restart NS-ARPA with NSINIT. ****************************************************************************** 21 November 1995 SR# 5003290015 No response to TCP segments with more than 4 bytes of options. PROBLEM TEXT: INPRO does not respond to TCP segments with more than 4 bytes of options. Nothing is written in the event log file when this occurs. This problem caused TCP connection requests from a system running OSF/1 to be ignored. Connection from other systems which use newer TCP options may also be affected by this problem. FILES: REL/INPRO.REL REV.6203 951121 INSTALLATION: See note at the beginning of this file regarding the requirements for installing previous patches. INPRO must be relinked for this patch. Place new INPRO.REL in /NS1000/REL. Shutdown NS-ARPA with NSINIT. OF INPRO ID WD /NS1000 LINK LOD/INPRO.LOD /PROGRAMS/INPRO.RUN RP /PROGRAMS/INPRO.RUN Restart NS-ARPA with NSINIT. ****************************************************************************** 19 December 1995 SR# 5003296202 FTP memory protects or UI aborts at 6.1 when connected to VAX PROBLEM TEXT: The FTP program aborts if the remote system type name is shorter than 5 characters. The system type is included in the reply to the FTP SYST command. Remote system types "RTE-A" and "UNIX Type: L8" will work, but shorter names such as "VMS" will cause FTP to abort. The abort occurs after FTP has connected to the remote system and the user has entered a username and password. FILES: REL/FTP.REL REV.6204 951219 INSTALLATION: Link FTP as follows. Place new FTP.REL (and FTPULIB.REL) in /NS1000/REL. Be sure that FTP is not being used. WD /NS1000 LINK LOD/FTP.LOD /PROGRAMS/FTP.RUN ****************************************************************************** 31 May 1996 SR# 4701-325415 The HP 1000 networking software drops TCP segments with a checksum PROBLEM TEXT: INPRO will drop a packet if the TCP checksum is 0. Since 0 is a valid TCP checksum, this should not occur. SR# 4701320085 PROBLEM TEXT: The program INPRO is logging error messages in the NS_event.log which are undefined in the Error Messages and Recovery Manual. Examples of the error messages are given below. Error PROSW 2 50 22461 161b INPRO/0 677 20 385 0 1 0 1 0 0 3840 22695 3840 22913 0 0 0 0 0 0 0 Error HP-IP 3854 0 0 161b INPRO/0 -1 FILES: REL/INPRO.REL REV.6204 950524 INSTALLATION: See note at the beginning of this file regarding the requirements for installing previous patches. INPRO must be relinked for this patch. Place new INPRO.REL in /NS1000/REL. Shutdown NS-ARPA with NSINIT. OF INPRO ID WD /NS1000 LINK LOD/INPRO.LOD /PROGRAMS/INPRO.RUN RP /PROGRAMS/INPRO.RUN Restart NS-ARPA with NSINIT. ******************************************************************** **************************************************************************** Date: 23 June 1995 SR # 5000094003 PROMT and CM are hardcode to priority 2 and 3. ENHANCEMENT TEXT: Enhancement request to be able to set priority of PROMT and CM. This enhancement will be done in the next PCO (6.21). PROMT and CM should be high priority programs, but the capability to reset the priorities is a reasonable expectation. FILES: %PROMT REV.6201 950623 %PM000 REV.6201 950623 INSTALLATION: Relink PROMT using #PROMT from the VC+ product directory. You will have to OF the system's current copy of PROMT to replace it and then reboot your system. OS Revs Supported: 6.1 and 6.2 USING THE NEW FEATURE: To change the priority of PROMT, RP it and set its priority with the PR command in the welcome file. Alternatively, you can use the LINK LK command to reset PROMT's priority permanently. INSTALLATION RECOMMENDATION: Install only if you need to lower the priorities of PROMT, LOGON, and CM for realtime considerations. In most cases, it is important for promt, logon, and cm to have high priorities so that they can inter- rupt processes and allow user control. **************************************************************************** 11/10/95 New relocatables and Absolute code for the STDVR in the diagnostic package 24612A correcting the error which fails to read in the diagnostic on DAT tape. PROBLEM DESCRIPTION: SR#5003265769 24612A diagnostic fail with STDVR error 9 on A990 using 2 Gbyte DDS. INSTALLATION: Replaced the old files of !stdvr and %stdvr with the new rev. 6201 version of !stdvr and %stdvr. FILES: stdvr.read (This file) !stdvr rev.6201 %stdvr rev.6201 OS Revs Supported: 6.1 6.2 Others: Keep factory inform if this corrected the problem. **************************************************************************** **************************************************************************** 2/20/96 New @.HEX files for use in VSCSI down-load tests to correct intermittent DVR error 28 problems. PROBLEM DESCRIPTION: SR#4701-313809 VSCSI down-load test reports error 28 An intermittent SCSI driver error 28 can be reported after a VSCSI down-load test, such as the ZRAMTST.HEX or ZLPBK.HEX test. After the error is reported the driver "downs" the disk LU and an up command then recovers from the error. The failure rate is low, it usually takes more than 1 day of looping on the test before the error is reported. In cases VSCSI looped for more than 4 days without reporting an error. FIX DESCRIPTION: There are two new files, ZRAMTST.HEX and ZLPBK.HEX, that correct the problem. INSTALLATION: Copy the new @.HEX files and replace the old files in the /PROGRAMS/ directory. FILES: ZRAMTST.HEX REV.6201 ZLPBK.HEX REV.6201 OS Revs Supported: 6.1 and 6.2