home *** CD-ROM | disk | FTP | other *** search
Text File | 1994-04-13 | 49.6 KB | 1,904 lines |
- *******************************************************************
- ** IMPORTANT: Please immediately replace the README.TXT file **
- ** on the DOS PATCH and OS/2 PATCH 1 diskettes with this file, **
- ** which contains essential, updated information. **
- *******************************************************************
-
- ------------------------------------------
- LAN Manager 2.2B--Patch to LAN Manager 2.2
- ------------------------------------------
-
-
- CONTENTS
- ========
-
- 1.0 Introduction
- 2.0 LAN Manager 2.2B Patch Installation Instructions
- 2.1 User Notes
- 2.2 Bugs Fixed in the LAN Manager 2.2B Release
- 2.3 Text of Bug Fixes
-
-
- 1.0 INTRODUCTION
- ================
-
- This release of LAN Manager 2.2B contains all components that have
- been fixed since the original release of LAN Manager 2.2.
-
-
- 2.0 LAN Manager 2.2B Patch Installation Instructions
- ====================================================
-
- For OS/2
- --------
-
- A. Verify you have backed up all of your important data before you begin
- the update.
-
- B. Verify disk integrity.
- 1. run CHKDSK on fat partitions
- 2. run CHKDSK /F:3 on HPFS partitions
-
- C. Backup your NET.ACC and ACLs. The detach/attach process
- handles this procedure automatically. However, it is helpful
- to do this manually to maintain current backups of the user accounts
- database as well as the permissions.
-
- 1. Run BACKACC x:\ /S. This will save your ACLs in a file
- called ACLBAKx.ACL. This file is located in <lanroot>\ACCOUNTS.
- Do this for every drive where x: represents the drive letter.
-
- D. Complete the following steps:
- 1. Type NET STOP WKSTA at an OS/2 command prompt.
- 2. Detach LAN Manager using the LAN Manager Setup utility on the
- hard disk.
- 3. Reboot the system.
- 4. Once the system has rebooted, insert LAN Manager 2.2B OS/2 Patch
- Disk 1 into the floppy drive.
- 5. At an OS/2 command prompt, change to that floppy drive (that is,
- A: or B:) and type SETUP.
- 6. Follow the instructions given on the screen.
- 7. If needed, install the OS2KRNL, OS2LDR and OS/2 drivers and files manually.
-
-
- For MS-DOS
- ----------
-
- 1. Insert LAN Manager 2.2B DOS Patch Disk 1 into the floppy drive.
- 2. At an MS-DOS command prompt, change to that floppy drive (that is,
- A: or B:) and type SETUP.
- 3. Follow the instructions given on the screen.
-
-
- 2.1 User Notes
- --------------
-
- Note One: Problem Loading LM 2.2B NETWKSTA.EXE into MS-DOS 6.2 UMBs
- --------------------------------------------------------------------
-
- When you install the LAN Manager 2.2B Patch onto an MS-DOS 6.2
- workstation, NETWKSTA.EXE will not correctly loaded into the upper
- memory blocks (UMBs) since LAN Manager initialization incorrectly
- processes the version number of MS-DOS 6.2.
-
- Make sure that there is adequate contiguous UMB space to accommodate
- the redirector. Do not load SMARTDRV or other drivers and utilities
- into the UMB prior to starting the workstation. Many personal
- computers have UMB space available in the E000 segment (verify that
- ROM and shared memory devices are not located in this region). To make
- this memory region available, use I=E000-EFFF as an option for
- EMM386.EXE in CONFIG.SYS. If you do not need expanded memory (EMS), use
- NOEMS as an EMM386.EXE option to save the 64k page frame for EMS
- support.
-
- Example line in CONFIG.SYS
-
- Device=C:\WINDOWS\EMM386.EXE NOEMS I=E000-EFFF
-
- Microsoft has confirmed this to be a problem and is testing a possible
- fix. Contact Microsoft Product Support Services for its status and
- availability. New information will be posted in the Microsoft
- Knowledge Base as it becomes available.
-
-
- Note Two: MS-DOS Low Density Install Diskettes not Available
- -------------------------------------------------------------
-
- The low density MS-DOS client patch install diskettes are not
- available.
-
- The following is a procedure for installing the Patch on a network workstation
- with a 360k floppy drive(s).
-
- 1. NET SHARE a directory on a network file server.
- 2. Copy all files on the LAN Manager 2.2B MS-DOS Patch high density
- disk to the shared directory.
- 3. Connect to this share from the workstation that you want to
- upgrade with the NET USE command.
- 4. Apply the patch to the workstation by running SETUP.EXE from the
- shared directory.
-
-
- Note Three: OS/2 Drivers and Files Included with Patch
- -------------------------------------------------------
-
- Once the upgrade process completes, and you determine that it is necessary
- to install any of the OS/2 drivers or system files, you must do so manually.
- The following paragraphs describe how to install these files.
-
- The files are (located on OS/2 DISK 2):
- OS2KRNL
- OS2LDR
- PMSPL.DLL
- PMSPOOL.EXE
- PSCRIPT.DRV
-
- These files are not automatically installed. To install these files on
- an OS/2 computer, you must manually replace them. Use the File Manager
- to change attributes, rename, and copy these files.
-
- Note: You must use capital letters only when copying these files. If this
- is not done, the server will fail to boot.
-
- The following steps detail how to replace the OS/2 drivers and system files:
-
- 1. Open File Manager.
- 2. Choose View, then Include, then File Flags. Select Show Read-Only,
- Archive, Hidden, and Other.
- 3. Choose the C: drive, then highlight OS2KRNL.
- 4. Choose File, Change Flags, deselect the Read Only, Archive, Hidden,
- and System flags, and click on Change.
- 5. Rename the file OS2KRNL to OS2KRNL.OLD.
- 6. Copy the new OS2KRNL from the floppy disk, and highlight it in File
- Manager.
- 7. Choose File, then choose Change Flags. Select the Read Only,
- Archive, Hidden, and System flags, then click on Change.
- 8. Repeat the steps 2-7 to replace the OS2LDR.
- 9. Reboot using the LAN Manager HPFS Recovery Disk 1
- 10. Replace the Recovery Disk with OS/2 Disk 2.
- 11. Copy PMSPL.DLL to c:\os2\dll\ subdirectory.
- a. From the A:\ use the command: DECOMP -f pmspl.dl$ c:\os2\dll\pmspl.dll
- 12. Copy PMSPOOL.EXE to c:\os2 subdirectory.
- a. From the A:\ use the command: DECOMP -f pmspool.ex$ c:\os2\pmspool.exe
- 13. Copy PSCRIPT.DRV to c:\Lanman\drivers subdirectory.
- a. From the A:\ use the command: DECOMP -f pscript.ex$ c:\lanman\drivers\pscript.drv
-
- 14. The following third-party .BID drivers are also supplied with this patch
- (located on OS/2 DISK 2):
- AHA174X
- CPQC710
- ULTRA24
-
- Copy the above drivers (or the specific files needed) to c:\
- a. copy a:\*.bid c:\
- 15. Shutdown and Reboot.
-
- These drivers are developed by third-party vendors independent of
- Microsoft. We make no warranty, implied or otherwise, regarding these
- drivers' performance or reliability. In addition, you may need to upgrade
- other drivers if these drivers have been updated by the manufacturer.
-
- Special Note:
- If your server traps while copying the files manually, it is likely
- you attempted to copy over a driver which is currently in use. Go to
- step 9 above.
-
-
- Note Four: More Information on Upgrading MS-DOS Workstations
- -------------------------------------------------------------
-
- If you encounter the following error during setup:
-
- Unable to read: NETWKSTA
- Error code: 2
-
- This can be corrected by including the following line in your CONFIG.SYS:
- DEVICE=c:<path>\SETVER.EXE
-
- Special Note:
- This upgrade should NOT be performed on Windows For Workgroup clients.
-
-
- Note Five: Named Pipe Server Applications Including Comm Server & SQL Server Fail
- ----------------------------------------------------------------------------------
-
- After the Microsoft LAN Manager 2.2B update is applied to a server
- running DCA/Microsoft Communications Server, Comm Server clients
- may not be able to open an emulation session (such as 3270
- or 5250). When the user tries to open an emulation session, the
- host sign-on screen will not appear and the emulator will hang.
-
- The server does not respond to the client's named pipe read request due
- to a failure in the redirector, netwksta.sys. This file can
- replaced with the LAN Manager 2.2 redirector as a workaround.
- Microsoft has confirmed this to be a problem and is testing a possible
- fix. Contact Microsoft Product Support Services for its status and
- availability. New information will be posted in the Microsoft
- Knowledge Base as it becomes available.
-
-
- 2.2 Bugs Fixed in the LAN Manager 2.2B Release
- ==============================================
-
- - Printsta Fails, Hangs MS-DOS Workstation When Printscreen Starts
- - More Than 32 Domain Controllers on a Domain Causes PDC GP Fault
- - Ill-Formed SMB From Client Causes Server to Die or GP Fault
- - Workstation Hangs After Period of Copying or Running HLAPI Code
- - Net Logon on MS-DOS Client Not Validated by Server
- - Attempt to Stop/Restart PDC Locks out BDC if No One Logged On
- - WinAdmin Creates Too Many ACL Entries, Making NET.ACC Too Big
- - FAT Server Stops Responding if Same Byte Range Repeatedly Locked
- - Replicator Copies Files Already in Sync or for HPFS386 Importer
- - API Call Holds Semaphores After Access Check on Non-Existent File
- - Later Logon Scripts Fail After Time-out Occurs When Logon Running
- - BACKACC Failure from AT Command Required Server Stop/Restart
- - NetBEUI Needs SRB On/Off Switch with 3Com NetBuilder SRT Bridge
- - FTBOOT Causes Trap D/Trap C When Recovering Boot Volume Mirror
- - Core Drive Error Causes Multiple Errors and Data Corruption
- - Failure of Network Request During Logon Hangs Time-out Period
- - MS-DOS Enhanced Workstation Runs Out of Netbufs and Hangs
- - File Seek Fails After Buffer Flushing Error on Seek Initiation
- - Net Password Command During Sync Deadlocks Server and Netlogon
- - WS Hangs Running TCP/IP, DLC and Windows Enh on IBM Token Ring
- - WfW Workstation Cannot Run FoxPro after a LM Workstation Does
- - Application Produces Occasional NET808 and NET805 Errors
- - Windows NetBIOS App Encounters SYS3175 Trap Under OS/2 2.0
- - System Hangs or Reboots Accessing File on Disconnected Drive
- - NET LOGOFF /Y Returns NET2402 if Files Open on a Connection
- - NETWKSTA.EXE Won't Run with MS-DOS 6.0 Without SETVER
- - Password Displayed in Share Info Dialog when Server Share-level
- - More than 300 Entries in RPL.MAP Causes RPLMGR to Trap 000d
- - NET STOP RDR /Y Followed by NET START RDR Hangs MS-DOS Client
- - Period ( . ) Not a Valid Character in Workstation Names
- - Logon Starts Failing after 16 Logons with a Logon Script
- - Running Local and Remote I/O Stress Tests Caused GP Fault
- - DosPeekNmPipe Call Returns Too Much Data, Causes Broken Pipe
- - SQL Queries Under RAS Cause Lock-up, Reboot or EMM Exception #06
- - Full Sync Initiated by Small Netlogon Service Buffer Request
- - LM Setup Does Not Recognize IBM DOS 5.02 as Valid DOS Version
- - Netlogon Service Encounters GP Fault as Array Memory is Overrun
- - DOSPeekNmPipe API Returns Incorrect Byte Count for Its Buffer
- - Issuing CTRL-C as BACKACC Writes to Hard Disk Causes GP Fault
- - OS/2 Server Hangs if Network Adapter Cannot Reset
- - Duplicate Name and Net Address on Token Ring Crashes Server
- - MS-DOS Applications Incorrectly Report "0 Bytes Free"
- - RESTACC Fails with "SYS0111 File Name Too Long" Error
- - Badly Formed Server Message Block Sent to Server Causes Trap
- - Netlogon Performs Continuous Full Synchronization
- - UPS Service Fails Compaq ProSignia if Low Battery Signal Enabled
- - Windows Hangs or Crashes to MS-DOS Prompt If Running Many Apps
- - Workstation Hangs with NET 805, Server Records Incomplete NCB Error
- - Trap D Occurs With HPFS386 when ATTRIB *.TXT /S is Entered
- - Trap D in TCPNB$
- - DOS TCP/IP Clients get Logged on Standalone Despite DC Responses
- - TCP/IP Server Won't Send Datagrams; BDCs Go Out of Sync
- - BCASTADDR Parameter in PROTOCOL.INI Ignored by OS/2 TCP/IP
- - All TCP/IP Servers on Net Trap in NBDRV.0S2
- - Server Crashes and Issues TRAP 0003
- - Sybase Server Doesn't Respond to TCP/IP "Cancel" Commands
- - TCP/IP MS-DOS Utilities Cannot Connect with IBM 3090 MVS
- - Information Packets Incorrectly Reassembled After Crossing Router
- - TCP/IP Returns Wrong Error Code on Failed Call
- - LM Server Trap D's in TCPDRV When Sent LLC XID
- - TCP/IP Server Routes Traffic from Token Ring to Ethernet, Crashes
- - Error 53--TCPCONNECTIONS Depleted, Connection Denied to Client
-
-
- 2.3 Text of Bug Fixes
- ---------------------
-
- ----------------------------------------------------------------
- Printsta Fails, Hangs MS-DOS Workstation When Printscreen Starts
- ----------------------------------------------------------------
-
- CSD: LM22.CSD01.037
-
- SYMPTOMS
- ========
-
- Printsta failed. If the user on an MS-DOS workstation attempted to
- perform print screen operations, the workstation operated properly
- until printscreen was initiated, then would hang.
-
- RESOLUTION
- ==========
-
- The timer stack was getting corrupted, so the size of the stack was
- increased. Also, some of the retry logic was changed so that on slow
- PCs requests no longer time out.
-
- Binary Affected
- ---------------
-
- PRINTSTA.EXE
-
-
- ---------------------------------------------------------------
- More Than 32 Domain Controllers on a Domain Causes PDC GP Fault
- ---------------------------------------------------------------
-
- CSD: LM22.CSD01.040
-
- SYMPTOMS
- ========
-
- The addition of more than 32 domain controllers (primary and backup)
- on a domain caused a GP Fault on the primary domain controller.
-
- RESOLUTION
- ==========
-
- The number of combined members (backup and primary domain controllers)
- allowed in a domain was changed from 32 to 128. A check was also added
- that keeps the primary domain controller from faulting if more than
- 128 domain controllers are added to the domain (although an error
- still is logged).
-
- Binary Affected
- ---------------
-
- NETLOGON.EXE
- NETAPI.DLL
-
-
- -----------------------------------------------------------
- Ill-Formed SMB From Client Causes Server to Die or GP Fault
- -----------------------------------------------------------
-
- CSD: LM22.CSD01.041
-
- SYMPTOMS
- ========
-
- An ill-formed server message block from a client sometimes caused the
- server to either GP fault or "go dead."
-
- RESOLUTION
- ==========
-
- The manner in which ill-formed server message blocks are handled was
- changed. Specifically, code was added that put a null at the end of
- their truncated or corrupted strings.
-
- Binary Affected
- ---------------
-
- HPFS386.IFS
- HPFS.386
-
-
- ---------------------------------------------------------------
- Workstation Hangs After Period of Copying or Running HLAPI Code
- ---------------------------------------------------------------
-
- CSD: LM22.CSD01.042
-
- SYMPTOMS
- ========
-
- After 15 minutes of activity such as copying files between a server
- and a workstation or running Attachmate HLAPI code that polls the
- server, the workstation would hang.
-
- RESOLUTION
- ==========
-
- The MS-DOS redirector stack was overflowing. Stack size was increased.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- --------------------------------------------------
- Net Logon on MS-DOS Client Not Validated by Server
- --------------------------------------------------
-
- CSD: LM22.CSD01.043
-
- SYMPTOMS
- ========
-
- A net logon on an MS-DOS client sometimes resulted in a standalone
- logon--the server refused to validate it.
-
- RESOLUTION
- ==========
-
- If the name-query packet was larger than usual (that is, if an
- additional field in the resource record was filled in) then a routine
- in TINYRFC.EXE failed to free a buffer. The code was modified to
- assure that the buffer is freed.
-
- Binary Affected
- ---------------
-
- TINYRFC.EXE
-
-
- -------------------------------------------------------------
- Attempt to Stop/Restart PDC Locks out BDC if No One Logged On
- -------------------------------------------------------------
-
- CSD: LM22.CSD01.044
-
- SYMPTOMS
- ========
-
- When no one was logged on at the backup domain controller and the
- primary domain controller was stopped and started, the backup
- attempted to connect to the primary using its computer name and some
- unknown password. This repeated, causing two effects:
-
- - the account was locked out (if lockout was set up)
-
- - the audit log was filled with spurious bad password attempts
-
- Netlogon synchronization was not affected.
-
- RESOLUTION
- ==========
-
- Code was changed so that a backup domain controller no longer attempts
- a single API if there is no one logged on: it simply gets the next
- request. This does not lessen functionality, as these calls would fail
- in any case, but it makes the failures local so they are not logged.
-
- Binary Affected
- ---------------
-
- NETLOGON.EXE
-
-
- -------------------------------------------------------------
- WinAdmin Creates Too Many ACL Entries, Making NET.ACC Too Big
- -------------------------------------------------------------
-
- CSD: LM22.CSD00.002
-
- SYMPTOMS
- ========
-
- WinAdmin's "permit tree" feature incorrectly walked the directory tree
- and created ACL entries for each file and for its subdirectory,
- quickly causing NET.ACC to become huge.
-
- RESOLUTION
- ==========
-
- Code was changed to create ACL entries only for subdirectories and to
- delete any existing ACL entries for files. The change also displays
- both explicit and inherited permissions.
-
- Binary Affected
- ---------------
-
- NETADMIN.EXE
-
-
- ----------------------------------------------------------------
- FAT Server Stops Responding if Same Byte Range Repeatedly Locked
- ----------------------------------------------------------------
-
- CSD: LM22.CSD00.003
-
- SYMPTOMS
- ========
-
- When the same byte-range was repeatedly locked, a FAT server stopped
- responding after some number of lock attempts.
-
- RESOLUTION
- ==========
-
- Code was changed to correct time-out determination.
-
- Binary Affected
- ---------------
-
- NETSERVR.EXE
-
-
- ---------------------------------------------------------------
- Replicator Copies Files Already in Sync or for HPFS386 Importer
- ---------------------------------------------------------------
-
- CSD: LM22.CSD00.004
-
- SYMPTOMS
- ========
-
- The replicator copied files already in sync from subdirectories that
- contained large numbers of files, and when the importer had an HPFS386
- disk.
-
- RESOLUTION
- ==========
-
- A larger buffer size was implemented in the replicator code.
-
- Binary Affected
- ---------------
-
- REPLCLI.EXE
-
-
- -----------------------------------------------------------------
- API Call Holds Semaphores After Access Check on Non-Existent File
- -----------------------------------------------------------------
-
- CSD: LM22.CSD00.005
-
- SYMPTOMS
- ========
-
- Checking the access on a file that did not exist (on a FAT partition)
- caused the API call to take but not release three semaphores.
-
- RESOLUTION
- ===========
-
- Code was changed so that it now sets the correct error and frees the
- semaphores before returning.
-
- Binary Affected
- ---------------
-
- NETAPI.DLL
-
-
- -----------------------------------------------------------------
- Later Logon Scripts Fail After Time-out Occurs When Logon Running
- -----------------------------------------------------------------
-
- CSD: LM22.CSD00.006
-
- SYMPTOMS
- ========
-
- If a Windows logon script was still running when the logon script
- time-out occurred, all subsequent logon scripts failed until Windows
- was exited.
-
- RESOLUTION
- ==========
-
- A variable that checks the time-out in a DLL was set to initialize
- each time the routine is called.
-
- Binary Affected
- ---------------
-
- NETAPI.DLL (Win16 version)
-
-
- ------------------------------------------------------------
- BACKACC Failure from AT Command Required Server Stop/Restart
- ------------------------------------------------------------
-
- CSD: LM22.CSD00.007
-
- SYMPTOMS
- ========
-
- BACKACC failed to complete when run from the AT command, and could not
- be run by AT again until the server was stopped and started.
-
- RESOLUTION
- ==========
-
- The server had too many open files, since remote APIs were not closing
- the UASCHG.LOG file once it was opened. Code was changed to close this
- file after each open.
-
-
- Binary Affected
- ---------------
-
- NETAPI.DLL
-
-
- ---------------------------------------------------------------
- NetBEUI Needs SRB On/Off Switch with 3Com NetBuilder SRT Bridge
- ---------------------------------------------------------------
-
- CSD: LM22.CSD00.008
-
- SYMPTOMS
- ========
-
- With the 3Com NetBuilder SRT Bridge, NetBEUI required a switch to turn
- instant flipping of the source routing bit on and off.
-
- RESOLUTION
- ==========
-
- Added [NETBEUI_XIF] parameter, LocalRing, which is 0 by default. When
- this is zero, the source routing bit is turned on immediately. When it
- is 1, it allows the packet to traverse the local ring three times
- before the SRB is turned on.
-
- Binary Affected
- ---------------
-
- NETBEUI.*
-
-
- --------------------------------------------------------------
- FTBOOT Causes Trap D/Trap C When Recovering Boot Volume Mirror
- --------------------------------------------------------------
-
- CSD: LM22.CSD00.010
-
- SYMPTOMS
- ========
-
- FTBOOT caused a Trap D or Trap C when trying to recover the boot
- volume mirror.
-
- RESOLUTION
- ==========
-
- Code was changed to clear the extended partition record if the boot
- volume mirror is the only partition on its disk.
-
- Binary Affected
- ---------------
-
- FTBOOT.EXE
-
-
- -----------------------------------------------------------
- Core Drive Error Causes Multiple Errors and Data Corruption
- -----------------------------------------------------------
-
- CSD: LM22.CSD00.011
-
- SYMPTOMS
- ========
-
- The core drive issued an error code which the BID handled incorrectly,
- causing multiple errors and data corruption. This affected other SCSI
- drives attached to an IBM SCSI adapter, not just the core drives.
-
- RESOLUTION
- ==========
-
- This was fixed by causing the BID to set a timer and simulate an
- interrupt, which allows it to handle the error condition correctly.
-
- Binary Affected
- ---------------
-
- ABIOS.BID
-
-
- -------------------------------------------------------------
- Failure of Network Request During Logon Hangs Time-out Period
- -------------------------------------------------------------
-
- CSD: LM22.CSD00.012
-
- SYMPTOMS
- ========
-
- In Windows, failure of a network request during a logon attempt (such
- as attempting to connect to a down server) caused the time-out period
- to hang for an excessively long time.
-
- RESOLUTION
- ==========
-
- Changed the logic to use a count to check for the elapsed time since
- the dialog was started and to abort the logon script based on this.
-
- Binary Affected
- ---------------
-
- NETAPI.DLL (Win16 version)
-
-
- ---------------------------------------------------------
- MS-DOS Enhanced Workstation Runs Out of Netbufs and Hangs
- ---------------------------------------------------------
-
- CSD: LM22.CSD00.014
-
- SYMPTOMS
- ========
-
- MS-DOS enhanced workstations would run out of netbufs and hang.
-
- RESOLUTION
- ==========
-
- Code was changed so that the terminating loop case no longer became
- incorrect. The MS-DOS redirector is now guaranteed not to run out of
- work buffers.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- --------------------------------------------------------------
- File Seek Fails After Buffer Flushing Error on Seek Initiation
- --------------------------------------------------------------
-
- CSD: LM22.CSD00.015
-
- SYMPTOMS
- ========
-
- File seeks failed if there was an error flushing a buffer when the
- seek was initiated.
-
- RESOLUTION
- ==========
-
- This was fixed by correcting the stack manipulation.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- --------------------------------------------------------------
- Net Password Command During Sync Deadlocks Server and Netlogon
- --------------------------------------------------------------
-
- CSD: LM22.CSD00.016
-
- SYMPTOMS
- ========
-
- Issuance of a net password command while a domain controller (backup
- or primary) was performing a full sync caused a deadlock that kept
- part of the server and part of Netlogon from functioning. Domain
- controllers made no progress syncing, the net password command stopped
- functioning, and the Netlogon process that gets and responds to
- requests would hang.
-
- RESOLUTION
- ==========
-
- The code was changed to fix the logic code so that the Netlogon
- process releases its semaphore before requesting the server's
- semaphore.
-
- Binary Affected
- ---------------
-
- NETAPI.DLL
-
-
- --------------------------------------------------------------
- WS Hangs Running TCP/IP, DLC and Windows Enh on IBM Token Ring
- --------------------------------------------------------------
-
- CSD: LM22.CSD00.017
-
- SYMPTOMS
- ========
-
- Workstations running TCP/IP, DLC, and Windows Enhanced mode with an
- IBM token ring card would randomly hang.
-
- RESOLUTION
- ==========
-
- The stack usage of the token ring driver was changed to alleviate the
- problem.
-
- Binary Affected
- ---------------
-
- IBMTOK.DOS
-
-
- -----------------------------------------------------------
- WfW Workstation Cannot Run FoxPro after LM Workstation Does
- -----------------------------------------------------------
-
- CSD: LM22.CSD00.018
-
- SYMPTOMS
- ========
-
- When a LAN Manager workstation was running a FoxPro application on a
- LAN Manager server, a Windows for Workgroups workstation that tried to
- run the FoxPro application received:
-
- Network Error Reading Drive X:
-
- followed by:
-
- File Read Error
-
- When a Windows for Workgroups workstation initialized the FoxPro
- application with a database file 2K or larger, both the first and the
- second workstations received the error messages.
-
- RESOLUTION
- ==========
-
- The FAT file server was changed so that it now correctly sets two
- fields in a server message block sent to the Windows for Workgroups
- client.
-
- Binary Affected
- ---------------
-
- NETSERVR.EXE
-
-
- --------------------------------------------------------
- Application Produces Occasional NET808 and NET805 Errors
- --------------------------------------------------------
-
- CSD: LM22.CSD00.019
-
- SYMPTOMS
- ========
-
- An application produced occasional NET808 and NET805 errors during
- command processing.
-
- RESOLUTION
- ==========
-
- Buffer flushing code was changed to correct this.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- ----------------------------------------------------------
- Windows NetBIOS App Encounters SYS3175 Trap Under OS/2 2.0
- ----------------------------------------------------------
-
- CSD: LM22.CSD00.020
-
- SYMPTOMS
- ========
-
- A Windows NetBIOS application encountered a SYS3175 trap while issuing
- a listen. The application worked correctly under Win 3.1 but not in
- the Win 3.1 environment under OS/2 2.0.
-
- RESOLUTION
- ==========
-
- OS/2 was not handling the call correctly. A workaround was implemented
- in NETVDD.
-
- Binary Affected
- ---------------
-
- NETVDD.SYS
-
-
- ------------------------------------------------------------
- System Hangs or Reboots Accessing File on Disconnected Drive
- ------------------------------------------------------------
-
- CSD: LM22.CSD00.021
-
- SYMPTOMS
- ========
-
- The system would hang or reboot when trying to access a file in a
- Windows application on a disconnected drive when Windows was being run
- over the net, usually in a token ring environment.
-
- RESOLUTION
- ==========
-
- The program was running out of network buffers. Code was changed to
- add checks to the code path that retrieves buffers.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- -----------------------------------------------------------
- NET LOGOFF /Y Returns NET2402 if Files Open on a Connection
- -----------------------------------------------------------
-
- CSD: LM22.CSD00.022
-
- SYMPTOMS
- ========
-
- The command NET LOGOFF /Y returned the error NET2402 if files were
- open on a connection. Subsequent attempts returned the same error.
-
- RESOLUTION
- ==========
-
- The system file table entry is now marked as invalid, allowing the
- command to complete. Applications receive an error if they try to
- access an open file's handle.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- -----------------------------------------------------
- NETWKSTA.EXE Won't Run with MS-DOS 6.0 Without SETVER
- -----------------------------------------------------
-
- CSD: LM22.CSD00.024
-
- SYMPTOMS
- ========
-
- NETWKSTA required the use of SETVER to run with MS-DOS 6.0.
-
- RESOLUTION
- ==========
-
- Code was changed. NETWKSTA now allows MS-DOS 6.0.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- ---------------------------------------------------------------
- Password Displayed in Share Info Dialog when Server Share-level
- ---------------------------------------------------------------
-
- CSD: LM22.CSD00.026
-
- SYMPTOMS
- ========
-
- Password is displayed in share info dialog when server is share-level.
-
- RESOLUTION
- ==========
-
- This has been corrected so that a read-only clear-text share password
- is displayed if the server is share-level. If you choose OK without
- changing anything, the system deletes the password on the share,
- leaving it blank.
-
- Binary Affected
- ---------------
-
- NETADMIN.EXE
-
-
- -----------------------------------------------------------
- More than 300 Entries in RPL.MAP Causes RPLMGR to Trap 000d
- -----------------------------------------------------------
-
- CSD: LM22.CSD00.027
-
- SYMPTOMS
- ========
-
- If the RPL.MAP file had more than 300 entries, RPLMGR issued Trap 000d
- on startup. After a workstation beyond number 300 was added,
- refreshing the workstation list or exiting and re-entering RPLMGR
- caused the program to Trap 000d.
-
- Session Title:
- RPLMGR.EXE
-
- SYS1943:A program caused a protection violation.
- TRAP 000D
- AX=0296 BX=EAF0 CX=0000 DX=1BF7 BP=1D90
- SI=052A DI=0000 DS=1957 ES=1BCF FLG=2212
- CS=18F7 IP=07D8 SS=01C7 SP=1D8A MSW=001B
- CSLIM=07F2 SSLIM=1FFF DSLIM=0CEF ESLIM=F019
- CSACC=FB SSACC=F3 DSACC=F3 ESACC=F3
- ERRCD=0000 ERLIM=**** ERACC=**
-
- RESOLUTION
- ==========
-
- Memory re-allocation code was modified so that the size of the memory
- block now is calculated correctly.
-
- Binary Affected
- ---------------
-
- RPLMGR.DLL
-
-
- -------------------------------------------------------------
- NET STOP RDR /Y Followed by NET START RDR Hangs MS-DOS Client
- -------------------------------------------------------------
-
- CSD: LM22.CSD00.030
-
- SYMPTOMS
- ========
-
- Performing NET STOP RDR /Y followed by NET START RDR sometimes hung
- an MS-DOS client or caused it to hang a few seconds after returning
- from the commands.
-
- RESOLUTION
- ==========
-
- Code was changed to handle cases wherein no protocols are loaded and
- the network card NDIS driver is sending a packet up to the protocol.
-
- Binary Affected
- ---------------
-
- PROTMAN.DOS
- PROTMAN.EXE
-
-
- -------------------------------------------------------
- Period ( . ) Not a Valid Character in Workstation Names
- -------------------------------------------------------
-
- CSD: LM22.CSD00.031
-
- SYMPTOMS
- ========
-
- Workstation names could not include '.' as a character.
-
- RESOLUTION
- ===========
-
- The list of valid characters for workstation names was changed to
- include the period ( . ).
-
- Binary Affected
- ---------------
-
- NETADMIN.EXE
-
-
- --------------------------------------------------------
- Logon Starts Failing after 16 Logons with a Logon Script
- --------------------------------------------------------
-
- CSD: LM22.CSD00.037
-
- SYMPTOMS
- ========
-
- Under Windows, logging in and out with a logon script 16 times in a
- row caused the logon to start failing.
-
- RESOLUTION
- ==========
-
- The LMSCRIPT.$$$ flag file was being left open after logon script
- processing, so that after about 16 logons too many files were open.
- Changes in code cause the file to be closed now before it is deleted.
-
- Binary Affected
- ---------------
-
- NETAPI.DLL
-
-
- ---------------------------------------------------------
- Running Local and Remote I/O Stress Tests Caused GP Fault
- ---------------------------------------------------------
-
- CSD: LM22.CSD00.038
-
- SYMPTOMS
- ========
-
- Running a local and a remote I/O stress test simultaneously caused a
- GP fault.
-
- RESOLUTION
- ==========
-
- Code was changed to allow HPFS386 and the LM OS/2 redirector to share
- a field in a common structure.
-
- Binary Affected
- ---------------
-
- NETWKSTA.SYS
-
-
- ------------------------------------------------------------
- DosPeekNmPipe Call Returns Too Much Data, Causes Broken Pipe
- ------------------------------------------------------------
-
- CSD: LM22.CSD00.040
-
- SYMPTOMS
- ========
-
- The OS/2 redirector returned more data than an MS-DOS client machine
- could handle in response to a DosPeekNmPipe call. The underlying
- transport returned error 06h (Incomplete Received Message) to the
- MS-DOS redirector on a RECEIVE ANY, causing it to hang the session and
- result in a broken pipe.
-
- RESOLUTION
- ==========
-
- Code was changed to correct circular buffer handling.
-
- Binary Affected
- ---------------
-
- NETWKSTA.SYS
-
-
- ----------------------------------------------------------------
- SQL Queries Under RAS Cause Lock-up, Reboot or EMM Exception #06
- ----------------------------------------------------------------
-
- CSD: LM22.CSD00.041
-
- SYMPTOMS
- ========
-
- During SQL queries under RAS the machine would occasionally lock-up,
- reboot, or report EMM Exception #06.
-
- RESOLUTION
- ==========
-
- The redirector's internal critical stacks were too small and were
- getting overrun when RAS would back up NCB processing. The stacks were
- amended to alleviate the problem.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- ------------------------------------------------------------
- Full Sync Initiated by Small Netlogon Service Buffer Request
- ------------------------------------------------------------
-
- CSD: LM22.CSD00.042
-
- SYMPTOMS
- ========
-
- For groups with large numbers of users, the Netlogon service was not
- asking for a buffer large enough to contain all users, resulting in a
- full sync between domain controllers.
-
- RESOLUTION
- ==========
-
- A larger buffer has been allocated for this request.
-
- Binary Affected
- ---------------
-
- NETLOGON.EXE
-
-
- -------------------------------------------------------------
- LM Setup Does Not Recognize IBM DOS 5.02 as Valid DOS Version
- -------------------------------------------------------------
-
- CSD: LM22.CSD00.044
-
- SYMPTOMS
- ========
-
- The LAN Manager Setup program did not recognize IBM DOS 5.02 as a
- valid DOS version.
-
- RESOLUTION
- ==========
-
- Checks were added for this DOS version.
-
- Binary Affected
- ---------------
-
- SETUP.EXE
-
-
- ---------------------------------------------------------------
- Netlogon Service Encounters GP Fault as Array Memory is Overrun
- ---------------------------------------------------------------
-
- CSD: LM22.CSD00.045
-
- SYMPTOMS
- ========
-
- The Netlogon Service was ending in a GP fault when memory usage for an
- array was overrun.
-
- RESOLUTION
- ==========
-
- Larger allocations were made for the array.
-
- Binary Affected
- ---------------
-
- NETLOGON.EXE
- NETAPI.DLL
-
-
- -------------------------------------------------------------
- DOSPeekNmPipe API Returns Incorrect Byte Count for Its Buffer
- -------------------------------------------------------------
-
- CSD: LM22.CSD00.046
-
- SYMPTOMS
- ========
-
- The DOSPeekNmPipe API returned an incorrect count for the number of
- bytes in its buffer.
-
- RESOLUTION
- ==========
-
- This was caused when the DOSPeekNmPipe API did not initialize
- correctly under certain circumstances. It has been changed so that it
- now initializes the number of bytes read to the correct value.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- -------------------------------------------------------------
- Issuing CTRL-C as BACKACC Writes to Hard Disk Causes GP Fault
- -------------------------------------------------------------
-
- CSD: LM22.CSD00.047
-
- SYMPTOMS
- ========
-
- Issuing CTRL-C as BACKACC was writing to the hard disk caused a GP
- fault.
-
- RESOLUTION
- ==========
-
- The signal handler was changed so that it now correctly releases
- semaphores.
-
- Binary Affected
- ---------------
-
- BACKACC.EXE
-
-
- -------------------------------------------------
- OS/2 Server Hangs if Network Adapter Cannot Reset
- -------------------------------------------------
-
- CSD: LM22.CSD00.048
-
- SYMPTOMS
- ========
-
- COMTOKR.OS2 went into a tight loop if the network adapter was in a
- state where it would not reset, causing the OS/2 server to hang.
-
- RESOLUTION
- ==========
-
- The driver was checking an incorrect flag. Code was changed so that it
- now checks the correct one.
-
- Binary Affected
- ---------------
-
- COMTOKR.OS2
-
-
- ------------------------------------------------------------
- Duplicate Name and Net Address on Token Ring Crash Server
- ------------------------------------------------------------
-
- CSD: LM22.CSD00.049
-
- SYMPTOMS
- ========
-
- A server would crash if it had the same NetBIOS name, local session
- number and locally administered net address as a workstation on the
- network.
-
- RESOLUTION
- ==========
-
- Since incorrectly administered workstations can cause problems on the
- network, the fix centered on detecting this situation and causing the
- offending workstation to hang.
-
- Binary Affected
- ---------------
-
- NETBEUI.EXE
- NETBEUI.OS2
-
-
- -----------------------------------------------------
- MS-DOS Applications Incorrectly Report "0 Bytes Free"
- -----------------------------------------------------
-
- CSD: LM22.CSD00.050
-
- SYMPTOMS
- ========
-
- On server drive geometries exceeding 2 GB, MS-DOS applications would
- report that there was no free space left on the disk, or "0 bytes
- free."
-
- RESOLUTION
- ==========
-
- MS-DOS applications (such as DIR) do 16-bit multiplication to
- determine the free space on the server disk, and on server drive
- geometries exceeding 2 GB the cluster size becomes to big to be
- contained in a 16-bit number. Checks were installed to prevent
- returning more information to MS-DOS clients than can fit into a 16-
- bit number.
-
- Binary Affected
- ---------------
-
- HPFS.386
-
-
- -----------------------------------------------------
- RESTACC Fails with "SYS0111 File Name Too Long" Error
- -----------------------------------------------------
-
- CSD: LM22.CSD00.052
-
- SYMPTOMS
- ========
-
- RESTACC failed with a "SYS0111 File Name Too Long" error under the
- following circumstances:
-
- - net access c:\data /a /f:all /s:write;delete;acl
-
- - net access c:\data /a /f:acl /s:write;delete
-
- - net access c:\data /a /f:acl;delete /s:write
-
- - net access c:\data /a /f:acl;delete;write /s:all
-
- followed by:
-
- backacc c: /s
- restacc c: /s
-
- RESOLUTION
- ==========
-
- The method by which the size of the structure is determined has been
- changed.
-
- Binary Affected
- ---------------
-
- NETAPI.DLL
-
-
- ------------------------------------------------------------
- Badly Formed Server Message Block Sent to Server Causes Trap
- ------------------------------------------------------------
-
- CSD: LM22.CSD00.053
-
- SYMPTOMS
- ========
-
- A badly formatted server message block sent to the server, caused the
- following trap:
-
- Trap 000D
- AX=0000 BX =0FF6 CX=0002 DX=0001 BP=EEF2
- SI=D2C4 DI=9BEA DS=0658 ES=0658 FLG=2202
- CS=0908 IP=174C SS=0038 SP=EEAC MSW=001B
- CSLIM=370B SSLIM=EB85 DSLIM=FFFF ESLIM=FFFF
- CSACC=9B SSACC=97 DSACC=93 ESACC=93
-
- RESOLUTION
- ==========
-
- The method by which the size of the incoming server message block is
- determined has been made more robust.
-
- Binary Affected
- ---------------
-
- HPFS.386
-
-
- -------------------------------------------------
- Netlogon Performs Continuous Full Synchronization
- -------------------------------------------------
-
- CSD: LM22.CSD00.054
-
- SYMPTOMS
- ========
-
- If a backup domain controller failed to set its password (which causes
- a new attempt to do so every 15 minutes) and a full sync took longer
- than 15 minutes, Netlogon performed a continuous full sync.
-
- RESOLUTION
- ==========
-
- A check was installed to make sure that a request to change the backup
- domain controller password is not dealt with during a user-accounts
- system update.
-
- Binary Affected
- ---------------
-
- NETLOGON.EXE
-
-
- ----------------------------------------------------------------
- UPS Service Fails Compaq ProSignia if Low Battery Signal Enabled
- ----------------------------------------------------------------
-
- CSD: LM22.CSD00.060
-
- SYMPTOMS
- ========
-
- Uninterruptible power supply (UPS) service failed when connected to
- the serial port of a Compaq ProSignia running Microsoft LAN Manager
- 2.1 or 2.1a and the low battery signal check was enabled in
- LANMAN.INI. The server shut down during initialization even under
- normal UPS operation, usually when the UPS service was started.
-
- RESOLUTION
- ==========
-
- Timing conditions were changed so that the UPS service does not fail
- with high speed machines.
-
- Binary Affected
- ---------------
-
- UPSDRV.OS2
- UPS.EXE
-
-
- --------------------------------------------------------------
- Windows Hangs or Crashes to MS-DOS Prompt If Running Many Apps
- --------------------------------------------------------------
-
- CSD: LM22.CSD00.061
-
- SYMPTOMS
- ========
-
- On a very low memory Windows 3.1 machine, with both Windows 3.1 and
- Windows applications running off of a server on the net, loading many
- applications caused Windows to hang or to crash to an MS-DOS prompt
- when pulling up print dialog boxes or exiting applications.
-
- RESOLUTION
- ==========
-
- Two instructions that prevent page faults (which invalidate the stack)
- from occurring in Windows have been reversed in order so that they now
- function more effectively.
-
- Binary Affected
- ---------------
-
- NETWKSTA.500
- NETWKSTA.401
- NETWKSTA.330
-
-
- -------------------------------------------------------------------
- Workstation Hangs with NET 805, Server Records Incomplete NCB Error
- -------------------------------------------------------------------
-
- CSD: LM22.CSD00.062
-
- SYMPTOMS
- ========
-
- A custom named pipes program, when performing raw block write
- commands, would cause the workstation to hang with a NET 805 error.
- The server would also record a NetBIOS "Incomplete NCB Error" message.
-
- RESOLUTION
- ==========
-
- Code was changed to check that the pipe was not in a closing state
- before performing the operation.
-
- Binary Affected
- ---------------
-
- NETWKSTA.SYS
-
-
- ----------------------------------------------------------
- Trap D Occurs With HPFS386 when ATTRIB *.TXT /S is Entered
- ----------------------------------------------------------
-
- CSD: LM22.CSD00.063
-
- SYMPTOMS
- ========
-
- A TRAP D would occur with HPFS386 when the user typed ATTRIB *.TXT /S
- at the root of a large (greater than 1 GB) partition with a huge tree
- structure.
-
- RESOLUTION
- ==========
-
- An invalid instruction has been changed.
-
- Binary Affected
- ---------------
-
- HPFS.386
-
-
- ----------------
- Trap D in TCPNB$
- ----------------
-
- SYMPTOMS
- ========
-
- The following trap screen would be displayed on the server:
-
- TRAP: 000D
-
- AX=17FE BX=4B54 CX=02CA DX=0000 BP=0744
- SI=0000 DI=0E18 DS=0D80 ES=0D80 FLG=2202
- CS=0DF0 IP=2716 SS=1920 SP=073C MSW=001B
- CSLIM=9850 SSLIM=07FF DSLIM=417A ESLIM=417A
- CSACC=9B SSACC=93 DSACC=93 ESACC=93
- ERRCD=000 ERLIM=**** ERACC=**
- Exception in device driver: TCPNB$
-
- The system detected an internal processing error at location 0228:37CD
-
- RESOLUTION
- ==========
-
- An invalid instruction has been changed.
-
- Binary Affected
- ---------------
-
- NBDRV.OS2
-
-
- ----------------------------------------------------------------
- DOS TCP/IP Clients get Logged on Standalone Despite DC Responses
- ----------------------------------------------------------------
-
- SYMPTOMS
- ========
-
- Occasionally workstations using the TCP/IP protocol stack would be
- logged on standalone, despite getting logon responses from the domain
- controller.
-
- RESOLUTION
- ==========
-
- Buffers controlling datagrams were being exhausted and not being
- recycled.
-
- Binary Affected
- ---------------
-
- TINYRFC.EXE
- TCPTSR.EXE
- TCPDRV.OS2
-
-
- -------------------------------------------------------
- TCP/IP Server Won't Send Datagrams; BDCs Go Out of Sync
- -------------------------------------------------------
-
- SYMPTOMS
- ========
-
- Servers running the TCP/IP protocol would, after a period of time,
- stop sending server announcements, which would in turn cause the
- domain controllers to fall out of sync.
-
- RESOLUTION
- ==========
-
- Code that was not freeing packets has been changed.
-
- Binary Affected
- ---------------
-
- NBDRV.OS2
- TCPDRV.OS2
-
-
- ----------------------------------------------------------
- BCASTADDR Parameter in PROTOCOL.INI Ignored by OS/2 TCP/IP
- ----------------------------------------------------------
-
- SYMPTOMS
- ========
-
- Under OS/2, TCP/IP was ignoring the BCASTADDR parameter in the
- PROTOCOL.INI file.
-
- RESOLUTION
- ==========
-
- This functionality was added.
-
- Binary Affected
- ---------------
-
- NBDRV.OS2
-
-
- -------------------------------------------
- All TCP/IP Servers on Net Trap in NBDRV.0S2
- -------------------------------------------
-
- SYMPTOMS
- ========
-
- A particular packet sent on the network as a broadcast was causing
- servers to crash with the following register information:
-
- CSLIM=9850
- IP=00E0
- BP=0738
- SI=0000
-
- RESOLUTION
- ==========
-
- The software now checks for a null pointer.
-
- Binary Affected
- ---------------
-
- NBDRV.OS2
-
-
- -----------------------------------
- Server Crashes and Issues TRAP 0003
- -----------------------------------
-
- SYMPTOMS
- ========
-
- A server crashed with the following output:
-
- TRAP: 0003
-
- AX=0D60 BX=2242 CX=1000 DX=DEAD BP=071E
- SI=2242 DI=0000 DS=0D68 ES=0DA0 FLG=2246
- CS=0D60 IP=017F SS=1920 SP=071E MSW=001B
- CSLIM=1680 SSLIM=07FF DSLIM=FECF ESLIM=36F7
- CSACC=9B SSACC=93 DSACC=93 ESACC=F3
-
- RESOLUTION
- ==========
-
- An invalid instruction has been changed.
-
- Binary Affected
- ---------------
-
- TCPDRV.OS2
-
-
- ---------------------------------------------------------
- Sybase Server Doesn't Respond to TCP/IP "Cancel" Commands
- ---------------------------------------------------------
-
- SYMPTOMS
- ========
-
- The Sybase SQL server did not respond to "Cancel" commands issued
- through the TCP/IP SQL server gateway.
-
- RESOLUTION
- ==========
-
- Support for the out-of-band data signal was added.
-
- Binary Affected
- ---------------
-
- SOCKETS.EXE
- SOCKETS.OS2
-
-
- --------------------------------------------------------
- TCP/IP MS-DOS Utilities Cannot Connect with IBM 3090 MVS
- --------------------------------------------------------
-
- SYMPTOMS
- ========
-
- Ftp and other TCP/IP MS-DOS DOS utilities cannot connect to IBM 3090
- MVS systems
-
- RESOLUTION
- ==========
-
- MVS will not reply unless the PUSH bit is set on each packet, so the
- ForcePushBit parameter has been added to the [TCPIP] section of the
- PROTOCOL.INI file. It defaults to 0, but when it is set to 1 it causes
- the PUSH bit to be set on every outgoing packet.
-
- Binary Affected
- ---------------
-
- TCPTSR.EXE
-
-
- -----------------------------------------------------------------
- Information Packets Incorrectly Reassembled After Crossing Router
- -----------------------------------------------------------------
-
- SYMPTOMS
- ========
-
- TCP/IP packets that were fragmented as they passed through routers
- were not being correctly reassembled.
-
- RESOLUTION
- ==========
-
- Changed the code to reassemble fragmented packets correctly.
-
- Binary Affected
- ---------------
-
- TCPDRV.OS2
- TCPTSR.EXE
-
-
- ----------------------------------------------
- TCP/IP Returns Wrong Error Code on Failed Call
- ----------------------------------------------
-
- SYMPTOMS
- ========
-
- When a server would go down for a client-server application (in this
- case, Schedule+), TCP/IP would cause the client application to hang.
-
- RESOLUTION
- ==========
-
- The NetBIOS return code was changed so that the application would
- recover gracefully.
-
- Binary Affected
- ---------------
-
- TINYRFC.EXE
-
-
- ----------------------------------------------
- LM Server Trap D's in TCPDRV When Sent LLC XID
- ----------------------------------------------
-
- SYMPTOMS
- ========
-
- A server running TCP/IP was crashing with the following output screen:
-
- TRAP 000D
- AX=0104 BX=0000 CX=0000 DX=07F8 BP=0770
- SI=085C DI=016C DS=07C0 ES=07F8 FLG=2206
- CS=07B0 IP=0FF9 SS=1918 SP=0764 MSW=001B
- CSLIM=98E0 SSLIM=07FF DSLIM=FECF ESLIM=5DCB
- CSACC=9B SSACC=93 DSACC=93 ESACC=F3
- ERRCD=694C ERLIM=**** ERACC=**
-
- RESOLUTION
- ==========
-
- Changed the code so that the packet was not accepted.
-
- Binary Affected
- ---------------
-
- TCPDRV.OS2
-
-
- -----------------------------------------------------------------
- TCP/IP Server Routes Traffic from Token Ring to Ethernet, Crashes
- -----------------------------------------------------------------
-
- SYMPTOMS
- ========
-
- A server running TCP/IP on token ring was crashing with a Trap D when
- traffic was routed between a multi-ringed token ring network and an
- Ethernet network. The IP register was 1629.
-
- RESOLUTION
- ==========
-
- Rather than continuing to re-use a buffer, a fix was installed to
- allocate a new buffer to ensure that space for routing information was
- always available.
-
- Binary Affected
- ---------------
-
- TCPDRV.OS2
-
-
- --------------------------------------------------------------
- Error 53--TCPCONNECTIONS Depleted, Connection Denied to Client
- --------------------------------------------------------------
-
- SYMPTOMS
- ========
-
- A server would run out of TCPCONNECTIONS and issue Error 53 to clients
- that attempted to connect to it.
-
- RESOLUTION
- ==========
-
- Timeout code was added for the TCP/IP FINWAIT-2 state.
-
- Binary Affected
- ---------------
-
- TCPDRV.OS2
-