Page of GC20-1807-7 As Updated April 1, 1981 by TNL GN25-0829
Saving the eMS System Only named systems can be saved. The NAMESYS macro must be used to name
a system. A discussion on creating a named system is found under
"Generatinq Named System" in "Part 2: Control Program (CP)". The DMKSNT module must have been configured (by coding the NAMESYS macro) when CP was qenerated. The DMKSNT module contains the system
name, size of the system, and its real disk location. The CMS system
may be saved by entering the command "SAVESYS name" as the first command
after the IPL command (that is, after the CMS version identification is
displayed), where "name" is the name to be assigned to the saved system.
The CMS S- , D-, and Y disks (and, optionally, the A-disk) should be
mounted and attached to the virtual machine, creating the saved system
before the SAVESYS command is issued. This ensures that the CMS file
directory is saved correctly. The status of this saved system, when activated by a subsequent IPL, is changed as though an IPL of a specific device had occurred. The one
exception to this procedure is the file directory for the system disk,
which is part of the nucleus. When a user IPLs CMS, CP loads the
directory for the CMS residence device from disk into virtual storage. Subsequent crpdates to this directory modify the copy that is in virtual
storage but not the copy that is on disk. To modify the copy that is on
disk, save the CMS system after updating the directory.
The CMSSEG Discontiguous Saved Segment
The CMSSEG discontiguous saved segment contains the CMS modules that
perform the CMS Editor, EXEC, or as simulation functions. These same
modules are also loaded on the S-disk of a CMS virtual machine. When CMSSEG has been generated and is in use during execution of the CMS virtual machine, CMS handles a call to the Editor or EXEC processors
by first searching for the requested Editor or EXEC load modules on all
accessed CMS disks, except the S-disk. CMS next attempts to attach the CMSSEG segment; CMSSEG may not be available, depending upon how the CMS virtual machine was generated. If this is the case, CMS attempts to
load the appropriate modules from the CMS S-disk. To handle as simulation routines, CMS first attempts to attach the CMSSEG segment. If the segment is not available, CMS searches all
accessed disks for the as simulation load modules and loads them into
hiqh user storage if they are found. The as simulation modules are then
kept in storage until CMS is reloaded or until a SET SYSNAMES command is
issued for a valid CMSSEG saved segment.
There is overhead associated with controlling discontiguolls saved
segments and with ensuring their integrity. In small systems, the
overhead associated with the use of the CMSSEG saved segment may not be
offset by the benefits of sharing storage among users. Therefore, the
us-e of CMSS-EGmust be deteI"mined-by the user for- his own- ell1l-LrOnlll-ent. Part 3. Conversational Monitor System (CMS) 313

April 1, 1981 CMSSEG USAGE OPTIONS At system
segment.
contained
Gu generation time, you may choose not to generate the CMSSEG If you choose to use it, information on how to generate it is
in the publication VM/370 Plan!!!M: gnd Once generated, users also have the option of choosing whether or not
to use CMSSEG. The IPL command provides the facility to either use or
not use CMSSEG. When you IPL your CMS virtual machine, you can request
that CMSSEG be used by specifying IPL CMS PARM SEG=CMSSEG SEG=CMMSEG is the default option.
To request that CMSSEG not be used for your virtual machine, specify
an invalid segment name in the 1PL command, for example, IPL CMS PARM SEG=DUMMY When the CMSSEG segment is not loaded, the routines that perform the
Editor, EXEC, and OS simulation functions execute in the CMS user area.
The SET command also can be used to control the use of the CMSSEG segment. If the CMS system is loaded (via IPL) with an invalid CMSSEG segment name specified, DMSSVT is loaded in the CMS user area to provide
support for OS simulation routines. In this case, the Editor and EXEC modules must be available on the S-disk or another accessed disk, in
this case.
In addition, for the CMSSEG segment only, you can indicate an
alternate segment to be loaded on the IPL command. The format of the IPL command to support this is: 1PL cuu PARM SEG=segmentname
systemname
SEG=seqmentname indicates the name of the saved segment to be loaded
whenever the CMS Editor, EXEC processor, or OS simulation routines
are needed. Eight characters must be entered for segmentname;
either assign an 8-character segment name when you code the NAMESYS macro for your installation, or be sure that the operator enters
trailing blanks if segmentname is less than 8 characters long.
The CMS Batch Facility loads whatever segment is specified
first 1PL command issued for the batch virtual machine. Thus,
first IPL command for a CMS Batch Facility machine is: IPL CMS PARM SEG=CMSSEG02 the same segment name (CMSSEG02) is loaded.
on the
if the When the command "SET SYSNAME CKSSEG segmentname" is specified (where
seqmentname is the name of a defined CMS discontiguous saved segment) ,
free storage containing OS simulation routines is released and OS routines contained in the CKSSEG segment are used by the
virtual machine to provide OS simUlation fUnctions.
314 IBM VM/3 7 0 System Programmer's Guide