Fast/CRAM: Difference between revisions
(manual cut/past/formatting of text from PDF) |
m (misc formatting and links) |
||
| Line 1: | Line 1: | ||
==Overview== | ==Overview== | ||
The Fast/CRAM program product is a functional replacement for the Model 204 Cross Region Access Method (CRAM) SVC. CRAM-SVC provides communications support between Model 204 and teleprocessing monitors, such as CICS or TSO. CRAM also provides support for host language interface (IFAM2) programs executed in batch as well as the BATCH2 utility program. | The Fast/CRAM program product is a functional replacement for the Model 204 Cross Region Access Method (CRAM) SVC. [[Cross-memory facilities CRAM and M204XSVC#Two CRAM types|CRAM-SVC]] provides communications support between Model 204 and teleprocessing monitors, such as CICS or TSO. CRAM also provides support for host language interface ([[HLI: Job requirements#IFAM2 jobs|IFAM2]]) programs executed in batch as well as the [[BATCH2 (CRAM)|BATCH2]] utility program. | ||
===Fast/CRAM advantages compared to CRAM=== | ===Fast/CRAM advantages compared to CRAM=== | ||
| Line 6: | Line 6: | ||
<li>Fast/CRAM consumes considerably less system resources. Total CPU savings range from 20% to 70%, depending upon the application. Savings realized by a CICS region can be significantly higher. </li> | <li>Fast/CRAM consumes considerably less system resources. Total CPU savings range from 20% to 70%, depending upon the application. Savings realized by a CICS region can be significantly higher. </li> | ||
<li>Fast/CRAM does not require any subtasks. This removes the single-threading of CRAM OPEN and CLOSE and dramatically reduces the overhead encountered by CICS when many CRAM connections with Model 204 are active. </li> | <li>Fast/CRAM does not require any subtasks. This removes the single-threading of CRAM OPEN and CLOSE, and it dramatically reduces the overhead encountered by CICS when many CRAM connections with Model 204 are active. </li> | ||
<li>Fast/CRAM makes extensive use of SRB routines, which take full advantage of multiple-processor complexes. </li> | <li>Fast/CRAM makes extensive use of SRB routines, which take full advantage of multiple-processor complexes. </li> | ||
<li>Fast/CRAM is totally compatible with MVS/XA and MVS/ESA. All storage is allocated | <li>Fast/CRAM is totally compatible with MVS/XA and MVS/ESA. All storage is allocated "above the line." 31-bit application programs are supported. </li> | ||
<li>Fast/CRAM is reliable. Extensive use of MVS recovery facilities, ensures that has Fast/CRAM has no | <li>Fast/CRAM is reliable. Extensive use of MVS recovery facilities, ensures that has Fast/CRAM has no "hanging states" or integrity exposures. Fast/CRAM adheres to the installation responsibilities for MVS integrity as outlined by IBM. </li> | ||
</ul> | </ul> | ||
| Line 20: | Line 20: | ||
Fast/CRAM comprises a single SVC load module and an interface module. The SVC load module may be installed as a Type 2, Type 3, or Type 4 SVC. The interface module may be installed in the Link Pack Area (LPA), or a private library, and need not be authorized. | Fast/CRAM comprises a single SVC load module and an interface module. The SVC load module may be installed as a Type 2, Type 3, or Type 4 SVC. The interface module may be installed in the Link Pack Area (LPA), or a private library, and need not be authorized. | ||
Fast/CRAM requires a subsystem name, reserved through an entry in an IEFSSNxx member of SYS1.PARMLIB. Both Fast/CRAM modules are totally re-entrant and refreshable and do not need to be fixed in real storage. The SVC load module is linked with AMODE 31,RMODE ANY, and always executes in 31-bit addressing mode | Fast/CRAM requires a subsystem name, reserved through an entry in an IEFSSNxx member of SYS1.PARMLIB. Both Fast/CRAM modules are totally re-entrant and refreshable and do not need to be fixed in real storage. The SVC load module is linked with <code>AMODE 31, RMODE ANY</code>, and it always executes in 31-bit addressing mode. The interface load module is linked with <code>AMODE ANY, RMODE 24</code>, and it executes in the addressing mode of its callers. | ||
Fast/CRAM allocates three different types of storage: SQA, CSA, and private area. All storage is allocated | Fast/CRAM allocates three different types of storage: SQA, CSA, and private area. All storage is allocated "above the line" in MVS/XA and MVS/ESA systems. | ||
<ul> | <ul> | ||
<li>The SQA storage is allocated in subpool 245 when a Fast/CRAM channel is opened. The amount of SQA storage allocated for each channel is 128 bytes plus the number of users times 560 bytes. </li> | <li>The SQA storage is allocated in subpool 245 when a Fast/CRAM channel is opened. The amount of SQA storage allocated for each channel is 128 bytes plus the number of users times 560 bytes. </li> | ||
<li>CSA storage is allocated in subpool 231 the first time that a particular Fast/CRAM subchannel is used. The size of each CSA allocation is determined by the blocksize for the Fast/CRAM channel. This is either IFAMBS for IFAM2 threads, LOUTPB for IODEV | <li>CSA storage is allocated in subpool 231 the first time that a particular Fast/CRAM subchannel is used. The size of each CSA allocation is determined by the blocksize for the Fast/CRAM channel. This is either <var>[[IFAMBS parameter|IFAMBS]]</var> for IFAM2 threads, <var>[[LOUTPB parameter|LOUTPB]]</var> for <var>[[IODEV parameter|IODEV]]</var> 11 threads, or <code>4 + MAX ([[OUTMRL parameter|OUTMRL]], [[INMRL parameter|INMRL]], 252)</code> for <var>IODEV</var> 29 threads. </li> | ||
<li>The private area storage is allocated in subpool 252, above the line. </li> | <li>The private area storage is allocated in subpool 252, above the line. </li> | ||
| Line 32: | Line 32: | ||
==Storage usage== | ==Storage usage== | ||
Fast/CRAM allocates storage in three distinct areas: CSA, SQA, and the private area. All of this storage is | Fast/CRAM allocates storage in three distinct areas: CSA, SQA, and the private area. All of this storage is "above the line" in MVS/XA and MVS/ESA systems. | ||
===SQA storage=== | ===SQA storage=== | ||
Fast/CRAM allocates SQA space only when Model 204 is initialized. The amount of SQA storage allocated for a Model 204 Online depends upon parameters specified for the job. Each Model 204 Online may establish up to three CRAM channels: | Fast/CRAM allocates SQA space only when Model 204 is initialized. The amount of SQA storage allocated for a Model 204 Online depends upon parameters specified for the job. Each Model 204 Online may establish up to three CRAM channels: | ||
<ul> | <ul> | ||
<li>One channel is for any host language threads (IODEV=23). </li> | <li>One channel is for any host language threads (<code>IODEV=23</code>). </li> | ||
<li>One channel is for any full screen remote threads (IODEV=11). </li> | <li>One channel is for any full screen remote threads (<code>IODEV=11</code>). </li> | ||
<li>One channel is for any non-full screen remote threads (IODEV=29). </li> | <li>One channel is for any non-full-screen remote threads (<code>IODEV=29</code>). </li> | ||
</ul> | </ul> | ||
SQA storage for each CRAM channel is allocated in a contiguous block when the channel is opened, and freed when the channel is closed. Channel open generally corresponds to Model 204 initialization, while channel close generally corresponds to Model 204 termination. The amount of SQA storage allocated depends upon the channel type, as follows (all quantities in bytes): | SQA storage for each CRAM channel is allocated in a contiguous block when the channel is opened, and freed when the channel is closed. Channel open generally corresponds to Model 204 initialization, while channel close generally corresponds to Model 204 termination. The amount of SQA storage allocated depends upon the channel type, as follows (all quantities in bytes): | ||
<table> | <table> | ||
<tr><td>IODEV=23</td> | <tr><td>IODEV=23</td> | ||
<td>If any IFAM2 threads are defined, one CRAM channel is opened and the following amount of SQA storage is allocated: | <td>If any IFAM2 threads are defined, one CRAM channel is opened and the following amount of SQA storage is allocated: | ||
<p class="code">((Number of IODEV= | <p class="code">((Number of IODEV=23s) * 560) + 128 </p></td></tr> | ||
<tr><td>IODEV=11</td> | <tr><td>IODEV=11</td> | ||
<td>If any remote full screen connections are defined, one CRAM channel is opened and the following amount of SQA storage is allocated: | <td>If any remote full screen connections are defined, one CRAM channel is opened and the following amount of SQA storage is allocated: | ||
<p class="code">(NOTERM * 560) + 128 </p></td></tr> | <p class="code">([[NOTERM parameter|NOTERM]] * 560) + 128 </p></td></tr> | ||
<tr><td>IODEV=29</td> | <tr><td>IODEV=29</td> | ||
| Line 60: | Line 60: | ||
</table> | </table> | ||
<p> | <p> | ||
Essentially, the SQA usage is 560 bytes per possible user. In an installation with twenty host language (IODEV=23) threads and fifty full-screen (IODEV=11) threads, the total SQA usage would be 39,456 bytes, or less than eleven pages. </p> | Essentially, the SQA usage is 560 bytes per possible user. In an installation with twenty host language (<code>IODEV=23</code>) threads and fifty full-screen (<code>IODEV=11</code>) threads, the total SQA usage would be 39,456 bytes, or less than eleven pages. </p> | ||
===CSA storage=== | ===CSA storage=== | ||
Fast/CRAM allocates CSA space for data buffers. This space is fetch-protected, unlike the space allocated by | Fast/CRAM allocates CSA space for data buffers. This space is fetch-protected, unlike the space allocated by Model 204 CRAM-SVC. This provides increased data security by preventing unauthorized users from examining Model 204 data streams which may contain passwords and other sensitive data. | ||
Fast/CRAM data buffers are not allocated when a channel is opened. Instead, they are allocated the first time that a particular connection is used. Fast/CRAM data buffers are only freed when the channel is closed, typically corresponding to Model 204 termination. Thus, the number of Fast/CRAM data buffers allocated is determined by the | Fast/CRAM data buffers are not allocated when a channel is opened. Instead, they are allocated the first time that a particular connection is used. Fast/CRAM data buffers are only freed when the channel is closed, typically corresponding to Model 204 termination. Thus, the number of Fast/CRAM data buffers allocated is determined by the highwater mark for the number of concurrent users. Model 204 CRAM-SVC allocates data buffers in the same way, and they are the same size. | ||
The amount of CSA storage allocated depends upon the channel type, as follows (all quantities in bytes): | The amount of CSA storage allocated depends upon the channel type, as follows (all quantities in bytes): | ||
| Line 79: | Line 79: | ||
<tr><td>IODEV=29</td> | <tr><td>IODEV=29</td> | ||
<td>The following amount of CSA storage is allocated for each remote non-full screen connection the first time it is used: | <td>The following amount of CSA storage is allocated for each remote non-full-screen connection the first time it is used: | ||
<p class="code">4 + MAX (OUTMRL, INMRL, 252) </p> | <p class="code">4 + MAX (OUTMRL, INMRL, 252) </p> | ||
</td></tr> | </td></tr> | ||
| Line 85: | Line 85: | ||
===Private area=== | ===Private area=== | ||
Each Fast/CRAM private area work area is 136 bytes long, and is allocated from subpool 252, with key zero. One work area is allocated in the Model 204 region for each Fast/CRAM channel open. This work area is freed when the Fast/CRAM channel is closed. No other private area storage is allocated or freed by Fast/CRAM. Fast/CRAM reduces the SPCORE requirement by approximately 200 bytes per IODEV=23, IODEV=29, or IODEV=11 thread. | Each Fast/CRAM private area work area is 136 bytes long, and is allocated from subpool 252, with key zero. One work area is allocated in the Model 204 region for each Fast/CRAM channel open. This work area is freed when the Fast/CRAM channel is closed. No other private area storage is allocated or freed by Fast/CRAM. Fast/CRAM reduces the <var>[[SPCORE parameter|SPCORE]]</var> requirement by approximately 200 bytes per <code>IODEV=23</code>, <code>IODEV=29</code>, or <code>IODEV=11</code> thread. | ||
One Fast/CRAM work area is allocated in each application region each time a | One Fast/CRAM work area is allocated in each application region each time a CRAM connection is established. This area is freed when the connection is broken. Individual Fast/CRAM data transfer calls do not allocate or free storage. | ||
===31-bit considerations=== | ===31-bit considerations=== | ||
When Fast/CRAM is used with MVS/XA or MVS/ESA, all storage is allocated "above the line." Application programs may execute in 31-bit addressing mode, provided that | When Fast/CRAM is used with MVS/XA or MVS/ESA, all storage is allocated "above the line." Application programs may execute in 31-bit addressing mode, provided that all data areas and the CRAM user control block are "below the line." | ||
==Coexistence considerations== | ==Coexistence considerations== | ||
Fast/CRAM may be installed and used in parallel with the basic Model 204 CRAM. However, successful coexistence requires that the correct operational procedures be carefully followed. Simplicity argues for just one version of CRAM-SVC. Using two versions of CRAM will require two separate SVC numbers and two separate subsystem names. | Fast/CRAM may be installed and used in parallel with the basic Model 204 CRAM. However, successful coexistence requires that the correct operational procedures be carefully followed. Simplicity argues for just one version of CRAM-SVC. Using two versions of CRAM will require two separate SVC numbers and two separate subsystem names. | ||
CRAM-SVC applications LOAD a copy of IGCLM244 to determine the correct CRAM or Fast/CRAM SVC number. The key to coexistence is two separate versions of IGCLM244, which in turn identify two different SVC numbers. The "standard" version of CRAM would be identified by a copy of IGCLM244 contained in the link pack area or linklib, while the "test" copy would be identified by a copy of IGCLM244 contained in a separate library. | CRAM-SVC applications LOAD a copy of [[Cross-memory facilities CRAM and M204XSVC#Component modules|IGCLM244]] to determine the correct CRAM or Fast/CRAM SVC number. The key to coexistence is two separate versions of IGCLM244, which in turn identify two different SVC numbers. The "standard" version of CRAM would be identified by a copy of IGCLM244 contained in the link pack area or linklib, while the "test" copy would be identified by a copy of IGCLM244 contained in a separate library. | ||
The test version would be invoked whenever Model 204 or CICS included the special IGCLM244 library in their STEPLIB concatenation. Care will be required to ensure that each partner has the correct sequence of STEPLIB libraries. If Fast/CRAM is installed in this fashion, IGCLM244 should be placed in an APF authorized library so that SNAPFAST will work correctly. | The test version would be invoked whenever Model 204 or CICS included the special IGCLM244 library in their STEPLIB concatenation. Care will be required to ensure that each partner has the correct sequence of STEPLIB libraries. If Fast/CRAM is installed in this fashion, IGCLM244 should be placed in an APF authorized library so that [[#SNAPFAST utility|SNAPFAST]] will work correctly. | ||
If this approach is used, then a special version of the TSO interface may be required. Suppose that Fast/CRAM is being installed in a test mode. A special library containing just the Fast/CRAM version of IGCLM244 would then be added to the STEPLIB concatenation for Model 204 and any of its partners (CICS, BATCH2, etc). | If this approach is used, then a special version of the TSO interface may be required. Suppose that Fast/CRAM is being installed in a test mode. A special library containing just the Fast/CRAM version of IGCLM244 would then be added to the STEPLIB concatenation for Model 204 and any of its partners (CICS, BATCH2, etc). | ||
| Line 157: | Line 155: | ||
==Installation== | ==Installation== | ||
For information about downloading and installing Fast/CRAM, see [[Downloading and installing Rocket M204 products#How should the Fast/CRAM object files be installed?|How should the Fast/CRAM object files be installed?]] | |||
===Defining subsystem name=== | ===Defining subsystem name=== | ||
| Line 198: | Line 166: | ||
===Zapping IFAM2 control commands=== | ===Zapping IFAM2 control commands=== | ||
Various infrequently used IFAM2 control commands, for example IFAMFORCE, IFAMDRAIN, and IFAMSTART, can cause timing-related problems. This is because these commands violate CRAM-SVC protocols and create race conditions. | Various infrequently used IFAM2 control commands, for example <var>[[IFAMFORCE command|IFAMFORCE]]</var>, <var>[[IFAMDRAIN command|IFAMDRAIN]]</var>, and <var>[[IFAMSTART command|IFAMSTART]]</var>, can cause timing-related problems. This is because these commands violate CRAM-SVC protocols and create race conditions. | ||
With CRAM-SVC, this can cause various hanging conditions or protection exceptions. Fast/CRAM intercepts these errors and takes an SVC dump and returns a completion code of 99 to the offending call. | With CRAM-SVC, this can cause various hanging conditions or protection exceptions. Fast/CRAM intercepts these errors and takes an SVC dump and returns a completion code of 99 to the offending call. | ||
Latest revision as of 21:44, 4 August 2017
Overview
The Fast/CRAM program product is a functional replacement for the Model 204 Cross Region Access Method (CRAM) SVC. CRAM-SVC provides communications support between Model 204 and teleprocessing monitors, such as CICS or TSO. CRAM also provides support for host language interface (IFAM2) programs executed in batch as well as the BATCH2 utility program.
Fast/CRAM advantages compared to CRAM
- Fast/CRAM consumes considerably less system resources. Total CPU savings range from 20% to 70%, depending upon the application. Savings realized by a CICS region can be significantly higher.
- Fast/CRAM does not require any subtasks. This removes the single-threading of CRAM OPEN and CLOSE, and it dramatically reduces the overhead encountered by CICS when many CRAM connections with Model 204 are active.
- Fast/CRAM makes extensive use of SRB routines, which take full advantage of multiple-processor complexes.
- Fast/CRAM is totally compatible with MVS/XA and MVS/ESA. All storage is allocated "above the line." 31-bit application programs are supported.
- Fast/CRAM is reliable. Extensive use of MVS recovery facilities, ensures that has Fast/CRAM has no "hanging states" or integrity exposures. Fast/CRAM adheres to the installation responsibilities for MVS integrity as outlined by IBM.
System requirements
Fast/CRAM executes within the IBM operating systems: MVS/SP Version 1 (MVS/370), MVS/SP Version 2 (MVS/XA), and MVS/SP Version 3 (MVS/ESA). One set of modules supports all of these system types.
Fast/CRAM comprises a single SVC load module and an interface module. The SVC load module may be installed as a Type 2, Type 3, or Type 4 SVC. The interface module may be installed in the Link Pack Area (LPA), or a private library, and need not be authorized.
Fast/CRAM requires a subsystem name, reserved through an entry in an IEFSSNxx member of SYS1.PARMLIB. Both Fast/CRAM modules are totally re-entrant and refreshable and do not need to be fixed in real storage. The SVC load module is linked with AMODE 31, RMODE ANY, and it always executes in 31-bit addressing mode. The interface load module is linked with AMODE ANY, RMODE 24, and it executes in the addressing mode of its callers.
Fast/CRAM allocates three different types of storage: SQA, CSA, and private area. All storage is allocated "above the line" in MVS/XA and MVS/ESA systems.
- The SQA storage is allocated in subpool 245 when a Fast/CRAM channel is opened. The amount of SQA storage allocated for each channel is 128 bytes plus the number of users times 560 bytes.
- CSA storage is allocated in subpool 231 the first time that a particular Fast/CRAM subchannel is used. The size of each CSA allocation is determined by the blocksize for the Fast/CRAM channel. This is either IFAMBS for IFAM2 threads, LOUTPB for IODEV 11 threads, or
4 + MAX (OUTMRL, INMRL, 252)for IODEV 29 threads. - The private area storage is allocated in subpool 252, above the line.
Storage usage
Fast/CRAM allocates storage in three distinct areas: CSA, SQA, and the private area. All of this storage is "above the line" in MVS/XA and MVS/ESA systems.
SQA storage
Fast/CRAM allocates SQA space only when Model 204 is initialized. The amount of SQA storage allocated for a Model 204 Online depends upon parameters specified for the job. Each Model 204 Online may establish up to three CRAM channels:
- One channel is for any host language threads (
IODEV=23). - One channel is for any full screen remote threads (
IODEV=11). - One channel is for any non-full-screen remote threads (
IODEV=29).
SQA storage for each CRAM channel is allocated in a contiguous block when the channel is opened, and freed when the channel is closed. Channel open generally corresponds to Model 204 initialization, while channel close generally corresponds to Model 204 termination. The amount of SQA storage allocated depends upon the channel type, as follows (all quantities in bytes):
| IODEV=23 | If any IFAM2 threads are defined, one CRAM channel is opened and the following amount of SQA storage is allocated:
((Number of IODEV=23s) * 560) + 128 |
| IODEV=11 | If any remote full screen connections are defined, one CRAM channel is opened and the following amount of SQA storage is allocated:
(NOTERM * 560) + 128 |
| IODEV=29 | If any remote non-full screen connections are defined, one CRAM channel is opened and the following amount of SQA storage is allocated:
(NOTERM * 560) + 128 |
Essentially, the SQA usage is 560 bytes per possible user. In an installation with twenty host language (IODEV=23) threads and fifty full-screen (IODEV=11) threads, the total SQA usage would be 39,456 bytes, or less than eleven pages.
CSA storage
Fast/CRAM allocates CSA space for data buffers. This space is fetch-protected, unlike the space allocated by Model 204 CRAM-SVC. This provides increased data security by preventing unauthorized users from examining Model 204 data streams which may contain passwords and other sensitive data.
Fast/CRAM data buffers are not allocated when a channel is opened. Instead, they are allocated the first time that a particular connection is used. Fast/CRAM data buffers are only freed when the channel is closed, typically corresponding to Model 204 termination. Thus, the number of Fast/CRAM data buffers allocated is determined by the highwater mark for the number of concurrent users. Model 204 CRAM-SVC allocates data buffers in the same way, and they are the same size.
The amount of CSA storage allocated depends upon the channel type, as follows (all quantities in bytes):
| IODEV=23 | The following amount of CSA storage is allocated for each IFAM2 thread the first time it is used:
IFAMBS |
| IODEV=11 | The following amount of CSA storage is allocated for each remote full screen connection the first time it is used:
LOUTPB |
| IODEV=29 | The following amount of CSA storage is allocated for each remote non-full-screen connection the first time it is used:
4 + MAX (OUTMRL, INMRL, 252) |
Private area
Each Fast/CRAM private area work area is 136 bytes long, and is allocated from subpool 252, with key zero. One work area is allocated in the Model 204 region for each Fast/CRAM channel open. This work area is freed when the Fast/CRAM channel is closed. No other private area storage is allocated or freed by Fast/CRAM. Fast/CRAM reduces the SPCORE requirement by approximately 200 bytes per IODEV=23, IODEV=29, or IODEV=11 thread.
One Fast/CRAM work area is allocated in each application region each time a CRAM connection is established. This area is freed when the connection is broken. Individual Fast/CRAM data transfer calls do not allocate or free storage.
31-bit considerations
When Fast/CRAM is used with MVS/XA or MVS/ESA, all storage is allocated "above the line." Application programs may execute in 31-bit addressing mode, provided that all data areas and the CRAM user control block are "below the line."
Coexistence considerations
Fast/CRAM may be installed and used in parallel with the basic Model 204 CRAM. However, successful coexistence requires that the correct operational procedures be carefully followed. Simplicity argues for just one version of CRAM-SVC. Using two versions of CRAM will require two separate SVC numbers and two separate subsystem names.
CRAM-SVC applications LOAD a copy of IGCLM244 to determine the correct CRAM or Fast/CRAM SVC number. The key to coexistence is two separate versions of IGCLM244, which in turn identify two different SVC numbers. The "standard" version of CRAM would be identified by a copy of IGCLM244 contained in the link pack area or linklib, while the "test" copy would be identified by a copy of IGCLM244 contained in a separate library.
The test version would be invoked whenever Model 204 or CICS included the special IGCLM244 library in their STEPLIB concatenation. Care will be required to ensure that each partner has the correct sequence of STEPLIB libraries. If Fast/CRAM is installed in this fashion, IGCLM244 should be placed in an APF authorized library so that SNAPFAST will work correctly.
If this approach is used, then a special version of the TSO interface may be required. Suppose that Fast/CRAM is being installed in a test mode. A special library containing just the Fast/CRAM version of IGCLM244 would then be added to the STEPLIB concatenation for Model 204 and any of its partners (CICS, BATCH2, etc).
The Fast/CRAM IGCLM244 should also be linked into SYS1.CMDLIB or the LINKLIB concatenation as FASTRLD. A second copy of the TSO interface should then be linked into SYS1.CMDLIB, with a special name such as FASTFS. Then use the IMASPZAP service aid to DUMPT this second copy of the TSO interface and locate all occurrences of the string IGCLM244. Then zap these to FASTRLD.
Any time FASTFS is used, it will load FASTRLD instead of IGCLM244, thus using the Fast/CRAM SVC instead of the SVC identified by the link pack version of IGCLM244. Care must be taken to ensure that the respective CRAM SVCs are only invoked by "their own" versions of IGCLM244.
Contents of distribution tape
Fast/CRAM is distributed on a magnetic tape which also contains all the other Sirius products you purchased.
The tape contains three files, each in IEBCOPY format, but only the first two are used for Fast/CRAM installation. The relevant members for Fast/CRAM in these two files are described in the following two sections.
SIRIUS.LIB
The first file, SIRIUS.LIB, contains the following members:
| Member | Description |
|---|---|
| FASTCUST | Example job for customizing and installing Fast/CRAM. |
| FASTOSIV | Example job to linkedit FASTSVC and FASTRLD on Fujitsu OSIV systems. |
| FASTVOS3 | Example job to linkedit FASTSVC and FASTRLD on Hitachi VOS3 systems. |
| FASTIBM | Example job to linkedit FASTSVC and FASTRLD on IBM MVS systems. |
| FASTSNAP | Example job for executing the SNAPFAST and SNAPSUBS utility. |
| SVCREP | Fast/CRAM installation utility, used to dynamically update SVC table. |
SIRIUS.LOAD
The second file, SIRIUS.LOAD, will contain the following members:
| Member | Description |
|---|---|
| IGCLM244 | Fast/CRAM interface module replacement. |
| FASTSVC | Fast/CRAM SVC module. |
| SNAPFAST | Fast/CRAM snapping utility (SNAPCRAM replacement). Also contains the SNAPSUBS utility. |
| FASTREP | Fast/CRAM installation utility, used to dynamically update SVC table. |
Installation
For information about downloading and installing Fast/CRAM, see How should the Fast/CRAM object files be installed?
Defining subsystem name
The IEFSSNxx member of SYS1.PARMLIB is used to define subsystems. Fast/CRAM requires a subsystem definition without any initialization routine. Refer to the "SPL: Initialization and Tuning" IBM manual for the format of IEFSSNxx entries. The utility SNAPSUBS can be used to get a listing of all defined subsystems.
Customizing SVC modules
Once the SVC number and subsystem name have been determined, the SVC load module and its interface module must be customized and then linkage edited into the appropriate library. This can be accomplished with the job stream FASTCUST, found in SIRIUS.LIB. This assumes that the Fast/CRAM load modules have already been restored into SIRIUS.LOAD, the SVC is to be installed in SYS1.LPALIB, the SVC number is 250, and the subsystem name is M204.
Note: This job first zaps the subsystem name and SVC number into the SVC load module and interface module, then links these two modules into SYS1.LPALIB with their appropriate names. For more information on SVC installation techniques, refer to the "SPL: Supervisor" IBM manual.
Zapping IFAM2 control commands
Various infrequently used IFAM2 control commands, for example IFAMFORCE, IFAMDRAIN, and IFAMSTART, can cause timing-related problems. This is because these commands violate CRAM-SVC protocols and create race conditions.
With CRAM-SVC, this can cause various hanging conditions or protection exceptions. Fast/CRAM intercepts these errors and takes an SVC dump and returns a completion code of 99 to the offending call.
The following job applies a zap to the Release 2.2 version of the Model 204 module IFII. This zap corrects most of the IFAM2 control command timing errors. This zap should not be applied to CMS systems.
//IFIIZAP JOB ,'FAST/CRAM',CLASS=A,MSGCLASS=A //* //* JOB TO ZAP IFII TO CORRECT CRAM PROTOCOL //* VIOLATIONS. //* //* DO NOT APPLY THESE ZAPS TO A CMS VERSION //* OF IFII, SINCE THEY REUSE A CMS-ONLY //* PARAMETER LIST //* //ZAPS EXEC PGM=SIRZAP,REGION=1024K //SYSPRINT DD SYSOUT=* //SYSLIB DD DSN=, //SYSIN DD DISP=SHR,DSN=SIRIUS.LIB(V22) //
SNAPFAST utility
The SNAPFAST utility is provided for listing the SQA storage allocated by Fast/CRAM, as well as the Fast/CRAM SVC load module and interface module. If this job is run with SNAPFAST linked into an APF-authorized library, the Fast/CRAM CSA data buffers are snapped as well. This utility can be invoked with the following job, assuming that the Fast/CRAM load modules have already been restored into SIRIUS.LOAD.
//SNAPF JOB 0,'SNAP SQA/CSA',CLASS=A, // MSGCLASS=X //FAST EXEC PGM=SNAPFAST,REGION=64K //STEPLIB DD DSN=SIRIUS.LOAD,DISP=SHR //SYSPRINT DD SYSOUT=* //SYSUDUMP DD SYSOUT=* //
FASTREP utility
The FASTREP utility is provided for dynamically installing a version of Fast/CRAM without requiring a system IPL. For this utility to work, FAST.CRAM.LOAD must be APF-authorized, and the correct SVC number must have been zapped into the FASTSVC load module.
FASTREP verifies that the identified SVC number is the correct type, and that the identified subsystem name has been defined and is not already in use. If these conditions are met, the FASTSVC load module is loaded into the (extended) SQA, and the SVC table is dynamically updated.
Note: An IPL will still be required to install a new version of IGCLM244 into the link pack area, but Fast/CRAM can be tested with IGCLM244 in a private library. Please refer to Coexistence considerations for more information on such a test environment.
The FASTREP utility can be invoked with the following job:
//FASTREP JOB 0,'UPDATE SVC TABLE',CLASS=A, // MSGCLASS=X //* //* UPDATE SVC TABLE WITH MEMBER FASTSVC //* FROM SYSLIB //* //SVCUPDTE EXEC PGM=FASTREP,REGION=64K,PARM=FASTSVC //STEPLIB DD DSN=SIRIUS.LOAD,DISP=SHR //SYSLIB DD DSN=SIRIUS.LOAD,DISP=SHR //SYSPRINT DD SYSOUT=A //SYSUDUMP DD SYSOUT=A //
SNAPSUBS utility
The SNAPSUBS utility is provided for listing all currently defined subsystems. This utility can be invoked with the following job, assuming that the Fast/CRAM load modules have already been restored into FAST.CRAM.LOAD. The output from this job is a dump of the storage for each subsystem CVT.
//SNAPS JOB 0,'SNAP SSCVTS',CLASS=A, // MSGCLASS=X //SSVCTS EXEC PGM=SNAPSUBS,REGION=64K //STEPLIB DD DSN=SIRIUS.LOAD,DISP=SHR //SYSPRINT DD SYSOUT=A //SYSUDUMP DD SYSOUT=A //