|
The parameter descriptions for REQUEST=WRITE_DATA are listed in
alphabetical order. Default values are underlined: - REQUEST=WRITE_DATA
- Use this input parameter to specify that the data item identified
by NAME be written from the areas specified by BUFFER or BUFLIST,
and/or ADJAREA to the cache structure.
- ,ADJAREA=NO_ADJAREA
- ,ADJAREA=adjarea
- Use this input parameter to specify a storage area to contain
the adjunct data to be written to an entry in the cache structure.
Specify ADJAREA only when ELEMNUM is greater than 0.
If the structure
supports adjunct data and ADJAREA is not specified or ADJAREA=NO_ADJAREA
is specified, binary zeros are written to the adjunct area. If the
structure does not support adjunct data and ADJAREA is specified,
ADJAREA is ignored. (Adjunct areas for a structure are established
through the IXLCONN macro.)
To Code: Specify the RS-type
name or address (using a register from 2 to 12) of a 64-byte area
that contains the adjunct data you want associated with this data
entry.
- ,ANSAREA=NO_ANSAREA
- ,ANSAREA=ansarea
- Use this output parameter to specify an answer area to contain
information returned from the request. The format of the answer area
is described by the IXLYCAA mapping macro. On
a successful completion of a request, the following information is
returned in the answer area:
- If changed data is written to the cache, the total number of entries
containing either changed or locked-for-cast-out data, that are assigned
to the storage class to which the entry was written, is returned (field
CAATOTCHANGED).
- If changed data is written to the cache, the total number of data
items assigned to the cast-out class to which data was just written
is returned (field CAACOCOUNT).
- If WHENREG=NO was specified and the request replaced a previously
specified vector index for the entry, the replaced vector index is
returned (fields CAAINVLCVINUM and CAAINVLCVI).
To Code: Specify the RS-type name or address (using
a register from 2 to 12) of an area (with a length of ANSLEN) where
the information returned from the request may be stored.
- ,ANSLEN=anslen
- Use this input parameter to specify the size of the storage area
specified by ANSAREA.
Use either CAALEVEL0LEN or CAALEVEL1LEN of the IXLYCAA mapping
macro to determine
the minimum size of the answer area. The answer area length must
be at least large enough to accomodate the level of the IXLYCAA mapping
appropriate to the requested function. When the value of PLISTVER
is 0 — 3, the minimum answer area length is CAALEVEL0LEN; when
the value of PLISTVER is 4 — 6, the minimum answer area length
is CAALEVEL1LEN.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 2-byte field that contains the length of the answer
area (ANSAREA).
- ,ASSIGN=YES
- ,ASSIGN=NO
- Use this input parameter to specify whether a directory entry
will be assigned for NAME if one does not currently exist.
- YES
- A directory entry will be assigned.
- NO
- No directory entry will be assigned.
ASSIGN=NO is meaningful
only for cache structures allocated in a coupling facility of CFLEVEL=9
or higher.
- ,BUFADDRSIZE=31
- ,BUFADDRSIZE=64
- Use this input parameter to specify whether a 31-bit or a 64-bit
address is specified by a BUFLIST entry.
- 31
- The entry in BUFLIST is 31 bits in size.
- 64
- The entry in BUFLIST is 64 bits in size.
- ,BUFADDRTYPE=VIRTUAL
- ,BUFADDRTYPE=REAL
- Use this input parameter to specify whether the buffer addresses
specified in the BUFLIST list are virtual storage or real storage
addresses.
- VIRTUAL
- The buffer addresses are virtual storage addresses. The virtual
storage can be pageable or nonpageable. See the PAGEABLE parameter
for information about managing storage binds when specifying virtual
storage addresses.
- REAL
- The buffer addresses are real storage addresses.
It is the caller's responsibility to manage the binds between the
data buffer virtual storage and the real storage addresses provided.
The caller must ensure that the data buffer virtual storage remains
bound to the real storage addresses provided until the request completes.
- ,BUFALET=NO_BUFALET
- ,BUFALET=bufalet
- Use this input parameter to specify an access list entry token
(ALET) to be used in referencing all of the buffers specified by BUFLIST.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 4-byte field that contains the ALET.
- ,BUFFER=buffer
- Use this input parameter to specify a buffer area to hold the
entry data to be written to the cache structure. Only
31-bit addressable virtual storage areas (below 2GB) are supported
by the BUFFER specification. High virtual
storage areas (above 2GB) can only be specified via the BUFLIST specification.
You can define the buffer size to be a total of up to 65536 bytes.
Other requirements depend on the size you select: - If you specify a buffer size of less than or equal to 4096 bytes,
you must ensure that the buffer:
- Is 256, 512, 1024, 2048, or 4096 bytes.
- Starts on a 256-byte boundary.
- Does not cross a 4096-byte boundary.
- Does not start below storage address 512.
- If you specify a buffer size of greater than 4096 bytes, you must
ensure that the buffer:
- Is a multiple of 4096 bytes.
- Is less than or equal to 65536 bytes.
- Starts on a 4096-byte boundary.
- Does not start below storage address 512.
See the BUFSIZE parameter description for defining the
size of the buffer.
See z/OS MVS Programming: Sysplex Services Guide for
more information on buffers.
To
Code: Specify the RS-type name or address (using a register from
2 to 12) of an area (with a length of BUFSIZE) to contain the data
to be written to the entry in the structure.
- ,BUFINCRNUM=bufincrnum
- Use this input parameter to specify the number of 256-byte segments
comprising each buffer in the BUFLIST list.
Valid BUFINCRNUM values are 1,2,4,8, or 16, which correspond to
BUFLIST buffer sizes of 256, 512, 1024, 2048, and 4096 bytes respectively.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 1-byte field that contains 1,2,4,8, or 16.
- ,BUFLIST=buflist
- Use this input parameter to specify a list of buffers to hold
the entry data to be written to the cache structure. BUFLIST specifies
a 128-byte storage area that consists of a list of 0 to 16 elements.
Either 31-bit addressable (below 2GB) or 64-bit addressable
(above 2GB) real or virtual storage areas are supported for the BUFLIST specification,
depending on the specifications for the BUFADDRTYPE and BUFADDRSIZE keywords.
However, pageable high shared virtual storage areas (above 2GB) may
not be used.
The format of the list is a set
of 8-byte elements. The BUFADDRSIZE keyword denotes whether four or
eight bytes of the element are used. - If BUFADDRSIZE=31 is specified, then the first four bytes of each
element are reserved space and the last four bytes contain the address
of the buffer.
- If BUFADDRSIZE=64 is specified, then the full eight bytes specify
the address of the buffer.
The BUFLIST buffers must:
See the BUFNUM and BUFINCRNUM keyword descriptions for
specifying the number and size of buffers.
For structures
allocated in CFLEVEL=0 through CFLEVEL=3 coupling facilities, one
of BUFFER or BUFLIST is required. For structures allocated in CFLEVEL=4
or higher coupling facilities, one of BUFFER or BUFLIST is required
when CHANGED=YES and optional when CHANGED=NO.
See z/OS MVS Programming: Sysplex Services Guide for
more information on buffers.
To
Code: Specify the RS-type name or address (using a register from
2 to 12) of a 128-byte storage area that contains a list of buffer
addresses.
- ,BUFNUM=bufnum
- Use this input parameter to specify the number of buffers in the
BUFLIST list. Valid BUFNUM values are from 0 to 16. A value of zero
indicates that no data is to be written to the entry. If you do not
want buffers, but want to specify this parameter, code BUFNUM=NO_BUFNUM.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 1-byte field that contains the number of buffers
(0 to 16) in the buffer list.
- ,BUFSIZE=bufsize
- Use this input parameter to specify the size of the BUFFER area.
See the BUFFER parameter description for valid buffer sizes.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a fullword field that contains the size of the buffer
(BUFFER) in bytes.
- ,BUFSTGKEY=CALLERS_KEY
- ,BUFSTGKEY=bufstgkey
- Use this input parameter to specify a storage key that you define
and use when referencing the buffers specified by BUFLIST or the buffer
specified by BUFFER.
If you do not specify BUFSTGKEY, or if you specify BUFSTGKEY=CALLERS_KEY,
all references to the buffer(s) are performed using the caller's PSW
key.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of an 8-bit field that contains the storage key in the
format B'kkkkxxxx', where kkkk is the key and xxxx is ignored.
- ,CHANGED=NO
- ,CHANGED=YES
- Use this input parameter to specify whether changed data is to
be written to an entry in the cache structure.
- NO
- The data to be written is unchanged:
- The cached copy is the same as the permanent storage copy.
- In a store-through process, the permanent storage copy will be
updated (by the user) to match the cached copy, either before the
cast-out lock is released (if GETCOLOCK=YES was specified), or before
serialization on the named data item is released.
If the entry already exists in the structure, the entry
must contain unchanged data or the request will fail. You cannot overwrite
changed data with unchanged data.
For structures allocated
in CFLEVEL=4 or higher coupling facilities, ELEMNUM=0 and one of BUFFER
and BUFLIST are optional when CHANGED=NO is specified.
- YES
- The data to be written is changed. The changed data will be assigned
to the specified cast-out class (COCLASS) superseding any previously
specified cast-out class for the data. With the exception of your
connection, all users with registered interest in the data will have
their interest deregistered such that their locally cached copies
of the data are invalidated.
- ,COCLASS=coclass
- Use this input parameter to specify the cast-out class to which
the data item being written is to be assigned. Any previous assignment
is updated to the new specification.
To Code: Specify the
RS-type name or address (using a register from 2 to 12) of a 2-byte
field that contains the cast-out class value.
- ,CONTOKEN=contoken
- Use this input parameter to specify the connect token that was
returned by the IXLCONN service in the IXLCONN answer area, mapped
by IXLYCONA. The connect token uniquely identifies your connection
to the cache structure, and must be specified on each IXLCACHE invocation.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 16-byte field that contains the connect token.
- ,CROSSINVAL=NO
- ,CROSSINVAL=YES
- Use this input parameter to specify whether cross-invalidate processing
should be performed when writing unchanged data.
- NO
- Cross-invalidate processing is not performed. All users with registered
interest in the data item remain registered, and the copies of the
data item in their local cache buffers remain valid.
- YES
- Cross-invalidate processing is performed. All users with registered
interest in the data item, with the exception of your connection,
have their interest deregistered and the copies of the data item in
their local cache buffers invalidated.
- ,ELEMNUM=elemnum
- Use this input parameter to specify the number of elements to
be allocated to the data entry. Valid ELEMNUM values are from 0 to
the MAXELEMNUM value that was returned to the connector in the connect
answer area. For structures allocated in a CFLEVEL=0 coupling facility,
ELEMNUM values can be in the range of 1 to 16. For structures allocated
in coupling facilities with a CFLEVEL between CFLEVEL=1 and CFLEVEL=3,
ELEMNUM values can be in the range of 1 to 255. For structures allocated
in coupling facilities with a CFLEVEL=4 or higher, ELEMNUM values
can be in the range of 0 to 255 where 0 is valid only when CHANGED=NO.
Considerations for specifying ELEMNUM are: - If this request creates a new entry, the number of elements is
set to the ELEMNUM value.
- If the entry already exists, the number of elements is updated
to this ELEMNUM value.
If the data passed in BUFFER/BUFLIST does not fill up
all the space in the elements, the extra space will be padded with
zeros. If the data passed in BUFFER/BUFLIST is greater than the amount
of space in the elements, the data will be truncated.
To
Code: Specify the RS-type name or address (using a register from
2 to 12) of a 1-byte field that contains the number of elements.
- ,GETCOLOCK=NO
- ,GETCOLOCK=YES
- Use this input parameter to specify whether the cast-out lock
should be obtained when writing unchanged data.
- NO
- The cast-out lock is not obtained.
- YES
- The cast-out lock is obtained.
- ,LOCALREGCNTLNO
- ,LOCALREGCNTL=YES
- Use this keyword to specify whether the WRITE_DATA request should
be suppressed (write suppression) when the user's connection (local
cache) is the only registered interest in the data item identified
by NAME in the cache structure, and no subsystem data (data entry)
for the data item is cached.
LOCALREGCNTL=YES is only meaningful
when write requests are issued to cache structures allocated in a
coupling facility that supports write suppression based on local cache
registration. If the cache structure is allocated in a coupling facility
that does not support write suppression based on local cache registration,
the WRITE_DATA request will be performed as if LOCALREGCNTL=NO was
specified.
- NO
- The WRITE_DATA request will be performed normally without write
suppression.
- YES
- Suppress the WRITE_DATA request when the user's connection is
the only registered interest in the directory entry and the data entry
does not have cached subsystem data. When the write operation is suppressed,
no data is written to the cache structure. The WRITE_DATA request
will complete with a return code IXLRETCODEWARNING, reason code IXLRSNCODELOCALREGWRTSUPPRESS.
If
LOCALREGCNTL=YES is requested on a WRITE_DATA request for a cache
structure that is currently in the system-managed Duplex Established
phase, the request for write suppression is ignored.
CHANGED=YES
is required when LOCALREGCNTL=YES is specified.
When the subsystem
data is not written to the cache structure, the local cache buffer
contains the only valid copy of the subsystem data. Hardening of the
subsystem data to permanent storage should be performed to ensure
availability and successful recovery of data in the event of a failure.
When
issuing a WRITE_DATA request with LOCALREGCNTL=YES to write data to
a cache structure that is in user-managed duplex mode, the following
programming technique should be followed: - When using registration, assignment or write suppression protocols
in the primary structure of a user-managed duplexed structure, unconditional
writes of changed data to the secondary structure specifying LOCALREGCNTL=NO
should be performed after successful write operations to the primary
structure. If the write operation to the primary structure resulted
in the write being suppressed, then the write operation should not
be issued to the secondary structure.
- ,MF=S
- ,MF=(L,mfctrl)
- ,MF=(L,mfctrl,mfattr)
- ,MF=(L,mfctrl,0D)
- ,MF=(E,mfctrl)
- ,MF=(E,mfctrl,COMPLETE)
- Use MF=S to specify the standard form of the macro, which builds
an inline parameter list and generates the macro invocation to transfer
control to the service.
Use MF=L to specify the list form of the macro. Use the list form
together with the execute form of the macro for applications that
require reentrant code. The list form defines an area of storage that
the execute form uses to store the parameters. Only the PLISTVER parameter
can be coded with the list form of the macro.
Use MF=E to specify the execute form of the macro. Use the execute
form together with the list form of the macro for applications that
require reentrant code. The execute form stores the parameters into
the storage area defined by the list form, and generates the macro
invocation to transfer control to the service.
- ,mfctrl
- Use this output parameter to specify a storage area to contain
the parameters.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of the parameter list.
- ,mfattr
- Use this input parameter to specify the name of a 1- to 60-character
string that can contain any value that is valid on an assembler DS
pseudo-op. You can use this parameter to force boundary alignment
of the parameter list. If you do not code mfattr, the system
provides a value of 0D, which forces the parameter list to a doubleword
boundary.
- ,COMPLETE
- Use this input parameter to require that the system check for
required parameters and supply defaults for omitted optional parameters.
Note: In the macro expansion you might see some defaults for optional
parameters that are not documented here. The ones that are not documented
do not have any effect on the macro. For example, if SMILE=var were
an optional parameter and the default is SMILE=NO_SMILE then it would
not be documented. However, if the default was SMILE=:-), then it
would be documented because a value would be the default.
- ,MODE=SYNCSUSPEND
- ,MODE=SYNCECB
- ,MODE=SYNCEXIT
- ,MODE=SYNCTOKEN
- ,MODE=ASYNCECB
- ,MODE=ASYNCEXIT
- ,MODE=ASYNCTOKEN
- Use this input parameter to specify:
- Whether the request is to be performed synchronously or asynchronously
- How you wish to be notified of request completion if the request
is processed asynchronously.
See z/OS MVS Programming: Sysplex Services Guide for
more information on understanding
synchronous and asynchronous cache operations.
- SYNCSUSPEND
- The request processes synchronously. If necessary, the request
is suspended until it can complete synchronously. To use this option,
your program must be enabled for I/O and external interrupts.
- SYNCECB
- The request processes synchronously if possible. If the request
processes asynchronously, the ECB specified by REQECB is posted when
the request completes.
- SYNCEXIT
- The request processes synchronously if possible. If the request
processes asynchronously, your complete exit is given control when
the request completes.
- SYNCTOKEN
- The request processes synchronously if possible. If the request
processes asynchronously, an asynchronous request token is returned
to the area specified by REQTOKEN. Use the returned request token
on the IXLFCOMP macro to determine whether your request has completed.
Note: ANSAREA is a required parameter when MODE=SYNCTOKEN is specified.
- ASYNCECB
- The request processes asynchronously. The ECB specified by REQECB
is posted when the request completes.
- ASYNCEXIT
- The request processes asynchronously. Your complete exit is given
control when the request completes.
- ASYNCTOKEN
- The request processes asynchronously. An asynchronous request
token is returned to the area specified by REQTOKEN. Use the returned
request token on the IXLFCOMP macro to determine whether your request
has completed.
Note: ANSAREA is a required parameter when MODE=ASYNCTOKEN is specified.
- ,NAME=name
- Use this input parameter to specify the name of the data item
to be written to the cache structure.
To Code: Specify
the RS-type name or address (using a register from 2 to 12) of a 16-byte
field that contains the name of the data item.
- ,NEWVERS=newvers
- Use this input parameter to specify the value that is to be assigned
to the entry version number.
NEWVERS is meaningful only for structures
allocated in a coupling facility of CFLEVEL=5 or higher.
To
Code: Specify the RS-type name or address (using a register from
2 to 12) of an 8-byte field that contains the value for the entry
version number.
- ,OLDNAME=NO_OLDNAME
- ,OLDNAME=oldname
- Use this input parameter to specify the name of the data item
for which your interest should be deregistered.
- For structures allocated in a coupling facility of CFLEVEL=4 or
lower, you must also specify the name of the data item for which interest
is to be registered after the interest is deregistered in OLDNAME.
Identify the data item to which interest is to be registered with
NAME. The VECTORINDEX currently associated with the entry specified
by OLDNAME will be reassigned to the name of the data item specified
by NAME.
- For structures allocated in a coupling facility of CFLEVEL=5 or
higher, you can specify that interest in the name of the data item
specified by OLDNAME is to be deregistered without registering interest
in the entry specified by NAME.
In either case, whenever deregistration of interest is requested,
VECTORINDEX must be specified.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 16-byte field that contains the name of the old
data item that was associated with VECTORINDEX.
- ,PARITY=00110000
- ,PARITY=parity
- Use this input parameter to update the directory entry parity
bits for the data item specified by NAME. The parity bits are updated
only when CHANGED=YES is specified and the initial value of the directory
entry parity bits is null (that is, B'11'). If these conditions
are not met, the PARITY value is ignored.
Bits two and three of
the 8-bit PARITY field are the parity bits. Valid parity values are:
- xx00xxxx - Parity equals 0
- xx01xxxx - Parity equals 1
- xx11xxxx - Null (parity is not set)
Note that when a directory entry is created, the parity
bits are initialized to the null value.
To Code: Specify
the RS-type name or address (using a register from 2 to 12) of a 1-byte
field that contains the parity bits.
- ,PAGEABLE=YES
- ,PAGEABLE=NO
- Use this input parameter to identify whether the storage areas
specified by BUFFER or BUFLIST are in pageable or potentially pageable
storage.
- YES
- Specify this option to indicate that the BUFFER or BUFLIST buffers
reside in pageable virtual storage. XES performs the required page
fixing to fix the buffers in real storage while the cache or list
request transfers data to or from the coupling facility.
This includes storage obtained from pageable subpools, disabled
reference (DREF) storage, and may include storage that has the potential
to become pageable during the processing of a request. (An example
is address space storage owned by any swappable address space, for
which a PGSER FIX has been successfully processed, but for which the
owning address space gets swapped during processing of a cache or
list request.) This does not include implicitly non-pageable storage
(for example, storage obtained from non-pageable subpools).
The system takes responsibility for managing binds to central storage
for the duration of the cache or list request, regardless of what
address space owns the storage or whether the storage-owning address
space is swappable or nonswappable. The storage can be owned by any
address space.
High shared virtual storage areas (above 2GB) may not
be used.
- NO
- Specify this option to indicate that the BUFFER or BUFLIST buffers
reside in non-pageable virtual storage. XES does not page fix the
buffers in real storage.
This includes implicitly non-pageable storage areas (for example,
storage obtained from non-pageable subpools), and may include storage
that has the potential to become pageable during the processing of
a request (An example is address space storage owned by any swappable
address space, for which a PGSER FIX has been successfully processed,
but for which the owning address space gets swapped-out during processing
of a cache or list request.)
The system takes responsibility for managing binds to central storage
for the duration of the cache or list request, if and only if the
non-pageable storage is owned by either the requestor's address space
or the connector's address space. If the storage is owned by any other
address space, then the invoker is responsible for ensuring that the
virtual storage remains non-pageable for the duration of the request
(including the case in which the storage is owned by a swappable address
space that is swapped during processing of an IXLCACHE or IXLLIST
request). Subject to this consideration, the storage can be owned
by any address space. See z/OS MVS Programming: Sysplex Services Guide.
- ,PLISTVER=IMPLIED_VERSION
- ,PLISTVER=MAX
- ,PLISTVER=plistver
- Use this input parameter to specify the version of the macro.
See Understanding IXLCACHE Version Support for a description of the
options available with PLISTVER.
- ,PROCESSID=NO_PROCESSID
- ,PROCESSID=processid
- Use this input parameter to specify a user-defined process identifier
to be compared with the cast-out lock along with the connection identifier.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 1-byte field that contains the user-defined process
identifier.
- ,REGUSER=YES
- ,REGUSER=NO
- Use this input parameter to specify, for structures allocated
in coupling facilities with CFLEVEL=5 or higher, whether the request
should register connection interest in the entry.
- YES
- Specify this option to indicate that connection interest registration
will be performed.
VECTORINDEX is required when REGUSER=YES is
either specified or selected by default.
- NO
- Specify this option to indicate that no connection interest will
be registered. Be careful when using REGUSER=NO because preference
to reclaiming is given to entries that have data but no registered
interest.
VECTORINDEX is required when REGUSER=NO is specified
along with OLDNAME for the deregistration of interest. VECTORINDEX
is ignored when REGUSER=NO is specified without the specification
of OLDNAME.
- ,REQDATA=NO_REQDATA
- ,REQDATA=reqdata
- Use this input parameter with MODE=SYNCEXIT or MODE=ASYNCEXIT
to pass any data you choose to the complete exit. The exit will get
control only if the request is processed asynchronously.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of an 8-byte field that contains the data to be passed
to the complete exit.
- ,REQECB=reqecb
- Use this output parameter with either MODE=SYNCECB or MODE=ASYNCECB
to specify the address of an ECB, which is to be posted when the request
completes if the request was processed asynchronously.
Before coding REQECB, you must ensure that:
- You initialize the ECB before you issue the request.
- The ECB resides in either common storage or the home address space
where IXLCONN was issued.
- Any tasks that wait for the ECB to be posted reside in the home
address space where IXLCONN was issued.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 4-byte field that contains the address of the ECB
to be posted when the request completes. The ECB must be aligned on
a fullword boundary.
- ,REQID=NO_REQID
- ,REQID=reqid
- Use this input parameter to specify a user-defined request identifier
to be associated with the request. You can specify this request identifier
on the IXLPURGE macro to cancel a request that has not yet been processed.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of an 8-byte field that contains the user-defined request
identifier.
- ,REQTOKEN=reqtoken
- Use this output parameter with either MODE=SYNCTOKEN or MODE=ASYNCTOKEN
to specify the address of a storage area to receive the request token
that is returned when the request will be processed asynchronously.
This token, which uniquely identifies the request, must be used as
input to the IXLFCOMP macro, which you use to determine if the request
has completed.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 16-byte field where the system will put the request
token.
- ,RETCODE=retcode
- Use this output parameter to specify a field to contain the return
code. (The return code is also returned in register 15.)
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 4-byte field that will contain the return code
when the request has completed.
- ,RSNCODE=rsncode
- Use this output parameter to specify a field to contain the reason
code returned, if applicable. (The reason code is also returned in
register 0.)
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 4-byte field that will contain the reason code
(if any) when the request has completed.
- ,STGCLASS=stgclass
- Use this input parameter to assign a storage class to the data
item being written. Any previous assignment is updated to the new
specification.
To Code: Specify the RS-type name or address
(using a register from 2 to 12) of a 1-byte field that contains the
storage class.
- ,USERDATA=NO_USERDATA
- ,USERDATA=userdata
- Use this input parameter to specify user-defined information to
be written to the directory entry for the data item specified by NAME.
The information is written only if either of the following is true:
- There is no entry data in the structure for NAME.
- There is unchanged entry data in the structure for NAME.
If one of these conditions is not met, USERDATA is ignored.
If you do not want to write user data, but want to specify this parameter,
code USERDATA=NO_USERDATA.
To Code: Specify the RS-type
name or address (using a register from 2 to 12) of an 8-byte field
that contains the user data.
- ,VECTORINDEX=vectorindex
- Use this input parameter to specify an index into your local cache
vector to be associated with the data item specified by NAME. The
vector index identifies a vector entry that cache services will use
to indicate both your interest in the data item and the validity of
the copy of the data item in your local cache buffer.
To Code: Specify the RS-type name or address (using a register
from 2 to 12) of a 4-byte field that contains the vector index for
NAME.
- ,VERSCOMP=NO_VERSCOMP
- ,VERSCOMP=verscomp
- Use this input parameter to specify a version number to be compared
to the version number of the entry designated by NAME.
VERSCOMP
is meaningful only for structures allocated in a coupling facility
of CFLEVEL=5 or higher.
If the condition specified by VERSCOMPTYPE
is not met, the request is terminated with no resultant change to
the structure.
Note that use of VERSCOMP is needed to ensure
that updates to the version number requested with the VERSUPDATE keyword
are not processed multiple times as a result of internal request
redrive logic. When VERSCOMP is requested along with VERSUPDATE to
update the version number, then if the initial execution of the request
succeeds, any subsequent internal redrive of the request will fail
due to a version number miscompare, preventing multiple updates from
occurring on the request. Conversely, if the initial execution of
the request was unsuccessful, then any subsequent internal redrive
of the request will be able to execute successfully and update the
version number only once.
In either of these cases, if the
request is internally redriven and experiences a version number miscompare
on the redrive, a return and reason code of IXLRSNCODESTATUSUNKNOWN
will be returned. This reflects the fact that it is not known whether
the observed version number miscompare resulted from the version number
update succeeding on the original issuance of the request (causing
the miscompare on the redriven request), or the observed version number
miscompare was present all along, or resulted from a version number
update made by another request. When this return and reason code is
returned, it is up to the user to determine whether or not the requested
update has actually occurred, and take the appropriate recovery action.
To
Code: Specify the RS-type name or address (using a register from
2 to 12) of an 8-byte field that contains the comparative version
number.
- ,VERSCOMPTYPE=EQUAL
- ,VERSCOMPTYPE=LESSOREQUAL
- Use this input parameter to specify how the structure entry version
number comparison is to be performed.
VERSCOMPTYPE is meaningful
only for structures allocated in a coupling facility of CFLEVEL=5
or higher.
- VERSCOMPTYPE=EQUAL
- The version number for the structure entry must be equal to the
value specified for VERSCOMP.
- VERSCOMPTYPE=LESSOREQUAL
- The version number for the structure entry must be less than or
equal to the value specified for VERSCOMP.
- ,VERSUPDATE=NONE
- ,VERSUPDATE=INC
- ,VERSUPDATE=DEC
- ,VERSUPDATE=SET
- Use this input parameter to specify how the entry version number
will be updated or, for those cases where an entry is created, initialized.
VERSUPDATE is meaningful only for structures allocated in a coupling
facility of CFLEVEL=5 or higher.
Note that, depending on the
VERSUPDATE specification, internal request redrive logic may cause
a request to INCrement the version number more than once, DECrement
the version number more than once, or SET the version number more
than once. If it is necessary to ensure that this kind of multiple
operation does not occur, version number comparison must also be requested.
See the description of the VERSCOMP keyword.
- VERSUPDATE=NONE
- The version number is not updated.
On a request that causes
an entry to be created, the version number is set to contain all binary
zeros.
- VERSUPDATE=INC
- The version number will be incremented.
On a request that
causes an entry to be created, the version number for the created
entry is set to contain all binary zeros except for the low order
bit, which is set to one.
- VERSUPDATE=DEC
- The version number will be decremented.
On a request that
causes an entry to be created, the version number for the created
number is set to contain binary ones.
- VERSUPDATE=SET
- The version number will be set to the value specified by NEWVERS,
including the case where an entry is created.
- ,WHENREG=YES
- ,WHENREG=NO
- Use this input parameter to specify whether previous registration
of interest by the requesting user in the data item specified by NAME
is a prerequisite to this request.
- YES
- The request is processed only if your interest in the data item
is already registered. If your interest is not registered, the request
will fail.
- NO
- The request is processed regardless of whether your interest in
the data item is already registered.
- If the entry already exists in the structure, the data and your
interest is updated.
- If the entry does not exist in the structure, a new entry is created,
and your interest is registered.
See the VECTORINDEX parameter description for how to register
interest in a data item.
|