top - download
⟦43c7d73be⟧ Wang Wps File
Length: 64587 (0xfc4b)
Types: Wang Wps File
Notes: CPS/SDS/026
Names: »1592A «
Derivation
└─⟦7985b6947⟧ Bits:30005810 8" Wang WCS floppy, CR 0117A
└─ ⟦this⟧ »1592A «
WangText
(…00……00……00……00…?…02……00……00…?
?…06…>…0a…>…0e…>…01…>…06…=…0b…=…0d…=…01…=…06…4…0c…4…0e…4…01…4…07…3…0b…3…0e…3…00…3
3…05…3…07…2…09…2…0b…2…00…2
2…06…1…08…1…86…1 …02… …02… …02…
…02…CPS/SDS/026
…02…BMN/840105…02……02…
TABLE MANAGEMENT
DETAILED DESIGN SPECIFICATION…02…ISSUE 1…02…CAMPS
4.2.1.4.3 G̲e̲n̲e̲r̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.1.4.3.1\F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This module contains general search procedures used
by TMP process.
The procedures are:
- Find Secondary
- Find Key
- Binary Search
- Compare
a) F̲i̲n̲d̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
Cf 4.2.1.7.g.
b) F̲i̲n̲d̲ ̲K̲e̲y̲
Cf. 4.2.1.7.h.
c) B̲i̲n̲a̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲
Cf. 4.2.1.7.i
d) C̲o̲m̲p̲a̲r̲e̲
4.2.1.7.j.
4.2.1.4.3.2 G̲e̲n̲e̲r̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
This module has four interfaces.
a) F̲i̲n̲d̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
Cf. 4.2.1.7.g.
b) F̲i̲n̲d̲ ̲K̲e̲y̲
Cf. 4.2.1.7.h.
c) B̲i̲n̲a̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲
Cf. 4.2.1.7.i
d) C̲o̲m̲p̲a̲r̲e̲
4.2.1.7.j.
4.2.1.4.3.3 G̲e̲n̲e̲r̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of four procedures
- FIND ̲SECONDARY
- FIND ̲KEY
- BINARY ̲SEARCH
- COMPARE
4.2.1.4.3.4 G̲e̲n̲e̲r̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.7.1
TABLE ̲DESCRIPTION ̲ARRAY cf. 4.1.4.2
ARRAY ̲DESCRIPTION cf. 4.1.4.1
KEY ̲DESCRIPTION cf. 4.1.4.1
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD
TABLE ̲DESCRIPTION ̲ARRAY
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.1.4.3.5 G̲e̲n̲e̲r̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
READ ̲DISK ̲BLOK c.f. 4.1.5.7
a) F̲i̲n̲d̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
All RECORDS in a whole BLOCK is scanned one by
one. Return is made if the key is found or end
of BLOCK is reached.
F̲l̲o̲w̲g̲r̲a̲m̲
Cf. figure 4.2.1.4.3.5-2
b) F̲i̲n̲d̲ ̲K̲e̲y̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
The KEY is found by calling BINARY ̲SEARCH or address
of searched KEY is determined arithmetically. The
way to find the KEY is chosen in accordance to
table organization type. All needed parameters
are found via TMP ̲COROUTINE ̲RECORD.
F̲l̲o̲w̲g̲r̲a̲m̲
None.
c) B̲i̲n̲a̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
Binary search method is used until only one RECORD
is left. ELEMENT is updated and returned to caller.
By comparing searched KEY with a RECORD KEY the
procedure COMPARE is called.
F̲l̲o̲w̲g̲r̲a̲m̲
None
d) C̲o̲m̲p̲a̲r̲e̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
The BYTES are compared one by one starting at the
lowest address. Comparison is continued until the
two compared BYTES are not identical or end of
strings are reached. Completion code is updated
and returned to caller.
F̲l̲o̲w̲g̲r̲a̲m̲
None
4.2.1.4.4. M̲e̲m̲o̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.1.4.4.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure performs search in a memory located
table.
The search type to be performed is selected in accordance
with TMP ̲OPERATION.PARAMS.FUNCTION.
The search consists of finding specified keys and delivery
of associated output.
All parameters needed for a search must be located
in, or referenced by TMP ̲COROUTINE ̲RECORD at invocation
time.
4.2.1.4.4.2 I̲n̲t̲e̲r̲f̲a̲c̲e̲
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
MEMORY ̲SEARCH
(R4 ; " c k Pointer to TMP ̲OPERATION
R5 ; " c k Pointer to TMP ̲COROUTINERECORD
R6); " c d LINK
R0-R3 - k
R7 - d
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
4.2.1.4.4.3 M̲e̲m̲o̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of the following procedures,
- MEMORY ̲SEARCH
- PRIMARY
This procedure searches the keys specified in KEY ̲LIST
and delivers the requested output from the records
with specified key.
The key must be described by TMP ̲COROUTINE ̲RECORD.KEY
and the table must be described by TMP ̲COROUTINE ̲RECORD.TABLE.
The keys in KEY ̲LIST are only searched in the field
containing primary key.
- NEXT ̲PRIMARY
This procedure searches the key specified by ENTRY
and delivers the requested number of records, starting
with the one next to the one found.
The key must be described by TMP ̲COROUTINE ̲RECORD.KEY
and the table must be described by TMP ̲COROUTINE ̲RECORD.TABLE.
COUNT specifies number of records to be delivered.
- IN ̲RECORD
This procedure validates if the keys in KEY ̲LIST are
contained in the record specified by primary key pointed
out by ENTRY.
The keys are validated against the record fields described
by TMP ̲COROUTINE ̲RECORD.SEARCH ̲FORMAT.
The keys existing in record will have associated SINGLE
̲KEY.INFO field set to TRUE at output.
- ONE ̲SECONDARY
This procedure searches the keys specified in KEY ̲LIST
and delivers the requested output from the records
with specified key.
The key must be described by TMP ̲COROUTINE ̲RECORD.KEY
and the table must be described by TMP ̲COROUTINE ̲
RECORD.TABLE.
The keys in KEY ̲LIST are searched in the fields described
by TMP ̲COROUTINE ̲RECORD.SEARCH ̲FORMAT.
For each key in KEY ̲LIST only one record may be delivered.
- ALL ̲SECONDARY
This procedure searches the keys specified in KEY ̲LIST
and delivers the requested output from the records
with specified key.
The key must be described by TMP ̲COROUTINE ̲RECORD.KEY
and the table must be described by TMP ̲COROUTINE ̲RECORD.TABLE.
The keys in KEY ̲LIST are searched in the fields described
by TMP ̲COROUTINE ̲RECORD.SEARCH FORMAT.
For each key in KEY ̲LIST all records having specified
key will be delivered.
4.2.1.4.4.4 M̲e̲m̲o̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD 4.1.4.1
TABLE ̲DESCRIPTION 4.1.4.1
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD (M)
TABLE ̲DESCRIPTION
TABLE ̲DESCRIPTION.STATISTICS (M)
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.1.4.4.5 M̲e̲m̲o̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
MAP ̲PARAMS ̲IN cf. 4.1.5.2
SORT ̲KEYS cf. 4.2.1.4.2
SEND ̲RESPONSE cf. 4.1.5.5
MAP ̲PARAMS ̲OUT cf. 4.1.5.3
FIND ̲KEY cf. 4.2.1.4.3
DELIVER ̲OUTPUT cf. 4.2.1.6.1
COMPARE ̲FIELDS cf. 4.2.1.6.2
FIND ̲SECONDARY cf. 4.2.1.4.3
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
None
F̲l̲o̲w̲g̲r̲a̲m̲s̲
C f. figure 4.2.1.4.4.5-1
Procedure MEMORY ̲SEARCH
EQUIVALENCE(RUNNING ̲COROUTINE,C:TMP ̲COROUTINE
̲RECORD)
Case MAP ̲PARAMS ̲IN : ERROR ̲OK :
- ERROR ? - C.RESPONSE.CC = PARAMETER
̲ADDRESS ̲ERROR
End Case MAP ̲PARAMS ̲IN
Case SORT ̲KEYS ( ) (SORTED ̲KEYS, CC)
- ERROR ? - C.RESPONSE.CC = CC
End Case SORT ̲KEYS
Case C.INPUT.FUNCTION
- SEARCH ̲PRIMARY ̲KEY ? - P̲r̲i̲m̲a̲r̲y̲
- SEARCH ̲NEXT ̲PRIMARY ̲KEY ? - N̲e̲x̲t̲ ̲P̲r̲i̲m̲a̲r̲y̲
- SEARCH ̲IN ̲RECORD ? - I̲N̲ ̲R̲E̲C̲O̲R̲D̲
- SEARCH ̲ONE ̲SECONDARY ̲KEY ? - O̲N̲E̲ ̲S̲E̲C̲O̲N̲D̲A̲R̲Y̲
- SEARCH ̲ALL ̲SECONDARY ̲KEY ? - A̲L̲L̲ ̲S̲E̲C̲O̲N̲D̲A̲R̲Y̲
End Case C.INPUT.FUNCTION
Update TMP Statistics
MAP ̲PARAMS ̲OUT
SEND ̲RESPONSE
Stop
Figure 4.2.1.4.4.5-1
4.2.1.4.5 D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.1.4.5.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This module performs search in tables located at disk.
a) D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲ ̲C̲o̲r̲o̲u̲t̲i̲n̲e̲
One invocation will perform one search request
and send response to caller when search is complete
or an error has been detected.
Depending on function code one of the following
search procedures is executed
- DISK ̲PRIMARY
cf 4.2.1.4.5.1.b
- DISK ̲NEXT ̲PRIMARY
This procedure delivers the No. of keys specified
by TMP ̲OPERATION.PARAMS.COUNT.
The actual primary key must be described by TMP
̲COROUTINE ̲RECORD.KEY and the table must be described
by the TABLE ̲DESCRIPTION identified by TMP ̲COROUTINE
̲RECORD.TD.
The first record to be delivered is the one having
primary key next to the key specified by the pointer
in TMP ̲OPERATION.PARAMS.ENTRY, if this pointer
is NIL delivery is started with first record in
table.
The records being delivered will have increasing
primary keys.
- DISK ̲IN ̲RECORD
This procedure validates if the keys in KEY ̲LIST
are contained in the record specified by primary
key pointed out by ENTRY.
The keys are validated against the record fields
described by TMP ̲COROUTINE ̲RECORD.SEARCH ̲FORMAT.
The keys existing in record will have associated
SINGLE ̲KEY.INFO field set to TRUE at output.
- DISK ̲ONE ̲SECONDARY
This procedure searches the keys specified in KEY
̲LIST and delivers the requested output from the
records with specified key.
The key must be described by TMP ̲COROUTINE ̲RECORD.KEY
and the records in table must be described by TMP
̲COROUTINE ̲RECORD.TABLE.
The table must be described by TMP ̲COROUTINE ̲RECORD.TD.
The keys in KEY ̲LIST are searched in the fields
described by TMP ̲COROUTINE ̲RECORD.SEARCH ̲FORMAT.
For each key in KEY ̲LIST only one record may be
delivered.
- DISK ̲ALL ̲SECONDARY
This procedure searches the keys specified in KEY
̲LIST and delivers the requested output from the
records with specified key.
The key must be described by TMP ̲COROUTINE ̲RECORD.KEY
and the records in table must be described by TMP
̲COROUTINE ̲RECORD.TABLE.
The table must be described in TMP ̲COROUTINE ̲RECORD.TD.
The keys in KEY ̲LIST are searched in the fields
described by TMP ̲COROUTINE ̲RECORD.SEARCH ̲FORMAT.
For each key in KEY ̲LIST all records having specified
key will be delivered.
b) D̲i̲s̲k̲ ̲P̲r̲i̲m̲a̲r̲y̲
This procedure searches the keys specified in KEY
̲LIST and delivers the requested output from the
records with specified key.
The key must be described by TMP ̲COROUTINE ̲RECORD.KEY
and the table must be described by TMP ̲COROUTINE
̲RECORD.TABLE.
The keys in KEY ̲LIST are only searched in the field
containing primary key.
4.2.1.4.5.2 I̲n̲t̲e̲r̲f̲a̲c̲e̲
a) D̲i̲s̲k̲ ̲C̲o̲r̲o̲u̲t̲i̲n̲e̲
Cf. 4.2.3.7.1
b) D̲i̲s̲k̲ ̲P̲r̲i̲m̲a̲r̲y̲
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
DISK ̲PRIMARY
(R4 ; " c k Pointer to TMP ̲OPERATION
R5 ; " c k Pointer to TMP ̲COROUTINE ̲RECORD
R6); " c d LINK
R0-R3 - d
R7 - d
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
c) F̲i̲n̲d̲ ̲D̲i̲s̲k̲ ̲K̲e̲y̲
Cf 4.2.1.7.f
4.2.1.4.5.3 D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of the coroutine
D̲I̲S̲K̲ ̲S̲E̲A̲R̲C̲H̲
and the procedures
D̲I̲S̲K̲ ̲P̲R̲I̲M̲A̲R̲Y̲
F̲I̲N̲D̲ ̲D̲I̲S̲K̲ ̲K̲E̲Y̲
G̲E̲T̲ ̲D̲I̲R̲E̲C̲T̲
This procedure finds a record in a direct organized
disk table. The table is identified by R5%TMP ̲COROUTINE
̲RECORD.TD and the record is identified by the key described
in R5%TMP ̲COROUTINE ̲RECORD ̲KEY.
At output R5%TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD contain
a pointer to the record identified by specified key.
A NIL pointer means that key is not found.
If key has been found OKAY exit is used otherwise ERROR
exit is used.
G̲E̲T̲ ̲I̲N̲V̲E̲R̲S̲E̲
This procedure finds the record identified by BLOCK
̲NO and RECORD ̲NO. At return specified block is read
into memory and TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD
will specify the address of actual record.
G̲E̲T̲ ̲S̲E̲Q̲U̲E̲N̲T̲I̲A̲L̲
This procedure finds a record in a sequential organized
disk table. The table is identified by R5%TMP ̲COROUTINE
̲RECORD.KEY.KEY.
If OVERFLOW ̲SEARCH is FALSE all records in overflow
blocks will be considered not existing.
R5%TMP ̲COROUTINE ̲RECORD.CURRENT ̲BLOCK specifies in
which block the key might be located.
If specified key is higher than highest key at specified
block, search will continue in the preceding blocks
until a key equal to or higher than specified key has
been found or end of tables is reached. R5%TMP ̲COROUTINE
̲RECORD.CURRENT ̲BLOCK will always specify number of
the block residing in memory.
At output R5%TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD contain
a pointer to the record identified by key or the record
having key next to specified key, a NIL pointer means
that key is higher than highest key in table.
If key has been found OKAY exit is used otherwise ERROR
exit is used.
4.2.1.4.5.4 D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1
TABLE ̲DESCRIPTION cf. 4.1.4.1
TMP ̲BUFFER cf. 4.1.4.1
SEARCH ̲BUFFER cf. 4.1.4.2
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD (M)
TABLE ̲DESCRIPTION
TABLE ̲DESCRIPTION.STATISTICS (M)
SEARCH ̲BUFFER
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
N/A
4.2.1.4.4.5 D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
MAP ̲PARAMS ̲IN cf. 4.1.5.2
SORT ̲KEYS cf. 4.2.1.4.2
SEND ̲RESPONSE cf. 4.1.5.5
MAP ̲PARAMS ̲OUT cf. 4.1.5.3
DELIVER ̲OUTPUT cf. 4.2.1.6.1
COMPARE ̲FIELDS cf. 4.2.1.6.2
FIND ̲SECONDARY cf. 4.2.1.4.3
READ ̲DISK ̲BLOCK cf. 4.1.5.7
SPECIAL ̲SEARCH cf. 4.2.1.4.6
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
The flowgrams show the high level control logic for
a disk search.
F̲l̲o̲w̲g̲r̲a̲m̲s̲
Cf. figure 4.2.1.4.5.5-(1 - 7)
Coroutine DISK ̲SEARCH
Initialize Data
EQUIVALENCE(RUNNING ̲COROUTINE,C:TMP ̲COROUTINE ̲RECORD)
FOREVER ̲Loop
WAIT ̲OPSEM (DISK ̲SEARCH ̲SEM)
WAIT ̲SEM (DISABLE ̲SEM)
Case MAP ̲PARAMS ̲IN: ERROR ̲OK:
- ERROR? C.RESPONSE.CC = PARAMETER ̲ADDRESS
̲ERROR
End Case MAP ̲PARAMS ̲IN
Case SORT ̲KEYS()(SORTED ̲ARRAY,CC): ERROR ̲OK
- ERROR? C.RESPONSE = CC
End case SORT ̲KEYS
P̲e̲r̲f̲o̲r̲m̲ ̲D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲
Update TMP Statistics
SEND ̲RESPONSE
MAP ̲PARAMS ̲OUT
SIGNAL ̲SEM (DISABLE ̲SEM)
End FOREVER Loop
Figure 4.2.1.4.5.5-1
P̲e̲r̲f̲o̲r̲m̲ ̲D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲
Case C.INPUT.FUNCTION
- SEARCH ̲PRIMARY ̲KEY? NOT SIC ̲SCD ̲TABLE
?
DISK ̲PRIMARY
SPECIAL ̲SEARCH
- SEARCH ̲NEXT ̲PRIMARY ̲KEY? D̲i̲s̲k̲ ̲N̲e̲x̲t̲ ̲P̲r̲i̲m̲a̲r̲y̲
- SEARCH ̲IN ̲RECORD? D̲i̲s̲k̲ ̲I̲n̲ ̲R̲e̲c̲o̲r̲d̲
- SEARCH ̲ONE ̲SECONDARY ̲KEY? D̲i̲s̲k̲ ̲O̲n̲e̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
- SEARCH ̲ALL ̲SECONDARY ̲KEY? D̲i̲s̲k̲ ̲A̲l̲l̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
End case C.INPUT.FUNCTION
Stop
Figure 4.2.1.4.5.5-2
Procedure DISK ̲PRIMARY: OK
EQUIVALENCE(RUNNING ̲COROUTINE,C:TMP ̲COROUTINE ̲RECORD)
Not Two level? FOR ̲ALL ̲KEYS loop
FIND ̲DISK ̲KEY
FOR ̲ALL ̲KEYS loop Deliver Output
Last key EXIT
FIND ̲DISK ̲KEY End FOR ̲ALL ̲KEYS
loop
Deliver Intermediate
Output in TMP Data Area
Last key EXIT
End FOR ̲ALL ̲KEYS loop
FOR ̲ALL ̲KEYS loop
Sort Keys In Intermediate Output
FOR ̲ALL ̲INTERMEDIATE ̲KEYS loop
FIND ̲DISK ̲KEY
Deliver Output In Caller Data Area
Last Intermediate Key EXIT
End FOR ̲ALL ̲INTERMEDIATE ̲KEYS
Last key EXIT
END FOR ̲ALL ̲KEYS loop
Stop
Figure 4.2.1.4.5.5-3
D̲i̲s̲k̲ ̲N̲e̲x̲t̲ ̲P̲r̲i̲m̲a̲r̲y̲
Find Key Specified by C.ENTRY
DELIVER loop
Find Next Key
Deliver Output
No. Of Found Keys EQ C.COUNT …0f…EXIT…0e…
OR End Of Table Reached
End DELIVER Loop
Stop
Figure 4.2.1.4.5.5-4
D̲i̲s̲k̲ ̲I̲n̲ ̲R̲e̲c̲o̲r̲d̲
Case FIND ̲DISK ̲KEY: ERROR ̲OK
- ERROR? - C.RESPONSE.CC = UNDEFINED ̲KEY
End case FIND ̲DISK ̲KEY
COMPARE ̲FIELDS
C.RESPONSE.CC = OK
Stop
Figure 4.2.1.4.5.5-5
D̲i̲s̲k̲ ̲O̲n̲e̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
FOR ̲ALL ̲BLOCKS loop
FOR ̲ALL ̲KEYS loop
Case FIND ̲SECONDARY (BLOCK ̲END)
(DUMMY): ERROR ̲OK
- ERROR?
- OK? DELIVER ̲OUTPUT
End case FIND ̲SECONDARY
Last Key EXIT
End FOR ̲ALL ̲KEYS loop
All Keys Found OR Last Block EXIT
End FOR ̲ALL ̲BLOCKS loop
Stop
Figure 4.2.1.4.5.5-6
D̲i̲s̲k̲ ̲A̲l̲l̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
FOR ̲ALL ̲KEYS loop
FOR ̲ALL ̲BLOCKS loop
FOR ̲ALL ̲RECORDS loop
Case FIND ̲SECONDARY (BLOCK ̲END)
(DUMMY): ERROR ̲OK
- ERROR? EXIT
- OK? DELIVER ̲OUTPUT
End Case FIND ̲SECONDARY
End ALL ̲RECORDS loop
LAST ̲BLOCK EXIT
END FOR ̲ALL ̲BLOCKS loop
LAST ̲KEY EXIT
End FOR ̲ALL ̲KEYS loop
Stop
Figure 4.2.1.4.5.5-7
4.2.1.4.6 S̲p̲e̲c̲i̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.1.4.6.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This module searches the SICs specified in TMP ̲OPERATION.PARAMS.BUFFER1.
The output will not be the SDL ̲NOs which is present
in the SIC table, but the SDLs associated to the SDL
̲NOs in the SDL table.
TMP ̲OPERATION.PARAMS.MASK1 defines which fields in
SIC record shall be used as read mask at the output
from SIC table ( SDL ̲NO ).
A SIC is existing if its 3 character code preceded
by a space is found or if its first 2 characters preceded
by 2 spaces are found or if its first character preceded
by 3 spaces are found, the SIC version with the highest
No. of significant characters matching with the searched
one will be used.
4.2.1.4.6.2 S̲p̲e̲c̲i̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
SPECIAL ̲SEARCH: OK
(R4 ; " c k Pointer to TMP ̲OPERATION
R5 ; " c k Pointer to TMP ̲COROUTINE ̲RECORD
R6); " c d LINK
R0-R3 - k
R7 - d
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
4.2.1.4.6.3 S̲p̲e̲c̲i̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of the two procedures:
- SPECIAL ̲SEARCH
- SIC ̲SCD ̲SEARCH
This procedure searches the SDLs associated to the
SICs in R5%TMP ̲COROUTINE ̲RECORD.PRIVATE. The SDL ̲NO
associated to each SIC is referenced by actual SINGLE
̲KEY.LINK.
The original SINGLE ̲KEY specified by INPUT%TMP ̲OPERATION.BUFFERR1
will be updated in accordance with TMP interface concept.
4.2.1.4.6.4 S̲p̲e̲c̲i̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1
TABLE ̲DESCRIPTION ̲ARRAY cf. 4.1.4.2
TABLE ̲DESCRIPTION cf. 4.1.4.1
RECORD ̲DESCRIPTION cf. 4.1.4.1
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD (M)
TABLE ̲DESCRIPTION ̲ARRAY
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.1.4.6.5 S̲p̲e̲c̲i̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
DISK ̲PRIMARY cf. 4.2.1.4.5
MODIFY ̲RECORD cf. 4.1.5.1
a) S̲p̲e̲c̲i̲a̲l̲ ̲S̲e̲a̲r̲c̲h̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
The procedure SIC ̲SCD ̲SEARCH is called.
F̲l̲o̲w̲g̲r̲a̲m̲
None
b) S̲I̲C̲ ̲S̲C̲D̲ ̲S̲e̲a̲r̲c̲h̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
Each SIC is searched by calling DISK ̲PRIMARY.
If it is not found third letter is changed to a
space and DISK ̲PRIMARY is called. If it is still
not found second letter is also changed to a space
and DISK ̲PRIMARY is called.
If a SIC is found requested SDLs is written in
TMP data area. All SDLs is converted to SCDs by
calling DISK ̲PRIMARY.
F̲l̲o̲w̲g̲r̲a̲m̲
None.
4.2.1.5 S̲e̲a̲r̲c̲h̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲D̲a̲t̲a̲
4.2.1.5.1 C̲o̲r̲o̲u̲t̲i̲n̲e̲ ̲R̲e̲c̲o̲r̲d̲s̲
Search subpackage consists of two coroutines and must
thus have two coroutine records.
VAR
SEARCH ̲COMMUNICATION ̲COREC,
DISK ̲SEARCH ̲COREC :
TMP ̲COROUTINE ̲RECORD,
cf 4.1.4.2
4.2.1.5.2 S̲e̲a̲r̲c̲h̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲D̲i̲s̲k̲ ̲b̲u̲f̲f̲e̲r̲
Search subpackage has a disk buffer used when searching
in disk located tables
VAR
SEARCH ̲BUFFER: TMP ̲BUFFER;
cf 4.1.4.2
4.2.1.6 C̲o̲m̲m̲o̲n̲ ̲S̲e̲a̲r̲c̲h̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
4.2.1.6.1 D̲e̲l̲i̲v̲e̲r̲ ̲O̲u̲t̲p̲u̲t̲
4.2.1.6.1.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure delivers one record to the application.
The record is written in the area specified by TMP
̲COROUTINE ̲RECORD.INPUT%TMP ̲OPERATION.PARAMS.BUFFER2.
The record to be written is specified by
TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD.
The record parts to be written are specified by
TMP ̲COROUTINE ̲RECORD.WRITE ̲FORMAT.
TMP ̲COROUTINE ̲RECORD.INPUT%TMP ̲OPERATION.PARAMS.BUFFER2
̲LENGTH is decremented by size of the delivered output.
TMP ̲COROUTINE ̲RECORD.INPUT%TMP ̲OPERATION.PARAMS.BUFFER2
is incremented by size of the delivered output.
4.2.1.6.1.2 I̲n̲t̲e̲r̲f̲a̲c̲e̲
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
DELIVER ̲OUTPUT
(R5 ; " c k Pointer to TMP ̲COROUTINE ̲RECORD
R6); " c d LINK
ERROR ̲OK ;
R0-R4 - k
R7 - k
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None.
4.2.1.6.1.3 D̲a̲t̲a̲
R̲e̲f̲e̲r̲e̲n̲c̲e̲d̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None.
4.2.1.6.1.4 D̲e̲l̲i̲v̲e̲r̲ ̲O̲u̲t̲p̲u̲t̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲ ̲D̲e̲s̲i̲g̲n̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
The fields described in TMP ̲COROUTINE ̲RECORD.READ ̲FORMAT
is written one by one until all of them have been written.
F̲l̲o̲w̲g̲r̲a̲m̲
None
4.2.1.6.2 C̲o̲m̲p̲a̲r̲e̲ ̲F̲i̲e̲l̲d̲s̲
4.2.1.6.2.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure compares specified fields in specified
record, to the keys in the KEY ̲LIST specified by TMP
̲COROUTINE ̲RECORD.INPUT%TMP ̲OPERATION.PARAMS.BUFFER1.
If SINGLE ̲KEY.KEY is identical to one of the specified
record fields the SINGLE ̲KEY.INFO is set to TRUE.
4.2.1.6.2.2 I̲n̲t̲e̲r̲f̲a̲c̲e̲
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
COMPARE ̲FIELDS
(R5 ; " c k Pointer to TMP ̲COROUTINE ̲RECORD
R6); " c d LINK
ERROR ̲OK ;
R0-R4 - k
R7 - k
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
4.2.1.6.2.3 D̲a̲t̲a̲
R̲e̲f̲e̲r̲e̲n̲c̲e̲d̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1
KEY ̲LIST cf. (a) 4.2
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD
KEY ̲LIST
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.1.6.2.4 C̲o̲m̲p̲a̲r̲e̲ ̲F̲i̲e̲l̲d̲s̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
BINARY ̲SEARCH cf. 4.2.1.4.3
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
For each SINGLE ̲KEY the RECORD is scanned until an
identical field is found or until end of RECORD.
F̲l̲o̲w̲g̲r̲a̲m̲
N/A
4.2.1.7 S̲e̲a̲r̲c̲h̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
a) R̲e̲c̲e̲i̲v̲e̲ ̲R̲e̲q̲u̲e̲s̲t̲
Requests are received from TMP Monitor via SEARCH
̲SYNCEL cf. 4.2.3.7.1.
b) R̲e̲t̲u̲r̲n̲ ̲R̲e̲s̲p̲o̲n̲s̲e̲
Responses are returned to TMP Monitor via an ANSWER
̲SYNCEL cf. 4.2.3.7.2.
c) R̲e̲c̲e̲i̲v̲e̲ ̲D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲ ̲R̲e̲q̲u̲e̲s̲t̲
A disk search request is received from DISK ̲SEARCH
̲SEM OPERATION ̲SEMAPHORE. The OPERATION received
is of TMP ̲OPERATION type cf. 4.1.4.1 TMP ̲OPERATION.DATA
is of TMP ̲REQUEST ̲TYPE cf. 4.1.4.1.
d) S̲e̲n̲d̲ ̲D̲i̲s̲k̲ ̲S̲e̲a̲r̲c̲h̲ ̲R̲e̲s̲p̲o̲n̲s̲e̲
A response after a disk search request is sent
to SUPPORT ̲RS ̲SEM in a TMP ̲OPERATION.
TMP ̲OPERATION.DATA is of TMP ̲RESPONSE type cf.
4.1.4.1.
e) D̲i̲s̲a̲b̲l̲e̲ ̲S̲e̲a̲r̲c̲h̲
To make it possible to suspend search, SEARCH ̲COMMUNICATION
and DISK ̲SEARCH COROUTINES will wait at DISABLE
̲SEM after they receive a request and signal DISABLE
̲SEM when the request has been completed.
f) F̲i̲n̲d̲ ̲D̲i̲s̲k̲ ̲K̲e̲y̲
F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure finds the KEY specified by TMP ̲COROUTINE
̲RECORD.KEY in the TABLE specified by TMP ̲COROUTINE
̲RECORD.TD.
If OVERFLOW ̲SEARCH is FALSE all records in overflow
blocks will be considered not existing.
If TWO ̲LEVEL ̲SEARCH is FALSE table will be considered
a normal sequential even if it is INVERSE. i.e.
the record being found is the one containing BLOCK
and RECORD No. of the physical record being searched.
TWO ̲LEVEL ̲SEARCH has only significance when searching
in inverse tables.
At return TMP ̲COROUTINE ̲RECORD.TABLE will describe
the RECORDS located in the TMP ̲BLOCK located in
the buffer associated to current coroutine. TMP
̲COROUTINE ̲RECORD.CURRENT ̲BLOCK will contain actual
block number.
TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD will contain
a pointer to the RECORD having specified KEY or
a pointer to the record having KEY next to specified
KEY if the KEY does not exist. A NIL pointer means
that key is higher than highest key in table.
If KEY is found OKAY EXIT is used otherwise ERROR
EXIT is used.
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
FIND ̲DISK ̲KEY
(R2 " c k TWO ̲LEVEL ̲SEARCH boolean specifying
that search in next table level
shall be performed. Has no meaning
for other tables than INVERSE tables.
;R3 " c k OVERFLOW ̲SEARCH boolean specifying
that search in overflow block shall
be performed
;R4 " c k Pointer to TMP ̲OPERATION
;R5 " c k Pointer to TMP ̲COROUTINE ̲RECORD
;R6) : " c d LINK
ERROR ̲OK;
R0-R3,R7 - d
F̲a̲t̲a̲l̲ ̲e̲r̲r̲o̲r̲s̲
None
g) F̲i̲n̲d̲ ̲S̲e̲c̲o̲n̲d̲a̲r̲y̲
F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure finds one record with specified
secondary key, okay return is used when the record
is found.
If no record with specified key is found error
exit is used.
The table to be searched in is specified by TMP
̲COROUTINE ̲RECORD.TABLE and the key to be searched
for is specified by TMP ̲COROURINE ̲RECORD.KEY.
The search always starts in the record specified
by TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD.
TMP ̲COROUTINE ̲RECORD.SEARCH ̲FORMAT specifies the
record fields in which the key could be present.
At output TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD points
to the record containing specified key, TMP ̲COROUTINE
̲RECORD.CURRENT ̲INDEX will contain index of this
record.
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
FIND ̲SECONDARY
(R0 ; " - r FIELD ̲COUNT
R5 ; " c k Pointer to TMP ̲COROUTINE ̲RECORD
R6); " c d LINK
ERROR ̲OK ;
R1-R4 - k
R7 - k
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
h) F̲i̲n̲d̲ ̲K̲E̲Y̲
This procedure searches the primary key specified
by TMP ̲COROUTINE ̲RECORD.KEY in the table specified
by TMP ̲COROUTINE ̲RECORD.TABLE.
At output a pointer to the record having specified
key or to the record having the key next to specified
key is delivered, a NIL pointer means that the
key is greater than highest key in table.
TMP ̲COROUTINE ̲RECORD.CURRENT ̲RECORD and TMP ̲COROUTINE
̲RECORD.CURRENT ̲INDEX are updated to actual values.
Okay exit is used if the key is found, otherwise
error exit is used.
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
FIND ̲KEY
(R5 ; " c k Pointer to TMP ̲COROUTINE ̲RECORD
R2 ; " - r Pointer to record or NIL
R6); " c d LINK
ERROR ̲OK ;
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
i) B̲i̲n̲a̲r̲y̲ ̲S̲e̲a̲r̲c̲h̲
F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure finds the record with specified
key in specified table.
The key is specified by means of a KEY ̲DESCRIPTION.
The table is specified by means of an ARRAY ̲DESCRIPTION.
If the key is found a pointer to the record containing
it is delivered at output, if the key is not found
the pointer identifies the record having the key
next to actual key.
If the table is empty or if actual key is higher
than the key of last record in the table a NIL
pointer is specified at output.
OKAY exit is only used in the case where actual
key is found.
This procedure works on both direct organized tables
and on tables where the records are organized by
means of a pointer array.
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
BINARY ̲SEARCH
(R4 ; " c k Pointer to KEY ̲DESCRIPTION
R7 ; " c k Pointer to ARRAY ̲DESCRIPTION
R0 ; " - k Index of current record or NIL
R2 ; " - r Pointer to current record or NIL
R6); " k d LINK
ERROR ̲OK ;
R1 - k
R3 - k
R5 - k
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
j) C̲o̲m̲p̲a̲r̲e̲
F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure compares two word strings and the
exit value specifies the result of the comparison.An
one word string is considered an INTEGER all other
length strings are considered BYTE ARRAYs.
The exit value will be one of the following:
- LOWER which means that STRING1 has a lower
integer value than STRING2
- SAME which means that STRING1 has same integer
value as STRING2.
- HIGHER which means that STRING1 has a higher
integer value than STRING2
NOTE that the string must not be of 0 words length
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
COMPARE
(R0 ; " c k COUNT
R4 ; " c k Pointer to STRING1
R7 ; " c k Pointer to STRING2
R6); " c d LINK
COMPARE ̲RESULT ;
R1-R3 - k
R5 - k
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
4.2.2 U̲p̲d̲a̲t̲e̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.2.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
Update Subpackage performs all functions causing recoverable
changes in TMP data. The functions are:
- Insert Record
- Remove Record
- Single Change Field Position
- Global Change Field Position
- Single Change Field Value
- Global Change Field Value
- Lock Table
- Unlock Table
- Reorganize
- Abandon Reorganize
- Backup
- System Start Up
- Set GSN flags
- Set System Parameters
For functional specification of a single function (b)
4.1.
When one of these functions except Lock Table have
been completed, the change will always be recoverable.
4.2.2.2 U̲p̲d̲a̲t̲e̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲S̲o̲f̲t̲w̲a̲r̲e̲ ̲S̲t̲r̲u̲c̲t̲u̲r̲e̲
In the following, the software components of each module
are identified.
4.2.2.2.1 U̲p̲d̲a̲t̲e̲ ̲C̲o̲m̲m̲u̲n̲i̲c̲a̲t̲i̲o̲n̲
The Update Communication module consists of the coroutine
UPDATE ̲COMMUNICATION.
4.2.2.2.2 S̲e̲a̲r̲c̲h̲
The Search module consists of one procedure, SEARCH.
4.2.2.2.3 U̲p̲d̲a̲t̲e̲ ̲R̲e̲c̲o̲r̲d̲
The Update Record module consists of nine procedures:
- PERFORM ̲UPDATE
- REMOVE ̲REC
- INSERT ̲REC
- SINGLE ̲FIELD ̲POS
- SINGLE ̲FIELD ̲VAL
- GLOBAL ̲FIELD ̲UPD
- PREP ̲TOTAL ̲REC
- PREP ̲FIELD ̲POS
- PREP ̲FIELD ̲VAL
4.2.2.2.4 U̲p̲d̲a̲t̲e̲ ̲D̲i̲s̲k̲
The Update Disk module consists of the following procedures:
- WRITE ̲WORDS
- WRITE ̲DISK ̲BLOCK
- COPY ̲TABLE
- COPY ̲FILE
- SAVE ̲BLOCK
- WRITE ̲TABLE ̲DESCRIPTION
- UPDATE ̲ORIGINAL ̲TABLE
- GET ̲BACKUP ̲DIR
- INITIALIZE ̲RECOVERY ̲RECORD
- DISABLE ̲SEARCH
- ENABLE ̲SEARCH
- MEMORY ̲RECOVER
- DISK ̲RECOVER
4.2.2.2.5 S̲u̲p̲p̲o̲r̲t̲
The Support module consists of one coroutine and two
procedures:
- SUPPORT Coroutine
- REORGANIZE Procedure
- BACKUP Procedure
4.2.2.2.6 S̲t̲a̲r̲t̲ ̲U̲p̲
Start Up module consists of five procedures:
- SYSTEM ̲START ̲UP
- INITIALIZE ̲TMP
- INITIAL ̲START
- BACK ̲UP ̲START
- RECOVERY ̲START
4.2.2.3 U̲p̲d̲a̲t̲e̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲C̲o̲n̲t̲r̲o̲l̲ ̲L̲o̲g̲i̲c̲
Fig. 4.2.2.3-1 shows the control logic of Update Subpackage.
Requests are received from UPDATE ̲SYNCEL.
The request may be processed by UPDATE ̲COMMUNICATION
coroutine itself or passed on to PERFORM ̲UPDATE procedure
or SUPPORT ̲RQ ̲SEM.
PERFORM ̲UPDATE performs an update by means of the procedures
shown.
SUPPORT receives a request from SUPPORT ̲RQ ̲SEM and
performs it by means of the procedures shown. UPDATE
̲COMMUNICATION and SUPPORT coroutines share a common
buffer by waiting at BUFFER ̲SEM for reserving the buffer
and signalling BUFFER ̲SEM by releasing buffer. UPDATE
̲COMMUNICATION will also pass an abandon reorganize
request on via BUFFER ̲SEM.
Search coroutines are deactivated by waiting for DISABLE
̲LOCK semaphore and then waiting twice for DISABLE ̲SEM
and activated again by signalling DISABLE ̲SEM twice
and then signalling DISABLE ̲LOCK semaphore.
REORGANIZE procedure requests Disk searches by signalling
DISK ̲SEARCH ̲SEM, and receives response from SUPPORT
̲RS ̲SEM.
Fig. 4.2.2.3-1
4.2.2.4 U̲p̲d̲a̲t̲e̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲s̲
4.2.2.4.1 U̲p̲d̲a̲t̲e̲ ̲C̲o̲m̲m̲u̲n̲i̲c̲a̲t̲i̲o̲n̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.2.4.1.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This module awaits requests to the Update Subpackage.
The action to be taken when a request is received depends
on the function referenced by the request.
- Requests requiring much time are passed on to the
Support Module.
- Table Update requests are passed on to the Update
Module.
- Simple Update requests are performed by this module
itself.
Access rights are checked and if they are accepted
function is performed. Completion Code is returned
to caller.
4.2.2.4.1.2 U̲p̲d̲a̲t̲e̲ ̲C̲o̲m̲m̲u̲n̲i̲c̲a̲t̲i̲o̲n̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
This module has following interfaces:
a) R̲e̲c̲e̲i̲v̲e̲ ̲R̲e̲q̲u̲e̲s̲t̲
c.f. 4.2.2.7.1
b) S̲e̲n̲d̲ ̲R̲e̲s̲p̲o̲n̲s̲e̲
c.f. 4.2.2.7.2
c) U̲p̲d̲a̲t̲e̲
The procedure UPDATE is called c.f. 4.2.2.4.3.2.
d) P̲a̲s̲s̲ ̲O̲n̲ ̲T̲o̲ ̲S̲u̲p̲p̲o̲r̲t̲ ̲M̲o̲d̲u̲l̲e̲
TMP ̲COROUTINE ̲RECORD.INPUT is sent to SUPPORT ̲RQ
̲SEM in an INPUT ̲OPERATION c.f. 4.2.2.4.5.2.
e) A̲b̲a̲n̲d̲o̲n̲ ̲R̲e̲o̲r̲g̲a̲n̲i̲z̲e̲
An ABANDON ̲REORGANIZE request is sent to BUFFER
̲SEM operation semaphore in a TMP ̲OPERATION.
4.2.2.4.1.3 U̲p̲d̲a̲t̲e̲ ̲C̲o̲m̲m̲u̲n̲i̲c̲a̲t̲i̲o̲n̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of the coroutine UPDATE ̲COMMUNICATION.
4.2.2.4.1.4 U̲p̲d̲a̲t̲e̲ ̲C̲o̲m̲m̲u̲n̲i̲c̲a̲t̲i̲o̲n̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1
TMP ̲OPERATION cf. 4.1.4.1
TMP ̲REQUEST cf. 4.1.4.1
TABLE ̲DESCRIPTION ̲ARRAY cf. 4.1.4.2
GSN ̲ARRAY cf. 4.1.4.6
SYSTEM ̲PARAMETER ̲ARRAY cf. 4.1.4.5
WORKING ̲STORAGE (a) 11
RECOVERY ̲RECORD cf. 4.2.2.5
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TABLE ̲DESCRIPTION ̲ARRAY (M)
GSN ̲ARRAY (M)
SYSTEM ̲PARAMETER ̲ARRAY (M)
WORKING ̲STORAGE (M)
RECOVERY ̲RECORD (M)
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.2.4.1.5 U̲p̲d̲a̲t̲e̲ ̲C̲o̲m̲m̲u̲n̲i̲c̲a̲t̲i̲o̲n̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
DEFINE ̲SERVICE ̲SYSTEM (b) 3.2.6.6
AWAIT ̲SYNCEL (b) 3.2.6.7
MAP ̲PARAMS ̲IN cf. 4.1.5.2
MAP ̲PARAMS ̲OUT cf. 4.1.5.3
CHECK ̲ADDRESSES cf. 4.1.5.4
SIGNAL ̲OPSEM (b) 3.2.5.7
SEND ̲RESPONSE cf. 4.1.5.5
PERFORM ̲UPDATE cf. 4.2.2.4.3
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
External requests are awaited at UPDATE ̲SYNCEL.
A TMP ̲REQUEST function invokes the UPDATE procedure.
The functions TMP ̲LOCK ̲TABLE
TMP ̲UNLOCK ̲TABLE
TMP ̲SET ̲GSN ̲FLAGS
TMP ̲SET ̲SYSTEM ̲PARAMETERS
are performed by this module itself.
Other functions are passed on to support module via
an OPERATION ̲SEMAPHORE.
F̲l̲o̲w̲g̲r̲a̲m̲
c.f. fig. 4.2.2.4.1.5-1
Start
EQUIVALENCE(RUNNING ̲COROUTINE,C:TMP ̲COROUTINE
̲RECORD)
Coroutine UPDATE ̲COMMUNICATION
Initialize Data
DEFINE ̲SERVICE ̲SYSTEM (UPDATE ̲SYNCEL, OBJECT)
FOREVER LOOP
WAIT ̲OPSEM (POOL ̲SEM)
AWAIT ̲SYNCEL(UPDATE ̲SYNCEL)(C.INPUT, COUNT,
CC): OK
FUNCTION = TABLE FUNCTION ? CHECK TABLE ID
Case C.INPUT.FUNCTION
TMP ̲REQUEST ? - PERFORM
̲UPDATE
TMP ̲LOCK ̲TABLE ? - Lock Table
TMP ̲UNLOCK ̲TABLE ? - Unlock
Table
TMP ̲SET ̲GSN ̲FLAGS ? - Set GSN
Flags
TMP ̲SET ̲SYSTEM ̲PARAMETERS ? - Set System
Parameter
TMP ̲ABANDON ̲REORGANIZE ? - Signal
BUFFER ̲SEM
OTHERWISE ? - Pass On
To Support
End case C.INPUT.FUNCTION
Fig. 4.2.2.4.1.5-1
L̲o̲c̲k̲ ̲T̲a̲b̲l̲e̲
If TABLE ̲DESCRIPTION.ATTRIBUTES.LOCK in TABLE ̲DESCRIPTION
of specified table is FALSE, it is set to TRUE and
TMP ̲COROUTINE ̲RECORD.RESPONSE.INPUT.CC is set to ALREADY
̲LOCKED.
Finally, the procedure SEND ̲RESPONSE is always called.
U̲n̲l̲o̲c̲k̲ ̲T̲a̲b̲l̲e̲
If TABLE ̲DESCRIPTION.ATTRIBUTES.LOCK in TABLE ̲DESCRIPTION
of specified table is TRUE, it is set to FALSE and
TMP ̲COROUTINE ̲RECORD.RESPONSE.INPUT.CC is set to OK.
If LOCK field is FALSE TMP ̲COROUTINE ̲RECORD.RESPONSE.INPUT.CC
is set to NOT ̲LOCKED.
Finally, the procedure SEND ̲RESPONSE is always called.
S̲e̲t̲ ̲G̲S̲N̲ ̲F̲l̲a̲g̲s̲
If access rights are accepted and GSN ̲ID is known,
requested update is performed and TMP SYSTEM DATA are
written to disk.
S̲e̲t̲ ̲S̲y̲s̲t̲e̲m̲ ̲P̲a̲r̲a̲m̲e̲t̲e̲r̲s̲
If access rights are accepted and PARAMETER ̲ID is known,
the data area of caller is mapped in by calling MAP
̲PARAMS ̲IN and specified parameter is updated and TMP
̲SYSTEM ̲DATA is written to disk.
S̲i̲g̲n̲a̲l̲ ̲B̲U̲F̲F̲E̲R̲ ̲S̲E̲M̲
A TMP ̲OPERATION is received from BUFFER ̲SEM. If TMP
̲OPERATION.FUNCTION is equal to REORGANIZE it is changed
to ABANDON ̲REORGANIZE and the Operation is returned
to BUFFER ̲SEM operation semaphore. Then an OK completion
code is returned to caller.
If TMP ̲OPERATION .FUNCTION is not equal to REORGANIZE
no reorganization is going on and a response with CC
set to REJECTED is returned to caller. Final the OPERATION
is returned to BUFFER ̲SEM.
P̲a̲s̲s̲ ̲O̲n̲ ̲T̲o̲ ̲S̲u̲p̲p̲o̲r̲t̲
The request is passed on by signalling the TMP ̲OPERATION
identified by TMP ̲COROUTINE ̲RECORD.INPUT to SUPPORT
̲RQ ̲SEM operation semaphore.
4.2.2.4.2 S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.2.4.2.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
The record identified by primary key specified by TMP
̲OPERATION.PARAMS.ENTRY is searched in the table specified
by TMP ̲COROUTINE ̲RECORD.TD.
The search procedures work in UPDATE ̲COMMUNICATION
̲COREC so all fields in this record describing actual
data being searched in will be updated on return. Cf.TMP
̲COROUTINE ̲RECORD 4.1.4.1.
OKAY EXIT is used if the record is found otherwise
ERROR EXIT is used.
4.2.2.4.2.2 S̲e̲a̲r̲c̲h̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
SEARCH
(R2 ; " c k TWO ̲LEVEL ̲SEARCH : BOOLEAN Specifying
if search shall be performed in second
level table if it exists.
R3 ; " c k OVERFLOW ̲SEARCH : BOOLEAN Specifying
is search shall be performed in overflow
block.
R4 ; " c k Pointer to TMP ̲OPERATION
R5 ; " c k Pointer to TMP ̲COROUTINE ̲RECORD
R6); " c d LINK
ERROR ̲OK ;
R0-R1 - k
R7 - d
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None
4.2.2.4.2.3 S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of the procedure SEARCH:
4.2.2.4.2.4 S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1.1
TABLE ̲DESCRIPTION cf. 4.1.4.1.2
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD (M)
TABLE ̲DESCRIPTION
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.2.4.2.5 S̲e̲a̲r̲c̲h̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
FIND ̲KEY cf. 4.2.1.7.h
FIND ̲DISK ̲KEY cf. 4.2.1.7.f
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
If the table to be searched in is a memory located
table FIND ̲KEY is called otherwise FIND ̲DISK ̲KEY is
called.
The search will go on until:
- Record is found
- A Record with higher key is found
- End of table is reached
F̲l̲o̲w̲g̲r̲a̲m̲
N.A
4.2.2.4.3 U̲p̲d̲a̲t̲e̲ ̲R̲e̲c̲o̲r̲d̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.2.4.3.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This procedure perform update of one or more records
in one table, if records in an associated table are
affected by the update these records will also be updated.
This procedure maintain execution of the functions
descriped by the 6 TMP UPDATE INTERFACES refer ICD/009
sec. 4.1.1.2.
After a succeeded update TMP ̲STATISTICS is updated.
4.2.2.4.3.2 I̲n̲t̲e̲r̲f̲a̲c̲e̲
C̲a̲l̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
PERFORM ̲UPDATE
(R4; "C K Pointer to TMP ̲OPERATION
R5; "C K Pointer to TMP ̲COROUTINERECORD
R0 - R3 - K LINK
R7 - D
4.2.2.4.3.3 U̲p̲d̲a̲t̲e̲ ̲R̲e̲c̲o̲r̲d̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of the nine procedures:
- PERFORM ̲UPDATE
- REMOVE ̲REC
This procedure removes the record specified by
primary key referenced by
TMP ̲OPERATION.PARAMS.ENTRY from the table specified
by TMP ̲COROUTINE ̲RECORD.TD
- INSERT ̲REC
This procedure inserts the record specified by
TMP ̲OPERATION.PARAMS.ENTRY in the table specified
by TMP ̲COROUTINE ̲RECORD.TD, in accordance to its
primary key.
- SINGLE ̲FIELD ̲POS
In the table specified by TMP ̲COROUTINE ̲RECORD.TD
the record identified by the primary key specified
by TMP ̲OPERATION.PARAMS.ENTRY is found.
Then the fields specified by TMP ̲COROUTINE ̲RECORD.WRITE
̲FORMAT having the value specified by TMP ̲COROUTINE
̲RECORD.PARAMS.BUFFER1 are substituted with the
value specified by
TMP ̲OPERATION.PARAMS.BUFFER2.
If actual fields are repeated fields BUFFER1 or
BUFFER2 may be a NIL pointer. This means respectively
that a new simple field shall be inserted or a
simple field shall be removed.
- GLOBAL ̲FIELD ̲UPD
In the table specified by TMP ̲COROUTINE ̲RECORD.TD
all records will be scanned and possibly updated.
The kind of update depends on function code.
- Function code equal to GLOBAL ̲FIELD ̲POS causes
record updates as in SINGLE ̲FIELD ̲POS
- Function code equal to GLOBAL ̲FIELD ̲VAL causes
record updates as in SINGLE ̲FILED ̲VAL
- PREP ̲TOTAL ̲REC
Input parameters for the function REMOVE RECORD
or INSERT ̲RECORD are checked and prepared for actual
function.
The pointer in TMP ̲OPERATION.PARAMS.ENTRY is converted
to TMP address and it is checked that it points
to an area big enough to contain the expected input
parameter.
If function is INSERT ̲RECORD actual RECORD-DESCRIPTION
is copied into TMP ̲COROUTINE ̲RECORD.WRITE ̲FORMAT.
- PREP ̲FIELD ̲POS
Input parameters for the function SINGLE ̲CHANGE
̲FIELD ̲POSITION or GLOBAL ̲CHANGE ̲FIELD ̲POSITION
are checked and prepared for actual function.
The pointer in TMP ̲OPERATION.PARAMS.BUFFERl is
converted to TMP address and it is checked that
it points to an area big enough to contain the
expected input parameter.
If function is SINGLE ̲CHANGE ̲FIELD ̲POSITION the
pointer in TMP ̲COROUTINE ̲RECORD.ENTRY is checked
and converted in same way as BUFFER1 pointer.
It is checked that MASK1 does not reference primary
key or an organized secondary key.
Actual RECORD ̲DESCRIPTION is modified in accordance
with MASK1 and copied into TMP ̲COROUTINE ̲RECORD.WRITE
̲FORMAT.
- PREP ̲FIELD ̲VAL
Input parameters for the function SINGLE ̲CHANGE
̲FIELD ̲VALUE or GLOBAL ̲CHANGE ̲FIELD ̲VALUE are checked
and prepared for actual functions.
The pointers in TMP ̲OPERATION.PARAMS.BUFFER1 and
BUFFER2 are converted to TMP addresses and it is
checked that they points to areas big enough to
contain the expected input parameters.
If function is SINGLE ̲CHANGE ̲FIELD ̲VALUE pointer
in TMP ̲COROUTINE ̲RECORD.ENTRY is checked and converted
in same way as BUFFER1 and BUFFER2 pointers.
It is checked that MASK1 does not reference primary
key or an organized secondary key.
Actual RECORD ̲DESCRIPTION is modified in accordance
with MASK1 and copied into TMP ̲COROUTINE ̲RECORD.WRITE
̲FORMAT.
4.2.2.4.3.4 U̲p̲d̲a̲t̲e̲ ̲R̲e̲c̲o̲r̲d̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1.17
TABLE ̲DESCRIPTION ̲ARRAY cf. 4.1.4.2.2
TMP ̲BLOCK cf. 4.1.4.1.4
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD (M)
TABLE ̲DESCRIPTION ̲ARRAY (M)
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.2.4.3.5 U̲p̲d̲a̲t̲e̲ ̲R̲e̲c̲o̲r̲d̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
MAP ̲PARAMS ̲IN cf. 4.1.5.2
MAP ̲PARAMS ̲OUT cf. 4.1.5.3
SEARCH cf. 4.2.2.4.2
SAVE ̲BLOCK cf. 4.2.2.4.4
DISK ̲RECOVER cf. 4.2.2.4.4
MEMORY ̲RECOVER cf. 4.2.2.4.4
READ ̲DISK ̲BLOCK cf. 4.1.5.7
CHECK ̲ADDRESSES cf. 4.1.5.4
MODIFY ̲RECORD cf. 4.1.5.1
FIND ̲SECONDARY cf. 4.2.1.4.3
INSERT cf. 4.1.5.6
BINARY ̲SEARCH cf. 4.2.1.4.3
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
The UPDATE procedure performs one update request at
a time.
When a single disk block has been updated, it is moved
to WORKING ̲AREA file by the procedure SAVE ̲BLOCK.
The update is effective after one of the procedures
MEMORY ̲RECOVER or DISK ̲RECOVER have been called.
F̲l̲o̲w̲g̲r̲a̲m̲s̲
c.f. fig. 4.2.2.4.3.5-1
Start
Procedure UPDATE ̲RECORD: OK
EQUIVALENCE(RUNNING ̲COROUTINE,C:TMP ̲COROUTINE
̲RECORD)
WAIT ̲OPSEM(BUFFER ̲SEM)(BUFFER ̲HANDLE)
BUFFER ̲HANDLE.FUNCTION EQ REORGANIZE? C.RESPONSE.CC
= REJECTED
Case MAP ̲PARAMS ̲IN: ERROR ̲OK:
- ERROR?
End case MAP ̲PARAMS ̲IN
Case C.INPUT.FUNCTION
- REMOVE ̲RECORD
? - REMOVE ̲RECORD
- INSERT ̲RECORD
? - INSERT ̲RECORD
- SINGLE ̲CHANGE ̲FIELD ̲POSITION
? - One Field Pos.
- GLOBAL ̲CHANGE ̲FIELD ̲POSITION
? - All Field Pos.
- SINGLE ̲CHANGE ̲FIELD ̲VALUE
? - One Field Value
- GLOBAL ̲CHANGE ̲FIELD ̲VALUE
? - All Field Value
End C.INPUT.FUNCTION
INCREMENT (TABLE ̲DESCRIPTION(C.INPUT.TABLE
̲ID).STATISTICS.
UPDATE ̲REQUESTS) AND UPDATE TIME STAMP
MAP ̲PARAMS ̲OUT
SIGNAL ̲OPSEM(BUFFER ̲SEM,BUFFER ̲HANDLE)
MOVE ̲WORKING ̲STORAGE
SEND ̲RESPONSE
Stop
Fig. 4.2.2.4.3.5-1
4.2.2.4.4 U̲p̲d̲a̲t̲e̲ ̲D̲i̲s̲k̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.2.4.4.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This module consist of procedures used for updating
data on disk
a) WRITE ̲WORDS
This procedure copies COUNT words from memory to
specified file.
b) WRITE ̲DISK ̲BLOCK
This procedure copies the TMP ̲BLOCK located in
the TMP ̲BUFFER specified by TMP ̲COROUTINE ̲RECORD.BUFFER
to its original location in the file specified
by TMP ̲BUFFER.FILE.
TMP ̲BUFFER.BLOCK ̲COUNT specifies no of actual block;
c) COPY ̲TABLE
Table identified by TMP ̲COROUTINE ̲RECORD.TD%TABLE
̲
DESCRIPTION is copied to specified output file.
The table is copied to same block numbers in output
file as they where read from in input file.
d) COPY ̲FILE
INPUT FILE is copied to OUTPUT FILE.
The copying are performed via the TMP ̲BUFFER specified
by TMP ̲COROUTINE ̲RECORD.BUFFER.
If an disk operation returns in error exit and
completion code is not EOF got by a 'read' command,
error exit is used and the IO system completion
code is returned, otherwise okay exit is used.
e) SAVE ̲BLOCK
This procedure copies the TMP ̲BUFFER specified
by TMP ̲COROUTINE ̲RECORD.BUFFER to UPD ̲RECOVERY
file.
UPD ̲RECOVERY.COMPLETE1 and UPD ̲RECOVERY.COMPLETE2
are set to the value of COMPLETE.
If COMPLETE is TRUE, UPD ̲RECOVER is written to
disk.
f) WRITE ̲TABLE ̲DESCRIPTION
This procedure copies RECOVERY ̲RECORD.ATTRIBUTES
into ATTRIBUTES of the TABLE ̲DESCRIPTION identified
by TMP ̲COROUTINE ̲RECORD.TD and into all TABLE ̲DESCRIPTIONS
associated to this one.
All affected TABLE ̲DESCRIPTIONS are written to
disk in both memory files. First in MEMORY ̲SAVE
and then in MEMORY ̲WORK.
g) UPDATE ̲ORIGINAL ̲TABLE
This procedure copies a table from UPD ̲WORK ̲AREA
into its original file.
The table to be copied is identified by TMP ̲COROUTINE
̲RECORD.TD
h) GET ̲BACK ̲UP ̲DIR
This procedure opens the directory containing backup
of TMP files. If LAST ̲ISSUE is TRUE the latest
backup version is opened, otherwise the oldest
version is opened.
If a disk access command returns a completion code
different from OK ERROR ̲EXIT is used and the completion
code is passed on.
i) MEMORY ̲RECOVER
This procedure copies the updated parts of memory
to disk into both memory files.
If SINGLE ̲ENTRY is TRUE only TMP ̲COROUTINE ̲RECORD.CURRENT
̲RECORD is copied. Otherwise the whole table is
copied.
TABLE ̲DESCRIPTION.ATTRIBUTES identified by TMP
̲COROUTINE ̲RECORD.TD is timestamped and copied to
disk.
j) DISK ̲RECOVER
This procedure copies the updated blocks of a disk
table into the file containing actual table.
TABLE ̲DESCRIPTION.ATTRIBUTES identified by TMP
̲COROUTINE ̲RECORD.TD and its associated TABLE ̲ATTRIBUTES
is timestamped and copied to disk.
UPD ̲RECOVER.COUNT specifies no of blocks to be
copied. UPD ̲RECOVER.R ̲TYPE specifies if the whole
table shall be copied or if only a number of single
blocks shall be copied.
k) INITIALIZE ̲RECOVERY ̲RECORD
This procedure initializes the specified RECOVERY
̲RECORD so that it is ready for a new update of
a disk table or a support function.
l) DISABLE ̲SEARCH
This procedure disables all search functions untill
the procedure ENABLE ̲SEARCH are called.
m) ENABLE ̲SEARCH
This procedure enables search functions.
It must only be called when the procedure DISABLE
̲SEARCH has been called.
4.2.2.4.4.2 U̲p̲d̲a̲t̲e̲ ̲D̲i̲s̲k̲ ̲M̲o̲d̲u̲l̲e̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
a) WRITE ̲WORDS
(R0 "C D COUNT specifying no. of words to
be written
;R1 "C D WORDS pointer the first word to
be written
;R2 "C D FILE ̲WORD ̲ADDRESS address of first
word in file to be modified
;R4 "C D FILE:TMP ̲FILE ̲ID ̲TYPE specifying
output file
;R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
;R6); "C D LINK
R3, R7 - D
b) WRITE ̲DISK ̲BLOCK
(R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
;R7 "- R CC
;R6); "C D LINK
R0-R4 - D
c) COPY ̲TABLE
(R4 "C D Pointer to file description of
OUTPUT FILE
R5; "C K Pointer to TMP ̲COROUTINE ̲RECORD
R6); "C D LINK
R0 - R3, R7 - D
d) COPY ̲FILE
(R3 "C D Pointer to file description of
INPUT FILE
R4; "C D Pointer to file description of
OUTPUT FILE
R5; "C K Pointer to TMP ̲COROUTINE ̲RECORD
R7; - R CC
R6; "C D LINK
ERROR ̲OK ;
R0 - R2, R7 - D
f) WRITE ̲TABLE ̲DESCRIPTION
(R4 "C D Pointer to RECOVERY ̲RECORD
;R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
;R6) ; "C D LINK
R0-R3 - D
g) UPDATE ̲ORIGINAL ̲TABLE
(R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
;R6); "C D LINK
R0-R4 - D
h) GET ̲BACK ̲UP ̲DIR
(R3 "C D LAST ̲ISSUE : BOOLEAN
;R4 "- R FDCB INDEX of searched directory
;R7 "- R CC completion code returned from
a system call
R6): "C K LINK
ERROR ̲OK ;
R0-R2, R5 - D
i) MEMORY ̲RECOVER
(R3 "C D SINGEL ̲ENTRY : BOOLEAN specifying
if more than one record shall be
written
(R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
R6); "C D LINK
R0 - R2, R4, R7 - D
j) DISK ̲RECOVER
(R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
R6); "C D LINK
R0 - R4, R7 - D
k) INITIALIZE ̲RECOVERY ̲RECORD
(R4 "C K Pointer to TMP ̲OPERATION
;R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
;R7 "C K Pointer to RECOVERY ̲RECORD
R6); "C D LINK
R0-R5 - D
l) DISABLE ̲SEARCH
(R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
;R6); "C D LINK
R0-R4, R7 - D
m) ENABLE ̲SEARCH
(R5 "C K Pointer to TMP ̲COROUTINE ̲RECORD
;R6); "C D LINK
R0-R4, R7 - D
4.2.2.4.4.3 U̲p̲d̲a̲t̲e̲ ̲D̲i̲s̲k̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of the thirteen export procedures
described in 4.2.2.4.4.(1-2).
4.2.2.4.4.4 U̲p̲d̲a̲t̲e̲ ̲D̲i̲s̲k̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1
TMP ̲BLOCK cf. 4.1.4.1
TABLE ̲DESCRIPTION ̲ARRAY cf. 4.1.4.2
FILE ̲DESCR ̲ARRAY cf. 4.1.4.2
All TMP Files (a) 11
RECOVERY ̲RECORD c.f. 4.2.2.5.1
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
TMP ̲COROUTINE ̲RECORD (M)
TABLE ̲ARRAY (M)
FILE ̲DESCR ̲ARRAY (M)
All TMP Files (M)
RECOVERY ̲RECORD (M)
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.2.4.4.5 U̲p̲d̲a̲t̲e̲ ̲D̲i̲s̲k̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
READ ̲BYTES (d) 4.3.6.1
MODIFY ̲BYTES (d) 4.3.6.2
4.2.2.4.5 S̲u̲p̲p̲o̲r̲t̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.2.4.5.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This module performs reorganization and backup of TMP
data.
- Reorganize
A reorganization means that all tables not reorganized
after each update is reorganized e.g. unused table
entries are removed and entries from overflow areas
are inserted at their proper positions.
Index tables and secondary key organized tables
are also reorganized.
After a completed reorganization every disk block
will contain the number of entries defined at system
generation.
- Backup
Backup means that all TMP data are copied from
online to SYS ̲GEN offline volume.
- SYSTEM ̲START ̲UP
From initialization time until this command has
been received, memory table updates will only effect
the memory version (i.e. no update of disk version
is performed).
4.2.2.4.5.2 S̲u̲p̲p̲o̲r̲t̲ ̲M̲o̲d̲u̲l̲e̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
The support module receives requests from BUFFER ̲SEM
or SUPPORT ̲RQ ̲SEM operation semaphores in an operation
of TMP ̲OPERATION type. Function code is REORGANIZE
or BACKUP or SYSTEM ̲START ̲UP.
The support module requests disk searches by sending
a TMP ̲OPERATION to DISK ̲SEARCH ̲SEM and awaits associated
response by SUPPORT ̲RS ̲SEM.
Support module reserves UPDATE ̲BUFFER by receiving
a TMP ̲OPERATION from BUFFER ̲SEM before it uses it.
Same TMP ̲OPERATION is returned to BUFFER ̲SEM after
each completed request and after each reorganized table.
During reorganization, the TMP ̲OPERATION received from
BUFFER ̲SEM may contain an ABANDON ̲REORGANIZE request,
all other requests are received from SUPPORT ̲RQ ̲SEM.
4.2.2.4.5.3 S̲u̲p̲p̲o̲r̲t̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of four components.
- SUPPORT Coroutine
- REORGANIZE TABLE
This procedure reorganize a table.
TMP ̲COROUTINE ̲RECORD.TD points to TABLE ̲DESCRIPTION
of actual table.
If an abandon request is received error exit is
used.
- REORGANIZE ̲INVERSE ̲TABLE
This procedure reorganizes an inverse table.
At call all the new inverse ref records must reside
in SUPPORT ̲RECOVERY file, with first record starting
in address null.
TMP ̲COROUTINE ̲RECORD.TD%TABLE ̲DESCRIPTION.INVERSE
̲TABLE shall point at TABLE ̲DESCRIPTION of inverse
table.
No of records must be MAX ̲ENTRIES-ATTRIBUTES.LOAD.
- BACKUP
This procedure copies all TMP files with, KIND
= K ̲
BACKUP, from the mirroed disk to the ofline volume
named SYSGEN.
4.2.2.4.5.4 S̲u̲p̲p̲o̲r̲t̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD cf. 4.1.4.1
FILE ̲DESCR ̲ARRAY cf. 4.1.4.2
SUBPROCESS ̲DESCR ̲ARRAY cf. 4.1.4.2
TMP ̲FILES (a) 11
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
FILE ̲DESCR ̲ARRAY (M)
SUBPROCESS ̲DESCR ̲ARRAY
TMP ̲FILES (M)
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.2.4.5.5 S̲u̲p̲p̲o̲r̲t̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
READ ̲BYTES (d) 4.3.6.1
MODIFY ̲BYTES (d) 4.3.6.2
WAIT ̲OPSEM (c) 3.2.5.7
SIGNAL ̲OPSEM (c) 3.2.5.8
SEND ̲RESPONSE cf. 4.1.5.5
a) S̲u̲p̲p̲o̲r̲t̲ ̲c̲o̲r̲o̲u̲t̲i̲n̲e̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
SUPPORT coroutine awaits requests at
SUPPORT ̲RQ ̲SEM when it is inactive.
If a reorganize request is received, the procedure
REORGANIZE is called.
If a backup request is received, the procedure
BACKUP is called.
After return from these procedures SEND ̲RESPONSE
is called and a new request is awaited.
F̲l̲o̲w̲g̲r̲a̲m̲
None
b) R̲e̲o̲r̲g̲a̲n̲i̲z̲e̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
During reorganization a copy of the table to be
reorganized is made. The tables in this file which
requires reorganization are reorganized one by
one.
S̲i̲m̲p̲l̲e̲ ̲T̲a̲b̲l̲e̲s̲
The records in table are reorganized by requesting
SEARCH ̲NEXT ̲PRIMARY from the DISK ̲SEARCH coroutine.
In this way, the records will be delivered in
sorted order independent of if they are located
in overflow area.
The number of keys requested at a time is the same
as keys per block read from TABLE ̲DESCRIPTION.
For each block associated index in index table
is updated to the value of last key in block.
These steps are repeated for all blocks in a table
and finally TABLE ̲DESCRIPTION is updated with LOAD
and OVERLOAD.
When all tables in a file are reorganized FILE
̲DESCRIPTIONs are changed so the reorganized copy
is referenced instead of the original file.
Now the copy is copied to its original file and
FILE ̲DESCRIPTIONS are changed to their original
versions.
During manipulation with FILE ̲DESCRIPTION Search
will be disabled.
After manipulation, Search is enabled.
I̲n̲v̲e̲r̲t̲e̲d̲ ̲T̲a̲b̲l̲e̲s̲
A special problem is when the table to be reorganized
is referenced by an organized secondary key.
In this case each block will require additional
manipulation as follows.
The pointer array referencing the records is initialize
for each block.
A list of all secondary keys and their associated
(block,record number) references is created.
The copy of the secondary key table is sorted and
formatted in blocks and its index table is updated.
A̲b̲a̲n̲d̲o̲n̲ ̲R̲e̲o̲r̲g̲a̲n̲i̲z̲e̲
Each time a table has been reorganized, a TMP ̲OPERATION
is signalled to BUFFER ̲SEM and a new TMP ̲OPERATION
is received.
Before the TMP ̲OPERATION is signalled to BUFFER
̲SEM,
TMP ̲OPERATION.FUNCTION is set to REORGANIZE.
If the new TMP ̲OPERATION requests an ABANDON ̲REORGANIZE
function, reorganization is stopped and a response
is sent.
c) B̲a̲c̲k̲u̲p̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
All TMP files at On Line volume are copied to associated
files at SYS ̲GEN Of Line volume by using the procedure
COPY ̲FILE.
The SYS ̲GEN volume contains two incarnations of
TMP files. A time stamp in MEMORY ̲SAVE ̲FILE tells
which of them is the oldest Back-up copy.
The new Back-up copy to be made is always written
at the oldest incarnation.
After Backup is completed CPS ̲HEADER file at SYS
̲
GEN volume is updated with time for this backup.
F̲l̲o̲w̲g̲r̲a̲m̲
None.
4.2.2.4.6 S̲t̲a̲r̲t̲ ̲U̲p̲ ̲M̲o̲d̲u̲l̲e̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
4.2.2.4.6.1 F̲u̲n̲c̲t̲i̲o̲n̲a̲l̲ ̲S̲p̲e̲c̲i̲f̲i̲c̲a̲t̲i̲o̲n̲
This module performs functions related to Start Up
and Initialization.
Three Start Up types are supported.
- Current Start
- Backup Start
- Initial Start
a) C̲u̲r̲r̲e̲n̲t̲ ̲S̲t̲a̲r̲t̲
Current Start is performed by the Start Up types
COLD, WARM1, WARM2.
Memory is reloaded from On Line volume. If TMP
RECOVERY ̲FILE contains consistent data, these are
copied to appropriate locations. Memory resident
data will be copied both to disk and into memory.
b) B̲a̲c̲k̲u̲p̲ ̲S̲t̲a̲r̲t̲
The Backup Start is performed by the Start Up type
DEAD2.
Online volume is reloaded from Of Line volume using
Last Backup version of TMP data.
Then memory is reloaded from online volume.
c) I̲n̲i̲t̲i̲a̲l̲ ̲S̲t̲a̲r̲t̲
The Initial Start is performed by the Start Up
type DEAD1.
On Line volume is reloaded from SYS ̲GEN volume
using Initial TMP data.
Memory is reloaded from On Line volume.
d) I̲n̲i̲t̲i̲a̲l̲i̲z̲a̲t̲i̲o̲n̲
The TMP process is initialized.
The components to initialize are:
- Coroutines
- Coroutine Semaphores
- Coroutine Operations
- Synchronization Elements
- File Descriptions
- SUBPROCESS ̲DESCR ̲ARRAY
4.2.2.4.6.2 S̲t̲a̲r̲t̲ ̲U̲p̲ ̲M̲o̲d̲u̲l̲e̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
SYSTEM ̲START ̲UP
(R0; "C D MODE : START ̲UP ̲TYPE
R6); "C D LINK
R0 - R3, R5 - D
R7 - D
F̲a̲t̲a̲l̲ ̲E̲r̲r̲o̲r̲s̲
None.
4.2.2.4.6.3 S̲t̲a̲r̲t̲ ̲U̲p̲ ̲M̲o̲d̲u̲l̲e̲ ̲C̲o̲m̲p̲o̲n̲e̲n̲t̲s̲
This module consists of four procedures:
- INITIAL ̲START
- This procedure reloads TMP initial data from SYS
̲GEN volume and loads system parameters and memory
table data into memory.
- BACKUP ̲START
This procedure reloads TMP backup data from SYS
̲GEN volume and loads system parameters and memory
table data into memory.
- RECOVERY ̲START
If UPD ̲RECOVER saved on disk specifies that an
update was going on and possibly was not completed
before the TMP process was stopped, this procedure
copies the updated blocks of a disk table into
the file containing actual table.
If SUPPORT ̲RECOVERY saved on disk specifies that
an reorganization was going on and one or more
tables may be inconcistent, this tables will be
made consistent, but it is not garantied that all
tables are reorganized.
If an error occur the TMP process is retired.
- INITIALIZE ̲TMP
First SUBPROCESS ̲DESCR ̲ARRAY is initialized with
data got from TMP ̲LOAD ̲TABLE.
Then data and code associated to the COROUTINE
̲MONITOR are initialized.
After this initialization TMP will await its requests
as specified in TMP interface specification.
Note that no requests can be received before the
procedure START ̲TMP has been called.
- RELOAD
This procedure copies all TMP files with, KIND
= K ̲BACKUP, from the ofline volume SYSGEN to the
mirrored disks
- READ ̲SYSTEM ̲PARAMETERS
This procedure reads an consistent version of system
parameters and global number series (system data)
into memory, and ensures that both copies on mirrored
disks contain this version.
An version having identical timestamps in start
and end will be considered consistent.
If RECOVERY is TRUE both memory save and memory
work may contain a version of system data.
If RECOVERY is FALSE only memory save version may
contain a version of system data.
If an disk read error occur or no consistent version
is found the TMP process will be retired.
- READ ̲MEMORY ̲DATA
This procedure reads an consistent version of memory
data into memory, initializes it and saves it in
both memory save and memory work files. If RECOVERY
is TRUE the data must be checked, otherwise the
data in memory save file will be considered consistent.
If RECOVERY is TRUE and data shall be checked it
is checked that time stamp in TABLE ̲DESCRIPTION.ATTRIBUTES
are identical with the associated one in memory
work file.
If they are not identical it is checked that all
associated tables in memory save file has identical
time stamp. If they have, memory save version is
used, otherwise memory work version is used.
If an error occur the TMP process is retired.
- OPEN ̲FILES
This procedure opens all TMP files descriped by
FILE ̲DESCR ̲ARRAY. The FDCB index received when
opening the files are saved in associated FILE
̲
DESCRIPTION.FILE.
If an disk access error occur the TMP PROCESS is
retired;
4.2.2.4.6.4 S̲t̲a̲r̲t̲ ̲U̲p̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲a̲t̲a̲
D̲a̲t̲a̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
TMP ̲COROUTINE ̲RECORD c.f. 4.1.4.1
FILE ̲DESCR ̲ARRAY c.f. 4.1.4.2
GENERAL ̲OPERATIONS c.f. 4.1.4.2
Coroutine Semaphores c.f. 4.1.4.2
SUBPROCESS-DESCR ̲ARRAY c.f. 4.1.4.2
UPDATE ̲BUFFER c.f. 4.1.4.2
E̲x̲t̲e̲r̲n̲a̲l̲ ̲D̲a̲t̲a̲
FILE ̲DESCR ̲ARRAY (M)
GENERAL ̲OPERATION (M)
Coroutine Operations (M)
SUBPROCESS ̲DESCR ̲ARRAY (M)
UPDATE ̲BUFFER (M)
L̲o̲c̲a̲l̲ ̲D̲a̲t̲a̲
None
4.2.2.4.6.5 S̲t̲a̲r̲t̲ ̲U̲p̲ ̲M̲o̲d̲u̲l̲e̲ ̲D̲e̲s̲i̲g̲n̲
E̲x̲t̲e̲r̲n̲a̲l̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
READ ̲BYTES (d) 4.3.6.1
MODIFY ̲BYTES (d) 4.3.6.2
LOOK ̲UP (d) 4.3.5.2
INIT ̲COMON (c) 3.2.5.10
INIT ̲COROUTINE (c) 3.2.5.11
INIT ̲COROUTINE ̲SEMAPHORE (c) 3.2.5.12
SIGNAL ̲OPSEM (c) 3.2.5.13
SEND ̲RESPONSE cf 4.1.5.5
a) S̲t̲a̲r̲t̲ ̲U̲p̲
N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲
The procedure OPEN ̲FILES is called.
The procedure INITIALIZE ̲TMP is called.
Then selected processing depends on MODE. If MODE
is DEAD1, INITIAL ̲START is called.
If MODE is DEAD2, BACK ̲UP ̲START is called.
If MODE is COLD, WARM1 or WARM2 RECOVERY ̲START
is called.
F̲l̲o̲w̲g̲r̲a̲m̲
None
4.2.2.5 U̲p̲d̲a̲t̲e̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲D̲a̲t̲a̲
Refer TMP.D*UPD.D*VAR.D
4.2.2.6 C̲o̲m̲m̲o̲n̲ ̲U̲p̲d̲a̲t̲e̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲P̲r̲o̲c̲e̲d̲u̲r̲e̲s̲
N/A
4.2.2.7 U̲p̲d̲a̲t̲e̲ ̲S̲u̲b̲p̲a̲c̲k̲a̲g̲e̲ ̲I̲n̲t̲e̲r̲f̲a̲c̲e̲
4.2.2.7.1 R̲e̲q̲u̲e̲s̲t̲ ̲F̲r̲o̲m̲ ̲T̲M̲P̲ ̲M̲o̲n̲i̲t̲o̲r̲
The Update Subpackage receives requests from TMP Monitor
subpackage via the synchronization element UPDATE ̲SYNCEL.
The requests received are:
- TMP Request
- Lock Table
- Unlock Table
- Reorganize
- Abandon Reorganize
- Backup
- System Start Up
- Set GSN Flags
- Set System Parameter
The info received from synchronization element is for
each request type described in 4.2.3.7.1.
The interface specifications are given in (b) 4.1.
4.2.2.7.2 R̲e̲s̲p̲o̲n̲s̲e̲ ̲T̲o̲ ̲T̲M̲P̲ ̲M̲o̲n̲i̲t̲o̲r̲
When a function is completed, a response is sent to
TMP Monitor.
The response is sent to the synchronization element
associated to the subprocess specified by the request.
The response will contain SOCB ̲INDEX and Completion
Code in a RESPONSE ̲INFO record c.f. 4.1.4.1.2.
4.2.2.7.3 S̲e̲a̲r̲c̲h̲ ̲R̲e̲q̲u̲e̲s̲t̲
The Support module requests disk search by Search Subpackage
by sending operations to DISK ̲SEARCH ̲SEM.
Response is received from SUPPORT ̲RS ̲SEM.
4.2.2.7.4 D̲i̲s̲a̲b̲l̲e̲ ̲S̲e̲a̲r̲c̲h̲
The Update Subpackage may disable search by waiting
for DISABLE ̲LOCK semaphore and then twice at DISABLE
̲SEM. The search is enabled again by signalling twice
to DISABLE ̲SEM and then signalling DISABLE ̲LOCK semaphore.
