071

Explanation
An error occurred when the operator initiated a
restart.
Register 15 contains a hexadecimal reason code that
explains the error: Code
Explanation
00
The operator replied "ABEND" to message BLW004A.
04
The operator replied "REPAIR" to message BLW004A.
08
A programming error caused the program that was currently running
to branch directly to the restart interruption handler.
0C
The operator selected RESTART option 0 on a processor that is
in a valid spin waiting for a global resource held by a second processor
that is in an erroneous disabled loop. A DSGNL RESTART command, directed
to the second processor, failed.
10
The system started writing logrec data set error records after
detecting an excessive spin loop condition. The interrupted unit of
work continues to run.
20
The system abnormally ended the program that was causing a system
excessive spin loop.  Recovery routines are not allowed to retry because
the recovery action for the spin loop was TERM.
30
The system abnormally ended the program that was causing a system
excessive spin-loop condition. Recovery routines are allowed to retry
because the recovery action for the spin loop was ABEND.

System action
The system does one of the following: For reason code X'00', X'08', X'0C', X'20',
or X'30', the system abnormally ends the program currently
running.
For reason code X'04', the system detects and repairs,
where possible, errors in critical system areas. The system refreshes
selected system control blocks and validates and repairs selected
system resources. The system logs the old values of some refreshed
control block fields in the variable recording area (VRA) of the system
diagnostic work area (SDWA). The system writes these old values in
the logrec data set error record. Then the system returns control
to the interrupted program.
For reason code X'10', the system records the excessive
spin loop in a logrec data set error record.

Operator response
If requested by the system programmer, set
a SLIP trap to obtain a dump.

System programmer response
Depending on the reason code, do the
following: Reason Code X'00'
Message BLW004A should have identified the work that was in
progress at the time of the interruption:If a logrec data set error record is not produced with
the abend, ask the operator to obtain an SVC dump.
Examine the program status word (PSW) at the time of
error in the logrec data set error record. The PSW points to the location
where RESTART was entered by the operator. If the program was in a
loop, the PSW address will be within the loop.
Examine the system trace table for evidence of a loop
or a wait. If a loop or a wait is suspected, use the address in the
PSW to obtain the name of the looping module. If the module is not an IBM® module,
continue diagnosis with the module.
If the module is an IBM module,
search problem reporting data bases for a fix for the problem. If
no fix exists, contact the IBM Support
Center. Provide the dump and the logrec data set error record.

Reason code X'08'
Determine which program branched incorrectly, as follows: If a logrec data set error record is not produced with
the abend, ask the operator to obtain an SVC dump.
Look at register 14 at the time of the error. If the illegal branch
was made by a BALR 14,15 instruction, register 14 points to the instruction
following the instruction that took the bad branch. If this instruction
is a branch instruction, then this instruction probably branched directly
into the restart interrupt handler.
Use the address in register 14 to get the name of the
module that issued the bad branch instruction. If the module is not an IBM module,
continue diagnosis with the module.
If the module is an IBM module,
search problem reporting data bases for a fix for the problem. If
no fix exists, contact the IBM Support
Center. Provide the dump and the logrec data set error record.

Reason code X'0C', X'10', X'20', or X'30'
Determine why an excessive spin condition occurred, as follows: Get the program status word (PSW) at the time of error
as follows: For reason code X'10', X'20', or X'30',
obtain the PSW from the logrec data set error record.
For reason code X'0C', X'20', or X'30',
ask the operator to obtain an SVC dump. Format the dump to see the
PSW at the time of error.

Use the address in the PSW to obtain the name of the
spinning module. If the module is not an IBM module,
continue diagnosis with the module.
If the module is an IBM module,
search problem reporting data bases for a fix for the problem. If
no fix exists, contact the IBM Support
Center. Provide the dump and the logrec data set error record.

See z/OS Problem Management for
information about diagnosing a loop.

Source
Loadwait/Restart