of dedicated resident foreground pages will be used. In any case, the pages start at physical address 64K.
After PASSO has executed, GHOSTl requests the current date and time key-ins with the following messages:
DA TE (MM/DD/YY)
=
TIME (HH:MM)
=
The operator must respond by typing the appropriate quan-tity in the indicated format, terminating the response as usual with
8, 9,
ore.If the system is being loaded on a machine for which it was not SYSGENed, one of the following messages will be dis-played on the OC device and the bootstrap operation will be terminated.
SYSTEM NOT SYSGENED FOR SIGMA 6
SYSTEM NOT SYSGENED FOR SIGMA 9
SYSTEM NOT SYSGENED FOR XEROX 560
If the system and target machines match and if I was speci-fied the following message is displayed:
C/LL/DC ASSIGN OK (YES/NO)
If the operator's response is YES or
8,
it is assumed that the device addresses for the control device, listing log, and system device are not to be changed from thoseestab-I ished when the monitor was defined. If the response is NO, then the following messages will be output to rede-fi ne these de vi ce addresses.
CRndd ~ CR
LPndd ~ LP
DCndd ~ DC
where each ndd is the current device identification and as many DC messages are output as there are swap devices.
36 Sequence of Operations
In response to each of these messages the operator must type two or three characters. If two characters are typed, they must be 'SA' and indicate no change for this device. If three characters are typed, they must be the channel and device designation codes (ndd) defining the address of the indicated device (see Appendix B, Tables B-2 and B-3).
If the DC or swapper assignment is incorrect, one of the following five messages will be displayed. Three of the messages request a new swapper device address.
NO RESPONSE FOR SWAPPER yyndd DCndd ~ DC
(The device address is unrecognizable by the hardware.)
DISC PACK BAD yyndd DCndd ~ DC
(The disk pack spindle is not initialized or cannot be ac-cessed due to I/O transmission errors.)
SWAPPER NOT 7212 DCndd => DC 7232 72DP
(The RAD indicated by 7212 or 7232 or disk pack indicated by 72DP was expected as the swap device.)
SWAPPER WRITE PROTECTED yyndd
(The swapper device is write protected. If a write protect violation is encountered during the reading and patching of the monitor overlays, the above message is output and the system stops. However, if this occurs during swapper initialization, then a new swapper device is requested as for the above messages. )
PSA TRACK FLAWED
(The swapper disk pack contains flaws. The boot process term i no tes. )
Before completing any of the above responses with a 90r
8,
the operator may cancel the response by striking the e key. Following this, or if a completed response is in error, the message??
wi II be output and the key-in request will be repeated.
If no characters are typed within 10 seconds, a 6) response is assumed.
After a II necessary responses have been received, the boot subroutine reads the system information record from tape and writes it on the LL and OC devices if Pond I are speci-fied, respectively.
The following sense switch information is then I isted on the OC device if I was specified.
SET SENSE SWITCHES AND TYPE N/L SSW1 =>CHECKWRITE DISK WRITES
SSW2 =>NO AUTOMATIC LOGON/LOGOFF SSW3 =>OPERA TOR RECOVERY ON DISK BOOT SSW4 =>SYSTEM SECURITY CHECKING
The system 'will continue when a NEW LINE or any other character is entered.
Next, the reconfiguration and partitioning commands (if any exist) are read and processed. A summary of the system's device will be output on the LL device (even if no :TYPE commands are encountered). Permanently down devices are not listed. the reconfiguration and partitioning commands in the patch deck. )
After the nonroot patches have begun, reconfiguration and partitioning commands are illegal. If any such commands appear in the deck, the following message is displayed on the OC device (and also on the LL device if P was speci-fied) and the bootstrap continues.
':' COMMAND NOT IN PATCH DECK PROPERLY
This message is displayed only one time, even if additional reconfiguration and partitioning commands are encountered.
It then copies the overlays, etc., to the swapping device, communicati ng the sizes and disk addresses to the resident root of the absolute monitor. Control then passes to another boot subroutine at WRTROOT. This second boot subroutine causes the monitor root to be copied to the disk, preceded by a disk bootstrap. At this point, the resident monitor is operational but the system has not yet been establ ished on the resident swapping device. The GHOSTl processor performs this function.
If P was specified, GHOSTl determines whether any devices or controllers are partitioned. If none are parti-tioned, the following message is displayed on the LL device:
*** NOTHING PARTITIONED
However, if devices and/or controllers are partitioned, the following message is displayed on the LL device:
***** ITEMS PARTITIONED *****
followed by messages identifying each device or controller which is partitioned. The messages have the followi ng formats:
DEV yyndd PARTITIONED
(for devices)
~
_________ C_O_N_T_y_y_nd_d __ PA_R_T_IT_IO __ N_E_D ________~1 ,
(for controllers)
When all partitioned items have been identified, the following message concludes the list:
** END OF PARTITIONED ITEMS **
When P was not specified or when GHOSTl has completed the above I isting, GHOST 1 starts the symbiont ghost, Fix ghost, ERR:FIL ghost, and fill ghost, and then exits.
BOOTING FROM DISK
Once the operating system has been bootstrapped from tape, it may thereafter be brought into core from the disk by
to determine whether the system debugger's memory should be released.
Partitioning information isdisplayedasdescribed previously, and the system ghost jobs {Fill, ERR:FIL, and Fix} are started. Normal operation may then be resumed.
BOOTSTRAP 110 ERROR RECOVERY
I/O error recovery during bootstrap is provided for the card reader, line printer, magnetic tape, and disk. However, error recovery is not possible until the tape boot and mon-itor root have been read from tape. The following error messages may appeor on the OC device:
xx INOPERA nVE
xx ERROR. no value TDV value
xx MA NUAL MODE
CHECKWRITE ERROR
where
xx is MT, CR, DC, or LP.
value indicates the TIO or TDV results.
When either of the first two messages above occurs, the wait state is entered. To continue, the operator must
Table 7.
place the CPU into IDLE, STEP, and then RUN state. The I/O will then be retried. If the third message above oc-curs, I/O will continue when the condition is corrected.
When an error occurs for a magnetic tape or disk opera-tion, the operation is retried ten times before an error message is output. If the fourth message above occurs, the wait state is entered. To continue, the operator must place the CPU into IDLE, STEP, and then RUN state.
This message will appear if the checkwrite on disk fails.
The checkwrite will be executed only if hardware sense switch 1 is set.