Table of Contents
Advertisement
however, automatic cylinder incrementing does not occur.
A seek operation must be issued to cross cylinder boundaries
to prevent an "unusual end" condition.
When the controller detects a flaw mark during a surface
operation, automatic surface address incrementing is in-
hibited to allow for an immediate Header Read operation
to obtain the alternate address.
TIO AND TDV INSTRUCTIONS
Frequent TIO and TDV instructions during surface opera-
tions may cause data overrun conditions.
SEEK AND RESTORE ORDERS
The Seek operation may be performed only when the disk is
on cylinder. A modifier bit in the Seek order causes an on-
cylinder interrupt to occur when the on-cylinder condition
has been reached. The interrupt should be used to max-
imize controller availability.
The on-cylinder interrupt has a window that is one sector
ahead of the addressed sector. Thus, it is possible for the
condition code bits, CCI and CC2, of an AIO instruction
to be set to a 11 binary configuration (indicating that no
interrupt was recognized). This "no interrupt recognized"
condition should not be considered as a spurious interrupt.
When the seek modifier is used during Seek operations that
are command chained to surface, header, or Sense opera-
tions, the on-cylinder interrupt does not occur.
When the Restore operation is performed, the on-cylinder
indication must be present before a following Seek opera-
tion can be performed. In order to command chain a
Restore and Seek order, a Header Read order should be
used between the Restore and the Seek orders.
1/0 OPERATION EVALUATION
By evaluating the status response and condition code bits
obtained by executing AIO and TDV instructions after an
1/0
interrupt, a decision may be made as to whether the
1/0
operation was completed successfully or terminated
due to a fault condition.
For Sigma 3 Computer Systems Only. A disk
1/0
opera-
tion may be considered successful if Overflow, Carry, and
all the status response bits are as follows:
1.
After an AIO instruction, the Overflow and Carry
indicators are both 0.
2.
After
a
TDV instruction, the Overflow and Carry
indicators are both 0.
3.
The Error flag in the odd
1/0
channel register
(assigned to disk) is 0.
I
4.
Bits 0, 1, and
4
(tjperational Status Byte) in the even
1/0
channel
regist~r
(assigned to disk) are all 0.
I
!
For Sigma 5-9
Compute~
Systems Onl • A disk
1/0
opera-
tion may be considered uccessful if all the status response
and condition code bits !are as fol lows:
!
1.
After an AIO instruction, condition code bits 1 and 2
are both 0.
I
2.
After an AIO instr~ction, Operational Status Byte
bits 8, 9, and 12 are all 0. Note that bit 8 may be
a 1 if the SIL (Supwess Incorrect Length) flag is set.
3.
4.
After a TDV instrudtion, condition code bits 1 and 2
are both 0.
i
I
After a TDV
instru~tion,
8-15areall 0.
'
I
Operational Status Byte bits
In either system (Sigma
3
or Sigma 5-9), if one or more of
the specified bits are nt 0, a TIO instruction must be
issued to obtain the stat of the device. Also, a Sense
order must be executed , o obtain 10 additional bytes of
status and fault informaflion.
I
It is recommended that Clln error log be maintained in which
appropriate status and ft. ult information (i.e., bytes 8 and
9 of Sense order, etc.) e recorded that will aid in hard-
ware maintenance and r pair.
I
I
I
1/0 FAULT
EVALUAll~N
AND RECOVERY PROCEDURES
Status information that 1s obtained by executing an AIO,
TDV, and TIO instructifn after an
1/0
interrupt is eval-
uated in a prescribed m¢mner as listed in Table 9. The first
column provides a summbry description of the fault when
the specified bits are n1t 0 (indicating generally that a
fault condition has
occ~rred).
Column 2 shows a specific
bit configuration only if two or more bits are used to dif-
ferentiate the fault indlcation and detailed description of
the fault condition.
T~us,
the first two columns of Table 9
permit a fault conditiori to be defined as a function of the
status response and
con~ition
code bits.
The Recommended
Corr~ctive
Action (column 9) for a
specific fault depends on the failing order (columns 3-8).
Generally, for a failing Seek order, a reference is made
to Recovery Procedure
$,
Table 10; for fai I ing Restore,
Release, and Sense ord$rs, a reference is made to Recovery
Procedure 2; and for fajling Read, Write, Check-Write, or
Header Read/VVrite ordfrs, a reference is made to Recovery
Procedure
4.
i
I
If
the device is not opelrational, program recovery is not
possible but operator in1tervention may be required (see
Chapter
4,
"Operator
~anel
Controls and Indicators"). If
the FAULT light on theloperator control panel is on, press
the FAULT switch to attempt recovery.
If
the light goes
off, repeat the previou' operation.
If
the I ight remains on,
call the customer servi4e engineer.
1/0
Feliu Its and Recovery Procedures
25
I
Advertisement
Table of Contents
