Google
 

Trailing-Edge - PDP-10 Archives - bb-bt99g-bb - ietv04.d09
There are 2 other files named ietv04.d09 in the archive. Click here to see a list.
                 EDIT DESCRIPTIONS FOR IBMCOM-10-ET-V4                          
  
  
                             EDIT 010    FOR D60SPD
  
No SDC file
********************************************************************************
  
  
                             EDIT 011    FOR DN60
  
  
  
  
[s]
DN20 reportedly crashes with large part of memory destroyed.
  
  
  
[d]
A test for an abort  condition  in  the  translate  task  is
incorrectly  done  indexing off of R0 containing the address
of the line  control  block  instead  of  off  of  R5  which
contains  the address of the appropriate task control block.
As a result an abort condition can  be  detected  when  none
exists   thus   causing  the  translate  task  to  grab  for
unavailable chunks and fail.  The code which  processed  the
failure  path, returns a 0 in R0 indicating no chunks on the
free  list.   Since  the  translate  task  is  guaranteed  a
successful  return  from  the  GETSTG  routine,  it  may not
protect itself from having to continue with contents  of  R0
detroyed.
  
  
  
[c]
Fix the code in GETSTG to do the correct test for the  abort
condition in a translate task.
********************************************************************************
  
  
                             EDIT 362    FOR IBMSPL
  
[SYMPTOM]
IBMSPL dies with an address space exhausted error.
  
  
[DIAGNOSIS]
The call to INIPAG in routine LPTJOB is not set up correctly and as a
result, accumulators S and J are not maintained.
  
  
[CURE]
Add edit 362.
  
********************************************************************************
  
  
                             EDIT 363    FOR IBMSPL
  
[SYMPTOM]
If an Emulation link is SHUTDOWN while receiving data, and is then reSTARTED
before the transmission completes, IBMSPL loops eating up JFNs.
  
  
[DIAGNOSIS]
After processing the SHUTDOWN command, QUASAR changes its database to
indicate that the node is shutdown.  However, if the link was receiving
data when the SHUTDOWN was given, the data transmission must complete before
IBMSPL does its shutdown processing.  Therefore, it is possible to issue
a START command to restart the node when in fact IBMSPL is not ready to
accept it.  When this occurs, IBMSPL forces the line down and then goes
into a loop trying to dispose of the hold file that it still has open.
Another JFN is acquired, the new open on the file fails, and IBMSPL tries
again.
  
[CURE]
When IBMSPL receives a START message for a node, have it check to see if
that node is still active and in the process of shutting down.  If this
is the case, the START should be rejected so that shutdown will complete
and normal cleanup activities will be performed.
  
********************************************************************************
  
  
                             EDIT 364    FOR IBMSPL
  
No SDC file yet
  
  
********************************************************************************
  
  
  
END OF  IBMCOM-10-ET-V4