Information regarding hyper fix SO15901 for Common Services r15.0
search cancel

Information regarding hyper fix SO15901 for Common Services r15.0


Article ID: 207079


Updated On:




Questions have arisen requesting details on SO15901, written for Common Services for z/OS r15.0 to prevent a potential CAIRIM overlay.  Although categorized as a hyper, customers are not sure if it applies to their environment and if they should apply it.

Fix defined as a HYPER due to the potential to overlay storage beyond the originally allocated storage. Storage is allocated once during the initial execution of CAS9 (CAIRIM).

• Impact of the overlay is the unknown 
• Times when no adverse symptoms witnessed 
• Other times when the system was adversely impacted
• As with all overlays, results are unpredictable and difficult to identify


Common Services for z/OS running on a z/OS image with more than 64 (x'40') processors  

Common Services for z/OS r15.0
Component :


Client was configuring on and off large numbers of CPU's (much larger than 64). The problem arises from the counts in the following fields returned from the CSCRI function.

1      SI22V2TotalLCPUCount   Current total CPUs
2      SI22V2ConfiguredLCPUCount Configured CPUs
3      SI22V2StandbyLCPUCount Standby CPUs
4      SI22V2ReservedLCPUCount Reserved CPUs
5      SI22V2DedicatedLCPUCount Shared CPUs
6      SI22V2SharedLCPUCount  Shared CPUs  

  • CAIRIM only structured to handle a value of 64 in any one of those fields
  • Storage is an array with a section for each of the above
  • Value above 64 in any one of sections 1 - 4 would overlay CAIRIM's storage  
    • Of no real consequence given the purpose of this storage
  • Value above 64 in one of sections 5 or 6 could potentially overlay the adjacent storage causing unpredictable results
    • If customer has never experienced this problem, without major changes to hardware configuration (ahead of time or dynamically), it is unlikely to occur.   
    • Note that these are counts, not necessarily CPU numbers.
    • If  large numbers of CPU's, greater than 64, are configured, there is a potential for a problem
  • How the CPU information is used by CAIRIM not easily explained.
    • Code design has existed for many years (>16 yrs)
    • Problem just recently reported by a single customer
    • CPU information not really used for any functionality within CCS or any CA MF product





Apply SO15901

Fix written in 2 parts

  • Part 1 - Expand GETMAIN performed on initial startup of CAIRIM
    • Necessitates the IPL 
  • Part 2- Bypass the code related to the CPU data collection
    • No IPL necessary
    • Bypass can be accomplished by re-running CAIRIM using
      initialization statement in CAIRIMPARM
    • Prevents the overlay from occurring until an IPL is performed with the fix applied  


Additional Information

Please refer to the Common Services for z/OS r15.0 for details on the CAIRIM(REFRESH(LMP)) Initialization Statement