On Wed, 2 Jul 2025 10:13:17 -0500, Thomas David Rivers <[email protected]>
wrote:
>So I can say "set the linkage-stack to level N" and then branch
>(where N is less than the current level.)
You need to understand abend recovery on the stack. Consider your example:
ESTAE recovery
WTO 'performed abend recovery'
BAKR (1st)
WTO 'abend ignored instead of percolated in 1st BAKR'
BAKR (2nd)
WTO 'abend ignored instead of percolated in 2nd BAKR'
BAKR (3rd)
DC '00DEAD00' causing an abend
Ooops - some error happened.
Your ESTAE is outside the 3 BAKRs and I don't believe your ESTAE is active
during those 3 BAKRs. I expect stack abend recovery to issue a PR for the 3rd
BAKR and resume control after the 2nd BAKR. Since I don't know BAKR, it may
behave differently and percolate the abend all the way to your ESTAE.
You can run the test above to verify BAKR behaviour. If you get to your ESTAE
and the SDWA if for 00DEAD00 then the ESTAE then does not work like other stack
entries. I expect you will get the WTO from the 2nd BAKR, no SDWA and probably
no indication that an abend occurred.
Run the test above to see BAKR's stack entry abend recovery. Your recovery
strategy will depend upon the abend recovery. You most likely won't have access
to the SDWA that you expect unless you implement recovery within each BAKR.
Alternatively, you could have flags indicating when each BAKR successfully
completes by resetting the flag before the PR.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
