> Hello: I saw this in a z/os 1.13 dump: > > 00 01B2 008BA3E8 SVC 9 00000000_0001A314 00000000 00040520 > 000400C0 Delete 22:25:22.551801 32 > 07741000 80000000 > 01 01B2 008BA1C8 SVC 78 00000000_24F4135E 0000E601 00001608 > 008789F8 Freemain 22:25:22.551810 28 > 07841000 80000000 > 00 01B2 008BA3E8 SSRV 78 8152C6B2 0000FF03 0000000C > 008AB050 Freemain 22:25:22.551810 32 > 01B20000 > 00 01B2 008BA3E8 SVCR 9 00000000_0001A314 00000000 00040520 > 000400C0 22:25:22.551812 32 > 07741000 80000000 > 00 01B2 008BA3E8 SVC 9 00000000_0001A314 00000000 00040530 > 000400C0 Delete 22:25:22.551814 32 > 07741000 80000000 > 01 01B2 008BA1C8 DSP 00000000_24F4135C 00000000 00001608 > 008789F8 00000001 00000000 01B2 01B2 22:25:22.551815 28 > 07841000 80000000 > 00 01B2 008BA3E8 SUSP 8174CF08 00708A30 LOCL > 00000000 00000000 00000000 22:25:22.551815 32 > 00000000 > 00000000 > 01 01B2 008BA1C8 SVC 78 00000000_24F4135E 0000E601 00001608 > 008789F8 Freemain 22:25:22.551816 28 > 07841000 80000000 > 01 01B2 008BA1C8 SVCR 78 00000000_24F4135E 00000000 00001608 > 008789F8 22:25:22.551831 28 > 07841000 80000000 > > It looks like the first SVC 78 at 24F4135E never completed and was > later retried. I have seen intervening trace entries from other > processors between the start and finish of a SVC 78, but have never > seen the SVC 78 repeated like that. It gives the illusion that > there was a double freemain, but I would think if that was the > case the second SVC 78 would have encountered some type of x78 abend. > > I am just curious if anyone else has seen a SVC 78 trace entry like above.
When the SVC 78 was issued, the local lock was not available (it was held by the Freemain being done by TCB 8BA3E8 on CPU 0). When the local lock is not available for a type 1 SVC, the SVC FLIH backs up the PSW address in the RB/XSB so that the SVC will be reissued when the local lock is available. Jim Mulder z/OS System Test IBM Corp. Poughkeepsie, NY ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
