Assertion failure in dbLock.c

[Andrew Johnson <anj@aps.anl.gov>]

> tux% pwd
> /local/anj/base-3.15
> tux% bin/linux-x86_64/softIoc -x anj
> Starting iocInit
> ############################################################################
> ## EPICS R3.15.0.1-DEV $$Date$$
> ## EPICS Base built Oct  6 2014
> ############################################################################
> iocRun: All initialization complete
> epics> dbpf anj:exit 1
> 
> 
> 
> A call to 'assert(pset->state == lockSetStateFree)'
>     by thread '_main_' failed in ../../../src/ioc/db/dbLock.c line 406.
> Dumping a stack trace of thread '_main_':
> [    0x7f6f12187af3]: /local/anj/base-3.15/lib/linux-x86_64/libCom.so.3.15(epicsStackTrace+0x73)
> [    0x7f6f121815ed]: /local/anj/base-3.15/lib/linux-x86_64/libCom.so.3.15(epicsAssert+0x4d)
> [    0x7f6f126166df]: /local/anj/base-3.15/lib/linux-x86_64/libdbCore.so.3.15(dbLockCleanupRecords+0x7f)
> [    0x7f6f12649b5a]: /local/anj/base-3.15/lib/linux-x86_64/libdbCore.so.3.15(iocShutdown+0xea)
> [    0x7f6f121791af]: /local/anj/base-3.15/lib/linux-x86_64/libCom.so.3.15(epicsExitCallAtExits+0x8f)
> [    0x7f6f12179268]: /local/anj/base-3.15/lib/linux-x86_64/libCom.so.3.15(epicsExit+0x8)
> [    0x7f6f128a32e5]: /local/anj/base-3.15/lib/linux-x86_64/libdbRecStd.so.3.15(process+0x205)
> [    0x7f6f12617ea1]: /local/anj/base-3.15/lib/linux-x86_64/libdbCore.so.3.15(dbProcess+0x211)
> [    0x7f6f12618bdf]: /local/anj/base-3.15/lib/linux-x86_64/libdbCore.so.3.15(dbPutField+0x4bf)
> [    0x7f6f1262c072]: /local/anj/base-3.15/lib/linux-x86_64/libdbCore.so.3.15(dbpf+0xb2)
> [    0x7f6f12174370]: /local/anj/base-3.15/lib/linux-x86_64/libCom.so.3.15(iocshBody+0xae0)
> [          0x4047c7]: bin/linux-x86_64/softIoc(main+0x1e7)
> [      0x3832c1ed1d]: /lib64/libc.so.6(__libc_start_main+0xfd)
> [          0x403c19]: bin/linux-x86_64/softIoc(_start+0x29)
> EPICS Release EPICS R3.15.0.1-DEV $$Date$$.
> Local time is 2014-10-06 12:10:11.169288800 CDT
> Please E-mail this message to the author or to tech-talk@aps.anl.gov
> Calling epicsThreadSuspendSelf()
> ^C

The IOC database should be able to shut down an IOC, preferably by just
calling exit() which is what the softIoc's exit subroutine record is
doing above. In 3.14 this worked fine, but now it causes the process to
suspend with the above message.

Fix is not urgent, but should be made before 3.15.1.

- Andrew