Experimental Physics and
Industrial Control System

Benjamin Franksen via Tech-talk <[email protected]> · Thu, 11 Apr 2019 14:30:36 +0200

Am 10.04.19 um 15:59 schrieb Konrad, Martin:
> I'm experiencing a crash when using the sequencer with
> 
> pvPut(foo, SYNC)
> 
> while it has no access to PV "foo" due to Access Security constraints
> (see backtrace.txt).
> 
> I can reproduce this with the latest development version of the
> sequencer. Replace all pvPuts in the demo example and insert
> 
> asSetFilename("does-not-exist.acf")
> 
> at the beginning of demo.stcmd to reproduce the problem.

I can reproduce this, using the steps you indicated. Thanks for the bug
report!

It looks to me as if the bug is in base, however, not the sequencer.
This is the backtrace I get from the core dump:

(gdb) bt
#0  0x00007f054d1ecc4f in dbProcessNotify (ppn=0x7f050c057d68) at
../../../src/ioc/db/dbNotify.c:351
#1  0x00007f054d200ed4 in dbPutNotifyBlocker::initiatePutNotify
(this=0x7f050c057d50, guard=..., notify=...,
dbch=dbch@entry=0x7f050c013298, type=type@entry=5, count=count@entry=1,
pValue=0x55e5bdff4f20) at ../../../src/ioc/db/dbPutNotifyBlocker.cpp:183
#2  0x00007f054d1feb41 in dbContext::initiatePutNotify
(this=0x7f050c000e30, guard=..., chan=..., dbch=0x7f050c013298,
type=type@entry=5, count=count@entry=1, pValue=<optimized out>,
notifyIn=..., pId=0x0) at ../../../src/ioc/db/dbContext.cpp:257
#3  0x00007f054d1ffdfd in dbChannelIO::write (this=0x7f050c013370,
guard=..., type=5, count=1, pValue=0x55e5bdff4f20, notify=..., pId=0x0)
at ../../../src/ioc/db/dbChannelIO.cpp:107
#4  0x00007f054c38b218 in ca_array_put_callback (type=5,
count=count@entry=1, pChan=0x7f050c000f10,
pValue=pValue@entry=0x55e5bdff4f20, pfunc=pfunc@entry=0x7f054b325010
<pvCaPutHandler>, usrarg=usrarg@entry=0x7f050c04fc90) at
../oldChannelNotify.cpp:432
#5  0x00007f054b32599b in pvVarPutCallback
(var=var@entry=0x55e5bdff5978, type=<optimized out>,
count=count@entry=1, value=value@entry=0x55e5bdff4f20,
arg=arg@entry=0x7f050c04fc90) at ../pv.c:122
#6  0x00007f054cd3e073 in seq_pvPutTmo (ss=0x55e5bdff5190, chId=0,
compType=SYNC, tmo=10) at ../seq_if.c:515
#7  0x00007f054cd3a8fd in ss_entry (arg=0x55e5bdff5190) at ../seq_task.c:361
#8  0x00007f054cd3ae11 in sequencer (arg=0x55e5bdff4d00) at
../seq_task.c:101
#9  0x00007f054cf8add8 in start_routine (arg=0x55e5bdff5f50) at
../../../src/libCom/osi/os/posix/osdThread.c:403
#10 0x00007f054c1256db in start_thread (arg=0x7f05491ec700) at
pthread_create.c:463
#11 0x00007f054ca6688f in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:95

Since this is a technical issue deep down in the EPICS core I have
cross-posted this to @core-talk.

Cheers
Ben

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System