Re: Running area detector simulator IOC on OS X aarch64 (M2) , segmentation fault

[Heinz Junkes via Tech-talk <tech-talk at aps.anl.gov>]

Hi Mark,

You are right, of course, and I found the place in the program code tonight as well:


The error is generated by NDArray.cpp:

"NDArray.cpp"  line 224

/** Calls NDArrayPool::release() for this object; decreases the reference count for this array. */
int NDArray::release()
{
  const char *functionName = "NDArray::release";

  if (!pNDArrayPool) {
    printf("%s: WARNING, no owner\n", functionName);
    return(ND_ERROR);
  }
  return(pNDArrayPool->release(this));
}

Unfortunately I don't understand why pNDArrayPool doesn't exist?

NDArray::NDArray()
  : referenceCount(0), pNDArrayPool(0), pDriver(0),
    uniqueId(0), timeStamp(0.0), ndims(0), dataType(NDInt8),
    dataSize(0),  pData(0)
{

Your reference to memory corruption is of course good. I run the software on a M2 (aarch64). Possibly this leads to the problem.

My first "quick shot" to mutexes was guided by the error message "no owner”.  I have had problems in the past with mutexes
 that were released from from a different thread than the one that had taken the mutex.

"Mutexes have ownership, unlike semaphores. Only the thread that locked a mutex, (that is, the owner of the mutex), should unlock it.”
e
Danke Heinz

> On 27. Sep 2023, at 23:35, Mark Rivers <rivers at cars.uchicago.edu> wrote:
>
> Hi Heinz,
>
>> NDArray::release: WARNING, no owner
>
> That message means that NDArray.pNDArrayPool is 0.  That could be a memory corruption problem.  What leads you to think that it might be a mutex problem?
>
> Mark
>
>
> -----Original Message-----
> From: Tech-talk <tech-talk-bounces at aps.anl.gov> On Behalf Of Heinz Junkes via Tech-talk
> Sent: Wednesday, September 27, 2023 8:30 AM
> To: Heinz Junkes via Tech-talk <tech-talk at aps.anl.gov>
> Subject: Running area detector simulator IOC on OS X aarch64 (M2) , segmentation fault
>
> Hello,
> I am trying to test on a new MAC with M2 processor areadetector.
> I was able to build it and start a simulator IOC. But when I start the simulator camera I get a segmentation fault. I haven't searched further for the error and wanted to ask here first if someone has alrerady  seen this (and maybe solved it).
>
> epics> dbpf 13SIM1:cam1:Acquire 1
> NDArray::release: WARNING, no owner
> DBF_STRING:         "Acquire"
> ./start_epics: line 2: 31835 Segmentation fault: 11  ../../bin/darwin-aarch64/simDetectorApp st.cmd.darwin
>
>
> I guess it has something to do with Apple's lack of (limited) support for POSIX mutexes?
>
> Danke Heinz
>

Attachment: smime.p7s
Description: S/MIME cryptographic signature