Re: StreamDevice device locking

[Dirk Zimoch via Tech-talk <tech-talk@aps.anl.gov>]

The question is: Is there a way to tell *asyn* to block only on the 
address level?
This may depend on the implementation of the particular asyn interface. 
I do not have any devices with addresses at the moment. (Only GPIB comes 
to my mind.) So I cannot test.
StreamDevice calls pasynManager->blockProcessCallback() with its 
asynUser handle. The asynUser handle is bound to an address on a port. 
So I had expected that asyn blocks only that address, not others on the 
same port.
I have to set up a GPIB system with at least 2 devices first before I 
can test this case.
Dirk


On 08.02.19 14:24, Joao Afonso via Tech-talk wrote:
> Hello,
>
> At the moment I am experimenting with StreamDevice, while connecting it 
> to a custom asynPortDriver (instantiated with ASYN_MULTIDEVICE, 
> ASYN_CANBLOCK).
> According to the documentation, each protocol command blocks the device 
> for the duration of the protocol, which makes sense.
> But an access to two devices on two different addresses should be possible:
> http://epics.web.psi.ch/software/streamdevice/doc/processing.html
>
>   * /"The first |out| command in the protocol locks the device for
>     exclusive access. That means that no other record can communicate
>     with that device. This ensures that replies given by the device
>     reach the record which has sent the request. _On a bus with many
>     devices on different addresses, this normally locks only one
>     device_. The device is unlocked when the protocol terminates.
>     Another record trying to lock the same device has to wait and might
>     get a |LockTimeout|." /
>
> However, I noticed that if I process two records with stream device, for 
> addresses #1 and #2, it first runs all callbacks for address #1, and 
> only after to #2.
> This is contrary to my expectation, which would be to see callbacks to 
> both devices being executed interleaved.
> Is there a way to tell StreamDevice to block only on the address level, 
> allowing to different addresses to be processes simultaneously?
> Thanks,
> Joao Afonso
> CERN TE-EPC-CCS