Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Wed, 31 Jan 2018 17:41:04 -0600

Hi Simon,

On 01/31/2018 04:54 PM, Simon Reiter wrote:
> Ok, I am pretty sure, that these test will only run for my host
> system. I guess, I need to copy the files somehow to the powerpc. And
> get perl running. I will investigate in this tomorrow.

Ah, you're cross-compiling. Don't worry about trying to get Perl to run
on the PowerPC then, but there are test programs in the following three
directories which check various aspects of the software and the OS that
might be helpful in finding problems:

> src/libCom/test
> src/ioc/db/test
> src/std/rec/test

These programs can be run individually by hand and the output is
human-readable, e.g.:

> tux% cd src/ioc/db/test/O.linux-x86_64
> tux% ./dbStateTest 
> 1..20
> ok  1 - Finding nonexisting state fails
> ok  2 - Create state 'red'
> ok  3 - Find 'red' returns correct id
> ok  4 - Create existing 'red' returns correct id
> ok  5 - Finding nonexisting state still fails
> ok  6 - Create state 'blue'
> ok  7 - Find 'blue' returns correct id
> ok  8 - Create existing 'blue' returns correct id
> ok  9 - Finding nonexisting state still fails
> ok 10 - Default 'red' state is 0
> ok 11 - Default 'blue' state is 0
> ok 12 - After setting, 'red' state is 1
> ok 13 - 'blue' state is 0
> ok 14 - After setting, 'blue' state is 1
> ok 15 - 'red' state is 1
> ok 16 - After clearing, 'blue' state is 0
> ok 17 - 'red' state is 1
> ok 18 - After clearing, 'red' state is 0
> ok 19 - 'red' state is 0
> ok 20 - Finding nonexisting state still fails
> 
>     Results
>     =======
>        Tests: 20 
>       Passed:  20 = 100.00%

Some of the later test programs will look for files from their
individual O.<arch> directory, their parent directory, and possibly even
from the O.Common directory as well so if possible try to make the whole
build tree accessible on the target (mounting it using NFS or SMB might
be simpler than trying to work out what files to copy). Each test
program displays a summary at the end telling you how many tests, passes
& failures were run.

If doing the above seems like too much work, I want to see what happens
if you replace the internals of your SET_PV macro with a simple printf()
of the state values, because currently you are monitoring the state
machine at the end of a long pipeline:

  SNL CA client → RSRV → dbAccess → record → dbEvent → RSRV → CA client

Calling printf() inside the SNL program would bypass that pipeline and
should prove whether the Sequencer is actually skipping states. Once we
know the answer to that we can decide if we need to investigate further
to explain the results you are seeing by monitoring it over CA.

HTH,

- Andrew

-- 
Arguing for surveillance because you have nothing to hide is no
different than making the claim, "I don't care about freedom of
speech because I have nothing to say." -- Edward Snowdon

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System